JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY ALFRED REHDER EDITOR JOSEPH H. FAULL ann CLARENCE E. KOBUSKI ASSOCIATE EDITORS VOLUME XVII JAMAICA PLAIN, MASS. 1936 Reprinted with the permission of the Arnold Arboretum of Harvard University KRAUS REPRINT CORPORATION New York 1968 DATES OF ISSUE (pp. 1-51, pl. 166-178) issued January 24, 1936, (pp. 53-114, pl. 179-181) issued April 16, 1936. (pp. 115-240, pl. 182-194, map) issued July 30, 1936. (pp. 241-389, pl. 195-200) issued October 20, 1936. Printed in U.S.A. TABLE OF CONTENTS STUDIES ON THE BIOLOGY OF GYMNOSPORANGIUM GLOBOSUM FAL. ‘Vith plates 166-175 and two text figures. By J.D. MacLachlan... 1 Tue Lire Histories or MILEs1A ScoLopeNpRII, M. Potypopi, M. VOGESIACA AND M. KRIEGERIANA. With plate 176. By Lillian M. A CONVENIENT SAND-CULTURE ee With plate 177 and one text tipure. (By Mover! WW Ord eee oases odoin eas See eee CHROMOSOME COMPLEX AT PREMEIOTIC ANAPHASE AND MEIOTIC METAPHASE. With two text figures. By Haig Dermen ........ BURRETIODENDRON, A NEW GENUS OF TILIACEAE. With ous 178. BY Ee OER Ne 2055550 6 a oA SoU eae ee eae Two NEw SPECIES OF GYMNOSPORANGIUM FROM AsIA. With two RONG UUM Se OOM UPON 6 oss bsead eae dae eaneee caer Notes ON THE LIGN PLANTS DESCRIBED BY LEVEILLE FROM EASTERN Asta. By “Alfred TCROCE «2 sisatecs Yip eae eee CHROMOSOME NUMBERS AND PHYLOGENY IN THE GYMNOSPERMS. With two text figures. By Walter S. Flory ........ccccceeeeeee APOSPORIC PARTHENOGENESIS IN A TRIPLOID since MALUS HUPE- HENSIS. With plates 179-181. By Haig Dermen .............. FERTILIZATION IN THE BALDWIN APPLE, A TRIPLOID VARIETY. By Ha ES PINON as 63 kh 4d ee nea SHR Oe 106 Two SpRUCE-INFECTING Rusts, CHRYSOMYXA PIPERIANA AND CHRYSOMYXA CHIOGENIS. By 7. POU a. Pee eae eee ae 109 MorPHOLOGY AND ONTOGENY OF THE SPERMOGONIA OF THE MELAMP- SORACEAE. With plates 182-188. By Lillian M. Hunter ........ 115 THE EXPERIMENTAL PRODUCTION OF PoLyPLompy. With plate 189. BE OR ats We ie oh aos oy + Wk e's Be GE STUDIES IN THE OAKS OF THE MOoUNTAINS OF NORTHEASTERN Mexico. By Cornelius H. Mueller .........ccccccccccseccceces 1 PHYTOGEOGRAPHIC STUDIES IN THE ATHABASKA—GREAT SLAVE LAKE Recion. With plates 190-194 and map. By Hugh M. Raup .... 180 PavroctocaaPsc STUDIES IN THE ATHABASKA—GREAT SLAVE LAKE REGION (contin.). With plates 195-200. By Hugh M. Raup .... 241 NoTES ON THE LiGNEOUS PLANTS DeEscRIBED BY LEVEILLE FROM EASTERN AsIA (contin.). By Alfred Rehder ......... 0.000005. 316 THE GeNuS CESTRUM IN GUATEMALA. By C. V. Morton ......... 341 A New SPECIES oF ACER FROM GUATEMALA. By Alfred Rehder .... 350 POLYPLoIDy AND GEOGRAPHIC DIsTRIBUTION IN SPIRAEA. By Karl Sax 352 Ai a A a eo Ae a a ee ee er ee Yer ee Ye Te a ee ee iv TABLE OF CONTENTS THe ARNOLD ARBORETUM DuRING THE FiscAL YEAR ENDED JUNE ct ie |X, ee re rer eee eer errr errs yor rer ery 57 BIBLIOGRAPHY OF THE PUBLISHED WRITINGS OF THE STAFF AND STUDENTS, JULY 1, 1935-JUNE 30, 1936 ....... cee cece eeceeeces 366 STAFF OF THE ARNOLD ARBORETUM, 1935-36 ....... 00sec eee ences 370 COMNORPINME ook bac ck pace ee cace P50 bd) wu RW e Rea ee areeTEes 371 > ee ee en ee ee ee eRe ee rr ree rr TT 373 JOURNAL OF THE ARNOLD ARBORETUM VotuME XVII. JANUARY, 1936 NuMBER | STUDIES ON THE BIOLOGY OF GYMNOSPORANGIUM GLOBOSUM FARL. Toa. MaAcLACHLAN With plates 166-175 and two text figures TABLE OF CONTENTS i TA TROUUCTION fe 64-24 6 0.05-05-556 8 E46 whe SRDS re es 1 II INGMENCEATOU RE op oiis Go ceed oi eee e teed e eee ene s 2 Ill RANGE AND EcoNnoMIc IMPORTANCE OF G. GLOBOSUM ........ 3 IV SYMPTOMATOLOGY OF THE DISEASES CAUSED BY G. GLOBOSUM.. 5 V Factors AFFECTING SPORE GERMINATION ......--eeeceeees 14 VI Tee Aree OF THE LELTAL SCORUS. .<sJ<0cns eae meee canes 16 Vil . INFECTION OF THE RED CEDAR 2.5 os cc bcss .citiew eee vet eee 18 APUEL PSUSMARY oo eee BNo 65k ea hs 5 45a SRR COO Oe 19 IX Pe) < (eh g ACL 6 1. 5 ¢ ee er err rere ark etree cera er era 20 x PLRIAOGE AT AY aoc se oe a Oe aes HE Ree awa 20 XI EXPLANATION OF LA DES crerettietcunts tite ceetaniechete sche aleus opeterere aye oe I. INTRODUCTION Stnce the pioneer work of Farlow (1880), who ascertained the alter- nate hosts of Gymnosporangium globosum Farl., many observations on various phases of the biology of this rust have been recorded; these observations, however, do not afford a complete survey of the biology of G. globosum, More than one hundred suscepts have been added, since 1880, to the host list, yet observations of the writer warrant the con- clusion that the number of suscepts, conservatively estimated, is more than six hundred; moreover, little information was previously available 2 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xv regarding their relative susceptibility. Many aspects of the progressive development of the symptoms and signs of the diseases caused by this rust have not been described in full; the incompleteness of this knowl- edge becomes more evident with the increase in the number of known hosts. No definite information has been available concerning the time when the aeciospores of G. globosum germinate or the method by which they produce infection on the Juniperus hosts; knowledge with respect to this question is a prerequisite to any attempt to control the rust on the red cedar. Aside from an academic consideration of the biology of G. globosum, this rust is of increasing economic importance and is causing great damage to both the ornamental and the orchard hosts in localized areas throughout the eastern part of the United States. These desiderata have led the writer to investigate more fully the biology of G. globosum. A determination of the hosts and their rela- tive susceptibility has already been published (MacLachlan, 1935). The numerous inoculations and field observations that were conducted at that time afforded an opportunity to carry on extensive studies on other aspects of the biology of G. globosum; the results of these investi- gations have been incorporated with those of other writers and are now presented. Il. NOMENCLATURE The following names, chronologically arranged, have been given to the rust now known as Gymnosporangium globosum Farl. The stage in the life cycle of the rust to which the name refers precedes the name. III Gymnosporangium globosum Farl. in Bot. Gaz. 11: 236, 239 (Sept. 1886). I Aecidium _ bes Oxyacanthae Schw. Syn. Fung. Carol. 66 40), no. 432 (1822). I Caeoma Cvlindrites Lk. var. Crataegi-punctatae Schw. Syn. Fung. Am. Bor. (Trans. Amer. Phil. Soc. Il. 4: 294, no. 2899a. 32). III Gymnosporangium fuscum DC. var. eal Farl. Gymno- sporan pl. 1, f. 7-11 (188 III Gymnosporangium Sabinae (Dicks. ) ie var. globosum Trel. rans. Wisc. Acad. 6: 133 (29) (July 1866). Ill es ee globosum Farl. in Bot. Gaz. 11: 236, 23 (Sept. 1886). I Roestelia lacerata y & zg Thaxt. ex Farl. in Bot. Gaz. 11: 240 (Sept. 1886). I Gy Wane heroin globosum I Thaxt. in Rept. Conn. Exper. Sta. 14 (20 891). ‘oO I Roestelia globosa Shear, N. Y. Fungi Exsicc. no. 79 (1893). Corrected label. 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 3 III Puccinia globosa Kuntze, Rev. Gen. Pl. 37: 507 (1898). III] Roestelia globosa II] Kuntze in Bot. Centralbl. 77: 300 (15 Feb. 1899). III Tremella globosa Arth. in Proc. Ind. Acad. 1900:136 (June 1901) I Aecidium globosum Farl. Bibl. Index N. Amer. Fungi, 49 (1905). III Aecidium globosum Arth. in Result. Sci. Congr. Bot. Wien, 1905: (1906). The names Gymnosporangium globosum Farl. and Roestelia globosa Shear stand as the authentic names for the III and I stages, respectively. The name Gymnosporangium globosum Farl. is now accepted as refer- ring to either stage of the rust and is so used throughout this presentation. III. RANGE AND ECONOMIC IMPORTANCE OF G. GLOBOSUM A difficulty in determining the exact range and economic importance of G. globosum arises due to the many reports in the literature which refer collectively to the three rusts, G. globosum, G. Juniperi-virginianae, and G. clavipes. Nevertheless, sufficient evidence is extant to make possible accurate determinations of both range and economic importance of G. globosum. Gymnosporangium globosum is confined in its range to the eastern and central parts of United States and to the southern parts of Ontario and Quebec. In Fig. 1 the distribution by states has been plotted; the circles and dots indicate the states in which the diseases caused by this rust have been reported on the Juniperus and pomaceous hosts, respectively. A review of the literature indicates that the prevalence of this rust is steadily increasing; several factors may be involved in this phenomenon. Jones and Bartholomew (1915) offer the suggestion that the alternate hosts, the red cedar and the wild pomaceous hosts such as crab-apples and hawthorns, multiply more rapidly in open pastures and waste cut- over lands than they did in the original forests. The orchardist and ornamentalist, as well as bringing the alternate hosts closer together, have introduced on their estates many susceptible pomaceous trees that may serve as hosts. Whether or not the rust is increasing in virulence is open to question.. The rapid increase in scientific investigation during recent years, which brings to light new host species and through the discovery of these has extended the known range of the diseases caused by this rust, may give an appearance of increased prevalence and virulence. Gymnosporangium globosum is causing more damage in some states 4 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xvi than in others; this, as well, may be due to the influence of man, who is constantly bringing into close range of each other, susceptible alter- nate hosts of the rust. In the eastern part of New York State, for example, this rust is almost on a par with G. Juniperi-virginianae with respect to the damage which it is causing. Stewart (1910) reports severe infection on Kieffer pears in an orchard at Long Island. According to W. D. Mills (Haskell, 1929) G. globosum was unusually prevalent in Duchess and Greene counties on the foliage of pears in 1928. Thomas and Mills (1929a) consider G. globosum as one of the three rusts (G. Fic. 1. DistRIBUTION OF GYMNOSPORANGIUM GLOBOSUM BY STATES globosum, G. Juntpert-virginianae and G. clavipes) destructive to apples in eastern New York State and report the occurrence of G. globosum on at least thirteen varieties of apples, with severe infection in some in- stances. In 1930 they (Thomas and Mills, 1930) list twenty-three varieties of apples on which the disease caused by this rust has been found. H. E. Thomas in a letter, dated Aug. 27, 1930, to the Plant Disease Reporter (Anonymous, 1930) writes that G. globosum was more common that year on the foliage of apples in Essex County than was G. Juniperi-virginianae. Another report (Anonymous, 1930) by W.S. Fields states that all the Crataegus plants of all kinds on an estate at Locust Valley, Long Island, were heavily infected. Other writers who have reported this rust in New York State include Grier (1925), 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 5 Martin (1925), Salisbury (1929), Crosby, Mills and Blauvelt (1929), Barrus, Boyd and Wood (1931) and Miller, Stevens and Wood (1933). Severe local infections have been observed by the writer in eastern Massachusetts. Loci of infection may be found on estates where suscep- tible ornamental Crataegi are planted in vicinities that embrace the red cedar, and in open pastures and waste lands where wild Crataegus and Malus species are within close range of the cedar. Apparently no seri- ous outbreaks of the diseases caused by this rust have occurred, as yet, in commercial orchards in Massachusetts. States in which severe to moderately severe local infections have been reported include Connecticut (Clinton, 1903), Florida (Martin, 1925), Illinois (verbal report from Morton Arboretum), Michigan (Martin, 1923), Minnesota (Martin, 1920) and Ohio (Martin, 1923). Bliss (1933) states that this rust is common on apple in Indiana. Reports would indicate that the rust is not of great economical im- portance in Alabama (Underwood and Earle, 1897), Iowa (Bliss, 1933), Kansas (Bartholomew, 1899), Maine (Morse and Lewis, 1910) (Miller, Stevens and Wood, 1933), New Jersey (Cook, 1917), Ontario (Con- nors, 1934), Wisconsin (Jones and Bartholomew, 1915) and Vermont (Anonymous, 1928). Martin (1925) reports G. globosum as occurring in Alaska. IV. SYMPTOMATOLOGY OF THE DISEASES CAUSED BY G. GLOBOSUM A. Own Pomaceous Hosts Observations were made concerning the progressive development of the symptoms and signs of the diseases caused by G. globosum as they appeared on the inoculations of the respective pomaceous hosts (Mac- Lachlan, 1935). For the sake of convenience the progressive develop- ment of the symptoms and signs will be considered from both the mor- phological and histological viewpoints as they appeared on the foliage of Crataegus; modifications of these symptoms and signs as they appeared on the foliage of other hosts and on the flowers, fruit and twigs, will follow; additional observations made by other writers will be included throughout the presentation. (a) On the foliage of Crataegus. The first morphological evidence of the disease caused by this rust is the exhibition of very small light-colored areas or flecks on the upper surface of the leaf. These appear within ten to twelve days after inocu- lation. In rare cases no further symptoms ever occur except that the 6 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xvi flecks may turn brown; normally, though, these flecks become more distinct and within approximately fifteen days after inoculation (see Table I) they show as bright yellow lesions varying from one to more than ten millimeters in diameter. Sections at this stage of development reveal intercellular hyphae among the palisade and mesophyll cells (Plate 168, Fig. 5) as well as densely intertwined masses of hyphae between the epidermis and palisade layer. These masses of hyphae are the spermogonia primordia which develop to form the mature spermo- gonia approximately twenty-three days after inoculation (see Table I). The mature spermogonia appear on the upper surface of the leaf, rarely on the lower, as small raised points in the center of the lesion. On rupturing the epidermis of the leaf the spermogonia exude a sticky fluid in which abundant spermatia may be found. In Plate 173, Fig. 1 may be seen a photograph of these spermatia (mag. « 545), obtained by placing a drop of the spermogonial fluid on a glass slide. Weimer (1917b) has illustrated the type of spermogonium in cross section. TABLE I DATA ON THE TIME OF OCCURRENCE OF THE SYMPTOMS AND SIGNS OF THE DISEASE CAUSED BY G. GLOBOSUM ON THE FOLIAGE OF CRATAEGUS! . days after inoculation No. Symptoms and signs minimum maximum — average species Ist appearance of lesion 13 Zz 15.1 38 Ist exudation of spermogonial uid 30 229 39 Spermogonia turned black 28 54 41.5 35 Ist appearance of swelling 47 59 530 at Ist appearance of aecia vf 111 96.0 48 Exudation of the spermogonial fluid continues for approximately twenty days (see Table I), following which the spermogonia turn black and are quite conspicuous on the yellowish background (Plate 171, Fig. 1). Other than the color change of the spermogonia, little change in the lesion can be noted except for the formation, in many instances, of reddish borders around the lesions. These borders while manifested to a greater extent on some hawthorns than on others are not consistently formed even on a single host. Severe infections where more than fifty percent of the leaf area is diseased may cause yellowing of the leaf and defoliation at this stage in the development of the rust; scattered in- 1These data were obtained from typical representatives of the various groups of Crataegus inoculated with G. globoswm Farl. in May, 1932 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 7 fections, however, cause only minor injury to the leaf except for reduc- tion of photosynthetic area. Swelling on the lower surface of the leaf, opposite the spermogonia may be observed approximately ten days after the spermogonia cease exudation (see Table I). Sections of the lesions reveal that this swell- ing is due to both hypertrophy and hyperplasia of the mesophyll tissue, resulting in long palisade-like cells closely packed together. Hyper- trophy of the leaf is confined to the area occupied by the lesion and may increase the thickness of the leaf to more than five times normal (Plate 168, Fig. 1). It is not until this stage in the development of the rust that severe leaf killing can be caused by relatively few lesions per leaf; this is especially true of vein infections (MacLachlan, 1935a). Following the time during which hypertrophy of the lesion takes place, minute swellings may be seen on the lower surface, rarely on the upper surface, which, within two or three days time, break through the epi- dermis and appear as greyish acuminate cylinders about one millimeter in diameter. These are the peridia of the developing aecia and are first evident approximately ninety-six days after inoculation (see Table I). The peridia are not arranged in any definite pattern except when they occur on veins or petioles; in that case they may be arranged in two rows, one on each side of the vein or petiole (Plate 167, Fig. 4); they may develop to a length of more than six millimeters and may vary in number from one to more than fifty per lesion. Within each peridium are the aeciospores which are released by the splitting and shredding of the peridium a short distance from the distal end (Plate 171, Fig. 2). Unlike G. Juniperi-virginianae the peridial cells do not recurve with changes in humidity to release the aeciospores; the latter are shed through the shredded region of the peridium. High winds as well as the effect of one leaf rubbing on another have a marked influence in break- ing up the peridium to release the aeciospores. The peridial cells, as described by Kern (1911), are “broadly lanceo- late in face view, 15-23 %* 60~90 uw, linear-rhomboid in side view, 13-19 u thick, outer wall about 1.5 p thick, smooth, inner and side walls 3—5 uw thick, rather densely rugose with ridge-like papillae of varying length.” As has been demonstrated by Thomas and Mills (1929b), the peridial cells are shorter and broader towards the distal end of the peridium and do not separate easily from each other. The aeciospores are borne in chains from the base of the aecial cup and when mature are, as described by Kern (1911), “‘globoid or broadly ellipsoid, 15-19 & 18-25 wu, wall light chestnut brown, 1.5—2 w thick, finely verrucose.” In Plate 173, Fig. 2 may be seen a photograph of 8 JOURNAL OF THE ARNOLD ARBORETUM [voL. xv aeciospores at the same magnification as the spermatia in Fig. 1. Two of the aeciospores (lighter color in the figure) were colorless; these colorless spores were found occasionally and proved to be viable. Sections reveal that the aecium itself is long and narrow and deeply embedded in the hypertrophied tissue, the base of the aecium approxi- mating that of the palisade layer of the leaf (illustrated by Weimer, 1917b). In Plate 168, Figs. 1 and 2 may be seen photographs of a longitudinal section through several aecia; the non-median sections give the appearance that the aecia begin at various levels throughout the hypertrophied tissue. The great majority of the aeciospores are distributed by the middle of September. Remnants of the peridia may remain on the lesion until the leaves normally drop, but usually the circular pits in the aecial cups are all that can be seen at this time. Sections through the lesions after the leaves have dropped reveal very few viable aeciospores; most of those that still remain in the base of the aecium are hollow and sterile. Insects have been observed feeding on the aeciospores during August. The peridia containing the aeciospores are eaten out first, then the insects, after gaining entrance by way of the pits into the aecial cups, proceed to break down the whole interior of the lesion. (b) On the foliage of pomaceous hosts other than Crataegus. Pyrus: The lesions on the ornamental species are smaller than those found on Crataegus. They vary greatly. On some host species they are very minute, being less than a millimeter in diameter and producing a few spermogonia only (e.g. P. serrulata Rehd.); others may exhibit two or three aecia on each lesion (e.g. P. Balansae Decne.) ; still others may exhibit lesions which are two to three millimeters in diameter with several aecia each (e.g. P. betulaefolia Bge.) (Plate 166, Fig. 4). On P. betulaefolia the red borders around the lesions are conspicuous. Stewart (1910) describes the lesions on Kieffer pear as being bright yellow and pin-head in size on June first. He states, further, that by June fifteenth the lesions had turned brown to black with conspicuous red borders on the upper surface of the leaf but without the red borders on the lower surface. He adds that a few aecia are formed on the upper side of the leaf, but most of them are on the lower side, frequently on each side of the mid-rib. Hesler and Whetzel (1917) describe the lesions as being orange-colored in June, one quarter to one half inch in diameter with red borders showing on the upper side of the leaf. In the fall the lesions turned dark on the under side of the leaf and at this time of year showed no red border. 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 9 Sorbus: The lesions in all cases were very small, rarely measuring more than one to two millimeters in diameter, with an average of three to five aecial horns per lesion (Plate I, Fig. 3). As in Pyrus the lesions on certain species died and turned brown shortly after the spermogonia appeared. Malus: Inoculation on ornamental apples resulted in small lesions that rarely measured more than one to two millimeters in diameter (Plate 166, Fig. 5). Certain species exhibited spermogonia only. Crosby, Mills and Blauvelt (1929), Sherbakoff (1932) and others have reported this small type of lesion on commercial apples and refer to it as one method of distinguishing G. Juniperi-virginianae from G. glo- bosum. Bliss (1931), by culture, obtained flecking only on nine vari- eties of commercial apples. (c) On the flowers and fruit. Infection of flowers and fruit of Crataegus is relatively rare when compared to the prevalence of foliar infection. All parts of the flower including pedicel, ovary, sepals, petals and even stamens may be attacked. Infections on the stamens and petals cannot persist but infec- tions of the ovary result in either dwarfing and killing of the fruit (Plate 167, Fig. 1) or premature ripening and dropping of the fruit. Abundant spermogonia and aecia may be produced (Plate 167, Fig. 2). A cross section of a diseased fruit reveals that only a portion of the fruit is affected. In Plate 168, Fig. 3 may be seen a photograph of such a section; the diseased area is represented by the darkened strip on one side of the section. On the opposite side of the section may be seen some of the diseased area of a second lesion. As in the leaf, hyper- trophied, palisade-like, densely packed host cells are evident. The fungal mycelium is intercellular and produces haustoria in the host cells. Some mycelium can be seen ramifying among the host tissue beyond the hypertrophied area. Fruit infection of Pyrus was not observed in the Arnold Arboretum. Stewart (1910) reports the Kieffer pear as suffering infection from this rust at Long Island, New York. In particular he finds the diseased fruits are very small and misshapen, usually exhibiting circular black areas devoid of aecia, although a few show aecia. Hesler and Whetzel (1917) report somewhat similar symptoms on the fruit of pear. Fruit infection of Malus has not been observed by the writer. Gar- man (1899), speaking collectively of G. Juniperi-virginianae and G. globosum, reported these rusts on the fruit of Malus; he found the in- fection to be most abundant on crab apples. 10 JOURNAL OF THE ARNOLD ARBORETUM [voL. XvII (d) On the twigs. Infection of the twigs of Crataegus is relatively rare and occurs on the current year’s growth only. Both spermogonia and aecia may be produced (Plate 167, Fig. 5). Similar infections of the thorns and axillary buds have also been observed (Plate 167, Fig. 3). A cross section of a diseased twig reveals that the infection is confined to the cortical region (Plate 168, Fig. 4). The same typical, densely packed, hypertrophied cells that occur in the leaves and fruit may be seen in the diseased area. The aecial cups, however, appear shorter and broader than those exhibited in diseased leaves. Several diseased twigs were tagged in the fall of 1932 to determine if the mycelium might be perennial in the twigs; such was not evident as the twigs were all dead the following spring. B. On JuNripeRus Hosts The advanced stage of the disease caused by G. globosum on the Juniperus host is expressed in the form of a woody gall attached firmly to a twig that may be more than two or three years old. This condition led early observers to conclude that infection originated in the stem. Farlow (1880) who named this fungus stated that the disease is first manifested by bursting through the stem at the point of attachment of the leaves. Kern (1911) described the gall as being caulicolous. Stewart (1915) gave an elaborate account of histological studies made, which he interpreted as showing that the gall originates from the axils of leaves and are evidently transformed axillary buds. Weimer (1917a), after examining Stewart's slides, showed that Stewart mistook an axillary bud for a gall. He (Weimer, 1917a) gave a full description of a normal juniper leaf and presented definite proof that the gall may originate in the leaf. Bliss (1933) has substantiated the results of Weimer and has illustrated stages in the progressive development of the gall on the leaf. Several brief descriptions of the mature gall as it occurs on the red cedar may be found in the literature; these descriptions are given, for the most part, as an aid in distinguishing the galls caused by G. globosum from those caused by G. Juniperi-virginianae. To obtain a more complete picture of the progressive development of the various symptoms and signs of the disease caused by G. globosum on the red cedar, the writer has carried on further studies both morpho- logically and histologically. Material for study was obtained from a locus of infection that was isolated from other Gymnosporangium rusts. The symptoms and signs will be considered as they occur on the red cedar in eastern Massachusetts. 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 11 The initial manifestations of the disease caused by this rust may be found on one year old leaves of infected red cedars about the first. of August. The exhibition of a yellowish chlorotic zone or band on the leaf is closely followed by the formation of a slightly raised area on the inner (upper) surface of the leaf; the lifting of the epidermis over the point of infection is due to the development of a very young gall underneath. Longitudinal splits in the epidermis immediately over the raised area allow the young gall to emerge. On long subulate leaves the gall may occur at any place throughout the length of the leaf; in Plate 169, Figs. 2 and 3, may be seen young galls developing near the tip, in the middle and near the base of the leaf. The usual type of leaf and young gall that is found is illustrated in Plate 169, Fig. 1; the leaves are short and the gall must of necessity be brought in close contact with the stem, a phenomenon which led early investigators to conclude that the gall originates in the stem at the base of the leaf. A longitudinal section of a portion of a twig bearing two normal leaves as well as a diseased leaf, at approximately the same stage of gall de- velopment that is shown in Plate 169, Fig. 1, is illustrated in Plate 170, Fig. 1. A comparison of the two normal leaves with the diseased leaf reveals marked hypertrophy and hyperplasia of the mesophyll of the latter. Distortion of the vascular system may also be seen. Figs. 4 and 5 of Plate 170 exemplify further the distortion of the vascular elements. In Fig. 4 parenchymatous cells may be seen separating the vascular elements. In Figs. 2 and 3 a corky layer, several cells thick and com- pletely surrounding the portion of the gall not in contact with the twig, may be seen. This corky layer usually cuts off the tip of the leaf; it is of interest to note in Figs. 3 and 4 the sharp severing of the vascular strand. The tip of the leaf that is cut off together with the mycelium of the rust that may be found in it dies; remnants of this dead portion may be found on the gall for two or three years afterwards. Intercellular mycelium is abundant throughout the gall tissue but haustoria were not observed at this stage. The gall as seen in late autumn and winter is smooth, shiny and mahogany red in color, rarely exceeding a diameter of half a centimeter. This is in striking contrast to the larger, greenish and more or less con- voluted gall of G. Juniperi-virginianae. About the last’of March of the following spring light orange colored markings may be seen over the surface of the gall. These markings represent early stages in the rupturing of the corky epidermis and are caused by the underlying masses of hyphae—the telia primordia—from 12 JOURNAL OF THE ARNOLD ARBORETUM [voL. XVII which the teliospores arise. Progressive stages in the maturation of the teliospores from this dense hyphal layer are shown in Plate 171, Figs. 5, 6 and 7. Dodge (1918) has published an extensive cytological study of the development of the teliospores and has illustrated progressive stages of their maturation. The teliospores mature progressively from the centre towards the periphery of the layer of telia primordia. The corky epidermis is ruptured over the maturing teliospores and the latter emerge as a cCinnamon-brown, pulvinate mass (Plate 171, Fig. 4). The irregu- lar outline of this rupture readily distinguishes the galls caused by this rust from those caused by G. Juniperi-virginianae since the telial sori of the latter emerge through circular openings. The teliospores are borne on pedicels and the extension of these pedicels, as the teliospores pro- gressively mature from the centre towards the periphery of the basal layer of mycelium, enlarges the sorus to form a tongue or wedge-shaped structure, the centre of which may be hollow. This cavity is due to the developing teliospores around the peripheral regions of the sorus; these push up the mature teliospores in the centre and as a result the pedicels of the latter teliospores are broken off. A mature telial sorus, if kept dry, is a tongue-like or wedge-shaped flange 1-3 mm. broad by 2—5 mm. long at the base and 6—12 mm. high. A section through such a sorus reveals a dense cinnamon-brown layer of teliospores, several spores thick, over the surface. Under the spore layer is a dense white mass composed entirely of long pedicels, one for each teliospore. The peripheral pedicels still remain attached to the layer of mycelium from which they arose; while the pedicels in the central region, which have been broken off from the layer of mycelium from which they arose, are suspended to form a conical amphitheatre-like cavity in the centre. The teliospores are typically two-celled and are, as described by Kern (1911), “ellipsoid, 16-21 % 37-48 u somewhat narrowed above and below, slightly constricted at the septum, wall pale cinnamon-brown, 1-2 uw thick; pores 2 in each cell, near the septum.” Pammel (1905) has described one-celled teliospores. No single-celled teliospores have been observed by the writer; three- and four-celled teliospores, how- ever, as well as four-celled teliospores with one cell abortive have been found (Plate 173, Fig. 3). Studies with respect to the nature of the pedicels of the teliospores and the effect of wetting will be considered later (page 16). Wetting of the telial sori by rains in May causes them to expand to many times their original size. Progressive stages in the expansion of sori on galls as well as on the small infections that take place on the long 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 13 subulate leaves are illustrated in Plate 172. The hygroscopical nature of the pedicels (see page 16) is responsible for this expansion. If only a portion of the teliospores are mature at the time of wetting, the telial mass may, on drying, shrink to somewhat of its original shape and size. If, however, all the teliospores are mature at the time of original wetting and the humidity remains high for a sufficient length of time to fully expand the telial mass, the latter, on drying, will shrink to an amorphous sheet or strand and later fall away from the gall. Under favorable weather conditions practically all the teliospores are mature within three to five weeks from the time of their initial appearance. Germination of the teliospores results in the production of basidio- spores. Progressive stages in the production of the latter are shown in Plate 174, Fig. 1. The basidiospores may be seen as a powdery, some- what velvety coating over the surface of the expanded telial sorus and are evident about five or six hours after the expansion takes place. These basidiospores will drop from the gall of their own accord but are usually carried away by the air currents. The basidiospores are ovate, flattened on one side, slightly tapering towards the end of attachment to the sterigmata; thin-walled (less than 1 u thick); protoplasm light orange- yellow in color containing numerous oil droplets which tend to aggregate (Plate 173, Fig. 4); size, 12-16 (average, 13.7 1) XX 7-11 (average, 8.9 1). Under favorable conditions they germinate immediately after their formation to form relatively long germ tubes (Plate 174, Fig. 4). If the spores are allowed to dry shortly after the germ tubes begin to develop, small secondary spores may be formed (Fig. 5). As a rule, the majority of the teliospores have germinated by May 25. The dried up remnants of pedicels and empty teliospores blow away leaving a smooth orange colored scar bordered by the lacerated edges of the broken corky epidermis. This scar soon turns greyish brown and may persist on the gall surface during the remainder of the life of the gall (Plate 171, Fig. 4). Cross sections through the gall reveal that the tissue immediately underlying this scar dies and a corky layer develops, segregating this dead area from the remaining living tissue of the gall (Plate 170, Fig. 5). The gall continues growth during the summer causing distortion of the neighboring twigs (Plate 169, Fig. 5) and in some cases killing of the twig beyond the gall (Plate 169, Figs. 4 and 6). The original single vascular strand of the leaf now shows as a many-branched structure (Plate 170, Fig. 5). New growth beyond the gall as well as the natural dropping of the older leaves and the formation of a corky epidermis on the twig surrounding the gall give the appearance that the gall originated 14 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI in the stem rather than in the leaf. This illusion is further substantiated by the tendency of the gall to become woody in appearance as it grows older (Plate 169, Fig. 7). The gall is commonly referred to as being perennial in nature. For how many years it will live and produce teliospores is not known; in Plate 169, Fig. 7 is shown a gall that produced teliospores for three con- secutive years. The telial sori are formed each year between the scars where teliospores were produced the previous years (Plate 171, Fig. 4). It is doubtful that many of the galls persist for more than three or four years, because, by that time the surface of a gall would be completely covered with dead tissue resulting from the formation of the telial sori. Long subulate leaves on which infection takes place usually drop after the first crop of spores; these infections, then, can persist for one year only. V. FACTORS AFFECTING SPORE GERMINATION A. Tue Time Factor The teliospores exhibit a low and inconsistent percentage of germina- tion when tested immediately after their appearance on the gall. If twigs bearing galls on which teliospores are just emerging are placed in water in the greenhouse, abundant germination can be obtained, regu- larly, in a week to ten days time. The maturation period of the telio- spores is longer, depending on the weather conditions, when the galls bearing them are left in the field. Normally, the teliospores will all have germinated within six to eight weeks after their formation. Under low temperature conditions, however, they have the potential ability to remain viable for a much longer time; in Plate 175, Fig. 3 may be seen severe infection obtained on leaves of Crataegus Jonesae Sarg. using, as inoculum, teliospores that had been kept in the refrigerator for one year. The basidiospores have the ability to germinate immediately after their formation. The length of time that they will live is still a question; basidiospores of G. Juniperi-virginianae will live for many days above a humidity of 25% and below a temperature of 25° C. (MacLachlan, 1935b Aeciospores will not germinate to the extent of more than 2 or 3% at the time of their formation. Nevertheless, if hawthorn leaves bear- ing aecia are collected in August before the aeciospores are shed and placed in a refrigerator at 0° C., abundant germination can be obtained within a month’s time and by the first week in October more than 80% germination can be obtained consistently. By keeping the infected leaves at this temperature, the same high degree of germination can be 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 15 obtained until the first of March of the following year; after this date, however, the percentage of germination that can be obtained falls off rapidly. Whether the infected leaves are kept in a dry container or a moist chamber makes no difference in the percentage germination obtained. It may be noted that on one occasion more than 75% germi- nation was obtained in August, 1932, using aeciospores taken from un- broken peridia of an infected leaf of Crataegus shirleyensis Sarg. B. Tue Humipity Factor Tests in the laboratory revealed that none of the three spore forms would germinate, even in saturated humidity (spores placed on cover glasses over water ), unless in direct contact with a drop of water. Excess water, however, causes irregularity in the percentage germination; basidiospores, especially, show a very low percentage of germination when immersed in large drops of water; under the same conditions aeciospores are not consistent in the percentage of germination that may take place; teliospores germinate in excess water but instead of produc- ing basidiospores the promycelia grow to great lengths, presumably to come in contact with the air, and may exhibit long side tubes (Plate 174, Fig. 3), or break up to form elongated spore-like bodies (Plate 174, Fig. 2). Farlow (1880) reported a type of spore, similar to that shown in Fig. 2, which he observed when the telial masses were quickly dried after moistening. He also found that, on remoistening, these spore-like bodies would send out germ-tubes similar to those of normal basidio- spores. In the laboratory, optimum humidity conditions for germination of any of the three spore forms was obtained by placing the spores on a cover glass and inverting the latter on a Van Tieghem cell in a petri-plate lined with wet filter paper; enough water of condensation would accu- mulate on the lower (spore) surface of the cover glass to give optimum humidity conditions for germination. In the field, very satisfactory humidity conditions for infecting the pomaceous host were obtained by painting the leaves with an aqueous suspension of the teliospores and enclosing the inoculated twig in a celluloid cylinder, the ends of which were plugged with moist sphagnum. C. Tue TEMPERATURE FACTOR Temperature has a marked effect on the percentage of germination that may be obtained. Miller (1932) found 24° C. to be the optimum temperature for the germination of teliospores and aeciospores, and 16° C, the optimum for basidiospores. In Table II may be found data 16 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XVI TABLE II DATA ON THE PERCENTAGE GERMINATION OF THE SPORES OF G. GLOBOSUM WHEN SUBJECTED TO DIFFERENT TEMPERATURES! Temp. Percentage germination of Ke Ge teliospores basidiospores aeciospores 2 0.0 —_ 0.0 5 0.3 0.0 4 7 0.7 7.2 8.2 10 80.4 68.7 11.3 15 90.9 79.2 70.2 20 96.3 83.9 88.7 25 98.6 63.1 60.7 30 59.8 14.5 35.3 35 0.9 0.9 0.6 38 0.0 0.0 0.0 on the percentage germination of the three spore forms when sub- jected to various temperatures; these data have been plotted in Fig. 2. Examination of Fig. 2 reveals that: (1) the teliospores have a very wide range of temperature within which more than 80% germination may take place, (2) the corresponding curve for basidiospore germination has the same general contour, except at a lower level, (3) the aecio- spores have a narrower range of temperature within which a high per- centage of germination may take place, the optimum being around 20° C., (4) below a temperature of 10° C. and above a temperature of 30° C., the percentage germination of all three spore forms drops off to almost zero. VI. THE NATURE OF THE TELIAL SORUS Reference was made under SyMp ocy (page 12) to the marked expansion of the telial excrescences on the galls during wet weather. In Plate 172 may be seen photographs of progressive stages in the expansion of these telia. Further studies were made with respect to this phenomenon. It was found that the pedicels of the teliospores, when wetted, are responsible for this expansion process. No estimation has been made with respect to the number of teliospores and the corresponding number of pedicels that may exist in a single telium, yet, they must number in the millions. Crowell (1934) has given estimations which indicate that four to five million teliospores may exist in a telium of G. Juniperi- 11500 spores were counted in each sample. Counts were made 24 hours after the cultures were set up. 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 17 virginianae ; the number that may exist in a telium of G. globosum would be as many or more as the telia are characteristically larger in the latter rust. The pedicels, as well as being numerous, are relatively long; some idea of their length may be obtained in Plate 175, Fig. 5. Under strong light and on a dark background they appear white and opaque; they do not stain readily in aqueous methylene blue. In Plate 175, Fig. 4 may be seen a photograph of a pedicel that was immersed in 25% alcoholic @-—TELIOSPORES 100 Q@-— BASIDIOSPORES O- AECIOSPORES r o 80] ra < z 2 60, uJ rT) 0 < 40) = z uJ ) tl a 20) ce) ll 7 Tt v ¥ ¥ as ) 5 10 15 20 25 30 35 40 alae A | DEGREESC. Fic. 2. THe INFLUENCE OF TEMPERATURE ON THE acai GERMI- NATION THAT MAY BE OBTAINED FOR THE THREE SPORE FoR methylene blue for fifteen hours; the pedicel itself did not stain to any extent but the stain was absorbed in the centre of the pedicel indicating the presence of a lumen. The hygroscopic nature of the pedicel is illus- trated in Plate 175, Figs. 1 and 2. The pedicels represented in Fig. 1 were subjected to 95% alcohol for twelve hours while those represented in Fig. 2 were placed in water for the same length of time. The two pictures are at different magnifications but when one compares the diameter of the pedicel with the size of the teliospore which it subtends, it is quite evident that the pedicels are capable of considerable expan- sion when wet; Fig. 1 illustrates the pedicel as it would appear when the telium is dry while Fig. 2 illustrates the pedicel as it would appear during 18 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XvII rainy weather in May. In Fig. 2 may also be seen small air bubbles existing in the region that was once the lumen but is now filled by the expanded wall of the pedicel. On the other hand, as may be seen in Fig. 1, the lumen is relatively wide with respect to the diameter of the desic- cated pedicel. Taking into consideration, then, the number of pedicels present in a telium as well as their length and the diameter they may attain when wet, one can readily see how a telium may expand to the size that is so common on infected cedars during rainy periods in May. VII. INFECTION OF THE RED CEDAR A. TIME oF INFECTION The actual time of year that infection of the red cedar by the aecio- spores takes place is still conjectural. About 2% of the aeciospores will germinate at the time of their dispersal in late August and when one realizes the large number of aeciospores that may be released from infected pomaceous hosts it can be presumed that possibly some infec- tion may take place immediately. Nevertheless, the fact that more than 80% germination can be obtained after the fresh aeciospores have been subjected to 0° C. for about six weeks time (cf. page 14) would indicate that both time and low temperature are involved in the matura- tion of the aeciospore. This would indicate that infection takes place in late fall after frosts have occurred. The fact that a high percentage of germination can be obtained in March of the following spring, using aeciospores that have been kept at 0° C. during the interim, would not necessarily mean that infection takes place in the spring; if the aecio- spores are mature in late fall, in all probability they will have all germi- nated long before March. B. INFECTION PROCESS Having access to an abundance of aeciospores which would germinate readily, attempts were made to trace the method by which infection of the red cedar by this rust takes place. Leaves of the current year’s growth were removed from potted red cedars, washed carefully and dusted on the stomatal (inner and upper) surface with aeciospores. The inoculated leaves were then inverted and placed on moist filter-paper in petri-plates. Examination 24 hours later revealed abundant germination of the aeciospores; the germ tubes were long and slender and sometimes spiralled; the protoplasm had all mi- grated to the end of the germ tubes. Within 12 hours more, the ends of the germ tubes had enlarged to form irregular and sometimes convoluted 1936 ] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 19 haustoria-like formations. On one occasion the germ tube was observed to bend at a sharp right angle to form the haustoria-like structure over a stoma; however, no consistent orientation of the germ tubes with respect to the location of stomata could be observed. The haustoria- like structures that happened to form over stomata were smaller than those that formed on the inter-stomatal regions of the leaf epidermis. Two samples of about fifty leaves each were removed from the culture and killed for embedding thirty-six and fifty-two hours, respectively, from the time of inoculation. These were sectioned (both transversely and longitudinally), stained with safranin and light green and exam- ined under the high power of the microscope in hopes of observing the method of penetration. The length of the germ tubes made it impossible to obtain even a series of sections that showed an aeciospore connected to its germ tube throughout its entire length. This led to a difficulty in the identification of germ tubes that were observed entering the leaf and consequently the results obtained can only be considered as indications and not conclu- sive proof. The characteristic intercellular mycelium of G. globosum was found around the mesophyll cells in the immediate vicinity of the stomata. Hyphae of the same diameter as the germ tubes were found passing through the stomata; in some instances three or four passed through the same stoma. Some of the haustoria-like structures were found to be still clinging to the surface of the leaf but no penetration from them was observed. Whether the germ tubes that were observed passing through the stomata were those of G. globosum or of some other fungus that hap- pened to be mixed with the aeciospores cannot be said for certain. It is quite safe to say, however, that the rust gains entrance through the upper epidermis of the leaf on which the stomata exist, as the character- istic intercellular hyphae of this rust were found around the mesophyll cells immediately under the stomata. Further studies on this problem will be carried out by the writer. VIII. SUMMARY At least thirteen names have been given to the rust now known as Gymnosporangium globosum Farl. Of these the authentic names for the III and O & I stages of the rust stand as Gymnosporangium globo- sum Farl. and Roestelia globosa Shear, respectively. The name Gymno- Sporangium globosum Farl. is now accepted as referring to either stages of the rust. 20 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Gymnosporangium globosum is confined in its range to the eastern and central parts of the United States and to the southern parts of Ontario and Quebec. The rust is increasing in prevalence and in local- ized areas is causing great damage to both the ornamental and orchard trees; this is especially true in eastern New York State. The diseases caused by G. globosum occur on at least ten genera of the Pomoideae and on at least three species of Juniperus. The symp- toms and signs of the diseases caused by this rust may be found on the foliage and to a lesser extent on the flowers, fruit and twigs of the poma- ceous hosts and on the foliage and twigs of the Juniperus hosts. The progressive development of these symptoms and signs as they occur on the aforementioned organs of the respective hosts have been described and illustrated. The factors of time, humidity and temperature have been considered with respect to the percentage germination of the teliospores, basidio- spores and aeciospores that may be obtained. The age of the three spore forms as well as the temperature to which they are subjected have a marked effect on the percentage germination that can be obtained. The amount of water present also modifies the percentage germination that can be obtained as well as the type of germ tubes that may be exhibited. All evidences indicate that infection of the red cedar by G. globosum occurs, primarily, in late autumn. From studies of the infection process it would appear that infection of this host takes place through the upper and stomatal surfaces of the leaves. IX. ACKNOWLEDGMENTS To Professor J. H. Faull, who afforded the writer the opportunity to study this problem and for guidance and supervision during the pursu- ance of its investigation, sincere appreciation is expressed. The writer is also grateful to Dr. Ivan H. Crowell for many helpful suggestions and to Mr. Francis M. Bryant for technical assistance. X. BIBLfLOGRAPHY ee (1928). Apple rust (Gymnosporangium spp.). (Plant Dis. ep. 7: (1930). Apple rust diseases in New York. (Plant Dis. Rep. 14: i nee Barrus, M. O. oyp and J. I. Woop (1931). Plant diseases in the United eee “geen Dis. Rep. Suppl. 81: 72.) BARTHOLOMEW, FE, (1899). The Kansas Uredineae. (Trans. Kan. Acad. : J Sci. 16: 184 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 21 Buss, D. E. (1931). Physiologic specialization in Gymnosporangium Casual ee ee 4 Te BB ye oo The pathogenicity and seasonal development of Gymno- een in Iowa. (Iowa State Coll. Agric. and Mech. Arts Res. CLINTON, Ge (1903). Diseases of plants cultivated in Connecticut. (Ann. Rep. Conn. Agric. Exper. Sta. Part 4: Connors, I. L. 19 4). VI. Diseases of ornamental plants. (Ann. Rep. Coox, M. T. (1917). Common diseases of apples, pears and quinces. . J. Agric. Exper. Sta. Circ. 44: 14. Crossy, C. R., W. D. Mitts and W. E. BiavuveEtt (1929). Protecting orchard crops from diseases and insects. (Cornell Univ. Agric. Exper Sta. Bull. 498: 36.) CroweE Lt, I. H. (1934). The hosts, life history and control of the cedar- apple rust fungus a ala Juniperi-virginianae Schw. (Jour. Arn. Arb. 15: 165-232 Dopce, B. O. (1918). Studies ¢ on the genus Gymnosporangium—III. The origin of the teleutospore. ipcole da: 10: 182-189. ) Fartow, W. G. (1880). The Gymnosporangia or cedar apples of the United States. (Anniv. Mem. Bost. Soc. Nat. Hist. pp. 9, 18, 34-35.) GarMAN, H. (1899). Some pests likely to be disseminated from nurseries. ieols) Grier, N. M. (1925). The native flora of the vicinity of Cold Spring Harbor, L. I., New York. (Amer. Midl. Nat. 9: 94. HaskeELt, R. J. (1929). Diseases of fruit and nut crops in the United Sets in 1928. (Plant Dis. Rep. Suppl. 70: 210. Hester, L. R. and H. H. Wuerzer (1917). Manual of fruit diseases. (MacMillan aad Co., New York, pp. 341- aay Jones, L. R. and E. T. BARTHOLOMEW (1915). pie rust and its control in Wisconsin. (Wis. Agric. Exper. Sta. Bull. ar Kern, F. D. (1911). A biologie and taxonomic ae of the genus Gym- nosporangium. (N. Y. Bot. Gard. Bull. 7: 468-469. MacLacutan, J. D. (1935a). The hosts of Gymnosporangium globosum Farl. and their relative susceptibility. (Jour. Arn. Arb. 16: 98-142.) 35b). The ne eal of viable basidiospores of Gymno- sporangium rusts. our. Arn. Arb. 16: a ee. MartTIN, G. H., Jr. (1920). mie of cotton, sugar cane, forest trees and miscellaneous plants in the United ieee in 1919. (Plant Dis. Survey Supp 19 VAN Diseases of forest and shade trees, ornamental and miscellaneous plants in the United oe in 1922. (Plant Dis. Survey Suppl. 418. — (1925). Diseases of forest and shade trees, ornamental and miscellaneous plants in the United States in 1923. (Plant Dis. Rep Miter, P. R. (1932). Pathogenicity of three red-cedar rusts that occur on apple. ae 22: 728, 73. Mitter, P. R E. STEVENS and J. I. Woop (1933). Diseases of plants in the naa: States i in 1931. (Plant Dis. Rep. Suppl. 84: 50. Morse, W. J. and C. E. Lewis (1910). Maine apple diseases. (Maine PAMMEL, L. H. (1905). The ‘cedar apple fungi and apple rust in Iowa. (Iowa Exper. Sta. Bull. 84: 5-16.) 22 JOURNAL OF THE ARNOLD ARBORETUM [voL. xvu SatisBury, W. S. (1929). Cedar ae (Plant Dis. Rep. 13: 7.) SHERBAKOFF, C. D. (1932). e more important diseases of apples in Tennessee. (Univ. of Tenn. pen Exper. Sta. Bull. 145: 34-37.) Stewart, A. (1915). Anatomical study of Gymnosporangium galls. (Amer. Jour. Bot. 2: 402-417.) Stewart, F. C. (1910). Notes on New York plant diseases, I. (N. Y. — Exper. Sta. Bull. 328: 376. Tuomas, H. E. and W. D. Miris (1929a). Rust diseases of the apple. ie 19: 87.) —— (1929b). Three rust diseases of the apple. (Cornell Univ. Agric. Exper. Sta. Memoir, 123. — (1930). Rust diseases of the apple. (Plant Dis. Rep. 14: 214-215.) Unperwoop, L. M. and F. S. Earte (1897). A preliminary list of the Alabama fungi. (Ala. Agric. Exper. Sta. Bull. 80: 212. Weimer, J. L. (1917a). The a and development of the galls pro- uced by two cedar rust fun (Amer. Jour. Bot. 4: 241-2 (1917a). hree ae rust fungi, their life histories and the diseases they produce. (Cornell Univ. Agric. Exper. Sta. Bull. 390: 527-538. ) XI. EXPLANATION OF PLATES PLATE 166 Aecial stage of G. globosum on the foliage of pomaceous hosts. Unless signified inoculations were artificia Fig. 1. On Crataegomes pilus andehien Bean. Note leaf killing. Fig. 2. On Crataegus pruntfolia (Marsh.) Persoon. Fig. 3. On Sorbus americana Marsh. The lesions are very small. Fig. 4. On Pyrus betulaefolia Bge. Fig. 5. On Malus sp. (Natural infection. ) PLATE 167 Aecial stage of G. globosum on various parts of Crataegus hosts. (Arti- ficial inoculations. ) Fig. 1. Dwarfing and killing of fruit. Fig. 2. Abundant aecia produced on basal portion of a fruit Fig. 3. Ona terminal bud that was forming in the axil of a leaf. Fig. 4. On the petiole of a leaf. Fig. 5. On twigs of the current year’s growth. PLATE 168 Histological sections of Crataegus showing diseased areas. Fig. 1. Cross section through a diseased Crataegus leaf showing the amount of hypertrophy that may occur in the lesion. Longitudi- The section also uae: a near-median section of an aecium. Fig. 3. Section through a diseased Crataegus fruit. Hypertrophied area of the lesion shows as a dark strip in the photograph. This area is composed of densely packed, eee like, cells similar to those which occur in the foliar lesions 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 23 Fig. 4. Cross section through a lesion on a Crataegus twig. ee tite and intercellular mycelium are confined to cortical regio The aecial cups are eet stierily shorter and broader aan those found on the foliage. Big so, Intercellular mycelium in the mesophyll of a diseased a eaf under a spermogonium before hypertrophy has s PLATE 169 s of G. Pa eat in various stages on leaves and twigs of Juniperus ie Redes -cedar ). ig. 1. The eee type of young galls that may be seen aia through Fig. 4. A one-year-old cal that has caused the death of the twig alate it. Fig. 5. A two-year-old gall that is causing distortion of the tw Fig. 6. A two-year-old gall that has killed the twig beyond the aon of hae ction. The dead portion of the twig has fallen away. (See g. 4.) Fig. 7. aL old gall (3 or 4 years) that gives the appearance of having originated in the stem. The infection that resulted in this gall took place when the twig bearing it was young and covered with green leaves. PLATE 170 Histological sections of Juniperus infected with G. paatieie ig. 1. Longitudinal section through a twig bearing two mal leaves n infected leaf. Note hypertrophy of the a Peo leaf as well as the distortion of the single strand of vascular elements. Fig. 2. A later stage than that shown in Fig. 1. A corky layer several cells thick has set in that surrounds the gall and cuts off the tip of the leaf. The tip of the leaf that is cut off together with the sail ralasehes that roa e found in it dies but may persist in the or three Fig. 3. ae portion of aii similar to that shown in Fig. 2. Note Fig. 4. Enlarged portion of section shown in Fig. 3. ote par renchym a- like cells that are forming within the vascular strand splitting up e latter. Fig. 5. A section through a two-year-old gall. The splitting up of the single vascular strand of the original leaf has now resulted in a spring; a corky layer had set in, segregating the dead portion from the remainder of the gall. PLaTE 171 Fruiting structure of G. globosum on Crataegus (Figs. 1 and 2) and Juniperus (Figs. 3-7). Fig. 1. Spermogonia in the center of a foliar lesion; they have turned and are quite conspicuous on the yellow background of the lesion. Fig. 2. Typical aecia on the foliage. Note method of shredding of the peridial cells for the release of the aeciospores. 24 JOURNAL OF THE ARNOLD ARBORETUM [voL. xvi ture the epi idermis readily Pike ge ier the galls caused by G. globosum from those caused by G. Juniperi-virginianae in fuo circular pit-like openings are formed. Note also the scars be- tween the sori; within the area occupied by these scars iat sori occurred in previous years Figs. 5, 6 and 7. Progressive stages in the development of the layer of Fie primordia that form under the corky epidermis of a gall in : corky layer is ruptured about the time that the Elie are formed. PiaTE 172 Gelatinization of the telial excrescences. Figs. 1 and 2. Before and after wetting of two infections on a long ead leaf. No gall formation was evident on this el Figs. 3, 4 and 5. a eal stages in gelatinization of three telial sori on ngle one-yea gall. Figs. 6 ae oe Before on after wetting of a single ecies on a long leaf. Note the formation of a gall on this Fig. 8. Gelatinization of the usual type of gall that is rae y PLATE 173 Spore forms of G. globosum (Magnification * 545). Fig. 1. Spermatia. Fig. 3. rae yas The usual type are two- celled. Occasionally three- ace celled spores may be found. Note the four-celled telio- ith one cell abortive. Fig. 4. Bee Note the conspicuous yellow oil droplets. PLATE 174 Germinating spores of G. globosum. Fig. 1. Progressive stages in the germination of a teliospore to the dase of basidiospores. (A portion of the promycelium of er cell was out of focus and does not show in the last ic Fig. 2. pictur promycelia of a teliospore that have broken up to form spore-like bodies ae of the normal production of basidio- WwW nating. They occur frequently if the teliospore is in too m water. Fig. 3. A promycelium of a teliospore that has produced long projections instead of uipeun aie This as well occurs frequently if the teliospore is in too m , Fig. 4. Germinating basidio Fig. 5. A germinated ay a that has produced a secondary spor Secondary spores are frequently found when ce ‘nies | ae es are allowed to dry shortly after germination begin Fig. 6. A germinated pre that has produced a nee branch, Fig. 7. Germinating aeciospor Jour. Arnotp- Ars. VoL. XVII PLATE 166 (GYMNOSPORANGIUM GLOBOSUM Farl. Jour. Arnotp Ars. VoL. XVII PLATE 167 GYMNOSPORANGIUM GLOBOSUM Farl. Jour. ArNoLp Ars. VoL. XVII PLATE 168 GYMNOSPORANGIUM GLOBOSUM Farl. Jour. ArNoLp Ars. VoL, XVII PLATE 169 ¥ ay sop hed i crs Sa GYMNOSPORANGIUM GLOBOSUM Farl. Jour. ArNotp Ars. VoL. XVII PLATE 170 Pe Set — ' BS 5 rhe ri EAN }' ~ , uy ee GYMNOSPORANGIUM GLOBOSUM Farl. Jour. Arnotp Ars. VoL. XVIT PuaTeE 171 1 Ais a Hy) 4 By 77, ae? | f tf, FA hi 4i4 i de ULI es ‘J Opt, ee Ss saeahy apt ® &e® 8 ‘v, & & GY MNOSPORANGIUM GLOBOSUM Farl. Jour. Arnotp Ars. VoL. XVII Piate 172 (8 ab ‘ GYMNOSPORANGIUM GLOBOsUM Farl. Jour. Arnotp Ars. VoL. XVII PLaTE 173 GYMNOSPORANGIUM GLOBOSUM Earl. Jour. AkNo_p Ars. VoL. XVII Prater 174 GYMNOSPORANGIUM GLOBOSUM Farl. Jour. ARNotp Ars. VoL. XVII PLATE 175 GYMNOSPORANGIUM GLOBOSUM Farl. 1936] MacLACHLAN, GYMNOSPORANGIUM GLOBOSUM 25 PuatTeE 175 Teliospores of G. globosum. Fig. Fig. Ls w&w aN wm edicels and teliospores subjected to ou alcohol for twelve hours. Note the slender thin-walled p Pedicels and teliospores subjected to ae for twelve hours. The photograph is at a larger magnification than that of Fig. 1 (com- pare sizes of teliospores) but marked swelling of the pedicels has occurred; it is the swelling of these pedicels that causes the telial sori to swell during rainy periods in May. ote the air bubbles in the pedicel subtending the spore; these are in what was once the lumen but the latter has now lost its identity due to the swell- ing of the pedicel wall. . Severe infection obtained on the foliage of Crataegus Jonesae Sarg., the result of inoculation using Me Ms buise that had been kept in a refrigerator at 0° or on . A teliospore and its pedicel that were sibieceal to 25% alcoholic methylene blue for fifteen hours. rae all of the pedicel did not stain but the stain is evident in the Portions of pedicels of teliospores. aoe idea of their length may be obtained from this photograph. LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, Harvarp UNIVERSITY. 26 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XVII THE LIFE HISTORIES OF MILESIA SCOLOPENDRII, M. POLYPODII, M. VOGESIACA AND M. KRIEGERIANA LittiAN M. HuNTER With plate 176 Heretorore the life histories of ten species of Milesia in all have been determined more or less completely by Klebahn (12), Faull (3), Kamel (9, 10, 11) and Mayor (14). Four more are now added. During the year 1933-4 it was my privilege to make studies on the genus Milesia in England. Through personal collecting and the help of others telial material of Milesia Kriegeriana (Magn.) Arth. on Dry- opteris spinulosa (O. F. Miiller) Kuntze and on Dryopteris spinulosa dilatata (Hoffm.) Underw., M. Polypodii White on Polypodium vul- gare L., M. Scolopendrii (Fuckel) Arth, on Scolopendrium vulgare Smith and M. vogesiaca (Syd.) Faull on Polystichum angulare Pres] was obtained from Ireland, and the following from England—M. Blechni (Syd.) Arth. on Blechnum Spicant (L.) With., M. carpatica (Wrob.) Faull on Dryopteris Filix mas (L.) Schott and D. spinulosa, M. Poly- podii on Polypodium vulgare, M. Scolopendrii on Scolopendrium vulgare and M. Whitei Faull on Polystichum angulare. My experiments were restricted to M. Scolopendrii, M. Polypodii, M. vogesiaca and M. Kriegeriana. All were obtained from overwintered fronds in the spring months of 1934. Teliospore formation began in late March and con- tinued through April and May. It is of interest to record that search was made from October 1933 to early January 1934 for the teliospores of M. Scolopendrii, M. Polypodii, M. Kriegeriana and M. Whitei; but uredospores only were found. The fern fronds bearing teliospores were sent in from the field and placed in a frigidaire, where they were kept at a temperature of approxi- mately 11° C. It was found that the range of temperature over which germination of the teliospores took place was between 16° and 20° C. Material was prepared for germination in the usual way. Pieces of rust- affected fronds, spread out on wet absorbent paper in petri dishes, were tested for germination in constant temperature chambers set at various temperatures ranging between 15° and 30° C. Material set up in a similar way was left in the laboratory and a thermograph record was kept of the room temperature. Records show that teliospore germina- tion was not secured when the temperature remained constantly above 1936] HUNTER, LIFE HISTORIES OF MILESIA 27 20°; but if it dropped to 20° for a part of the time some germination was secured. After the rusted fronds had been stored in the frigidaire they dried out noticeably; moistening before setting the teliospores to germinate was necessary. To do that the leaves were placed in tap water for a few minutes to an hour and then surplus water was removed by placing them between layers of absorbent paper before arranging them in petri dishes. Excellent germination was secured when the correct temperature and moisture conditions were maintained and basidia appeared within 14 hours to slightly over 3 days. Germination of teliospores was also obtained in Van Tieghem cells set up in the way described by Ashworth (1). The basidiospores germinated equally well on the cover glass in a thin film of moisture or on a plain agar film. With bean agar the results were not so satisfactory; normal germination did not take place and grotesque bodies formed on germination of the spores. Basidiospores of the species of Milesia described in this paper are slightly attenuated at the point of detachment from the sterigma of the basidium. A very conspicuous oil drop is usually present in the basidiospore. Basidiospores of M. Polypodii (Plate 176, Fig. 1) tend to be more spherical than those of M. Scolopendrii (Fig. 2) and M. Kriegeriana (Fig. 3). A table of measurements made from living basidiospores is given below: Species Shape of of rust basidiospore Breadth Length Milesia Polypodii Spherical 8-12 u 9-12 u Milesia Scolopendrii Elongated 5-7 wu 10-12 u Milesia Kriegeriana Elongated 5-8 wu 10-12 u In normal germination of basidiospores a germ tube is usually sent out laterally from the spore. This tube may be only a few microns long, ending in a short, bulbous, conidial-like body; or it may grow to a length of more than 90 u (Fig. 4). Occasionally the contents of the basidiospore pass into the conidial-like body which increases in size as the protoplasm enters it, and the basidiospore wall collapses. A photo- micrograph of a stained preparation showing basidiospores and basidio- spore germination of M. Polypodii is shown in Fig. 5. Short germ tubes were put out two days after basidiospores were set to germinate and in from three to four days the tubes had attained a length of more than 90 u. No further development was observed to occur when experiments were left much longer than four days. Occasionally, in addition to the usual long germ tube extending from one side of the basidiospore, a TABLE I. RESULTS OF SUCCESSFUL INFECTION No. of No. of needles First First Peridermia Date of appearance appearance of rst Date needles with peri- inoculation Host of lesions spermogonia ruptured harvested infected dermia g 27. 1V.34 Abies alba 26 days 26 days 75 days 94 days rd us| 27.1V.34 Abies alba 25 days 25 days 67 days 70 days 14 22 2.V.34 Abies alba 21 days 21 days 89 days 101 days 5 mee, 30. V 1.34 Abies alba 23 days 23 days — — Tree wilted before 235 peridermia were fully a 3 develo Es.| 4.V1.34 Abies concolor 29 days 29 days 75 days 75 days Peridermia 8 insect eaten. < 2.V.34 Abies alba 26 days 26 days 89 days 101 days s=2] 28.V.34 Abies concolor 21 days 25 days — 82 days Peridermia Sv insect eaten S$s mostly immature. g&] 1.V1.34 Abies concolor 26 days 26 days gs 8 588 30.1V.34 Abies alba 23 days 23 days — 62 days 4 Tree wilted before aS peridermia developed. S's = 2.V.34 Abies alba 21 days 21 days 99 days 108 days 28 WOLAIYOdUV GIONYY AHL JO TVNYNOL IIAX “IOA] HUNTER, LIFE HISTORIES OF MILESIA 1936] ¢ @ skepgg skep TZ skepgg step yz skep Zz — sarpuvsb saiqp reAtz | se Oc €2 skep¢g skepog skep fo skep og skep gg — 40J02u09 saigy bf A'8Z ge . 86 se shep zg skep gg = skep fg skep [zs Skep [Z 400909 Saigy bEA'8Z [SS ‘paivaddesip yeay — lis — = skep [z skep [z pgp saiqp PEA ge ‘Ua}va JOVSUT =— — — — sXep [Z sep [Zz Dq}D saiqyv PE TAT _ E ZI 61 = Shep zg = sAep pp skep [Z_— sdep [Z sipupdb saiqy bf A'6 ¥ 8 €1 sdepQg shep gp skep se skep][z skep [Zz — sipuvsB sargy rea |S ‘dojaaap S28 jou pip eIullepllag — — — —_ sXep pz = sdep $7 sipuvsb saigp re A’9 ness pl +t = =6Shep cg — — skep 6p = skep7Z skep ZZ = 40j02u02 saiqy = HETA'S $8 : S 9 sAep ¢g = Shep fo SKEP 6F mad _ sh ode | POAC § PIWdap pa}IjUE pajsoArey peinqdna PIWJeplied eruoz0useds SUOIS2] JO 1s0H] uoljR|NDOUr -led YyIM Sa[psou ajyeq jsiy, jo aouvivodde jo aduvievadde aduesvedde joayrqg — sa[paau = JO “ON PIULIIPLI9 J Is WI WIJ JO "ON (panunu0s) NOILOAANI INASSAOQINS AO SLIASAN ‘I AIAVL 30 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XVII bulbous body was put out from the other side of the spore (Fig. 4). Normal germination of basidiospores was as high as 90 per cent. The basidiospores stuck so fast to the surface on which they fell, regardless of whether they alighted on agar or on the cover glass in a thin layer of moisture, that they remained on the slides without requir- ing the use of egg albumen or other adhesive to secure them during the process of staining in iron alum haematoxylin. A study was made of stained germinated basidiospores and it was found that from two to four nuclei were in the germ tubes (Fig. 6). Abies alba Mill., A. concolor Lindl. & Gord., A. grandis Lindl., A. Nordmanniana (Stev.) Spach, A. Veitchii Lindl., A. firma Sieb. & Zucc. and Picea pungens Engelm. were used for inoculation purposes. ‘Trees, three and four years old, were secured through the Forestry Commis- sion, London, and through reliable nurseries. Some of these were planted in pots and placed in greenhouses at the Chelsea Physic Garden, London, and some in the open at the Biological Station, Imperial Col- lege, Slough, Bucks. Hubert’s tube method as re-described by Faull (3) was used in making inoculations. In some cases the trees were so small that it was impossible to flex the short branches without breaking them, so in connection with these trees the teliospore inoculum was suspended from above so that the basidiospores fell directly upon the upper rather than the lower surface of the leaves. Experiments carried on in both ways gave successful results. Large celluloid tubes stuffed with wet sphagnum at either end to ensure a moist atmosphere were used to entirely enclose the smaller trees. Of all the species used, A. alba, A. concolor and A. grandis only gave positive results. The in- fected needles of A. concolor were slightly distorted but those of A. alba and A. grandis were not (cf. Faull, 3). Successful infection results were obtained by the writer (7) for the ‘following species:—(1) Milesia Scolopendriti on Abies alba and A. con- color; (2) M. Polypodii on A, alba and A. concolor; (3) M. vogesiaca on A, alba; (4a) M. Kriegeriana (from Dryopteris spinulasa) on A. alba, A. concolor and A. grandis; (4b) M. Kriegeriana (from Dryop- teris Filix mas) on A, alba, A. concolor and A. grandis. Data on the experiments are assembled in Table I. One fir tree, simultaneously and under what were essentially identical conditions, bore four successful infections—each from a different rust species. The variations shown with regard to the periods of time that elapsed between inoculation and the first appearance of spermogonia and peridermia were similar to those recorded in the tables for the respective species. Many more inoculations than those recorded were later set up between June 6 and July 6 but no visible infection was observed. 1936] HUNTER, LIFE HISTORIES OF MILESIA 31 TABLE II. Milesia Scolopendrii from Abies to Various Ferns Time from inoculation Fern From Date of Date to appearance inoculated Abies inoculation harvested of uredinia Results Scolopendrium A. alba 7.VI1.34 30.VIII.34 54 days Uredinia vulgare s ri 7.V1I.34 30.VIII.34 54 days Uredinia 7 _ “ 24.VI1.34 15.1X%.34 53 days Uredinia 24 : “ “ 24.VIL.34 = 30. VIII.34 37 days Uredinia “ a “ VLVITIL.34 =18.1X.34 38 days Uredinia a ag “ TLVIIL.34 ) =18.1X.34 38 days Uredinia Blechnum « 24.VI1.34 = 19.X.34 — No infection Spicant Dryopteris spinulosa eo "11.VIL.34 = :19.X.34 ~ No infection dilatata Polystichum a 7.VII.34 = 19.X.34 — No infection angulare a . « -11.VIIT.34 = 19.X%.34 _- No infection TABLE III. Milesia Polypodii from Abies to Various Ferns fob me from oculation Fern Fr Date of Date my appearance inoculated Abies inoculation harvested of uredinia Results Polypodium A.alba 30.VI1.34 15.1X.34 47 days Uredinia are ea « « — 30.VIL.34 15.1X.34 47 days eons «4 “ “11, VIIL.34 18.X.34 68 days. Uredinia ue “ —‘11.VIIL.34 19.X.34 69 days No uredinia a“ a ~ a Polystichum 11.VIII.34 19.X.34 — No infection angulare n a ~ ~ Scolopendrium 30.VII.34 19.X.34 _ No infection vulgare JOURNAL OF THE ARNOLD ARBORETUM [voL. XVII TABLE IV. Milesia Kriegeriana (from Dryopteris Filix mas) from Abies to Various Ferns Time from a Fern From Date of Date to appear inoculated Abies inoculation harvested of iredinis Results Dryopteris A. con- 30.VI1.34 = 15.1X.34 47 days Uredinia Filix-mas color = « « 30.VII.34 = 15.1X.34 47 days erie nsect eaten = « 18.VIT1.34 15.1X.34 28 days Uredinia Drvyopteris « « 30.VI1.34 ©9330. VIII.34 31 days Uredinia spinulosa “ «© -18.VII1.34 = 19.%.34 — No infection TABLE V. Milesia Kriegeriana (from Dryopteris) from Abies to Various Ferns inoculation Fern From Date of Date to appearance inoculated Abies inoculation harvested of uredinia Results A. From D. spinulosa and D. spinulosa dilatata Dryopteris A. gran- 30.V1.34 31.VIII.34 64 days Uredinia Filix-mas dis Dryopteris A. gran- 30.V1.34 31.VIII.34 64 days Uredinia, spinulosa dis very slight infection B. From D. spinulosa dilatata Dryopterts A. alba 27.V1.34 31.VIII.34 — No infection spinulosa dilatata . . A. con- 30.VII.34 31.VIII.34 32 days Uredinia color “ “ A. con- 30.VII.34 19.X.34 — No infection col Dryopteris A. con- 30.VII.34 31.VIII.34 32 days Uredinia Filix-mas color Dryopteris A.con- 30.VII.34 15.1X.34 47 days Uredinia spinulosa color intermedia 1936] HUNTER, LIFE HISTORIES OF MILESIA 33 Ferns for inoculation experiments were procured from Perry’s Hardy Plant Farm and from the field in Devon. Those from Devon were obtained between April 1 and 23. An endeavor was made to select plants that were not already rusted. All the overwintered fronds were cut away fairly close to the underground stem but the young unfolded fronds attached to the stem were left untouched. They were planted in pots at the Biological Station, Slough, in a shaded part of the garden well removed from firs. No rust was observed on any of the ferns as late as August 19, so it is thought that if the ferns from Devon bore natural infection there should have been evidence of it before that date. No naturally occurring ferns bearing rusts were found at the Biological Station. Aéciospores from culture experiments were used in inoculating vari- ous ferns with the result that uredospores were obtained for the follow- ing species—(1) Milesia Scolopendrii on Scolopendrium vulgare; (2) M. Polypodii on Polypodium vulgare; (3a) M. Kriegeriana (from Dryopteris spinulosa) on Dryopteris Filix-mas, D. spinulosa, D. spinu- losa dilatata and D. spinulosa intermedia (Muhl.) Underw.; (3b) M. Kriegeriana (from Dryopteris Filix-mas) on Dryopteris Filix-mas and D. spinulosa. Data on the experiments are assembled in Tables II to V. The successful culturing of Milesia Scolopendrii, M. Polypodii, M. vogesiaca and M. Kriegeriana on their alternate hosts now makes pos- sible a completion of the description of these rusts. The uredinial (2, 4-6) and telial (2) phases have already been adequately described. Heretofore the spermogonial and the aecial phases of M. Scolopendrii, M. Polypodii and M. vogesiaca have remained wholly unknown; and the same phases of M. Kriegeriana have been described by Mayor (14) from field collections only, a part of the material of which was used in making successful cultures on the fern host—material not free from the uncertainty of intermixtures. The spermogonia and the aecia of these four species are described below from the writer’s culture material on Abies, that is, from what is known to be authentic, representative material. (1) Milesia Scolopendrii O. Spermogonia on needles of current season, epiphyllous and hypo- phyllous, immersed, abundant, inconspicuous, colorless, plane, hemi- spherical to slightly flask-shaped in sectional view, subcuticular, 120- 228 uw broad by 100-188 u high; spermatia hyaline, narrowly elliptical, 1.5-2.0 K 4-5 u. 34 JOURNAL OF THE ARNOLD ARBORETUM [vVOL, XVII I. Aecia hypophyllous on needles of current season, in two irregular rows, one on each side of midrib, white, cylindrical, 0.4—0.5 mm. in diameter by 0.7-1.5 mm. high; ese tlcire colorless, delicate, rupturing at the apex; peridial cells polygonal, elongated vertically, overlapping, in a single layer, 20-36 28-56 ut, with outer walls smooth, and inner walls finely and densely warted, the warts arranged in elevated, short lines; aeciospores ellipsoid, ovoid or globose, mostly elongated, white, 22-44 28-48 u, very densely and rather coarsely verrucose with warts irregular in outline, tapering to a blunt point and somewhat decidu- ous; walls hyaline, thin, about 1 p thick. (2) Milesia Polypodii O. Spermogonia on needles of current season, epiphyllous and hypo- phyllous, immersed, abundant, inconspicuous, colorless, plane, hemi- spherical to slightly flask-shaped in sectional view, subcuticular, 120— 228 uw broad by 105-194 uw high, usually broader than high; spermatia hyaline, narrowly elliptical, 1.5-2.0 & 4-5 u. I. Aecia hypophyllous on needles of current season, in two irregular rows, one on each side of midrib, on slightly yellowish discolored por- tions of affected needles, white, cylindrical, 0.5-0.7 mm. in diameter by 1.0-1.5 mm. high; peridium colorless, delicate, rupturing at the apex; peridial cells polygonal, elongated vertically, overlapping, in a single layer, 22-42 >< 28-60 u, with outer walls smooth and inner walls with elevated, coarse, short, irregularly oriented ridges; aeciospores ellipsoid, ovoid or globose, mostly elongated, white, 20-36 28-54 u, densely and rather coarsely warted, warts irregular in outline, tapering to a very blunt point and somewhat deciduous; walls hyaline, thin, about 1 u thick. (3) Milesia vogesiaca O. Spermogonia on needles of current season, epiphyllous and hypo- phyllous, mostly epiphyllous, very abundant, immersed, inconspicuous, colorless, plane, 154-241 uw broad by 168-214 wu iat: spherical to slightly flask-shaped in sectional view, subcuticular; spermatia hyaline, narrowly elliptical, 1.5-2.0 « 4-5 wu. I. Aecia hypophyllous on needles of current season, in two irregular rows, one on each side of midrib on slightly yellowish discolored portions of affected needles, white, cylindrical, 0.5-0.7 mm. in diameter by 0.6-1.0 mm. high; peridium colorless, delicate, rupturing at the apex; peridial cells polygonal, elongated vertically, overlapping, in a single layer, 20-32 & 32-48 u, with outer walls smooth and inner walls warted or with elevated, coarse, short, irregularly oriented ridges; aeciospores 1936] HUNTER, LIFE HISTORIES OF MILESIA 35 ellipsoid, ovoid or globose, mostly elongated, white, 24-30 32-46 u, densely warted, warts very irregular in outline and tapering to a very blunt point, somewhat deciduous; walls hyaline, thin, about 1 p thick. (4) Milesia Kriegeriana O. Spermogonia on needles of current season, epiphyllous and hypo- phyllous, mostly epiphyllous, numerous, irregularly scattered, incon- spicuous, colorless, plane, abundant, immersed, hemispherical in sec- tional view, subcuticular, 98-168 u broad by 94-168 u high; spermatia narrowly elliptical, 1.5—-2.0 « 3.5—5.0 wu. I. Aecia hypophyllous on needles of current season, in two irregular rows on slightly -yellowish discolored portions of affected needles, sub- circular, ellipsoid or compressed laterally in transverse section, cylindri- cal, 0.3-0.8 mm. in diameter by 0.5-1.3 mm. high; peridium colorless, delicate, rupturing at the apex; peridial cells polygonal, elongated verti- cally, overlapping, in a single layer, 16-52 32-68 u, with outer walls smooth and inner walls with fine elevated ridges, irregularly oriented ; aeciospores ellipsoid, ovoid or globose, mostly elongated, white, 20-36 22-48 u, finely warted, warts tapering to a blunt point, somewhat decidu- ous; walls hyaline, thin, about 1 p thick. Mayor (14) refers to the late appearance of the aecia from natural infection in the field and believes that the younger needles of Abzes are not susceptible to Milesia Kriegeriana. It would seem from the writer’s experience that the late appearance of the aecia of M. Kriegeriana is because of the long period of development of the fungus from the time of inoculation until the peridermia make their appearance and that young leaves are more susceptible to the rust than older ones. A comparison of the foregoing descriptions shows that these rusts can- not be separated on morphological differences in their peridermia but, on the other hand, that the spermogonia do afford some help. The spermo- gonia of Milesia Scolopendrii and M. Polypodii are hemispherical to slightly flask-shaped while those of M. vogesiaca are spherical to slightly flask-shaped, and are on the average much larger. The spermogonia of Milesia Kriegeriana are hemispherical in shape and are on the average consistently smaller than those of the other three species. The sub- cuticular areas directly overlying the spermogonia of M. Kriegeriana, M. Polypodii and M. Scolopendrii are relatively greater than that in M. vogesiaca. Indeed in M. vogesiaca the area is so small that in lateral vertical sections the spermogonium appears to be subepidermal. The spermogonia of M. Scolopendrii and M. Polypodii are much alike in form and size, although those of the latter are on the average a little larger. The detailed data are recorded in a paper to be published soon. 36 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xvi SUMMARY 1, By means of controlled cultures the O I stages of Milesia Scolo- pendriu, M. Polypodti, M. vogesiaca and M. Kriegeriana have been obtained for the first time. O I hosts II III hosts (1) Milesia Scolopendrii Abies alba, Scolopendrium vulgare A. concolor (2) Milesia Polypodi Abies alba, Polypodium vulgare A. concolor (3) Milesia vogesiaca Abies alba Polystichum angulare (4a) Milesia Kriegeriana Abies alba, Dryopteris spinulosa, (from Dryopteris A. concolor, D. spinulosa dilatata, Spinulosa) A. grandis D. spinulosa intermedia, D. Filix-mas (4b) Milesia Kriegeriana Abies concolor Drvyopteris spinulosa, (from Dryopteris D. Filix-mas Filix-mas ) 2. The spermogonia and the aecia of M. Scolopendrii, M. Polypodii and M. vogesiaca are described for the first time. The spermogonia and the aecia of M. Kriegeriana are described for the first time from culture material, 3. The spermogonia of M. Kriegeriana and M. vogesiaca are dis- tinguishable by their form and size from one another and from those of M. Scolopendrit and M. Polypodii. The spermogonia of M. Scolopen- drii and M. Polypodii are similar in form and size. ACKNOWLEDGMENTS Studies recorded in this paper form part of an investigation carried out by the writer under the direction of Professor J. H. Faull of the Arnold Arboretum, Harvard University. This particular phase of the work was made possible by a fellowship granted by the Canadian Fed- eration of University Women and was mostly done in England under the sponsorship of Professor William Brown of the Imperial College of Science and Technology, London University. Sincere thanks are ten- dered to Professors Brown and Faull. The author also gratefully acknowledges various services rendered by Mrs. L. Allen, Mr. D. Blockey, Dr. G. D. Darker, Dr. E. Milton, Mr. H. B. S. Montgomery, Dr. Muriel P. Hall, Dr. P.O’Connor, Mr. A. H. G. Alston, Mr. W. Hales, Mr. W. Williams, Mr. J. Ramsbottom, Mr. A. D. Cotton, Miss E. M. Wakefield, Dr. E. J. Butler, Mr. E. W. Swanton, Mr. J. Ogilvie, the Torquay Museum of Natural History and the Forestry Commission of Great Britain. Jour. ARNoLD Ars. VoL. XVII PLATE 176 Mites1a Potypopit, M. ScoLopENpDRII, M. KRIEGERIANA AND M. VOGESIACA FULL-TONE— MERIDEN oe OO a ee 1936] HUNTER, LIFE HISTORIES OF MILESIA 37 Ne oO. © NW np w bo LITERATURE CITED AsHuwortH, D. Puccinia Malvacearum in monosporidial culture. (Trans. Br. Myc. Soc. 16: 177-202. 1931. FauL., J. H. Taxonomy and geographical distribution of the genus Milesia (Contrib. Arnold Arb. 2: 1-138. The biology of Milesian rusts. (Jour. Arnold Arb. 15: 50-86. 1934. Grove, W. B. The British rust fungi. (Cambridge, 1913.) — —— The British species of Milesina. (Jour. Bot. 59: 109. 1921.) ——— Mycological notes VI. (Jour. Bot. 59: 311-313. 1921.) Hunter, L. M. preliminary note on life history studies of Euro- pean species of Milesia. (Jour. Arnold Arb. 16: 143. 1935. ———— The genus ie in Great Britain and Ireland. (Trans. . Kamel, S. Note Sve na vogesiaca Sydow on Polystichum ana Miyabe et Kudo, A. firma Sieb. et Zucc. and A. sachalinensis ast. yas, Sapporo Nat. His. Soc. 11: 141-147. 1930. ———. A new species of Milesina Sig yee on ae vulgare L Cras Sapporo Nat. His. Soc. 12: 27-33. On new species of heteroecious yee rusts. teoas Sapporo Nat. a Soc. 12: 161-174. 1932. . KieBpaAun, H. Kulturversuche mit Rostpilzen. (Zeitsch. Pflanzenkr. 26: 357.277, 1916. Maanus. P. Weitere ee uber die auf Farnkrautern auftre- tenden Uredineen. (Ber. Deutsch. Bot. Ges. 19: 578-584. 1901.) . Mayor, Euc. Notes be Le VIII. (Bull. Soc. Neuchat. Sci. Nat. 58: 23-26. 1933.) HITE UCHANAN. Note on the zoology and botany of Glen Tilt. (Scottish Naturalist, 4: 160-163. 1877-8. ) EXPLANATION OF PLATE 176 ig. 1. Basidiospores of Milesia Polypodii B. White. Note variation in form and germination. x 495. ioc: Basidiospores of Milesia apts (Fuckel) Arth. Note variation in form and germinati 3. Basidiospores of Milesia Kriegeriana (Magn. ) Arth. from Dry- opteris Filix- ions ia Schott. Note variation in form and ermination. 4. Germinating Sed cere of Milesia Polypodii. Note the vari- ous stages in germination and the formation of a secondary spore. 5. . 5. Germinating basidiospores of Milesia Polypodii. X 200. . 6. Germinating basidiospores of Milesia Polypodii. Note nuclei in germ tubes. he . 7. Young tree of Abies alba Mill. infected with four species of Milesia, namely,—(1) M. Polypodii on upper left branch (inocu- lation experiment No. 7); (2) M. vogesiaca (Syd.) Faull on up- per right branch (inoculation experiment No. 8) ; (3) M. Krieg- ertana on lower left branch (inoculation experiment No. 9); (4) M. ehatia’ on lower right branch (inoculation experiment No. 10). xX %. LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, HARVARD Lwiv ERSITY. 38 JOURNAL OF THE ARNOLD ARBORETUM [vVoOL. Xvi A CONVENIENT SAND-CULTURE APPARATUS Rosert W. Warp With plate 177 and one text figure One of the essential requirements in nutritional studies on higher plants is a culture apparatus in which replications of an extended series may be studied simultaneously. It is also always desirable, if not actu- ally necessary, that the time expended in the routine operations of water- ing and feeding the plants be reduced to a minimum. The most im- portant consideration, however, from the physiological point of view is that the component parts of the apparatus with which the plants, medium, or nutrient come in contact be constructed of materials which are chemically resistant, as well as non-porous, and capable of affording leak-proof connections. An apparatus which seems to satisfy these requirements better than any heretofore described has been designed and put into use by the writer in a study on physiological disorders of apple trees. The apparatus consists primarily of an inverted, bottomless, one- gallon jug connected to an inverted, standard, one-half-gallon jug by means of rubber- and glass-tubing (Fig. 14). The bottomless jug B, in which the plants are grown in white silica sand, is of standard dimensions but lacks both bottom and handle and has the neck grooved so that the rubber stopper may be wired in position. It is glazed externally accord- ing to standard requirements with a white glaze and internally with a special acid-resistant glaze. The sand found satisfactory for apple cul- ture is a mixture of 5 parts of Columbia No. 4, a coarse, angular, silica sand, and 6 parts of Ottawa No. 20, a silica sand having almost spherical grains. The lower part of the jug, as shown in the illustration, is pro- vided with the Columbia No. 4 sand to ensure good drainage. The sand is prevented from entering the glass tubing by the use of a small circle of Monel metal screening, No. 32 mesh, laid against the rubber stopper. The nutrient solution is placed in the one-half-gallon jug A, from which it is flooded into the sand by elevating this jug to the proper height and into which it is later drained by lowering again. The connecting tubings H are of heavy-walled gum rubber and standard Pyrex glass. As much of the latter is used as is conveniently possible (Plate 177). The tubing J, which facilitates movement of air in the nutrient jug, may also serve as a suction tube to which a very weak suction pump may be attached 1936] WARD, SAND-CULTURE APPARATUS 39 if more aération of the sand culture medium is desired than can be obtained by normal drainage of the nutrient solution. As many of these pieces of apparatus as are necessary or can be used conveniently are assembled on a wooden stand, as shown in Fig. 1, in which the culture jugs B...B are placed in twin beds C, C and the nutrient jugs A...A in a movable rack E. The writer was thus enabled to set up twenty-four cultures in triplicate, seventy-two cultures in all, on one greenhouse bench. XX C x» Go PPT Figure 1. SANp CULTURE ApparRATus. 1, 2, 3. Working drawings of stand and assembled jugs. 4. Vertical section of jugs in flooding position. The wooden stand which holds the jugs has been designed to fit on an ordinary greenhouse bench (Fig. 11:3; Plate 177). It is composed of twin beds C, C resting on horses D...D, placed about four feet apart, and a movable median rack E, which, in the raised or flooding position, rests on removable horses of a predetermined height, placed about eight feet apart (Plate 177) and which later, in the lowered or draining posi- tion, comes to rest on two stout spiral-spring blocks (Fig. 11?-F), The beds are fastened firmly to their respective horses and the ensemble is made immovable by use of the upright structures G...G (Fig. 13; Plate 177), which also serve as guides for the rack E. The lumber used 40 JOURNAL OF THE ARNOLD ARBORETUM (VOL. Xvi in the construction of the stand is ‘‘western fir,” 2 inches by 4 inches, and white pine, 1 inch by 4 inches. All exposed parts are painted with a good white waterproof paint. Although details of construction of the stand are obvious in the accompanying working drawings (Fig. 1) there are several important features to be noted. First, the scantlings supporting the inverted bottomless jugs B...B in the beds C, C are mitered instead of being inserted at right angles or parallel to the sides of the bed. This makes it possible to adjust the beds to the proper width so that the ensemble of beds and rack may be accommodated on the greenhouse bench. e mitered angle of the scantlings depends naturally on the width of bed desired. The scantlings are also spaced accurately, both in the beds and rack, so that the inverted jugs may rest snugly on their shoulders between adjacent pieces. Secondly, an intermediate portion of the upright part of the beds and the entire bottoms are covered with gal- vanized wire netting of a coarse mesh. This permits the bottomless jugs in the beds to be packed in wet sphagnum moss, an arrangement which tends to keep the cultures at a uniform temperature due to the regulatory properties of evaporating water. Thirdly, the horses D...D supporting the beds must be of sufficient height to allow proper and adequate drainage of the cultures. Finally, it is imperative that at least two sets of heavy spiral springs F...F, of the shock-absorber type be placed in such a position under the rack E as to give the latter maxi- mum support when fully loaded. This precaution is necessary to pre- vent shock and probable damage to the glass tubing when the rack with its heavy load of nutrient jugs and attached glass- and rubber-tubings are lowered to the draining position. — The operations of raising and lowering the rack and nutrient jugs are accomplished by means of two quadruple, self-locking block and tackle sets. These are attached to the rack at the proper supporting points and to a solid superstructure directly above the rack, the slings in each case being of Manila rope or some equally reliable uecanl, The single operation of raising or lowering the rack can be accomplished by one operator and occupies about three minutes. The height to which the rack should be raised for proper flooding of the cultures is determined in the following manner. Two liters of water are placed in the nutrient jug A, which is then set in the rack and con- nected to its respective culture jug B by means of the rubber- and glass- tubing H described above. The rack is then raised until the water level reaches exactly to the surface of the sand (Fig. 14). This height is carefully recorded and light-weight wooden horses are constructed to support the rack when raised to this position (Plate 177). Jour. ArNoLp Ars. VoL. XVII PLATE 177 SAND CULTURE APPARATUS WITH APPLE TREES IN CULTURE. (Photos by A. B. Hatch) a oe oe a oe Bn re a 7 a —— 7 a 7 — — oe a _ _ - 7 1936] WARD, SAND-CULTURE APPARATUS 41 At this time tests are made with several of the units to determine the average volume of water retained in the sand after drainage. Later, when the culture solutions are made up, this volume is taken into consideration. The routine daily care of the cultures resolves itself briefly into the following operations. The rack is raised to the flooding position and, after a lapse of a few minutes to allow adjustment of the liquid level, distilled water is added directiy to the sand to compensate for loss by evaporation and transpiration, until the original level at the surface of the sand is reached. Thus the original volume of the culture solution is retained throughout the course of the experiment. When the cultures have been flooded for a sufficient length of time the rack is lowered, allowing the sand to be drained and aérated naturally at the same time. For purposes of convenience, the writer placed a carboy of distilled water on a high stand at the rear of the beds. Leading from this, along either of the outer sides of the beds, were two lines of 12 mm. glass-tubing provided with several outlets (Plate 177). A short length of rubber- tubing was attached successively to these outlets and by this means the complement of distilled water was added to the cultures directly from the carboy. The frequency with which the culture solutions should be renewed depends upon the technique of the operator and the plants in culture. Since all of the nutrients can be accounted for, either in the plants or in the solutions, it is possible to make periodic analyses of the latter and adjust the time of renewal according to the interpretation of the results obtained. This apparatus was designed and constructed in the laboratory of Professor J. H. Faull of the Arnold Arboretum, Harvard University, in connection with investigations carried on under his direction. This research was made possible through scholarships from the Harvard Forest and the Arnold Arboretum. The writer wishes to express his sincere thanks to Professor Faull for the facilities afforded and his con- tinued interest. Much credit is due Professor P. R. Gast of the Harvard Forest, who contributed the basic principle of the apparatus devised from his own researches and assisted in many ways with time and equipment. Grateful appreciation is expressed to my former labora- tory associates, Drs. A. B. Hatch and J. D. MacLachlan for their many constructive criticisms. LABORATORY OF PLANT PATHOLOGY, RNOLD ARBORETUM, HARVARD UNIVERSITY AND Division OF HorTICULTURE, ENTRAL EXPERIMENTAL FARM, OTTAWA, CANADA. 42 JOURNAL OF THE ARNOLD ARBORETUM [voL. XvII CHROMOSOME COMPLEX AT PREMEIOTIC ANAPHASE AND MEIOTIC METAPHASE Haic DERMEN With two text figures A NUMBER of workers have reported in the case of plants that the individual chromosomes at anaphase are at least double. In this labora- tory we have seen this duality in the root-tips (Dermen 1933), during all stages of microsporogenesis in Rhoeo, and in the telophase stage of embryo sac tetrad nuclei in Lilium regale. In the present paper we are presenting conclusive evidence indicating that this duality also exists in Fic. 1. Ficure 1. Preme1otic ANAPHASE FROM RHOEO DISCOLOR.—FIGURE 2. Metotic First METAPHASE FROM A Hysrip TRADESCANTIA (T. REFLEXA x T. PALUDOSA). the last premeiotic anaphase chromosomes which are destined to become meiotic leptotene threads. In addition to this the meiotic metaphase chromosomes were studied and were found to be structurally eight-parted. The anthers of a young bud which is the third in series from that in the reduction stage in Rhoeo discolor were used for the study of pre- meiotic anaphase (Fig. 1). The cells of these anthers are for the most part in early prophase of meiosis but a few may be found in somatic metaphase or anaphase stage. The material showing the structure of meiotic first metaphase chromosomes (Fig. 2) was from a_ hybrid Tradescantia, T. reflexa * T. paludosa. 1936] DERMEN, CHROMOSOME COMPLEX 43 MATERIAL AND TECHNIQUE The anthers in either case were smeared and pretreated for 5 to 10 seconds in 30% alcohol, 40 cc. in a staining jar, to which was added approximately 4 to 6 drops of ammonia water (26%), and the smear was then flooded with aceto-carmine 2 or 3 times for about 30 seconds. The slide was drained, any residual anther material that remained on the slide was removed, and a drop of aceto-carmine was added and a cover glass placed on the smear. The material was first located and studied, and if the preparation was successful the slide was gently heated over an alcohol lamp until small bubbles began to form. The cover glass was then pressed using a piece of filter paper for the purpose. To make these temtporary preparations permanent either McClintock’s acetic-alcohol or Metz’s acetic-alcohol-xylol method may be used. For a more permanent smear using the crystal violet-iodine stain, the following method is prescribed: Smear and pretreat as described above and place the slide, smear side down, in a small shallow dish containing a few drops of % to 34 strength Flemming’s modified solution for %4 to 1 minute. Rinse in water in a staining jar for one to two minutes, then stain in crystal violet for 15 seconds or longer depending upon the source of the material. Some plant smears may require over one minute of staining. If material so stained is thoroughly washed with xylol future destaining is prevented and the stain is as permanent as haemotoxylin preparations. DESCRIPTION A photomicrograph of an anaphase genom at last premeiotic stage from Rhoeo discolor is shown in figure 1. It is one of the two anaphase groups seen from a slightly oblique polar view. It shows on the lower right the short arm of a chromosome pointing down. From this posi- tion the structure of this arm could be clearly made out indicating unmistakably that it is split. The halves seem to be independently coiled and are not intimately associated. Where these chromosomes are twisted they appear as if constricted. A meiotic first metaphase from a hybrid Tradescantia (T. reflexa T. paludosa) is shown in figure 2, also a photomicrograph. A bivalent chromosome is seen in a horizontal plane. This pair of chromosomes at the connection point shows two kinds of split, one major and one minor. This type of opening was not confined to the rod-bivalents alone. In other cells in the ring-bivalent chromosomes at both ends of their connections, these double, large and small, diamond shaped open- ings could be observed, but for the purpose of illustration the rod- bivalent type was found best, for it is easily flattened out by pressure on the cover glass to show this feature. 44 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII DISCUSSION The duality in the premeiotic chromosomes has been shown by Kauf- mann (1926) in plant material. In animals, McClung (1928) and Robertson (1931) have demonstrated this duality of chromosomes at a corresponding stage. Koshy (1934) has illustrated the duality of pre- meiotic anaphase chromosomes and leptotene threads, though it appears from his drawings that he may have confused leptotene with pachytene stage. Nebel (1932 and 1933), Stebbins (1935) and Goodspeed, Uber and Avery (1935) are in agreement that telophase chromosomes are four- parted. It is indicated that the leptotene threads, although optically single, are actually four-parted. Goodspeed et al. believe further, by estimation, that the meiotic metaphase chromosomes are sixteen-parted, thus homologizing with Drosophila salivary chromosomes which are suggested to be sixteen-parted. Huskins (1932) and Huskins and Smith (1934 and 1935) have ad- mitted the duality of anaphase chromosomes elsewhere but deny it in the premeiotic anaphase. They also have reported eight-parted meiotic metaphase chromosomes. Darlington (1935) on the other hand has denied the duality of anaphase chromosomes in all regions. He believes that anaphase chromosomes are sing!e in both somatic and premeiotic phases because of the fact that leptotene threads appear to be single. Our own evidence (Figs. 1 and 2) shows conclusively that premeiotic anaphase chromosomes are at least two-parted and eight-parted at mei- otic metaphase. The indication is that mitotic and meiotic phases are essentially of the same fundamental nature, with this difference, that instead of each chromosome remaining as a unit and perpetually dividing as in mitosis, in meiosis similar chromosomes (homologues) pair to- gether (barring the crossing-over phenomenon from the discussion) and during the first metaphase move to opposite poles. In mitosis homo- logues do not form intimate association while in meiosis their associa- tion is quite intimate and complex. The mitotic chromosomes are four- parted at metaphase, in meiosis they are eight-parted in the bivalents, and correspondingly there are four in each half of the bivalents. I tried to analyze the same situation of duality in Lilium ovules where as in any other plant there is usually only one cell that becomes meiotic. The preparations I made (sectioned material) were of a somewhat advanced stage.- However, there is one obvious thing that should be considered here, that this meiotic cell is a product of a somatic cell division. Here one cell only becomes meiotic while its sister cell con- tinues the-somatic cycle. It is hard to imagine if the anaphase chromo- 1936] DERMEN, CHROMOSOME COMPLEX 45 somes in the somatic cell are split, as Huskins admits (1932), how the anaphase chromosomes of one such group become normally split while those in the opposite group remain unsplit to justify Huskins’ assump- tion that leptotene threads are single by origin. In both sister cells chromosomes are either split or unsplit since these halves are supposedly mirror images of each other. If the chromosomes were single in this stage there should be two cells alike and both meiotic. However, there is normally only one such cell. The obvious conclusion is that if chromosomes are split in the somatic cells at anaphase elsewhere, they must be split here too and that undoubtedly late premeiotic anaphase chromosomes of the egg-mother-cell are of a dual nature. SUMMARY Conclusive evidence is presented showing that last premeiotic ana- phase chromosomes in microsporogenesis are split and that metaphase chromosomes in pollen-mother-cells are optically eight-parted. It is argued that the same is true in the ovule. A method which enabled us to bring out the coiled structure in meiotic chromosomes is described in some detail. This method involves pre- treatment of smear preparations with ammonia in alcohol. LITERATURE CITED DarLINGTON, C. D. (1935). The old Sige reer Ae and the new analysis of mae aaa ee r. (Ann. 49: 579-586. ) DeRMEN, H. (19 Origin and behavi or of the nucleolus in plants. (Jour. Arnold res 14: 282-323. ) GoopspEED, T. H., F. M. User, & P. Avery (1935). Application of the Altmann freezing- drying technique to plant cytology. III. Chromosome structure in Lilium longiflorum. (Univ. Calif. Publ. Bot. 18: 33-44.) Husxins, C. L. (1932). Observations ‘bearing o on the mechanism of meiosis and crossing over. (Proc. Sixth Int. Congress Genetics, 2: 95-96. —— & S. G. SmitH (1934). Chromosome division and pairing in Fritillaria Meleagris: The mechanism of meiosis. (Jour. Genetics, 28: 6.) ————— ————. (1935). Meiotic chromosome structure in Trillium erectum L. (Ann. Bot. 49: 119-150.) KAuFMANN, B. P. (1926). Chromosome structure and its relation to the chromosome cycle. I. Somatic mitosis in Tradescantia pilosa. (Amer. 0 Kosuy, T. K. (1934). Chromosome studies in Allium. II. The meiotic McCiune, C. E. (1928). ’ Differential chromosomes of Mecostethus gracilis. (Zeit. Zellf. Microsk. Anat. 7: 756-778). 46 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Neset, B. R. (1932). Chromosome structure in Tradescantiae. I. Method and morphology. (Zeit. Zellf. Microsk. Anat. 16: 251-284. —— (1933). Chromosome structure in Tradescantiae. IV. The Rogertson, W. R. B. (1931). Chromosome studies. II. Synapsis in the Tettigidae with special reference to the presynapsis split. (Jour. ree i -_ Physiol. 51: 119-146. STEBBINS, G. (193 ). Chromosome structure and the mechanism of eo hi in ae (Am. Nat. 69: 81.) CyTOLoGIcAL LABORATORY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1936] REHDER, BURRETIODENDRON 47 BURRETIODENDRON, A NEW GENUS OF TILIACEAE ALFRED REHDER With plate 178 Burretiodendron, gen. nov. Flores dioeci, 5-meri, in cymis axillaribus paucifloris bracteatis pedunculatis petiolo multo brevioribus dense stellato-pilosis interdum in apice ramulorum brevium in paniculam brevem congestis; flores masculi breviter pedicellati, basi bracteolis 2 late ovalibus vel 3 ellip- ticis obtusiusculis concavis quam sepala fere dimidio brevioribus extus dense stellato-pilosis glandulis interspersis, intus glabris ante anthesin alabastrum arcte includentibus instructi; sepala 5, libera, anguste elliptica, acuta, extus dense stellato-pilosis, intus glabra basi area oblonga elevata dense glanduligera trientem sepali aequante instructa; petala 5, libera, alba aestivatione contorta, sepalis paulo longiora, unguiculata, lamina late obovata, apice fere truncata et plus minusve erosula et cilio- lata basi in unguem tenuem fere dimidiam laminam aequantem attenu- ata, margine ciliolata excepta glabra; stamina circiter 30, petalis triente breviora, filamentis filiformibus infra medium in phalanges 5 et basi in tubum connatis, antheris lineari-oblongis bilocularibus basifixis, loculis rima longitudinali dehiscentibus; rudimentum ovarii parvum tubo staminali inclusum ovoideum rostratum. Flores feminei non _ visi. Capsula oblonga, stipitata, 5-alata, septicide 5-cocca, alis capsulae papyraceis bilamellatim partitis, coccis verticaliter bialatis indehiscenti- bus monospermis; semen oblongo- obovatum, basin versus attenuatum, apice obtusum et chalaza ovali notata, parce albuminosum; pee ae rectus, cotyledonibus foliaceis latis tenuibus plicatis sacreule recta.— Arbor ramulis hornotinis dense pilis stellatis vestitis, annotinis glabris. Folia alterna, decidua, ovata vel late ovata, acuminata, basi cordata, palminervia nervis basalibus 5—7 et insuper utrinsecus nervis 4—5 angulo angusto a costa media divergentibus et ascendentibus, nervis lateralibus a costa media et a nervis basalibus divergentibus ad marginem ceterum integram in mucronem exeuntibus, utrinque pilis stellatis satis dense, in costa et venis densius vestita, costa nervisque supra fere planis subtus distincte elevatis et trabeculis reteque venularum prominulis conjunctis; petioli satis graciles, dense stellato-pilosi; stipulae caducae, vestigia tantum adsunt. Species unica Chinae occidentalis incola. 48 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xvi Burretiodendron Esquirolii (Lévl.), comb. nov. Pentace Esquirolii Léveillé in Fedde, Rep. Spec. Nov. 10: 147 (1911) ; Fl. Kouy-Tchéou, 419 (1915).— Burret in Notizbl. Bot. Gart. Mus. Berlin, 9: 620 (1926). Eriolaena Esquirolii Léveillé, Fl. Kouy-Tchéou, 405 (1915). Character generis. Petiolus 2.5—5 cm. vel interdum ad 11 cm. longus; folii lamina 6-12 cm. longa et 4-8 cm. lata vel interdum ad 28 20 cm. magna. Inflo- rescentia pedunculo brevi incluso circiter 3 cm. longo; bracteae 7 mm. longae; flores circiter 2 cm. diam.; sepala circiter 10 mm. longa et 4 mm. lata; petala ungue 3 mm. longo incluso circiter 11-12 mm. longa, lamina circiter 7-8 mm. lata; stamina 7 mm. longa antheris 2 mm. longis; ovarii rudimentum 1 mm. longum. Capsulae 3—3.5 cm. longa et circiter 1.5 cm. lata, alis circiter 6 mm. latis; semina 8 mm. longa et 3-4 mm. lata. DISTRIBUTION: Southeastern Yunnan and Kweichou. Cuina. Kweichou:. ouest de Lo-fou (Kouai-kou), J, Cava- leri, no. 2648, Nov. 1905, “arbre moyen, moucilagineux” (holotype of Pentace Esquirolii; photo. in A. A.); same locality, J. Esquirol, no. 817, and Yang-ly, J. Esquirol, no. 2717, both cited in Fl. Kouy-Tchéou under P. Esquirolii; Yang-ly, J. Esquirol, no. 2717, Aug. 1911, “arbre 8-10 métres, fleur blanche” (holotype of Eriolaena Esquirolii; photo. in A. A.); Lohu, 170 m. in lightly shaded ravine, Y. Tsiang, no. 7290, Oct. 3, 1930, “tree 10 m. high, diam. of trunk (breast high) 20 cm., leaves dull green above, light green below, fruit yellowish green” (New York Bot. Gard.) Yunnan: Red River, A. Henry, no. 9572 (New York Bot. Gard.); Red River, near Manhao, A. Henry, no. 9573, “tree 25 ft., flowers all fallen off, young green fruit” (New York Bot. Gard.). This new genus does not seem to be closely related to any of the genera of the Tiliaceae or Sterculiaceae, but is best perhaps placed in the Tiliaceae near Luehea; from the Sterculiaceae it differs in the dry winged fruit dividing septicidally into one-seeded carpels broadly winged all around, and in the androecium which in the latter family has the episepalous stamens sterile or lacking, while in Burretiodendron they are all fertile. With Luehea it agrees in the lack of an androgynophore and in pentadelphous stamens, but differs markedly in the unisexual flowers, in the slender-clawed petals without glandular spot at the base and in the fruit separating into 5 one-seeded carpels. In the presence of a nectary at the base of the sepals Burretiodendron differs from all Tiliaceae. In the fruit separating into one-seeded winged carpels, it resembles Colona, but is easily distinguished by the unisexual flowers in axillary short cymes, the lack of an androgynophore, the clawed petals Jour. ARNoLp Ars. VoL, XVII PLATE 178 2B oP be roe Sogee nok (eer? J J Chol aor. FP care G 6 OU. Det Alfred Rehder ARXHLD Antounrow SY, war BURRETIODENDRON Esgurrotit (Lévl.) Rehder. 7 - - OO Be BS 1936] REHDER, BURRETIODENDRON 49 without nectary and the pentadelphous stamens. Eriolaena in which Lé veillé placed the flowering specimen, differs among other characters chiefly in its ligneous loculicidally dehiscent capsule with many winged seeds. Pentace, to which he referred the fruiting specimen, is easily distinguished by the rather small bisexual flowers in large terminal panicles and the much smaller indehiscent one-seeded fruit. The specimens collected by Henry near Manhao, about 25 kilometers north of the border of Tonkin, agree exactly with Cavalerie’s and Esquirol’s specimens, except that the former have the leaves generally more deeply cordate like those of Tsiang’s specimen. The genus may possibly extend into northern Tonkin and northwestern Kwangsi. I take pleasure in naming this new and very distinct genus in honor of Dr. M. Burret who has made important contributions to the knowl- edge of the Tiliaceae. EXPLANATION OF PLATE 178 1. Fruiting specimen, Y. Tsiang, no. 7290. x 2/5. . 2. Single carpel of fruit from the ventral and dorsal side. x 2/5. Fig. 3. Axillary cyme with one open flower, from Henry, no. 9572. Natural size. 4. Terminal panicle of flower buds, from Henry, no. 9572. Natural size. HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 50 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xvu TWO NEW SPECIES OF GYMNOSPORANGIUM FROM ASIA Ivan H. Crowe. With two text figures Gymnosporangium magnum Crowell, sp. nov. (Fig. 1). Spermogonia foliicolae, subepidermalia. Aecidia hypophylla, cupuli- formia, tumoribus subglobosis insidentia; cellulae peridii rhomboideae, hyalinae, pariete interiore marginibusque subspinulosae, 22.5-30.6 50.0-64.4 ut, plus minusve 24.3 59.2 4; aecidiosporae globosae, pallide brunneae, verruculosae, 24.2-29.0 > 30-32 u, plus minusve 26.1 29.8 w. y H ry tis* s*s) pate G,,2. . §202,2 383 pa 4 = f é & v 17 ¢, V. e®%o. & o ye ee i) i erat’ & Slete la SRR > pur Gees Kio — Fic. 2. IGURE 1. GYMNOSPORANGIUM MAGNUM a —Ficure 2. GyYMm- NOSPORANGIUM LEVE Crowell.—AECIOSPORE ND PERIDIAL CELLS. Hab. in foliis Crataegi sp. in China centrali. Ex Herb. G. Lagerheim, Botaniska Andelning Riksmuseet, Stockholm, Sweden (Smith, no. 917) The specific name is suggested by the unusual size of the aeciospores, a feature in which they are surpassed by G. juniperinum (L.) Mart. only. Gymnosporangium leve Crowell, sp. nov. (Fig. 2). Spermogonia foliicolae, sibepideuans, Aecidia hypophylla, roes- telioidea, tumoribus tumidis insidentia, 1-2 mm. alta, 0.25—0.5 mm. lata; cellulae peridii elongatae, hyalinae, pariete interiore spinulosae, mar- 1936] CROWELL, NEW SPECIES OF GYMNOSPORANGIUM 51 ginibus et pariete exteriore gradatim plus breviter spinulosae, 12.8-18.2 X 80-120 u, plus minusve 15.2 102 uw; aecidiosporae globosae, pallide brunneae, leves, 19.2-22.4 20.8-24.0 u, plus minusve 20.8 xX 22.6 wu. Hab. in foliis Mali Sieboldii (Reg.) Rehd. in China. Ex Herb. G. Lagerheim, Botaniska Andelning Riksmuseet, Stockholm, Sweden (Smith, no, 4221). The material from which this species is described was collected by Smith at Karlong, Szechuan, China, altitude 3200 m., Aug. 27, 1922. The specific name is suggested by the smooth surface of the aeciospore wall, a character of rare occurrence in this genus of rusts. Comparatively few species of Gymnosporangium have so far been described from Asia. One feature common to them, generally speaking, is the occurrence of spinulose embossments on the peridial cells of the aecia. Among the American species this feature is found in G. exiguum Kern only. The specimens which now become the types of Gymnosporangium magnum and G. leve were loaned from the Herbarium of the Natur- historika Riksmuseet, Stockholm, Sweden, through the courtesy of Professor Dr. Gunnar Samuelsson. LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. : _ _ _ oo BS a - Oo Be 7 JOURNAL OF THE ARNOLD ARBORETUM VoLtuME XVII APRIL, 1936 NUMBER 2 NOTES ON THE LIGNEOUS PLANTS DESCRIBED BY H. LEVEILLE FROM EASTERN ASIA’ ALFRED REHDER IN THE COURSE of my revision of the ligneous plants described by Léveillé from Eastern Asia, it became apparent that after once going over the ground in systematic order, I would have to publish a supplement to include those plants which for various reasons, had been omitted in the first account. When I started this work in 1929, after a short sojourn in Edinburgh in the summer of the preceding year, I based it chiefly on notes and photographs I had taken and on duplicates of types generously given to me whenever feasible, by Sir William Wright Smith, Regius Keeper of the Royal Botanic Garden. Later in the course of my work, Sir William kindly offered to send me on loan all the types I had not examined while in Edinburgh; this enabled me to treat the later families more fully. Another reason for the omission of certain species in their proper places was the fact that in many instances, species referred by Léveillé to cer- tain genera, did not belong there but to families previously treated, e.g. Clerodendron Esquirolii which turned out to be a Tacca. There remains also to be cited important additional literature dealing with Léveillé’s plants and not yet mentioned in my previous publications. After having finished my first attempt of an enumeration of Léveillé’s ligneous plants in systematic order from the Taxaceae to the Compositae, I journeyed last year once more to Edinburgh to go over the Léveillé herbarium, and again Sir William facilitated my work in every possible way, so that I was able to accomplish much in the limited time at my disposal. It is to be expected that with the following supplement, the critical enumeration of the ligneous species described by Léveillé from eastern Asia will be fairly complete with the exception, perhaps, of a few species which could not be located in Léveillé’s herbarium. 1Continued from Vol. 16: 311-340; for preceding parts see Vols. 10: ag 132, 164- 196; 12: 275-281; 13: 299-332; 14: 223-252; 15: 1-27, 89-117, 267-3 54 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII TAXACEAEF! Podocarpus neriifolius D. Donn. — Rehder in Jour. Arnold Arb. 10: 108 (1929). Alyxia Schlechteri Léveillé, Fl. Kouy-Tchéou, 30 (1914), quoad Cavalerie, no. 3463; not Léveillé (1911). Cutna. Kweichou: Lo-fou, riviére de louest, J. Cavalerie, no. 3463, March 1909, “petit arbre, 4—5 m.” (sub Alyxia Schlechteri in herb. Lévl.; duplicate in A. A.) Leaves of Cavalerie no. 3463 were sent in 1916 to the Arnold Arbore- tum by Léveillé in answer to our request for a fragment of Alyxia Schlechteri. For typical A. Schlechteri see Jour. Arnold Arb, 15: 316 (1934). Of P. macrophylla var. Argyi Léveillé in Mem. R. Acad. Ci. Barce- lona, ser. 3, 12: 547 (1916), nom. nud., I have seen no description and no specimen; it is probably not different from P. macrophylla. Taxus chinensis (Pilger) Rehder in Jour. Arnold Arb. 1: 51 (1919). Tsuga Mairei Lemée & Léveillé in Monde PI. ser. 2, Sit 20 (1914). — Léveillé, Cat. Pl. Yun-Nan, 58 (1916). — Synon CHINA. Yunnan: commun sur les collines be ae 700-800 m., E. E. Maire, May, 1912 (holotype of Tsuga Mairez; cola: and inane in A. A.). The type specimen is sterile. To 7. chinensis probably also belongs T. cuspidata in Léveillé, Fl]. Kouy-Tchéou, 112 (1914). PINACEAE? Pinus Massoniana Lambert. — Rehder in Jour. Arnold Arb. 10: 109 (1929). Pinus Argyi Lemée & Léveillé in Fedde, Rep. Spec. Nov. 8:60 (1910). — Léveillé in Mem. R. Acad. Ci. Barcelona, ser. 3, 12: 547 (Cat. Pl. Kiang-Sou, 7) (1916 Pinus Argyi var. longevaginans Léveillé j in aes lc. and Mem. L.c. Pinus Cavaleriei Lemée et Léveillé in Fedde, Pinus tabulaeformis Carr. sensu Rehder in . “Arnold Arb. 10: 109 (1929) quoad syn. P. Argyi; non Hort. ex Carriére. CHInA. Kiangsu: Che-sang and Zuosé, Ch. d’Argy (syntypes of P. Argyi,; photos. in A. A.); locality illegible, Ch. d’Argy (holotype of P. Argyvi var. longevaginans ; photo.in A. A.). Kweichou: Pin-fa, montagnes, J. Cavalerie, no. 1695, March 22, 1902 (holotype of P. Cavaleriei; photo. in A. 1See Vol. 10: 108. 2See Vol. 10: 108. 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 55 When I enumerated Léveillé’s species cited above in 1929, I had not seen any of the type specimens. I have since had the opportunity to examine the types, and find that they all belong to P. Massoniana and bore already on the sheets identification labels, two (P. Cavaleriei and P. Argyi var. longevaginans) by V. Komarov, and the other two signed by M. Y. O[rr]. Cunninghamia lanceolata (Lamb.) Hooker in Bot. Mag. 54: t. 2743 (1827). Cunninghamia sinensis var. prolifera Lemée & Léveillé in Monde PI. ser. 2, 16: 20 (1914). — Léveillé, Cat. Pl. Yun-Nan, 56 CuinA. Yunnan: Tong-tchouan, pagode, alt. 2500 m., E. E. Maire, March [1910-13] ‘“‘rameaux secondaires pendants, naissant tous du centre d’un fruit” (holotype of C. sinensis var. prolifera; photo. in A. A.) The abnormality that the axis of the cone continues growing into a leafy branch is found occasionally in this species, e.g. in Wilson’s no. 4076 from Szechuan. It can hardly be considered a variety; at best it is a monstrous form. Cryptomeria japonica (L.f.) D. Don in Trans. Linn. Soc. 18: 166, t. 13, fig. 1 (1841). — Léveillé, Fl. Kouy-Tchéou, 111 (1914); Cat. Pl. Yun-Nan, 56 (1916). Cupressus Mairei Léveillé, Cat. Pl. Yun-Nan, 56 (1916). —Synon. nov. ae er Duclouxiana Hickel sensu Rehder in Jour. Arnold Arb. 110 (1929), quoad syn. Cupressus Mairei; non Hickel. a, Yunnan: plaine de Tong-tchouan, pagodes, alt. 2500 m., E. E. Maire, Jan. [1910-13], “grand et superbe resineux, petits rameaux decourbants.” (holotype of Cupressus Mairei; photo. and merotype in A. A.) In 1910 when I referred Cupressus Mairei doubtfully to C. Du- clouxiana Hickel, I had not seen the type specimen. Thuja orientalis Linnaeus, Sp. Pl. 1002 (1753). — Léveillé, FI. Kouy-Tchéou, 112 (1914); Cat. Pl. Yun-Nan, 56 (1916). Thuja orientalis L. var. Argyi Léveillé & Lemée in Monde PI. ser. 2, 17:15 (1915). CutInA. Kiangsu: Ch. d’Argy, [1848-66] (holotype of Th. orientalis var. Argyi; photo. and merotype in A. A.). The specimen which bears staminate flowers differs slightly from the usual form in its more slender and distant branchlets, but the differences are hardly sufficient to distinguish it as a variety or form. 56 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Juniperus squamata Lambert, Descr. Gen. Pinus, 2: 17 (1824). — Rehder & Wilson in Jour. Arnold Arb. 9:17 (1928).— Handel- Mazzetti, Symb. Sin. 7: 8 (1929). Juniperus ike ih Léveillé, Cat. Pl. Yun-Nan, 57 (1916), nom. seminud. — Komarov in a Syst. Herb. Hort. Bot. Reipub. Ros 5: 30 (1924), cum ds CHINA. Yunnan: au an du col de Yentzchay, J. Delavay, no. 3112, July 19, 1887 “arbrisseau a tiges redressées” (ex Komarov; syntype in herb. Paris, photo. in A. A.). Juniperus Franchetiana, of which the type could not be located in the Léveillé herbarium is known to me only from a photograph of a Delavay specimen in the Paris herbarium. Komarov distinguishes it from J. Fargesii and from J. squamata chiefly by its fruit and seed char- acters, while Handel-Mazzetti l.c. places it under J. squamata. Juniperus Franchetiana is based by Léveillé on J. recurva Franchet, non Hamilton, which in Plantae Wilsonianae (2: 58. 1914) was cited as a synonym of J. squamata, Franchet (Pl. David. 1: 292. 1884) cites under J. recurva a specimen collected in Shensi without any further remarks; the type of J. recurva Franchet, non Hamilton, and conse- quently the type of J. Franchetiana, would therefore be the David specimen from Shensi and not Delavay’s specimen from Yunnan. Juniperus squamata Lamb. var. Fargesii (Komar.) Rehder & Wilson in Sargent, Pl. Wilson. 2: 59 (1914). Juniperus Fargesii Komarov in sched. Herb. Paris, Feb. 1911; in Not. Syst. Herb. Hort. Bot. Reipub. Ross. 5: 30 (1924), cum descript. Juniperus Lemeeana Léveillé & Blin in Léveillé, Fl. Kouy-Tchéou, (1914). — Handel-Mazzetti, Symb. Sin. 7:7 (1929). Cutna. Kweichou: Gan-chouen, J. Cavalerie, no. 3922 (holotype of J. Lemeeana; photo. in A. A.) By Komarov and by Handel-Mazzetti this juniper is considered a distinct species, but it seems closely connected with J. sguamata by inter- mediate forms, and I prefer to keep it, at least for the present, as a variety of the latter. If considered a distinct species, the correct name will be J. Lemeeana Lévl. & Blin which has priority. GNETACEAE Gnetum Cavaleriei Léveillé, Fl. Kouy-Tchéou, 186 (1914).— Markgraf in Engler & Prantl, Nat. Pflanzenfam. ed. 2, 13: 440 (1926). Curna, Kweichou: Pin-fa, sur un arbre au sud a Pai-tchen, J. Cavalerie, no. 2336, March 30, 1905 (holotype; merotype in A. A.). The specimen is fragmentary; it consists of bare branches and broken, 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 57 immature, catkin-like inflorescences covered with a ferrugineous woolly tomentum. By Markgraf (in Bull. Jard. Buitenzorg, ser. 3, 10: 490. 1930) it was referred to Populus, but it certainly does not belong to that genus. From a certain resemblance of the branches and the pubescence to Lannea, I thought it might belong to the Anacardiaceae, and Dr. I. W. Bailey who kindly examined the wood structure, made the following statement: ‘The plant has five traces and five gaps at each node and the wood is clearly ring porous. The various anatomical characters, as a whole, would appear to place the plant in either the Anacardiaceae or the Burseraceae. I should be inclined to place it in the Anacardiaceae except for the fact that this family is stated to have trilacunar nodes. The Burseraceae have five traces, but I know of no species which is ring porous. The transition from three to five traces is an easy one and it may be that certain Anacardiaceae have five traces.” To the Gnetaceae Léveillé (l.c.) doubtfully refers Cavalerie no. 2764 from Pin-fa, Li-tséou-gai, April 1906, “grand arbre a tronc écailleux, les rameaux coupés laissent couler une sorte de colle.” The specimen consists of branches with long pendulous staminate catkins which slightly resemble the spikes of Gnetum. It belongs to Carpinus and is probably referable to C. Fangiana Hu with which it agrees in the large winter-buds with strongly striate scales. ARACEAE Rhaphidophora Dunniana Léveillé in Fedde, Rep. Spec. Nov. 9: 325 (1911); Fl. Kouy-Tchéou, 39 (1914 CuinaA. Kweichou: Kuao-siang, lieux ombreux et frais, foréts profonds et ruisseaux, J. Esquirol, no. 246, Jan. 1905, “gros lierre qui monte aux arbres et rochers” (holotype; photo. in A. A.). The specimen consists only of some bare branches, an immature in- florescence and detached leaves which measure 8-15 & 5-—9.5 cm. and are abruptly contracted into a short acumen about 5 mm. long. It is similar to Rh. Calophyllum Schott from Khasia and Sikkim, but may possibly belong to Scindapsus. STEMONACEAE Stemona japonica (Bl.) Miquel, Prol. Fl. Jap. 386 (1867). Helwingia Argyi Léveillé & Vaniot in Bull. Herb. Boissier, ser. 2, 6: 506 (1906). Stemona ae (Lévl. & Vant.) Léveillé in Fedde, Rep. Spec. Nov. 1 (1912); in Mem. Acad. Ci. Barcelona, ser. 3, 12: 560 (Cat. Pl. Kiang-sou, 20) (1916). — Synon. nov. 58 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XVII Cumna. Kiangsu:_ Sou-tchéou-fou, Ch. d’Argy [1848-66] (syntype of Helwingia Argyi; merotype in A. A.). Anhwei: Ning-koue-fou, Ch. d’Argy [1848-66] (syntype of H. Argyi; ex Léveillé). Neither C. H. Wright (in Jour. Linn. Soc. Bot. 32: 493. 1896), nor Schlechter (in Notizbl. Bot. Gard. Mus. Berlin, 9: 193. 1924) credit S. japonica to China, but Hemsley (in Jour. Linn. Soc. Bot. 36: 95. 1903) cites a specimen from Kiangsu collected by Carles (no. 357) on the hills near Shanghai. The two specimens in the Léveillé herbarium are in a cover which bears on the outside the localities “Sou-tchéou Ning-koue-fou,” but there is no indication from which locality each specimen came, nor any reference to Anhwei (Ngan-hoei). LILIACEAE! Smilax herbacea L. var. acuminata Wright in Jour. Linn. Soc. Bot. 36: 97 (1903). Smilax herbacea L. var. foetida Léveillé in Mem. Pontif. Accad. Nuov. Lincei, 23: 351 (Liliac. 23) (1905); Fl. Kouy-Tchéou, 256 14 914). Smilax aa L. var. heterophylla Léveillé l.c. (1905). CuinaA. Kweichou: Pin-fa, bois, bords des torrents, lieux ombrageux, J. Cavalerie, no. 1314, May 25, 1902 (holotype of S. herbacea var. foetida; photo. in A. A.); same locality and collector, no. 1314, May 22, 1902 (holotype of S. herbacea var. heterophylla; photo. in A. A.). Both of Léveillé’s varieties cited above are referable to var. acuminata Wright with which they agree in the gradually acuminate leaves, but they differ widely in the width of their leaves, var. heterophylla having on one branch leaves about 8 cm. wide, on the other 2 cm. in width, while var. foetida is intermediate with about 3—4 cm. width. Smilax glabra Roxburgh, Fl. Ind. ed. 2, 3: 792 (1832). — Léveillé, Fl. Kouy-Tchéou, 256 (1914). — Rehder in Jour. Arnold Arb, 10: 110 (1929).— Wang & Tang in Sinensia, 5: 417 (1934). Smilax glabra Roxb. var. maculata Bodinier ex Léveillé in Mem. Pontif. Accad. Nuov. Lincei, 23: 351 (Liliac. 23) (1905). Smilax Dunniana Léveillé in Fedde, Rep. Spec. Nov. 9: 446 (1911). Smilax Blinii Léveillé, Fl. Kouy-Tchéou, 256 (1914). Cutna. Kweichou: Ma-jo, J. Cavalerie, no. 2973, July 1908 (holotype of S. Dunniana and S. Blinii; merotype in A. A.). Hong- 1See Vol. 10: 110. 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 59 kong: dans les rocailles d’un torrent 4 mi-céte du Pic Victoria, E. Bodinier, Feb. 24, 1893 (holotype of S. glabra var. maculata; photo. in Smilax Dunniana is not cited in Flore du Kouy-Tchéou, but S. Blinii is described as a new species in that work, based on the same number. The name S. Dunniana does not appear on the type specimen in the Léveillé herbarium. Smilax glabra var. maculata is hardly different enough to be main- tained as a variety; it is at best a slight form. Specimens of S. glabra with leaves variegated with irregular white spots occasionally occur, e.g. T. N. Hsiung 494 from Kiangsi, J. B. Norton 1252 from Fukien and U. Faurie 950 from Formosa. Smilax microphylla Wright in Kew Bull. 1895: 117.— Wang & Tang in Sinensia, 5: 418 (1934). Smilax Labordei Léveillé & Vaniot in Mem. Pontif. Accad. Nuov. Lincet, 23: 355 (Liliac. 27) (1905). — Léveillé, Fl. Kouy-Tchéou, 257 (1914). Smilax gracillima Léveillé & Vaniot, 7 354 (Liliac. 26) (1905). — Léveillé, FI. Kouy-Tchéou, 256 (1914). Smilax Darrisii Léveillé in Fedde, Rep. Spec. Nov. 12: 533 (1913); Fl. Kouy-Tchéou, 256 (1914).—Rehder in Jour. Arnold Arb. 10: 111 (1929).— Wang & Tang in Sinensia, 5: 419 (1934).— Synon. nov. Smilax Mairei Léveillé in Bull. Géog. Bot. 25: 39 (1915); Cat. PI. Yun-Nan, 169 (1916).—Rehder in Jour. Arnold Arb. 10: 111 (1929). Cuina. Kweichou: environs de Kouy-yang, mont. du Col- lége, rocailles, June 2, 1898 (flowers), environs de Tsin-gay, Dec. 9, 1897 (fruit), J. Laborde in herb. Bodinier, no. 2318 “petit buisson épineux” (holotype of S. Labordei; photo. in A. A.); Tsin-gay, bois, J. Cavalerie, no. 1156, July 15, 1903 (holotype of S. gracillima, photo. in A. A.); steppes et bois de Gan-chouen, J. Cavalerie, no. 3819, Dec. 1909, J. Esquirol, no. 3145, May, 1911 (syntypes of S. Darrisii; mero- type of Cavalerie’s and sitioks: of Esquirol’s specimen in A. A.). Yunnan: rochers a mi-mont de La-kou, 2400 m., E. E. Maire, June 1912 (holotype of S. Mairei; photo. in A. A.). Smilax Darristi is enumerated as a distinct species by Wang & Tang (l.c.) as it was by me (l.c.), but I do not see now that it can be sepa- rated from S. microphylla. On the other hand, Wang & Tang refer S. castaneiflora Lévl. to S. microphylla, but that is undoubtedly a different species. 60 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Smilax scobinicaulis Wright in Kew Bull. 1895: 117. Smilax brevipes Warburg in Bot. Jahrb. 28: 256 (1900).— Wang & ang in Sinensia, 5: 419 (1934). Smilax ocreata Léveillé & Vaniot apud Léveillé in Mem. Pontif. Accad. Nuov. Lincei, 23: 354 (Liliac. 26) (1905). — Léveillé, FI. Kouy-Tchéou, 257 (1914).— Non A. De Candolle (1878). Smilax Martini Léveillé & Vaniot in Léveillé in Mem. Pontif. Accad. Nuov. Lincei, 23: 355 (Liliac. 26) (1905). — Léveillé, Fl. Kouy- Tchéou, 257 (1914). Smilax Cavalerici Léveillé & Vaniot apud Léveillé in Mem. Pontif. Accad. Nuov. Lincei, 23: 355 (Liliac. 27) (1905). — Léveillé, Fl. Kouy-Tchéou, 257 (1914) ; Cat. Pl. Yun-Nan, 168 (1916). Cutna. Kweichou:_ environs de Gan-pin, au Grandes Rocailles, trainant sur les rochers, L. Martin in herb. Bodinier, no. 1871, Sept. 17, 1897 (holotype of S. ocreata; photo. in A. A.); Pin-fa, bois, J. Cavalerie, no. 1315, May 3, 1902 (holotype of S. Martini; photo. in A. A.); Pin-fa, montagnes, J. Cavalerie, no. 1313, May 3, 1902 (holo- type of S. Cavaleriei; photo. in A. A.). The three species of Léveillé’s had been referred by Wang & Tang to S. brevipes Warb. which can hardly be separated from S. scobinicaulis as a distinct species. Smilax leucocarpa Léveillé & Vaniot apud Léveillé in Mem. Pontif. Accad. Nuov. Lincei 23: 354 (Liliac. 26) (1905); 24: 344 (Liliac. 10) (1906). — Wang & Tang in Sinensia, 5: 420 (1934). Smilax Esquirolii Léveillé, Fl. Kouy-Tchéou, 256 (1914). Curna. Kweichou: environs de Kouy-yang, dans les bois de Kien-lin-chan, E. Bodinier, no. 2124, Apr. 14 (flowers) and July 7 (fruit), 1898, ‘“arbuste épineux, a tiges dressées, non sarmenteuses”’ (holotype of S. leucocarpa; photo. in A. A.); plateau de San-tou, 1000 m., J. Esquirol no. 2109, May 1, 1910 (type of S. Esquirolii; photo. in A. A.); Gan-pin, J. Cavalerie, no. 2120 (op. cit. 24: 344 under S. leuco- carpa; photo, in A, A.) This species resembles somewhat in its small leaves S. microphylla Wright, but the leaves are cuneate and obtuse and the fruits are yellow; from the related S. megalantha Wright and S. China L. it is at once dis- tinguished by the very small obovate to obovate-oblong obtuse leaves and smaller fruits on shorter peduncles. In the Flore du Kouy-Tchéou Léveillé cites S. leucocarpa as a syno- nym of S. Esquirolii, a name which appears on the sheet of Esquirol 2109, but does not seem to have been published before 1914. Smilax megalantha Wright in Kew Bull. 1895: 118.— Wang & Tang in Sinensia, 5: 420 (1934). 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 61 Smilax loupouensis Léveillé in Bull. Géog. Bot. 25: 38 (1915) ; Cat. Pl. Yun-Nan, 169 (1916). Smilax ferox Wall. sensu Rehder in Jour. Arnold Arb. 10: 110 (1929), p-p., quoad synon. S. loupouensis; non Wall. ex DeCandolle. Cuina. Yunnan: “brousse des mont. a Lou-pou, 3000 m., E. E. Maire, Sept. 1912, ‘“arbuste grimpant, un peu épineux, feuilles blanches au dessous, fleurs grises” (holotype of S. loupouensis ; mero- type in A. A.). Smilax loupouensis resembles S. discotis Warb. in its stipules extend- ing the entire length of the short petiole, but that species has black fruits. From S. ferox, to which it seems closely related, it differs in the only slightly spiny branches, in the leaves being abruptly contracted at the apex and very glaucous beneath and in the stipules. It apparently agrees best with S. megalantha, but differs from the type in the shorter petioles with the stipules extending the whole length. Smilax China Linnaeus, Spec. Pl. 1029 (1753).— Léveillé, FI. Kouy-Tchéou, 256 (1914), p.p. excl. syn. S. Bodinieri; Cat. Pl. Yun- Nan, 168 (1916). — Norton in Sargent, Pl. Wilson. 3: 4 (1916). Smilax Taquetii Léveillé in Fedde, Rep. Spec. Nov. 10: 372 (1912). Smilax China L. f. obtusa Léveillé in Mem. Pontif. Accad. Nuov. Lincei, 24: 344 (Liliac. 10) (1906). Korea. Quelpaert: in silvis Hogno et Hallisan, E. Taquet, nos. 3306, 3307, 3308, Jan., May, Aug. 1909 (syntypes of S. Taquetii; photo. of 3306 and isotypes of nos. 3306, 3307 and 3308 in A. A.). Cuina. Kiangsu: Tou-gou-lin, Ch. d’Argy, [1848-66] (holo- type of S. China f. obtusa; photo. in A. A.). Smilax Taquetii was first identified with S$. China by Norton (l.c.). Smilax China f{. obtusa does not seem to be different from typical S. China. Smilax glauco-china Warburg in Bot. Jahrb. 29: 255 (1900).— Wang & Tang in Sinensia, 5: 420 (1934). Smilax Bodinieri Léveillé & Vaniot apud Léveillé in Mem. Pontif. Accad. Nuov. Lincei, 23: 355 (Liliac. 27) (1905). Smilax China “L.” sensu Léveillé, Fl. Kouy-Tchéou 256 (1914), quoad syn. S. Bodinieri; non Linnaeus. Cutna. Kweichou: monts de Hin-y-fou et Hin-y-hien, dans les haies, E. Bodinier, no, 1581, Apr. 13, 1897 (holotype of S$. Bodiniert ; photo. in A. A.). Smilax Bodinieri was cited as a synonym of S. China L. by Léveillé in his Flore du Kouy-Tchéou, but Wang & Tang referred it to S. glauco- china. 62 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Smilax menispermoidea A. De Candolle, Monog. Phan. 1: 108 (1878). — Wang & Tang in Sinensia, 5: 420 (1934). Smilax luteocaulis Léveillé in Fedde, Rep. Spec. Nov. 13: 339 (1914) ; Cat. Pl. Yun-Nan, 169 (1916). Cuina. Yunnan: brousse du mont To-chan, 3200 m., E. E. Maire, June, 1913 (holotype of S. /uteocaulis ; merotype in A. A.). The specimen agrees perfectly with typical S. menispermoidea. Smilax stemonifolia Léveillé & Vaniot apud Léveillé in Mem. Pontif. Accad. Nuov. Lincei 23: 356 (Liliac, 28) (1905).— Wang in Bull. Fan Mem. Inst. Biol. 5: 117 (1934). Cuina. Hongkong: bois le long de Bowenroad, trés rare, E. Bodinier, no. 804, Aug. 4, 1894 (holotype; photo. in A. A.). The type specimen of this species is sterile, but Wang (l.c.) refers to it two other specimens, one from Hunan collected by H. F. Chow and P. Ou-yang, no. 46102, with young fruit, and one from Anhwei col- lected by F. T. Wang, no. 23860, sterile. Wang states that it is closely related to S. Bockii Warb. which differs in its much narrower leaves rounded at the base and in longer pedicels. It may also be near S. longipes Warb. Smilax castaneiflora Léveillé in Bull. Acad. Géog. Bot. 25: 39 (1915); Cat. Pl. Yun-Nan, 168 (1916).— Rehder in Jour. Arnold Arb. 10: 111 (1929). Smilax microphylla “Wright” sensu Wang & sea in a 5: 418 (1934) quoad synon. S. castanetflora; non Wri CHINA. Yunnan: brousse des mont. a nitie ee, alt. 2700 m., E. E. Maire, June 1912, “arbuste grimpant, fleurs chocolat, fruits noirs” (holotype; merotype in A. A In general appearance this species resembles S. Bockii Warb., but the branches are angled and quite spiny, the leaves are short-cordate, lanceolate, broad at the base and partly constricted below the middle into the narrow long-acuminate upper part. By Wang & Tang the species was referred to S. microphylla Wright, but it differs from that species in the long peduncle, the gradually acuminate cordate leaves, often somewhat restricted below the middle and green beneath. I have been unable to identify this species with any Chinese species I know. Handel-Mazzetti, no. 10379, from Kweichou may belong here, but it has broader leaves and is not spiny. Smilax micropoda A. DC. var. reflexa Norton in Sargent, Pl. Wil- son. 3: 6 (1916). Smilax Lebrunii Léveillé, Fl. Kouy-Tchéou, 257 (1914). Pehies & Tang in Sinensia, 5: 421 (1934), “Lebranii.” — Synon 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 63 CHINA. Kweichou: bois sacré de Ban-gnien, 1100 m., J. Esquirol, no. 3518, March 19, 1912 (holotype of S. Lebrunii; photo. in Wang & Tang keep S. Lebrunii as a distinct species and refer to it three specimens from Kwangsi, R. C. Ching 5846, 6992 and 7201 which I have not seen. Smilax tortopetiolata Léveillé & Vaniot apud Léveillé in Mem. Pontif. Accad. Nuov. Lincei 23: 354 (Liliac. 26) (1905); Fl. Kouy- Tchéou, 258 (1914). — Wang & Tang in Sinensia, 5: 422 (1934), excl. synon. Smilax opaca (A. DC.) Norton in Sargent, Pl. Wilson. 3: 11 (1916). — Synon. nov. CutInA. Kweichou: Pin-fa, bois, J. Cavalerie, no. 1312, March 18, 1902 (holotype of S. tortopetiolata; photo. in A. A.). This species had not yet been reported from western China; it seems common in southeastern China and is represented in this herbarium by many specimens from Fukien, Kwangtung, Hainan and Kwangsi. I have been unable to find any specific difference between S. opaca and S. tortopetiolata which has to replace the former on account of priority. Smilax cocculoides Warb. var. lanceolata Norton in Sargent, Pl. Wilson. 3: 11 (1916). Smilax Pinfaensis Léveillé & Vaniot apud Léveillé in Mem. Pontif. Accad. Nuov. Lincei, 23: 355 (Liliac. 27) (1905). — Léveillé, Fl. Kouy-Tchéou, 257 (1914). — Synon. nov. Smilax tortopetiolata “Lévl. & Vant.” sensu Wang & Tang in Sinensia, 5: 422 (1934), p.p., quoad synon. S. pinfaensis. Cutna. Kweichou: Pin-fa, J. Cavalerie no. 1311, March 18, 1902, “fl. verte” (holotype of S. pinfaensis ; photo. in A. A.). Smilax pinfaensis is apparently identical with S. cocculoides var. lanceolata Norton, the type of which is in fruit, and which I have before me. The flowers agree with those of typical S. cocculoides, as repre- sented by Wilson, no. 3252. Wang & Tang (l.c.) refer S. pinfaensis as a synonym to S. tortopetiolata, but it is easily distinguished from the latter by the narrow-lanceolate, long-acuminate leaves, the slender peduncles much longer than the petioles and the short stamens. Smilax Lyi Léveillé in Fedde, Rep. Spec. Nov. 8: 171 (1910); Fl. Kouy-Tchéou, 257 (1914). — Wang & Tang in Sinensia, 5: 421 (1934). Cuina. Kweichou: Pin-fa, montagnes, J. Cavalerie, no. 1406, Sept. 27, 1903 ‘fl. vert-jaune” (holotype; photo. in A. A.). 64 JOURNAL OF THE ARNOLD ARBORETUM [voL. xv This species was kept distinct by Wang & Tang. The rather meagre specimen which consists of a branch with two leaves and an inflo- resence in bud, resembles S. stenophylla A. Gray from the Liukiu Islands and Formosa, but with the scant material at hand, I hesitate to consider these geographically widely separated species identical. Smilax perfoliata Loureiro, Fl. Cochinch. 622 (1790).— Wang & Tang in Sinensia, 5: 422 (1934). Smilax perulata Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 9:78 (1910). — Léveillé, Fl. Kouy-Tchéou, 257 (1914). Cuina. Kweichou: Lo-fou, J. Cavalerie, no. 3648, March 1909 (holotype of S. perulata; merotype in A. A.). This species was identified first by Wang & Tang with S. perfoliata Lour. (S. ocreata A. DC.). Heterosmilax Gaudichaudiana (Kunth) Maximowicz in Bull. Acad, St. Pétersb. 17: 176 (1872); in Mél. Biol. 8: 415 (1872).— A. DeCandolle, Monog. Phan. 1: 44 (1878). — Léveillé in Mem. Pontif. Accad. Nuov. Lincei, 23: 350 (Liliac. 22) (1905). Heterosmilax Gaudichaudiana var. latifolia Bodinier ex Léveillé 1.c. 1905). Cuina. Hongkong: bord de l’Aquéduc, prés Richmond Ter- race, FE. Bodinier, Aug. 22, 1894 (holotype of H. Gaudichaudiana var. latifolia; photo. in A. A.). In the shape and measurements of the leaves Léveillé’s variety agrees with DeCandolle’s var. « Gaudichaudiana (1.c.), which represents the type of the species. It seems doubtful whether this and var. 3 hong- kongensis (Seem.) A. DC., (l.c.) are sufficiently distinct to be main- tained as varieties. TACCACEAE Tacca Esquirolii (Lévl.), comb. nov. Clerodendron Esquirolii Léveillé in 7 Rep. Spec. Nov. 11: 298 (1912) ; Fl. Kouy-Tchéou, 439 (1915). Tacca Paxiana Limpricht f. in Engler, Pflanzenreich, IV. 42: 16 (1928).— P’ei in Mem. Sci. Soc. China, 1, no. 3, p. 162 (1932). Cuina. Kweichou: bois de Ta-Thamn, trés ombreux et chaud, J. Esquirol, no. 2802, May 1912, “toute l’inflorescence rouge, lie de vin, feuilles radicales, fleur 1 m. de haut sur tige simple” (holotype of Clero- dendron Esquirolii; photo. in A. A.). Léveillé’s species was first identified by the writer with Tacca Paxtana (see P’ei, l.c.). 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 65 SALICACEAEF} Populus adenopoda Maximowicz in Bull. Soc. Nat. Moscou, 54: 50 (1879). — Léveillé, Fl. Kouy-Tchéou, 380 (1915); Cat. Pl. Yun-Nan, 250 (1917). — Handel-Mazzetti, Symb. Sin. 7: 58 (1929). Populus macranthela Léveillé & Vaniot in Bull. Soc. Bot. France, 142 (March 1905); in Monde Pl. 12:9 (1910); in Fedde, ep. Spec. Nov. 8: 446 (1910).— Rehder in Jour. Arnold Arb. 10: 111 (1929). Populus Duclouxiana Dode in Bull. Soc. Nat. Hist. Autun, 18: 190, t. 11, f. 329 (Extr. Monog. ined. Populus, 32) (1905). Populus alee aa Griff. var. Duclouxiana (Dode) Gombocz in Mat mesz. Kozl. 30: 130 (Monog. ae Populi) (1908). — Schneider i in Sargent, Pl. Wilson. 3: 25 (19 Populus rotundifolia Griff. « macranthela ae in Bot. Kozl. 10: 25, (7) (1910). — Léveillé, Fl. Kouy-Tchéou, 380 (1915). Cuina. Kweichou: Pin-fa, J. Cavalerie, no. 974, April 1, 1903, “grande arbre” (holotype of P. macranthela; photo. in A. A.). Yunnan: _ Ko-kouy prés Tchao-tong, Pére Marc Mey in herb. F. Ducloux, no, 667, in 1906 (in hb. Léveillé and in hb. Univ. Calif. sub P. macranthela ; isotypes of P. Duclouxiana Dode; photos. in A. A.). In my earlier account of P. macranthela, I had considered it a species distinct from P. adenopoda, chiefly on account of the absence of the glands at the base of the blade of the leaves, their glabrousness and coarser serration. In comparing a large series of specimens of P. adeno- poda one finds a great variability in these characters; though the glands are usually conspicuous, they may occasionally be small and partly lacking, as in Wilson’s no. 724 from Hupeh, and in Steward, Chiao and Cheo, no. 294, from Kweichou. One of the two mature leaves of Duclours specimen in the herbarium of the University of California has an almost suborbicular shape, but on the material from trees cultivated in the Arnold Arboretum similar leaves can be found. Both specimens I have seen of Ducloux’s no. 667 bear the name P. macranthela in Lé- veillé’s handwriting. Neither Dode nor Gombocz give an exact citation of the specimen upon which P. Duclouxiana is based, but I assume that the species is based on Ducloux’s no. 667. Salix Cavaleriei Léveillé. — Rehder in Jour. Arnold Arb. 10: 113 (1929). — Handel-Mazzetti, Symb. Sin. 7: 61 (1929). Salix erioclada Léveillé.— Rehder, l.c. 115 (1929). — Handel- Mazzetti, l.c. 64, fig. 1,1, 2 (1929). Handel-Mazzetti gives a detailed description of the species and figures staminate and pistillate flowers. 1See Vol. 10: 111. 66 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI Salix luctuosa Léveillé.— Rehder, lc. 115 (1929). — Handel- Mazzetti, I.c. 68 (1929). Handel-Mazzetti gives a full description of the pistillate flowers and refers S. dyscrita Schneid. as a synonym to this species. Salix Duclouxii Léveillé. — Rehder, l.c. (1929).— Handel-Maz- zetti, lc. 87 (1929). Salix kouytchensis (Lévl.) Schneider in Sargent, Pl. Wilson. 3: 171 (1916). — Rehder, I.c. 117 (1929). Handel-Mazzetti unites S. kouwytchensis (Lévl.) Schneid. with S. Duclouxii and discusses the close relation of the latter to S. Bocki Seemen and S. variegata Franch., but prefers to keep these species dis- tinct for the present. MYRICACEAE! Myrica esculenta Hamilton ex D. Don, Prodr. Fl. Nepal, 56 (1825). Litsea myricopsis Léveillé, Cat. Pl. Yun-Nan, 150 (1916).— Synon. nov. Cutna. Kweichou: prés Tchang-pou, J. Cavalerie, no. 2697, Nov. 3, 1905, “petit arbre” (holotype of Litsea myricopsis ; merotype in A. A.). Myrica esculenta has been collected in Kweichou also by Handel- Mazzetti (no. 10852). Myrica adenophora Hance in Jour. Bot. 21: 357 (1883). — Lé- veillé, Fl. Kouy-Tchéou, 281 (1914). Litsea Mairei Léveillé, Cat. Pl. Yun-Nan, 150 (1916).— Synon. nov. Cutna. Yunnan: collines arides incultes de Tché-hay, alt. 2600 m., E. E. Maire, March, 1912 “arbousier, arbuste rameuse, fleur blanc-verdatre” (holotype of Litsea Mairei; merotype in A. A.). Myrica adenophora does not seem to have been recorded from Yunnan before, nor from Kweichou, for I feel sure that Esquirol no. 3296 cited by Léveillé in his Flore du Kouy-Tchéou under that species (1.c.) does not belong here, but I have not seen it. BETULACEAF? Alnus Fauriei Léveillé in Bull. Soc. Bot. France, 51: 423 (1904). — Rehder in Jour. Arnold Arb, 10: 118 (1929). APAN. Hondo: in monte Gurvassan, U. Faurie, no. 783, Sept. 28, 1897 (holotype; photo. in A. A.). The type specimen bears an identification label by Koidzumi with 1See Vol. 10: 118. 2See Vol. 10: 118. 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 67 the legend “Alnus cylindrostachya (Winkl.) Koidz.”” Though Wink- ler’s name, A. glutinosa var. cylindrostachya, was published June 17, 1904, while Léveillé’s name did not appear before November 1904, Léveillé’s binomial has precedence over Winkler’s varietal name. Alnus nepalensis D. Don, Prodr. Fl. Nepal. 58 (1825). — Léveillé, Fl. Kouy-Tchéou, 124 (1914); Cat. Pl. Yun-Nan, 66 (1916). Alnus Mairei Léveillé in Bull. Géog. Bot. 24: 283 (1914). Cuina. Yunnan: vallées et montagnes de Tong-tchouan, 2500— 2700 m., E. E. Maire, Nov. 1913, “grand arbre, chatons jaunes” (holo- type of A. Mairei; photo. in A. A.). Alnus Mairei was referred to A. nepalensis by Léveillé himself in 1916. Carpinus Seemeniana Diels in Bot. Jahrb. 29: 279 (1900).— Schneider in Sargent, Pl. Wilson. 2: 430 (1916). Carpinus pinfaensis Léveillé & Vaniot in Bull. Soc. Bot. France, 52: 142 (1905). Carpinus pubescens Burkill sensu Winkler in Bot. Jahrb. 51: 501 (1914), p.p., quoad syn. C. pinfaensis et C. Seemeniana. — Léveillé, Fl. Kouy-Tchéou, 125 (1914).—Hu in Sunyatsenia, 1: 119 (1933). Cuina. Kweichou: Pin-fa, J. Cavalerie, no. 1011, May 28, 1903, “arbre” (holotype of C. pinfaensis; photo. in A. A Carpinus pinfaensis was first referred to C. pubescens (including C. Seemeniana) by Winkler in 1914, but in 1916 it was placed under C. Seemeniana by Schneider, where it certainly belongs if C. Seemeniana and C. pubescens are considered distinct. These two species and also C. austrosinensis Hu are closely related and perhaps best united under C. pubescens with one or two varieties, but at present I prefer to keep C. Seemeniana as a species, since the extreme forms are quite distinct. Carpinus austrosinensis is nearer to the latter. Carpinus laxiflora Blume, Mus. Bot. Lugd.-Bat. 1: 309 (1850) :— Schneider in Sargent, Pl. Wilson. 2: 438 (1916). ? Carpinus laxiflora Bl. var. chartacea Léveillé in Bull. Soc. Bot. France, 51: 424 (1904) JAPAN: without locality and collector (holotype of var. chartacea ; ex Léveillé). No specimen named C. laxifolia var. chartacea can be found in the Léveillé herbarium, but I have no doubt that it is but a slight variation of C. laxiflora under which it was placed by Schneider in 1916. 68 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI FAGACEAF! Castanea mollissima Blume in Mus. Bot. Lugd.-Bat. 1: 286 (1850). Castanopsis yunnanensis (Franch.) Léveillé, Cat. Pl. Yun-Nan, 66 (1916), Syn.: “(C. vulgaris L., p.p.)” — Camus, Chataigniers, 485 The combination Castanopsis yunnanensis was published by Léveillé without any citation of a specimen. It is based on Castanea vulgaris var. yunnanensis Franchet which is a synonym of C. mollissima; also the citation by Léveillé of C. vulgaris L., p.p., as a synonym seems to indicate that it is a true Castanea. It can, therefore, hardly be referred to any other species than C. mollissima which is common in Yunnan; probably to this species belongs also C. sativa Mill. cited in Cat. PI. Yun-Nan, 66. Castanopsis cuspidata (Thbg.) Schottky in Bot. Jahrb. 47: 625 (1912). Castanea Fauriei Léveillé & Vaniot in Bull. Soc. Bot. shige 52: 142 (1905). — Camus, Chataigniers, 241 (1929). — Synon. nov. JapAN. Kiushu: autour de Nagasaki, U. eRe no. 3681, June 1899 (holotype of C. Fauriei,; isotype in A. A.). Miss Camus enumerated C. Fauriei under the doubtful species of Cas- tanea with a note that it is probably a Castanopszs. Castanopsis hystrix (Hook. f. & Th.) A. DeCandolle in Jour. Bot. 1: 182 (1863); Prodr. 167: 111 (1864). — Léveillé, Cat. Pl. Yun-Nan, 66 (1916). — Rehder in Jour. Arnold Arb, 10: 118 (1929). Castanea de Léveillé & Vaniot in Bull. Soc. Bot. France, 52: 142 (1905). — Léveillé, Cat. Pl. Yun-Nan, 66 (1916).— Camus, Chataigniers, 241 (1929). Castanopsis Bodinieri (Lévl. & Vant.) Koidzumi in Tokyo Bot. Mag. 30: 100 (1916) ‘“Boodinieri.” Quercus brunnea Léveillé in Fedde, Rep. Spec. Nov. 12: 364 (1913) ; Fl. Kouy-Tchéou, 127 (1914). — Camus, Chataigniers, 482 (1929) ; Atlas, pl. 28, fig. 9-11, 1x. f. 9-17 (1928). Castanopsis brunnea (Lévl.) Camus, Chataigniers, Atlas, pl. 28 (1928). Cu1na. Kweichou: Pin-fa, bois, J. Cavalerie, no. 2299, Apr. 13, 1905 (holotype of Q. brunnea; photo.in A.A.). Yunnan: en- virons de Yun-nan-sen, bois de la pagode de Kiang-tchou-se, E. Bo- dinier, Feb. 2, 1897, “grand arbuste ou petit arbre, fruits en long épis a glands sessiles serrés” (holotype of Castanea Bodinieri ; photo. in A. A.). Castanea Bodinieri is enumerated by Miss Camus (l.c.) under in- 1See Vol. 10: 118. 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 69 completely known species of Castanea. Quercus brunnea she describes under doubtful species of Castanopsis and says “probablement var. ou sous-espece du C. Hystrix.” Castanopsis Eyrei (Champ.) Tutcher in Jour. Linn. Soc. Bot. 37: 68 (1905). — Chun in Sunyatsenia, 1: 217, pl. 35 (1934), excl. syn. Q. Castanopsis. Castanopsis caudata Franchet in Nouv. Arch. Mus. Paris, ser. 3, 7: 87 (Pl. David. 1:277) (1884).—Rehder in Jour. Arnold Arb. 10:119 (1929), excl. syn. Q. Castanopsis. — Handel-Mazzetti, Symb. Sin. 7: 28 (1929). Quercus trinervis Léveillé in Fedde, Rep. Spec. Nov. 12: 364 (1913) ; Fl. Kouy-Tchéou, 128 (1914 Quercus cepifera Léveille, l.c. 364 (1913) ; lc. 127 (1914), pro parte, quoad specimina foliifera. Castanopsis asymetrica Léveillé, Fl. Kouy-Tchéou, 125 (1914).— A. Camus, Chataigniers, 472 (1929) “asymmetrica”; Atlas, pl. 72, fig. 15, xxxiv, fig. 17-19 (1928). Castanopsis neo-Cavaleriet A. Camus, ee ai 375- (1929): Atlas, pl. 72, fig. 1-12, xxx1v, fig. 8-15 (1928). Castanopsts trinervis (Lévi.) A. Camus, op. cit. 400 (1929); Atlas, pl. 70, fig. 1-7, xxx111, fig. 1-7 (1928). Castanopsis tribuloides var. echidnocarpa King sensu Rehder in Jour. Arnold Arb. 10: 119 (1929) ; non King apud Hooker f. (1888). Cuina. Kweichou: J. Cavalerie, no. 3275, Aug. 20, 1908 (holotype of Q. trinervis ; photo. in A. A.) ; sud de Pin-fa, rare, J. Cava- lerie, no. 2341 (in part, except fruit), June 8, 1905 ‘fruits ci-joints trouvés par terre; apartiennent-ils 4 cet arbre” (holotype of Q. cepifera, except fruit; photo. in A. A.); Kouy-yang, monts du College, J. Chaffan- jon in herb. Bodinier, no. 2235, Apr. 15, 1898, ‘‘arbre, fleurs blanches”’ (syntype of Castanopsis asymetrica; merotype in A, A.); Kouy-yang, bois d’une pagode, J. Cavalerie, no. 2078 [cited as 2178 by Léveillé], “arbre” (syntype of C. asymetrica; photo. in A. A.); Kien-lin-chan, J. Esquirol, no. 17, May 2, 1904, “feuilles de camélia” (syntype of C. asymetrica; photo. in A. A.). I agree with Chun in considering Castanopsis Eyrei and C. caudata identical. Miss Camus places C. asymmetrica, C. Eyrei and C. caudata among the incompletely known species and describes C. trimervis as a distinct species. Handel-Mazzetti cites Castanopsis asymetrica as a synonym of C. caudata Franch. Castanopsis neo-Cavaleriei A. Camus in Bull. Bimens. Soc. Linn. Lyon, 8: 87 (1929); Chataigniers, 375 (1929) ; Atlas, pl. 72, fig. 1~12, XXIV, fig. 8-15 (1928). 70 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Quercus Cavaleriei Léveillé & Vaniot in oe Soc. Bot. France, 52: 142 (1905); Fl. Kouy-Tchéou, 127 (191 Quercus Castanopsis Léveillé in rt om Spec. Nov. 12: 363 (1913); Fl. Kouy-Tchéou, 127 (191 Castanopsis caudata Franchet sensu See in sbeakd aaa Arb. 10: 119 (1929), quoad syn. Q. Castanopsis; non Fra Castanopsis tribuloides var. echidnocarpa King sensu | Rehder, lc. (1929) ; non King ex Hooker f. Castanopsis eyrei (Champ.) Tutcher sensu Chun in Sunyatsenia, 1: 217 (1934), quoad syn. Q. Castanopsis; non (Champ.) Tutcher. Cutna. Kweichou: Pin-fa, hautes montagnes, J. Cavalerie, no. 57 (in part; young fruit), July 15, 1902, “petit arbre” (holotype of Quercus Cavaleriei; photo. in A. A.); Pin-fa, Ma-jo, hautes montagnes, rare, J. Cavalerie, no. 57 (in part; flowers), May 1902 (syntype of Q. Castanopsis; merotype in A. A.); same locality, J. Cavalerie, no. 57, Aug. 1908 (syntype of Q. Castanopsis; merotype in A. A.); same lo- cality, J. Cavalerie, no. 1268, July 1908 (syntype of Q. Castanopsis ; ex Léveillé; ? merotype in A. A.). This species is closely related to C. Eyrei, but differs in the thinly scurfy-tomentulose brownish under side of the leaves — quite glabrous in C. Eyrei — and in the larger, very oblique cupula with more distinct fascicles of spines nearly wanting on the ventral side. From C. Fargesii Franch, it differs in the broader leaves less densely tomentulose be- neath, — in C. Fargesii the dense brown tomentum obscures the vein- lets, — and in the larger apparently indehiscent cupula. Two fruiting spikes are figured by Miss Camus on pl. 72; fig. 3 apparently represents the young fruits of Q. Cavaleriei, while fig. 4-6 represent mature fruits of Q. Castanopsis collected in August. The name C. neo-Cavaleriei for this species has to stand, since there exists already a Castanopsis Cavaleriei Lévl. of 1913 which, however, is not a Castanopsis, but represents Sloanea sinensis (Hance) Hu (see Jour. Arnold Arb. 15: 89. 1934), and the specific epithet of the second synonym cannot be transferred to Castanopsis, since this would result in a tautonym. Castanopsis tibetana Hance. — Rehder in Jour. Arnold Arb. 10: 119 (1929), Quercus Franchetiana Léveillé, Fl. Kouy-Tchéou, 128 (1915).— Camus, Chataigniers, 485 (1929). Miss Camus states that C. Franchetiana is perhaps identical with C. tibetana ; in a note she says that the leaves of the former are quite differ- ent from those of Hance’s type in shape, are very lustrous above and more waxy and less pubescent beneath but that she has seen an inter- mediate form. 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 71 Castanopsis cryptoneuron (Lévl.) Rehder in Jour. Arnold Arb. 10: 119 (1929).— Camus, ChAataigniers, Atlas, pl. 54, fig. 1-4 (1928), nom, tantum; text, 418 (1929). Miss Camus cites the combination as dating from her Atlas of 1928, but since the original binomial is not cited there, it cannot be considered as validly published. Castanopsis spec. Rehder in Jour. Arnold Arb. 10: 120 (1929). Quercus Argyi Léveillé in Mem. Acad. Ci. Art. Barcelona, ser. 3, 12: 548 (Cat. Pl. Kiang-Sou, 8) (1916). Cuina. Kiangsu: Sonosé, Ch. d’Argy, nos. 842, 856 [1846— 66]. In the cover of Q. Argyi in Léveillé’s herbarium, there are two speci- mens, no. 842 with pistillate and no. 856 with staminate flowers, both belonging apparently to the same species related to C. Fargesii Franch., but they have broader and more serrate leaves. Cavalerie’s no. 2078 (not 2178) of which I stated (l.c.) that it may belong here, belongs to C. Eyrei (Champ.) Tutch. together with the other specimens of Casta- nopsis asymetrica, Castanopsis spec. Myrica Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 12: Sa (1983); pro parte, quoad specim. Q Cuina. Kweichou: haut plateau, route de Pin-fa a Kouy- yang, J. Cavalerie, no. 3184, June 1908 (holotype of Myrica Cavaleriet, in part; photo. in A. A.). Cavalerie’s no. 3184 consists of two leafless branches, for which see Quercus spec. (p. 72), and a leafy branch with short axillary spikes of pistillate flowers and with lanceolate slender-petioled leaves 10-12 cm. long and about 2 cm. broad. Quercus serrata Thunberg. — Rehder in Jour. Arnold Arb. 10: 120 (1929). Since the name Q. serrata Thbg. which correctly belongs to the species named Q. glandulifera Bl. has been applied by almost all authors, fol- lowing Siebold and Zuccarini, to Q. acutissima Carruthers, it is advis- able to consider Q. serrata a nomen ambiguum and use Q. acutissima and Q. glandulifera for the two species involved. Quercus guyavaefolia Léveillé in Fedde, Rep. Spec. Nov. 12: 363 (1913); Cat. Pl. Yun-Nan, 67 (1916) “guyavifolia.” — Rehder in Jour. Arnold Arb. 10: 121 (1929).— Camus, Chénes, Atlas, 1: t. 35 (1934). 72 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xvi Quercus semecarpifolia Sm. var. longispica Handel-Mazzetti, Symb. Sin. 7: 39 (1929). Handel-Mazzetti cites Q. guyavaefolia as a synonym of his Q. seme- carpifolia var. longispica, but does not enumerate the type of Léveillé’s species, and apparently bases his description on other specimens from Yunnan and also from Szechuan; Miss Camus’ illustrations also are partly based on other specimens. The type specimen of QO. guyavaefolia has all the leaves entire. Quercus Prainiana Léveillé. — Rehder in Jour. Arnold Arb. 10: 121 (1929), This species is figured by Léveillé in the manuscript work Cat. III. Pl. Seu-Tchouan on pl. 23 under the name Q. Prainii. Quercus glauca Thunberg, Fl. Jap. 175 (1784).— Rehder in Jour. Arnold Arb. 10: 121 (1929). — Handel-Mazzetti, Symb. Sin. 7: 50 (1929). Quercus Blakei Skan var. Vaniotii (Lévl.) Chun in Jour. Arnold Arb. 9: 153 (1928). Chun’s new combination is based on Q. Vaniotii Lévl. and on Cava- lerie, no. 3274 (not 3264, as cited), which undoubtedly belongs to Q. glauca, he evidently placed too much faith in Léveillé’s statement, that Q. Vaniotii is very closely related to Q. Blakei Skan. The type speci- men consists only of sterile branches; the leaves show the closely appressed pubescence of Q. glauca, while Q. Blakei has the leaves gla- brous and green beneath, more gradually attenuate at base and with shorter appressed and fewer teeth. Handel-Mazzetti cites (l.c.) QO. Vaniotti as a synonym of Q. glauca. Quercus spec. — Rehder in Jour. Arnold Arb. 10: 122 (1929). Myrica Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 12: 537 (1913), pro parte, quoad specim. ¢. Curna. Kweichou: haut plateau, route de Pin-fa 4 Kouy- yang, J. Cavalerie, no. 3184 in part (4), April, 1908 (holotype of Myrica Cavaleriei, in part; photo. in A. A.). To Quercus belong two leafless branches with immature staminate catkins; for the pistillate specimen see Castanopsis spec. (p. 71). MORACEAE! Vanieria tricuspidata (Carr.) Hu in Jour. Arnold Arb. 5: 228 (1924). Cudrania_ tricuspidata Bureau in Lavallée, Arb. Segrez. 243 (1877). — Nakai, Fl. Sylv. Kor. 19: 114 (1932). 1See Vol. 10: 123. 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 73 Morus integrifolia Léveillé & Vaniot in Bull. Acad. Intern. Géog. Bot. 17, no. 210-11, p. iii (1907).— Schneider in Sargent, Pl. Wilson. 3: 301, in nota (1916) Cuina. Hopei: Tchao-tchao, cimetiére, L. Chanet, no. 49, June 25, 1905. “fruit rouge a la maturité” (holotype of Morus integri- folia; photo. in A. A.) Nakai (l.c.) seems to have first referred Morus integrifolia to this species. As Cudrania Trecul, though proposed as a nomen conservan- dum, did not receive the unanimous vote of the Committee, I retain here the oldest generic name Vanieria Lour. As Chanet collected his specimen in a cemetery, it was probably from a planted tree the specimen has not been recorded as growing wild in Hopei, but we have specimens in the herbarium from the adjoining provinces — Shansi, Honan and Shantung. Ficus gibbosa Bl. var. cuspidifera King. — Rehder in Jour. Arnold Arb. 10: 123 (1929). Ficus cuspidifera Miquel in London on Bot. 7: 434 (1848).— Handel-Mazzetti, Symb. Sin. 7:92 (19 Handel-Mazzetti cites F. rhomboidalis tél, & Vant. as a synonym of F. cuspidifera. Ficus glaberrima Blume, Bijdr. 451 (1825). Ficus suberosa Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 8: 549 (1910). — Léveillé, Fl. Kouy-Tchéou, 433 Cee — Rehder in Jour. Arnold Arb. 10: 131 (1929).— Synon. n Ficus Blinti Léveillé ba Vaniot, I.c. 550 (1910). wn Leveillé, lc. 429 (1915). — Synon. n Ficus Feddei Léveillé & ‘Vanier op. cit. 9: 19 (1910). — Synon. n Ficus Kingiana Léveillé, Fl. Kouy-Tchéou, 431 (1915), pro ae non Hemsley. — Synon. nov Cuina. Kweichou: Lo-fou, J. Cavalerie, no. 3597, March 1909, (holotype of F. suberosa; fragments in A. A.); Pia-ouai-lo, alt. 600 m., J. Esquirol, no. 2197, Sept. 1910 (cited under F. suberosa in FI. Kouy-Tchéou; duplicate in A. A.); Lo-fou, J. Cavalerie, no. 3595, March 1909 (holotype of F. Blinii; merotype in A. A.) ; Lo-fou, J. Cava- lerie, no. 3634, Oct. 1908 (holotype of F. Feddei; photo. in A. A.) ; Hua- kiang, J. Cavalerie, no. 2172, June 3, 1904 (syntype of F. Kingiana; photo. in A. A.). In his Flore du Kouy-Tchéou, Léveillé refers F. Fedde as a synonym to F. suberosa, but keeps F. Blinii distinct. Cavalerie’s no. 2172 is cited as a syntype of F. Kingiana by Léveillé, but the characters under which F. Kingiana appears in the key, exclude it, and therefore the other syntype which belongs to F. pyriformis Hook. & Arn. must be considered the real type of the species. 74 JOURNAL OF THE ARNOLD ARBORETUM [voL. xvi Ficus retusa Linnaeus, Mant. 129 (1767). — Léveillé, Cat. Pl. Yun- Nan, 275 (1917). — Handel-Mazzetti, Symb. Sin. 7: 100 (1929). Ficus retusiformis Léveillé in Fedde, Rep. Spec. Nov. 8: 549 (1910); Fl. Kouy-Tchéou, 433 (1915). Cuina. Kweichou; Lou-fou, J. Cavalerie, no. 3601, March 1909 (holotype of F. retusiformis; merotype in A. A.). Ficus retusiformis was first referred to F. retusa as a synonym by Handel-Mazzetti (l.c.). Ficus lacor Hamilton.— Rehder in Jour. Arnold Arb. 10: 124 (1929 Ficus pili Miq. sensu Handel-Mazzetti, Symb. Sin. 7: 92 (1929), uo . F. Tent; vix Mig ie pial unnan: * Kiao-kia, Siméon Ten, no. 733, Jan. 5, 1906, “grand arbre” (holotype of F. Tenii in hb. Lévl., isotype in hb. Ducloux in hb. Univ. Calif.; photos. in A. A.). Ficus Tenu Lévl. was referred by Handel-Mazzetti (l.c.) to F. superba. Ficus pseudoreligiosa Lévl. has smaller longer-stalked recep- tacles and distinctly cuneate leaves. Ficus obscura Blume, Bijdr. 474 (1825). — Gagnepain in Lecomte, Fl. Gén. Indo-Chine, 5: 792 (1928). Ficus asymetrica Léveillé & Vaniot in Mem. Acad. Ci. Art. Barcelona, ser. 3, 6: 147 (Ficus Spec. Chin. 9) (1907); in Fedde, Rep. Spec. Nov. 4:82 (1907). — Léveillé, Fl. Kouy-Tchéou, 429 (1915).— Rehder in Jour. Arnold Arb. 10: 128 (1929). — Synon. nov CuHInA. Kweichou: bord du Hoa-kiang, L. Martin in hb. Bodinier, no. 2577, Feb. 18, 1899 (holotype of F. asymetrica ; merotype in A. A.); Lo-fou, J. Cavalerie, no. 3596, March 1909 (cited in Fl. Kouy- Tchéou [as 3796]; duplicates in A. A.); bord du Ta-ras riviére, J. Esquirol, nos. 2692, July 20, 1911, and 3549 (cited in Fl. Kouy-Tchéou; duplicates in A. A.). Ficus asymetrica agrees well with F. obscura which has been reported from Kweichou by Gagnepain (l.c.) apparently based on Cavalerie 3596 named in the Paris Herbarium F. obscura. It is a species of wide dis- tribution extending from the Malayan Archipelago to Tonkin, southern China and the eastern Himalayas. Ficus longepedata Léveillé & Vaniot.— Rehder in Jour, Arnold Arb, 10: 127 (1929). — Handel-Mazzetti, Symb. Sin. 7: 93 (1929). Handel-Mazzetti gives (l.c.) a complete description of the species, to which he refers F. trichopoda Léveillé and F. sordida Hand.-Mazz. as synonyms. 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 75 Ficus laevis Blume, Bijdr. 437 (1825). Ficus Jamini Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 8: 550 (1910). — Léveillé, Fl. Kouy-Tchéou, 430 (1915).—Rehder in Jour. Arnold Arb. 10: 129 (1929). — Synon. nov. Cutna. Kweichou: Lo-fou, J. Cavalerie, no. 3601, March 1909 (holotype of F. Jamini; merotype in A. A.). The specimen cited represents the typical glabrous form to which be- longs also Henry’s no, 11823 from Yunnan. Ficus scandens Roxburg, Fl. Ind. 3: 536 (1832). Ficus heapicagenigt Bodinier ex Léveillé in Mem. Acad. Ci. Art. Bar- celona, ser. 3, 6: 148 (Ficus Spec. China, 10) etry: in Fedde, Rep. Spec. a 4: 83 (1907). — Synon. nov. Cuina. Hongkong: introduit 4 Hongkong, ot il est main- tenant trés commun sur les murs, E. Bodinier, no, 1253, July 7, 1895 (holotype of F. cantoniensis ; merotype in A. A.). Ficus scandens is not recorded in Hemsley, Index Florae Sinensis, nor in Dunn & Tutcher, Flora of Kwangtung and Hongkong, but it has been collected more recently in southern Yunnan and in Hainan. Ficus foveolata Wall.— Rehder in Jour. Arnold Arb. 10: 124 (1929). — Handel-Mazzetti, Symb. Sin. 7: 95 (1929). Handel-Mazzetti (l.c.) cites F. rufipes Lévl. as a synonym and points out that of the three syntypes cited only Cavalerie 340 belongs here, while Esquirol 75 and 76 represent Rubiaceae; the two last-named syn- types I have not yet seen. Ficus foveolata var. Thunbergii (Maxim.) King.— Rehder in Jour. Arnold Arb. 10: 125 (1929). Ficus stipulata Thunberg, es Ficus, 5, 8 (1786).— Nakai, FI. Sylv. Kor. 19: 124, t. 40 (1932). Ficus hederera Léveillé. — Rehder, l.c. 129 (1929 Ficus Fauriei var. macrocarpa Léveillé in Fedde, Rep. Spec. Nov. 11:65 (1912). Korea. Quelpaert, add: scandens in rupibus, E. Taquet, no. 5972, Dec. 1911 “feuilles et fruits doubles du F. Fauriez’ (holotype of F. Fauriei var. macrocarpa; photo. in A. A.); Hong-no, in declivibus altis, repens in rupibus et tapetem formans, E. Taquet, no. 4425, July 2, 1910 (holotype of F. hederifolia Lévl.; merotype in A. A.). Nakai enumerates under F. stipulate. (l.c.) as synonyms F. hedert- folia Lévl. and F. Fauriei var. macrocarpa Lévl. Ficus Martini Léveillé & Vaniot.— Rehder, l.c. 127 (1929).— Handel-Mazzetti, Symb. Sin. 7: 95 (1929). 76 JOURNAL OF THE ARNOLD ARBORETUM [VOL. xviI Ficus botryoides Léveillé & Vaniot in Mem. Acad. Ci. Art. Barcelona, ser. 3, 6: 148 (Ficus Sp. Chin. 10) (1907); in Fedde, Rep. Spec. Nov. 4:83 (1907). — Léveillé, FI. ane Tchéou, 429 (1915), — Rehder in Jour. Arnold Arb. 10: 128 (1929). Ficus lacrymans Léveillé, Fl. Kouy-Tchéou, 431 (1915).— Rehder in Jour. Arnold Arb. 10: 130 (1929). Ficus kwangtungensis Merrill in Jour. Arnold Arb. 8:3 (1927).— Synon. nov. Cuina. Kweichou, add: Ta-rin, alt. 900 m., J. Esquirol, no. 5373, April 12, 1913, “figuier pleureur” (holotype of F. lacrymans ; merotype in A. A.); environs de Tsin-gay, rochers au bord de la riviére, a Cha-teou-tchay, FE. Bodinier, no. 2653, June 27, 1899 (holotype of F. botryoides ; merotype in A. A.). Handel-Mazzetti |.c. places F. lacrymans with F. Martini, and I now think that he is right; leaves of thinner texture with slight reticulation occur occasionally on specimens from various localities and the habit seems to be extremely variable; on many specimens the branchlets are very long and slender and probably hang down from scandent stems, while other specimens are described as upright shrubs, as is the type of F. kwangtungensis. Though I have not seen the type of the last-named species, I have seen specimens from Kwangtung and Hainan determined by its author. Ficus tikoua Bureau in Jour. de Bot. 2: 213, pl. 7 (1888). — Lé- veillé, Cat. Pl. Yun-Nan, 275 (1917). Ficus Bonatit Léveillé in Fedde, Rep. Spec. Nov. 6: 112 (1908). — 8 ynon. nov. Ficus Bonatiana Léveillé, Cat. Pl. Yun-Nan, 274 (1917). CuinaA. Yunnan: Yunnan-sen, ravins du mont Tchong-chan, F. Ducloux, no. 732, Aug. 18, 1905 (holotype of F. Bonati in herb. Bonati in Univ. Calif.; photo. in A. A.; isotype in herb, New York Bot. Gard.) The specimen cited above differs somewhat from most of the material in this herbarium in the uniformly short petioles not exceeding 1 cm. Ficus erecta Thunberg. — Rehder in Jour. Arnold Arb. 10: 125 (1929). — Handel-Mazzetti, Symb. Sin. 7: 100 (1929). — Nakai, FI. Sylv. Kor. 19: 121 (1932). Nakai cites F. Taquetii as a synonym of F. erecta, and Handel- Mazzetti cites as synonyms besides F. pseudopiriformis and F. Taqueti also F. Mairei Lévl. which according to fragments from the type speci- men belongs to F. heteromorpha Hemsl. (see Rehder, op. cit. 126). 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 77 Ficus erecta Thunb. var. Sieboldi (Miq.) King in Ann. Bot. Gard. Calcutta, 1: 142 (1888).— Nakai, Fl. Sylv. Kor. 19: 123, t. 39 (1932). Ficus Bhi piriformis Léveillé & Vaniot in Fedde, Ree. Spec. Nov. 5: (1908).— Nakai in Jour. Coll. Sci. Tokyo, 31: 199 (FI. ah (1911). Ficus erecta Thunb. — Rehder in Jour. Arnold Arb. 10: 125 (1929), in Korea. Quelpaert: Hong-no, secus torrentes, U. Faurie, no. 1991, July 1907; in petrosis, U. Faurie, 2024, July 1907 (syntypes of F. pseudo-piriformis ; fragments of no. 1991 in A. A.; no. 2024 ex Lé- veillé) ; secus torrentes, alt. 400 m., E. Taquet, no. 4422, July 28, 1910 (in hb, Léveillé sub F. pseudo-piriformis; merotype in A. A.). According to Léveillé’s description Faurie no. 2024 belongs here, but an isotype of this number in the A. A. herbarium is typical F. erecta. On the other hand, a duplicate of Taquet no. 900, which is cited by Léveillé under F. Taqueti, belongs to var. Sieboldi. Ficus Taqueti, according to Léveillé’s description, represents typical F. erecta, while F. pseudo-piriformis agrees with var. Sieboldi. Nakai (l.c. 1932) cites F. pseudo-piriformis under F. erecta var. Sieboldi. Ficus hirta Vahl. — Rehder, l].c. 126 (1929). — Handel-Mazzetti, Symb. Sin. 7: 99, 100 (1929). Handel-Mazzetti (l.c. 100) identifies F. Porteri Lévl. & Vant. with F. hirta Vahl. Ficus pandurata Hance in Ann. Sci. Nat. ser. 4, 18: 229 (1862). Ficus cuneata Léveillé & Vaniot in Mem. Acad. Ci. Art. Barcelona, ser. 3, 6: 149 (Ficus Sp. Chin. 11) (1907); in Fedde, Rep. Spec. Nov. 4:84 (1907). — Léveillé, Fl. Kouy-Tchéou, 429 (1915).— Rehder in Jour. Arnold Arb. 10: 128 (1929).— Pro parte, quoad specim. Bodinier 2363. — Non Blume, nec Wallich, nec Miquel. CHInA. Kweichou: environs de Kouy-yang, mont du Col- lege, gorge de Yan-pa, rocailles (May 16, 1898), environs de Tou-chan (May 31), E. Bodinier, no. 2363, “arbuste” (syntype of F. cuneata; photo. in A, A.). On the label of Bodinier’s no. 2363, two localities are given, but there is only one specimen on the sheet; the name F. cuneata does not appear on the label. Here belongs probably also F. comata Hand.-Mazz. as suggested by F. P. Metcalf in a note on one of the sheets of that species. Ficus heteromorpha Hemsley. — Rehder in Jour. Arnold Arb. 10: 126 (1929). — Handel-Mazzetti, Symb. Sin. 7: 98, 100 (1929). 78 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XVI Ficus cuneata Léveillé & Vaniot in Mem. Acad. Ci. Art. Barcelona, ser. 3, 6: 149 (Ficus Sp. Chin. 11) (1907); in Fedde, Rep. Spec. Nov. 4:84 (1907). — Léveillé, Fl. Kouy-Tchéou, 429 (1915). — Rehder in Jour. Arnold Arb. 10: 128 (1929).— Pro parte, quoad specim. Cavalerie 1351. — Non Blume, nec Wallich, nec Miquel. — Synon. nov. Cuina. Kweichou, add: Pin-fa, J. Cavalerie, no. 1351, Sept. 8, 1903, (syntype of F. cuneata; photo. in A. A.). The other syntype of F. cuneata differing in the stalked receptacles seems referable to F. pandurata Hee. Handel-Mazzetti (l.c. 100) refers F. pinfaensis to his F. comata, but the type of F. pinfaensis, Cavalerie 532, differs from F. comata in the sessile fruits, longer petioles and in the venation of the generally oblong and quite glabrous leaves rounded at base. To F. heteromorpha Handel- Mazzetti refers F. Stapfii Lévl., but Léveillé’s species, though apparently related to F. heteromorpha, differs in its leaves being rather densely hirsute on both sides and scarcely acuminate. Ficus formosana Maximowicz. — Rehder in Jour. Arnold Arb. 10: 127 (1929). — Handel-Mazzetti, Symb. Sin. 7: 98, 100 (1929). Ficus ane Léveillé, FI. Kouy-Tchéou 433 (1915), p.p., quoad specim. Esquirol 3267. — Non Léveillé & Vaniot (1907). Cuina. Kweichou: Teng-Tchéou, jardin du P. Marchand, Esquirol, no. 3267, June 22, 1902 (cited in Fl. Kouy-Tchéou; duplicate in A, A.). The type of F. rhomboidalis belongs to F. gibbosa var. cuspidifera (Miq.) King. Ficus lageniformis Lévl. & Vant. was referred by Handel-Mazzetti (l.c. 100) to F. formosana, as already identified by the writer (1.c.) Ficus laceratifolia Léveillé & Vaniot.— Rehder in Jour. Arnold Arb. 10: 130 (1929). This species has been referred tentatively by Handel-Mazzetti (Symb. Sin. 7: 100) to his F. comata as a form with lacerated leaves, but looks quite distinct. It is represented in this herbarium by a photograph of the type with fragments and by two specimens collected by W. P. Fang in Szechuan, no. 2415, from Mt. Omei and no. 5645 from Nanchuan Hsien. Ficus Cavaleriei Léveillé & Vaniot.— Rehder in Jour. Arnold Arb. 10: 128 (1929). This species seems related to F. pyriformis Hook. & Arn., but it is quite glabrous and the veins of the leaves diverge at nearly right angles; 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 79 it also resembles F. stenophylla Hemsley, but the leaves are broader and gradually narrowed from the middle to the ends. Ficus pyriformis Hooker & Arnott, Bot. Beechey Voy. 216 (1841). — Léveillé, Cat. Pl. Yun-Nan, 275 (1917), “piriformis.” — Handel-Mazzetti, Symb. Sin. 7: 98, 100 (1929). Ficus Nerium Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 4: 66 (1907). — Léveillé, Fl. Kouy- oc 432 (1915).—Rehder in Jour. Arnold Arb. 10: 130 (1929 Ficus Kingiana Léveillé, FI. nee 431 (1915), pro parte; non Hemsley. — Synon. nov. CuInA. Kweichou: without locality, J. Esquirol (holotype of F. Nerium,; fragments in A. A.); Tchen-fong, J. Esguirol, no. 914 (cited in Fl. Kouy-Tchéou; duplicate in A. A.); Ly-po, J. Cavalerie, Feb. 15, 1900 (syntype of F. Kingiana; merotype in A. A.) Ficus Nerium was first referred to F. pyriformis by Handel-Mazzetti (l.c. 100). Ficus Kingiana appears in Fl. Kouy-Tchéou without de- scription, but as its characters can be partly ascertained from the key, it is not a nomen nudum; of the two syntypes, only the specimen from Ly-po belongs here and must be considered the type (lectotype) since it fits the characters given in the key, while the other specimen does not. Ficus pyriformis Hook. & Arn. var. ischnopoda King in Ann. Bot. Gard. Calcutta, 1: 158, pl. 201c (1888). Ficus macropodocarpa Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 4:66 (1907). — Léveillé, Fl. Kouy-Tchéou, 431 ee — Rehder in Jour. Arnold Arb. 10: 130 (1929). — Synon Cuina. Kweichou: without locality, a hae. no. 838 (holotype of F. macropodocarpa; merotype in A. A.); riviére de Lou- fou, J. Esquirol, no. 2216, Sept. 1910 (cited in Fl. Kouy-Tchéou; dupli- cate in A. A.). Ficus macropodocarpa was first referred by Handel-Mazzetti to F. pyriformis (l.c. 100); on account of the long-peduncled receptacles, it seems referable to var. ischnopoda. Ficus Esquirolii Léveillé & Vaniot in Mem. Acad. Ci. Art. Barce- lona, ser. 3, 6: 150 (Ficus Spec. Chin. 12) (1907); in Fedde, Rep. Spec. Nov. 4: 84 (1907). — Léveillé, Fl. Kouy-Tchéou, 430 (1915). Ficus stenophylla Hemsley in Hooker’s Icon. 26: t. 2536 (1897), p-p., excl. parte typ. — Handel-Mazzetti, Symb. Sin. 7: 98 (1929). Cut1na. Kweichou: préfecture de Hin-y-fou, J. Esquirol in hb. Bodinier, no. 2588, April 1899 (holotype of F. Esquirolii ; photo. in A. A.) 80 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII This species seems to be closely related to F. pyriformis Hook. & Arn. and to F. stenophylla Hemsl.; from the former it differs in the narrower long-caudate leaves, subsessile subglobose receptacle covered with short appressed hairs, and from the latter by the thinner caudate leaves slightly hairy below and the subsessile subglobose and hairy receptacle. Hemsley based his species mainly on Henry’s no. 8716 from Hainan which has to be considered the type of his species and is the plant figured on the plate except figure 7 representing a single receptacle which is from Faber’s no. 446 and apparently belongs to F. Esquirolii. The de- scription also applies mainly to the Hainan plant if under “Reptacula” the words “‘subsessilia vel . . . vel sphaeroidea . . . vel leviter puberula” are taken out. Typical F. stenophylla is represented in this herbarium by plants from Hainan (W. T. Tsang no. 378, McClure, no. 9584), Kwangtung (C. L. Tso, no. 21048, Levine & Kwok, no. 875), Fukien (L. Y. Tai, no. 11673), Chekiang (R. C. Ching, no. 1692) and from Kweichou (Steward, Chiao & Cheo, no. 877); the last named specimen seems to approach F, Esquirolii in the shorter stalk of the receptacle. Ficus Esquirolii is represented besides by the photograph of the type also by Handel-Mazzetti, no. 10695, Y. Tsiang, no. 6053 and Steward, Chiao & Cheo, no. 147, all from Kweichou. Under Ficus Esquiroliana Léveillé (in Bull. Geog. Bot. 24: 252) says of F. Esquirolii “extra genus probata est” but he still enumerates it the following year in his Flore du Kouy-Tchéou as F.. Esquirolii, while he changes F, Esquiroliana to F. laus-Esquirolit. Ficus silhetensis Miquel in Ann. Mus. Bot. Lugd.-Bat. 3: 223, 291 (1867). — Léveillé, Cat. Pl. Yun-Nan, 275 (1917). Ficus Cyanus Léveillé & Vaniot.—Rehder in Jour. Arnold Arb. 10: 129 (1929) — Synon. nov. Ficus Cyanus var. viridescens Léveillé & Vaniot in Mem. Acad. Ci Art. Barcelona, ser. 3, 6: 149 (Ficus Spec. Chin. 11) (1907); in Fedde, Rep. Spec. Nov. 4: 84 (1907). sh! i la Léveillé & Vaniot in Léveillé, Fl. Kouy-Tchéou, 429 15).— ae in Jour. Arnold Arb. 10:128 (1929).— Seay no Cu1na. Kweichou:_ Si-liéou-gay, trous profonds, J. Cava- lerie, no. 169, July 3, 1902 (holotype of F. Cyanus;; merotype in A. A.) ; arrivée a Tong-kai (ruisseau), J. Esquirol, no. 3031, July 25, 1911 (cited in Fl. Kouy-Tchéou under F. Cyanus; photo, and duplicate in A. A.); Pin-fa, J. Cavalerie, no. 1099, June 23, 1903 (holotype of F. Cyanus var. viridescens; photo. in A. A.); Tché-chou, bois, J. Esquirol, no, 2506, Sept. 1909 (holotype of F. congesta; merotype in A. A.). Ficus Cyanus differs slightly from F. silhetensis as figured and de- 1936] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 81 scribed by King (in Ann. Bot. Gard. Calcutta, 1: 154, pl. 194, 1888) in the shorter petioles and cuneate base of the leaf, but otherwise agrees. The variety viridescens does not seem to differ from typical F. Cyanus, and in Flore du Kouy-Tchéou, Léveillé enumerates it under the species. Ficus congesta resembles closely Cavalerie’s no. 1099 (F. Cyanus var. viridescens) except that the leaves are somewhat smaller. The species has also been collected in Kweichou by Y. Tsiang, no. 4404, determined as F. silhetensis Miq. by E. D. Merrill. Ficus Stapfii Léveillé.— Rehder in Jour. Arnold Arb. 10: 131 (1929). Ficus Stapfi is represented only by a sterile specimen (photo. and fragments in A. A.). It was referred by Handel-Mazzetti (Symb. Sin. 7: 100) tentatively to F. heteromorpha Hemsl., but differs in its rather dense pubescence; in this respect it resembles F. silhetensis Miq., but the leaves are oblong and obtusish to acutish, not acuminate at the apex. Ficus Abelii Miquel in Ann. Mus. Bot. Lugd.-Bat. 3: 281 (1867). Ficus Schinzii Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 8: 550 (1910). — Léveillé, Fl. Kouy-Tchéou, 433 (1915).—Rehder in Jour. Arnold Arb, 10: 131 (1929).— Non F. Schinziana Warb. — Synon. nov. Cuina. Kweichou: Lo-fu, J. Cavalerie, no. 3592, Apr. 1909 (holotype of F. Schinzii; merotype in A. A.); Tsao-ly, bord de riviére, alt. 600 m., J. Esquirol, no. 3503, Feb. 5, 1912 (cited in Fl. Kouy- Tchéou; duplicate in A. A.). By Handel-Mazzetti (l.c. 100) F. Schinzii was referred to F. pyri- formis and its allies. It certainly belongs to F. Abelii which has not yet been recorded from Kweichou, though it has been collected in Szechuan and Hunan. Ficus Roxburghii Wallich. Num. List. 4508 (1832).— Handel- Mazzetti, Symb. Sin. 7: 100 (1929). Ficus macrocarpa Léveillé & Vaniot in Mem. Acad. Ci. Art. Barcelona, ser. 3, 6: 152 (Ficus Spec. Chin. 14) (1907); in Fedde, Rep. Spec. Nov. 4:85 (1907). — Léveillé, Fl. Kouy-Tchéou, 431 (1915), sub F. macrocarpa Wight.—Rehder in Jour. Arnold Arb. 10: 130 (1929).— Non Wight. Cuina. Kweichou: without precise locality, J. Cavalerie in 1899 (holotype of F. macrocarpa; ex Léveillé); ouest de Lo-fou, J. Cavalerie, no. 3593, March 1909, “bel arbre” (cited in Fl. Kouy-Tchéou; photo. in A. A.). Handel-Mazzetti refers to F. Roxburghii besides F. macrocarpa also F, Letaqui Lévl. & Vant. which see below. 82 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xv Ficus Roxburghii Wall. must be considered a valid publication of a new name, since he cites as a synonym “F. macrophylla Hb. Roxb. (haud Desf.)” a species described in Roxburgh, Fl. Ind. 3: 556 (1832). Ficus Letaqui Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 8: 550 (1910). — Léveillé, Fl. Kouy-Tchéou, 531 (1915).— Rehder in Jour. Arnold Arb. 10: 130 (1929). This species was referred by Handel-Mazzetti (Symb. Sin. 7: 100) to F. Roxburghii, but it differs markedly from F. Roxburghii in the much smaller subglobose and subsessile receptacles, borne mostly in pairs on the nodes of long and slender leafless shoots; the leaves are similar in shape, but scabrid above and borne on longer and slenderer petioles. Ficus Vanioti Léveillé, Fl. Kouy-Tchéou, 434 (1915), pro parte. Cutna. Kweichou: bois sacré au Che-chen de Na-é, J. Esquirol, nos. 3077 and 3078, end of June and July 1 (syntype of F. Vanioti; photo. in A. A.). The name F. Vanioti was published in Flore du Kouy-Tchéou with F. ouangliensis as a synonym and was intended apparently as a new name for that species which is based on Cavalerie no, 2568, and is not a Ficus, but an Aglaia. The two specimens cited above, however, repre- sent a species of Ficus, but the material is so fragmentary, consisting only of bare branches, broken leaves and a few broken receptacles, that exact identification is not possible; the material seems to belong in the affinity of F. formosana Maxim. LORANTHACEAE Tolypanthus Maclurei (Merr.) Danser in Bull. Jard. Bot. Buiten- zorg, ser. 3, 10: 355 (1929). nme Maclurei Merrill in Philipp. Jour. Sci. 21: 494 (1922). Loranthus Esquirolii Léveillé, Rev. Ann. Chine, 1916: 21 (MS).— Nomen non rite sabliontus, Cuina. Kweichou: Pin-fa, Heou-tcheng, J. Esquirol, no. 52, “liane parasite sur le bibassier” (type of Loranthus Esquirolii; photo. in A. A.). Léveillé’s name Loranthus Esquirolii appeared in a manuscript publi- cation, of which only a few copies were distributed and cannot be con- sdeced validly published; Merrill’s specific epithet therefore, remains vali The name Loranthus Esquirolii does not appear on the original label a the type specimen which was identified by Danser as Tolypan- thus Maclurei. (To be continued) HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1936] FLORY, CHROMOSOMES IN GYMNOSPERMS 83 CHROMOSOME NUMBERS AND PHYLOGENY IN THE GYMNOSPERMS WALTER S. FLory With two text figures CRITICAL sTuDIES of chromosome morphology in most groups of living gymnosperms have been published during the past few years. Geitler (1929) and Florin (1932) have described the chromosomes of Wel- wttschia and Ephedra of the order Gnetales. Sax and Beal (1934) inves- tigated the chromosome situation in all nine genera of the Cycadales. The most comprehensive chromosome survey in the gymnosperms is that of the Coniferales by Sax and Sax (1933). Fifty-three species from six- teen genera of the subfamilies Cupressineae, Taxodineae, Abietineae, and Taxineae were studied. The sole surviving member of the Gink- goales was included in the last mentioned analysis and has been the object of cytological studies by other workers as well. This leaves only the sub- families Araucarineae and Podocarpineae and the genus Gnetum with representatives unanalyzed in recent cytological papers. With regard to two of these groups Sax and Beal (1934) write, ‘“ .. . recent studies ae that Podocarpus has 20 pairs of chromosomes’; and, also . Gnetum has a large number of relatively small eheoiiosomnien in erie: accounts Burlingame, Ishikawa and Schurhoff (from Sax and Sax, 1933, p. 368) have found 12 chromosomes in several species of Podocarpus. Only two genera are included in the subfamily Araucarineae. They are Araucaria and Agathis. An early morphological study (Lopriore, from Coulter and Chamberlain) on one species of Araucaria indicated the presence of 12 pairs of chromosomes here. No references have been found to critical chromosome work on either genus of this subfamily. Questions concerning the phylogenetic sequence of conifer subfamilies involve the Araucarineae at important points. One question concerns the comparative antiquity of the Araucarineae and Abietineae; another the relationship between the Araucarineae and the Podocarpineae. This investigation was made to see if a knowledge of chromosome number and morphology in the Araucarineae would aid in solving problems of the nature mentioned. To further that end, the chromosomes of several species of Podocarpus have been examined, and the results obtained will be briefly discussed. 84 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Material for this work has been secured from several sources. The greenhouses of the Bussey Institution, the Arnold Arboretum, and the Biological Institute of Harvard University, have been sources of some species of Agathis and Araucaria. Root tips of two species of Podocarpus and of two species of Araucaria were obtained at Wellesley College. Other material of Araucaria and Podocarpus was collected at the Brook- lyn Botanic Garden and at the New York Botanical Garden. We realize that the species names of such specimens may not be correct, but it was not possible to verify the classification of these immature plants. Absolutely accurate chromosome counts of conifer species are difficult to secure in root tip cells, — endosperm material being much preferred for preparations of somatic chromosomes. However, root tips were the only material available for the present study. The chromosomes of all species examined were studied in aceto-carmine smears of root tip cells. Some recent advances in technique allow somewhat more critical prepa- rations to be made than were formerly possible. A modification of Warmke’s (1935) method of making temporary root tip smears was used. Root tips were fixed in aceto-alcohol overnight; treated with a hydrochloric acid and 95% alcohol mixture; hardened in aceto-alcohol ; then sliced and macerated. After the preparations had been allowed to stand for from 5 to 15 minutes, they were heated almost to boiling; then considerable pressure was exerted upon the cover glass, and this was fol- lowed by a second heating, after which the mounts were sealed. By this method metaphase plates were secured, in several species, in which the chromosomes were well scattered, enabling them to be readily counted. Rehder’s (1927) classification has been followed with regard to divi- sion of the Coniferales, and elsewhere when possible. This has been sup- plemented by use of Pilger’s (1926) classification for the forms not gen- erally grown in the temperate United States. DESCRIPTION Agathis robusta, Araucaria brasiliana, and A, Cunninghamii each con- tains 13 pairs of chromosomes. Material from a plant of Agathis bearing a doubtful specific name also showed 26 somatic chromosomes. Numer- ous plates from Araucaria Bidwilli were also studied. The chromosome picture appeared to be identical with that found in the other two species of Araucaria studied. Counts of a number of metaphase figures seen in side view indicated the presence of 26 somatic chromosomes in this spe- cies. The chromosomes of species belonging to the same genus were very similar. The chromosomes of Agathis robusta (Fig. 1) and of Araucaria brasiliana (Fig. 2) were studied in more detail than those of the other 1936] FLORY, CHROMOSOMES IN GYMNOSPERMS 85 species. In each of these species there were 9 pairs of chromosomes in which the arms were about equal in length, or in which the longest arm was, at most, not twice the length of the shortest. In addition to these chromosomes with median or submedian attachments, both species apparently have 4 pairs with nearly terminal attachment points. The chromosomes of Agathis and Araucaria are similar not only in morphol- ogy but also in size. Chromosomes of Agathis, in aceto-carmine, vary in length from 13 u to 22 uw; those of Araucaria from 12.5 uto 20 u. In Araucaria the elements appear to be slightly greater in diameter than in Agathis, but the difference is too small to be significant. TEXT FIGURE 1 Drawings of chromosomes of root tip cells. x 900. Fig. 1. Agathis robusta, 26 chromosomes. Fig. 2. Araucaria brasiliana, 26 chromosomes. Fig. 3. Podocarpus falcatus, 24 chromosomes. Fig. 4. Podocarpus macro- phyllus, 38 chromosomes. Fig. 5. Podocarpus neriifolius, 36 chromosomes. The chromosomes of five forms of Podocarpus were examined. These were P. macrophyllus, P. falcatus, P. neriifolius, P. chinensis, and P. andinus. Pilger indicates that P. chinensis is merely a form of P. macro- phyllus, which is cultivated in botanic gardens, and is not a separate species. The first four forms listed belong in the generic section Eupodo- carpus. Podocarpus andinus is placed in Pilger’s section Stachycarpus. In these five forms there are two different types of genoms. The first type of genom was found in a species growing in the greenhouse at Wel- lesley College and tentatively identified by Dr. Alice Ottley as P. fal- 86 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII catus R. Br. Several plates quite clearly showed 24 chromosomes in root tip cells of this plant (Fig. 3). These plates, quite apparently, were almost complete. Each, however, had its elements quite scattered, owing to pressure; and it is possible that a few chromosomes were lost. Of the 24 chromosomes observed, 2 pairs were approximately isobrachial; the other 10 pairs were distinctly heterobrachial. Chromosome measure- ments were, for the longest, 15 ; for the shortest, 5.5 u; and the average length was approximately 11 u. In morphology the complement is similar to that found in many conifers. However, the chromosomes are smaller than those found in most species. In P. macrophyllus many plates were found which showed all but a few chromosomes distinct and unentangled (Fig. 4), but no complete plate was beyond question at some point. The number is about 38, many plates seeming to have this count. In several cases there was a suggestion of 39 chromosomes. No plate seemed to have 40 but per- haps, for safety, the number should be placed as between 38 and 40. The best figures, those appearing to contain 38 elements, were analyzed. Ten pairs are heterobrachial, 3 of these approaching the isobrachial con- dition. Nine pairs appear to have attachment points that are terminal, or very nearly so. Chromosome measurements give, for the longest pair, 12.5 uw; for the shortest, 3.5 4; and the average length was about 7.5 uw. The P. nertifolius complement has much the same appearance as that of the species just described. Approximately 38 chromosomes are pres- ent. However, the majority of these appear terminal or nearly so, the only apparent exceptions being one short isobrachial pair, and the longest pair, which have submedian attachment constrictions. The longest chromosomes are 12.5 ut, the shortest 3 u. Figure 5 shows 36 well-spaced chromosomes in a plate from which a few have been lost. Podocar pus chinensis and P. andinus (belonging to a different section from the other podocarps mentioned here) were seen to have the same general type of complement as P. macrophyllus, having nearly 40 chro- mosomes, many with terminal or near terminal attachment points. CHROMOSOME NUMBERS AND PHYLOGENY Numerous morphological studies have been made on gymnosperms. These, with discussions of suggested evolutionary tendencies, are best summarized by Chamberlain (1935 a and b). A diagram (Text fig. 2) representing the phylogeny of gymnosperm orders and families, as rep- resented by the best available evidence, has been made. In this diagram ideas and material from. a number of sources (chiefly: Chamberlain, 1935b, diagram p. 184; Jeffrey, 1917, especially pp. 355-6; Sax and Sax, 1936 | FLORY, CHROMOSOMES IN GYMNOSPERMS 87 1. c.; Sax and Beal, 1.c.; Florin, 1. c.; Matsuura and Suto, 1935; and the present work) are incorporated in one figure. Since representatives of all major gymnosperm groups have now been subjects of critical cyto- logical examination, it seemed appropriate that some effort be made to see how the cytological facts fit in with those of grosser morphology. A description of Text figure 2 follows. Beneath each major division are placed its genera from which representatives have been cytologically Zami 8 i) Pinus 12 Cryptowmerial iy St 8 = 5 : : her ia a; = noi" Cedrus IL Taxodium [ut eratozama 6 3 Encephalartos i Un 8 3 "3 7F aes : ; ‘. phals g 4 2 v o|tl Picea " f]3 Boweni 5 0 ¥ 12 = Ts uga lo “ Dioon q wW F = Foo Abie 5 : ™Macreza icf 0 2 ; ; +2 a mua Oo w Oe Poevdotsu 0) A Cycas st S c z 7 al Yala Y ca £|'3 £ evdolarix]22 sla 9} 39 Microcycas 13 a ALC c ee ras "a c : | = v © ) q y 3 = E rie : £ a 2 v 3 ad “SE ‘ Cycadales 8B Coni fev ates leo TEXT FIGURE 2 Phylogenetic relationships of the living gymnosperms based on morpho- logical characters, and the basic chromosome numbers of the genera in the various families. analyzed. The chromosome count is given in basic haploid numbers. Extinct orders are indicated by four small perpendicular lines, placed on the main evolutionary lines at about the geological level at which ex- tinction occurred. The most doubtful points in the suggested relation- ships are shown by dotted lines or by question marks. There are apparently three basic groups of chromosome numbers in 88 JOURNAL OF THE ARNOLD ARBORETUM [voL. xvi the living gymnosperms. The basic numbers are predominantly 8 or 9 for the Cycadales, 7 for the Gnetales, and 12 for the Ginkgoales and Coniferales. The deviations from these basic numbers in these groups can be attributed to elimination of one or several chromosomes following segmental interchange or polyploidy, and an increase in chromosome number resulting from fragmentation or chromosome duplication. Polyploidy is rare in the gymnosperms. The chromosomes of species of gymnosperms are much more uniform in number, size, and general morphology than those in the species of angiosperms. The count of 38 somatic chromosomes in certain species of Podocarpus at first suggested a basic number of 19. This is the basic number in certain primitive angiosperms, and Anderson (1934) has suggested an allopolyploid origin of such angiosperms from primitive gymnosperms with basic chromosome numbers of 7 and 12. The presence of about 12 pairs of chromosomes in one species of Podocarpus suggests, however, that the forms with about 38 somatic chromosomes may be triploids with 13 chromosomes as the basic number, or that chromosome fragmenta- tion at the fiber attachment point has effected the change in these genoms. The cytological analysis of the Coniferales seems to show some corre- lation with the morphological evidence regarding family relationships. The Cupressineae and Taxodineae have the same basic chromosome numbers. The basic chromosome number is 13 in the Araucarineae and is 12 or 13 for the Podocarpineae. The presence of 12 chromosomes in most genera of the Abietineae supports the theory that this may be the main line from which the other families have been derived. The basic number of 12 chromosomes in the Ginkgoales also supports the sugges- tion that 12 is the basic number from which both families, Taxaceae and Pinaceae, originated. SUMMARY The basic chromosome number is 13 for Agathis and Araucaria, and the genoms are similar in morphology. Chromosome numbers were obtained in five species of Podocarpus. In one species about 24 somatic chromosomes were counted, but in the other species there were about 38 chromosomes in the root tip cells. The basic chromosome number is 12 or 13, and the four species with about 38 chromosomes presumably are the result of polyploidy or chromosome fragmentation. A survey of chromosome numbers in the gymnosperms indicates three basic numbers, — 8 or 9 for the Cycadales, 7 for the Gnetales, and 12 for the Ginkgoales and Coniferales. There is some correlation between 1936] FLORY, CHROMOSOMES IN GYMNOSPERMS 89 chromosome numbers and morphological relationships of the various taxonomic units. This work was done at the Bussey Institution of Harvard University while the author was a National Research Council Fellow. LITERATURE CITED ANDERSON, EpcGar (1934). Origin of the Angiosperms. (Nature, 133: 462. CHAMBERLAIN, C. J. (1935). Gymnosperms, structure and evolution. Univ. of Chicago Press a 5). The Gymnosperms. (Bot. Rev. 1: 183-209.) Coutter, J. M. and C. J. CHAMBERLAIN (1910). Morphology of Gymno- sperms. one of Chicago Press. FLorIn, R. (1932). Die ee ean ae bei Welwitschia und einigen Ephedra-Arten. (Svensk Bot. Tidskr. 26: 205-214. GEITLER, L. (1929). Zur Zytologie von Ephedra. (Oester. Bot. Zeitschr. 8: 242 ff. ) JEFFREY, E. C. Ces The anatomy of woody plants. University of Chic cago Matsuura, H. and T. Suto or pas eget to the idiogram aired in Bh oe Heke plants. I. (Jour. Fac . Hokkaido Imp. Uni 5: 33-75. Pricer, R. (1926). Gymmospermae (in Engler and Prantl’s “Die nattr- en Pflanzenfamilien,” 2d ed. REHDER, ALFRED (1927). Manual of cultivated trees and shrubs. The Macmillan Co., New York. Sax, Kari and J. M. BEAL (1934). Chromosomes of the Cycadales. (Jour. Arnold Arb. 15: 255-258. Sax, Kart and Hatty J. Sax (193 3). Chromosome number and morphol- ogy in fe conifers. (Jour. Arnold Arb. 14: 356-375.) CYTOLOGICAL LABORATORY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 90 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII APOSPORIC PARTHENOGENESIS IN A TRIPLOID APPLE, MALUS HUPEHENSIS Haic DERMEN With plates 179-181 A NUMBER of years ago Dr. Karl Sax, in his attempt to make inter- generic crosses in Pomoideae, found that whenever Malus hupehensts was used as the female parent, seeds were always formed. Preliminary experiments were made to discover the basis of this peculiarity (Sax, 1931) and it was found that this species behaved as a typical partheno- genetic plant. I was designated to investigate the problem and the present report is a cytological analysis of this phenomenon. Malus hupehensis (Pamp.) Rehd. (M. theifera Rehd.) (Rehder, 1933) is known to be a pure species, distributed widely in mountain regions of China extending south into Assam. It was introduced into this country in 1900. All plants of this species growing in the Arnold Arboretum, some of which have been propagated from seeds here and some brought in from outside sources, fruit heavily; the fruits con- taining the usual number of seeds. All the seeds appear to be deformed in outline in comparison with those of other species. MATERIAL AND METHODS In order to have a sounder basis for our conclusions concerning par- thenogenetic behavior in M. hupehcnsis it was thought advisable to make parallel observations on another species, the embryonic development of which was normal. Malus arnoldiana (Rehd.) Sarg. (M. baccata X floribunda), a hybrid, chosen as a control plant, produces an abundance of fruit, and normal seeds in contrast to M. hupehensis. In the summer of 1931 a preliminary examination was made of the chromosome number from root-tips of a M. hupehensis seedling and it was found to be a triploid form. The analysis of the ovules was made beginning the first week of May and continuing through the summer during the years 1932, 1933 and 1934. The study of this problem dur- ing the first two years was exploratory in character and it was in 1934 that a more complete analysis of the ovules of both M. arnoldiana and M. hupehensis was made. Specimens at different stages of ovular development were fixed sepa- rately. The fixing of material was begun during the first week of May 1936 ] DERMEN, PARTHENOGENESIS IN MALUS 91 when the buds began to turn pink and the development of the female meiotic phase was initiated. This phase is completed just before shed- ding of petals when the egg apparatus in the embryo sac (ES) is formed and fertilization has taken place (normally). From the pink stage to the time of shedding of petals six separate collections were made. Two collections were made between shedding of petals and remaining parts of perianth. After that the collections were continued at 5 to 7 day intervals in order to follow up the sequence of embryo development. In fixing material for early stages, the fleshy part of the ovary is trimmed off with a sharp knife in order that the fixative may penetrate the ovules quickly. For later stages young seeds are removed from the fruit and are fixed directly. For quick penetration of fixative in older seeds the seed coat may be slightly slashed. For more mature seeds the whole testa may be removed. The first year Flemming’s solution was used as fixative but was found unsuitable for this material. Navashin’s solution was found preferable for the meiotic and earlier stages of embryo sac formation. However, being doubtful of the effect of the latter fixative on later stages, Lewitz- ky’s solution (5% formalin and 0.5% chromic acid) was used for those, since I had found earlier (Dermen, 1933) that it facilitates at least the staining of nucleoli with either crystal violet-iodine or haematoxylin. In this way I hoped to have at least the nucleoli to go by as landmarks, so to speak, in determining the nuclear number and other features in the development of embryo sacs. The results eventually showed this assump- tion to be justified. To prevent severe distortion of embryo sacs in changing fixed material into alcohol for paraffin sectioning, the material was allowed to remain in the fixative for twelve hours or more, washed in water and run up in alcohol gradually beginning with 5%. The results later showed that there was some shrinkage of the embryo sac and endo- sperm layer away from the nucellar tissue, though not enough to make observations confused. Paraffin sections of early stages to the beginning of ES formation were cut 20 u thick and from ES formation on up 35-40 u, thick enough to obtain whole embryo sacs and young embryos, in order that confusion might be avoided in studying the origin and sequence of development. Meiotic stages were stained with crystal violet-iodine, and ES and later stages with haematoxylin. Root-tips were fixed in Lewitzky’s solution and stained with crystal violet-iodine; while for meiosis in the anther aceto-carmine technique was used. Due to technical difficulties in illustrating various phases, it was necessary to use different scales. The scale used is indicated under each drawing. 92 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII DESCRIPTION The features illustrated which were considered important in the analysis of parthenogenetic behavior in M. hupehensis are presented in 38 figures in three plates. The drawings are made in diagrammatic form to bring out the points pertinent to this study and to give the most essential details of origin and development of both embryo sac and embryo. The comparative illustrations of the control plant, M. arnold- iana, and M. hupehensis run from Figs. 18 and 19 to Figs. 34 and 35 in- clusive. Other figures are from M. hupehensis alone. The somatic chromosome number in M. hupehensis, studied from root-tips, was 3n = 51 (Fig. 1), a triploid number, the diploid being 2n — 34. The next figure (Fig. 2) represents the chromosome set from a tetraploid seedling of M. hupehensis with 4n = 68, the discussion of which will be found elsewhere in this paper. The size of these chromo- somes is about one micron in length. The meiotic development in anthers took place about 10 days earlier than the corresponding phase in the ovules. In the anthers the homolo- gous chromosomes form trivalents and bivalents and a few remain uni- valents (Sax 1932). Figure 3 is II M from anthers with 25+ chromo- somes at each plate. No mature pollen grains are found in the anthers. Some cells begin to degenerate at I M and others at later stages, and most degenerate after they have reached the tetrad stage when the individual microspores fail to break apart although they are fully differ- entiated and each has its own wall. A similar process of degeneration was found in the ovules. Here the chromosomes either remain univalents at I M or pair variously and divide at random, Figure 4 (a and b) represents I M in an egg mother cell (EMC). The chromosomes here are in univalent form. In the two sections, a and b, the total is exacly 51. It cannot be stated with cer- tainty when, if ever, in univalent form there was any division in the EMC. However, in a few instances observed where the chromosomes were unpaired the cell was in a stage of degeneration. Such a cell is shown in Fig. 11. An anaphase stage is given in Fig. 5 (a and b). The b group of chromosomes from the lower pole was in a following serial section. Here also the total number of chromosomes could be accurately counted. Altogether in M. hupehensis over two hundred ovules in the early stages of meiosis were examined and only rarely were there found ovules lacking a meiotic cell, either in prophase, in division stages, or in a de- generative process. There were found some older ovules which lacked embryo sacs. One young ovule was found with two egg mother cells. 1936] DERMEN, PARTHENOGENESIS IN MALUS 93 There were observed many ovules divided into two nucellar regions, a twin ovule, each with a simple or a compound ES, enclosed together within a common integument. An EMC is shown in Fig. 6 at pachytene stage. Usually around a meiotic cell there are one or more somatic cells each with a nucleolus as large or larger than that in the meiotic nucleus, which, as will be pointed out later, may develop into embryo sacs. Very rarely there were found ovules with two egg mother cells. Such a case is shown in Fig. 7. Around these meiotic cells are found a number of large nucleolated somatic cells. Figure 9 is a small scale drawing of the same cell to indicate the distance of a meiotic cell from the upper end of the nucellus. Here the EMC is 6 cell layers deep, while in other ovules the number of layers may vary. In Lilium regale (original observation) this layer was always one layer deep from an early stage to the time of fertilization (later stages were not observed). Figure 10 illustrates a typical young ovule with two layers of integumental tissue as it appears in longitudinal section. The cell in it is the same as in Figs. 8and 9. The tissue around this cell is the nucellus. This figure shows the approximate location of the EMC in the ovule taken as a whole. As was mentioned above the EMC may either degenerate at I M or later as in the anther. A degenerating cell at I M is shown in Fig. 11. Here the cell was found compressed from the sides and the chromosome stain was blurry showing these bodies more or less clumped together. As a whole its degenerated feature could be easily distinguished when compared with other cells at the same stage. As well as could be made out the chromosomes in this cell were not paired, which fact, I believe, may be correlated with degeneration of the cells at this stage. Figures 12 and 13 show a parallelism in division with EMC and in development of one of the resulting cells into, supposedly, an ES. Nor- mally an EMC should divide twice and from this division four cells should result. One of these, which is the furthest from the micropyle should be the megaspore destined to become an ES. However, as will be shown below, this expectation is not fulfilled in M. hupehensis. In Fig. 12 we find three cells. The one nearest to the micropyle has degen- erated. This cell corresponds to one in Fig. 13 which was in metaphase and in a stage of degeneration. In this latter figure the second division was found to be irregular; some of the chromosomes at opposite poles were at a distance from the main chromosome groups. If this cell divided it would result in two unequal sized cells, because these two groups of chromosomes are not found at equal distance from the ends of the cell. This is what is found in Fig. 12. Here one cell is degenerating. 94 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII One of the sister cells from the second division is small, the other large, and in this large cell there is a small chromosome mass outside the main nucleus, indicating that this cell is not an ordinary somatic cell but the product of an irregular second division of an EMC. Direct as well as indirect evidence shows that meiotic cells at some stage of their development degenerate while somatic cells in their imme- diate vicinity usually grow in size and develop into embryo sacs. Figure 14 illustrates this point clearly. Here is shown a rather large somatic cell with a large nucleolus, above which there is a meiotic cell in a stage of degeneration. In Fig. 15 is shown such a cell that has grown in size. The nucleus has divided and the cell is in a two-nucleate stage and beginning to develop into an ES. At the base of this are shown four cells in a row that have annular xylem thickenings, an unexpected mani- festation in an ovule, this type of cell being characteristic of vegetative tissues. Similar cells were found at later stages of ES development and in one case in a complex ES. Serial sections from either side of this were scrutinized and no trace of degenerating meiotic cells was found in the neighborhood of this binucleate ES. Therefore, it may be that these four annulated cells represent the tetrad resulting from two meiotic divi- sions. Their degeneration was expressed by transformation into xylem elements. Such cases were found only in 1932 material. In the majority of cases the embryo sacs were structurally complex. These were found to have resulted from the fusion of two or more embryo sacs. A simple case is shown in Fig. 16 where two embryo sacs were found growing simultaneously. In other cases the number of these young embryo sacs growing side by side may reach six or more. In this figure one is four-nucleate and the other below it two-nucleate. In Fig. 23 two adjacent sacs are shown in mature egg apparatus stages. Figure 17 represents the middle section of five serial sections in which a total of 30 nuclei could be counted which may have resulted from fusion of four embryo sacs, boundaries of which could not be made out clearly. Illustrations from Fig. 18 to Fig. 35 inclusive are diagrammatic rep- resentation of comparative development of embryo sacs and embryos in M. arnoldiana and M. hupehensis. Figures 18 and 19, the former from M. arnoldiana and the latter from M. Aupehensis, are from material fixed at the time of full bloom. Both represent the stage when the embryo sacs have completed the cycle of the development with three antipodal nuclei cut off, each having formed a separate small cell. The main portion of the ES now functions as a distinct body in which there is an egg cell with two synergid cells together forming the egg apparatus, below which the two polar nuclei have approached each other but com- 1936 ] DERMEN, PARTHENOGENESIS IN MALUS 95 plete fusion is delayed, perhaps till joined with a sperm. A similar situation was observed in Lilium regale (original observation) where first the upper polar nucleus is joined with a sperm and these two jointly meet half way with the lower polar nucleus. By the time of fertilization the egg cell is larger than the synergids (Fig. 20, M. arnoldiana). In these drawings the upper limits of the egg and synergid cells are not shown because they were vague in my preparations. In Fig. 19 (M. hupehensis) a dotted nucleolus of one of the polar nuclei and a dotted synergid cell are superimposed on the main portion from an adjacent section in a single drawing. The outer broken line represents the inner limit of the nucellus. A more mature ES is shown in Fig. 20 of M. arnoldiana from material fixed at the time of calyx shedding (the last remnant of the perianth with dried up styles). Here only nucleoli are shown. At this stage the antipodal cells had disappeared and the sac had grown longer but not much wider. At this time of the ES develop- ment fertilization has usually taken place, but perhaps due to climatic conditions some cells remain unfertilized. In the ES shown in Fig. 21 (M. arnoldiana) there is a clear indication that fertilization has taken place, judging by the amount of nucleolus in the egg nucleus in compari- son with the same in the synergids. In the other figures (Figs. 18-20, taking these individually) the size of the nucleoli in the egg and the synergids is remarkably similar. Fertilization in Fig. 21 is further indicated by the fact that there appears the beginning of endosperm nuclear division, due to fusion of the polar nuclei with a male gamete. The embryo sacs shown in Figs. 22, 23 and 24, all of M. hupehensis, are from material fixed at the time of calyx shedding as in Figs. 20 and 21 of M. arnoldiana. These five figures are drawn to the same scale. In all the antipodals have disappeared. Judging from the size of the nucleoli of the cells in the egg apparati in the three figures (22, 23 and 24), as in Fig. 20 pointed out above, there is no indication of fertiliza- tion and these cells will remain unfertilized since the fruits at this time are quite advanced and the styles have long since dried up and dropped off with the remaining part of the perianth. In Fig. 22 we see the beginning of endosperm formation after, perhaps, polar nuclei have fused together and divided without being fertilized. The egg cell grows larger, its nucleus moves to the tip of the cell, the synergids eventually disappear and long after there sets a cell division at the tip of the egg cell. From now on the ES begins to grow lengthwise and extends through the nucellar tissue in a manner similar to pollen tube growth. The nucellar cells in the neighborhood of the sac begin to break down, undoubtedly due to some sort of enzymic action, offering no resistance 96 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xvi to the tubular growth of the ES. In Fig. 23 we find two embryo sacs side by side. The missing parts of these are given at the side of the twin sacs drawn from an adjacent section. Here a fused polar nucleus is shown under one of the cells of the egg apparatus, indicating that polar nuclei may fuse before the endosperm development is set. Figure 24 represents a single ES; one of the cells shown at the side is drawn from an adjacent section. Tubular growth of the ES in M. arnoldiana is in proportion to the embryo growth. About a week after fertilization the number of cells in longitudinal section of the embryo is about six. At this time the tip of the ES is 4 to 2% the length of the nucellus. In M. Aupehensis, when the ES is a derivative of only one somatic cell, the width of the sac is about the same as that of M. arnoldiana, but in the great majority of the ovules examined, the ES has grown disproportionately wider and deeper in the nucellus after reaching its basal limit, independently of the amount of endosperm as well as embryo growth which at this time may still be in the one cell stage. This is undoubtedly due to the compound nature of the ES, as shown for example in Figs. 17 and 23. Figure 25 represents a very young embryo (proembryo) of M. arnoldiana about two weeks after fertilization. At this time the number of cells in the proembryo in longitudinal section averages ten. The tip of the ES has barely reached the lower limit of the nucellus. Figure 26 shows diagrammatically this stage. The broken line in this figure indicates the extent and the limit of the ES while the dotted line at the base shows the basal tissue which is firm and not penetrated by the ES. In Fig. 27 is shown the beginning of proembryo formation in M. hupehensis, two weeks after shedding of petals. The embryo development in M. arnoldiana is regular and rapid com- pared with that of M. hupehensis. About three weeks after fertilization the proembryo in M. arnoldiana differentiates into a suspensor and a true embryo (Fig. 28). In this figure the endosperm, composed of a uninucleate layer, is separated from the nucellar tissue due to shrinkage from fixation. In M. hupehensis, if any embryo development is initiated, the embryos are still in proembryonic form at a corresponding period. Such a stage is shown in Fig. 29. This drawing represents a compound ES in which two proembryos were found, one at the micropylar corner and the other in the free space, away from the nucellus, each undoubt- edly belonging to a separate ES, the limits of which cannot be easily distinguished. A four weeks old embryo of M. arnoldiana is illustrated in Fig. 30 showing the suspensor, the true embryo and more extended growth of endosperm tissue, the broken line indicating its limit of growth. 1936] DERMEN, PARTHENOGENESIS IN MALUS 97 The central region of the ES is still devoid of this tissue. In M. Aupe- hensis, four weeks from the time of flower shedding, the proembryo is still undifferentiated. Such a stage is illustrated in Fig. 31. In this case a single proembryo was found in the middle of a complex ES. Eventu- ally other embryos may develop in the same sac. This particular ES was chosen for illustration merely to indicate that in M. hupehensis embryos develop, as seen in prepared material, directly from the ES and never from any other tissue surrounding or near the ES; otherwise, more often, the embryos are located at or near the micropylar corner. Figure 32 represents a six weeks old embryo in M. arnoldiana when the coty- ledons begin to differentiate. The broken line indicates the limit of a dense endosperm layer next to the nucellar tissue. This layer seems to separate the suspensor from the nucellus and extend between these. The embryo is imbedded in looser endosperm. Figure 33 shows the stage of embryo development in M. hkupehensis six weeks from flower shedding. Here there is the beginning of differentiation of the proembryo into a suspensor and a true embryo with few endosperm nuclei in the sac. Figure 34, drawn to a smaller scale, shows a five weeks embryo from a wide view of the cotyledon in M. arnoldiana. The broken line shows the limit of the dense endosperm, the embryo being imbedded in loose endo- sperm. The outer two solid lines indicate the limit of the integumental tissue. The extent of embryo development in M. hupehensis at a cor- responding period is shown in Fig. 35 drawn to the same scale used in figures preceding Fig. 34. Broken lines indicate the extent of endosperm growth which is shrunk away from the nucellus. The middle area is still devoid of such growth. Often in M. hupehensis there were found twin ovules, each with its own ES, both enclosed in a common testa. In some the demarcation between the twins is quite distinct, in others vague. In Fig. 35 only the upper limits of such twin ovules are shown, where the demarcation, shown by a dotted line, could hardly be distinguished. At the left side of the suspen- sor in this figure there is a bulge which may indicate that at this point an- other true embryo may be initiated resulting in a “siamese twin.” In more mature stages I have observed structurally complex embryos which may have been originated from such a process of development. The above ends the description of comparative illustrations for M. arnoldiana and M. hupehensis. The following description is based on M. hupehensis material given in Figs. 36-38. Figure 36 represents an embryo from material fixed two months after shedding of flowers, which is still in a premature stage; while in M. arnoldiana the embryos were fully developed and filled the whole 98 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII nucellar area. The sac in this particular ovule is divided into two por- tions indicated by a dotted line. The side which contains the embryo is filled with loose endosperm cells. The other side has only a narrow strip of endosperm shown in the drawing with broken lines. Examining the serial sections, at the upper corner of this portion there was found a very young embryo. As a whole the embryos at this late period of develop- ment did not vary much from the one shown in this figure. An enlarged drawing of this embryo is given in Fig. 37. The cells in the suspensor are drawn loosely since their actual limits could not be made out clearly. The round tip represents the beginning of development of the true embryo. A three months old embryo, from a narrow view of cotyledons, is given in Fig. 38 drawn to the same scale as Fig. 34. Between the cotyledons the epicotyle is beginning to be formed. In this drawing the testa which was removed from the seed for better penetration of fixative is not shown. At this period the embryo is 4 to % way down in the sac and has not yet replaced the endosperm as well as the nucellus. DISCUSSION The title of this paper is meant to suggest that embryogeny in M. hupe- hensis occurs by means of development of unfertilized egg cells and further, that embryo sacs in the ovules develop from other cells than megaspores or megaspore mother cells (egg mother cells). The terms apospory and parthenogenesis, as here applied are meant in the following sense: Apospory signifies the development of an embryo sac from a cell other than a megaspore or megaspore mother cell; Parthenogenesis signi- fies the development of an embryo from an egg cell without fertiliza- tion. The two processes are included in the term Apomixis which sig- nifies the absence of sexual fusion in the reproductive process. Almost without exception in the ovules of M/. hupehensis there is one typical meiotic cell that can be recognized by its synezetic chromosomal appearance in early stages. Rarely a meiotic cell is not recognized in this or in an advanced stage. Both in the ovules and in the anthers there is an attempt at reduction in meiotic cells. In the anther the great majority of cells reach the tetrad stage with a wall around each of the resulting cells. The component microspores of the tetrads, however, do not break away and their development ceases at this stage and eventually the tetrads degenerate. Some cells do not even go through reduction but stop at metaphase. A comparable situation exists in the ovules where, if meiotic division is at all successful, one of the resulting megaspores may tend to grow somewhat, while others have ceased to grow or have already disintegrated. However, the author’s impression 1936] DERMEN, PARTHENOGENESIS IN MALUS 99 is that probably never.in this species is a megaspore developed into a functional ES. All the evidence available tends to show that the embryo sacs are of aposporic origin. As was described in the text, in the ovule, in the immediate vicinity of a meiotic cell, there is usually more than one somatic cell with a large nucleolated nucleus which in almost all ovules develops into a typical eight-nucleate embryo sac which sooner or later forms the egg appa- ratus. A few ovules were found which were devoid of embryo sacs. The ovules themselves continued to grow. In most cases the embryo sacs were complex in appearance and were found to have resulted from the fusion of two or more aposporic embryo sacs. In such embryo sacs more than one egg apparatus is formed, from which the appropriate number of embryos and appropriate number of endosperm layers could be formed. No indication whatever was found of development of embryo sacs and embryos from any other tissue than described in the text. The exact origin of these parts could be fully demonstrated since a very large number of ovules at all significant stages was examined. Furthermore, these stages were carefully checked by using as control a diploid form of apple in parallel observations and there were found no exceptions to the rule indicated above. Kobel (1931) refers to “diploid parthenogenesis” in an apple variety, ‘Transparent von Croncels.” No reference was found in the literature on apples giving any account of such a phenomenon. Liljefors (1934) has shown apospory to be present in two tetraploid species of Sorbus. Rosenberg (1908) describes a similar phenomenon in Hieracium. He finds that the greater number of the embryo sacs in the ovules in Hiera- cium are of aposporic origin, and that “Simultaneously with the develop- ment of this vegetative cell into an embryo sac, the embryo sac mother cell is, as usual, divided into tetrads with the reduced number of chromo- somes, and a normal embryo sac commences to develop. This however, most usually, is sooner or later destroyed by the encroaching aposporic embryo sac. Still in some cases two embryo sacs could be fully devel- oped: one probably typical and the other aposporic. In the same head some ovules have a typical and others an aposporic embryo sac.” As in Hieracium, in M. hupehensis there is a typical meiotic cell present, but apart from that only aposporic cells develop into functional embyro sacs. Rosenberg has made no mention of this point. To show definitely whether reduced cells play any role in embryo development, a large number of seeds was planted from which a very small number of seedlings was obtained. Of a total of 39 seedlings,.two were tetraploids, one had some roots showing the triploid number and 100 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII some hexaploid (this plant was destroyed after transplanting; if grown it might have developed into a chimeral plant since such an effect was already noticeable in the seedling), and 36 were triploids, one of which showed a more vigorous growth than others. Tetraploid as well as a few other sister triploid seedlings are now saved for future study. To be positive whether any reductional megaspore actually entered into the formation of a mature ES, the chromosome number in the ES nuclei or anywhere else during embryo development should be studied at the time when some nuclei or cells may be found dividing. Such an exami- nation was not made since in this respect the fixation of my material was not as successful as could be desired. However, based on the observed facts recorded from studying numerous ovules at early and later stages of ES development and the fact that no aneuploid chromosome number could be found in the seedlings examined, we may be justified in con- cluding that although there is usually a meiotic cell present in the ovules, the embryo sacs, however, develop only from somatic cells in the immediate vicinity of the meiotic cell, at which time meiotic cells degenerate. When the rate of embryo development in M. arnoldiana is compared with that of M. Aupehensis, as was pointed out in the text, the embryo development in the latter was very slow without a single excep- tion. Very few seeds at maturity in M. hupehensis were at all fully filled; for the most part they were empty or contained diminutive embryos indicating their somatic origin. It is a noteworthy phenomenon that in M. hupehensis, although an egg cell with a somatic number of chromosomes can develop into an embryo parthenogenetically to insure the stabilization as well as the propagation of the species, this same egg cell is at times receptive to fertilization with a haploid male gamete derived from a diploid species. Considering it in more practical terms, this type of hybridization if successful, may be used as a method of generation of new species arti- ficially, considering that this combination occurred normally with no application of artificial methods. It may be assumed that some of the tetraploid species found in nature have originated in this fashion. Triploidy in itself is an indication that M. hupehensis is a derivative either from a similar species but of diploid form, or from a hybrid be- tween two species and later, through one haploid gamete being fused with a diploid gamete, or from a cross between a tetraploid and diploid species. In any case, the polyploidy in this plant seems to have brought about a genetic unbalance by virtue of which normal sexuality was sup- pressed but not enough to suppress the inherent property of the ovule to produce a gametophyte, although of sporophytic origin, thus insuring 1936] DERMEN, PARTHENOGENESIS IN MALUS 101 alternative process of generation fundamental in almost all the plant kingdom. A genetic unbalance due to triploidy seems to have brought about a vegetative tendency of which parthenogenesis is the result, which in itself may be considered as being a vegetative process. In extreme cases the same tendency is expressed in the differentiation of nucellar cells into conductive (xylem) elements (Fig. 15) which are characteris- tic only of vegetative tissues while nucellar cells serve as a food source for a gametophytic growth. The same factor seems to be at play in the disintegration of sexual cells in the ovules where these cells begin to degenerate at times even before they have begun to divide. The destruction of these cells in M. hupehensis is not a physical process of encroaching of an aposporic ES over a normal one, as Rosenberg reports for Hieracium, as if there were a competitive growth between these embryo sacs, but a process likely physiological, since egg mother cells begin to degenerate before an apo- sporic ES has commenced to develop. Usually simple hybridization between any two species is not sufficient to lead to parthenogenesis or apomixis, in general, and such a method of propagation seems to be commonly linked with genetic unbalance through polyploidy. The data presented by Gustafsson (1935) in Table 8 seem to substantiate the validity of such an assumption. He presents the data of ten genera all of which in diploid form, with the exception of Potentilla, are sexual. Apomixis appears to be prevalent whenever the genom is in odd numbers, 3x, 5x,. . ., 15x the basic number of chromo- somes. On the other hand in polyploids with an even number there are both tendencies, apomictic as well as sexual, present. The prevalence of apomixis in polyploids with odd chromosome num- bers may be usually due to a cytogenetic unbalance brought about by random and unequal distribution of chromosomes so that in some species such megaspores may not develop into functional embryo sacs. The triploid apple species studied in this paper differs from ordinary triploids in the fact that sexual cells may degenerate even before a first division is accomplished and not as a result of random and unequal distribution of chromosomes. In this laboratory we analyzed the cytological situa- tion in the Baldwin apple, a triploid form, and found an approximately normal distribution of chromosomes from germinating seeds (see page 106). Here neither the polyploidy nor unequal distribution of chromo- somes had any apparent effect on the normal sexual processes. If specific genes were responsible for the apomictic phenomenon, as Gustafsson seems to argue, then the same phenomenon should have been 102 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII equally common in diploid hybrids, which is not the case judging from the table he presents. The present author’s contention is that the parthenogenetic process is fundamentally a vegetative act on the part of the egg cell growing into an embryo without uniting or being stimulated by a sperm. Asa whole, in plants, parthenogenesis seems to be linked with polyploidy which accelerates vegetative growth and decreases normal sexual development. Crane and Lawrence (1930) have remarked that in apple “‘triploids are invariably vigorous.” Ina hybrid of Mahonia Berberis, two closely related genera in Berberidaceae, there is a complete suppression not only of sexuality but even of formation of sexual parts such as flowers. The whole plant is completely vegetative and vigorous in growth. This hybrid has the same number of chromosomes as its parents with 2n = 28 (Dermen 1931). Obviously here polyploidy is not the cause of this peculiarity but it may be due to extreme vigor brought about by com- bining the two species of closely related genera. We have made attempts to make some artificial hybridization between the above two genera; if successful, the study of the intergeneric plants as well as the full analysis of tetraploid forms derived in M. hupehensis mentioned above should throw a good deal of light on the attempts to determine some of the causes of parthenogenesis. Taxonomically M. hupehensis must be considered as a very stable species since very rarely will it fail to breed true from seeds. Most diploid species are more or less cross fertile with other species within the genus and when grown together they will fail to breed true from seeds; hence their propagation must be insured by grafting or by other vegetative methods. Malus hupehensis in this respect cannot be con- sidered as a typical species since it differs from the normal in that it may safely be propagated from seeds although grown side by side with other species. As we have pointed out the seeds here develop parthenogeneti- cally and very seldom through fertilization, in which case we may obtain entirely new cytological forms, as was described in the text. It is safe to assume that M. hupehensis is of hybrid origin and is consequently an allopolyploid form cytologically, since we assume that mere autopoly- ploidy through fusion of a haploid and a diploid gamete in a pure species will not bring about a genetic unbalance which would eventually have resulted in parthenogenesis. Some species in Pomoideae that are found to breed true under similar conditions, from seeds, may be suspected of similar phenomena as those described for M. hupehensis. This analogy may be extended to other species of similar behavior in other families. 1936] DERMEN, PARTHENOGENESIS IN MALUS’ 103 SUMMARY Malus hupehensis (Pamp.) Rehd. is a triploid species with 3n = 51 chromosomes. Sexuality is completely suppressed through degeneration of sexual cells in both anthers and ovules; however, the propagation of the species is carried on by parthenogenetic method. Evidence of hybrid origin of this species is indicated. In the ovules embryo sacs are probably entirely formed by apospory in which normal egg apparati are formed. It is the egg cells alone that develop into embryos, and embryos never are formed from any other source. Polyembryony in M. hupehensis is found to be linked with the development of complex embryo sacs. Based on the fact that parthenogenesis is commonly associated with polyploidy, a suggestion is made that the expression of such a phenome- non in M. hupehensis may be due to a physiological change brought about by triploidy and that this factor accelerates the vegetative growth in the plant and decreases or entirely suppresses its normal sexual development. LITERATURE CITED Crane, M. B. and W. J. C. eligi ey. Fertility and ee of apples in seri to chromosome num Jour. Genet. 22: 153-163). DERMEN, H. ( ). A study aa number in two genera of Berberidaceae ; Mahonia Be "Bee (Jour. Arnold Arb. 12: 281- 287). —— (1933). Origin and behavior of the nucleolus in plants. (Jour. Arnold Arb. 14: oe 325. GUSTAFSSON, A ete Studies on the mechanism of parthenogenesis. Kose., F. (1931). Lehrbuch des Obstbaus auf physiologischer Grund- lage. (Berlin, ney Sprin Lit_jerors, A. 34). Ueber normale und apospore Embryosackent- wicklung in Ps ea iat Sorbus, nebst einigen Bermerkungen tiber die Chromosomenzahlen. (Svensk Bot. Tidskr. 28: 290-299). Renper, A. (1933). New species, varieties and combinations. (Jour. Arnold Arb. 14: 199-252. DAK, Ke C1931). The origin and relationship of the Pomoideae. (Jour. + -—- i. Chronineanie relationships in the Pomoideae. (Jour. Arnold Arb. 13: 363-367. ) EXPLANATION OF PLATES 179-181 Figs. 1-17 Malus hupehensis 1. Triploid number 3n = 51 chromosomes from root-tip sections. Fig. 2. Tetraploid number 4n = 68 chromosomes from root-tip sections. 3. II M from anther with 25+ chromosomes at each pole. Aceto- carmine preparation. Material prepared between last week of April and first week of May, about ten days before buds turn pink. 104 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Fig. 4. Egg mother cell (EMC) at I M with 51 univalent chromosomes from two sections a and b. Ovaries fixed from buds at pink stage, ee . oe, Fig. 5. EMCatI Ai in two sections. Pink stage, 5/10/32. Fig. 6. EMC pachytene. Pink stage, 5/ Fig. 7. Two EMCs at late pachytene. Pink stage, 5/10/32. Fig. 8. EMCatIM. Pink stage, 5/9/32. Fig. 9. mi as Fig. 8 showing cells in vicinity of the top of the Fig. 10. Outline ‘drawing of an bry showing the EMC in Fig. 9, the en Fig. 11. An EMC at IM pean Pink stage, 5/10/3 s. U meiotic derivative is indicated by a small nuclear mass outside the large nucleus. Petals falling stage, 5/10/32. Fig. 13. Two daughter cells from an EMC division, one degene agin ds and the other at anaphase stage. Some of the c osomes W the plates, showing meiotic origin of an cells. Pole Susi opening stage, 5/10/32. Fig. 14. A somatic cell with a large nucleolus growing into an ES apos- porically, | while the meiotic one above is degenerating. Petal 32. Fig. 15. Two-nucleate aposporic ES and below it a tetrad of cells (of meiotic origin?) with xylem annular thickenings. Open bloom stage, 1Pink stage buds usually a 5th of May in the Arnold Arboretum. Fig. 16. Two aposporic embryo sacs, one two-nucleate and the other four-nucleate. Flowers just bloomed stage, ig Fig. 19. M. hupehensis. A complete ES as in Fig. 18. One polar nucle- olus and one synergid cell, dotted, from an adjacent ie superimposed in a single drawing. Before shedding of petals, Fig. 20. M. arnoldiana. An unfertilized ES. Antipodals have dis- 34. Fig. 21. M.arnoldiana. A fertilized ES. Otherwise same as in Fig. 20. Fig. 22. M. hupehensis. Endosperm development and tubular growth of ES. At shedding of calyx, me Fig. 23. M. hupehensis. Twin ES. Parts drawn at the side from an adjacent section. Antipodals ie disappeared. From same ay ee Fig. 24. M. hupehensis. A single ES. One of the cells of the egg apparatus drawn from an adjacent section. From same material as Fig. Fig. 25. M. na Fe i stage with scant endosperm. Moaed eee 5/3 Fig. 26. M. arnoldiana. pee tudinal section of a whole young seed at the time of proembryo about 12 cells long. his shows ae extent of tubular growth of ES. Material same as in Fig. oLp Ars. VoL. XVII PLATE 179 a” ie a JY 4 vs 7) Jen ~ WRGS> 257 00 e) mes eine Sage 8, ’ a? ie 05 Bb 1 a an ee RQ ApoOSPORIC PARTHENOGENESIS IN MALUS HUPEHENSIS Jour. ArNotp Ars. VoL. XVIT Prate 180 / ee — aa 5 = .~_ re i. % /00¢" Hs /000 26 27 ApOSPORIC PARTHENOGENESIS IN MALUS HUPEHENSIS Jour. ArNotp Ars. VoL. XVII Pate 181 — i) 20" gg ™ 50r APOSPORIC PARTHENOGENESIS IN MALUS HUPEHENSIS : a _ 7 —— oe ; | ; 7 - | oe : — _ 7 . : a : | — . ; eS : 7 7 — OS Oo a TO a > - : : oe _ oe BS Oo >) > Oo a a OO ee - en _ es ee oe 7 _ : : - a 7 a : : OB : : 7 a - 7 a 7 _ —_ - : Ww W ao nN DERMEN, PARTHENOGENESIS IN MALUS 105 . M. hupehensis. Proembryo and beginning of endosperm de- 34. velopment. Material fixed 5/31/34 . arnoldiana. Suspensor and true embryo. Material fixed 6/5/34. . M.~ hupehensis. Two proembryos in one complex ES. Lower one is away from nucellus. Material fixed 6/5/34. . M. arnoldiana. Suspensor and embryo at a later stage with con- 6/11/34. noldiana. ns siderable amount of endosperm. Material foe . M. hupehensis. Still in proembryonic stage. This also chosen to show the development of embryo away from nucellus. Material fixed 6/11/34. . M. arnoldiana. Older embryo with suspensor. Material fixed 6/18/34. : ae ye Sa ensis. Just the beginning of proembryo differentia- uspensor and true embryo. Endosperm still scarce Material fixed 6/18/34. M. arnoldiana. Advanced stage of embryo. The nucellus is not yet er aclcee: replaced by endosperm. The outer two solid lines indicate the limits of testa. Material fixed 6/26/34 . M. hupehensis. Twin ovule marked with a dotted line. ‘In the pensor and embryo differentiation compared with the one in ig. 33. Broken lines indicate the extent of endosperm Sevetop: ment. Material fixed 6/26/34. . M. hupehensis. Twin ES, one with advanced stage of suspensor and embryo. This sac is completely filled with loose endosperm tissue. The other has in another section a ghia just ier ta ning to differentiate into suspensor and true embryo. latter sac has a little growth of endosperm. aed al eee 7/14 ; a hupehensis. Enlarged drawing of the embryo in Fig. 36 to ow the development of embryo independent of nucellus. : M. hupehensis. The stage of embryo development tee months after meiotic stage. The embryo is not yet halfway down in the nucellus. The testa is not shown in this figure. Material fixed 8/7/34. CytToLocicaL LABORATORY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 106 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XVII FERTILIZATION IN THE BALDWIN APPLE, A TRIPLOID VARIETY HaiGc DERMEN THE DATA presented in Table I were obtained by studying sectioned root-tips from germinating seeds of the Baldwin apple. Several hun- dred seeds were placed in a box mixed with well moistened peat and frozen in a refrigerator for more than two months, from 1/9/34 to 3/19/34. They were then planted in soil in a flat. The seeds began to germinate during the latter part of March and one primary root-tip from each individual seed was cut and fixed in Lewitzky’s solution. Four separate collections were made, 3/30/34, 4/2/34, 4/3/34 and 4/18/34. Fourteen seeds were analyzed from the collection of 3/30/34, 41 from that of 4/2/34, 45 from that of 4/3/34 and 36 from that of 4/18/34. Each root-tip was numbered and each group was recorded separately. Studying each group separately, no appreciable difference in the distribution of chromosome numbers was found, indicating that there was no segregation of chromosome numbers due to difference in seeds germinating at different times. Moffett (1931) has analyzed the somatic chromosome numbers of 32 seedlings obtained from four differ- ent 3n & 2n crosses and has found 10 seedlings having 41 chromosomes, other seedlings sharply varying from this number. Based on this observation he remarks that ‘41 chromosome number is the secondary diploid number (thirty-four). It is to be imputed that the high fre- quency of seedlings with 41 chromosomes is due to the greater viability of the 17 + 7 gametes in the triploid rather than to their zygotic via- bility, since seedlings having other chromosome numbers are frequently more vigorous than the 41 type.” Moffett has observed secondary pairing in the meiotic metaphase chromosomes of all the diploid types of the Pomoideae, forming sexiva- lent and quadrivalent groups and in extreme cases seven groups — three sexivalents and four quadrivalents. Because of this and the fact that he found more seedlings with 41 chromosomes than with other numbers, he has concluded that the 34 chromosomes of the diploid Pomoideae are of seven types, of which four are presented four times and three are pre- sented six times, and that the haploid number 17 is a secondary basic number of-7, the ancestral basic number. A critique of this point of view is presented by Sax (1933). 1936] DERMEN, FERTILIZATION IN THE BALDWIN APPLE 107 At Dr. Karl Sax’s suggestion the present writer examined the somatic number of chromosomes from 136 germinating seeds. Of these seeds 130 showed numbers ranging from 35 to 51 (Table I) and six seeds had the following numbers: 1 = 53, 1 = 57, 1 = 58, 1 = 59, 1 = 62, and 1 = 65 chromosomes (these are not included in the table). Here we notice that the class with 40 chromosomes is higher than the class with 41 chromosomes. The distribution here recorded is quite random and does not suggest that a gamete with 17 + 7 chromosomal combina- tion has any advantage over the other combinations. Therefore, the fact that the class 41 singly is not larger than other classes, but conforms quite nearly to the expectation on the basis of normal distribution, invalidates Moffett’s theory that a 17 + 7 chromosomal combination is more viable than others. The only peculiarity observed in this table is that the class with 39 chromosomes has only 9 seeds, fewer than ex- pected, in which case perhaps for some reason some eggs with 22 chromo- somes were either too few or failed to be fertilized. TABLE I CHROMOSOME DISTRIBUTION IN BALDWIN APPLE SEEDS Chromosome | | | number in 34|35|36/ 37/38/39 40| 4142/43/44 als ed 49|50|51 pce ela tale & laatnll ae israule la balers ok of seeds : | | Statistica: datas Mean =- 41.36 = 0.173: o-= 293° = €122° CYV..= 7.09 + 0.296. Pollen germination in the Baldwin apple is about 11-12% (Crane and Lawrence, 1930). By making breeding tests Sax (1922) has shown that the Baldwin is self sterile in Maine. The data presented here are from seeds of fruit set by open pollination. If the seeds had resulted from self pollination or cross pollination with another triploid form, the class center should have been around 51 and not, as in the table, far below this and with only five seeds above it. This strongly indicates that the Baldwin apple is self sterile in Massachusetts as well as in Maine and that it is easily cross fertile with diploid varieties. In this case, over 95% of the seeds must have resulted from pollination with haploid gametes from diploid varieties. The six seeds probably were the result of fertilization of eggs that contained more than 34 chromo- somes which may have been due to restricted reduction division when at the first metaphase chromosomes fail to reduce and the great majority 108 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII of chromosomes go into resting stage again, while a few form small microcytes. Such cases were observed in Rhoeo discolor when treated with cold or heat. The fact that there were only three seeds with 51 chromosomes and none immediately below makes remote the proba- bility of the six seeds with higher numbers having resulted from pollina- tion with another triploid variety. On the basis of 3n > 3n cross with the union of gametes with 25 + chromosomes, seeds with chromosomes ranging around 51 as a central point should have been numerous which obviously is not the case. LITERATURE CITED Crane, M. B. and W. J. C. Lawrence (1930). Fertility and vigor of apples in relation to chromosome number. (Jour. Genet. 22: 153-163.) Morrett, A. A. (1931). The chromosome constitution of the Pomoideae. (Proc. Roy. Soc., B. 108: 423-446. Sax, K. (1922). Sterility relationship in Maine apple varieties. (Annual report of the Maine Agric. Ex — (1933). The origin of the Pomoideae. (Proc. Amer. Soc. Hort. Sci. 30: 147-150.) CyTOLOGICAL LABORATORY, ARNOLD ARBORETUM, Harvarp UNIVERSITY. 1936] FAULL, TWO SPRUCE-INFECTING RUSTS 109 TWO SPRUCE-INFECTING RUSTS CHRYSOMYXA PIPERIANA AND CHRYSOMYXA CHIOGENIS ). A. Fars Chrysomyxa Piperiana (Arthur) Sacc. & Trott. is a common rust on the Pacific slope of the northwestern United States, and within that region has been collected in Washington, Oregon and northern Cali- fornia. Up to the present it has been recognized on Rhododendron macrophyllum Don (Rh. californicum Hook.) only, and on that host in its uredial stage only. The earlier collections were usually labelled Chrysomyxa Rhododendri DeBary, but evenually Arthur diagnosed it aS a new species and described it from its uredia under the name Melampsoropsis Piperiana. Recent successful infection experiments, however, conducted in the greenhouses of the Arnold Arboretum, demon- strate that Peridermium Parksianum Faull on Picea sitchensis (Bong.) Carr. is the haploid phase of this rust. Moreover, abundant telia, telio- spores and basidiospores have been discovered in material collected in Oregon by Dr. G. D. Darker. So it is now possible to present a complete taxonomic description of Chrysomyxa Piperiana. The cultures referred to were made by sowing the aeciospores of Peridermium Parksianum on two small potted plants of Rhododendron macrophyllum ; two similar ones served as controls. All four were grown from seed of British Columbia origin, sown February 8, 1928 by Mr. William H. Judd at the Arnold Arboretum. The greenhouses were entirely free from Rhododendron rust of any kind. The aecial material used as inoculum was collected by Mr. H. E. Parks at Trinidad, Humboldt County, California, November 23, 1933 and sent in fresh condition by air mail. The aeciospores were sown five days later. A relatively small percentage of the spores germinated, but with the liberal amount employed the number was amply sufficient. Within a month to six weeks many small yellow lesions developed on the inoculated plants. Thereafter until early April the lesions, some of them slightly swollen, seemed to remain in a somewhat stromatic resting condition. Then hypophyllous uredia, usually about one to a lesion, began to form on several of the lesa The first discharge of uredio- spores was noted April 27, 1934. Careful studies of both uredia and urediospores were promptly made. Fortunately, herbarium material (a single rusted leaf) was harvested prior to serious fumigation injury 110 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII in June. The control plants remained throughout without any sign of infection. The discharged urediospores were yellow, ellipsoidal or linear, often attenuate at one or both ends, straight or slightly curved. Their walls were hyaline, closely tuberculate and 1.5-3.0 uw thick. Forty spores from the same sorus measured 16.0—27.4 & 27.4-68.5 uw and averaged 22 * 46 uw. Razor sections of the uredia showed that the spores were borne catenulately and separated by intercalary cells. They also re- vealed the presence of a well-marked peridium. The peridium consists of a single layer of laterally adherent cells, the terminal cells of the urediosporogenic, primordial columns. The finding of a peridium clashes with current descriptions of the uredium of C. Piperiana. But sections of type material or other field-collected specimens show exactly the same kind of peridium as found in my cultures. Indeed, the rust as obtained in cultures answers in all respects to C. Piperiana. With regard to the telia, lack of acquaintance with them is probably due to the fact that they are of brief duration, their occurrence coincid- ing with a short period of susceptibility to infection of young spruce needles; and so they have been missed by collectors. They surely can- not be rare. Once familiar with their appearance, remains of them can be detected in many existing collections. It is likewise probable that this rust is not infrequent on spruces within its range. Recognition is easy because of the striking microscopic characters of both spermogonia and aeciospores. Chrysomyxa Piperiana (Arthur) Sacc. & Trott. in Syll. Fung. 21: 716 (1912). (II.) O, I, II and III. Melampsoropsis Piperiana Arthur in North Am. Flora, 7: 120 (1907). II Caeoma Piperianum (Arthur) Sacc. & Trott. in Syll. Fung. 21: 787 (1912). IL. Peridermium Parksianum Faull in Jour. Arnold Arb. 15: 86 (1934). O, I. O. Spermogonia on needles of current season, hypophyllous, numer- ous, biseriate, conspicuous, yellowish, then reddish-brown, slightly ele- vated, aparaphysate, immersed, subtriangular in vertical section, subepidermal, 228-315 u broad and 125-155 uw deep, averaging 260 « 140 wt; spermatiophores unbranched; spermatia globose to broadly ellipsoidal, 2.7-3.2 3.0—-5.0 u. I. Aecia (peridermia) on needles of current season, yellow, hypo- phyllous, irregularly biseriate, on yellowish discolored portions of affected needles, circular or somewhat laterally compressed in transverse section, cylindrical, 0.4-1.2 mm. in diameter and 1.0-2.2 mm. high; 1936 ] FAULL, TWO SPRUCE-INFECTING RUSTS 111 peridium colorless, rupturing at the apex; peridial cells polygonal, elongate vertically, subimbricate, in a single layer, 18-26 48-88 u, with outer walls smooth, 0.7—1.0 u thick, and inner walls finely verru- cose, about 1.5 uw thick; aeciospores yellow, fusiform or narrowly ellip- soidal, straight or curved, rarely subglobose, 13-24 25-111 , usually fusiform and averaging about 16 70 u; walls of aeciospores densely and finely warted, hyaline and about 1 wu thick. II. Uredia hypophyllous, pustular, waxy before rupture, subepi- dermal, 0.2—0.4 mm. in diameter, on solitary or aggregate areolae that constitute the pale green to yellowish brown lesions, one each to an areola; peridium distinct, 10-14 u thick; peridial cells in a single layer, angular, thin-walled, 13-19 u in diameter; urediospores yellow, catenu- late, with intercalary cells, pulverulent as discharged, or extruded in a cylindrical column (resembling the telial column of a Cronartium) from the apically ruptured uredium, ellipsoidal to linear, often narrowed towards each end, straight or slightly curved, 14-27 & 27-100 uw, aver- aging 16-21 41-67 4; walls of spores hyaline, 1.5-3.0 u thick, coarsely and closely verrucose. III. Telia on overwintered leaves, hypophyllous, yellow, pulvinate, waxy, rupturing widely at maturity and eventually becoming concave, subepidermal, one or two to an areola on solitary or aggregate areolae that constitute the pale green to yellowish brown lesions, subcircular, 0.2-0.5 mm. in diameter; no peridium; teliospores yellow, catenulate, 3 to 6 or more spores in a chain, without intercalary cells, smooth, thin- walled, 11-16 13-22 u. The teliospores germinate in situ at maturity and in centripetal order. Basidia slightly curved to strongly arched, 4-celled, 7-8 u thick and up to 40 u long. Basidiospores subglobose, 8-11 uw in diameter. The foregoing description of telia is based on two collections of over- wintered leaves — (a) Still Creek, Clackamas Co., Ore., elev. 3500 ft., July 7, 1929, G. D. Darker (2682), Arnold Arb. Path. Herb. no. 179; (b) Rhododendron, Clackamas Co., Ore., elev. 2000 ft., July 9, 1929, G. D. Darker (2697), Arnold Arb. Path. Herb. no. 194. The sequence of spore production on overwintered Rhododendron leaves seems to be — uredia in spring, telia during period of suscepti- bility to infection of new spruce needles, then a copious crop of uredia. Chrysomyxa Chiogenis Dietel is a rust long known on Chiogenes his- pidula (L.) Torrey & Gray, both as to its uredia and telia. Though it has been reported from Wisconsin, Newfoundland and northern New York only, it is much more widely distributed. Thus, herbaria in Har- vard University contain, in addition, collections from Ontario, Quebec, 112 JOURNAL OF THE ARNOLD ARBORETUM TABLE 1 [VOL. XVII Chrysomyxa Chiogenis from Chiogenes hispidula to Picea glauca ( Moench) Voss. Date of appearance appearance of Date bee pi ee = inoculation of lesions peridermia harvested infected peridermia 23.v1.24 15 days 29 days 5.vili.24 53 47 24.vi.24 in x) a 5.viti.24 20 19 24.vi.24 — 30. 5.viil.24 1 1 17.vi.25 — a. * 5.viil.25 — 5 17.vi.25 —— a 5.viii.25 — 25 17.vi.25 — ras L.viii.25 — 38 16.vi.28 12 days — 31.vis. 28 14 Z 16.vi.28 - — 31.vii. 28 92 42 TABLE 2 Chrysomyxa Chiogenis from Chiogenes hispidula to Picea mariana (Mil 1.) B.S.P. First appearance of Date of appearance Date needles needles with inoculation of lesions peridermia harvested infected peridermia 25.vi.24 ae 10.ix. 24 2 2 17.vi.25 — 35 days 5.viii.25 20 12 17.vi.25 — ic 17.viii.25 10 + TABLE 3 Chrysomyxa Chiogenis from Picea glauca to Chiogenes hispidula Experiment 1 Source of inoculum Experiment 1 of Table 1 August 8, 1924 September, 1924 Late May, 1925 Date of inoculation First indication of lesions First appearance of sori Date harvested June 5, 1925 No. of inoculated branches 2 Leaves on inoculated branches 113 Leaves infected and with sori 19 Kinds of sori Uredia and young telia (7) Experiment 2 Experiments 4 and 5 of Table 1 August 5, 1925 September, 1925 About June 1, 1926 June 20, 1926 5 Uredia and telia 1936] FAULL, TWO SPRUCE-INFECTING RUSTS 113 New Brunswick and New Hampshire. Doubtless its range is much greater still. In some localities it is common; but even in such places, because of its inconspicuousness, it easily escapes detection unless the prostrate branches of its host are turned over for inspection. So far no host other than the creeping snowberry has been recognized for this rust, but successful infection experiments summarized below demonstrate that at least two species of Picea can serve as hosts for the haploid phase. These experiments were conducted at the field labora- tory of forest pathology on Bear Island in the Timagami Forest Reserve, Ontario; and in connection with them I gratefully acknowledge the assistance afforded by Mr. W. R. Watson, Professor E. H. Moss and Dr. G. D. Darker. The minbess of uredia on the infected leaves of experiment 1 were, respectively — 18, 16, 35, 24, 17, 45, 20, 26, 51, 38, 34, 75, 40, 56, 36, 24, 32, 28, 19. Experiment 1 was conducted on a potted plant at the laboratory; Experiment 2, on an undisturbed plant in a carefully selected spot in the field in which no rust could be found and which was well protected from chance infection. Uninoculated branches of the experimental plants served as controls, and in addition surrounding plants in experiment 2. All controls remained free from rust except that in experiment 1 seven uredia were found on a 4-leaved twig on a branch immediately below the inoculated branches. Chrysomyxa Chiogenis Dietel in Bot. Gaz. 19: 303 (1894). (II.) I, II and III. Melampsoropsis Chiogents psy Arthur in Rés. Sci. Congr. Bot. Vienne, p. 338 (1906). O. Spermogonia unknown; Sere lacking. I. Aecia (peridermia) on needles of current season, yellow, hypo- phyllous, biseriate, on yellowish discolored portions of affected needles, lenticular to subcircular in transverse section, 0.5-1.2 mm. long and 0.5-1.5 mm. high; peridium colorless, rupturing at the apex; peridial cells polygonal, elongate vertically, not imbricate, in a single layer, 16-32 30-45 uw, with outer walls smooth, 0.5-1.0 u thick, and inner walls verrucose, the warts often in short lines, 2.5-3.0 thick; aecio- spores yellow, ellipsoidal to subglobose, 15-18 19-30 u, averaging about 16 X 24 uw; walls of aeciospores closely and rather finely to sub- coarsely warted, hyaline and 2-3 u thick. II. Uredia hypophyllous, pustular, subepidermal, circular to ellip- tical, 0.2-0.7 mm. in major axis; peridium distinct, 13-16 uw thick; peridial cells in a single layer, angular, thin-walled, 10-13 in diameter ; 114 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XVII urediospores yellow, catenulate, with intercalary cells, pulverulent, ellip- soidal, rarely subglobose, 12-19 & 18-32 u, averaging about 16 22 w; walls of spores hyaline, closely and rather finely verrucose, 2.0-2.5 wu thick. III. ‘Telia on overwintered leaves, hypophyllous, yellow, flat, waxy, rupturing widely at maturity, subepidermal, subcircular to elliptical, 0.30.7 mm. in major axis; no peridium; teliospores yellow, catenulate, up to 6 or more in a chain, without intercalary cells, smooth, thin-walled, 11-15 & 12-19 wu. The teliospores germinate in situ at maturity. Basidia slightly curved to strongly arched, 4-celled, 7-8 uw thick and up to 43 uw long. Basidiospores subglobose, 7—9 u in diameter. The description of aecia given above is from culture material in Herbarium J. H. Faull: (a) on Picea glauca (Moench) Voss (P. cana- densis [Mill.] B.S.P.), nos. 7305, 7304 and 7302; (b) on Picea mariana (Mill.) B.S.P., nos. 7307 and 7308. LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. JOURNAL OF THE ARNOLD ARBORETUM VoLuME XVII. JULY, 1936 NUMBER 3 MORPHOLOGY AND ONTOGENY OF THE SPERMOGONIA OF THE MELAMPSORACEAE Litt1An M. HUNTER With plates 182-188 THE MORPHOLOGY and the ontogeny of the spermogonia of rust fungi have thus far received relatively scant attention. Contributions of greater or less significance have been made by Unger (73), Meyer (59), Tulasne (72), DeBary (29, 30), Sappin-Trouffy (66), Blackman (19), Rosen (65), Neuman (60), Maire (55), Allen (3, 5, 6, 8), Ludwig (54), Hunter (45), Jackson (47, 48) and others; but they have been so inci- dental, fragmentary or introductory in nature that the subject still re- mains largely unexplored. It would seem that such topics must be of importance with respect to spermogonia just as they are of other repro- ductive organs, and especially to students of systematic uredinology. Motivated by this viewpoint I have studied the spermogonia of a wide range of species. The paper here presented records part of this work. It is restricted to the Melampsoraceae, and particularly to representatives of the various genera of the Melampsoraceae that occur in temperate North America and Europe. The material of all Melampsora species discussed in this paper was obtained from culture experiments made by W. P. Fraser, J. H. Faull, H. P. Bell, E. H. Moss, G. D. Darker, E. H. Bensley, W. R. Watson and the writer, and from supplementary field material, except that no natu- rally occurring specimens of Melampsora Larici-Capraearum were avail- able. All of the cultured material, exclusive of that obtained from Professor Fraser, was fixed when in fresh condition. Treatment of the material was similar to that described by the writer (45), with the exception that fresh material of Uredinopsis, Milesia and Melampsora species was sectioned by hand and studied in an aqueous solution and after mount- ing in lacto-phenol. A combination differential staining method, using 116 JOURNAL OF THE ARNOLD ARBORETUM (VoL. Xv safranin and light green in lacto-phenol, was introduced for material of Milesia polypodophila and Coleosporium species after embedding in celloidin. This method was devised without knowledge of Lepik’s (52) method of double staining in lacto-phenol. The ethyl-butyl alcohol series was preferable to the ethyl alcohol series used previous to em- bedding. The data obtained on each of the species dealt with here are pre- sented in detail with reference to (1) morphology of mature spermo- gonia, (2) their ontogeny. This procedure has been adopted because of the fact that detailed descriptions of these organs by other authors have been meager at best, and most frequently are non-existent. The species selected belong to the following genera and subfamilies: — Subfamily MELAMPSOREAE — Melampsora II Subfamily PucctnrastREAE — Melampsoridium, Melampso- rella, Pucciniastrum, Thecopsora, Calyptospora, Hyalopsora, Milesia and Uredinopsts. III Subfamily CRoNARTIEAE — Cronartium IV Subfamily CHRYSOMYXEAE — Chrysomyxa V_ Subfamily CoLEospor1EAE — Coleosporium From this it will be seen that the only genera of the Melampsoraceae as recognized by Dietel under the family Melampsoraceae (Engler and Prantl’s Natiirliche Pflanzenfamilien, Band 6, 2nd ed. pp. 35—48, 1928) not included here are Chnoopsora and five genera the haploid stages of which are unknown — Mesopsora, Phakopsora, Crossopsora, Mikroni- geria and Bubakia. I. Subfamily MELAMPSOREAE The Melampsoreae comprise three genera, namely, Melampsora, Chnoopsora and Bubakia. Three species of Chnoopsora are known; they occur in India and Central Africa; their spermogonia are said to be subepidermal; no material was available for study. 1. Melampsora Abieti-Capraearum Tubeuf Arthur (16) considers Melampsora Abieti-Capraearum Tubeuf on Abies alba Mill. in Europe and M. americana Arthur on A. balsamea (L.) Mill. in America to be the same species. My studies on their spermogonia afford support to this conclusion. The morphology of the spermogonia of the American rust has al- ready been described by Hunter (45). Subsequent examination of additional specimens reveals no essentially divergent data. Quite re- cently my studies have been extended to the spermogonia of the Euro- pean rust. The similarities of the American and the European materials 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 117 are so close that a summarized presentation of the data for both, as found in Table IT below, will suffice. It should be noted, however, that in both of them certainty as to the location of mature spermogonia with reference to the epidermis is often impossible. Some are plainly sub- epidermal; others appear to be subcuticular. Here then is a case in which the correct description of the spermogonium requires a knowledge of its ontogeny. To clear up this point I have made a thorough study of the ontogeny of the spermogonia of this rust on A. balsamea, both from culture and field materials. The spermogonial primordia in all cases examined lay immediately under and in contact with the epidermis. Usually it forms in an air space (commonly a substomatal cavity), but sometimes between the epidermis and the otherwise contiguous palisade tissue. The _pri- mordium consists essentially of closely adhering, vertical, actively grow- ing hyphae, interpreted as spermatiophores, with a suggestion at times of a thin, basal, hyphal mat. Figure 1 (drawn from a lateral, vertical section of a mature spermogonium) adequately illustrates a typical primordium and also initial stages of disruption of overlying epidermal cells. Subsequent developmental stages comprise a thickening of the basal mat (Fig. 2) accompanied by some destruction of underlying mesophyll, and a maturing of the spermatiophores accompanied by a variable amount of destruction of overlying epidermal cells or none at all. These different effects on the epidermis, as revealed by serial sec- tions, may occur within the limits of a single spermogonium. Destructive action on the mesophyll is indicated in Fig. 3 by a break in the continuity of three layers of its tissues, and by the dark-staining remains of palisade cells embedded in the spermogonium. Among the epidermal cells the guard cells are the most resistant (Fig. 2). They may be invaded; but even so, their thick walls persist and the cells retain their contour. The effect on other epidermal cells is shown in Fig. 1, where it is seen that two hyphae are pushing inwards the lower wall of one cell, and others have already invaded the cell to the right the protoplasm of which is now dead. Almost all conditions of the epi- dermis overlying a mature spermogonium, as described above, are repre- sented in Fig. 1 of Hunter (45). The conclusion is that the spermogonium of M. A bieti-Capracarum is properly described as subepidermal, and not subcuticular as asserted by Ludwig (54). It consistently originates under the epidermis; and in the course of its development there is more or less destructive action of underlying tissues, and often of the epidermis even out close to the cuticle. 118 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII As an addendum it is of interest to note that a few sections showed long hyphae extending beyond the surface of the spermogonium. They were found centrally located in spermogonia that had just matured, never in old spermogonia. Spermatia were massed around them and, unlike paraphyses, they were ephemeral. These I interpret as the “‘flexu- ous hyphae” of Craigie (27). 2. Caeoma Faulliana, n. sp. ©. 4; Pycnidia amphigena, inconspicua, flavida, pustulata, non immersa, hemisphaerica vel conoidea, subcuticularia, 42-145 wu lata et 27-49 u alta, plus minusve 80 & 37 u; pycnidiophora simplicia, non septata vel juxta bases septata, ad apicem dehiscentia, poris 12-15 x 23-30 wu. Aecidia hypophylla, maculis flavidis insidentia, secus series duas ir- regulares disposita, aecidiosporis globosis vel subglobosis, subtiliter verruculosis, 16-21 u diam. The material studied was collected by Mr. W. R. Watson on the needles of the current season of Abies lasiocarpa (Hook.) Nutt. at Banff, Alberta, July 8, 1925. It consisted of pieces fixed in Carnoy’s fluid for study of the spermogonia and dried herbarium specimens (Herb. J. H. Faull, no. 8849). The spermogonia (Fig. 4) at once dis- tinguish this rust from M. Abieti-Capraearum which it superficially resembles. They are smaller, more elevated (pustular) and plainly subcuticular, No doubt it is a species of Melampsora ; but until someone makes appropriate cultures the identity of its diploid phase will remain unknown. So it is tentatively referred to as a new species under the name given above. 3. Melampsora Abietis-canadensis (Farl.) Ludwig The spermogonia of Melampsora Abietis-canadensis are borne on leaves of the current season, young cones and young twigs of Tsuga canadensis (L.) Carr. The affected leaves, especially those of the terminal buds, may be distorted and Adams (1) states that the same is true for affected twigs. The spermogonia are abundant, small, sub- cuticular, amphigenous, pustulate, and hemispherical-flattened to co- noidal in vertical section (Fig. 6). The base of the spermogonium is a flat-bottomed stroma. The epidermal cells at the base of the spermo- gonium are separated more or less by mycelium but they are not greatly displaced; they are eventually killed and may be disrupted, though fre- quently their walls remain unbroken, Thirty-four spermogonia occur- ring on needles measured from 35—98 wt broad and 15-38 u high, averag- ing 66 X 25 u. On the cones seventy-seven spermogonia measured 53— 128 uw broad and 13-30 u high, averaging 78 & 23 u. The spermatia are discharged through a long slit, 8-30 u wide. The foregoing description 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 119 is based upon culture material supplied by Professor W. P. Fraser. Considerable confusion has existed in the past with respect to the identity of M. A bietis-canadensis on Tsuga (4, 37, 38, 39, 16, 68, 69), but the conclusion as expressed by Ludwig (54) and Arthur (16) is ac- cepted. It is pertinent, however, to call particular attention to a rust (Caeoma dubium Ludwig, 54) on Tsuga heterophylla (Raf.) Sarg. from western North America. The spermogonia of this rust are distinguished from those of M. A bietis-canadensis as well as from other known species of Melampsora. The spermogonia of C. dubium are subepidermal and measure 80-115 uw broad and 50-90 : high. This affords another ex- ample of the diagnostic value of spermogonia. 4. Melampsora Bigelowii Thim. The spermogonia of Melampsora Bigelowii are found on needles of Larix laricina (Du Roi) K. Koch and L. leptolepis Murr. They are small, abundant, usually discrete but sometimes confluent, pustulate, amphig- enous, subcuticular, and hemispherical to conoidal in vertical sections (Fig. 7). The epidermal cells below the spermogonium, except for slight separation by the mycelium, are usually left im situ; but occasionally they are widely separated and sometimes so completely obliterated that the spermogonium occupies the space where the epidermal cells were located, as well as extending above them. Forty-two spermogonia meas- ured 50-102 u broad and 15-38 uw high, averaging 69 & 27 u. The spermatiophores may be nonseptate or one-septate towards the base. The spermatia measure 1.5-2.0 & 3-4 u. The opening through which the spermatia are exuded is 15-30 uw wide and extends over the greater part of the spermogonium in the longitudinal direction of the leaf. Both spermogonia and caeomata were obtained on Larix leptoleptis and L. laricina from inoculation experiments in which the teliospores of North American species of Salix were employed. The material on L. laricina only was sectioned. 5. Melampsora Medusae Thiim. The spermogonia of Melampsora Medusae occur on needles of Larix laricina and L. leptolepis. They are indistinguishable from those of M. Bigelowti. Both spermogonia and caeomata were obtained on Larix leptolepis and L. laricina from inoculation experiments in which telio- spores from North American species of Populus (including P. tremu- loides) were used. The material on L. laricina was sectioned. 6. Melampsora Larici-Capraearum Kleb. The spermogonia of Melampsora Larici-Capraearum are borne on needles of Larix laricina. They are abundant, small, discrete or occa- 120 JOURNAL OF THE ARNOLD ARBORETUM [vVoL. xvi sionally confluent, pustulate, amphigenous, subcuticular and in vertical section conoidal to hemispherical (Fig. 9). The epidermis is somewhat depressed by the overlying spermogonium. Its cells are separated and sometimes killed by the hyphae which have passed through to form the basal stroma. Sometimes the area formerly occupied by these cells is filled with a mass of interwoven hyphae. Thirty-one spermogonia meas- ured 30-101 pt broad and 18-38 uw high, averaging 65 & 28 u. The sper- matiophores may be septate towards their bases. Usually they are single-celled but they may be bicellular and the cells are uninucleate. The spermatia are catenulately produced and they measure 1.5—2.8 3—4 ww. The pore through which the spermatia are exuded measures 7.5 28.0 u wide and may extend over the entire length of the spermo- gonium. Figure 8 represents an immature spermogonium. Here the cuticle may be seen stretched beyond the free ends of the spermatio- phores. This condition has been observed frequently. Spermogonia and caeomata were obtained on Larix decidua and L. laricina from the European rust in its telial stage on Salix caprea. The spermogonia on both species of Larix were studied. II. Subfamily PUCCINIASTREAE 7. Melampsoridium betulinum (Pers.) Kleb. The spermogonia of Melampsoridium betulinum occur on needles of Larix decidua. They are abundant, minute, amphigenous, subcuticular, pustulate and in vertical section hemispherical flattened to conoidal ‘(Fig. 5). The epidermal cells may remain in situ or may be separated by mycelium. The spermogonia measured 45—53 u broad and 12-15 u high. The spermatiophores arise from a stroma which forms over the epidermal cells. The opening through which the spermatia are emitted is a slit 8-11 up wide. The spermogonium of M. betulinum is of the same type as that of Pucciniastrum species and of such species of Melampsora as M. Bigelowii and M. Medusae. Arthur (15) reports the haploid stage on Larix laricina (Du Roi) Koch from Connecticut and Wisconsin. An examination of specimens from the materials on which the American records are based (supplied through the courtesy of Dr. G. P. Clinton) shows the rust is Melampsora sp. True, a few weak peridial cells are present in the aecia but nowhere is there any indication of a well-defined peridium. Since it is now known that fragile, evanescent peridia are present in certain Melamp- sora species, this may account for the confusion which has resulted in error in determination. So far then as verified records go, the haploid stage of Melampsoridium betulinum has yet to be found in North America. 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 121 In 1898 Klebahn (51) established the genus Melampsoridium. He successfully cultured the rust on Larix and identified the haploid stage (bearing typical peridermia) with the Caeoma Laricis of Plowright (63). The writer has cultured from the telial stage of M. betulinum (from Betula verrucosa) to Larix decidua and L. laricina, obtaining scanty in- fection of both spermogonia and aecia. The latter are true peridermia. 8. Melampsorella Cerastii (Pers.) Schroet. A description of the rhorphology and of the ontogeny of the spermo- gonium of Melampsorella Cerastii is already recorded by the writer (45) under the name Melampsorella Caryophyllacearum. Figure 10 adds an excellent photegraphic illustration. 9. Pucciniastrum Epilobii (Pers.) Otth. 10. Pucciniastrum Abieti-Chamaenerii Kleb. The spermogonia of one of these species, the identity of which was unknown at the time, has already been described by the writer (45) under the name P. Epilobii. Subsequently, authentic culture materials of both were made available through the courtesy of Professor Faull, and their spermogonia have been studied in detail. No constant differences were evident. Professor Faull, however, pointed out that there are spe- cific differences in the aecia of the two species, and my examination of his herbarium specimens leads me to the same conclusion, 11. Pucciniastrum americanum (Farl.) Arth. The spermogonia of Pucciniastrum americanum occur on leaves of the current season of Picea glauca (Moench) Voss. They are small, dis- crete, amphigenous, irregularly scattered, subcuticular, pustulate, and in vertical section hemispherical flattened in shape. The epidermal cells at the base of the spermogonium are separated by hyphae passing be- tween them into the spermogonium. As the rust nears maturity these cells are somewhat depressed and frequently die; but there is no great disruption of host tissue. Twenty-nine spermogonia were found to measure 71-185 broad and 18-62 w high, averaging 124 35 nL. The spermatiophores are usually either nonseptate or one- to two- septate in the lower third of their length. They branch, just as do those of P. arcticum (Fig. 13), from enlarged basal hyphae. The spermatia form catenulately and they measure 1.5 & 3.5 win my sections. Meas- urements of spermatia from fresh material according to Darker (28) are 2.0-2.4 * 3.9-5.9 u. Frequently the opening through which the sperma- tia emerge measures 7.5—45.0 broad and extends as a slit over the greater part of the spermogonium. The cuticle may curl back exposing a considerable area of the upper surface of the spermogonium. At ma- 122 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII turity “flexuous hyphae” have been found extending beyond the tips of the spermatiophores through the opening of the spermogonium. 12. Pucciniastrum arcticum (Lagerh.) Tranz. The spermogonia of Pucciniastrum arcticum occur on first year needles of Picea glauca and resemble those of Pucciniastrum americanum so closely that it is probably impossible to separate them on a morphologi- cal basis (Fig. 12). It should be noted, however, that the sperma- tiophores of P. arcticum are commonly one-septate near their bases (Fig. 13). The spermatia, too, are possibly distinctive. In prepared sections they were found to measure 1.5-3.0 4.5-6.0 u. Darker’s measure- ments (28) made from fresh material were 1.3-1.6 & 3.9-4.7 U. 13. Thecopsora minima (Schw.) Syd. The spermogonia of Thecopsora minima are borne on the leaves and young cones of T’suga canadensis. Spermogonia on the leaves only were examined by the writer. They are small, usually discrete but occasion- ally confluent, amphigenous, irregularly scattered over the leaf surface, subcuticular, pustulate and in vertical section conoidal to hemispherical flattened. Thirty-four spermogonia measured 38-120 uw broad and 15- 25 uw high, averaging 79 & 20. The opening through which the sper- matia are exuded measures 14-15 u wide and extends as a slit over the spermogonium for most of its length. The materials studied were ob- tained from Professor W. P. Fraser. He secured them by inoculating Tsuga canadensis with basidiospores from Rhodora canadensis. A difference of opinion exists (37, 16, 38, 39, 64, 1, 70) as to whether or not 7. minima and T. Myrtilli are distinct species. It is doubtful if the spermogonial characters afford any aid in answering this question. It should be added that in all of the experimental work carried on in connection with Thecopsora minima and T. M yrtilli the infection experi- ments necessary to prove that the uredial stage on the various telial hosts Vaccinium, Gaylussacia, Rhodora and Azalea have never been per- formed. Gross inoculations with urediospores from one diploid host plant to the other would be of some value; but the experiments should be carried on by inoculating from the telial stages on various hosts to Tsuga, and when the aecia are secured inoculations should be made from each source back to all the various telial hosts. It would be well to in- clude Thecopsora Hydrangeae in such a set of experiments. 14. Thecopsora Myrtilli (Schum.), n. comb. Aecidium ? Myrtilli Schum. My study of the spermogonia of Thecopsora Myrtilli is based on cul- ture material obtained by Fraser (38) as the result of inoculations from 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 123 Vaccinium canadense Kalm on to Tsuga canadensis (L.) Carr. No sig- nificant differences were found between those of T. Myrtilli (Fig. 14) and T. minima, 15. Thecopsora Hydrangeae (B. & C.) Magn. The spermogonia of Thecopsora Hydrangeae are borne on first year needles of Tsuga canadensis. They are small, usually discrete but some- times confluent, amphigenous, subcuticular, pustulate and in vertical section hemispherical flattened to low conoidal. The stroma of the spermogonium does not usually depress the underlying epidermal cells. Thirty-five spermogonia measured 75-126 uw broad and 15-26 uy high, averaging 99 <-24 uw. The aperture through which the spermatia are discharged is a slit 15—23 w wide. 16. Calyptospora Goeppertiana Kuhn Since an earlier study of the spermogonia of this rust was made by the writer (45) examination of another lot of cultured material confirms the fact that spermatia do not ordinarily form and that the cuticle above the spermogonium remains unbroken. More than five hundred spermogonia have been examined. In these, spermatia were never found to have emerged. In many of the sections mature aecia were present. Fre- quently the spermogonia are dwarfed. Occasionally growth proceeds little further than is shown in Fig. 11. 17. Hyalopsora Aspidiotus (Pk.) P. Magn. The morphology of the spermogonia of Hyalopsora Aspidiotus has already been adequately described (45) and due reference made to Mayor’s findings in European material (57). Recently Kamei (50) has described similar but apparently somewhat bulkier spermogonia of his new species H. aculeata, cultured by him from Blechnum Spicant var. nipponicum (K.) M. & K. to Abies Mayriana Miy. & Kudo. In con- tinuation of my studies on H. Aspidiotus on Abies balsamea I am now able to present for the first time an account of the ontogeny of its sper- mogonia. The young spermogonium arises subepidermally from a very loosely scattered weft of mycelium in the outer intercellular spaces of the mesophyll (Fig. 16). At a very early stage in its development, hyphae (potential spermatiophores) branch off anticlinally from this weft; and very soon they come to form a continuous layer directly under the epidermis. During the latter process the intervening mesophyll cells are loosened from their connection with the epidermis and they and their neighbors in the outer layer of mesophyll are pressed downward into the spaces below. Frequently a number of the mesophyll cells become in- 124 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xvi volved in the growth of the spermatiophores. They are surrounded and eventually disrupted (Fig. 15). At maturity the epidermal cells above the spermogonium are usually killed; occasionally their walls are broken down by the mycelium (Fig. 16); and at times they may be penetrated by haustoria (Fig. 16). During the formation of the spermogonium the host leaf tissue becomes much swollen (Fig. 15) and drops of tannin and resinous matters form in affected cells — the former being especially abundant in the immediate neighborhood of the spermogonium. Within this region, too, starch grains fill the cells in abnormal amounts. 18. Milesia intermedia Faull In an earlier paper the mature spermogonium of Milesia intermedia on Abies balsamea was described by the writer (45) under the name Mulesina Kriegeriana Magnus. Faull (35) has recently shown that the latter is a different species and that it is not known to occur in America. The spermogonia of M. intermedia, just as is true of other known Mi- lesian spermogonia, are depressed in the host tissues at maturity. But an examination of mature spermogonia of M. intermedia, just as with most species of Milesia, leaves one in doubt as to their relation to the epidermis, that is, as to whether they are subepidermal, intraepidermal or subcuticular. To arrive at the correct interpretation in such species ontogenetic studies are necessary. These have been made for several of them by the writer. An account of the ontogeny of the spermogonium of M. intermedia follows. It plainly is of subcuticular origin. i, @ spermogonium of Milesia intermedia (Fig. 17) originates in the ou:<r epidermal wall. Hyphae separate the granular layer underlying th: cuticle from the wall strata below and growth of the young spermo- gonium continues in that region. As growth progresses the cuticular and granular strata are elevated and the underlying epidermal cells are depressed. The formation of the pore in the cuticle, through which later the spermatia are exuded, is of interest. A spermogonium about half de- veloped is shown in Fig. 18. At a central point above the row of spermatiophores, the granular stratum which normally occurs immedi- ately below the cuticle has disappeared. It is at this point that the cuticle breaks, the rupture being caused by pressure exerted by lateral growth of the fungus, and also by the filling of the cavity in the spermogonium with spermatia and spermatial fluid. At this intermediate stage (Fig. 18) in the growth of the spermo- gonium the spermatia are well developed. The spermatiophores are fewer and much more separated than in the mature stage. The growth of the spermatiophores is from the stroma. They increase in number by 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 125 branching from the cells of the stroma. As the spermatiophores multi- ply the underlying epidermal cells are pressed more deeply into the host tissue and some of them completely disintegrate. The mature spermogonium, which, in an early phase was raised and lenticular, is approximately plane with the leaf surface on its outer facies. Its body is hemispherical and deeply immersed as the result of prolific basal growth. 19. Milesia fructuosa Faull Milesia fructuosa was produced in Faull’s (36) cultures on various species of Abies (Table I), and always on leaves of the first season. In these experiments the rust developed scantily on A. nephrolepis Maxim., A, amabilis (Dougl.) Forbes and A. balsamea and sometimes only sper- mogonia were borne; on A. concolor Lindl. & Gord. the rust was more abundant; but on A. magnifica Murr., A. Fraseri (Pursh) Lindl. and A, Fraseri prostrata Rehd. it was not only abundant but it distorted the leaves of the host. The distorting effect was very marked on A. mag- nifica. The writer studied the spermogonia obtained in these cultures; the results are recorded in Table I. TABLE I. Number of Range in size Average size spermogonia Breadth X Height Breadth X Height Fir host measured in in Remarks A. amabilis 8 100-180 « 68-84 134 x 70 Mostly hypophyllous A. balsamea 17 112-140 x 76-96 137 x 81 Mostly hypophyllous A. cephalonica — Dead areas only Ss A. concolor 32 80-160 52-92 121 67 Mostly hypophyllous A. Fraseri 64 72-152 « 40-100 117 x 76 Mostly hypophyllous A. Fraseri prostrata 68 88-144 « 40-100 116 x 68 Mostl hypophyllous A. magnifica 198 80-172 « 48-96 121 60 Amphigenous A. nephrolepsis 6 108-180 « 40-80 132 XK 66 Mostly hypophyllous It will be noted that the measurements for the spermogonia of M. fructuosa vary somewhat from those given for M. intermedia; yet, if the mean or average measurements of the former are considered, they closely approach those given for the latter. Partly for this reason Faull (36) is inclined to regard these two as the same species (M. fructuosa). Certainly their spermogonia are, in general, similar in form and measure- 126 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII ments; and both are of subcuticular origin. The greatest aberration was found on Abies magnifica — on that host three percent of the spermo- gonia proved to be subpustulate. 20. Milesia marginalis Faull & Watson The spermogonia of Milesia marginalis have been described by Hunter (45), Faull (35) and Arthur (16) as being subcuticular; and that is what mature spermogonia, with very rare exceptions, appear to be. But a study of their ontogeny shows that they originate subepidermally; and then, as development proceeds, by partially breaking down the covering epidermal cells from below, they come to lie in a subcuticular position much as though they had originated directly under the cuticle. Stages shown in this interesting sequence are represented in Figs. 19-21; and the condition at maturity is illustrated in Fig. 23 of this paper and Fig. 10 of an earlier paper by the writer (45). Figure 19 portrays a spermogonial primordium—a thin mat of hyphae that is separating the epidermis from the underlying layer of mesophyll cells. On its outer facies it has projected haustoria into the epidermal cells and has already begun disintegration of epidermal cellu- lar contents and lower cell walls. Indeed, hyphae now completely oc- cupy a median epidermal cell shown and press against the inner face of its outer wall. The central part of the pad has forced the underlying mesophyll cells downward — the beginning of a process that terminates in a deeply depressed type of spermogonium. Figure 20 represents a later stage, but a still very young condition of a spermogonium. The spermatiophoric layer is now clearly defined; destructive action from below outwards on the epidermis is proceeding apace; and the depression of the spermogonium into the mesophyll is much more marked. Figure 21 is drawn from a more advanced stage of an immature spermogonium. All that now remains of the overlying epidermal cells are the cuticle, outer strata of the superior walls and tooth-like portions based on them of the anticlinal partitions. Depression of the basal stroma is continuing, but the final hemispheric form has not yet been attained. Perhaps the most striking new feature is the space now lying between the spermatiophoric layer and the reduced epidermal covering. This is the beginning of a hemispherical, fluid-filled cavity seen in a typical mature spermogonium. As yet spermatia are not forming but migrations of nuclei into the tips of a few of the spermatiophores and appearances of subterminal constrictions indicate that the first spermatia are being organized. Figure 23 (cf. Hunter, 45) is a good photomicrograph of a section 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 127 through a mature spermogonium. Spermatia are being produced abundantly and the central space is doubtless filled with spermatial fluid. The position of the slit through which the spermatia are discharged is evident. It is of particular interest to note that the epidermal membrane consists of the cuticle and a dark-staining layer (the middle of the outer epidermal wall). On the latter are based projections representing middle lamellar substance of the anticlinal partitions. In conclusion two observations are worthy of record. (1) Very oc- casionally the overlying epidermis may remain partially (Fig. 22) or wholly intact. (2) A central core of hyphae much longer than sper- matiophores has been seen in spermogonia that have reached about three-quarters of their full size. These are interpreted as being “flexu- ous hyphae,” that is, hyphae which possibly are receptive organs with respect to the phenomenon of fertilization. 21. Milesia polypodophila (Bell) Faull The morphology of the spermogonium of Milesia polypodophila (Fig. 26) has already been described by Hunter (45). Its ontogeny has not been studied, but note should be made of the fact that ‘“‘flexuous hyphae”’ are present at the time the spermogonium has matured (Fig. 24). 22. Milesia Blechni (Syd.) Arthur The spermogonia of Milesia Blechni occur on leaves of the current season of Abies alba (Fig. 30) and A. cephalonica Loud. They are amphigenous but usually hypophyllous, immersed and more or less flask-shaped in vertical sectional view. Nine spermogonia measured 110-175 uw broad and 105-150 u high. In the absence of ontogenetical studies it is not possible to decide definitely whether they are of sub- cuticular or subepidermal origin. 23. Milesia Kriegeriana (Magnus) Arthur The spermogonia of Milesia Kriegeriana are borne on leaves of the current season of Abies alba, A concolor and A. grandis Lindl. They are inconspicuous, colorless, abundant, amphigenous, irregularly scat- tered, subcuticular, immersed, extending well into the mesophyll and hemispherical in sectional view (Fig. 31). No discoloration of the leaf or presence of any lesion can be detected with a hand lens until the spermogonia reach maturity, that is, at the time the spermatial fluid begins to be exuded. Then the affected areas become slightly lighter in color and of a pale yellowish tint. Sixteen spermogonia measured 98— 168 uw broad and 94-168 u high, averaging 129 & 126 u. The sperma- tiophores are septate at a distance of about one-third of their length from the base of the spermogonium. They are rarely two-septate. The 128 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xvi cells are uninucleate. The spermatia are catenulate, hyaline and nar- rowly cylindrical, measuring 1.5—-2.0 & 3.5-5.0 uw. They are exuded through a slit-like pore measuring 7.5-15.0 & 25-80 np. It should be added that while the spermogonia of M. Kriegeriana on A. alba and A. grandis are identical, those on A. concolor are shallower on the average. Mayor (58) describes the spermogonia of M. Kriegeriana on A. alba as small, rounded, immersed in the tissues and 34—42 u in diameter — a description so different from that of the writer’s that the question arises as to the identity of Mayor’s materials. My studies on the ontogeny of the spermogonia of M. Kriegeriana demonstrate their subcuticular origin. Previous to the formation of the young spermogonium, hyphae separate the epidermal cells laterally and then extend periclinally within the outer walls of the epidermal cells. The lateral extensions lie in a plane just below the intermediate wall stratum. Before long the epidermal cells appear to be surrounded by mycelium. The primordial stroma of the spermogonium originates, however, from the layer of mycelium in the outer epidermal wall. Young spermatiophores are formed anticlinally on this stroma. The cuticle is raised above the surrounding leaf surface as in M. intermedia. After this stage is reached, basal growth of the spermogonium takes place rapidly. The epidermal cells are depressed and eventually disrupted. The basal stroma expands, the spermatiophores increase in number and the spermogonium as a whole begins to take on a hemispherical form. These processes, accompanied by flattening or destruction of contiguous mesophyll cells, continue until maturity is reached. The eventual form of the spermogonium is very similar to that of M. intermedia. 24. Milesia Polypodii White The spermogonia of Milesia Polypodii are borne on leaves of the current season of Abies alba and A. concolor. They are inconspicuous, abundant, amphigenous, subcuticular, colorless, immersed in the mes- ophyll, in vertical section hemispherical to slightly flask-shaped. No lesion or discoloration of the leaf is present until very shortly before the spermogonia open; then the affected areas become slightly lighter in color and of a pale yellowish tint. Thirty-four spermogonia measured 120-228 u broad and 105-194 u high, averaging 177 & 162 u. They are usually broader than high. The spermatiophores are frequently septate at about one-third of their length from the basal stroma. They may be non-septate or two-septate. The spermatia are narrowly cylindrical, hyaline, catenulate, and measure 1.5—2.0 & 4-5 uL. Although many of the spermogonia on Abies concolor were deeply immersed and reached a size almost as great as those on A. alba, 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 129 nevertheless, they were frequently found to be raised somewhat above the surrounding epidermis and in such instances were much shallower than the immersed type on A. alba. Milesia Polypodti on A. concolor caused much distortion of the leaves, whereas on A. alba there was no noticeable distortion except that the leaf was distended where spermo- gonia were present. In both hosts the mesophyll cells of the region in which they occur are destroyed in the course of development. The ontogeny of the spermogonia of M. Polypodit was studied from culture materials on Abies concolor and A. alba. They form essentially in the same way as those of M. Kriegeriana. Swollen hyphae separate the mesophyll cells immediately underlying the epidermal cells, and these are also separated laterally from one another by strands of mycelium. Eventually a mat of hyphae, by growth in a periclinal direction, cleaves the outer epidermal cell walls, thus coming to lie just below the layer that consists of the cuticle and the granular stratum. It is mainly from the mat developed in this location that the spermogonium originates. On the upper facies of the mat spermatiophores grow out towards the cuticular covering and the isolated epidermal cells are ir- regularly distributed near its base. This stage is illustrated in Fig. 25, taken from a spermogonium about one-third grown. When the spermo- gonium has reached from one-half to two-thirds of its mature size (Fig. 32) a very interesting morphological phenomenon is to be observed. The spermatiophores, not yet fully grown and not sporulating, are pres- ent in a semi-circular ring; while much longer, broader, non-tapering, non-septate, uninucleate hyphae, apparently rising from the same stroma as the spermatiophores, form a central core. The spermatiophores stain very deeply in contrast with the hyphae of this core. At an earlier stage in development than that just described, the central core of hyphae may be in contact with the overlying membrane of the host; but it soon be- comes separated and when the “‘slit” is formed the hyphae of the core emerge and grow out 20 u or more beyond the covering of the spermo- gonium. Sections of nearly mature spermogonia in which the spermatia have formed, but in which the cuticle is not yet ruptured, show that the spermatia have no particular tendency to cling to the long hyphae; but once the spermogonium has ruptured and there has been opportunity for contact with spermatia from other spermogonia the long hyphae may be found with spermatia attached. The dark band lying over the spermogonium shown in Fig. 33 is made up of spermatia apparently massed around the long hyphae. Occasionally a hypha is seen pro- truding from among the exuded spermatia and sometimes one can be recognized when cut transversely in section. Just what these long 130 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII hyphae are is not certain, but according to my interpretation they are “flexuous hyphae” as described by Craigie (27) for certain Pucciniaceae. 25. Milesia Scolopendrii (Fuckel) Arthur The spermogonia of Milesia Scolopendrii are borne on leaves of the current season of Abies alba and A. concolor. They are inconspicuous, amphigenous, apparently subcuticular, colorless, immersed in the mes- ophyll, in vertical section conspicuous and hemispherical to slightly flask-shaped. Twenty-eight spermogonia measured 120-228 wp broad and 100-188 u high, averaging 156 & 140 uy. These measurements are within close range of those given for M. Polypodii and it is difficult to distinguish these rusts by their spermogonia. The spermatiophores are usually one- or two-septate near their bases; but occasionally they are entire. The spermatia are catenulate, hyaline, narrowly cylindrical, 1.5— 2.0 < 4-5 u. The opening through which they are exuded measures 5-38 & 40-110 u. No early stages in the development of the spermogonia were studied. But interesting observations were made on “‘flexuous hyphae” in mature or submature spermogonia. Spermogonia, twenty-seven to thirty days following inoculation on Abies alba and from which spermatia were just beginning to discharge, contained unusually long ‘“flexuous hyphae” protruding far beyond the apical “slit.” Hand sections of fresh mate- rial were made, then mounted in water, and finally preserved in lacto- phenol tinted lightly with acid fuchsin stain. Some of the “flexuous hyphae” in these sections were turgid and filled with cytoplasm, while others appeared empty and flaccid. One section was of especial interest (Figs. 28, 29). In it the spermatiophores were of normal length and spermatia were present; but standing out in the central cavity there was a long flexuous hypha with spermatia attached, extending slightly be- yond the aperture and about 105 4 beyond the tips of the spermatio- phores (Fig. 27). Another hypha was present to which spermatia were also attached, and these, as in the former instance, had remained ad- herent throughout the processes of washing and staining. In this same section there were at least six other ‘“‘flexuous hyphae” in various condi- tions of turgidity or flaccidity. It may be added that “‘flexuous hyphae” were never found in old spermogonia. While the writer is not prepared to venture an opinion on the part played, if any, by “flexuous hyphae” in the sexual phenomena of rusts, yet, in continuation of such studies as have been made on this topic by Craigie (24-27), Pierson (62), Andrus (10) and Allen (7-9) it is suggested that species of Milesia would ap- pear to afford excellent material. Their spermogonia are relatively large, they lack confusing paraphyses and they produce striking “flexu- ous hyphae.” 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 131 26. Milesia vogesiaca (Sydow) Faull The spermogonia of Milesia vogesiaca (Fig. 34) occur on leaves of the current season of Abies alba. They are inconspicuous, very abund- ant, amphigenous but mostly epiphyllous, apparently subcuticular, plane on the upper surface, deeply immersed in the mesophyll, and in vertical section spherical to slightly flask-shaped. Thirty-four spermogonia measured 154-241 u broad and 168-214 w high, averaging 201 & 195 np. The spermatiophores branch from enlarged hyphal cells of the stroma which forms the base of the spermogonium. They are usually one- septate towards their bases but they may be two-septate, the second septum being about one-third of the length of the spermatiophore from its base. The spermatia are narrowly cylindrical, hyaline, 1.5-2.0 4-5 u. The opening through which they are exuded is a slit-like pore in the cuticle, 5-35 & 30-70 uw. “Flexuous hyphae,” 6—9 wu in diameter, occur in mature or submature spermogonia. The spermogonia of M. vogesiaca are much larger than those of any of the other European species of Milesia. They more nearly approach those of M. polypodophila in size, although the latter are generally larger. The spermogonia of M. polypodophila are subspherical, being elongated slightly along the vertical axis; while those of M. vogesiaca, also subspherical, are most frequently slightly elongated along the horizontal axis. Moreover, the spermogonia of M. polypodophila are distinctly subepidermal and also differ from those of M. vogesiaca in their much longer incubation period. Confusion as to the position of the latter in the host may result because the subcuticular area above the spermogonium is so narrow that in a lateral section there is an ap- pearance of being subepidermal. Ontogenetic studies alone can de- termine whether or not the spermogonia are really of subcuticular origin. 27. Peridermium rugosum H. S. Jackson The morphology of Peridermium rugosum was described by the author in an earlier paper (45). 28. Uredinopsis mirabilis (Peck) Magnus The spermogonia of Uredinopsis mirabilis occur in Abies balsamea on leaves of the current season. They are inconspicuous, rarely confluent, hypophyllous, subcuticular, colorless, immersed in the mesophyll, in vertical section hemispherical and either quite flat on top, slightly raised above the surrounding epidermis or, in late maturity, centrally depressed. Forty-nine spermogonia measured 58-123 w broad and 35-54 u high, averaging 89 & 44 u. The same condition which exists in Mdesia intermedia is found in U. mirabilis, namely, the combined cuticle and 132 JOURNAL OF THE ARNOLD ARBORETUM [voL. xvi the intermediate wall layer remain over the spermogonium at maturity. The spermatiophores are one-septate, or occasionally two-septate, towards their bases. The basal cells are long and tubular, narrowing toward their upper ends. Formerly (45) this was thought to be an aid in distinguishing Uredinopsis species from Milesia species; but equally large basal tubular cells have since been found in the spermatiophores of Milesia species. The spermatia are oval, hyaline, catenulate and meas- ure 1.5-2.0 & 3.5—4.2 u. The opening in the cuticle through which they are exuded is a pore or short slit, 4-9 « 10-40 LL. Doubt has existed as to whether the spermogonia of this and various other species of Uredinopsis are subcuticular or subepidermal. Thus Arthur (15), in referring to the mature spermogonia of Uredinopsis, speaks of them as “extending between and depressing the tissues be- neath, giving the appearance of being subepidermal.” Indeed, it seems from the mature spermogonia as if this might be the case; but investiga- tions of the earlier phases of their development in U. mirabilis show them to arise in a subcuticular fashion. After a certain amount of growth, the haploid mycelium begins to be directed towards the leaf surface and passes between the cells, along the middle lamellae, occa- sionally completely separating the epidermal cells from one another. Then the hyphae grow laterally and separate the outer epidermal wall into two layers. There is an interweaving of these hyphae and a stroma is organized; simultaneously, branches of mycelium from the stroma are sent out anticlinally and gradually the cuticle is raised. At the same time the underlying epidermal cells are slightly depressed due to pres- sure from the growth of the developing spermogonium. Figure 35 is an illustration of an early stage in the development of the spermogonium. It shows that the epidermal cells at the base have been separated and at least partly surrounded by the mycelium. Although in its very early stages the spermogonium is superficial and raised above the epidermis, it later becomes immersed in the leaf tissue. The young spermatiophores increase in number and lengthen. The epi- dermal cells gradually become crushed and finally they may disappear entirely. Usually the mesophyll cells immediately underlying the sper- mogonium are pushed downward into the leaf, probably filling up an intercellular space; occasionally they may be involved in the growth of the stroma and disappear. As the spermogonium becomes immersed the overlying cuticle resumes its former position and is more or less on a level with that covering the surrounding epidermis. 29. Uredinopsis Atkinsonii Magn. (Fig. 36) 30. Uredinopsis Osmundae Magn. 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 133 31. Uredinopsis Phegopteridis Arthur 32. Uredinopsis Struthiopteridis Stormer The spermogonia of all the four preceding species except Uredinopsis Struthiopteridis, have already been described by the writer (45). All resemble those of U. mirabilis so closely that these species cannot be distinguished from one another by their spermogonia. A recapitulation of their morphological data is found in Table IT. 33. Uredinopsis Pteridis D. & H. Little can be added to the account of the spermogonia of Uredinopsis Pteridis already given by Hunter (45) except to say that the literature on U. Pteridis-and Peridermium pseudo-balsameum Arthur and Kern is a maze of confusion (cf. Schmitz, 67; Weir and Hubert, 74; Rhodes, Hedgcock, Bethel and Hartley, 64; Hunter, 45; Kamei, 49). 34. Peridermium balsameum Peck Peck (61) discovered a white-spored rust on the leaves of Abies bal- samea Mill., which in 1875 he named Peridermium balsameum. For many years, as pointed out by Faull (34), this name has been applied to the aecial stage of various North American species of the two genera Milesia and Uredinopsis. Fraser (38) was the first to demonstrate that five species of Uredinopsis, namely, U. Struthiopteridis, U. Osmundae, U. Atkinsonii, U. Phegopteridis, and U. mirabilis have their aecial stage on Abies balsamea. He considered that the peridermia of all of these would pass current for the P. balsameum of Peck. The occurrence of these five species on Abies was later confirmed by culture experiments carried out under the direction of Professor Faull (34). The Sydows (70) in referring to Fraser’s work stated that the peridermia of the various North American species of Uredinopsis recorded in their Mono- graphia cannot be distinguished from Peck’s P. balsameum but they cite P. balsameum as a synonym of U. Struthiopteridis only. Arthur (15) chose to refer P. balsameum to U. mirabilis; but he also stated that the aecial stages of U. Osmundae and U. Copelandi Sydow are similar to those of U. mirabilis. Arthur and Kern (17) redescribed P. balsameum Peck. Their material comprised numerous specimens, including what they considered Peck’s type material from the Adirondack Mts. and without realizing that in it more than one species was involved. Inci- dentally they ventured the opinion that Aecidium pseudocolumnare Kihn, as maintained by Farlow, might be identical with P. balsameum. Faull (34) discovered that three American species of the genus Milesia on Abies balsamea, as well as those of Uredinopsis, occur on the leaves of Abies balsamea. Of these, M. polypodophila (Bell) Faull, 134 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xvi easily distinguished from the others by the occurrence on leaves 3-8 years old and by its strikingly distinctive morphological characters, apparently had been overlooked by collectors; but the remaining species referred to by him and occurring on leaves of the current season (as do those of the Uredinopsis species mentioned), had no doubt passed under the name P. balsameum. It is now known that the genera Milesia and Uredinopsis in North America, as they occur on Abies, can be distinguished from one another by their spermogonia, and that similarly the species Milesia intermedia (perhaps more properly named M. fructuosa Faull), M. marginalis and M. polypodophila possess specifically distinctive spermogonia (45). Hence it is now possible to more exactly determine the identity of what Peck himself preserved in his herbarium under the name Peridermium balsameum. With this object in view the writer has critically examined Peck’s collections stored in the Herbarium of the New York State Museum at Albany, N. Y. and made available through the courtesy of Dr. H. D. House. The writer found two sheets of exposed specimens and two packets, all ascribed to Professor Peck as collector. There are various later collec- tions in the Herbarium but these are not of pertinent interest here. The first sheet, regarded as the oldest (dating back possibly to 1873), carries several exposed specimens and bears the label, ‘““Peridermium balsameum Peck, Adirondack Mts., New York State, July and Aug.” Presumably it is the type sheet, though neither it nor any other is so marked. The specimens on this sheet prove to be predominantly Milesia intermedia; but one of them is a species of Uredinopsis. The second sheet, with several specimens mounted similarly to the first, bears the label, ‘“Peri- dermium balsameum Peck, Charlotteville Swamp, N. Y. State, Aug.” It is M. intermedia. ‘The third collection was made at N. Elba, N. Y. State, in Aug. 1910, according to the label; it is also M. intermedia. The fourth collection is recorded as from L. Avalanche, Adirondacks, N. Y., dated Aug.; but as with the first two sheets, no mention is made of the year. It is a species of Uredinopsis. The writer’s conclusions are: (1) Peck’s own collections of white- spored peridermial rusts on Abies comprised species of both Milesia and Uredinopsis, and were designated by him as Peridermium balsameum ; (2) P. balsameum cannot accurately be designated as a synonym of any particular species of Uredinopsis, as has been done by Sydow and Arthur; (3) what is possibly the type sheet of Peck’s P. balsameum carries a large proportion of Milesia intermedia but some material of an undeterminable species of Uredinopsis. 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 135 In the study of these plants serial sections were made of selected leaves, and comparisons were made with similar preparations from authentic materials (obtained from controlled cultures) of the various species of Milesia and Uredinopsis mentioned above, made available through the kindness of Professor Faull. 35. Aecidium pseudocolumnare Kiihn The first collections of Aecidium pseudocolumnare were made by Kuhn in the Black Forest, Germany, in August and September, 1883. The spermogonia were not described but a detailed description of the aecia and aeciospores was given on labels attached to herbarium speci- mens. The aeciospores were said to be white, finely warted and to measure 18.5—25.7 uw broad and 22.8—-37.2 uw long. Winter (76) in 1884 published the description of this Aecidium word for word with the description signed by Kithn. Kihn’s collections were distributed in Rabenhorst-Winter, Fungi Europaei under the name of Aecidium pseu- docolumnare J. Kiihn, nov. sp. number 3027. It is taken for granted then that this material represents the type for A. pseudocolumnare. This name, however, came into general use in Europe as a designation of white-spored peridermia on Abies alba. That this has been done uncritically is illustrated by the results of comparative examinations made by the writer of spermogonia from several collections found in the herbaria of the Royal Botanic Gardens, Kew, and the British Museum of Natural History — all regarded as A. pseudocolumnare. Aeciospores and peridial cells did not prove sufficiently characteristic to be of diag- nostic value. Descriptions of the spermogonia of the various collections follow: (1) Rabenhorst-Winter, Fungi Europaei no. 3027, Aecidium pseudo- columnare J. Kihn, nov. sp. (bearing Kihn’s original description). The spermogonia from this collection are subepidermal and measure 160-200 u broad and 152-180 up deep. They closely resemble those of Milesia vogestaca. (2) Krieger, Schadliche Pilze unserer Kulturgewdchse no. 74, Aecidium pseudocolumnare J. Kiihn. The spermogonia are subcuticular and measure 92-140 u broad and 112-160 u deep. They most closely resemble those of Milesia Kriegeriana. (3) Krieger, Fungi Saxonici no. 1419, Aecidium pseudocolumnare Kiihn nov. sp. Very little of this material was available but enough was obtained to show a few subcuticular spermogonia measuring 130-132 u broad and 92-120 uw deep. They also resemble those of Milesia Krieger- tana. 136 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XvII (4) C. E. Broome, Peridermium columnare on Picea Nordmanniana (Abies Nordmanniana), Torquay, ex Herb. M. J. B. Aug. 1867. Ac- cording to J. Ramsbottom, Keeper of Botany, British Museum of Natural History, this material is originally from the herbarium of Miles Joseph Berkeley. Ten spermogonia from a fragment of the specimen were examined. Nine of these are definitely subcuticular while one is subcuticular only in the center with remains of epidermal cells at the sides overlying the spermogonium. They measure 152-208 u broad and 140-200 p deep. The evidence is that the subcuticular spermogonia may belong to either or both Milesia Polypodii and M. Scolopendrii. (5) M. C. Cooke, Fungi Britannici exsiccati no. 314, Peridermium columnare A. & S., near Torquay, Sept. 1869. E. Parfitt. The spermo- gonia are deep-seated, subcuticular, measuring 128-180 uw broad and 132-180 deep. They also may be referable to either or both Milesia Polypodii and M. Scolopendrii.’ (6) Aecidium pseudocolumnare Kiihn on Abies alba. Kelso, Aug. 1925, Malcolm Wilson. Only a small fragment of an infected needle was examined. The sections were not sufficiently satisfactory to warrant a definite conclusion, since the spermogonia seemed to be distorted. They are deep-seated, large, hemispherical, subcuticular and probably belong to the genus Milesia. (7) Aecidium pseudocolumnare Kihn. Powerscourt, Co. Wicklow, P. O’Connor. One collection is dated 11.11.31 and 6.2.31, while a second collection is dated 11.11.31. Although this material was collected late in the year, it is in excellent condition. The peridermia even in the dried state are long and unbroken, and almost every leaf on a branch of con- siderable size is rusted. For convenience the two collections are referred to respectively as (a) and (b). Two spermogonia examined from (a) are deep-seated, hemispherical and measure 101 u broad and 168 u deep. These most closely resemble the spermogonia of Milesia Kriegeriana. Four spermogonia from (b) were examined. They are deep-seated, sub- epidermal, elongate and measure 174-181 u broad and 181—188 wu deep. They closely resemble spermogonia of the type material of Aecidium pseudocolumnare. A study of the spermogonia in these collections, therefore, clearly in- dicates that the latter are referable to several distinct species. Hence, just as in North America, until recently, several white-spored rusts on Abies have been passing under the name Peridermium balsameum Peck, so it would seem that the same is true in Europe for Aecidium pseudo- columnare. 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 137 III. Subfamily CRONARTIEAE 36. Cronartium ribicola Fischer The spermogonia of Cronartium ribicola occur on Pinus Strobus L. and other five-needle pines. They are caulicolous, subcortical, forming small blisters. They are located between the outermost layer of cortical cells and the periderm, effuse or indefinite in type. Six spermogonia measured 1.0-3.5 mm. broad and 34-67 high. Figure 38 represents a mature spermogonium. Sydow (70) states that the spermogonia are about 2.5 mm. long. Colley’s (22) observation that “occasionally long filaments grow some distance out beyond the tips of the spermatio- phores” is of especial interest when viewed in the light of recent dis- coveries in connection with the fusion of spermatia and haploid hyphae. The spermatia according to Colley are pyriform and measure approxi- mately 2.5 & 3.5 u. IV. Subfamily CHRYSOMYXEAE 37. Chrysomyxa Ledi (A. & S.) DeBary The spermogonia of Chrysomyxa Ledi occur on first year needles of Picea mariana (Mill.) B.S. P. and P. glauca (Moench) Voss. They are amphigenous, subepidermal and flask-shaped in section (Fig. 37). If the spermogonium happens to form below a stoma the guard cells are carried above the spermogonium. Frequently the epidermal layer, though raised slightly, is left intact; but sometimes the lower walls are attacked and the cells killed. The hypodermal cells within the area of the spermogonium are disrupted, and disappear. The mesophyll cells below the spermogonium are depressed. Twenty-nine spermogonia measured 60-128 uw broad and 45-125 uw high, averaging 105 & 82 u. The spermatiophores arise from a basal stroma and are directed to a slit-like pore in the epidermis. They are usually one-septate near their bases. The spermatia measure 1.5-2.0 & 3.0-3.5 u. They are slightly elongate and form catenulately. The pore through which they are emitted measured 4-35 u broad and 20-80 u long. V. Subfamily COLEOSPORIEAE 38. Coleosporium Helianthi (Schw.) Arth. The spermogonia of Coleosporium Helianthi are borne on leaves of Pinus virginiana Mill. and P. echinata Mill. They are numerous, amphigenous and of the applanate, lenticular type; they resemble those of Hyalopsora Aspidiotus in shape except that they are elevated more above the surrounding epidermis and do not depress as much the mesophyll cells of the leaf at maturity. The spermogonia form between the mesophyll and the overlying tissues. The latter consist of an 138 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII epidermis, a hypodermis and from one to two layers of thick-walled cells irregularly placed under the hypodermis. The spermogonia are most frequently discrete; but they are sometimes confluent. Twelve spermo- gonia measured 323-458 wt broad and 92-116 uw high, averaging 361 X 103 u. The spermatiophores are usually single-celled. The spermatia are discharged through a slit 34-92 4 wide. A mature spermogonium showing its relation to the leaf is represented in Fig. 39. In addition to other features, attention should be drawn to long, centrally located hyphae that extend beyond the spermatiophores and the level of the epidermis. These are “flexuous hyphae.” Such hyphae may extend as much as 15 u beyond the epidermal level. When the spermogonia have passed full maturity and the spermatia have been discharged there is no trace left of these hyphae. (cf. Arthur 12, 14, Hedgcock 42, Hedgcock and Hunt 43.) The materials for the study of this and the other species of Coleo- sporium described below were kindly supplied by Dr. G. G. Hedgcock. 39. Coleosporium inconspicuum (Long) Hedgc. & Long The spermogonia of Coleosporium inconspicuum are borne on leaves of Pinus virginiana etc. The same general description of form and loca- tion applies as for the spermogonia of C. Helianthi, except that the spermatiophores are of slightly smaller diameter. Eleven of them measured 226-317 broad and 73-85 « high, averaging 284 X 81H. The immature spermogonium of C. inconspicuum depresses the mesophyll cells lying immediately under it; whereas in the mature spermogonium the latter have regained their normal position with the exception that they may be separated by the strands of mycelium. The cells above the young spermogonium (Fig. 40) are raised and flattened ; but on either side of the spermogonium they have retained their normal size and position. Differential staining shows that two layers of meso- phyll cells immediately under the spermogonium, in contrast with those more distantly located, have in some way been affected by the parasite. (cf. Arthur 14, Hedgcock 42, Hedgcock and Hunt 43.) 40. Coleosporium Ipomoeae (Schw.) Burr. The spermogonia of Coleosportum I pomoeae occur on needles of Pinus rigida Mill. etc. They are subepidermal and of the applanate type. They measure 335-490 ut broad and 104-159 uw high. According to Hedgcock (42) they are olivaceous-black to brownish-black, on slightly chlorotic areas. They are amphigenous in P. rigida and hypophyllous in P. echinata. 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 139 41. Coleosporium delicatulum (A. & K.) Hedge. & Long The spermogonia of Coleosporium delicatulum occur on needles of Pinus rigida etc. They are amphigenous, subepidermal, applanate, and measure 354 ut broad and 110-122 high. According to Hedgcock (42) they are orange chrome to English red on reddened chlorotic areas. 42. Coleosporium Laciniariae Arth. The spermogonia of Coleosporium Laciniariae occur on needles of Pinus rigida etc. They are discrete, epiphyllous, subepidermal, appla- nate and measure 323-397 uw broad and 110-122 u high. According to Hedgcock (42) they are olive to olivaceous-black on yellow chlorotic areas. The spermatiophores usually are one-septate towards their bases. 43. Coleosporium Solidaginis (Schw.) Thiim. The spermogonia of Coleosporium Solidaginis are borne on needles of Pinus rigida etc. They are usually discrete. They measure 244-366 UL broad and 98-104 uw high. In P. rigida they are hypophyllous. Accord- ing to Hedgcock (42) they are grenadine red on slightly reddened chlo- rotic areas. 44. Coleosporium Terebinthinaceae (Schw.) Arth. The spermogonia of Coleosporium Terebinthinaceae occur on needles of Pinus virginiana etc. They are epiphyllous and measure 335—427 UL broad and 80-101 u high. According to Hedgcock (42) they are “orange rufous to mummy brown, on yellowed chlorotic areas.” 45. Gallowaya pinicola Arth. The spermogonia of Gallowaya pinicola occur on the leaves of Pinus virginiana, They are of the applanate type, lenticular in vertical sec- tion, subepidermal and amphigenous. Seven spermogonia measured 244-427 w broad and 48-92 u high. Figure 41 represents a mature spermogonium. The spermogonia are nearly always centrally located beneath the guard cells in stomatal cavities and develop between the mesophyll and the epidermis. The underlying mesophyll cells are not usually depressed and the overlying epidermis is very slightly elevated. The contiguous epidermal cells are never flattened as they may be in Coleosporium species; but differential staining shows that their lower walls have been affected by the fungus. The stroma is made up of inter- woven hyphae overlying the mesophyll cells; from it spermatiophores branch and are arranged periclinally, curving slightly towards a central point under the epidermis. The spermatiophores are attenuate, fre- quently curved at their tips and usually one-septate, about one-third up from the base. Spermatia do not form. 140 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XvII Arthur (16) places Gallowaya pinicola in the genus Coleosporium. He states that the spermogonia are rudimentary and not externally visible; Dodge (32) previously described them as vestigial or abortive. Dodge further stated that host tissues above the spermogonia are not fully ruptured and that spermatia are seldom formed. Here we have a parallel case to that of Calyptospora Goeppertiana. In G. pinicola then, as in C. Goeppertiana, the spermogonia cannot function in the way described for Puccinia species by Craigie (27). CONCLUSIONS AND SUMMARY 1. The morphology of the spermogonia of representatives of all the genera, except Chnoopsora, of the Melampsoraceae (Dietel, 1928) has been studied in the course of the investigations recorded in this paper. 2. From these studies it is amply demonstrated that spermogonia are of more or less diagnostic value (just as are other organs) in defining or determining species and genera. The features of chief importance in that connection are the form-type, size and position in the host tissues. It should, of course, be recognized that these features can be employed in diagnosis with greater assurance where the complete life history is known and where the spermogonia have been studied and defined from authentic culture material. This may be illustrated by reference to those species of Melampsoraceae whose haploid stages occur on the Pinaceae. In eastern North America the life histories of many of them have been worked out and the spermogonia have been studied from authentic cul- ture material. It so happens that the aecial characters of many of these species are not sufficiently distinctive to enable one to recognize the species or in some cases even the genus from haploid material. The spermogonial characters, on the other hand, in many such instances have proved to be of diagnostic value and their characters have been incorporated in definitions of the species. It is now possible, therefore, to recognize many of the Melampsoraceae in eastern North America from haploid material alone and without having recourse to cultures. As for western North America, by contrast, littlke work has been done on life histories of coniferous rusts. An examination of the spermogonia in western collections of coniferous rusts shows that the spermogonia of some of these rusts are surely of diagnostic value; but the delimitation of species is bound to be tentative until life histories are worked out and the exact morphological characters of their spermogonia are incorporated in descriptions of the species. Much the same condition prevails with respect to the European Melampsoraceae. 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 141 3. The spermogonia of the Melampsoraceae may be classified ac- cording to their form and position in the host tissues, as follows: A. Subcuticular — (a) pustulate — (1) hemispherical to conoidal (b) immersed — (1) hemispherical; (2) sub- spherical to spherical B. Subepidermal — (a) applanate (b) immersed — (1) elliptical; (2) spherical to flask-shaped C. Subcortical 4. Spermogonia have been found for the main part to be character- istic in type for each genus. The one outstanding exception is that of Melampsora. In Melampsora, as now commonly recognized, two very diverse types of spermogonia are comprised — (1) species in which the spermogonia are robust and deep-seated; (2) species in which the sper- mogonia are small and pustulate. Another feature in this connection refers to the importance of giving attention to the spermogonia in surmises as to the genetic position of coniferous rusts where the life histories are unknown. This may be il- lustrated in such a rust as Peridermium coloradense. Certain uredinol- ogists have suggested that it is the haploid stage of a Melampsorella; indeed, some have gone so far as to refer it to Melampsorella Cerastii (75, 16). An examination of the spermogonia of this rust shows that it cannot possibly be a Melampsorella. Its type is that of a Chrysomyxa. 5. The spermogonia are of value in establishing or confirming views as to the phylogenetic relationship of genera. An excellent example of this is afforded by the genera Milesia and Uredinopsis. The similar characters of their spermogonia lend support to the view that these genera are closely related. Another example is afforded by Puccinia- strum, Thecopsora, Calyptospora and Melampsoridium, in all of which the spermogonia are morphologically similar. 6. Spermogonia have, as indicated above, been found to be of value in the determination of species from haploid material; in some such instances no help is afforded by aecial characters. For example — Me- lampsora americana and Caeoma Faulliana; Milesia intermedia, M. marginalis and certain other species of Milesia as distinct from one an- other and from species of Uredinopsis; and Calyptospora Goeppertiana as distinct from various species of Pucciniastrum. 7. Attention is called to the non-spermatia-forming rusts, Calyptos- pora Goeppertiana and Gallowaya pinicola, as examples of rusts in which spermatia cannot take part in initiating the diploid phase such as is presumed to be the case for Puccinia graminis, P. Helianthi etc. 142 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII 8. The ontogeny of the spermogonia has been studied more or less completely for the following species — Melampsora A bieti-Capraearum, M. Larici-Capraearum, Melampsorella Cerastii, Calyptospora Goepper- tiana, Hyalopsora Aspidiotus, Milesia intermedia, M. marginalis, M. Kriegeriana, M. Polypodii, Uredinopsis mirabilis, U. Atkinsonu, U. Osmundae and Coleosporium incons picuum., The value of the ontogenetical studies of some species lies in the circumstance that it is only in this way that the exact type can be de- termined. Thus a mature spermogonium may appear to be subcuticular when it is really subepidermal or subepidermal when it is really sub- cuticular. 9. The spermogonia of the Melampsoraceae lack paraphyses. But “flexuous hyphae,” possibly receptive organs, extending beyond the aperture of the spermogonium have been found in numerous species of the Melampsoraceae, namely, Milesia marginalis, M. polypodophila, M. Kriegeriana, M. Polypodii, M. Scolopendrii, M. vogesiaca, Melampsora Abieti-Capraearum and Coleosporium Helianthi. In Milesia Scolo- pendrii spermatia were occasionally attached to “‘flexuous hyphae” but the nature of the connection was not determined. ‘Flexuous hyphae” are greater in diameter for most of their length than spermatiophores and are not so much attenuated. 10. Branching of spermatiophores from the stroma has been demon- strated for Uredinopsis Atkinsonii and Pucciniastrum arcticum, During the development of the spermogonium the distal ends of the spermatio- phores remain free from the overlying epidermis or cuticle. At their maturity the spermatiophores are only about one-half the length of the “flexuous hyphae.” The spermatiophores decrease in length as they age. Their cells are always uninucleate. Spermatia are borne catenu- lately. The “slit” through which spermatia are discharged is, in gen- eral, parallel to the long axis of the leaf on which the spermogonium is borne. 11. One new species, Caeoma Faulliana, is described in this paper. 12. Morphological data on the spermogonia of the species herein studied are summarized below in tabular form (Table IT). 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 143 TABLE II SPERMOGONIA OF MELAMPSORACEAE* No. Size Shape meas- (breadth X height in microns) Species (vertical section) ured Limits Average Remarks I. SUBFAMILY MELAMPSOREAE Melampsora Elliptical 67 | 79-154 42-71 | 9948 | Subepidermal, A bieti- depressed Capraearum (M. americana) Melampsora Elliptical 54 | 83-138 45-67 |11054 | Subepidermal, A bieti- subdepressed Capraearum Caeoma Hemispherical 45 | 42-145 «27-49 | 80x37 | Subcuticular, Faulliana to conoidal pustulate Melampsora Hemispherical 34 | 35-98 15-38 | 6625 | Subcuticular, A bietis- flattened to (on needles) pustular canadensis conoidal 77 | 53-128 13-30 | 78x23 (on cones) Melampsora Hemispherical 42 | 50-102 15-38 | 6927 | Subcuticular, Bigelowtt to conoidal pustulate Melampsora Hemispherical 17 | 51-105 15-30 | 75X23 | Subcuticular, Medusae to conoidal ustulate Melampsora Hemispherical 31 | 30-101 x 18-38 | 6538 | Subcuticular, arici- to conoidal pustulate Capraearum II. SUBFAMILY PUCCINIASTREAE Melampsoridium Hemispherical — |} 45-53 «12-15 —— | Subcuticular, tulinum flattened t ustulate conoida] Melampsorella Hemispherical 72 | 99-317 27-59 |18438 | Subcuticular, Cerastii flattene ustulate Pucciniastrum Hemispherical 81 | 45-212 16-34 | 95x21 | Subcuticular, Epilobii flattened ustulate Pucciniastrum Hemispherical 35 | 67-196 16-31 |102 23 | Subcuticular, A bieti- flattened pustulate Chamaenerit *Notes: 1. Spermogo 2. Spermogo nia on | f t pt as indicated Jer ““Remarks’’. nia “‘amphigenous” except as indicated under ‘“‘Remarks”. The term ‘‘amphigenous”’ is 4 } i 3 £ a } +h J 1 t I 4 4) £ whether + +1 7 i or not they may usually 144 JOURNAL OF THE ARNOLD ARBORETUM [voL. XvII TABLE II. (Continued) No. Size Shape meas- (breadth X height in microns) Species (vertical section) ured Limits Average Remarks II. SUBFAMILY PUCCINIASTREAE. (Continued) Pucciniastrum Hemispherical 29 | 71-185 x 18-62 |124% 35 | Subcuticular, americanum flattene pustulate Pucciniastrum Hemispherical 29 |100-167 x 22-51 {142 32 | Subcuticular, arcticum attene ustulate Thecopsora Hemispherical 34 | 38-120 15-25 79 x 20 | Subcuticular, minima flattened to pustulate conoidal Thecopsora Hemispherical 15 | 41-105 «15-23 | 7220 | Subcuticular, M yrtilli to conoidal pustulate, hypophyllous Thecopsora Hemispherical 35 | 75-126 15-26 | 99x 24 | Subcuticular, Hydrangeae flattened to pustulate conoidal Calyptos pora Hemispherical 105 | 42-137 « 13-30 73 X21 | Subcuticular, Goeppertiana to conoidal pustulate, hypophyllous Hyalopsora Lens-shaped, 25 | 311-496 x 86-117 |432 x 102} Subepidermal, As pidiotus much flattened applanate, hypophyllous On second year needles Milesia Hemispherical 53 | 84-137 59-84 |11071 | Subcuticular, intermedia to subspherical immersed, hypophyllous Milesia Hemispherical 393 | 72-180 40-100 |123 71 | Subcuticular, fructuosa to subspherical immerse Milesia Subspherical to 50 | 129-168 « 92-134 |147 x 106} Subepidermal, marginalis almost spherical immerse Milesia Spherical 17 |175-234 x 175-243)}199 x 212} Subepidermal, polypodophila immersed, hypophyllous On leaves 3-8 years old Milesia Somewhat 9 }110-175 «& 105-150) t——— | Subcuticular, Blechni flask-shaped immerse Milesia Hemispherical 16 | 98-168 « 94-168 |129 x 126] Subcuticular, Kriegeriana immersed 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 145 TABLE II, (Continued) No. Size Shape meas- (breadth X height in microns) Species (vertical section) ured Limits Average Remarks II. SUBFAMILY PUCCINIASTREAE. (Continued) Milesia Hemispherical 34 |120-228 x 105-194)177 162} Subcuticular, Polypodii to slightly immersed flask-shaped Milesia Hemispherical 28 |120-228 x 100-188|156 x 140} Subcuticular, Scolopendrii to slightly immersed flask-shaped Milesia Hemispherical 34 [154-241 « 168-214/201 « 195} Subcuticular, vogesiaca to slightly immersed flask-shaped Peridermium Spherical 11 |144-223 x 144-200/170 x 169} Subepidermal, rugosum immersed, hypophyilou On __ second- year leaves Uredinopsis Hemispherical 49 | 58-123 «35-54 | 8944 | Subcuticular, mirabilis immersed, hypophyllous Uredinopsis Hemispherical, 65 | 67-123 x 36-51 | 9648 | Subcuticular, Atkinsonii shallow immersed, hypophyllous Uredinopsis Hemispherical, 72 | 71-134 40-57 {10451 | Subcuticular, Osmundae shallow immersed, hypophyllous Uredinopsis Hemispherical, 68 | 56-125 34-58 | 9348 | Subcuticular, Phegopteridis shallow i hypophyllous redinopsis Hemispherical, 46 | 71-129 « 45-58 94> 50 | Subcuticular, Struthiopteridis shallow immersed, hypophyllous Uredinopsis Hemispherical, | —— |100-159 85-110 | ——— | Subcuticular, Pteridis vertically immersed, elongated hypophyllous On leaves of 1st or 2nd season (?) Aecidium Subspherical — |160-200 x 152-180] ———— | Subepidermal, pseudo- immersed columnare 146 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII TABLE II. (Continued) No. Size Shape meas- (breadth X height in microns) Species (vertical section) ured Limits Average Remarks III. SUBFAMILY CRONARTIEAE Cronartium Flattened 6 {1000-3500 « 34-67 ——— | Caulicolous, Di. hemispherical subcortical IV. SUBFAMILY CHRYSOMYXEAE Chrysomyxa Flask-shaped 29 | 60-128 x 45-125 |105 x 82 | Subepidermal, Ledi immersed V. SUBFAMILY COLEOSPORIEAE Coleosporium Lens-shaped, 12 |323-458 x 92-116 |361 x 103) Subepidermal, Helianthi elevated applanate Coleos porium Lens-shaped, 11 |226-317 73-85 |284 81 | Subepidermal, incons picuum elevated ; hypophyllous Coleos porium Lens-shaped, —- |335-490 x 104-159} ———— | Subepidermal, I pomoeae evated applanate Coleos porium Lens-shaped, — |354- 110-122) ——— | Subepidermal, delicatulum elevated applanate Coleos porium Lens-shaped, — |323-397 x 110-122) ——— | Subepidermal, Laciniariae elevated applanate, epiphyllous Coles porium Lens-shaped, —— |244-366 x 98-104 | —-—— | Subepidermal, Solidaginis elevated pplanate hypophyllous Coleos porium Lens-shaped, — |335-427 x 80-101 | ———— | Subepidermal, Terebinthinaceae | elevated ; epiphyllous Gallowaya Lens-shaped, 7 \244-427 x 48-92 ——— | Subepidermal, pinicola elevated applanate ACKNOWLEDGMENTS Studies recorded in this paper were carried out by the writer over a period of years in the University of Toronto, the Imperial College of London University and Radcliffe College. The year of study at the Imperial College of London University was made possible by a fellow- 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 147 ship granted by the Canadian Federation of University Women. The studies were initiated and finally completed under the direction of Pro- fessor J. H. Faull and, during the intervening periods, supervised by Professor H. S. Jackson and Professor William Brown. Acknowledg- ments are also made to the following for contributions of material — Messrs. J. H. Faull, G. G. Hedgcock, H. P. Bell, G. D. Darker, W. P. Fraser and J. F. Adams. BIBLIOGRAPHY 1. Apams, J. F. (1919). I. Rusts on conifers in peti ons Il. Sex- ual fusions and development of the sexual or in the Peri- dermiums. (Penn. State Coll. Agri. Expt. Sta Bu iL 160: 3-77.) 2. ———— The alternate stage of Pucciniastrum Hydrangeae. (Mycologia, 12: 33-35. FB. (193 o: ALLEN, 3s, (1930). A cytological eet of Fo acamamaeal in Puccinia graminis. (Jour. Agr. Res : 585-614. ) 4, ————- (l Heterothallism in Puce cca (Science, 74: 462-463. ) 5. ——— (1932). A cytological study 2 heterothallism in Puccinia triticina. oo Agr. Res. 44: 733- 6. ——— (1932). A rhe sae - heterothallism in Puccinia oa (Jour. Agr. Res. 45: 513-541. 7. —— (1933). The spermatia of the corn rust, Puccinia Sorghi. (Phytopath. 23: 923-925. 8, ———— 1933). ea cytological — of heterothallism in Puccinia graminis. gr. Res =) 9. ——— (1933). The Spree of flax ne (Phytopath. 23: 487.) 10. Anprus, C. F. (1931). The mechanism of sex in Uromyces appendi- 87. 11. Artuur, J. C. (1904). Taxonomic importance of the spermogonium. Bull. Torr. Bot. Club, 31: 113-123.) 12, ———— ia? North sabes Flora, 7, Part 2. 13. ———— (1910). New speci of Uredineae. (VIII) (Bull. Torr. Bot. Club, ah 569-580 14. ———— (1924). North American Flora, 7, Part 9. 15. ———— (1925). North American Flora, 7, Part 10. 16. ———— (1934). Manual of the rusts in United States and Canada. . —— and D. F. Kern (1906). North American species of Peri- dermium. (Bull. Torr. Bot. Club, 33: 403-438. 18. Beti, H. P. (1924). The rusts of Abies. (Bot. Gaz. 77: 1-31.) 1904 19. BLACKMAN, Va da . On the fertilization, alternation of gen- erations and general cytology of the Uredineae. (Ann. Bot. 18: 323-373.) 20. ————— and H. C. I. Fraser (1906). Further studies on the sex- uality of the Uredineae. (Ann. Bot. 20: 35-48 21. Boyce, J. S. (1928). A conspec ctus of es rusts on balsam firs in North America. ‘{Phytopath. 18: 705-708. ) 22. Cottey, R. H. (1918). Parasitism, eo cisey and cytology of ea ribicola. (Jour. Agr. Re es. 15: 619-660. ) 23. Craicige, J. H. (1927). Experiments on sex in rust fungi. (Nature, 120: 11 6-117.) aN _ Arb. 15: 50-86 . Fraser, W. P. (1912). Cultures of heteroecious rusts. (Mycologia, : 175-193 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII — (1927). Discovery of the function of the pycnia of the rust fungi. ee 120: 76 7a —— (1928). On the occurrence of pycnia and aecia in certain rust fungi. (Phytopath, 18: 1005-1015. ————. (193 An experimental investigation of sex in the rust fungi. CPhytopath 21: 1001-1040. ) ). Union of pycniospores and haploid hyphae in Puc- cinia Tslianthé Schw. (Nature, 131: . Darker, G. D. (192 ultures of Pucciniastrum americanum (Farl.) Arth. and P. arcticum (Lagerh.) Tranz. (Jour. Arnold Arb. 10: 156-167.) . DeBary, A. Bath Die Brandpilze, pp. 24-144. 865). Neue Untersuchungen uber die Uredineen, insbe- sondere dic Entwicklung der Puccinia graminis und den Zusam- menhang derselben mit Aecidium Berberidis. (Monatsber. K. Preuss. Acad. Wiss. 1865: 15-49.) ——— (1867). Uber den Krebs und die Hexenbesen der Weiss- tanne. re Zeit. 25: 257-264. Donce, B. (1925). Organization of the telial sorus in the pine rust, i pinicola. our. Agr. Res. 31: 641-651. . Fartow, W. G. (1885). Notes on some species of Gymnosporangium and Chrysomyxa of the United States. (Proc. Am. Acad. 20: 322 . Faut., J. H. (1929). The morphology, biology and phylogeny of the Pucciniastreae. [Proc. Internat. Congr. Plant Sci. Ithaca (1926), 2: A735 1745,] omnes (1992 . Taxonomy and geographical distribution of the genus thi (Contrib. Arnold Arb. 2: 1-1 ). The biology of Milesian rusts, (Jour. Arnold 1913), Further cultures of heteroecious rusts. (Mycolo- 39, . Notes on Uredinopsis mirabilis and other rusts. ich ta! 6: 25-28 . From YT t3 912). Sexual fusions and spore development in the ee rust. (Bull. Torr. Bot. Club, 39: 113-131. . GuINieER, P. (1931). Note sur deux Pucciniastrum nuisibles aux oe es coniféres. (Travaux Cryptog. dédiés a L. Mangin, Museum i 73-375. . Hepccock, G. G. ( ). Key to the known aecial forms of Col sporium occurring in the United States and a list of the host Rey ee 20: 97-100. N. Rex Hunt (1922). Notes on some species of Coleo- 257. ccna Tt (Mycologia, 14: . Herspercer, C. (1928). Uber rite Vorkommen einer Aecidienperidie bei ion ae rae (Mitt. Nat. Ges. Bern. 192 . Hunter, L. M. (1927). Comparative study of spermogonia of rusts 23; of Abies. (Bot. Gaz. 83: 1- (1935). A preliminary note on life history studies of Euro- pean tgs of ee (Jour. Arnold Arb. 16: 143 . JacKxson, H. (1931). The rusts of South America based on the Holway ieee III. (Mycologia, 23: 96-117.) 1). The rusts of South America based on the Holway collections IV. (Mycologia, 23: 332-364. 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 149 49. enkr. 26: 257-277. KaMEI, S. (1930). On the life history of Uredinopsis Pteridis with a nary bearing on its peridermial stage. (Ann. Phytopath. Soc. Jap. 2: 1-22. (1932). Notes on the cultural aged of a new species of His 34. Hyalopsora. (Trans. Sapporo Nat. . Soc. 12: 27-34 51. KLeBaun, H. (1899 Pla eeecnche a “heterécischen Rostpilzen. (1899). (Zeitsch. Pflanzenkr. 9: Kulturversuche mit Rostpilzen. (Zeitsch. Pflan- Do. eae E. (1928). Differential staining of Peronosporaceae. (Phy- opath 18: 869-872. 54. Lupwia, C. A. (1915). Notes on some North American rusts with Caeoma-like sori. (Phytopath. 5: 273-284.) 55. Maire, R. (1900). L’évolution nucléaire chez les Endophyllum. (Jour. de Bot. 14: 80-92. 56. Mayor, E. (1919-1921). Etude peat cary de Melampsora Abieti- Capraearum Tubeuf. (Bull. S ycol. 36: —2 57. ————— Etude a a d’Urédinées hétéroiques. (Bull. Soc. Neuchat. Sci. Nat. 47: 67-73. ) 58. 193 Notes mycologiques VIII. (Bull. Soc. Neuchat. Sci. Nat. 58: 23-26. 59. Meryen, F. J. F. (1841). Pflanzen-Pathologie, pp. 141-150 60. NeuMAN, R. (1894). Ueber die Entwickelungsgeschichte der Aeci- dien und er ee der Uredineen. Ses wigia, 33: 346-361.) 61. Peck, C. (187 N. Y. State Mus. 2 4. 62. Pierson, R. K. (1933), Fusion of pyc niospores with filamentous hyphae in the : rhite Ate blister rust. (Nature, 131: 728-729. 63. PLowRIGHT, C. 891). Ejinige ficients mit Rostpilzen. (Zeitsch. oe 1: 130-131. 64. Ruoaps, A. S., G. G. Hepccock, E. BerHet and C. Hartrey (1918). Host relationships of the North American rusts, other than Gymno- sporangiums, which attack conifers. (Phyt topath. 8: 309-352. 65. Rosen, F. (1893). eitrage zur Kenntniss der anzenzellen. hn’s Beitrage zur Biologie der Pflanzen, 6: 237-266. 66. Sappin-Trourry, M. (1896). Recherches histologiques sur la famille des Urédinées. (Le Botaniste, 5: 59-244. 67. Scumitz, H. (1916). Preliminary note on the occurrence of Peri- dermium balsameum in Washin ngton. (Phytopath. 6: 369-371.) 68. SpAuLpING, P. (1911). Rusts of Tsuga canadensis. (Science, 33: Ox 69, ————— (1911). The rusts of Tsuga canadensis. (Phytopath. 1: 94-96.) 70. Sypow, P. a H. (1915). Monographia Uredinearum, ITT. 71. THUMEN, (1880). Die Blasenrost-Pilze der Coniferen. (Mitteil. Forstl. ee Oesterreichs, II. 320.) 72. TuLasne, L. R. (1854). Second mémoire sur les aie et les Ustilaginées. (Ann. Sci. Nat. Bot. Sér. 4. 2: 77-196.) 73. Uncer, F. (1833). Die Exantheme der Pflanzen ee einige mit diesen verwandten Krankheiten ee Gewachse, pp. 297-301.) 74. Weir, J. R. and E. E. Huppert iy 7). Observations on forest tree rusts. (Amer. Jour. Bot. 4: Poi 918). Notes | on forest tree rusts. (Phytopath. 18.) 8: 114 . Winter, G. (1884). Repertorium Rabenhorstii fungi europaei et X I extraeuropaei. Cent. XXXI et XXXII cura Dr. G. Winter. (Hedwigia, 23: 169.) 150 bo Ww aN wm n N ioe) ‘oO _ oO — — —" bo — Ww —" is JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII EXPLANATION OF PLATES PLATE 182 . Lateral vertical section of mature spermogonium of Melampsora f Abieti-Capraearum from longitudinal section of leaf of Abres a he 700. dian vertical section of an immature spermogonium of oes Abieti-Capraearum located under guard cell in longitudinal section of leaf of A. balsamea. X 480. Note the es. Median vertical section of an genres spermogonium of Melampsora roid Mage Andee approximately mature size, from transverse section of leaf of ris Soiaies. x 2 Dark areas wits spermogonium are remains of mesophyll cells. Median vertical section of the mature spermogonium of Caeoma Faulliana from a transverse section of leaf of Abies lasiocarpa. Median vertical section of mature spermogonium of Melampso- af ridium betulinum from transverse section of leaf of Larix decidua, X Me dian vertical section of mature spermogonium of Melamp- ee Abietis-canadensis from cone scale of Tsuga canadensis. Median vertical section of mature spermogonium of — sora Bigelowii from longitudinal section of leaf of Lartx la ri- cima. X 280, Note mass of iubdigr above spermogonium. mog x IS. Median vertical section of mature s sen of ages en Larici-Capraearum from transverse section of leaf of Lari decidua. X 280. PLATE 183 . Median Wokeg section of a mature Ss of Melamp- sorella Cerastii from longitudinal section of lea A. balsamea. x 56 i spora Goeppertiana from transverse section of leaf of A. bal- samea _ Median vertical section of mature spermogonium of Puccinias- trum arcticum from transverse section of leaf of Picea glauca. x 2 _ Median vertical section of mature spermogonium of haga trum arcticum from transverse section of leaf of Picea glauca. x 625. Note cells in stroma from which eer cern branch. Median vertical section of spermogonium of Thecopsora Myr- tilli from transverse section of leaf of Tsuga pees ec x 280. ae — of leaf of A. balsamea infected with Hyal- opsora Aspidiotus. Note abundance of myce ium in intercellu- lar spaces, ictal and spermatia strewn over leaf surface. x 90, 1936] HUNTER, SPERMOGONIA OF THE MELAMPSORACEAE 151 PLATE 184 Fig. 16. Median vertical section of spermogonium of Hyalopsora Aspidi- otus at very early stage in development, from longitudinal sec- tion of leaf of A. balsamea. 290. Fig. 17. Median vertical section of immature spermogonium of Muilesia intermedia from transverse section of leaf of A. balsamea. x 288. Note depressed epidermal cells, spermatiophores and spermatia and overlying unruptured cuticle, with intermediate layer, darkly sta tained, beneat M nv ction of immature spermogonium of Muilesia intermedia age ra ee section oO a A, balsamea. xX 888. e depressed socal cells, separated spermatio- phores, eeensts cuticle minus intermediate layer in centre above spermogoniun Fig. 19. Median vertical ee of very early stage in the development f the spermogonium of Muilesia marginalis from transverse section of leaf of A. balsamea. X 550. Note central cell in nhl has been inv a . _ ee) o Ee jet) et Lay Ce} SES wn Fig. Fig. 20. Median vertical section early stage in the development of the spermogonium of Milesia marginalis from longitudinal sec- tion of A. balsamea. X 666. Note depressed mesophyll cells, SE caaatoahe and central ea cell invaded by enlarged hyphae. PLATE 185 Fig. 21. Median vertical section of young spermogonium —a later stage than that shown in Fig. 23 — of Milesia marginalis from trans- verse section a net of A. balsamea. X 374. Note epidermal cells have dis red. Median eal: peal of spermogonium of Milesia marginalis, ee cells over left, from transverse section of leaf of A x 280. = 79 I) nN Fig. 23. Median naps section of mature spermogonium of Miles marginalis from transverse section of leaf A. balsamea. x ae teeth from outer ey wall projecting into spermogoni . Median Seri section of mature eee ane of Miles polypodophila from transverse vite ion eaf of A. Walden. m2 Note two “‘flexuous hypha Pity 2o; Median vertical eae, of an iit spermogonium of Mile- sia Polypodii, showing an early stage of development, from transverse section of leaf of Abies concolor. X 280. Note a a bo aS Fig. 26. Median vertical section of mature spermogonium of Muilesia polypodophila from transverse section of leaf of A. balsamea. Note shortened spermatiophores and chains of sperma- tia. Fig. 27. Milesia Scolopendrit, “flexuous hypha” with sh ie attached. x 1000. This is a higher magnification of Fig PLATE 186 Figs. 28 and 29. Reiger of median vertical section of mature sper- different focus. Note flexuous hyphae with spermatia attached. Fig. WwW WW BR ies) Ww ies) Q Ww WwW oO Ow & S a paar JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII . Median vertical section of mature spermogonium of Muilesia Blechni from transverse section of leaf of Abies alba. X 300. m A. concolor. X 372. ge short Sade ipo and central core of long e spermogonium of “Milesia Polypodiu from transverse No hae . Ma wer of leaf of A. alba. X 280. te “flexuous hyp in central cavity PLATE 187 . Median vertical section of mature spermogonium of Muilesia x 300. vogesiaca, from transverse section of leaf of A. alba. . Median vertical section of immature spermogonium o of Uredi nopsis mirabilis, from transverse section of leaf of A. balsamea. x _ Median vertical section of mature spermogonium of Uredinop- e sts Atkinsonii, from transverse section of leaf of A. balsamea to show branching of fh adil ag seat from stroma. 420). . Median vertical section of mature spermogonium of Chryso- myxa Ledi, from jens anc fe of leaf of Picea mariana. x 350 PLATE 188 . Median — section of mature spermogonium of Cronartium ribicola from stem of Pinus sles x 170. . Median vertical section of mature spermogonium of Coleospo- rium Helianthi, from npoaatmesiy section of leaf of Pinus vir- giniana. = _ Median vertical section of immature spermogonium of Coleo- Sportum inconspicuum, from longitudinal section of leaf of P. virginiana. 170. . Median vertical section of spermogonium of Gallowaya pinicola f 142. from transverse section of leaf of P. virginiana. LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, HARVARD i aseeaaie: Jour. Arnotp Ars. VoL. XVII PLATE 182 SPERMOGONIA OF THE MELAMPSORACEAE FULL-TONE— MERIDEN Jour. ARNoLD Arb. VoL. XVII j g ar Ky | oe ee 4 m ye 2 ‘ pF EIR ge east tag *£p.~ ae Pitty so “a a , < Oi ¥ era): ie, oy SPERMOGONIA OF THE Mert AMT PCOPA FULL-TONE— MERIDEN PLATE 183 Jour. ARNoLtp Ars. Vo. XVII PuaTE 184 BS ow! ¢ <4 if ; Ne mais EP "ram Pa \ °E Saat 0 ° : MNES = PPS 1G eee Ab) a DN We He Nik \ SH 3 ; SINUS ie Zoe Or] aad Vf ey alee Cn ONS g Pe at (77ic ( ee a AH a Toys sh pee Sega): cost Wy ee Os SPERMOGONIA OF THE MELAMPSORACEAE FULL-TONE— MERIDEN Jour. ARNOLD Ars. VoL. XVII PLATE 185 SPERMOGONIA OF THE MELAMPSORACEAE FULL-TONE— MERIDEN Jour. ARNoLD Ars. VoL. XVII PLATE 186 4 A 4 fe = pee, &, ap ’ ety arn 1 Mog Bh’ A is wp) SPERMOGONIA OF THE MELAMPSORACEAE FULL-TONE — MERIDEN Jour. ARNotp Ars. VoL. XVII PLATE 187 “> Rae ni ~ ay fs» ns pe SPERMOGONIA OF > THE MELAMPSORACEAE FULL-TONE—— MERIDEN Jour. ARNotpD Ars. Vor. XVII PLATE 188 SPERMOGONIA OF THE MELAMPSORACEAE FULL-TONE—— MERIDEN 1936] SAX, EXPERIMENTAL PRODUCTION OF POLYPLOIDY 153 THE EXPERIMENTAL PRODUCTION OF POLYPLOIDY KarRL Sax With plate 189 PoLtyPLorpy has played an important role in the evolution of the angiosperms, and it is estimated that about half of the angiosperm species are polyploids. Two types of polyploids are known: autopoly- ploids, involving the reduplication of the same genom; and allopoly- ploids, involving the duplication of different genoms. The first type of polyploidy is not directly effective in speciation, although the poly- ploid differs from the diploid in certain morphological and physiological characters. Autopolyploidy does, however, serve to isolate the new form, since it is sterile with, and has a greater range than the diploid ancestor. Ina recent survey of autopolyploidy Mintzing (1936) shows that such polyploids are generally more vigorous and hardier than the diploid forms and that the chromosome races are ecologically different. There is also a strong tendency for the polyploids to be perennials. Allopolyploidy is directly effective in producing new species, since such polyploids result from species hybridization, followed by chromosome doubling. The duplication of each of the parental genoms restores fertility, and the hybrid breeds true. Thus the allopolyploid is a con- stant species hybrid which has the characteristics of a true species, and in the case of generic crosses the allopolyploid may merit generic rank. Allopolyploids are known to occur in nature and in cultivation, and have been produced experimentally in several different families. The various types of allopolyploids have been described by Winge in 1932; and since that time a number of new cases have been discovered, so that we now have records of about fifty allopolyploids, of which eight are known to have originated in nature. It is significant that many of our most important crop plants are of known or supposed allopolyploid origin. The general occurrence of polyploidy in the angiosperms has led to a detailed study of the causes of chromosome doubling in gametes and egg cells. In 1923 De Mol found that diploid pollen grains could be induced in Hyacinthus by subjecting the bulbs to abnormal environ- mental conditions. Belling (1925) found that low temperatures caused the production of diploid pollen grains in Stizolobium, Uvularia, and 154 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Datura; and similar results were found in Epilobium and Oenothera by Michaelis (1928). The experimental work of Sakamura and Stow (1926) showed that high temperatures are also effective in producing diploid pollen grains. Shimotomai (1927) was able to produce giant pollen grains in Liriope and in Scilla by cold treatment. The effect of temperature treatment is produced by inhibiting the normal meiotic division so that no reduction of chromosomes occurs; but the factors involved are not well known. The production of diploid gametes is undoubtedly an important fac- tor in the formation of polyploid races or species in nature, even though sterile triploids would be formed more frequently than the relatively fertile tetraploids, especially in cross-pollinated species. The origin of polyploidy by the union of diploid gametes is dependent upon the fre- quent occurrence of chromosome doubling; but chromosome doubling in the fertilized egg, or at a later stage in somatic development, is directly effective in producing fertile polyploids. This mechanism is known to occur naturally, and has been used under experimental con- trol. Randolph (1932) has been able to double the chromosome num- ber in the fertilized egg cell of Zea by subjecting the plant to a tempera- ture of 43°C. for about half an hour at the time the fertilized egg cell was presumably in the resting or early prophase stage. This technique has also been successful with wheat and rye (Dorsey 1936). EXPERIMENTAL PRODUCTION OF POLYPLOID POLLEN GRAINS During the past few years we have found occasional diploid pollen grains in plants of Rhoeo discolor grown in the greenhouse, where they were subject to moderate variations in temperature. Plants kept in the cold pit, at a temperature of about 5°C. for several days and then returned to the greenhouse, produced many diploid pollen grains. Last year we installed constant temperature chambers where the plants could be kept under control and studied more carefully. We found that plants of Rhoeo discolor would produce some diploid pollen grains when subjected to either high or low temperatures for several days; but more consistent results were obtained when the plants were kept at a temperature of about 10°C. for two or three days and then kept at a temperature of 36°C. for one day. After treatment the plants were returned to the greenhouse and examined every few days. Under normal conditions the chromosome pair at meiosis to form a ring or chain of 12 chromosomes. Adjacent chromosomes often pass to opposite poles so that each daughter cell receives six chromosomes. 1936] SAX, EXPERIMENTAL PRODUCTION OF POLYPLOIDY 155 These divide equationally to produce microspores with six chromo- somes (Fig. 2). There is usually some irregularity in the chromosome distribution at meiosis, so that about 80 per cent of the pollen grains are sterile and undeveloped. The viable pollen grains contain an elongated dark-staining generative nucleus and a degenerate tube nucleus (Fig. 2A). The first effect of the temperature treatment is the production of asynapsis. At metaphase the chromosomes are unpaired, and the twelve univalents lie on an equatorial plate (Fig. 1). The equational division of the univalent chromosomes produces a dyad with twelve chromosomes in each cell. In most cases these dyads develop inde- pendently to produce diploid microspores (Fig. 3) which develop into diploid pollen grains (3A). Since each microspore contains all twelve chromosomes, there is no deficiency of chromosomes or chromosome seg- ments, and the microspores produce normal pollen grains. There is very little sterility of diploid pollen grains, and certain anthers show almost perfect pollen development as contrasted with about 80 per cent sterility in the normal diploid. The temperature variation also produces tetraploid microspores. Meiosis is suppressed, and the twelve univalents divide, but there is no nuclear division. The chromosomes pass into the resting stage, and the entire microsporocyte develops as the microspore. There are twenty- four chromosomes in the microspore which divide to form the two nuclei of the pollen grain (Fig. 4). The differentiation of the generative and tube nucleus is seldom complete, although a few giant pollen grains are produced which appear to be normal (Fig. 4A). These tetraploid pollen grains are not produced regularly, and a single anther never contains only this class of pollen grains. They are found in small numbers asso- ciated with haploid pollen grains. In rare cases the univalent chromosomes divide, and the daughter nuclei are formed, but no cell wall is formed to produce a dyad. The microsporocyte functions as a microspore with two nuclei, each with twelve chromosomes. In microspore development each nucleus divides (Fig. 5), and a pollen grain is produced which has two generative and two tube nuclei (Fig. 5A). Diploid pollen grains have also been produced in species of Trade- scantia, but only when the plants have been subjected to low and high temperatures. Cold or heat treatment alone is rarely effective in pro- ducing polyploid gametes in this genus. The first indication of reaction to temperature treatment is partial asynapsis; further reaction causes complete asynapsis with irregular division of chromosomes; and with 156 JOURNAL OF THE ARNOLD ARBORETUM (VOL. xvi extreme treatment, the univalents divide equationally to form diploid pollen grains. The effect of temperature treatment appears to be caused by dis- turbing the synchronization of nuclear and cytoplasmic activities. This effect is especially clear in the development and differentiation of the generative and tube nucleus in microspore development in Tradescantia (Sax 1935). In the microsporocyte the temperature change prevents chromosome pairing. The development of the chromosomes is accel- erated in relation to other nuclear and cytoplasmic activities so that the univalents are ready to divide at the first instead of the second meiotic division. They may divide before nuclear division is possible so that the chromosome number is doubled. This is the mechanism involved in the production of tetraploid pollen grains of Rhoeo and the tetraploid egg cells of Zea and Triticum. THE ORIGIN OF POLYPLOIDY IN NATURE The occurrence of polyploidy in nature has been attributed to extreme temperatures. The tetraploids may have a more northern range than the diploids (Tischler 1935, Anderson and Sax 1936), or a more alpine distribution (Manton 1934), or may show a greater distribution along the seacoast (Tischler 1935). In such habitats a species would be sub- jected to greater temperature variations. The extreme temperature changes may cause polyploidy as the diploid races extend into unfavor- able territory, but it is also possible that the polyploids have originated before the extension of the range, and because they are hardier, they extend their range into more extreme environments. There are a number of cases where the forms in the northern areas or in extreme environments are more frequently diploid than the related varieties or species in more favorable environments (Bocher, 1936; Rohweder, 1936). One might expect that any extreme change in habitat might induce polyploidy, and that polyploidy would be expected at the periphery of the range, either northern or southern, or at low or high elevations. There is some evidence that different genera vary greatly in response to temperature changes. In many cases there is a definite cycle in gamete development which may serve as an adaptation to ex- treme diurnal temperatures. In the cereals, for example, the meiotic divisions tend to occur between 6 A. M. and 8 A. M., and it is possible that the more critical prophase stages are synchronized so that they are not often subjected to great variations in temperature. The experimental evidence obtained with Rhoeo and Tradescantia indicates that polyploidy may be induced by rather moderate tempera- 1936] SAX, EXPERIMENTAL PRODUCTION OF POLYPLOIDY 157 ture changes in certain genera, but that rather extreme and sudden changes are more effective. Neither constant heat nor constant cold seems to be effective. Since extreme variations in temperature are most likely to cause polyploidy, one might expect polyploidy in nature to occur in regions of high altitude, in semi-arid regions, or along the sea- coast, where the changes in temperature are often rapid and extreme. The polyploid economic plants appear to have originated in such regions. Of the seven centers of distribution described by Vavilov (1932), five are regions of high elevation. Most of the economic plants of polyploid origin—including apples, pears, wheat, oats, potatoes, tobacco, and upland cotton—have originated in the mountains or foothills in the subtropics or tropics. The increased vigor and hardiness of the polyploids are known to be associated with an increase in range of distribution in certain genera. We should expect that polyploids might be more frequent than diploids at the periphery of the range. Many species of plants in North America originated in Asia. A comparison of Asiatic and American species grown in the Arnold Arboretum shows that polyploidy is more frequent in the American species of Malus, Rosa, Acer, Staphylea, and Ulmus, while the reverse is true for Lonicera and Fraxinus. Many genera, such as Rhododendron, are represented entirely, or almost completely, by diploid species in both the old and new world. The data are hardly adequate for a critical survey, but the significance of polyploidy un- doubtedly varies in different genera. EXPERIMENTAL POLYPLOIDY Work on experimental polyploidy is being conducted at a number of research laboratories. Belling (1925) has called attention to the value of triploids in ornamental plants where seed production is not desired. The use of diploid gametes, formed without a reduction division, should be of value in breeding heterozygous horticultural varieties. In the apple, for example, the varieties are so heterozygous that breeding new varieties is largely a matter of chance; but the union of diploid gametes should combine all the characters of the parental trees. In the production of auto- and allopolyploids Randolph’s method of doubling the chromosome number in the fertilized egg seems to be the most practicable. Our experimental work indicates, however, that it is advisable to precondition the plant in a cold chamber before subjecting the fertilized egg to high temperatures. I have been told that the heat treatment alone is not effective in doubling the chromosomes in Zea in the southern states. The failure of chromosome doubling in southern 158 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII regions presumably is caused by the relatively high temperatures of the normal environment, so that there is insufficient range of temperature to cause chromosome doubling. Under such conditions the cold treatment is necessary, and probably most plants would show a better response when subjected to an extreme temperature range. Either the transition from cold to heat or from heat to cold seems to be effective in causing polyploidy. In experimental work the cell activities can be timed somewhat better if the plant is first subjected to cold treatment. SUMMARY Plants of Rhoeo discolor subjected to sudden temperature changes produced many diploid pollen grains, and a few tetraploid pollen grains. Chromosome doubling is caused by the suppression of meiosis and an equational division of the chromosomes to produce dyads which develop into diploid pollen grains. Occasionally the chromosomes divide with- out nuclear division to produce a monad which develops into a tetraploid pollen grain. Double diploid pollen grains may be produced which have two generative and two tube nuclei. Chromosome doubling appears to be caused by disturbing the synchronization of nuclear and cyto- plasmic activities. The effect of temperature variations in causing polyploidy in nature is discussed in relation to plant distribution and the centers of origin of cultivated plants. In the experimental production of polyploidy extreme temperature changes are more likely to be successful. The temperature chambers used in this work were paid for, in part, by a grant from the American Academy of Arts and Sciences. LITERATURE CITED ANDERSON, E., and Kart Sax (1936). A cytological monograph of the American species of Tradescantia. (Bot. Gaz. 97: 433-476. BeLuinG, J. (1925). Production of triploid and tetraploid plants. (Jour. Hered. 16: BocuHer, T. W. (1936). Cytological studies on Campanula rotundifolia. ( 69-277. Dorsey, E. (1936). Induced polyploidy in wheat and rye. (Jour. Hered. 27: 155-160. ) Manton, I. (1934). The problem of Biscutella laevigata. (Zeitschr. Indukt. Abst. benas 67: 41-57. ) Micwagtis, P. (1928). Ueber die experimentelle Erzeugung hetero- ploider Pflanzen bei Epilobium und Oenothera. (Biol. Zentralbl. 48: 370-374. Mot, W. E. E (1923). Duplication of generative nuclei by oe of aeeloei cal stimuli and its significance. (Genetica, 5: 225-272.) Jour. ARNotp Ars. Vor. XVII PLaTE 189 Ba EXPERIMENTAL PRODUCTION OF POLYPLOIDY FULL-TONE™™ MERIDEN OO a 1936] SAX, EXPERIMENTAL PRODUCTION OF POLYPLOIDY 159 Mutntzine, A. (1936). The evolutionary significance of autopolyploidy. (Hereditas, 21: 263-378. RANnboLPH, L. F. (1 932). Rect oa of high temperature on ee and other variations in maize. (Proc. Nat. Acad. Sci. 18: 2-229.) aa a acta Die Peas der Po lyploidie fiir die eee der men an die Kalkgebiete Schleswig-Holsteins. (Beibl. Bot. ee 54: 507-519. SAKAMURA, T., und I. Stow (1926). Ueber die experimentell veranlasste Entstehung von keimfahigen Pollenkornern mit abweichenden Chromo- ee tse (Jap. Jour. Bot. 7: 111-137.) Sax, K. (1935). The effect of temperature on nuclear differentiation in microspore development. (Jour. Arnold Arb. 16: 301-310.) SHimotomal, N. (1927). Ueber sa ee der meiotischen Teilungen durch niedrige Pa att (Bot. Mag. Tokyo, 41: 149-160. TiscHLER, G. (1935). Die Bedeutung der Polyploidie ftir die Verbreitung r An Ca asi (Bot. Jahrb. 67: 1-36. Neaveioe. N. I. (1932). The ie ‘of evolution in cultivated plants. (Proc. Int. cae eatee 1: 331-342. MN O. (1932). On the origin of constant species-hybrids. (Svensk ot. Tidskr. 26: 107-122.) DESCRIPTION OF PLATE 189 Photographs oh microspores and pollen grains of Rhoeo discolor. Aceto- carmine preparations. x 900. Fig. 1. Pollen ee cell with induced asynapsis. The twelve uni- valents divide be peer Figs. 2 an : normal haploid ane with 6 chromosomes, and the normal haploid pollen grai Figs. 3 and 3A. A diploid RIGROSROLE: ‘produced by temperature treat- ment, and a heey diploid pollen grain. Figs. 4 and aploid microspore induced by alo aia treat- ment, and a wall ‘differentiated tetraploid pollen g ; Figs. 5 and 5A. A double diploid microspore rodneee ed equational ell division of the 12 univalents and nuclear division without ce wall formation. ae resulting pollen grain has two generative and two tube nuc CYTOLOGICAL LABORATORY, ARNOLD ARBORETUM, HarvVARD UNIVERSITY. 160 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xv STUDIES IN THE OAKS OF THE MOUNTAINS OF NORTHEASTERN MEXICO CorNELIUS H. MUELLER In Two collections of plants made for the Arnold Arboretum in the mountains of Nuevo Leon, Mexico, by the author and Mary Taylor Mueller in 1934 and by the author in 1935 there are 250 numbers of oaks. A field study of considerable intensity from about May 6 to August 15, 1934 and from June 22 to August 15, 1935 has been supple- mented by an herbarium study of the specimens collected. The collecting and field study were carried on in the Sierra Madre Oriental south of Monterrey, Nuevo Leon. The localities cited in the treatment of the species are based upon names appearing on the topo- graphic maps of the republic published by the Secretaria de Agricultura y Fomento, Departmento de Geografia, Meteorologia e Hidrologia, Tacubaya, D. F., Mexico. The maps are known as “Carta de la Repub- lica Mexicana en hojas a la 100,0002.” Many of the names of ranches and cafions are local names not appearing on any maps. The term municipio (Mun.) refers to a political unit comparable to a county. A municipio takes its name from the town in which the seat of government is located. In many of the species there was noted a high degree of variability which undoubtedly has in part led to the description of numerous poorly distinguished species. It will be noted that there are proposed many new names for forms not identical with, but surely conspecific with, previously known entities. These are given the rank of forma merely to call attention simultaneously to their differences from, and their con- specificity with, the typical forms with which they are allied, in the hope of avoiding their being regarded as distinct species, as they might readily be if one had not the intergrading forms to study and could not see the trees in the field. The types of the new species and forms are deposited in the herbarium of the Arnold Arboretum (cited A. A.), Harvard University. The author wishes to acknowledge the excellent codperation and financial aid of the Arboretum which made possible the collection of these plants. 1936] MUELLER, STUDIES IN THE OAKS OF MEXICO 161 Subgenus LEUCOBALANUS Engelmann Series PoLyMorRPHAE Trelease Quercus polymorpha Chamisso & Schlechtendal in Linnaea, 5: 78 (1830). Nuevo Leon: Mun. de Monterrey, Cafion Diente, alt. 700 to 1000 m., common on the wooded cafion floor and adjacent slopes, Vo. 252, May 6, 1934; Mun. de Galeana, cafion below Alamar on Hacienda Pab- lillo, common on the densely wooded cafion floor, alt. 1900 m., Nos. 608 & 609, May 29, 1934, No. 632, May 30, 1934, No. 1143, July 21, 1934. Quercus polymorpha f. angustifolia, f. nov. A typo recedit foliis anguste lanceolatis (3-4 & 10-14 cm.) integris vel versus apicem angustatum leviter dentatis, acumine plerumque obtuso. Nuevo Leon: Mun. de Galeana, cafions about Alamar on Hacienda Pablillo, alt. 2000 m., No. 610, May 29, 1934, No. 1144 (type in A. A.), July 21, 1934; Mun. de Villa Santiago, arroyo bottom between Las Adjuntas and Potrero Redondo, alt. 1300 m., No. 2076, July 3, 1935. Series GLAUCOIDEAE Trelease Quercus clivicola Trelease & Mueller in Bull. Torr. Bot. Club, 63: 149 (1936). Nuevo Leon: Mun. de Galeana, Cafion San Francisco on Hacienda Pablillo, abundant in the pine and oak wood of the cafion floor and slopes, alt. 2000 to 2200 m., No. 286, May 10, 1934, Nos. 305, 306, & 309, May 11, 1934, No. 452, May 18, 1934, Nos. 1280, 1285, 1285A, 1286, & 1292, July 31, 1934; common in the oak wood about Taray on Hacienda Pablillo, alt. 2200 m., No. 423, May 17, 1934, No. 947, July 2, 1934; in Cafon Santa Barbara on Hacienda Pablillo, alt. 2000 m., No. 584, May 25, 1934, Nos. 1097 & 1098, July 19, 1934: common about Alamar on Hacienda Pablillo, alt. 2000 m., No. 974, July 3, 1934, Nos. 1103 & 1105, July 20, 1934; Puerto de los Pastores near Galeana, alt. 1500 m., common on openly wooded slopes, No. 1007A, July 5, 1934, No. 1295, August 2, 1934; Mun. de Monterrey, Cafion Diente, alt. 900 m., common on the wooded canon floor and slopes, Vos. 1306 & 1306B, August 8, 1934; Mun. de Villa Santiago, Cafion Guajuco above Villa Santiago, alt. 1300 m., common on openly wooded slopes, No. 1331, August 12, 1934, No. 1356, August 15, 1934. Quercus clivicola f. consanguinea C. H. Mueller in Bull. Torr. Bot. Club, 63: 150 (1936). Nuevo Leon: Mun. de Galeana, Cafion Santa Barbara on Hacienda 162 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XVII Pablillo, alt. 2000 m., common in the pine and oak wood of the canon walls, No. 583, May 25, 1934, No. 1096, July 19, 1934; about Alamar on Hacienda Pablillo, alt. 2000 m., scattered in the oak wood of rela- tively open slopes, No. 1104, July 20, 1934; Cafion San Francisco on Hacienda Pablillo, alt. 2100 m., abundant in the pine and oak wood of gentle slopes, Vo. 1285B, July 31, 1934. Quercus clivicola f. crenifolia Trelease & Mueller in Bull. Torr. Bot. Club, 63: 150 (1936). Nuevo Leon: Mun. de Galeana, Cafion San Francisco on Hacienda Pablillo, alt. 2100 m., common in open oak wood, No. 388, May 14, 1934: about Taray on Hacienda Pablillo, abundant on openly wooded, gentle slopes, alt. 2200 m., Nos. 421, 422, & 425, May 17, 1934; near Alamar on Hacienda Pablillo, abundant in open oak wood on gentle slopes, alt. about 2000 m., No. 614, May 29, 1934, No. 701, June 2, 1934, No. 1145, July 21, 1934, Nos. 1173 & 1176, July 22, 1934; Mun. de Villa Santiago, Potrero Redondo, alt. 1300 m., sparse in the luxuriant pine wood, No. 2128, July 5, 1935. Quercus clivicola f. dentata Trelease & Mueller in Bull. Torr. Bot. Club, 63: 150 (1936). Nuevo Leon: Mun. de Galeana, Cafion San Francisco on Hacienda Pablillo, alt. 2100 m., abundant constituent of the pine-oak forest, No. 322, May 12, 1934, Nos. 1284 & 1284A, July 31, 1934; Canon Santa Barbara on Hacienda Pablillo, alt. 2200 m., common in the pine-oak wood, No. 1093, July 19, 1934; Puerto de los Pastores near Galeana, alt. 1500 m., scattered on dry, openly wooded slopes, No. 996, July 5, 1934: Mun. de Linares, Cafion Santa Rosa near Iturbide. alt. about 1200 m., scattered in wooded arroyos, No. 1294, August 2, 1934. Quercus clivicola f. elongata, f. nov. A formis aliis differt foliis magis elongatis (2-3 > 10-12 cm.) plerisque supra crenatis vel dentatis. Nuevo Leon: Mun. de Villa Santiago, Cafion Guajuco above Villa Santiago, alt. 1300 m., common in the low oak wood, No. 1323 (type in A. A.), August 10, 1934; Mun. de Galeana, Taray on Hacienda Pablillo, alt. 2200 m., No. 582, May 25, 1934, No. 1095, July 19, 1934. This form is close to both f. crenifolia and f. dentata in that the leaves are usually crenate or dentate above. Quercus glaucophylla von Seemen in Bot. Jahrb, 29: 95 (1900). Nuevo Leon: Mun. de Galeana, Taray on Hacienda Pablillo, alt. 2200 m., sparse on the openly wooded slopes, No. 590, May 29, 1934; near Alamar on Hacienda Pablillo, alt. 2000 m., common on sparsely 1936] MUELLER, STUDIES IN THE OAKS OF MEXICO 163 wooded, dry slopes, No. 596, May 29, 1934, No. 1121, July 21, 1934; Puerto de los Pastores near Galeana, alt. 1500 m., scattered on sparsely wooded, dry slopes, No. 1002, July 5, 1934. Quercus glaucophylla f. lobata, f. nov. A formis aliis differt foliis typicis e basi ad apicem profunde lobatis, sinubus plerisque litteram U simulantibus. Nuevo Leon: Mun. de Galeana, Cafion San Francisco on Hacienda Pablillo, alt. 2100 m., scattered on openly wooded limestone bluffs, Vo. 1281 (type in A. A.), July 31, 1934; Puerto de los Pastores near Gale- ana, small tree sparse on openly wooded, gentle slopes, No. 1003, July 5, 1934. Quercus glaucophylla f. longifolia, f. nov. Ab aliis formis differt foliis elongatis (1.5-3 > 5-7.5 cm.) supra tantum leviter rotundo-lobatis. Nuevo Leon: Mun. de Galeana, Taray on Hacienda Pablillo, alt. 2200 m., small tree, sparse in open oak wood on slopes, No. 589 (type in A. A.), May 29, 1934. Quercus glaucophylla f. macropetiolata C. H. Mueller in Bull. Torr. Bot. Club, 63: 151 (1936). In some cases this form has much longer petioles than those first described for it from Monterrey, some of them being as long as 16 mm. Nuevo Leon: Mun. de Galeana, Canon San Francisco on Hacienda Pablillo, alt. 2100 m., common on openly wooded slopes, No. 307, May 11, 1934; Alamar on Hacienda Pablillo, alt. 2000 m., common on openly wooded, dry slopes, No. 1101, July 20, 1934, No. 1200, July 23, 1934; Puerto de los Pastores near Galeana, alt. 1500 m., scattered on dry, openly wooded slopes, No. 1001, July 5, 1934; Mun. de Villa Santiago, Potrero Redondo, alt. 1300 m., common on steep, openly wooded slopes, No. 2087, July 4, 1935. Quercus glaucophylla var. subrotundifolia, f. nov. A formis aliis differt foliis majoribus (ad 5 * 6 cm.) ovalibus integris vel supra inconspicue crenatis. Nuevo Leon: Mun. de Galeana, Cafon San Francisco on Hacienda Pablillo, alt. 2100 m., sparse on dry cliffs in the oak wood, No. 1282 (type in A. A.), July 31, 1934, Quercus verde, sp. nov. Arbor parva ramis gracilibus (1-2 mm.) sulcatis incano-tomentosis, annotinis pubescentibus, lenticellis sparsis minutis vel majusculis con- spicuisque; gemmae primum ovoideae et vix 1 mm. diam., demum acute ovoideae ad 2.5 & 3.5 mm. magnae, glabrae; stipulae subulatae, 5 mm. 164 JOURNAL OF THE ARNOLD ARBORETUM [VOL? XVII longae, plerumque persistentes. Folia decidua, oblonga vel obovata, obtusa, saepe mucronulata, leviter cordata, integra vel apicem versus dentibus paucis latis brevibusque mucronulatis, margine minute revoluto plano vel crispo, supra lucidula et sparsissime stellata, subtus pallidiora, pilis stellatis canescentia, venis utrinsecus 8 vel 9 ramosis et anastomo- santibus supra non sed subtus manifeste elevatis; petioli circiter 0.5—1 < 5-9 mm., dense stellato-pilosi. Amenta mascula 2.5—3.5 cm. longa, laxa, stellato-pilosa, antheris glabris breviter exsertis; amenta feminea 2—4-flora, pedunculo incano 1—2.5 cm. longo. Fructus annuus, non visus. Nuevo Leon: Mun. de Galeana, Cafion San Francisco on Hacienda Pablillo, alt. 2100 m., scattered in low oak wood on a hilltop, No. 373 (type in A. A.), May 14, 1934, No. 451, May 18, 1934; Cieneguillas on Hacienda Pablillo, alt. 2500 m., sparse in scrubby oak wood, No. 886, June 17, 1934. A moderately small tree with rather gnarled branches and scaly gray bark. It occurs infrequently in open oak wood. From all other species of the section GLAUCOIDEAE except Q. sororia Liebm. (Overs. Dansk. Vidensk. Selsk. Forhandl. 1854, p. 175) this species differs in having twigs stellately incanous the first year and still quite hairy the second. From Q. sororia it differs in its very slender twigs and leaves hairy be- neath. Its relationship to any one species in the series is problematic. % Quercus pastorensis, hybr. nov. (Q. glaucophylla von Seemen x Q. clivicola Trel. & C. H. Muell.). Ramuli, gemmae et fructus ut in Q. glaucophylla; ramuli et folia subtus persistenter pubescentes ut in Q. clivicola; folia nervatione Q. glaucophyllam referentia, sed textura et forma (oblonga margine cre- nata) Q. clivicolae similia. Nuevo Leon: Mun. de Galeana, Puerto de los Pastores near Galeana, alt. 1500 m., common on openly wooded, dry slopes, No. 1296 (type in A. A.) & 1297, August 2, 1934, Nos. 1000 & 1005, July 5, 1934. Nos. 1000 & 1297 seem to have resulted from a back cross with Q. glauco- phylla, In them the characters of Q. clivicola, especially the hairiness of the leaves, are somewhat less evident. Although the parents of this hybrid occur mixed over much of the range covered by this report, as well as in the type locality of the pro- posed hybrid, nowhere else was there noticed any evidence of hybridi- zation. Since the type locality is a rather poor site bearing little woody vegetation, it may be that the hybrid is in this case less vigorous than the parents and therefore confined to an area in which the rigors of drought have much reduced the competition. It is doubtless true that many of the species reported in Mexico are hybrids which cannot be 1936 ] MUELLER, STUDIES IN THE OAKS OF MEXICO 165 recognized as such because of the extremely small amount of field study which has been carried on. Series AURANTIACAE Trelease Quercus monterreyensis Trelease & Mueller in Bull. Torr. Bot. Club, 63: 151 (1936). NueEvo Leon: Mun. de Monterrey, Cafion Diente, alt. 600 m., sparse in the dense oak wood, No. 256, May 6, 1934; Mun. de Villa Santiago, Cafion Guajuco above Villa Santiago, alt. 1300 m., common in rather dense oak wood along arroyos and on slopes, Vo. 1315, August 10, 1934, Nos. 1328 & 1329, August 12, 1934; Potrero Redondo, alt. 1300 m., scattered in open oak wood on gentle slopes, No. 2080, July 3, 1935. Of these specimens No. 1328 differs considerably from the other specimens in having leaves about half normal size both surfaces of which are sparsely stellate hairy. These differences hardly merit a name even as a form, for they often occur on the same tree as typical leaves and are to be regarded as a part of the normal variation of the species. Series RETICULATAE Trelease Quercus reticulata Humboldt & Bonpland, Pl. Aequinoct. 2: 40 (1809). Nuevo Leon: Mun. de Galeana, Canon San Francisco on Hacienda Pablillo, alt. 2100 m., abundant along wooded waterways, Nos. 303 & 304, May 11, 1934, No. 1279A, July 31, 1934; above Alamar on Haci- enda Pablillo, alt. 2300 m., sparse in low oak scrub, No. 747, June 4, 1934; Canon de Cieneguillas above Pueblo San Francisco, alt. 2400 m., scattered along wooded waterways, Nos. 1271 & 1272, July 27, 1934. This polymorphic species is quite difficult to delimit, for there appears to be as much variation within the species as there is between it and some others of the series, notably Q. rhodophlebia Trelease (in Mem. Nat. Acad. Sci. 20: 74. 1924), QO. Uhdeana Trelease (1. c. 75), and Q. pilicaulis Trelease (1. c. 67). Quercus reticulata f. squarrosa Trelease in Mem. Nat. Acad. Sci. 20: 75 (1924). Nuevo Leon: Mun. de Galeana, Cafion San Francisco on Hacienda Pablillo, alt. 2100 m., sparse with the species along arroyos, No, 1279, July 31, 1934. Quercus reticulata f. longa Trelease in Mem. Nat. Acad. Sci. a 7511924); Nuevo Leon: Mun. de Galeana, Cafon San Francisco on Hacienda 166 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVIL Pablillo, alt. 2100 m., sparse with the species along arroyos, No. 1279B, July 31, 1934. This form, like the above, is not distinct from the species but grades into it. Quercus reticulata f. pungens, f. nov. A typo recedit foliis 1.5 & 3 cm. vel minoribus profunde mucronato- dentatis in dimidio superiore, margine revolutis et manifeste undulatis. Nuevo Leon: Mun. de Derrumbadero, Canon de los Capulines above San Enrique on Hacienda San Jose de Raices, alt. 2500 m., shrub, rare on dry, shrub-covered slopes, No, 2417 (type in A. A.), August 6, 1935. Quercus revoluta Trel.' f. acuta, f. nov. A typo recedit foliis magis elongatis (1-1.5 2-3.5 vel. 2 & 5 cm.) apice mucronatis acutissimis, dentatis vel integris, tomento subtus satis sparso. Nuevo Leon: Mun. de Galeana, Puente de Dios near Galeana, alt. 1200 m., abundant shrub on dry, shrub-covered hills, No. 2169 (type in A. A.), July 12, 1935. Quercus Loeseneri Trelease in Mem. Nat. Acad. Sci. 20: 79 (1924). Nuevo Leon: Mun. de Galeana, slopes about Cieneguillas on Hacienda Pablillo, alt. 2400 m., sparse in the shrubs of the dry hills, No. 882, June 17, 1934; Cerro Potosi near Galeana, alt. 2700 to 3300 m., abundant low shrub in the pine forest, No. 1263, July 26, 1934, Nos. 2225 & 2239, July 20, 1935, No. 2286, July 24, 1935. The distinction of this species from Q. reticulata seems very doubtful. Series MIcROPHYLLAE Trelease Quercus microphylla Née in Anal. Cienc. Nat. 3: 264 (1801). Nuevo Leon: Mun. de Galeana, Canon San Francisco on Hacienda Pablillo, alt. 2100 m., sparse shrubs on low-wooded slopes, Nos. 1291B & 1291X, July 31, 1934. The preceding specimens are doubtfully referred to this species. Probably the most confused group of white oaks and the last which will be satisfactorily understood is comprised of the various scrubby species of the desert mountain areas of northern Mexico and adjacent southwestern United States. One group, which seems to center about O. microphylla and Q. chihuahuensis Trel., is represented in these col- lections by twenty numbers. The specimens show definite affinities to several species, but by reason of their high degree of intergrading can 1Trelease in Mem. Nat. Acad. Sci. 20: 79 (1924). 1936] MUELLER, STUDIES IN THE OAKS OF MEXICO 167 only tentatively be referred to any species. It seems apparent that one of two conditions must be recognized. There might be an intricate system of closely related species which would many times multiply the number now known. To follow such an assumption would defeat the purpose of taxonomy. The only alternative which seems at present applicable is that there has been a high incidence of hybridization coupled with the natural variations of a profoundly fluctuating group of organisms. To classify satisfactorily such an entanglement will require field study of literally thousands of specimens and can hardly be undertaken in the study of a limited area such as the one now under consideration. Series CHIHUAHUENSES Trelease Quercus chihuahuensis Trelease in Mem. Nat. Acad. Sci. 20: 85 (1924). Nuevo Leon: Mun. de Galeana, Canon San Francisco on Hacienda Pablillo, alt. 2100 m., abundant low shrub on low-wooded slopes, Nos. 312 & 320, May 12, 1934, Nos. 1036 & 1037, July 15, 1934, Nos. 1288, 1288A, 1291, & 1291A, July 31, 1934; about Pablillo on Hacienda Pab- lillo, alt. 2000 to 2100 m., Nos. 535 & 536, May 21, 1934, No. 885, June 17, 1934, No. 1091, July 18, 1934; above Alamar on Hacienda Pablillo, alt. 2200 m., common low shrub in dry, open wood, No. 748, June 4, 1934, No. 1158A, July 22, 1934; Cerro Infernillo on Hacienda Pablillo, alt. 3000 m., a common shrub in the pine forest, Vo. 847, June 17, 1934; Cafion de las Placetas above Pueblo San Francisco, alt. 2300 m., abund- ant as an undergrowth in pine and oak wood, No. 1274, July 27, 1934; Cerro Potosi above Derramadero, alt. 2700 m., abundant low or creep- ing shrub in the pine forest, No. 2280, July 22, 1935; Arroyo Hondo near San Enrique on Hacienda San Jose de Raices, abundant on dry, shrub-covered slopes, Vo. 2311, July 31, 1935; Mun. de Villa Santiago, pass above Potrero Redondo, alt. 2100 to 2300 m., abundant in the shrub zones, No. 2134, July 7, 1935. The preceding specimens have been referred only tentatively to this species. Series OpacaE Trelease Quercus Pringlei von Seemen in Bot. Jahrb. 29: 96 (1900). Nuevo Leon: Mun. de Galeana, about Pablillo on Hacienda Pablillo, alt. 2000 to 2200 m., sparse on dry, shrub-covered hills, a low shrub, No. 880, June 17, 1934, No. 1090, July 18, 1934; Puerto de los Pastores, alt. 1500 m., rare on dry, openly wooded slopes, No. 995A, July 5, 1934; Mun. de Derrumbadero, hills above San Juanito, alt. 2300 m., abundant shrub on dry slopes, No. 2420, August 10, 1935. 168 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII Quercus Pringlei f. dentata, f. nov. A typo recedit foliis circiter triente inferiore excepto dentatis dentibus incurvis mucronatis, subtus initio flavo-stellato-tomentosis demum glabratis. Nuevo Leon: Mun. de Galeana, about Pablillo on Hacienda Pab- lillo, alt. 2000 m., rare shrub on dry slopes, No. 881 (type in A. A.), June 17, 1934. Series UNpULATAE Trelease Quercus undulata Torr. Vaseyana Rydberg in Bull. N. Y. Bot. Gard. 2: 218 (1901). Nuevo Leon: Mun. de Rayones, Casillas, alt. 1400 m., scattered on very dry and sparsely shrub-covered slopes, No. 2167, July 9, 1935. TaMAuLipas: Mun. de Villa Mainero, bajada east of the mountains and north of Villa Mainero, alt. 500 m., rare in the dense scrub thorn forest, Vo. 2431, August 14, 1935. The specimens agree perfectly with Q. sillae Trelease (in Mem. Nat. Acad. Sci. 20: 102. 1924) which is described from the vicinity of Monterrey, Nuevo Leon, a short distance to the north of the localities here cited. There is no apparent morphological distinction and no reason why Q. undulata should not range south over similar habitats from western Texas. Series PRINOIDEAE Trelease Quercus Muehlenbergii Engelmann in Trans. Acad. St. Louis 3: 391 (1877). Nuevo Leon: Mun. de Galeana, Alamar on Hacienda Pablillo, alt. about 1900 m., large tree abundant in the densely wooded canons, Nos. 623 & 651, May 30, 1934, No. 1120, July 21, 1934. Series VIRENTES Trelease Quercus fusiformis Small in Bull. Torr. Bot. Club, 28: 357 (1901). Nuevo Leon: Mun. de Villa Santiago, Cahon Denuncio above Rancho La Bola, alt. 750 m., abundant small tree in the more xerophytic oak wood, No. 2013, June 23, 1935. Subgenus ERYTHROBALANUS Oersted Series VIMINEAE Trelease Quercus duraznillo Trelease in Mem. Nat. Acad. Sci. 20: 122 (1924). Nuevo Leon: Mun. de Galeana, Cafion San Francisco on Hacienda Pablillo, alt. 2100 m., shrub abundant on open slopes, Vos. 369 (with definite indications of evergreen habit) & 374, May 14, 1934, Nos. 1289, 1936] MUELLER, STUDIES IN THE OAKS OF MEXICO 169 1289A, & 1290, July 31, 1934; Cerro Infernillo above Cieneguillas on Hacienda Pablillo, alt. 2700 m., common in low oak wood, No. 845, June 17, 1934; Mun. de Derrumbadero, hills above San Juanito, alt. 2500 m., common on very dry, open slopes, Vo. 2419, August 10, 1935. In most of the collections the leaves tend to be toothed, but not so deeply as in f. pinetorum Trelease (in Mem. Nat. Acad. Sci. 20: 123. 1924). The male catkins are scarlet when they first appear. The fruit is annual, paired on a short stalk, the cup scales shiny, light brown, almost glabrous, rounded or even truncate at the apex. Series CRASSIFOLIAE Trelease Quercus errans Trel.’ f. graciliramis, f. nov. A typo differt ramis gracilibus 2 mm. crassis et petiolis brevioribus; praeterea foliis manifeste sempervirentibus saepe ad basin dentatis et apice minus frequenter acutis. Nuevo Leon: Mun. de Galeana, above Pablillo on Hacienda. Pab- lillo, alt. about 2200 m., low shrub common in the dense, low oak wood of steep slopes, Nos. 394 (syntype in A. A.), 395, and 401, May 14, 1934; Cafion de las Placetas above Pueblo San Francisco, alt. 2000 m., common low shrub on densely shrub-covered steep slopes, No. 1273 (syntype in A. A.), July 27, 1934; Cerro Potosi above Derramadero, alt. 2800 m., abundant shrub 20 to 50 cm. high on pine-covered slopes, 1 to 2 m. high in arroyos, No. 2224, July 20, 1935. The specimens show distinctly the biennial fruiting habit, and in some the catkins are well preserved. The latter have not been de- scribed for the species: staminate catkins 4—7.5 cm. long, the peduncle loosely hairy, rather densely flowered above and loosely below, the stamens exserted. Pistillate catkins 10-20 mm. long, 3—5S-flowered, usually only the basal one developing into a single subsessile fruit. Series ARISTATAE Trelease Quercus Endlichiana Trelease in Mem. Nat. Acad. Sci. 20: 141 (1924). NvueEvo LEon: Mun. de Galeana, slopes above Alamar on Hacienda Pablillo, alt. 2200 m., small tree about 7 m. tall, forming a dense wood, Nos. 719 & 720, June 4, 1934, Nos. 1159, 1161, & 1162, July 22, 1934; Mun. de Villa Santiago, Canon Guajuco above Villa Santiago, alt. 1400 m., small tree or shrub 3 to 6 m. tall, abundant in pine and oak forest, No. 1330, August 12, 1934. Quercus Endlichiana f. minor, f. nov. A typo differt foliis parvis (2-3.5 4-6.5 cm.) leviter dentatis denti- 1Trelease in Mem. Nat. Acad. Sci. 20: 131 (1924). 170 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVII bus setaceo-aristatis vel supra profunde incisis. Nuevo Leon: Mun. de Galeana, slopes above Alamar on Hacienda Pablillo, alt. 2200 m., small tree about 4 m. tall, scattered in the pine and oak wood, Nos. 718, 728, & 729, June 4, 1934, Nos. 1157 (type in A. A.) & 1158, July 22, 1934; Mun. de Villa Santiago, Potrero Redondo, alt. 1400 m., moderate tree some 8 m. tall, growing on a dry, openly wooded slope, No. 2112, July 4, 1935, No. 2127, July 5, 1935. Quercus Endlichiana f. serrata, f. nov. A typo differt folius profunde incisis supra acute serratis. Nuevo Leon: Mun. de Galeana, slopes above Alamar on Hacienda Pablillo, alt. 2200 m., small tree about 4 m. tall, scattered in the pine and oak wood, No. 727, June 4, 1934, No. 1160, July 22, 1934; near Taray on Hacienda Pablillo, alt. 2200 m., moderate tree about 7 m. tall, sparse in the pine and oak wood, Nos. 585 & 587, May 25, 1934; Mun. de Villa Santiago, Cafion Guajuco above Villa Santiago, alt. 1400 m., moderate tree abundant in oak and pine wood, No. 1330A (type in A. A.), August 12, 1934; Potrero Redondo, alt. 1400 m., moderate tree about 8 m. tall, abundant in the pine and oak forests, No. 2126, July 5. 1945; These specimens and those of f. minor grade into one another almost imperceptibly. Certain numbers of the series representing this species (notably Nos. 1160, 1330 and 1330A) very pointedly indicate a close relationship to Q. aerea Trelease (in Mem. Nat. Acad. Sci. 20: 135. 1924.) in general appearance, leaf form, and “. . . resiniferous beneath so as to be of the color of ‘bronzy old gold’. . .”.. The only point upon which cer- tain of these specimens can be excluded from Q. aerea is that of the small round-ovoid buds of that species, and even so, certain of the less mature specimens of this series fit that description. Perhaps a com- parison of this material with the types of Q. Endlichiana and Q. aerea ‘would prove them identical. The description of the latter, however, is fragmentary, and it is to be hoped that the more fully described Q. Endlichiana be allowed to stand. Series RysopHYLLAE Trelease Quercus rysophylla Weatherby in Proc. Amer. Acad. 45: 423 (1910). Nuevo Leon: Mun. de Galeana, Alamar on Hacienda Pablillo, alt. 1900 m., tall tree, 20 to 25 m., common on densely wooded canon floors, Nos. 625 & 633, May 30, 1934, and No. 1123, July 21, 1934; Mun. de Villa Santiago, Potrero Redondo, alt. 1400 m., common on densely wooded cafion floors, 15 to 25 m. tall, No. 2113, July 4, 1935. 1936] MUELLER, STUDIES IN THE OAKS OF MEXICO 171 Series EUGENIAEFOLIAE Trelease Quercus pinnativenulosa, sp. nov. Arbor parva ramulis glabris gracilibus sulcatis, lenticellis fulvis con- spicuis; gemmae subfusiformes, circa 1.5 & 3 mm. magnae, acutae. Folia ut videtur decidua, subcoriacea, integra, anguste ovata vel pleraque lanceolata, utrinque acuta, raro basi late cuneata, satis parva (2 «& 5 to 3 & 12 cm.), supra et infra glabra et lucida, venis utrinsecus circa 18 utrinque prominulis, tenuioribus alternantibus, fere horizontaliter diver- gentibus, juxta marginem minute revolutum manifeste anastomosanti- bus; petioli supra satis applanati, circa 1 & 3-5 mm., apicem versus lamina decurrente alati. Amenta mascula 4—5