JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY ALFRED REHDER EDITOR JOSEPH H. FAULL ann CLARENCE E. KOBUSKI ASSOCIATE EDITORS VOLUME XVI JAMAICA PLAIN, MASS. 1935 Reprinted with the permission of the Arnold Arboretum of Harvard University KRAUS REPRINT CORPORATION 2 rk 1968 S46 F4 DEG * 21968 DATES OF ISSUE No. 1, (pp. 1-143, pl. 119-128) issued January 25, 1935. No. 2, (pp. 145-271, pl. 129-139) issued April 24, 1935. No. 3, (pp. 273-365, pl. 140-154) issued July 10, 1935. . 4, (pp. 367-483, pl. 155-165) issued October 25, 1935. Printed in U.S.A. TABLE OF CONTENTS STUDIES IN BORAGINACEAE, X. THE ee OF NORTHEASTERN SoutH A Iva MERICA. By Ivan M. Johnston ....... 0c cc ccc cece ccc cee 1 HANDELIODENDRON, A NEw GENUS OF ee With plate 119 and one text figure. By Alfred Rehder .........0. ccc eee cuccee 65 NOTES ON SOME OF THE es ENACEAE AND VERBENACEAE OF THE SOLO MON ISLANDS COLLECT HE ARNOLD ARBORETUM Expres: 1930-1932. With Sie 120-122. By R. C. Bakhuizen van den Brink 68 An ENDEMIC SOPHORA FROM RuMANIA. With plates 123 and 124 and one text figure. By Edgar Anderson ........cccccccccccccececs 76 SUPPLEMENT TO THE SPONTANEOUS FLORA OF THE ARNOLD ARBORE- et ESE er NET oy ci oc cass by aw a cee eee 81 THE Hosts oF GYMNOSPORANGIUM GLOBOSUM FARL. AND THEIR RELA- TIVE SUSCEPTIBILITY. With plates 125-128 and. four text figures. yh tae OC EON ica ee ce da bles Boe ewe ome ere es 98 A PRELIMINARY NOTE on Lire History Stupies oF EuroPEAN SPECIES OF MixEsia. By Lillian M. Hunter ........0.. 0000 cece 143 STUDIES IN THE BoRAGINACEAE, XI. By Ivan M. Johnston ......... 145 LoRANTHACEAE puiee: IN THE SOLOMON IsLANps By L. J. Bras AND Kaj KI ON THE ARNOLD crore Eee Cae 1930-1932. With | Site 129. BVO. tt) DGHSEP ilesceds cee eeeesees 206 CHROMOSOME NUMBERS IN THE HAMAMELIDACEAE AND THEIR PHYLO- GENETIC SIGNIFICANCE. With three text figures. By Edgar Ander- SOM-GNG ICMP NOS 2c aaa es wets ws 88a sacs oe eek 210 CHROMOSOME STRUCTURE IN THE MEIOTIC CHROMOSOMES OF RHOEO DISCOLOR Hance. With plates 130 and 131. By Karl Sax ........ 216 ar he ap IN Iris: A MorpuHotocicaL Stupy. With plates 132- ys WOE Faas ee ae een hee irs es ae es 225 NoTES ON , With plates 138 and 139. By Susan Delano TREE, O82 od acs See ee EP OR wae: 268 THE VISIBLE STRUCTURE OF THE SECONDARY WALL AND ITS pen CANCE IN PHYSICAL AND CHEMICAL INVESTIGATIONS OF TRACHEA CELLS AND FIBErs. With Ree 140-149. By J. W. Bailey er TORRES LIT” w= 554i ee RE Ee Ee ee Ae ad ea fa THE Errect oF TEMPERATURE ON NUCLEAR DIFFERENTIATION IN MIcrosPpoRE DEVELOPMENT. With one text figure and plate 150. By POO SO iho buch ce ee eek e s eas 301 NOTES ON THE - IGNEOUS PLANTS DESCRIBED BY LEVEILLE FROM Fast- ERN ae > — CO SCHOO Sema Pe ene, os ous sweatoees 311 Hu w GENUS OF STYRACACEAE. With one text figure ad pore “is! me 1152 By Allred Keer fons oes ie coe ee cess 341 STUDIES IN THEACEAE, I. EuryA SUBGEN. TERNSTROEMIOPSIS. With plate 153. By Clarence E. PLODUIE aie bn al oils aoe ws ¥ VAR Bae 347 Two New SPECIES OF CRATAEGUS FROM Missouri. With two text heures, By Lrnest J, PaNMet <9 we a ei he a oa SR ee $53 iv TABLE OF CONTENTS New Hysrips FROM THE ARNOLD ARBORETUM. By Edgar Anderson Ri ATM AIREY oedig es ho 4565 a FESR EO PAREN RN ERE 358 HypopDERMELLA HiIRATSUKAE, A NEw SPECIES OF HYPODERMATACEAE FROM JAPAN. With as 154. By Grant D. Darker ............. 364 Tue Hosts, Lire His ND CONTROL OF GYMNOSPORANGIUM CLAVIPES C. AND P. With Ree 155-160. By Jvan H. Crowell ... 367 Tue DISPERSAL OF VIABLE BASIDIOSPORES OF THE GYMNOSPORAN- GiuM Rusts. With two ae figures. By J. D. MacLachlan ...... 411 CHROMOSOME STRUCTURE AND BEHAVIOR IN Mitosis AND MEIOSIS. With plates 161-164. By Hally Jollivette Sax and Karl Sax ...... 423 THe Fiora oF SAN Fetix Istanp. With plate 165. By Ivan M. J OWMWSTOW: Seco x Sto ts a As Bae ai ord SORIA ERA a aie os Laie arate ahs 440 SoME New TREES AND SHRUBS FROM Mexico. By Alfred Rehder.. 448 New Facts ConcERNING CEPHALOSPORIUM WILT OF Ets. By D. B. CPGRAEE 55565 sa Kd HAA AER ORES ES ROR ES ERO SR RRR EATERS 453 Tue ARNOLD ARBORETUM DURING THE FISCAL YEAR ENDED JUNE 30, 1935; The Arboretum; The Pathological ieee ory; The Cyto- genetic Laboratory ; The Herbarium; The Library; rie uate of the published writings of the staff and sheets July 1, 1934—June DU. P08 ca Nae RECA SR RISES ASRS ODENSE VERE TESTA Sy ba ETS 455 STAFF OF THE ARNOLD ARBORETUM 1934-1935 ............. 20s eee 469 PRBARA 6 oases hese aaah ees eee ATR EEEES Week Ob See eee eRt 470 JOURNAL OF THE ARNOLD ARBORETUM VOLUME XVI JANUARY, 1935 NuMBER 1 STUDIES IN BORAGINACEAE, X THE BORAGINACEAE OF NORTHEASTERN SOUTH AMERICA Ivan M. JoHNSTON THE present paper is a critical account of the Boraginaceae known from British, Dutch and French Guiana and the adjoining portions of Brazil, north and east of, the Amazon and the Rio Negro. A general account, it is preliminary to a treatment of the Dutch Guianan species of the family which Prof. A. A. Pulle has invited me to prepare for his “Flora of Surinam.” The borages of the Guianas have been long neglected. Such frag- mentary work as has been done on them has been restricted to the nar- row political boundaries. Though various species of the group have been described from the Guianas, some of them among the first based upon South American material, the identity of the types has remained obscure, and material in herbaria has continued to be named largely by guess or has been left to accumulate unidentified. The great reference works, such as DeCandolle’s Prodromus, or Martius’s Flora Brasiliensis, resolve little of the confusion that seems always to have enveloped our knowledge of the Guianan Boraginaceae. They added little to the observations all ready long available in the writings of Lamarck and Poiret. Indeed, so little known and confused were the Guianan species of Cordia and Tournefortia, that a few years ago, during my studies of the Brazilian species of these genera, Contr. Gray Herb. 92: 1-89 (1930), I was forced to pass over, undiscussed, the very evident rela- tions existing between the species of the Guianas and those of northern Brazil, and forced to admit that certain of the obscure species (several of them not even listed in the Index Kewensis) might be identical and older than the ones I was forced to accept. A careful study of the Guianan Boraginaceae has been long needed. 2 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI The conspicuous relationship evident among the Guianan Boragi- naceae is that with Brazil, most of the species extending into and about the Amazon Basin or having their immediate relatives there. The affinities westward in Venezuela are not so numerous nor so pronounced. Except for Trinidad (which after all is floristically close to that of east- ern Venezuela) direct relations to the northward are negligible. Among the Guianan Boraginaceae only the group Cordia § Pilicordia has devel- oped any number of local species. The relations of these local endemics are in the Amazon Basin where the species of this group are not local but widely distributed. The Guianas are a marked endemic center for Pilicordia comparable with the centers of that group found in south- eastern Brazil, northern Venezuela and adjacent Colombia, and the West Indies. I have treated in this paper all the borages known north and east of the Amazon, the Rio Negro and the eastern boundary of Venezuela. The monotypic Lepidocordia is endemic to this area. Of the 38 species definitely known from this large area only two, Cordia multispicata and Cordia naidophila are at present unknown from British, Dutch or French Guiana. Several other species approach our area, reaching the Orinoco Valley from the westward. Among these species those which may eventually be found in northwestern British Guiana are Cordia globosa (Jacq.) HBK., Cordia alba Jacq. and Bourreria cumanensis (Loefl.) O. E. Schulz. The writings of Schomburgk, Aublet and others have listed various West Indian species from the Guianas. Some of these records are evidently based upon misidentifications, others, how- ever, I am convinced, are simply unfortunate guesses as to what the authors believed might be found there. Most of these questionable records relate to species common and widespread at low altitudes in the Antilles. This group of plants is poorly represented on the Guianan coastal area, probably because of adverse winds and currents and the unfavorably humid conditions. In the preparation of this report I have examined practically all the types concerned and have studied most of the large or important Gui- anan collections in Europe and United States. Studies have been made at Kew, London, Leiden, Utrecht, Copenhagen, Berlin, Munich, Geneva and Paris. Large loans of critical material have been obligingly sent for further, more leisurely study at the Arnold Arboretum and the Gray Herbarium from Kew, London, Leiden, Utrecht, Berlin, Paris and New York. Particular mention, however, is to be made of the large loan from the Botanical Museum at Utrecht. This material, assembled through the inspiration of Prof. Pulle and kindly made available to me 1935] JOHNSTON, STUDIES IN BORAGINACEAE 3 by him, consists of numerous series of copious specimens collected over a number of years, at different seasons, from various numbered individual trees or shrubs, in the Forest Reserves of Dutch Guiana. Through the examination of this remarkable record of seasonal varia- tion I have been able to establish unquestionably the specific identity of certain seasonal forms heretofore troublesome to identify. Of great help in the preparation of this report the collections have been generally instructive to me personally. It has been a privilege to have such a convincing demonstration of the nature and extent of seasonal variation in individual trees and shrubs of the Tropics. The following abbreviations have been used in designating the source of the material cited. B. W.— collections by the Forest Service (Boschwezen) of Dutch Guiana; AAA — Arnold Arboretum; BD — Botanical Museum at Berlin; BM — British Museum of Natural His- tory; DC — Prodromus Herbarium of DeCandolle at Geneva; Del — Delessert Herbarium at Geneva; G — Gray Herbarium; K — herba- rium at Kew; Leid — herbarium at Leiden; NY — New York Botani- cal Garden; US—U. S. National Herbarium; Utr — Utrecht Herbarium. KEY TO THE GENERA Stigmas 2 or 4, simple; inflorescence cymose-paniculate or spicate or globose, the branches not distinctly scorpioid; erect broad-leaved trees and shrubs. Stigmas 4, capitate or clavate, borne on a conspicuous slender, twice lobed OF lett Style. cect nes Gag cd coh ea 1. Cordia, p. 3. Stigmas 2, conic, sessile on the apex of the ovary, short and IMECUMSDICUOUS: .5<'bs ese ee yas ee ed oe . Lepidocordia, p. 45. are t eae, of an anulate fertile base and a more or less eloped sterile frequently bifid apical portion; inflo- rescence with distinctly scorpioid branches or the flowers cauline and solitary in the internodes. Fruit baccate; clambering shrubs ............. 3. Tournefortia, p. 46. Fruit dry; herbs or small suffrutescent plants ..4. Heliotropium, p. 57. 1. Cordia [Plumier] Linnaeus, Gen. 87 (1754). Trees or shrubs, usually with broad leaves. Inflorescence a loosely paniculate or glomerate or capitate or spicate cyme. Calyx usually 5-toothed or 5—10-lobed, usually persistent. Corolla white, yellow or reddish, small to conspicuous, usually 5-merous, rarely 6—15-merous, salvetform or subrotate to funnelform or subtubular. Stamens as many as the corolla-lobes, exserted or included, filaments short or long. Ovary 4-celled, ovules 1-4. Style terminal, well developed, 2-lobed or 2-parted, the branches each 2-lobed. Stigmas 4, capitate or clavate, small. Fruit unlobed, a drupe with a bony pit and mucilaginous or dry exocarp, or 4 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI the walls dry and papery, 1—4-celled. Endosperm none. Cotyledons plicate. A very large genus of diverse habit and structures; widely distributed throughout the Tropics. Centering in America. Type Species: C. sebestena L. On the grounds that the original “Cordia” of Plumier, Nov. Pl. Amer. Gen. 13, tab. 14 (1703), which was accepted and validated by Lin- naeus, is not a member of the Boraginaceae, the generic name Cordia has been recently discarded for the present concept by Dr. von Friesen, Bull. Soc. Bot. Genéve, sér. 2,24: 131-4 (1933). With this I can not agree. It is pointed out by von Friesen that Plumier’s illustration shows a 2- celled ovary and a simply bifid style and that the generic description of Cordia (based upon Plumier’s plate and description) given by Linnaeus in the Genera Plantarum of 1754, pg. 87, also calls for these structures. Dr. von Friesen believes they are structures of some genus outside of the Boraginaceae. I believe they are structures of Cordia sebestena faultily described from inaccurate drawings. I have had the privilege of studying, in the library of the Natural History Museum at Paris, the amazing series of volumes of plates and manuscripts accumulated by Plumier during his visits (1689-97) to the West Indies. In one of these volumes of manuscript, 6: tab. 64-66, are found fine drawings labeled “Cordia nucis iuglandis folio, flore purpureo.” The best of these original drawings, made in the West Indies by Plumier, fills a folio page and shows a characteristic branch of Cordia sebestena bearing leaves, flowers and fruit. In the corner of the page are the details of flower and fruit, differing only in arrangement from those printed in Plumier’s Genera. The structures of style and ovary are quite alike in both. This may be verified by a comparison of the small plate in Plumier’s Genera and the good copy of the original folio plate published in Burmann’s edition of Plumier’s Plantarum Ameri- canarum, fasc. 5, tab. 105 (1757). The later plate is identified as of sebestena by Urban, Rep. Spec. Nov. Beiheft 5: 60 (1920). Von Frie- sen, I. c. 135, however, believes that only the leafy, flower-bearing branch is C. sebestena and that the disputed unattached analytic details belong to some other genus. Since, however, the disputed details are an inte- gral part of the original drawing of Plumier, which consists mainly of a flowering and fruiting branch unquestionably of C. sebestena, and since the details, as far as one can compare them, are quite like homologous structures shown growing attached to the flowering and fruiting branch of C. sebestena, I feel there is every reason for believing that, however inaccurate, they were intended to show the structures of that species. 1935] JOHNSTON, STUDIES IN BORAGINACEAE 5 It is to be recalled that Plumier’s drawings were made long before the work of Linnaeus on the Sexual System directed particular attention and gave special importance to the number and structure of the internal parts of the flower and fruit. Plumier, and Linnaeus who copied from him, may have given erroneous descriptions of the fruit and style of Cordia sebestena but since they were trying to describe that species I believe we should retain their name for the genus containing it. Plumier’s drawing was made on the island of St. Thomas. The fol- lowing quotation from his manuscript gives the type-locality of C. sebes- tena in some detail. ‘‘Martio plantarum florentem frutusque maturos ferentem adinveni apud insulam Sancti Thomae, juxta Littus quoddam La Baye du nord vocitatum, sinum scilicet ad septenttrionalem plagam ipsi Arci oppositum.”’ Key TO THE SPECIES Corolla marcescent; fruit cylindrical, dry, with a fibrous charta- ceous coat, not bony, closely invested by the tube of the per- sistent corolla and the strongly ribbed cylindrical calyx, at maturity flower disarticulating from the inflorescence and the calyx and corolla and the enclosed fruit falling away together, with the spreading corolla-lobes acting as a para- chute; pubescence stellate ; axis of : ai usually tun- nelled or inflated and inhabited by ants ............-. 1. C. alliodora. Corolla withering after anthesis and soon ear nae fruit usu- nd aa with a bony ovoid or globose stone ; pubescence Corolla em or orange, large; calyx becoming fleshy and com- pletely enveloping the dry fruit and even adnate to it . C. sebestena. aw 6. We oe SS ee 66 wale fe Lol eee, ek eo Oe ae 8 ee SNe ee Oe 6) eee Corolla white or yellow; fruit juicy, not adnate to the dry calyx Gores se longer than broad; calyx explanate at ma- inflorescence usually large and loosely branched ; ed or shrubs with usually horizontal branches and flat tops. Petioles of well developed leaves 15-40 (usually 20-30) mm. - ovary and fruit glabrous; inflorescences terminating leafy branchlets. Leaves glabrous and lustrous above, 15-40 cm. long; calyx ca. 5 mm. long; stone obliquely ovoid, ca. 18 mm. long; explanate calyx 10-13 mm. ne Leaves oe and rather dull above, 10-27 cm. ane calyx 2.5-3.5 mm. long; stone transversely com- pressed-ovoid, ca. 10 mm. - long; explanate calyx 57 thm; DrOdd .a5.eaveer cermin ae se <2 4. C. Petioles of well developed leaves 3-15 (usually 5-10) C. fallax. tetrandra. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XvI mm. long ; sucesso usually borne at the forks of the dichotomous stem Ovary and style hairy; fruit mostly pubescent. Leaves glabrous above or practically so; veins less conspicuously rebranched than in next; calyx usually apiculate, opening somewhat irregularly. Stems with conspicuous subnodal swellings that erve as ant-domatia; calyx tending to disinte- nas at maturity and showing a fibrous struc- ture; plant usually ieee “— LAKE EOERER SEWER eee BES we .5. C. nodosa. Stems: eee subnodal swellings, nae myrmeco- philous ; calyx not with fibrous structure ; plant not bristl Fruit glabrous ; calyx with a fine minute strigose- puberulence; lower leaf-surfaces ee or practically so ............... : . laevifrons. Fruit iain calyx ees ou i sur- es pubescent. ee surfaces of leaves evidently bearing numerous erect slender hairs; inflo- rescence stiffish but loose and open...7. C. Sprucei. Lower surfaces of leaves apparently glabrous, but really bearing scattered minute in- conspicuous very short ascending hairs: inflorescence dense with short rigid crowded branches .............0... . C. nervosa. Leaves hacer! hairy above, veins repeatedly re- ached; calyx opening by 5 triangular lobes. Calyx prominently and regularly 10-ribbed, 4—5 mm. long; fruit ee stone transversely IE “ees cass coe cea nwo seu seni sansa 9. C. fulva. Calyx not ribbed, 2-4 mm. long. Upper —— of leaves shies with very abund- ant slender erect or ascending hairs: leaves ene ancl dimorphic about the stem- forks (the normal elongate leaves usually opposed by much smaller suborbicular ones); fruit clothed bo ahaa slender appressed usually tawny hairs; stone ascend- AS, DVONG 60d Boa coke tn i eehs eee 0. C. toqueve. Upper surface of leaves simply strigose or mi- nutely scabrous ; leaves homomorphic. Lower surface of leaves green boutons brown), scabrid with very short sparse hairs; leaves lanceolate; mae Pee strigose; stone ascending ovo siaip ths au maeid baare ae weed Ged E es 11. C. scabrifolia. Lower sic of leaves pallid, with a felty overing of abundant appressed slender 1935] JOHNSTON, STUDIES IN BORAGINACEAE 7 hairs; leaves broadly lanceolate to ovate; fruit glabrescent; stone transversely ovoid piohssaatuesarea ead eras oeeeeee 12. C. bicolor. Ovary, style, and fruit glabrous. Lower surface of leaves pallid with a felty covering of abundant slender appressed hairs ...... 12. C. bicolor. Lower surface of leaves not felty with a pallid indument. Stone globose or depressed globose, quite rugose; calyx with abundant long slender hairs on the inner surface which project beyond the edge of the calyx-lobes and appear as a dense pale cili- ation on their margins. Calyx outside covered with abundant slender silky hairs; leaves rather thin, more or less dimorphic at the stem-forks, lower surface much paler than the upper ....... 13. C. sericicalyx. Calyx sparsely strigose outside; leaves rigid, homomorphic, lower surface ee paler THANGUDDEE once enn be nee ee sors C. panicularis. Stone ovoid or ellipsoid, smooth, erect, ee calyx strigose on the inner surface, the hairs projecting beyond the calyx-lobes sparse and dark if present Hairs on lower leaf-surface erect. Lower leaf-surface somewhat scabrid with minute stout hairs; leaves large, 15-28 cm. long, with evidently falcate midrib; branchlets with short erect hairs ...15. C. Sagotit. Lower leaf-surface velvety with long sender hairs; leaves moderate-sized, 8-18 cm. long, midrib weakly falcate; branchlets DrOWHISH POM ais eee chase wan 16. C. hirta. Hairs on lower leaf-surface appressed. Flower-buds elongate, obovoid, 4-5 mm. long; leaves 8-20 cm. long; inflorescence large and _stiffish; Lower Amazon and the uianas. Leaf-blades broadest at or above middle, drying olivaceous or muddy brown BP ok er ae 17. C. exaltata. Leaf-blades broadest at or below the middle drying a bright warm brown ee Oe ee 17a. var. melanoneura. Flower buds subglobose, 2-3 mm. long; leaves 6-11 cm. long; inflorescence slender and usually small; Upper Amazon ...18. C. naidophila. Corolla-lobes distinctly broader than long; calyx cupulate or cylindrical at maturity; inflorescence dense, globose or spicate, or exceptionally a small loose cyme; shrubs 8 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI with erect or ascending branches, frequently sub- scandent Corolla very large, 35-50 mm. long, the slender tube abruptly and much expanded into a coarse cylindrical throat ; inflorescence capitate; tip of calyx-lobes con- spicuously long-attenuate ................ 19. C. grandiflora. saat small, about 5 mm. long or less, tube and throat eakly aimmpenad, jee aee glomerate or cymose ......... 20. C. polycephala. Inflorescence distinctly spicate. Leaves not hairy above, merely more or less verru- cose or muriculate, elongate; spikes terminal; petioles not decurrent on the peduncles Hise Ha eV sae Se ere t seamen eevaeeels C. wmaerostachya. Leaves hairy above, strigose or velvety, broad; spikes axillary with the base of the petiole apparently decurrent on the subtended peduncle. Calyx-lobes in bud lacking free tips; inflorescence becoming quite loosely flowered and elongate ; upper surface of leaves strigose, somewh¢ it lustrous ........... eee seen e 22, C. Schomburgkii. Calyx-lobes in the bud with projecting free tips; inflorescence dense si stout; upper surface of leaves with stiff erect or ascending hairs iar arise from tulbaae bases, surface not lustrous Calyx deigh = over, lobes narrowly triangular, on BOOS fey teas exe M oe ve 23. C. tomentosa. Calyx with nde nearly glabrous, lobes broadly triangular, short acuminate .....24. C. multispicata. 1. Cordia alliodora (R. & P.) Chamisso ex DeCandolle Prodr. 9: 472 (1845); Johnston, Contr. Gray Herb. 92: 13 (1930). Cerdana alliodora Ruiz & Pavon, Fl. Peruv. 2: 47, tab. 184 (1799). Cordia trichotoma sensu Sandwith, Kew Bull. 1933: 335 (1933). Tree up to 20 m. tall; branchlets sparingly to densely stellate- pubescent; leaves oblong or lanceolate to elliptic, usually broadest at or above the middle, 3-8 cm. broad, 1-2 dm. long, base acute or obtuse, apex acuminate, margin entire, upper surface stellate-pubescent or glabrate, lower surface paler, stellate-tomentose or glabrescent, 5—7 pairs of veins, petiole 1-3 cm. long; inflorescence terminal, loosely and widely branched, 1-3 dm. thick, the flowers crowded on the branches, the axis commonly inflated, gall-like, irregular, usually serving as an ant-domatium; calyx cylindrical, with ten prominent ribs, densely stellate-tomentose, 4-6 mm. long, 2—2.5 mm. thick, lobes 5, inconspicu- ous; corolla white, drying brown, marcescent, lobes oblong, 5-7 mm. long, 1.5-3.5 mm. broad, spreading; fruit sausage-shaped with fibrous g acum 1935] JOHNSTON, STUDIES IN BORAGINACEAE 9 chartaceous wall, ca. 5 mm. long, completely enveloped by the tube of the persistent corolla and by the ensheathing calyx-tube and falling away enclosed by them. Headwaters of the Rio Branco in northern Brazil and adjacent south- ern British Guiana; northern Venezuela and Colombia and southward along the Andes and northward in Central America and the West Indies. BritTisH GUIANA: north side of Kanuku Mts., ca. 10 miles east of the Takutu River, ca. 135 m. alt., small tree, 4.5 m. tall, trunk ca. 8 cm. thick, in secondary forest near edge of savanna, fl. pure white, Oct. 10, 1931, Forest Dept. Brit. Guiana D230/2221 (K); Pirara (Marakanata), Rupu- nuni Savannas, ca. 120 m. alt., tree ca. 20 m. tall, trunk 11 m. to fork, 4 dm. thick, in sandy soil on patch of savanna-forest on top of ridge, Oct. 21, 1931, Forest Dept. Brit. Guiana D195/2186 (K). BraziL: Mniam, tributary of Suruma River, Nov. 1909, Ule 8290 (K, BD, Del) ; Limao, lower Cotinga River, Sept. 1927, Tate 140 (NY). Although previously I have cited one of the above collections as C. trichotoma, Contr. Gray Herb. 92: 15 (1930), I am now of the opinion that all the material from the upper Rio Branco watershed is more closely related to C. alliodora. The corolla-lobes in our plants are 3—3.5 mm. broad. The stems are simply tunneled by ants. There are no dis- torted, gall-like thickenings in the axis of the inflorescence. Compared with large series of C. trichotoma and C. alliodora our plants seem most like the latter species in gross aspect. The colony on the Rio Branco was probably derived from northern Venezuela where only C. alliodora is known. Strangely C. alliodora seems to be rare or absent in the Ori- noco Valley and in the other parts of the wet tropical forests of north- eastern South America. Previously I have attempted to maintain the Argentine, Paraguayan and Brazilian plant, ranging to the east and south of the Amazon Basin, as a species distinct from C. alliodora. I am now of the belief that this plant, called C. trichotoma in my treatment of the Brazilian species, is distinguished from C. alliodora only by its larger flowers, and that it had best be classified as a variety of that latter species. The correct tri- nomial for the large-flowered Brazilian form is Cordia alliodora var. tomentosa A. DC. 2. Cordia sebestena Linnaeus, Sp. Pl. 190 (1753). Cordia spe- ciosa Salisbury, Prodr. 111 (1796); DeCandolle, Prodr. 9: 476 (1845) ; Pulle, Enum. Pl. Surinam 397 (1906). Tree or shrub I-7 m. tall; branchlets with a fine soft curly pu- bescence and scattered much coarser appressed hairs; leaves ovate to elliptic or subcordate, 9-16 cm. long, 5—14 cm. broad, broadest below the middle, base obtuse or rounded or subcordate, apex obtuse to coarsely 10 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI short-acuminate, margin entire, upper surface with scattered short stiff appressed hairs, the hairs usually arising from minute pustulate disks, lower surface glabrescent or sparsely strigose, with 5—6 pairs of veins; petiole slender, 1-4 cm. long; inflorescence corymbose, usually ter- minal, ascendingly branched; calyx firm, strigose and densely brown puberulent, elongate in the bud, 12-15 mm. long, 3-5 mm. thick, open- ing by several unequal teeth ca. 2-3 mm. long, at maturity becoming much expanded by the enlarging fruit which it encloses, 3-4 cm. long; corolla orange or scarlet, funnelform, tube twice length of the eylindrical calyx; fruit bony, dry, ovoid, pointed, 1-2 cm. long, completely and tightly invested by the juicy white accrescent calyx. Native on the islands of the Caribbean and probably also along the coasts of Venezuela, Colombia and Central America; frequently culti- vated in the Tropics. BritisH GuIANA: Botanic a Georgetown, cultivated, Aug. 1905, collector not given 7976/6915 (BD). DutcH GUIANA: Surinam, Od tree 6-9 m., fl. red, Dec. 1837, Splitgerber 312 (Leid) ; Cottica mee near oe Alhance, Aug. 1901, Went 280 (Utr); Paramaribo, Focke 1371 FRENCH GUIANA: indefinite, Gabriel on 3. Cordia fallax, sp. nov. Cordia guianensis Klotzsch ex Schom- burgk, Fauna u. Fl. Brit. Guian. 960 (1848), nomen; not C. gujanensis (Desv.) R. & S. (1819), nor C. guianensis R. & S. ex DC. (1845). Arbor 5-10 m. alta; ramulis brunnescentibus cum pilis abundantibus brevibus erectis velutinis; foliis homomorphis ellipticis vel obovato- oblongis 15-40 cm. longis 6-18 cm. latis ad medium vel paullo supra medium latioribus minute glanduloso-punctulatis, basi rotundis vel sub- cordatis ad obtusis vel late acutis, apice breviter acuminatis, margine integerrimis vel rariter leviter sinuatis, supra lucentibus in costa et nervis primariis pilos inconspicuos gerentibus ceteris glabris vel sub- glabris, subtus pilis plus minusve abundantibus gracillimis ascendentibus molliter vestitis, nervis 7—8-jugatis, costa falcato-curvatis, petiolis 15-30 mm. longis; cymis ramulos foliatos terminantibus laxissime ramosis ad 3 dm. diametro; calyce in alabastro obovoideo extus indumento brun- nescente velutino molli vestito, intus glaberrimo, ad anthesin ca. 5 mm. longo (lobis deltoideis 5), fructifero explanato 10-14 mm. lato; corolla 1 cm. longa, lobis obovatis extus glabris, filamentis basim versus pilosis; ovario et stylo glaberrimo; fructu glabro; nuce valde rugoso oblique ovoideo acuminato ca. 18 mm. longo. Endemic to British Guiana. BritisH GUIANA: Issorora, Aruka River, wet forest, tree 9 m. tall, Jan. 1935] JOHNSTON, STUDIES IN BORAGINACEAE 11 1920, Hitchcock 17563 (TYPE, pug Herb. ; shee NY); upper Rupununi River near Dadanawa, ca. lat. 2° 45’ N., tree 5 m. tall, June 10, 1922, La Cruz 1484 (NY); indefinite, aa Sou 875 /1510b (K); indefi- nite, 1841, Schomburgk 875 (BD, Del, P); indefinite, Schomburgk 1510 (BD, type of C. guianensis; G). Although bearing various numbers and different data, the material from Schomburgk cited above agrees so completely in details of ma- turity, pressing, etc., that one may recognize it as consisting of parts of a single collection. The material at Berlin bears Klotzsch’s binomial. This, however, has never been associated with a description and is fur- ther invalid by reason of being a homonym. Schomburgk, I. c., reported C. guianensis Kl. only from the banks of the Barama River and there is every reason for believing that this is indeed the source of the Schom- burgk material mentioned. The species has been confused with C. tetrandra, although it is readily distinguished from that species by having glabrous upper leaf-surfaces and in being noticeably larger in all its parts. Its relations are with that group of species of Venezuela and Colombia which is exemplified by C. bogotensis Benth. Its very large leaves, hairy beneath, quickly dis- tinguish it from these much more westerly species. 4. Cordia tetrandra Aublet, Hist. Pl. Guian. Fr. 1: 222, tab. 87 (1775); Poiret, Encyc. 7: 42 (1806); Pulle, Enum. Pl. Surinam 397 (1906); Johnston, Contr. Gray Herb. 92: 55 (1930). Lithocardium tetrandrum (Aubl.) Kuntze, Rev. Gen. 2: 976 (1891). Cordia cordi- folia Humboldt, Bonpland & Kunth, Nov. Gen. et Sp. 3: 70 (1818); DeCandolle, Prodr. 9: 483 (1845). Lithocardium cordifolium (HBK.) Kuntze, Rev. Gen. 2: 976 (1891). Cordia muneco Humboldt, Bonpland & Kunth, Nov. Gen. et Sp. 7: 207 (1825); DeCandolle, Prodr. 9: 486 (1845). Lithocardium muneco (HBK.) Kuntze, Rev. Gen. 2: 977 (1891). Borellia asper Rafinesque, Sylva Tellur. 41 (1838). Cordia umbraculifera DeCandolle, Prodr. 9: 484 (1845); Schomburgk, Fauna u. Fl. Brit. Guian. 960 (1848); Fresenius in Martius, Fl. Bras. 8': 16 (1857). Lithocardium umbraculiferum (DC.) Kuntze, Rev. Gen. 2:977 (1891). Tree, 3-12 m. tall; branchlets pallid, tomentose with abundant curved spreading short hairs; leaves homomorphic, ovate to elliptic or oblong or lance-ovate, broadest either below or just above the middle (usually the latter), 1-1.7 dm. long, 5-14 cm. broad, base more or less oblique, obtuse or rounded or subcordate, apex obtusish to acute, the very tip blunted (not acuminate), under surface green, sparsely strigose, sec- ondary venation obscure, lower surface much paler, more or less brown- ish with rather abundant short slender curved hairs which spring from 12 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI the much rebranched veins, usually velvety, with 7-10 pairs of veins; petioles well developed, 2—5 cm. long; cymes usually terminating leafy branchlets, loosely branched, 1-3 dm. broad; calyx obovoid in bud, densely covered with fine appressed hairs, inside sparsely strigose or hispidulous; calyx at anthesis 2.5—3.5 mm. long, with 4-5 more or less equal deltoid lobes, in fruit explanate and 5—7 mm. broad; corolla white, prevailingly 5-merous, 4-5 mm. long, glabrous, lobes elongate, filaments exserted, hairy at base; ovary and style glabrous; fruit glabrous; stone very rugose, transversely compressed-ovoid, ca. 1 cm. long, pulp white and mucilaginous. Northeastern coast of Brazil (Maranhao and Para), northern South America and southward along the Andes to Bolivia; frequently cul- tivated. BritisH GUIANA: Rockstone, banks of the Essequibo, 1921, Gleason 865 (K); Demerara River, May 1889, Jenman 4878 (KK); Demerara, Parker (K, DC); Platburg Creek, Canje River, fruit glutinous, yellowish green, 1914, Hohenkerk 631 (K); indefinite, large tree, flowers vellowish white, 1837, Schomburgk 408 (DC, Tyre of C. wmbraculifera; isotypes, G, By BOP) DutcH GUIANA: upper Nickerie River, Feb. 1915, B. W. 1074 (Utr); near Paramaribo, 1910, native collector (Utr); near Paramaribo, tree, fl. white, 1844, Kappler, ed. Hohenacker 1619 (Utr, P); Plant. Jagtlust, 5 m. tall, 1913, Soeprato 6E (Utr); Plant. Osembo-Onverwacht, 1913, B. W. 6229 (Utr); Plant. Slootwijk, tree 5 m. tall, Soeprato 10H (Utr); Watramiri, tree no. 1568, fruit edible, mucilaginous, June 4, 1916, B. IV. 1836 (Utr); Watramiri, tree no. 1568, Feb. 7, 1917, B. W. 2659 (Utr); Watramiri, tree no. 1568, Feb. 18, 1920, B. W. 4551 (Utr) ; Watramiri, tree no. 1568, Dec. 7, 1920, B. WV. 4974 (Utr) ; Surinam, tree 9-12 m. tall with broad horizontal branches, usually cultivated, fl. white, Nov. 1837, Splitgerber 123 (Leid); Surinam, 1841, Berthoud-Coulon 553 (BM); Surinam, Hostmann 355 (K, BM, BD, Del, P); Surinam, Hostmann (Utr). FRENCH GUIANA: Mana, March 1854, Mélinon 215 (P); Mana, 1857, Sagot (P); Hes du Salut, fruit white, glutinous, 1854, Sagot 445 (K, BM, P); Ile de Cayenne, 1851, Sagot (P); Cayenne, Aublet (BM, Type of C. tetrandra) ; Mahoury near Cayenne, Sagot (P); indefinite, Martin ex herb, Rudge (BM), LePrieur 252 (Del, P), Perrottet (P), Gabriel (Del) and Potteau (K). Aublet reports C. tetrandra from the Ile de. Cayenne and from the mainland of French Guiana. I have examined specimens from his per- sonal herbarium, now at the British Museum, as well as a duplicate from it now in the Swartz herbarium at Stockholm. His description, his illustration, and the two specimens, leave no doubt as to the exact iden- tity of C. tetrandra. Aublet’s name is inapt, the species is practically always pentandrous. 1935] JOHNSTON, STUDIES IN BORAGINACEAE 13 The species appears to be indigeneous only in a broad band of wet tropical forest about the northern margin of South America, where it seems to be most common at low altitudes on the coastal plain. On the east base of the Andes it is known from Peru and Bolivia. I have seen no material from the dryer portions of Brazil to the south of the Amazon Basin that is indubitably from wild plants. Brade, Bol. Mus. Nac. Rio Janeiro 8: 35 (1932), however, has recently reported it from Manaos. The collections by Spruce from the mouth of the Rio Negro, reported (sub C. umbraculifera) in the Flora Brasiliensis, 8': 16 (1857), is not C. tetrandra, but the material subsequently made the type of C. Sprucet Mez. I have reported, Contr. Gray Herb. 92: 55 (1930), C. tetrandra from Ceara and Pernambuco. This was incorrect and is the result of a bad clerical error. The specimens actually represent C. toqueve. I am indebted to Mr. Killip, in lit., and to Mr. Brade, |. c. 34, for the correc- tion of this unfortunate error. The wood of C. tetrandra has been described by Pfeiffer, De Hout- soort. v. Surinam. 1: 444 (1926). In the herbarium at Utrecht the col- lection, cited above, from the Plantation of Osembo-Onverwacht bears the annotation, “‘Pfeiffers Woods of Surinam no. 59.” Aublet states that the species is called “Bois Margarite” and “Arbre a parasol.” The following vernacular names are associated with the specimens cited above, Kakuru or Clammy Cherry—Hohenkerk 631; Kakhoro’ (Arow.), Tafrabom (Nig. Eng.) and Alatoeloeka (Kar.)— B. W. 1074; Tafelboom—B. W. 6229 and Splitgerber 123; Tafelboom (Sur. Dutch), Tafraboom (Nig. Eng.), Boggi lobbi (Saram.), Toen- balobbi (Saram.), Kakhoro (Arow.), Araatroekoe (Kar.)—Watramiri tree no. 1568; Bois parasol—Sagot 445 ; Roquei—Sagot at Cayenne and Mahoury. 5. Cordia nodosa Lamarck, Tab. Encyc. 1: 422 (1791); Poiret, Encyc. 7: 43 (1806); Schomburgk, Fauna u. Fl. Brit. Guian. 960 1848); Fresenius in Martius, Fl. Bras. 8: 16, tab. 5 (1857); Bailey, Bot. Gaz. 77: 32-49, tab. 6-7 (1924); Johnston, Contr. Gray Herb. 92:46 (1930). Lithocardium nodosum (Lam.) Kuntze, Rev. Gen. 2:977 (1891). Cordia hirsuta Willdenow, Sp. Pl. 1: 1076 (1798); Meyer, Prim. Fl. Esseq. 114 (1818). Firensia hirsuta (Willd.) Rafinesque, Sylva Tellur. 40 (1838). C ordia formicarum Hoffmannsegg ex Roemer & Schultes, Syst. 4: 800 (1819). C ordia miranda DeCan- dolle, Prodr. 9: 475 (1845). Lithocardium mirandum (DC.) Kuntze, Rev. Gen. 2:977 (1891). Cordia hispidissima DeCandolle, Prodr. 9: 475 (1845). Lithocardium hispidissimum (DC.) Kuntze, Rev. Gen. 2: 977 (1891). Cordia nodosa var. hispidissima (DC.) Fresenius in 14 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI Martius, Fl. Bras. 8': 17 (1857). Cordia nodosa var. angustifolia Fre- senius in Martius, Fl. Bras. 8': 17 (1857). Cordia umbrosa Spruce ex Rusby, Bull. Torr. Bot. Cl. 26: 147 (1899). Cordia volubilis Pittier, (Explor. Bot. Cuenca de Maracaibo p. 41) Bol. Comer. e Indust. 4: ? (1923); Jour. Wash. Acad. Sci. 19: 184 (1929). Cordia collococa sensu Aublet, Hist. Pl. Guian. Fr. 1: 219, tab. 86 (1775) Shrub or tree, 2-11 m. tall; stems bearing stiff spreading brownish bristles which are usually abundant but may be sparse or nearly absent; the stems below each fork abruptly and asymmetrically enlarged and con- taining a cavity usually serving as an ant-domatium; leaves usually subopposite or whorled, more or less heteromorphic, somewhat lustrous on both surfaces, lanceolate to nearly elliptic, broadest near the middle, 10-35 cm. long, 3-28 cm. broad, base obtuse, apex acuminate, margin entire, upper surface with impressed veins, more or less bullate, with a few hairs along the midrib, lower surface paler, with very scattered bristles on the veins, with 6-10 pairs of veins, these repeatedly re- branched and anastomosing, petiole 2-5 mm. long, bristly; inflorescence cymose-paniculate, loose or dense, 2-10 cm. in diameter, bristly and usually also with minute curly brownish pubescence, borne at the forks of the stem; calyx usually somewhat puberulent and strigose, more or less bristly especially about the apiculate apex, papery in texture and very obscurely ribbed, opening irregularly to form several very irregular lobes, frequently persisting and eventually breaking up into fibers; calyx in the bud ovoid or ellipsoid, ca. 5 mm. long; corolla white, tube 4—6 mm. long, lobes 2~3 mm. long, filaments hairy at base, 3-4 mm. long; style and ovary hairy; fruit usually more or less bristly; stone transversely ovoid, 13-17 mm. long. In British, Dutch and French Guiana and widely distributed in the Amazon Basin; also in the headwaters of the Orinoco (in southern Vene- zuela and eastern Colombia) and in northwestern Venezuela. Britisu GUIANA: Amakura River, 5 m. tall, March 1923, La Cruz 3430 (G); Barima River, March 1896, Jenman 7055 (KK); Kamakusa, upper Mazaruni, ca. long. 59° 50’, 1922-23, La us 2887 and 4231 (G); Macouria River, Nov. 1886, Jenman 2391 and 2392 (K); Tumatumari, dense upland forest, shrub 2.5-6 m. tall, 1921, api 311 (WK); Kaieteur Falls, Potaro River, 1923, La Crus 4407 (G); island in Cuyuni River below Kamaria Falls, 18 dm. tall, 1920, Bailey iy (G); Kartabo region, second growth forests, 1920, Bailey 29, 41 and 42 (G); Bonasika Creek, at sea-level, Anderson 66 (K); Moraballi Creek, small tree up to 6 m. tall, in low brush and clearings in mixed forest, fl. white, fruit bristly, becoming pale red, Aug. 15, 1929, apes me - wel aera dense upland forest, 2.5-3 m. tall, 1921, Gleason 583 (K); e Mts., en 1erara, fruit red, ‘hairy, sweet and asaiay: an oe eee 272 (K); Demerara 1935] JOHNSTON, STUDIES IN BORAGINACEAE 15 River, May 1889, Jenman 4924 (K); Malaroo Creek, Corantyne River, small tree, 3-6 m. tall, Oct. 1879, im Thurn (K, P); indefinite, Schom- burgk 904 (K) ae 984 (K, BM, BD, P). Dutcu GuIANaA: Kaboeri Reserve, Corantyne River, tree no. 684, Nov. 1920 and Aug. 1922, B. W. 4835 and 5986 (Utr); mouth of Lucie River, Corantyne River, 1910, Hulk 315 (Utr); way to Kwatta, Paramaribo, June 1916, Samuels 237 (G, Leid, BD, P); Station at Groningen, forest, May 1916, Samuels 123 (G, K, Leid, BD, P); Watramiri reserve, Sara- macca River, June 1918, B. WW. 3864 (Utr); Watramiri reserve, tree no. 1652, mature fruit yellow, soft and sweet, Dec. 1916, B. W. 2488 (Utr) ; Watramiri, tree no. 1652, used for tea, May, 1916, B. W. 1911 (Utr); Watramiri reserve, tree no. 1652, April 1917, Oct. 1917, July 1918, and Feb. 1920, B. W. nos. 2756, 3309, 3872 and 4541 (Utr) ; Watramiri reserve, tree no. 1652, fl. white, B. W. 4012 (Utr); Watramiri reserve, tree no. 1652, March 1919, ripe fruit sordid white, B. W. 4301 (Utr); Sectie O. reserve, upper Para River, tree no. 800, fl. light green, leaves used for tea, Aug. 1916, B. W. 2306 (Utr); woods near Poelebantji, tree 4-6 m. tall, Feb. 1845, nce 691 (Utr) ; Brownsberg, Surinam River, tree 10 m. tall, trunk 1 dm. thick, fl. sordid white, Sept. 1915, B. W. 727 (Utr) ; Browns- berg reserve, tree no. 1174, fl. white, dried leaves used as a substitute for tea, fruit yellow, globose, soft and juicy, Nov. 1916, B. W. 2498 (Utr); Brownsberg reserve, tree no. 1174, fl. white, Sept. 1918, B. W. 4002 Utr); Brownsberg reserve, tree no. 1174, March 1917, Feb. 1919 and March 1921, B. W. nos. 2721, 4265 and 5075 (Utr) ; Brownsberg reserve, tree no. 1174, fl. sordid white, with strong odor, Sept. 1923, B. W. 6227 (Utr); Brownsberg reserve, tree no. 1174, fl. sordid white, odor strong, Nov. 1924, B. W. 6684 (Utr); Brownsberg summit, July 1924, fl. white, B. W. nos. 6634 and 6722 (Utr); woods near Raleigh Falls, Coppename River, hispid tree, fl. yellowish white, fruit white, hispid, Sept. 11, 1933, Lanjouw 788 (Utr). FRENCH GUIANA: Maroni, 1864, Mélinon (G, Del, P); Ile Portal, Maroni River, Sagot (P); Ile Portal, fruit white, soft, size of a grape, June 1857, Sagot 446 (P); Acarouani, tree, corolla pale yellowish, sepals 4, stamens 4, Oct. 1854, Sagot 446 (P); Acarouani, 1854, Sagot (P); Acarouani, fruit red, April 1858, Sagot 446 (P); Acarouani, 1859, Sagot P); Acarouani, fruit pale yellow, 1854, Sagot 446 (P); Godebert, Wachenheim 410 (P); in loco Macaya ad praedicem Patuis, Richard (P) Saget (FP); ‘Cayenne, Patris (BM, Del); indefinite, 1850, Leprieur (BM, P); indefinite, Aublet (BM, type of C. nodosa) ; indefinite, Perrottet 214 (Del, DC) and Poitteau (K, BD, P). BraziL: Carmo, Rio Branco, Sept. 1, 1924, Bequaert (G); Surumu, Serra do Mairary, Rio Branco, tree or shrub 2-8 m. tall, fl. white, Nov. 1909, Ule 8456 (K, BD); Rio Negro below mouth of Xibaru, betw. Bar- cellos and Sao Gabriel, Dec. 1854 pies 3790 (NY, K, Del) ; Rio Cumina, Sampaio 5136 and 5148 (BD); r Montalegre, Nov. 24, 1873, Traill 561 (K); Prainha, Dec. 17, 1873, "Trail 562 (K). 16 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI VENEZUELA: Casiquiari, in shade along streams, fl. white, tree 6 m. tall, Jan. 1853, Spruce 3281 (G). This is a classic ant-plant. The results of an anatomical study and a review of the more important literature on this plant have been published recently by I. W. Bailey, Bot. Gaz. 77: 32—49, tab. 6-7 (1924). Accord- ing to this author the peculiar subnodal structures serving as ant-domatia are “formed by an invagination of epidermal, cortical, and fibrovascular tissues which originate in the axil of one of the leaves of the false verti- cil, and which develops into the interior of a more or less symmetrical or unilateral, subnodal enlargement of the cauline axis.” The attention drawn by the complex ant-domatia has, I believe, blinded students to the evident relationships of this remarkable plant. The flowers and fruit are very similar to those found in C. Sprucei and its relatives. The apiculate, papery, irregularly disrupted calyx, the hairy ovary and style, and the transversely ovoid stone, not to mention the glabrous upper leaf-surfaces, comparatively stiff and contracted inflorescence, etc., all indicate close relations with that group of upper Amazon and Guianan species. Cordia nodosa has only three notable peculiarities, its subnodal swellings, its bristly indument, and its fibrous calyx. In the past the species has been placed in a special section, Physoclada, of the genus Cordia. I have become so impressed with its obvious relations with C. Sprucei and allies, however, that I now am quite content to associate it with these species in the section Pilicordia. The species is very variable both in the size and shape of its leaves, and in the abundance of the bristles on its herbage. This variation seems to be ecological in origin. In any case I can find no evidence that it is in any way geographically correlated. It should be noted that collectors have given the fruit in British Guiana as red. In Dutch and French Guiana the fruit is given as white or whitish in numerous cases, and once as yellow. Aublet gives the Carib name for the plant as ‘‘Achira-mourou.”’ The following vernacular names are associated with specimens cited: Coura- belli ants plant—Anderson 66; Yluri-hee-lévi-kot—Parker 272; Hurneyreyroko—Sandwith 12; Awelemoeloe (Kar.)—B. W. 727; Marribonsoehoedoe (Neg. Eng.), Horowejoreroko (Arow.), Arreuonoe (Kar.)—Tree no. 1652 at Watramiri; Mattoe toenbalobbi (Sar.), Horowé, joee lokko, Hoereuereroko (Arow.), and Awali emoeloe, Aloeko uonoré (Kar.)—Tree 800 at Sectie O. 6. Cordia laevifrons, sp. nov. Arbor minor vel frutex, dichotome ramosus; ramulis fuscis, apicem versus dense puberulis mox glabrescentibus; foliis vix crassis ellipticis 1935] JOHNSTON, STUDIES IN BORAGINACEAE 17 ad lanceolato-oblongis vel oblongo-obovatis 12-25 cm. longis 6-14 cm. latis saepe ad medium vel supra medium latioribus, basi obtusis vel plus minusve rotundis vel late acutis, apice saepe abrupte acuminatis, supra lucentibus saepe in costa pilos paucos adpressos gerentibus ceteris glaberrimis, subtus pallidioribus glabris vel sparsissime minutissimeque ascendenter adpresseque pubescentibus, nervis primariis 6—8-jugatis, nervis tertiariis obscuris, petiolis 5-18 mm. longis; cymis saepissime in furcis ramulorum ortis, laxe graciliterque ramosis 3-15 cm. diametro, pedunculo gracili; corolla alba glabra, tubo 6 mm. longo calycem superante, lobis 2 mm. longis rotundis latis, filamentis 4 mm. longis longe exsertis basim versus pilosis; calyce in alabastro anguste obovato 4-5 mm. longo extus dense puberulento (intus subglabro) obscure lateque 10-costato, apice plus minusve apiculato ad anthesin in lobos irregulares lacerulatos disrupto, fructifero explanato; stylo et apice ovarii sparse minuteque hispidulo; fructu glaberrimo; nuce transverse ovoideo 10-14 mm. longo. Endemic to French and Dutch Guiana. DutcH GuiIANA: Lucie River, a small tree 6 m. tall, fl. white, April 12 1926, B. W. 6999 (Utr); forest near Abontjeman, May 1910, native fob lector 236 (TyprE, Utrecht). FRENCH GUIANA: Maroni River, 130 km. upstream, fruit edible, 1877, Crevaux (P); along the Maroni, 1861, Mélinon 16, 59, 254 and 271 (P); along the Maroni, 1863, Mélinon 283 (P); Maroni, aon road to St. Lau- rent, clearings, 15 dm. tall, fl. white, Oct. 1876, Mélinon 225 (P); St. Jean, 2 m. tall, fl. white, May 16, 1914, Benoist 1230 (P). A relative of C. Sprucei notable chiefly for its rather thin, nearly glabrous leaves, puberulent obscurely ribbed calyces, and quite glabrous fruits. As with other relatives of C. Sprucei the veins of the leaves are not so finely rebranched as is common in this section of the genus. Crevaux gives the bush-negro name of the plant as ‘Tiki Topichi.” 7. Cordia Sprucei Mez, Bot. Jahrb. 12: 549 (1890); Johnston, Contr. Gray Herb. 92:53 (1930). Lithocardium Sprucei (Mez) Kuntze, Rev. Gen. 2: 977 (1891). Tree 4-5 m. tall, branching dichotomous; branchlets dark, sparsely short-hirsute or with short incurved hairs; leaves drying brown, sub- homomorphic, elliptic or oblong-obovate, 12-23 cm. long, 6—-11.5 cm. broad, broadest at or above the middle, apex acuminate, base acute to truncate, upper surface somewhat lustrous, hairy along the midrib and with scattered hairs along the principal veins but otherwise glabrous, lower surface with rather abundant short soft erect hairs, petiole 5-10 mm. long, veins in 6-9 pairs, tertiary veins tending to be obscure; in- 18 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI florescence rather loosely though rigidly branched, ca. 1 dm. thick, usually borne at the forks of the stem, rarely terminal; calyx well cov- ered with short incurving tawny hairs, obscurely ribbed, obovate in the bud and apiculate, 4-5 mm. long, ca. 2.5 mm. thick, bursting rather irregularly at the apex into 2—5 broad rather thin teeth; corolla white, tube 4-5 mm. long, lobes broad ca. 1.5 mm. long, filaments hairy at base, 4 mm. long; ovary densely hairy above the middle; fruit yellow, strigose; stone transversely ellipsoid, 1-1.5 cm. long. Known only from the Rio Negro of Brazil and from French Guiana. FRENCH GUIANA: “in Sylvis doeciduis Fluvii Kourou, ad casam indi Felix,” Nov., Richard (P). BraAziL: Barra do Rio Negro, 1850-51, Spruce 1019 prey TYPE; BD, frag; G, photo.) ; vicinity of Barra, 1850-51, Spruce (G, K, BM); Barra to Matiriho, Jan. 1851, Spruce 1234 (K, BM); Barra, fruit yellow, transversely oblong, April 1851, Spruce 1234 (K, BM, Del) ; Sao Gabriel, Rio Negro, ca. 90 m. alt., 1930-31, Holt & Blake 608 (G). Richard’s label gives the following field data concerning his collection from French Guiana,—‘‘frutex 3—4 ped., ramis diffuse patentibus, dicho- tomis; fl. albidi; ramillis cymae recurvis et, inexpansis floribus, revo- lutis.” The collections are remarkably similar to Spruce’s material from the lower Rio Negro, except that in one of Richard’s two sheets the branches of the inflorescence are somewhat tawny tomentulose. Since discussing the type of C. Sprucei, 1. c., I have examined the actual type-specimen at Munich. The specimen has the following familiar printed label reading, “In vicinibus Barra, Prov. Rio Negro, coll. R. Spruce, Dec.-March, 1850-51.” The collector’s number, in script, is “1019.” The specimen is that cited under C. umbraculifera in the Flora Brasiliensis. It is evidently part of the same collection as the unnumbered specimens I have seen at the Gray Herbarium, at Kew, and at the British Museum. This species not only has relatives in C. nervosa and C. laevifrons of the Guianas but also in undescribed trees of the Putumayo and the Huallaga of eastern Peru. The glabrous upper leaf-surfaces and the somewhat papery irregularly disrupted apiculate calyces are characters of this group of species. 8. Cordia nervosa Lamarck, Tab. Encyc. 1: 422 (1791); Poiret, Encyc. 7: 47 (1806); DeCandolle, Prodr. 9: 484 (1845). Lithocardium nervosum (Lam.) Kuntze, Rev. Gen. 2: 977 (1891). Cordia calophylla Vahl, Ecolog. 3: 5 (1807); DeCandolle, Prodr. 9: 486 (1845). Litho- cardium calophyllum (Vahl) Kuntze, Rev. Gen. 2: 976 (1891). Shrub or small tree, up to 5 m. tall; branchlets closely and antrorsely strigose; leaves homomorphic, stiff and coriaceous, with an arcuate mid- 1935] JOHNSTON, STUDIES IN BORAGINACEAE 19 rib, broadly lanceolate to elliptic or lance-oblong, 10-25 cm. long, 4-10 cm. broad, margin weakly recurved, apex acuminate, base rounded to acute and usually more or less oblique and asymmetrical, upper surface glossy, smooth and quite glabrous, lower surface drying brown, dull, somewhat scabrous with abundant short inconspicuous hairs, with 8-10 pairs of primary veins, these connected by simple branches, the secondary branches of the veins absent or very obscure; petiole canaliculate, stiff, 5-10 mm. long; inflorescence small and compact, 1-4 cm. long, peduncles very short or none, branches slender, strictly forked, bearing flowers on only one side and in age studded with the elevated pedicellar flower- attachments, becoming rigid and woody in age and more or less spread- ing or deflexed, persistent long after the falling of the fruit; calyx obo- void in bud, 4—5 mm. long, minutely short-strigose, more or less apicu- late, not at all ribbed, sparsely strigose inside, bursting apically and the lobes torn and irregular, in fruit explanate; corolla white, tube ca. 5 mm. long, lobes broad, ca. 3 mm. long, filaments very hairy; ovary glabrous or sparsely hairy towards the apex; style usually sparsely hairy; fruit minutely and abundantly strigose, pulp bright red, glutinous, insipid; stone transversely ovoid, 10-13 mm. long. French Guiana and adjacent Brazil; British Guiana. BritisH GUIANA: Kaieteur Savanna, spreading shrub 18 dm. tall, 1881, Jenman 1062 (K). FRENCH GUIANA: Cayenne, 1857, Mélinon (P); Gourdonville, Kourou River, shrub, fl. white, Sept. 25, 1914, Benoist 1618 (P) ; in umbrosis sylvis praedii Dm. [?] Patuis, Richard (P); indefinite, herb. Lamarck (Paris, TYPE of C. nervosa); indefinite, von Rohr 152 (herb. Vahl, type of C. calophylla; BM, isotype); indefinite, 1859, Leprieur (Del); indefinite, 1819-21, Poiteau (K, Del). BraziL: Counany, Oct. 13, 1895, ““Chapeo del Sol,” Huber 1032 (Boiss). The type of C. nervosa in the Lamarck herbarium is so very similar to the material collected by Richard (in the General Herbarium at Paris) that I believe they are parts of a single collection or, in other words, that the type of C. nervosa was collected. by Richard. Unfortunately, I have been unable to identify Richard’s locality with any degree of con- fidence. At Paris I found on the label of a very different species the following more explicit mention of the probable locality, i. e. “in loco Dm. [spelling?] Macaya ad praedicem Dm. [?] Patuis.” There was formerly a sugar plantation called Macaya on the Ile de Cayenne sev- eral kilometers east of Matoury. This may have been that referred to by Richard, for he is known to have collected extensively about the Island of Cayenne as well as over most of the French Guianan coastal region, 20 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XVI It is interesting to note that Richard has appended to his specimen a manuscript name, under Collococcus, in which the same specific epithet is used as was subsequently published by Vahl. Richard and von Rohr were both in the Guianas about 1785. These facts naturally make one wonder if there may not have been some meeting or some exchange of material between these two botanists and possibly if Vahl’s type may not have had the same source as that of Lamarck. The species is an unusually distinct one, being notable because of its suppressed tertiary leaf-veins and small dense subsessile inflorescence. Its closest relations are with C. Sprucei and C. laevifrons which have similar somewhat papery, irregularly dehiscent apiculate calyces. In C. Sprucei the veins are more repeatedly branched than in C. nervosa, though generally less so than in other species of the section Pilicordia. 9. Cordia fulva, sp. nov. Arbor vel frutex, dichotome ramosus; ramulis brunneis cum pilis brevibus divergentibus abundantissimis velutinis; foliis subhomomor- phis crassiusculis ellipticis vel ovatis 10-22 cm. longis 7-11 cm. latis saepe ad medium vel supra medium latioribus, basi obtusis vel rotundis vel rariter acutis, apice breviter acuminatis, supra scabris pilis brevibus rigidusculis ascendentibus vestitis, subtus saepe brunnescentibus velu- tinis in nervis et nervulis elevatis et numerosis pilos graciles erectos abundantes gerentibus, nervis 5—7-jugatis, petiolis brevibus; cymis in furcis ramulorum ortis vel rariter terminalibus, rigidis brunneo-velu- tinis laxe ramosis, floribus in ramulis plus minusve congestis; calyce in alabastro obovato 4-5 mm. longo 2—3 mm. crasso evidenter 10-costato extus brunneo-velutino intus supra medium strigoso; lobis 5 deltoideis; corolla alba, tubo 5 mm. longo, lobis ca. 2.5 mm. longis et latis, filamentis 4-5 mm. longis basim versus pilosis; ovario apicem versus dense longeque pilosis; fructu evidenter velutino; nuce transverse ovoideo 1—1.5 cm. longo. Known only from northern Dutch Guiana and adjacent French Guiana. DutcH GuIANA: near Abontjeman, in forest, May 1910, native col- lector 227 (Utr); near Gold Placers, April 14, 1910, native collector 103 Utr). FreNcH GuIANA: Maroni, shrub 3 m. tall, in clearings, fl. white, branches horizontal, 1877, Mélinon 137 (typrE, Gray Herb.; isotype, Paris) ; Maroni, Mélinon 455 (G, P); Maroni, Wachenheim 75 (G, P); indefinite, 1862, Mélinon 82 (P); indefinite, 1821, Perrottet (P). Related to C. trichoclada DC. and C. Chamissoniana Don, of eastern Brazil, this species differs from the former in its velvety rather than 1935] JOHNSTON, STUDIES IN BORAGINACEAE 21 bristly stems, more softly hairy scarcely bullate leaves, looser less stiffly branched larger inflorescences and slightly smaller calyces, and from C. Chamissoniana in its more hairy leaves and conspicuously ribbed, more tawny calyces. 10. Cordia toqueve Aublet, Hist. Pl. Guian. Fr. 1: 228, tab. 90 (1775); Poiret, Encyc. 7:44 (1806); DeCandolle, Prodr. 9: 488 (1845); Johnston, Contr. Gray Herb. 92: 52 (1930). Lithocardium toqueve (Aubl.) Kuntze, Rev. Gen. 2: 977 (1891). Cordia hetero- phylla Poiret, Dict. Sci. Nat. 10: 409 (1818); Willdenow ex Roemer & Schultes, Syst. 4: 800 (1819); Chamisso, Linnaea 4: 480 (1829); DeCandolle, Prodr. 9: 487 (1845). Lithocardium heterophyllum (Poir.) Kuntze, Rev. Gen. 2: 977 (1891). Cordia pubescens Will- denow ex Roemer & Schultes, Syst. 4: 800 (1819). Lithocardium pubescens (Willd.) Kuntze, Rev. Gen. 2: 977 (1891). Toquera to- mentosa Rafinesque, Sylva Tellur. 40 (1838). Cordia hebecarpa DeCandolle, Prodr. 9: 488 (1845). Lithocardium hebecarpum (DC.) Kuntze, Rev. Gen. 2: 977 (1891). Tree; branchlets velvety with abundant soft more or less curled brown hairs; leaves strongly dimorphic, upper surface with rather abund- ant short straight ascending hairs, lower surface velvety with curved soft slender spreading hairs from the prominent and numerous veins and veinlets; larger principal leaves very broadly lanceolate to lance- ovate, broadest towards the base, 1-3 dm. long, 6-15 cm. broad, above the middle contracted to an acute or acuminate apex, base obtuse to rounded, somewhat oblique; smaller sort of leaves more or less orbicular, 5-12 cm. long, 5-11 cm. broad, broadly obtuse or even subcordate at base, rounded or acuminate at apex; inflorescence loosely and slenderly branched, 1—1.5 dm. broad; calyx obovoid in bud, 2.5—-3 mm. long, open- ing by 5 triangular lobes, unribbed, strigose inside, outside covered with a dense indument of appressed slender curved hairs; calyx becoming somewhat cupulate at maturity, ca. 1.5 mm. deep; corolla white, tube 2.5 mm. long, lobes 1.5 mm. long, filaments 2 mm. long, hairy near base; ovary densely hairy at apex; fruit abundantly tawny-strigose, style- base forming a short eccentric beak; stone ovoid, strictly ascending, ca. 1 cm. long. Confined to French Guiana and eastern Brazil. FRENCH GUIANA: vicinity of Cayenne, small tree, fruit yellowish, May 16, 1921, Broadway 201 (G); near Cayenne, fl. yellow, 1897, Soubiron (P); near Cayenne, July 1841, Mélinon 243 (Leid, P); Cayenne, Feb. 1859, Sagot (P); Cayenne, Martin (K); Cayenne, Leblond, ex Mus. Paris 348 (BD, P); Cayenne, ex Mus. Paris sine no. (G, BD); Cayenne, 22 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Herb, Willd. sub no. 4574 (BD, type of C. heterophylla Willd.) ; Cayenne, Herb. Poiret (P, TyPE of C. heterophylla Poir.) ; indefinite, Aublet (BM, TYPE of C. toqueve); indefinite, Poiteau (K); indefinite, 1859, Leprieur (Del). The original material of this species was collected by Aublet in clear- ings made by the Caribs about 15 leagues up the Sinnamary River. The tree was called ‘“Toquévé” by these inhabitants of French Guiana. I have seen Aublet’s specimen at the British Museum. This material, taken in conjunction with Aublet’s illustration and lengthy description, leaves no doubt as to the correctness of the present application of the name. At Paris among Poiret’s specimens (in the Cosson collections) there is a fragmentary specimen of the present species accompanied by a label in Poiret’s script reading, Cordia heterophylla, folia altera majora et minora, rami asperi, hirti. Caienne. Added to the label in another, and unrecognized hand is “dict. des Sc. nat. herb. Poiret.”’ Poiret stated that Cordia heterophylla was seen in the Desfontaines herbarium. The specimen in the Poiret collections at Paris is, I believe, a fragment of the type of C. heterophylla Poir., now probably conserved at Florence. The binomial “C. heterophylla” is found on a specimen of C. toqueve in Willdenow’s herbarium at Berlin and was published by Roemer & Schultes a year after Poiret’s published use of the name. The specimen is also given as from “Cayenne” and may be a duplicate of the specimen described by Poiret. Cordia toqueve is readily distinguished among the South American species by its tawny velvety indument, strikingly dimorphic leaves, and conspicuously hairy fruit. It is known only from Bahia, Ceara and Pernambuco in Brazil, and from near Cayenne in French Guiana. The Brazilian plant, which is quite like that from the Guianas, has been de- scribed as C. hebecarpa DC. 11. Cordia scabrifolia A. DeCandolle, Prodr. 9: 485 (1845); John- ston, Contr. Gray Herb. 92: 53 (1930); Brade, Bol. Mus. Nac. Rio Janeiro 8: 34 (1932), Tree or shrub up to 15 m. tall, branching dichotomous; branchlets dark, abundantly and minutely antrorse-strigose; leaves homomorphic, ovate- to oblong-lanceolate, 11-18 cm. long, 4-7 cm. broad, broadest near the middle, apex acuminate, base acute to somewhat rounded; upper surface drying dark, abundantly and very minutely antrorse-strigose, lower surface drying light, bearing numerous very minute very short appressed hairs on the abundant veins and veinlets, the hairs tending to be directed centripetally towards the middle of the veinlet-areoles, 1935] JOHNSTON, STUDIES IN BORAGINACEAE 23 petiole 4-9 mm. long, veins in 5-8 pairs, repeatedly rebranched; in- florescence usually borne at the forks of the stems, pedunculate, slender, loosely branched, 4-10 cm. broad; calyx obovoid in the bud, ca. 3 mm. long, densely strigose, opening by 5 subequal triangular lobes, strigose inside, base substipitate; corolla white, tube ca. 3 mm. long, lobes ovate, ca. 2 mm. long, filaments exserted, hairy at base; style and at least the apex of the ovary hairy; fruit densely and minutely strigose; stone ovoid, ascending, ca. 1 cm. long. Probably restricted to the Amazon Basin; doubtfully reported from British Guiana. BRITISH GUIANA: indefinite, Schomburgk 911 (Boiss, TYPE; isotypes, K, BM, BD, P). Brazii: Prainha, Noy. 1873, Traill 563 (K); Rio Cumina, Oct.-Nov. 1928, Sampaio 5367, 5505 and 5510 (BD); Rio Negro-gap6 above Cabu- quena, Dec. 1851, Spruce 1942 (K, BM). Except for the type-collection, Schomburgk 911, which is given as from British Guiana, all known collections of C. scabrifolia come only from within the Amazon Basin. I suspect that the type also came from the Amazon watershed and from what is now Brazil. Schomburgk, Fauna u. Fl. Brit. Guiana 960 (1848), reports the species from the upper Essequibo. It is significant, however, that no other collectors have found it in British Guiana and that, in the notes of Robert Schom- burgk, plant no. 911, which certainly seems to apply to this species, refers to a collection almost certainly from the Rio Negro watershed. In the list no. 911 has no locality indicated but the adjacent numbers, where the locality is indicated in several instances, all do come from Barcellos on the Rio Negro. 12. Cordia bicolor A. DeCandolle, Prodr. 9: 485 (1845); Pulle, Enum. Pl. Surinam 397 (1906). Lithocardium bicolor (A. D Kuntze, Rev. Gen. 2: 976 (1891). Cordia dichotoma Klotzsch ex Schomburgk, Fauna u. Fl. Brit. Guiana 1084 (1848), nomen; not Forst. (1797). Lithocardium Lockartii Kuntze, Rev. Gen. 2: 438 (1891). Cordia Lockartii Kuntze, Rev. Gen. 2: 438 (1891), in synonymy. Cordia trichostyla Pittier, Contr. U. S. Nat. Herb. 18: 252, fig. 102 (1920). Cordia carnosa Rusby, Three Hundred N. Sp. So. Amer. PI. 104 (1920). Cordia coriacea Killip, Jour. Wash. Acad. Sci. 17: 329 (1927). Cordia sericicalyx sensu Johnston, Contr. Gray Herb. 92: 54 (1930). Shrub or small tree up to 6 m. tall, branching dichotomous; branch- lets angulate, velvety with very abundant spreading usually tawny short hairs; leaves homomorphic, ovate to more or less broadly lanceolate, 24 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI broadest at or below the middle, 8-16 cm. long, 2.5—7 cm. broad, apex acuminate, base acute to rounded, upper surface dark, finely strigose, lower surface very pale, covered by appressed slender hairs that are borne on the veins and veinlets and which converge over and cover the veinlet-areoles; veins in 5—7 pairs, repeatedly rebranched; petiole 5—10 mm. long; inflorescence usually at the forks of the stem, loosely branched; calyx in bud ca. 4 mm. long, obovoid, clothed with abundant short appressed more or less tawny hairs, inner surface with numerous appressed longer white hairs, opening by 5 subequal triangular lobes; corolla white, tube ca. 3 mm. long, lobes ca. 2.5 mm. long, filaments hairy near base; style and upper part of ovary hairy or these exception- ally glabrous; fruit glabrescent; stone transversely ovoid, over 1 cm. long. Occurring in the Amazon headwaters of Bolivia; east and south of the Amazon Basin in Brazil; and across northern South America from Dutch Guiana to eastern Colombia; apparently very sporadic in occur- rence, widely distributed but not common. Also in Central America and in the southern-most West Indies. BRITISH GUIANA: Oreala, Corantyne River, Oct. 1879, Jenman 7 (K); indefinite, N. Cos As Boe 0th ee “sat Anders VoL. XVI. ARB. ARNOLD N. Jour LOTIDES VAI EC SoOPHORA ALOP FE. > SOPHORA PRODANII \F ( LEAFLET 1935] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 81 SUPPLEMENT TO THE SPONTANEOUS FLORA OF THE ARNOLD ARBORETUM ERNEST J. PALMER SINCE THE PUBLICATION in 1930 of the list of plants growing spon- taneously in the Arboretum’ observation and collecting has been con- tinued, and as a result so many additions have been made to the flora that it now seems desirable to publish a supplementary list. In 1931 I had an opportunity for the first time to remain at the Arboretum throughout the spring months and to make a thorough ex- ploration of the native and introduced plants at that season. As a result of this and of subsequent investigation a large number of plants not recorded in the first list have been found and additional information about some of the rarer species previously recorded has also been secured. Of the 173 new species and varieties reported in this supplement 94, or nearly 55 percent, are plants native in the Boston area and presum- ably in the Arboretum, and the remaining 79, or 45 percent, are intro- duced. Seven plant families and 43 genera are added to the spon- taneous flora in this supplementary list. The grasses, sedges, and composites, as might be expected, furnish the largest number of additions. The really surprising thing about the present list is the relatively large number of native plants that have been found. Many of these are now quite scarce or rare in the Arboretum. Several of the introduced plants are probably recent introductions. Amongst the more interesting discoveries are: Polystichum acrosti- choides, Aristida dichotoma, A. gracilis, Carex communis, C. Goode- nowti, C. longirostris, Erythronium americanum, Luzula nemorosa, Cypripedium acaule, < Quercus Rehderi, Anemone quinquefolia, Poten- tilla canadensis var. villosissima, Polygala sanguinea, Lechea intermedia, L. tenuifolia, Viola pedata var. lineariloba, V. sagittata, Pyrola amert- cana, Trientalis americana, Epifagus virginiana, Houstonia caerulea, Liatris scariosa, Aster acuminatus, Helenium nudiflorum, Senecio aureus, Hypochaeris radicata, Sonchus arvensis var. glabrescens, Hieracium florentinum, and H. vulgatum. A single weak plant of the Christmas fern was discovered in 1931 by Dr. Grant D. Darker on a wooded slope of the North Woods, where it had probably survived from a native colony. Later, sev- 1Jour. ARNOLD Ars. 11: 63-119 (1930). 82 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI eral plants were found on the north side of Hemlock Hill. At a little lower level in the latter locality ‘a small group of the stemless lady’s slipper was found growing under the hemlock and pine trees. The star flower is also found here, as well as on the top of the hill and as a greater rarity along the base of a gravelly ridge in the North Woods. The little bluets or innocence grows sparingly among the laurels and other shrubs at the foot of the hill, and a little higher up is found the round-leaved wintergreen. As the hemlock grove on this rocky elevation is probably the only bit of practically virgin timber left in the Arbore- tum and as the native flora has been little disturbed even in a few spots along the base of the hill where there is an accumulation of richer soil, a number of interesting plants have been able to survive here that at present are not found elsewhere. The March lily has not yet been seen flowering in the Arboretum, but a small colony of sterile plants comes up each year in a moist shady spot on the south side of Hemlock Hill. The plants apparently lack sufficient vitality to produce blossoms, probably due to the increased shade. A colony of plants blooming freely is found just outside the Arboretum area on a wooded bank in the grounds of the Adams Nervine Hospital, and only a few yards from the division fence. The wood anemone survives sparingly at the base of wooded hills and even on open banks that were cleared of native trees only in recent years. It has been found at the edge of the North Woods, on the slopes of Bussey Hill in the oak group, along a bank of Bussey Brook below the junipers, and in the South Woods near Peters Hill. The purple milkwort grows on an open grassy bank on the east side of Peters Hill, and near the same spot a single plant of the bird-foot violet was found. Several plants of the latter were also found in open rocky woods on the top of Hemlock Hill, but they are likely soon to be exterminated by careless picking and trampling. A specimen of the beech-drops was collected on the south side of Hemlock Hill, but it has not been observed elsewhere in the Arboretum. The hairy bush-clover and the pinweed (Lechea tenui- folia) grow together in dry gravelly or rocky soil at the edge of the Central Woods, on the north slope of Bussey Hill, and near a small aban- doned quarry in the South Woods. Lechea intermedia is also found near the top of Peters Hill. Amongst the native woody plants that have been added to the list the low juniper is one of the rarest. Two or three small plants of this are growing on conglomerate outcrops in the Central Woods, where they are probably indigenous. The scrub chestnut oak grows sparingly near the same spot as well as on top of Hemlock Hill, where a single plant was seen. At this locality in the Central Woods was also found the inter- 1935] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 83 esting natural hybrid between the bear oak and the black oak (Quercus Rehderi) growing with both of the parent species. The choke berry is another native shrub found in the Central Woods and a few plants of it have survived the repeated mowings of the grass on the northeast slope of Peters Hill. The spice-bush and high-bush blueberry grow in lower and richer ground at a few places on the borders of the woods. It is gratifying to find evidence that a second species of thorn was native in the Arboretum. Specimens of Crataegus rotundifolia collected many years ago by Mr. C. E. Faxon and others in Bussey Wood and on Peters Hill were found in the herbarium, and sprouts of this species were found to be still growing at the latter locality. Mr. Faxon also col- lected specimens of the native blackberries and other plants in the Arboretum, some of which are preserved in this herbarium and others at the Gray Herbarium or in that of the New England Botanical Club. It is interesting to note the introduction of new plants appearing spontaneously in the Arboretum and how they succeed and spread or fail to establish themselves and disappear after a single season or in a few years. Before the publication of the original list a few leaves of what appeared to be a sterile plant of Senecio aureus were seen between the Linden group and the bridle path. In 1931 a vigorous colony came up here and bloomed freely, making a conspicuous show. The following year only two or three plants remained and a search during the past summer failed to reveal any trace of it. A few plants of this species have also been found in the poplar group near Peters Hill. The king devil (Hieracium florentinum) has also recently appeared there, but in greater abundance farther up on the slopes of the hill among the thorns. This year it was also noted in the Celtis group near the North Woods. Fool’s parsley has become more common at two localities, at the foot of the hills near the Leitneria group and along the base of Bussey bank near the Forsythia planting. A vigorous plant of the blue weed came up along the bridle path opposite the Horsechestnut group last year. It was blooming freely when cut down by the mowers, but this year no trace of it could be found. Two weedy grasses, Eleusine indica (the goose grass) and Eragrostis cilianensis have recently appeared on dumps and in waste ground at the old quarry along Bussey Street, and the latter also in the South Street nursery. The flower-of-an-hour, Japa- nese knotweed, four-o-clock, Cyperus esculentus and other cultivated and weedy things have also turned up here, and this and the South Street tract continue to be the chief sources for plants of this class. At the latter place, where a considerable tract of low fertile land surround- ing the pond is still unoccupied and grown up with weeds, a real plant succession has been taking place. A number of plants that appeared 84 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI here soon after the construction of the pond have already been crowded out by the more aggressive weeds but other immigrants arrive from time to time. Last year Boltonia asteroides and Galium asprellum were noted here for the first time, and the latter at least has become more abundant. Last year a large colony of the smooth form of the perennial sow-thistle made a conspicuous show with its large yellow flowers. The yellow Canada lily also sent up a number of tall spikes above the other weeds at one side of the tract. On and about the rubbish dumps here several other weeds as well as escapes from cultivation have appeared. Amongst them are the gourd and jimson weed, as well as another species of thorn-apple, Datura Metel. A number of herbaceous plants, as well as a few shrubs and trees, persist in the Arboretum from the old gardens formerly planted here, and some of these appear to be holding their own or increasing in num- ber. The Virginia spiderwort and ox-eye have spread into the meadow near the old Dawson House, and Scilla, tulips and crocuses of various colors spring up in the grass each year making a pretty display. A bank near the barn of the State Laboratory is also carpeted with the brilliant blue of the Scilla blossoms in spring, and it is found more sparingly in other localities. The star of Bethlehem, day lily, narcissus and European bellflower are all well established in different parts of the Arboretum. More restricted are the white-flowered form of Campanula persicifolia, which is growing along the lilac border and in the open ground on the east side of Bussey Hill, and the English violet, abundant but local along a bank near the Jamaica Plain gate. A few plants of a small perennial pea (Lathyrus pannonicus var. versicolor) come up and bloom each year on the east side of Bussey Hill below the Overlook, and Corydalis bulbosa is growing near the top of the bank below the Bussey greenhouses, where Dr. Edgar Anderson reports having seen it at least ten or twelve years ago. Several plants reported in the first list have already disappeared from the Arboretum, or at least have not been seen again. Most of these were waifs escaped from cultivation, such as the cock’s-comb, candytuft, sweet alyssum, beef-steak plant, sneeze weed, corn flower and Nicotiana, or weeds of chance introduction, such as the jointweed, small bindweed and Bassa, but amongst them are also the cardinal flower, blue lobelia, wild senna, beard-tongue and Venus’ looking-glass. It is quite possible that some of these will be introduced again at some time. The European smoke-tree, mentioned as having formerly been seen on Hemlock Hill, has been rediscovered growing there amongst the rocks, and a specimen of the moth mullen was collected during the present summer among the lilacs at the foot of Bussey Hill. Hepatica has been reintroduced at 1935] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 85 the place where it formerly grew near the edge of the North Woods. A few plants of both Hepatica americana and of H. acutiloba have been set out and it is hoped they will survive. It is probable that it was the latter species that was formerly native here and not H. americana as reported on the list. The presence of certain native plants persisting in places scarcely suitable to them at present offers some evidence as to former conditions in parts of the Arboretum and of the changes that have taken place, and this may have some value as a guide or check in future planting, since it affords a clue to both past and present soil and drainage conditions. Skunk cabbage continues to come up every year along what appears to be now a well-drained bank below the stone foot bridge over Bussey Brook and near the bald cypresses, as well as along the bridle path oppo- site the lindens, among the Chinese apples near Peters Hill, and at sev- eral other places. Sensitive fern, royal fern, and the lance-leaved violet coming up in the edge of the maple group, at the foot of the hills near the Ilex and Aesculus groups, along the Meadow Road by the laurels, and elsewhere, indicate former boggy areas and show that the water table even now is very near the surface in wet seasons. Along the edge of the path near a planting of Aesculus parviflora the water pennywort has even managed to survive and still sometimes to produce fruit. The persistence also of certain shade-loving plants in open sunny situations where they are gradually being exterminated furnishes evidence, in some case no longer available from records, that the protecting woods have not long been removed. As the Arboretum has developed, the natural drainage has been modi- fied or changed in many places. A brook formerly entered the Arbore- tum area from the west through a gap in the low hills near the Aesculus and Linden groups. A small tributary which drained the ponds near the Forest Hills gate joined it as it flowed across the level ground at the foot of the hills and into the low ground across the Meadow road. The water from this brook is now carried under ground and only a small fragment of the course of the smaller stream can be made out in the somewhat boggy area where the corkwood is now growing. The course of these streams is shown on old maps and their history helps to account for the presence here of such native plants as Carex crinita, Scirpus rubro- tinctus, Pilea pumila, Callitriche palustris, Ludvigia palustris, Hypert- cum majus, Hydrocotyle americana and Scutellaria lateriflora, as well as suggesting the great changes that must have taken place in the char- acter of the flora and the many plants that must have disappeared from the area since the time when these brooks flowed across the fields and into the low meadow and bog. 86 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI The course of Bussey Brook has also been diverted or straightened at several points, and the amount of water that it formerly received from seepage and springs has been greatly diminished by the clearing off or thinning of the forests on the hills and of thickets along its margins, the water now running off rapidly after a rain instead of sinking into the humus and soil. The diversion of its permanent water supply has also been made almost complete by the construction of ditches and sewers along its upper course beyond the Arboretum. It is evident from a study of the surviving native plants as well as from the topography that a small swamp or bog formerly occupied the low ground a little way above the stone foot bridge and between the base of Hemlock Hill and the slopes now occupied by the conifer groups. A spring and little rivulet carrying water except in the dryest seasons still feed the brook on the north side, and small areas are kept wet by seepage water here for a considerable part of the year. But even beyond these moist places some traces of the palustral flora still remain. Such plants as Onoclea sensibilis, Lycopodium complanatum var. flabelliforme, Carex lurida, and sprouts of Salix pedicellaris, Alnus incana, Vaccinium corymbosum and of an undetermined species of Rhododendron have been found here. On similar evidence it can be seen that certain parts of Bussey Hill and of Peters Hill were covered with forest until recent years. An early map of the Arboretum shows native woods extending over a large part of Bussey l1ill, and Bussey Woods is mentioned on some of Mr. Faxon’s plant lalels, but I have seen no similar record in regard to Peters Hill. This hill was probably at one time covered with forest, but from the pres- ent composition of the flora it may be inferred that much of the sur- face was, cleared and used for pasture or other purposes at an early day, but that remnants of the forest remained along the east and north sides until quite recently. Sprouts of a number of characteristic forest trees and shrubs continue to come up here in spite of annual mowing, and stumps of several large trees are still in evidence. A single large white oak survives on the east slope and in its protection a number of plants are growing that are not found in the open ground. Others still huddle rather pathetically about the decaying stumps or in the meager shade afforded by the small thorn trees. Several groups of sprouts of the tremblinyz aspen and of the large-toothed aspen are found on the hill- sides as well as scattered specimens of various species of oak, hickory, chestnut, birch, elm, wild-cherry and ash; and among shrubs are the bayberry, sweet fern, meadowsweet, dwarf juneberry (Amelanchier oblongifolia), choke-cherry, sheep laurel, panicled dogwood and several species of wild rose, blackberry, dewberry, raspberry and blueberry. A few depauperate plants of the ground pine still survive in one spot, and 1935] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 87 a large colony of false Solomon’s seal (Smilacina stellata) is growing about one of the old stumps, with the wood aster, false lily of the valley and other plants that are evident relics of a woodland flora. The com- plete removal of the forest or thickets on rather steep slopes has resulted in the loss of the humus and in the leaching out and removal of the soil, which is reflected in the slow growth and poor condition of some of the Crataegus trees on this side of the hill. The original native flora has almost entirely disappeared from most of the Arboretum, and increasing inroads upon such fragments of it as remain will necessarily be made as the planting of cultivated trees and shrubs continues and as they come to occupy the ground more fully, and as the surface and soil are further modified by drainage, grading, and the bringing in of outside soils and fertilizers. Such traces of it as still remain have considerable significance in a number of ways, and a record of it should be of increasing interest and value in the future. The introduction of weeds and other exotic plants from various sources is certain to continue, and specimens of them should be collected and rec- ords kept as they appear or are discovered, and it may be thought worth while to issue another supplement to the Spontaneous Flora at some time in the future. I wish to express my thanks to the members of the staff and other friends who have shown an interest in the native and introduced plants of the Arboretum through the contribution of specimens and other assistance. Mr. Frederic W. Grigg has examined a number of the grasses, sedges, and other plants and has aided in their determination. I am also under obligation to Mr. C. A. Weatherby for assistance on points of nomenclature, and to Professor J. G. Jack and Professor Alfred Rehder of the Arboretum staff for information about early conditions in the Arboretum and for other suggestions, as well as to several others who have brought in specimens of plants found in the Arboretum. ENUMERATION OF THE ADDITIONAL PLANTS COLLECTED! Polystichum acrostichoides (Michx.) Schott. CHRISTMAS FERN. One plant found by Grant D. Darker on east slope of gravelly ridge, North Woods, probably surviving from a former native colony; also several plants on rocky wooded slopes on north side of Hemlock Hill. Nos. 40273, 42588. Rare. Thelypteris spinulosa (O. F. Muell.) Nieuwland var. intermedia (Willd.) Nieuwland. SPINULOSE SHIELD FERN. Base of wooded hills, near Aesculus group. No. 39678. 1Introduced plants are marked by an asterisk (*). 88 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI Juniperus communis L. var. depressa Pursh. Low JUNIPER. Rocky ground, Central Woods. No. 36405. Rare. Typha latifolia L. Common Cat-taiL. Borders of ponds and brooks. No. 39605. Sagittaria latifolia Willd. f. gracilis cane Robinson. Muddy mar- gins of Pond, South Street tract. No. 4 Potamogeton ere Raf. var. Nuttallii me & Schlecht.) Fer- nald. In shallow water along muddy margins of pond, South Street tract. No. 42723. Panicum philadelphicum Bernh. Waste and cultivated ground, No. 39669. Uncommon. Panicum dichotomiflorum Michx. Waste and cultivated ground. Nos. 38229, 39704. Panicum depauperatum Muhl. var. psilophyllum Fernald. Rocky slopes and ledges. Nos. 25608, 25627. Panicum linearifolium Scribn. Rocky ledges, conglomerate outcrops. No. 40172. Panicum lanuginosum Ell. var. Lindheimeri (Nash) Fernald. (P. Lindheimeri Nash). Common in dry open woods and meadows. Nos. 39588, 39621, 39638, 42646. Panicum lanuginosum var. septentrionale Fernald. Border of woods. No. 39635. Panicum commutatum Schultes var. Ashei (Pearson) Fernald. Dry, open woods, South Street tract. No. 42675 Panicum oligosanthes Schultes var. Scribnerianum (Nash) Fernald (P. Scribnerianum Nash). Dry gravelly banks between Shrub Collec- tion and Arborway wall. Nos. 39627, 39694. Panicum latifolium L. Edge of North Woods, near Celtis group. No. 39637. Rare. *Echinochloa crusgalli (L.) Beauv. f. longiseta (Trin.) Farwell. Culti- vated and waste ground, with the typical form. No. 28102a. Echinochloa muricata (Michx.) Fernald. Cultivated and waste ground. No. 42212. Aristida dichotoma Michx. Poverty Grass. Sterile gravelly banks, between Shrub collection and Arborway wall, and also on conglom- erate outcrops in Conifer group. Nos. 38190, 39742. Aristida gracilis Ell. Sterile gravelly banks, between Shrub Collection and Arborway wall. No. 38191. *Agrostis canina L. BROWN BENT Grass. Dry open ground, slopes of Bussey Hill. No. 39578. *Eragrostis cilianensis (All.) Link ex Lutati. (E. megastachya Link.) 1935] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 89 Waste ground, old quarry near Bussey Street, and also as a weed in South Street nursery. Nos. 38197, 38227. Eragrostis pectinacea (Michx.) Steud. Meadows and dry banks. Nos. 39687, 40234. Glyceria septentrionalis Hitchc. Margins of Bussey Brook, near Coni- fer group. No. 39661. Festuca ovina L. SHEEP’s FEscuE. Gravelly slopes, south side of Peters Hill. Nos. 36461, 36506. Festuca rubra L. var. commutata Gaud. Open ground, border of Aescu- lus group. No. 40199. *Cyperus esculentus L. Rich waste ground, old quarry along Bussey Street. No. 42202. Scirpus rubrotinctus Fernald. Along little brook, near Aesculus group. No. 36586 Carex Crawfordii Fernald. Open banks and meadows. No. 40162. Carex tenera Dewey (C. straminea of Gray’s Man.). Dry ground, bor- ders of woods and meadows. Nos, 42609, 42624. Carex laxiflora Lam. var. gracillima Boott. Moist banks of pond, near Forest Hills gate. No. 36501. Carex laxiflora var. leptonervia Fernald. Shaded banks and borders of woods. Nos. 39589, 42650. Carex canescens L. var. disjuncta Fernald. Springy ground, near base of Peters Hill. No. 36531. *Carex caryophyllea Lat. Dry slopes and gravelly banks, Peters Hill, Bussey bank, and near Dawson House. Nos. 36455, 40132, 40165. Carex angustior Mackenzie. (C. stellulata Good. var. angustata Carey). Local in boggy ground about spring, along southeast side of Peters Hill. Nos. 36532, 40128; also a specimen collected by Mary E. Gil- breath, June 6, 1892, in herb. of New England Botanical Club. Carex panicea L. Grassy slopes of Peters Hill, in Crataegus group. Nos. 36423, 36460, 36529, 36558. Carex pennsylvanica Lam. var. lucorum (Willd.) Fernald. Rocky banks, near top of Hemlock Hill. No. 36578. Carex varia Muhl. Dry rocky ledges, south side of Hemlock Hill, and along base of hills, North Woods. Nos. 40029, 40279a, 40281a. Carex Goodenowii J. Gay. Wet ground about spring, southeast side of Peters Hill, and also in low meadows near Administration Building. Nos. 40129, 40159, 40177. Carex communis Bailey. Specimen in the herbarium of the New Eng- land Botanical Club, collected by C. E. Faxon, May 30, 1878; also found on Hemlock Hill. No. 36455. 90 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XvI Carex brevior (Dewey) Mackenzie. Dry open woods, Oak group. No. 42621. Carex longirostris Torr. Shaded ground, at foot of Hemlock Hill, on south side. No. 40276 Carex crinita Lam. var. gynandra (Schwein.) Schwein. & Torr. Wet rocky ground along Bussey Brook at foot of Hemlock Hill. No. 42622. Carex lupulina Muhl. Margin of small pond west side of Bussey Hill. Nos. 40188, 40205. *Tradescantia virginiana L. VIRGINIA SPIDERWORT. Freely escaped into meadow, near Dawson House. Nos. 39646, 39675. *Luzula nemorosa (Poll.) Mey. Open grassy border, near Dawson nursery. No. 40185. Erythronium americanum Ker. YELLOW ADDER’S-TONGUE. Under trees at base of Hemlock Hill, on southeast side, where leaves come up each year from a small colony, but it has not been found flowering. There is also a colony which flowers freely on a wooded hillside of the Adams Nervine Hospital grounds, only about 20 feet from the Arbore- tum boundary, where the flowering specimen was collected. Nos. 36387, 36415. Allium canadense L. Witp Garuic, Open wooded banks, South Street tract. No. 42620. *Scilla sibirica Andr. Well established on banks near State Laboratory barn, and at top of Bussey Bank. No. 36361 Sisyrinchium. angustifolium Mill. Dry open slopes of Peters Hill, in Crataegus group and on slopes of Bussey Hill. Nos. 36497, 40026, 40144. Cypripedium acaule Ait. STEMLESS LApy’s SLIPPER. Under hemlocks and pines, near the base of Hemlock Hill, on the northeast side. No. 40272 Salix lucida Muhl. SHINING WiLLow. Along small spring brook, Coni- fer group. No. 39604. Salix pedicellaris Pursh. Boc Wi1Lttow. In boggy ground about spring, southeast side of Peters Hill, and margins of Bussey Brook near Conifer group. Nos. 36384, 36418. Salix humilis Marsh. PRAIRIE WILLow. Open slopes of Peters Hill, in Crataegus group. Sprouts coming up after repeated mowing. No. 39706 *Salix alba L. var. calua G. F. W. Mey. Wet ground about spring, Pop- lar group. No. 36530. *Salix fragilis L. Crack WiLLtow. Slopes of Peters Hill; sprouts per- sisting after mowing. No. 39718 1935] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 91 *Salix pyrifolia Anders. Waste ground near pond, South Street tract. No. 40250. Populus tremuloides Michx. QuUAKING ASPEN. Several large colonies of sprouts persistent after repeated mowings, on Peters Hill, in Cra- taegus group. No. 36413. Carya ovalis (Wang.) Sarg. Broom Hickory. North Woods. No. Sla. Quercus prinoides Willd. Scrusp CuEestNuT Oak. Rocky banks near top of Hemlock Hill and on conglomerate outcrops, Central Woods. Nos. 36456, 39692. Rare asa native plant. Quercus Rehderi Trelease. (Q. ilicifolia velutina). Rocky slope, Central Woods. No. 39682. *Ouercus Leana Nutt. (Q. imbricaria velutina). Rocky slope, Cen- tral Woods; also sprouts that appear to be this hybrid coming up spontaneously in Oak group. No. 39683. Polygonum Careyi Olney. Waste ground and cultivated beds, near State Laboratory barn. Nos. 42689, 42708. Polygonum lapathifolium L. Moist waste ground and borders of ponds. Nos. 42693, 42709. Polygonum punctatum Ell. var. leptostachyum (Meisn.) Small. No. 42711. *Polygonum Sieboldii De Vriese. Waste ground, old quarry near Bus- sey Street. No. 42201. *Rumex crispus < obtusifolius. Margins of small spring brook, Conifer group. Growing with the parent species. No. 40202. *Mirabilis Jalapa L. Four-o-cLock. Waste ground, as a waif, old quarry near Bussey Street. No. 39703. * Aristolochia Kaempferi Willd. JAPANESE BrrTHWoRT. Open ground near Administration Building, and also along Meadow Road near rock spring. No. 42673. Aristolochia Clematitis included in the original list, without collection, was probably based on young sprouts of this species, and should therefore be dropped. *Spergula arvensis L. Corn Spurry. In cultivated ground among laurels and other shrubs, near Hemlock Hill. Nos. 38171, 38190. *Sagina decumbens (Ell.) T. & G. Peartwort. Grassy borders and waste ground, old quarry and along Valley Road. Nos. 42633, 42660. *Silene antirrhina L. SLEEPY CaTCHFLY. Waste and cultivated ground, Maple group. No. 32635. *Silene Armeria L. SWEET WILLIAM CATCHFLY. Grassy borders, near Administration Building. No. 42688. *Saponaria officinalis L. Bouncinc Bet. Meadows and waste ground, near Dawson nursery. No. 39684. 92 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI *Ranunculus bulbosus L. Bu_Bous BuTTERCUP. Common in meadows. Omitted through oversight from first list. Nos. 36438, 36526, 36587. Anemone quinquefolia L. Woopv ANEMONE. Local in open woods, North Woods, slopes of Bussey Hill, near Oak group, and South Woods, near Peters Hill. Nos. 36372, 36398, 36410. *Clematis paniculata Thunb. Escaped into meadows, in Tilia group, No. 40182. *Aquilegia vulgaris L. GARDEN COLUMBINE. Bussey bank, in partial shade. No. 42607. *Liriodendron Tulipifera L. Tutte Tree. There is a large tree of this species in the edge of the woods along the base of Hemlock Hill that appears to be spontaneous. *Berberis Thunbergti DC. Established in woods near top of Peters Hill, and also on Hemlock Hill. No. 36380. Benzoin aestivale (L.) Nees. SptcE Busy. Near base of gravelly ridge, North Woods. No. 36437. *Corydalis bulbosa DC. Shaded bank, near Bussey Greenhouse. No. 40005. *Rorippa sylvestris (L.) Bess. YrELLOw Cress. Waste ground and cultivated borders. Nos. 39562, 39657. *Diplotaxis muralis (L.) DC. Waste ground about pond, rich soil, South Street tract. No. 38233. *Ribes sativum Syme (R. vulgare Lam.) RED CURRANT. Wooded bank near Forest Hills gate, also in woods near top of Peters Hill. Nos. 36394, 36412a. *Gillenia trifoliata (L.) Moench. INDIAN PHuysic. Open woods, edge of Oak group. No. 36602. Rare. Aronia arbutifolia (L.) Ell. CHOKEBERRY. Rocky open woods, Cen- tral Woods, and also on north slope of Peters Hill. No. 36404. *Malus baccata Borkh. var. mandshurica Schneider. Northeast side of Peters Hill. No. 40017. *Malus hupehensis (Pamp.) Rehder (M. theifera Rehder). Northeast slopes of Peters Hill. No. 40253. Amelanchier stolonifera Wiegand. Open woods, top of Hemlock Hill. No. 40265. Crataegus rotundifolia Moench. Bussey Woods, C. E. Faxon, June 6, 1882; Geo. Engelmann, Aug. 27, 1882; Peters Hill, C. E. Faxon, Oct. 1, 1883, Sept. 21, 1889; J. G. Jack, May 23, 1900. Sprouts of this plant still persist in the edge of the woods near the top of Peters Hill. Rubus Idaeus L. var. strigosus (Michx.) Maxim. Northeast slope of Peters Hill, in Crataegus group. No. 39725. 1935] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 93 *Rubus parvifolius L. Escaped and well established in Quercus group. No. 4 ; Rubus Jeckylanus Blanchard. C. E. Faxon, June 7, 1913. Rubus allegheniensis Porter. Open woods and banks. Nos. 36564, 40145 Rubus flagellaris Willd. DrwsBErry. Common in open rocky woods and on conglomerate outcrops. C. E. Faxon, July 19, 1909. Nos. 36451, 37689. Rubus Randi (Bailey) Rydb. C. E. Faxon, July 21, 1912; July 31, 1912; July 7, 1913. *Potentilla canadensis var. villosissima Fernald. Cultivated ground. No. 28014. (See Rhodora, 33: 187. 1931). *Prunus Cerasus L. More_tto CuHerry. Rocky woods on Hemlock Hill, and in woods near top of Peters Hill. No. 36493. *Prunus pumila L. var. susquehanae Jaeg. SAND CHERRY. Open woods, Hickory group, and persisting as sprouts after repeated mowing on Peters Hill. Nos. 36493, 40259. *Colutea media Willd. BLADDER SENNA. Open banks near Arborway wall and on Overlook. W. H. Judd, July, 1931. *Amorpha fruticosa L. Fatsrt Inptco. Borders of pond, near Forest Hills gate. No. 38240. Lespedeza hirta (L.) Hornem. Harry BusH-cLover. Rocky banks and borders of woods, Central Woods, near chestnuts, slopes of Bussey Hill and South Woods. Nos. 39653, 39735. *Medicago hispida Gaertn. Bur CLover. Waste ground, old quarry near Bussey Street. No. 38201. Apios americana Med. (A. tuberosa Moench). GRrounp Nut. Open grassy slopes, in Malus group, near foot of Peters Hill. *Lathyrus pannonicus (Kramer) Garcke var. versicolor (Gmel.) Maly. Open bank near small pond, on slope of Bussey Hill. Nos. 39660, 40008. Oxalis europaea Jord. f. villicaulis Wiegand. Cultivated and waste ground, with the typical form. No. 39748. Oxalis stricta L. Rocky open ground, South Woods. No. 38211. Polygala sanguinea L. PurpLte Mitkwort. Open slopes of Peters Hill, in Crataegus group. No. 39655. Callitriche palustris L. Moist banks and borders of ponds and brooks. Further study of more mature specimens show that plants reported in original list as Callitriche heterophylla Pursh are this species, and the latter should therefore be dropped from the list. *Evonymus obovatus Nutt. TRAILING STRAWBERRY BusH. Moist ground at base of hills, North Woods. No. 36569. 94 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI *Celastrus orbiculatus Thunb. Rocky slopes, south side of Hemlock Hill. No. 4 *Ampelopsis humulifolia Bunge. Dumps and waste ground. South Street tract. No. 42694. *Hibiscus Trionum L. FLOWER-OF-AN-HOUR. Waste ground, old quarry near Bussey Street. No. 38199. *Malva parviflora L. Waste ground, old quarry, near Bussey Street. No. 38200. *Malva verticillata L. var. crispa L. CurLtED MaLttow. Waste ground, old quarry south side of Bussey Street. No. 42794. *Sida hermaphrodita (L.) Rusby. Vircinta MALLow. Rocky open ground, near Centre Street gate. No. 25893. This was incorrectly identified in original list as Napaea dioica L., a plant which it closely resembles. Lechea tenuifolia Michx. PINweEep. Dry gravelly banks and borders of woods, Central Woods and South Woods. Nos. 39653, 40216. Lechea intermedia Leggett. Gravelly banks, near top of Peters Hill. No. 42701. Viola pedata L. var. lineariloba DC. Brrp-roor VioLET. Rocky open woods, top of Hemlock Hill and a single plant found on east slope of Peters Hill. Nos. 36557, 40256. Viola sororia Willd. Meadows, Aesculus group. No. 36477. Viola latiuscula Greene. Common in meadows and on open banks. Nos. 36366, 36388. Viola sagittata Ait. ARROW-LEAVED VIOLET. Moist grassy ground, near Arborway wall, in Maple group. Nos. 40268, 42670. Rare. *Viola odorata L. ENGLISH or SWEET VIOLET. Shaded bank, near Jamaica Plain gate. Nos. 36417, 39597, 40007. Viola fimbriatula * papilionacea. Base of hills, near Leitneria group. No. 42682. *Epilobium hirsutum L. Low ground along brook, near Arborway wall and opposite Administration Building, C. H. L. Gebfert. *Aralia hispida Vent. BRisTLY SARSAPARILLA. Waste ground, old quarry along Bussey Street. *Cornus stolonifera Michx. Rep Oster. About small abandoned quarry, South Woods. Nos. 36513, 36524. Pyrola americana Sweet. ROUND-LEAVED WINTERGREEN. Woods, north side of Hemlock Hill. Nos. 39565, 39591, Vaccinium corymbosum L. HicH BLUEBERRY. Open woods, Central Woods, slopes of Peters Hill, and along small brook near leitnerias. Nos. 39585, 39631, 40143. 1935] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 95 Trientalis americana (Pers.) Pursh. STAR FLowER. Woods, top and north slopes of Hemlock Hill, and base of hills, North Woods. Nos. 36454, 40030, 40271. *Ligustrum vulgare L. PriveT. Escaped in thickets and open woods. Peters Hill and South Woods.: No. 38184. *Syringa vulgaris L. ComMMon Litac. Persistent and spreading from cultivation in several places. J. G. Jack, *Phlox paniculata L. GARDEN PuHtox. Rich waste ground, South Street tract. Nos. 40242, 42676. *Echium vulgare L. Biuet WerEp. Weedy and grassy border along bridle path, near Aesculus group. Nos. 40183, 40236. Lycopus rubellus Moench. Wet meadow, between Administration Building and Arborway wall. No. 39691. *Physalis heterophylla Nees var. nyctaginea (Dunal) Rydb. Rich open ground, South Street tract. No.28149. * Datura Stramonium L. JimMsoN WEED. On dump, South Street tract. No. 38236. *Datura Metel L. THorn AppLte. On dump, South Street tract. No. 42208 *Lycopersicon esculentum Mill. Tomato. Not rare in waste and culti- vated ground and sometimes producing fruit and self-seeding. No. S732, Veronica peregrina L. NECKWEED. In waste and cultivated ground. Nos. 36442, 36516. Epifagus virginiana (L.) Bart. BrEcH Drops. On superficial roots of beech trees, south side of Hemlock Hill. No. 38185. Plantago major L. CoMMoN PLANTAIN. A common weed in waste ground. Nos, 39685, 39749, 40214. *Galium asprellum Michx. RoucH BrepstrAw. Moist weedy ground, South Street tract. Nos. 40251, 42678. Houstonia caerulea L. Buiurets. Among shrubs and laurel bushes, foot of Hemlock Hill, northeast side, and also one plant collected amongst laurels near South Street gate. Nos. 40260, 40270. Diervilla Lonicera Mill. Busu HonrysuckLE. Rocky open woods and ledges, Central and South Woods and slopes of Hemlock Hill. No. 36571, *Lonicera alpigena L. Rocky ground, south slope of Hemlock Hill. No. 36452. *Lonicera dioica L. HoNEySUCKLE. Open woods, Oak group. No. 42216. *Viburnum trilobum Marsh. HicH-BUSH CRANBERRY. Woods near top of Peters Hill. No. 38246. 96 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI *Lagenaria leucantha (Duch.) Rusby. Gourp. On dump, South Street tract. No. 42209. Liatris scariosa Willd. Bazine Star. Open woods, low hills, Maple group. No. 42204. Rare. Solidago ulmifolia Muhl. Open woods, South Street tract. No. 42206. There is also a specimen of this species in the herbarium of the New England Botanical Club, collected by C. E. Faxon, on “Bussey Moun- tain,” Sept. 4, 1887. Solidago odora Ait. SwEET GOoLDEN-RoD. Open woods, slopes of Bus- sey Hill near Oak group and in South Woods. Nos. 38213, 39736. Rare. *Boltonia asteroides (L.) L’Her. In low weedy ground, South Street tract. No. 42211 Aster multiflorus Ait. Dry open slopes, near top of Peters Hill. No. 38721. Aster linariifolius L. f. leucactis Benke. Slopes of Peters Hill, in Cra- taegus group. A form with smaller heads and white rays, growing with the species. No. 29730. Aster acuminatus Michx. Under shade of apple trees, on hillside north- west of Administration Building. No. 42213. Rare. Antennaria plantaginifolia (L.) Richards. Dry grassy slopes and meadows. Nos. 36362, 36386, 36444. Common. Antennaria canadensis Greene. Grassy slopes, gravelly soil, between Arborway wall and Shrub Collection, and also on slopes of Peters Hill. Nos. 36393, 40014. Antennaria neglecta Greene. Dry open slopes of Peters Hill. Nos. 40014a, 40014b. *Heliopsis helianthoides (L.) Sweet. Ox-Eyr. Spreading from cultiva- tion into open ground, near Dawson House. No. 42681. *Helenium nudiflorum Nutt. SNEEZEWEED. Open grassy ground, near South Street gate and also near Administration Building. Nos. 42205 42655, 42680. *Chrysanthemum segetum L. Corn Maricotp. Waste ground and dump, field near Dawson House. Senecio aureus L. GOLDEN Racwort. Moist shaded ground between Tilia group and bridle path, and also in Poplar group near Peters Hill. Nos. 36476, 36511. *Senecio viscosus L. CLAMMy GROUNDSEL. Waste and cultivated ground, South Street tract and Conifer group. Nos. 38232, 39750. *Hypochaeris radicata L. Cat’s-EarR. Grassy open ground northwest of Administration Building. No. 42672. ? 1935] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 97 *Sonchus arvensis L, var. glabrescens Guenth. Wimm. & Grab. PEr- ENNIAL Sow THISTLE. Rich waste ground, near pond, South Street tract. Nos. 42665, 42691, 42699. *Sonchus asper (L.) Hill. Sow TuistLte. Waste and cultivated ground. Nos. 35710, 35927. These numbers were listed through error as Son- chus oleracea in original list. Both species are found, and the latter has been collected under numbers 39739a and 40219. *Crepis capillaris (L.) Wallr. Hawk’s Brarp. Grassy open slopes near Administration Building and also along Meadow Road near Hemlock Hill. Nos. 42192, 42629, 42671. *Lactuca scariola L. Prickty Lettuce. Waste ground, South Street tract. No. 40245. *Lactuca sagittifolia Ell. Waste ground, South Street tract. No. 4 eiartues spicata (Lam.) ones var. integrifolia (Gray) Britton. Open woods, Carya group. No. 39648. *Hieracium florentinum All. =. Devit. Grassy slopes of Peters Hill in Crataegus group, near Roslindale Gate and in Celtis group. Nos. 36581, 39563, 40170, 40180. *Hieracium vulgatum Fries. Grassy open ground near Platanus nurs- ery. No. 42795. HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI THE HOSTS OF GYMNOSPORANGIUM GLOBOSUM FARL. I, IT. IIT. AND THEIR RELATIVE SUSCEPTIBILITY J. D. MACLACHLAN With plates 125 to 128 and four text-figures TABLE OF CONTENTS RAC irs go 05.9695 6 a beso 4 bean Ged acenee ae tue Relative susceptibility of hosts within genera of the Pomoideae PIGOEY Sodnos pase as ease ed ie ce web eee ese wae Methods used to determine susceptibility ................. CPOE SACRE ian ooo aee ¥en ewes ey seater nce ees Recording of data obtained from inoculations ............ Investigations and conclusions with respect to the various 1 PCNA: CONSIINTO os geceia bh -0 60 es ae OE EE PRES Crataegus Presentation of — obtained by observations on NEAT! DATOCION. uy 5 vn see e ee eee l eae cnn B. Presentation of an obtained by serial inoculations. . C. Factors influencing the relative susceptibility of Crataegus 1. The geographical distribution of Crataegus.... e 2. The role of the foliar cuticle ................ 3. The degree and the duration of the period of at , SUACEOUDUUY oiavdssacscuwsered cen eee ens D. Correlation of the data to classify the groups of Crataegus with respect to their relative susceptibility E. oT for the selection of resistant species and ieti 1 ic susceptibility as indicated by serial 1 yrus — IMOMIACONG. hc ieee cases eke Hepa EE eee Rey rene ene Sorbus — Relative susceptibility as indicated by serial TOCHIDHOUS: Asi vaascdevnessine Cooke et ecen ands leeks Malus — asians susceptibility as indicated by serial INGCUIAOOS cise eed wei ee deh eee eek Hh68S THT ee decieteatide Cydonia, Comptonia, rou de Aa Mes- 1 pilus, Myrica, Photinia, Sorbaronia, Sorbopyrus ........ Relative susceptibility of hosts within the genus Tales ... 128 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 99 IV. The hosts of Gymnosporangium globosum Farl. Peete tor tne) el Stage ss «x55 cduaniow dee aniehie vee ye 129 Prete 200 Ee 11 SEARS svc oe fk eas oawarte arma eee ee 137 Le Ne ye es Leo aae 650} Se oa bb eae Een ae Se 137 eo RN ENTRIES Nw Sov oes fp aedam sae EU ee ee ee 139 We DONOR TAONY Wi ntsc recr ce os 45 Le uee es eee aN CU eee 139 Ny RO Pee IANO CRIES a xd bviecuie's pats Wak Kee ees ees 141 I. INTRODUCTION GYMNOSPORANGIUM GLOBosUM Farl., a heteroecious rust, is restricted in its telial phase to a limited number of species and varieties of Junt- perus. To the aecial phase, however, representatives of at least ten genera within the Pomoideae may serve as hosts; and certain of these genera, especially Crataegus, include a large number of species and varieties. In spite of the number of hosts hitherto reported for G. globosum, very little information is available regarding the relative susceptibility of the hosts. This is a question of considerable importance because of the great damage done by the rust. A determination of the immune species and the comparative resistance of susceptible species within the various relevant host genera constitute the major part of this paper. Concurrently with the investigations which led to a determination of the relative susceptibility of the hosts, the writer was enabled to compile a more nearly complete list of the known hosts, from which it appears that, instead of the approximately one hundred hosts hitherto reported, the number of hosts should be conservatively estimated at more than six hundred. This list constitutes the latter portion of the paper. The work on the problems outlined above has been carried on at the Arnold Arboretum of Harvard University, where may be found one of the finest collections in the world of living representatives of species and varieties of Juniperus and of the Pomoideae. II. RELATIVE SUSCEPTIBILITY OF HOSTS WITHIN GENERA OF THE POMOIDEAE HISTORY The earliest successful attempt to determine pomaceous hosts of G. globosum Farl. by means of cultures may be credited to Farlow (1880) who, in 1876-7, using teliospores from Juniperus virginiana, obtained spermogonia on Crataegus tomentosa. Farlow (1885) also made suc- cessful cultures on leaves of Crataegus Douglasii, Crataegus Oxyacantha 100 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI and apple seedlings; but he obtained spermogonia only, because his ex- perimental leaves molded before the aecial stage could develop. Thax- ter (1887) obtained spermogonia on Crataegus coccinea, Malus pumila, Sorbus americana, and Amelanchier canadensis; and in 1887-8 (Thaxter, 1889) obtained spermogonia on Sorbus americana and Cy- donia oblonga (= Cydonia vulgaris), and both spermogonia and aecia on Crataegus coccinea and Malus pumila. In a later report Thaxter (1891) confirmed the previous results on Malus pumila and records suc- cessful infections of Crataegus crus-galli and Sorbus americana, both resulting in aecial development. In 1906, cultures were made by Arthur (1907) on Crataegus Pringlei and Sorbus americana resulting in spermogonia and aecia, and on Malus coronaria giving spermogonia only. In 1908 Arthur (1909) confirmed his results on Crataegus Pringlei, and in 1909 (Arthur, 1910) those of Farlow on Crataegus coccinea. Since Arthur’s work more than one hundred suscepts have been added to the host list, mostly by observations made in the field. Authors who have contributed or made significant reference to this list include Clin- ton (1904 and 1934), Stewart (1910), Kern (1911), Stevens and Hall (1910), Arthur (1921, 1924, 1926 and 1927), Burnham and Latham (1917), Hesler and Whetzel (1917), Jackson (1921), Hunt (1926), Anonymous (1930), Thomas and Mills (1930), Sherbakoff (1932), and others. Bliss (1931), by culture, obtained abundant spermogonia and aecia on Crataegus mollis, but obtained only flecking on nine varieties of commercial apples. These previous reports, together with the investigations made by the writer, warrant the conclusion that the genera involved as suscepts for the aecial phase of G. globosum are confined to the sub-family Pomoi- deae, and include Amelanchier, Crataegus, Cydonia, Malus, Mespilus, Pyrus, Sorbus, and the hybrid genera Crataegomes pilus, Sorbaronia and Sorbopyrus. METHODS USED TO DETERMINE SUSCEPTIBILITY Two methods of approach were utilized in the determination of the various hosts and their relative susceptibility within each genus, namely, (1) quantitative observations on natural infection, and (2) serial arti- ficial inoculations during the progressive development of the foliage to determine both the degree and the duration of the period of suscepti- bility. These methods of approach were especially applicable to Cratae- gus which is by far the largest genus susceptible to G. globosum. All cultures and observations were made on trees in the Arnold Arboretum. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 101 CULTURAL TECHNIQUE The cultural technique adopted was similar to that described by Crowell (1934). The inoculum was collected either the previous eve- ning, or in the morning prior to inoculating, from Juniperus virginiana. Galls bearing abundant telial flanges were soaked in water until maxi- mum swelling had taken place; then the gelatinous mass was crushed to form a thick aqueous suspension of teliospores. Fresh inoculum was prepared every two hours during inoculation in order to eliminate any possibility of crushing the promycelia emerging from the germinating teliospores, since microscopic examination revealed that the latter would germinate within that time. All inoculations were carried out in dupli- cate. For each test six to ten leaves on a twig were inoculated; the re- mainder of the tree served as a control. The spore suspension was painted on both sides of the leaves with a camel’s hair brush; then a celluloid cylinder was slipped over the twig and the ends of the cylinder were plugged with moist sphagnum. Care was taken that the inocula- tion should not be exposed to direct rays of the sun; otherwise burning of the leaves within the cylinder might occur. The sphagnum-plugged cylinder formed an excellent moist chamber; on removal of the tube two days later the sphagnum was always found to be still wet, and both the inside of the tube and the surfaces of the leaves were moist. Thus, with a heavy sowing of spores, a moist atmosphere in the inoculation tube, and a temperature below 25°C. the conditions for optimum spore germination and infection exceeded any that might occur in nature. Plate 127 fig. 5 illustrates a type set-up. RECORDING OF DATA OBTAINED FROM INOCULATIONS In recording data the inoculated plants were classified according to four categories or degrees of susceptibility, based on the number of sori, their relative size, and the pathogenic effect on an average-sized leaf. They are designated and defined as follows: O0—ImMungE; no visible infection obtained. 1—RESISTANT; one to five lesions which are relatively small, which cause very little leaf killing and no leaf drop; with or without aecia. This is a type of infection which causes no material. harm to the host. 2—MopERATELY SUSCEPTIBLE; five to twenty-five lesions per leaf with an intermediate pathogenic effect between categories 1 and 3; aecia always produced. This is a type of infection which, while reducing the photosynthetic leaf area and causing some leaf kill- ing, does not result in defoliation. 102 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI 3—VERY SUSCEPTIBLE; twenty-five or more lesions which are usually large or fuse to form large masses and which cause severe leaf killing and leaf drop; abundant aecial horns produced in each lesion. This is a type of infection which ruins the foliage. While these definitions are, in general, applicable in allotting a suscept to any one category, they can not be considered as absolute criteria. Within the genus Crataegus, for example, as will be shown later in this paper, variation in susceptibility is for the most part not physiological but is dependent primarily upon a natural barrier, the cuticle; the probability that the basidiospore can produce infection varies inversely with the thickness of the foliar cuticle. Again, the amount of leaf kill- ing is dependent upon whether infection takes place on main veins or elsewhere on the leaf. Consequently, for Crataegus at least, the actual number of lesions per average-sized leaf was given major significance. In other genera, the type of infection was accorded major consideration. In the genus Pyrus, for example, certain species exhibited very small lesions which died shortly after spermogonia appeared, while other spe- cies of this genus showed larger lesions producing aecia. In general, however, the foregoing definitions were employed as the bases for plac- ing the various species within the different categories of susceptibility. INVESTIGATIONS AND CONCLUSIONS WITH RESPECT TO THE VARIOUS GENERA CONSIDERED For the sake of convenience the various host genera will be considered individually with respect to their relative susceptibility. All the known hosts within each genus will be listed at the end of the paper. Crataegus The Arnold Arboretum with almost one thousand trees comprising about five hundred and fifty named species and varieties, spread over twenty-four groups, afforded an excellent opportunity to study the relative susceptibility of the Crataegi. But, since the species of this genus hybridize so freely, and since the specific classification is still in an unstable condition, the time and labor involved in making inocula- tions for each of those species and varieties (especially in the large very susceptible groups where an abundance of natural infection was ob- served) would not justify the results that might be obtained; conse- quently typical representatives of each of the twenty-four groups were selected and the results were used as a basis of comparison by groups rather than by species. Likewise the data obtained on all the species and varieties by observations on natural infection were treated by groups rather than by species. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 103 A. PRESENTATION OF DATA OBTAINED BY OBSERVATIONS ON NATURAL INFECTION In July, 1932, a general spread of G. globosum was observed through- out the entire plantation of Crataegi in the Arnold Arboretum.’ De- tailed observations were warranted by the fact that, within each group, the degree of infection was consistently uniform regardless of where the tree happened to be situated; likewise a sharp line of demarcation could be seen between the number of foliar lesions per tree in a relatively resistant group, such as the CRUS-GALLI, and the number per tree in a more susceptible group, such as the CoccINEAE or ANOMALAE. The amount of infection on any one tree while uniform was slight enough to allow fairly accurate counts to be made of the number of lesions per tree. While these data would hardly be adequate to permit comparison among species within any one group of Crataegus, they were sufficient for comparing the relative degree of susceptibility of the various groups represented in the Arboretum. As stated above, about one thousand trees were available for examination. Observations were made at the spermogonial stage, and again at the aecial stage of the rust. In order that the amount of infection per tree might be fairly compared the trees were graded as to size, five size- classes being used.!| Counts were made of the number of foliar lesions per tree at both stages of the rust; where the counts exceeded one hun- dred per tree, the degree of infection was termed “100-+-”.* A collection of herbarium material was assembled for permanent reference. In correlating the data obtained a method had to be devised by means of which a tree, for example size I, could be fairly compared with a tree, for example size V. The CocctNeag, a group containing 46 species rep- resented by 82 trees, exhibited the highest percentage of infection lesions per unit-sized tree. This group was classed as having severe infection, and the values obtained for this group were selected as a basis of com- 1This plantation is a pure, open stand SS on an heey hillside ; furthermore, the groups within the genus are arranged in contiguous blocks. Rust-infected cedars were so remote fhe there was Eaconeediy a Tiiforma distribution of inocu- lum over the Crataegus trees. 1The five size-classes were arbitrary gradings involving the relative amount of foliage as well as the actual tree size. 2A tree with “100+” lesions was considered as having 150 lesions. However, with the exception of those trees that were obviously very susceptible, such occurrences were so rare that deviation from this estimation would make no significant differ- ences in the correlations. 104 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI parison for all the other groups. It was found that for the CoccrNEAE: Size I (9 trees) averaged 24.3 lesions per tree. Size II (33 trees) averaged 51.7 lesions per tree. Size III (35 trees) averaged 75.7 lesions per tree. Size IV (5 trees) averaged 120.0 lesions per tree. Size V (0 trees).* If, for the sake of convenience, the ratio of the number of lesions per tree be changed from 24.3: 51.7: 75.7: 120.0: — to 25: 50: 75: 125: 200, for the respective tree sizes, and these values be considered as units for classifying a tree as having severe infection, then by taking arbi- trary averages for the number of lesions required to class a tree as having moderate infection, mild infection, or no infection, the scheme as pre- sented in Fig. 1 for classifying the trees of all the groups may be formulated. Number of lesions per tree within the respective tree sizes I II III IV V Severe infection . 25 50 75 125 200 20 40) 60 100 160 Moderate infection . 15 30 45 75 120 10 20 30 50 80 Mild infection . 5 10 15 25 40 No infection . 0 0 0 0 0 Fic. 1. An ARBITRARY SCHEME TO DETERMINE THE RELATIVE DEGREE OF INFECTION ON TREES OF DIFFERENT SIZES. From this scheme any tree of any size with any number of lesions may be classified according to the relative amount of infection present. On a tree size I, for example, one to ten lesions would be classed as mild infection, ten to twenty as moderate infection, and more than twenty as severe infection. As may be noted in Fig. 1, the ratio of the average number of lesions for any sized tree for the four degrees of infection is 5: 3: 1: 0. If, then, we multiply the number of trees classed as having severe infection by 5, moderate infection by 3, mild infection by 1, and no infection by O, take the total of these products and divide by the number of trees considered, a unit is obtained by which the relative susceptibility of any group may be fairly and quite accurately compared 1The CoccineEaE did not include any trees of size V; as a matter of fact there are only six trees of this size in the plantation. From actual measurements of the vari- ous tree sizes and from the table given above, it was estimated that a tree of size V must necessarily have at least 200 lesions to be classed as having severe infection. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 105 with a similarly derived unit for any other group. To illustrate this, let us consider a moderately susceptible group, the MACRACANTHAE, and a resistant group, the CRUS-GALLI: MACRACANTHAE (see Table II): Severe infection .. 7 trees XK 5 = 35 Moderate infection .. 10 trees KX 3 = 30 Mild infection .. 78 trees K 1 = 78 No infection .. 4 trees kK O= O Total .. 99 trees 143 Susceptibility unit of comparison 143 = 1.44 99 Crus-GALLI (see Table IT): Severe infection .. O trees KX 5 = O Moderate infection .. 2 trees KX 3 = 6 Mild infection .. 46 trees * 1 = 46 No infection .. 80 trees K 0 = O Total ..128 trees 52 Susceptibility unit of comparison 52 = 0.41 128 The groups of Crataegus examined, the number of species examined in each group and the number of trees representing these species, the numbers of trees classed according to the different degrees of infection, and finally the units of comparison, which may now be considered as the relative degrees of susceptibility as indicated by natural infection, are presented in Table II. These values for the degrees of susceptibility have been plotted in Fig. 4. B. PRESENTATION OF DATA OBTAINED BY SERIAL INOCULATIONS Serial artificial inoculations were made at the following stages in the foliar development: (a) on April 23 and 24, 1934, at which time very little foliage was evident, a few buds had begun to unfurl, the majority were just breaking through the winter scales, while in many instances it was necessary to part the winter scales and insert the inoculum; (b) on May 7 and 8, 1932 and 1934, respectively, at which times (the foliar conditions being approximately the same in both years) the leaves in practically all cases were in an advanced stage of expansion but were still tender, exhibiting relatively little cuticular development; (c) on May 22 and 23, 1933 and 1934, respectively, at which times the leaf 106 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI cuticle was fairly well developed and most of the trees were in an ad- vanced stage of blossom;' (d) on June 28, 1934, when the foliage was, for all practical purposes, fully mature and certain of the groups ex- hibited a very heavy cuticle on the leaves.” The number of species inoculated in each group and the percentage of these falling into the different classes of susceptibility for each of the four serial inoculations are presented in Table III. The correlation of these data will be found under sub-section D. DOUGLASIANA MAC RACANTHR Fic, 2. DIstRIBUTION OF THE GENUS CRATAEGUS IN NortTH AMERICA. 1In certain groups, for example the Crus-GALLt, pyenayeroa could be observed in the type of rake ge pointy y two ica of the sa e species, in which case both were “nocula = — ~ mine if variation in susce otibility existed within a single Exc uch i instances ee different representatives were used in the te je yee (b) and (c). eg ob tage for ghee (c) and (d) had been kept in a refrigerator at 0°C., wher will in a subsequent publication, the teliospores will retain their Viability ns more than ninety percent germination for at least a year. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 107 C. Factors INFLUENCING THE RELATIVE SUSCEPTIBILITY OF CRATAEGUS 1. The geographical distribution of Crataegus Of the twenty-nine groups as given by Rehder (1927), twenty-three are of American origin; the remainder have been introduced from Eurasia. With the exception of the MAcRACANTHAE, which extend into the middle west, and the DoucLasiIANag, which are typically western, the American groups, as indicated by the dotted area in Fig. 2, are confined to the eastern part of North America. While certain of these groups are typically more northern than others they overlap to such an extent that no correlation could be made between the distribution and the relative susceptibility of the respective American groups. Although none of the Eurasian groups proved to be very susceptible, no differences from the type of infection produced on American groups could be observed. Con- sequently, the distribution of the genus gave no information that proved of value in determining the relative susceptibility of the various Cratae- gus groups, 2. The role of the foliar cuticle By using herbarium material collected in the Arnold Arboretum from natural infection in 1932 a detailed comparison was made between one of the largest and most resistant groups, CRUS-GALLI, and one of the largest and very susceptible groups, TENUIFOLIAE, in an attempt to correlate the susceptibility of the host plant with the mechanical struc- ture of the leaf. As a check on the results obtained, the CoccINEAE, another very susceptible group, was examined in a similar manner. The following observations were made: (a) Distribution of lesions on the leat. i. Number of lesions primarily associated with mid and main lateral veins of the leaf. ii. Number of lesions on the chlorenchyma which, for present purposes, may be defined as the leaf area other than that occupied by the mid and main lateral veins. (b) Spermogonial stage. i. Number of spermogonia per lesion. ii. Diameter of lesion. (c) Aecial stage. i. Number of aecial horns per lesion. ii. Diameter of lesion producing aecia. ili. Length of mature aecial horns. iv. Number of lesions actually producing aecia. (d) Detailed notes on thickness of foliar cuticle, degree of hyper- trophy and amount of leaf-killing. 108 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XVI In addition to the above data separate measurements and counts were made for chlorenchyma and vein infections in the CoccinEAE. Table I gives the results obtained for these three groups. TABLE I PRESENTING DATA ON BIOMETRICS AS OBTAINED FROM HERBARIUM MATERIAL FOR THREE GROUPS OF THE CRATAEGI a cal a Pa asl Oo a 7) g — 0 ° Lo] a ° a8 “ oS 2 ¢ 2 38 gs - “ Oo 3) Ss hU*~é oy os A ® & wo oO ° 0 A oO on Th wet 0 n i o 2 ° o w =] fo S . g 8 8 © Ss .b 88 Sh 88 28 ° o £ A “A “4 “_ 5 7 2) o o o 0 ” ns) ao * W Le] ow By be .) o o a oo “4 oO oo “ © od Aas 7) » a cq ) “4 fA mo h aa wh As - fe © @ o “ Hb . . . e pty > ° te .> . ft b> *o oa ° fe} oO °o > oO - Oo Pr oO br Oo Po vo Group a x a a wR Rw am <0 <= B < 0 <0 wae Crus-galli 76 121 2216 216 3? 83 43 3.0 31 3. 3.2 66 Tenuifoliae 83 183 1283 1361 96 47 25 2 2.5 16;55 |] 1.732. BT] 2edsee0 Coccineae 42 75 638 717 97 31 25 2.2 2.4 Crus-galli ~ thick, coriaceous, waxy leaves. Tenuifoliae and Coccineae - thin,non- tgs leaves Within the CoccINEAE the pairs of values (separated by a semi-colon) refer to chlorenchyma and vein infections respectively; the averages are given below the pairs. All measurements were made to the nearest millimeter A comparison of these data brings out three significant facts: (1) Practically all the Crus-GALLI have thick coriaceous leaves with a very heavy cuticle. The TENUIFOLIAE and COcCcINEAE, on the other hand, have thin leaves with little cuticle. This condition was checked for all the other groups, and while the thickness of the leaf itself did not show consistent correlation with the relative susceptibility of the re- spective groups, there was a surprisingly consistent correlation on the part of cuticular thickness. Groups that finally fell into the moderately susceptible class exhibited an intermediate deposition of cuticle, the degree of which varied somewhat in different species within the respec- tive groups. All the species within the groups which were classed as resistant had consistently heavy cuticle and those classed as very suscep- tible had consistently little cuticle. (2) The Crus-GALit leaves have more than eighty percent of the infections on veins, the TENUIFOLIAE approximately fifty percent and the CoccINEaE about thirty percent. By correlating these data with the relative susceptibility of the three groups, it appears that the degree of 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 109 susceptibility varies inversely as the percentage of infections primarily associated with the main veins. (3) Although the Crus-GaALti exhibit the lowest percentage of trees infected, and thus would seem the most resistant, the individual lesions on the leaves of this group have the greatest diameter, and the largest number of spermogonia and aecia per lesion. When these facts are fitted into the picture of the relative suscepti- bility of any host tree to the rust, they definitely indicate that the differ- ence in susceptibility is purely mechanical, the cuticle being the deciding factor. The basidiospores of G. globosum, while able to produce infec- tion from the lower surface of the leaf, germinate and gain entrance primarily through the upper side. Thus, spores carried by the wind and alighting on the smooth waxy surface of the Crus-GALLI leaf are not so liable to adhere, and if they do remain and germinate, a large percentage of the germ-tubes die before they can penetrate the heavy cuticle. Many instances illustrating this phenomenon occurred during investigations of the waxy-leaf types. Within the Crus-catii, for example, a much higher degree of susceptibility relative to the groups with non-waxy leaves was indicated by artificial inoculation where conditions were optimum for the infection process, than by natural infection where the basidiospores must necessarily withstand a certain amount of desicca- tion before infection can take place. Again, in many cases waxy leaves infected by natural inoculation were found on very low branches only, that is, branches almost touching the ground. Here the leaves were kept cool and moist for longer periods of time by the tall grass that hap- pened to be growing around these trees; such an environment afforded a better opportunity for spore germination and germ-tube penetration. The distribution of lesions on the leaves gives further evidence of the cuticle acting as a natural barrier. In the Crus-GALLI eighty-three per- cent of the lesions were primarily associated with the main veins. The little grooves over these veins afford lodging places for the basidio- spores; moisture tends to remain longer along these areas, rendering a more favorable environment for the infection process. When making artificial inoculation by painting the leaves with an aqueous suspension of basidiospores, it was very difficult to get a film of the suspension to lie uniformly over waxy leaves. The water would form into droplets, and either roll off the leaf entirely or else remain in the little grooves over the veins. One can readily picture the same thing happening when the basidiospores are brought naturally. Inoculation usually takes place during wet weather, as it is then that the telial flanges on the galls swell and the teliospores germinate to produce basidiospores. The latter are 110 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XVI then carried aerially, either directly to the Crataegus leaf by the wind, or else washed out of the air by falling rain onto the host leaf. Here, as in the case of artificial inoculation, the moisture necessary for spore ger- mination accumulates in droplets and these either roll off the waxy leaf or remain in the grooves over the veins, carrying the basidiospores with them. With a non-waxy leaf we have an altogether different picture. A film of water readily spreads over the surface of the leaf in a uniform layer, in which case the basidiospores are more apt to remain where they hap- pen to alight on the leaf. Here the germinating basidiospores have no heavy cuticle with which to contend and can successfully penetrate the leaf surface almost as easily at any place over the chlorenchyma as over the veins. Since the area occupied by chlorenchyma far exceeds that occupied by the main veins, one can readily see why only thirty-one per- cent of the lesions on the CoccINEAE leaves were vein infections as com- pared with eighty-three percent on the CRUS-GALLI leaves. The fact that within the Crus-GALLI group the rust flourished even better than within the more susceptible groups, producing larger lesions with a larger number of spermogonia and aecia per lesion, can also be attributed to the relatively high percentage of vein infections. Regard- less of leaf type the very large lesions, some seven to ten millimeters long, with more than one hundred spermogonia and fifty to one hundred aecia per lesion, were vein infections. In the CoccINEAE measurements of vein and chlorenchyma infections were kept separate, in order to ob- ‘tain quantitative data on the relative size of the lesions and the number of spermogonia and aecia per lesion in the two types of infection. As may be seen from the foregoing table, the lesions are much larger in vein infections, producing almost twice as many spermogonia and aecia. All evidences indicate that G. globosum is capable of establishing a much more efficient nutritional regime when in direct contact with one of the veins. In the early spermogonial stage of even chlorenchyma infec- tions one can see yellowish lines of fungal hyphae, radiating out along the vascular bundles from the centre of the lesion, as shown in Plate 125, Fig. 2. Again, in Plate 125, Fig. 1, the infection appears systemic, extending the entire length of a lateral vein. Plate 125, Fig. 3 shows a main lateral vein infection branching out along one of the sub-lateral veins. In fact, in every vein infection observed (eight hundred and eighteen), as may be seen in Plate 125, Fig. 4, the lesion was typically long and narrow, the long axis corresponding with that of the vein. Vein infections appeared to produce aecia later in the season than chlo- renchyma infections. Many cases were found among the former where 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 111 the aecial horns were just emerging or else were very short when the leaves were collected, while nearly all the chlorenchyma lesions had fully developed aecia, with peridial cells ruptured and aeciospores emerging. It would seem, then, that the time of spore production is correlated with the availability of food supply. An infection not primarily associated with a main vein utilizes all the available nutrient and then produces spores. Vein infections, on the other hand, have a greater and more lasting nutrient supply from the host, develop more mycelium and, when they finally do sporulate, have a greater supply of reserve food to pro- duce aecia. Thus chlorenchyma infections produce relatively smaller and fewer aecia over a smaller lesion and at an earlier date than vein infections. This fact would account for the higher percentage of the lesions within the TENUIFOLIAE and CoccINEAE actually producing aecia at the time the herbarium material was collected. Severe leaf killing, where relatively few lesions per leaf were involved, was due in practically all cases to infections primarily associated with the main veins, the amount of leaf killing depending on how far back from the edge of the leaf the vein was attacked. Plate 126, Fig. 2 shows one lesion on the mid-vein resulting in the death of over half of the leaf. On the other hand, in Plate 126, Fig. 1, may be seen a chlorenchyma infection where leaf killing extends from the point of infection to the margin of the leaf but does not extend beyond the enclosing lateral veins. A purely chlorenchyma infection nearer the center of the leaf rarely causes killing beyond the area of actual infection. If the degree of susceptibility is in any way physiological, one would necessarily expect that within the resistant groups the rust would have greater difficulty in establishing a satisfactory nutritional regime, and if once established would produce small lesions with relatively few fruit- ing bodies due to some antagonistic physiological reaction on the part of the host. Crowell (1934) found.such to be the case when he determined the relative susceptibility of the genus Malus to Gymnosporangium Juni- peri-virginianae Schw. In European species of Malus. the lesions were very small, in some cases producing a few spermogonia but no aecia. Somewhat similar instances were found by the writer in determining the relative susceptibility of species of Pyrus to G. globosum. In the Cra- taegi a few rare instances were found that might suggest differential physiological antagonism on the part of the host. In Plate 126, Fig. 5 is shown a lesion that produced abundant spermogonia but died be- fore any hypertrophy or production of aecia took place; the host tissue may have been hypersensitive to the rust mycelium, the latter taking such a heavy toll on the nutrient content of the leaf that the host tissue 112 JOURNAL OF THE ARNOLD ARBORETUM [voOL. XvI was killed and as a result the fungus died. Plate 126, Fig. 4 illustrates a case of leaf killing extending below the area of infection; this suggests the existence of a toxic agent secreted by the rust. In a few of the col- lections very small lesions not more than a millimeter in diameter that never produced even spermogonia were found. In Plate 126, Fig. 3 may be seen a small lesion that exhibits no hypertrophy and produced only one aecial horn. However, such instances as the foregoing were rare and not consistent even on a single host, and may be considered as insignificant factors in determining the relative susceptibility of the genus Crataegus. Indeed, from examination of the herbarium material the writer found the exact opposite to any physiological antagonism on the part of the host to be true; G. globosum is apparently able to estab- lish itself more satisfactorily in the most resistant groups, due to the relatively high percentage of vein infections. This condition would indicate that the basis for differences in susceptibility is for the most part mechanical, involving primarily the cuticle as the deciding factor. The Crus-GaLti is a difficult group for the rust to invade, except for a very short period in the spring before the foliar cuticle has developed to any extent. However, once the rust has successfully penetrated this cuticle it is just as much at home and can do just as much damage or even more in the CRUS-GALLI than it can in the TENUIFOLIAE, COCCINEAE or any other very susceptible group. 3. The degree and the duration of the period of susceptibility The role of the cuticle also explains the significant phase in the dura- tion of the susceptibility of any host. There is a definite duration to this period of susceptibility for all the groups, the degree of which rises rapidly during the unfurling of the leaves and reaches a maximum dur- ing and immediately after the period of leaf expansion, then falls off gradually at a rate depending, in part at least, on the rapidity of deposi- tion of foliar cuticle. In PLatEe 127, Fics. 1-4 are shown the results obtained from the four respective serial inoculations on Crataegus Pringlei. At the time of ini- tial inoculation, April 25, 1934, the leaves, approximately one quarter of an inch long, had just begun to unfurl and a very small amount of infec- tion at the tip of one leaf resulted (Fig. 1). The inoculation on May 9, after the leaves had fairly well expanded, produced severe infection (Fig. 2). Inoculation two weeks later resulted in scattered lesions (Fig. 3), while the inoculation on June 28 gave negative results (Fig. 4). The same phenomenon but from a different approach is evident in Plate 128, Figs. 1 and 2, which demonstrate the results obtained from inoculations on Crataegus Jonesae on May 7 and June 4 respectively. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 113 All the leaves in both inoculations received approximately the same amount of inoculum per unit area of leaf. At the time of the first inoculation the five basal leaves were well expanded, while the two upper leaves were just beginning to expand. As may be noted in Fig. 1, much heavier infection occurred on the older leaves. (The large irregular white areas on the younger leaves are holes caused by insects.) In Fig. 2, showing the results of the later inoculation, the reverse situation is seen; on the younger leaves at the end of the twig abundant infection was obtained, while the older leaves exhibited cnly scattered lesions. It is quite evident, therefore, that the cuticle cannot be the sole deter- mining factor fer variation in susceptibility throughout the entire life of the foliage; certain physiological factors may also be involved. For example, the leaves apparently are not so susceptible during the period of emergence from the winter scales until they are in a moderately ad- vanced stage of expansion, a period prior to any heavy deposition of cuticle. It is possible that the rust is unable to establish itself in the very young leaf. However, since this rust is not primarily of a systemic nature, probably the dilution effect on the number of lesions resulting from the intussusceptional type of foliar growth and consequent expan- sion, as well as the relatively small leaf area exposed at the time of inoculation, will account for the major part of this phenomenon. Again, even the most susceptible groups, for example, ANOMALAE or COCcCcI- NEAE, are apparently quite resistant to the rust by the latter part of June, at which time the leaves have by no means the amount of cuticle that is formed on the CRUS-GALLI even in the early part of May. It is possible that the rust is unable to establish a nutritional regime in the mature leaf as exhibited in the latter part of June, a point in favor of assuming a physiological antagonism on the part of the host. The rela- tively high temperature may also be a factor, by inhibiting spore germination. Nevertheless these two periods play an insignificant part in any deter- mination of the amount of infection that may accumulate on a host, regardless of the group. In the former case the period is relatively short and the leaf area exposed to the basidiospores by the unfurling buds would be small in comparison with that exposed after the leaves have expanded. As for the latter case practically all the teliospores on the red cedar have germinated by May 25, and the degree of susceptibility of any pomaceous host after the last of May would have no significance in determining the amount of infection that might occur. Thus, for prac- tical purposes in the field the significant period within which infection might take place is between the time the leaves are fairly well expanded 114 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI and the end of basidiospore dispersal. During this time the thickness and rapidity of deposition of the cuticle are the deciding factors. For this reason the inoculations in April and June, respectively, are not con- sidered in determining the relative susceptibility of the various groups. DURATION OF SUSCEPTIBILITY PERIOD > - 4 AW COCCINEAE a B. MACRACANTHAE a C_ CRUS-GALLI o. 3 A Dj ENTIRE GENUS oO 1) = 7p) w 2 8 fe) uJ LJ a = | W q fa) q (@) T ' a « ~ Nu © cat > N N J < > rf] . = 5 < i) TIME IN DAYS BETWEEN INOCULATIONS Fic. 3. ILLUSTRATING THE DEGREE AND THE DURATION OF THE PERIOD OF SUSCEPTIBILITY WITHIN THE GENUS CRATAEGUS TO G. GLOBOSUM. To illustrate further the degree and duration of susceptibility within the different groups, values may be obtained for the relative degrees of susceptibility of the various groups by taking the sum total of the values as expressed by the symbols 0, 1, 2 and 3, and dividing by the number of representatives inoculated.'. These were obtained from Table I for the CoccrNEAE, MACRACANTHAE, and CRUS-GALLI, which are, respec- tively, typical of the classes very susceptible, moderately susceptible 1The objection arises that such a method of correlation utilizes arbitrary qualita- i hho y to designate quantitative entities. Nevert seep its usage here is not nsidered from a statistical stan dpoin nt and it does sent a clearer picture 8 oat ane both the sane and the duration of Pe e per ries of Se within any one group of Crat t is ee ang to note that if such a metho employed in erie a oy ‘data obtained from serial geal ere in this genus (under sub-sect ai one will arrive at precisely the same conclusions as in the method finally a rare 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 115 and resistant, and have been plotted in Fig. 3. A similar curve (in heavy line) is given for all the inoculated representatives of the genus. The area enclosed by the respective curves would, to a certain extent, be a measure of both the degree of susceptibility and its duration. The TABLE II PRESENTING DATA ON THE RELATIVE SUSCEPTIBILITY OF CRATAEGUS TO G. GLOBOSUM, AS INDICATED BY NATURAL INFECTION No. trees in the classes Group No. | No. Rel. degree of Sps. | Trees o | Mild . | Sev. | susceptibility infect.| infect.| infect.| infect. IAMOMal ae meta si ho Ma ates 19 40 0 19 3 18 2.95 ATO ee Cele ue oe Nyda ee 1 1 1 0 0 0 .00 Bracteatae............:. 2 2 1 1 0 0 .50 Gocemede 22h nits .cage ke 46 82 2 24 20 36 3.22 Crs-gallic wither ts oe 452 71 128 80 46 2 0 41 Dilatatae s.athtateonea 4 11 1 6 0 4 2.36 Douglasianae) 25-34. 8 19 3 16 0 0 84 BUC kigcr mice anes 10 11 11 0 0 0 .00 Intricataes.7 ee. eens 10 11 8 3 0 0 SX Macracanthae........... 68 99 4 78 10 7 1.44 Microcarpaés.ac6 seen: 1 1 1 0 0 0 .00 MOllesr ice cent ha SF 86 10 40 11 25 2.30 INIOTACRE onsen tee 2 2 0 2 0 0 1.00 Oxycanthdes ns 10 17 10 6 0 1 .65 Pinnatiidae--.. 2. 2a. 2 4 0 4 0 0 1.00 IPTUINOSAE. 3. eer ee 58 98 36 57 3 2 TOL Bunctatae. sea. oer: 34 37 7 26 y 2 1.14 Rotundifoliae........... oF 66 14 | 37 8 7 1.45 Sanguinae ween ayers 4 4 0 + 0 0 1.00 Silvicolae.2c+, atte ees 35 57 3 42 8 4 51 Tenuifoliae............. 81 175 7 104 32 37 2.06 erillorae=.. 6. ence 2 2 2 0 0 0 .00 Wniflorae.. since eos ee 2 2 2 0 0 0 .00 Mini ese yon ey eens 18 30 20 9 1 0 .40 CoccINEAE, characterized by little foliar cuticle, exhibit a much higher degree of susceptibility over a relatively longer period of time than the CRUS-GALLI which have consistently heavy cuticle on the leaves, whereas the MACRACANTHAE, with an intermediate and varying amount of cuticle, assume an intermediate position. 116 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI D. CoRRELATION OF THE DATA TO CLASSIFY THE GROUPS OF CRATAEGUS WITH RESPECT TO THEIR RELATIVE SUSCEPTIBILITY Bearing in mind that the thickness of the cuticle and its rapidity of deposition on the leaves are the primary factors in determining the rela- tive susceptibility of any host, while geographical distribution and physiological antagonism on the part of the host play a very minor part, if any, it is now possible to evaluate.the data obtained by the two pre- viously described methods of approach and determine the relative susceptibility of the various groups within the genus Crataegus. 325 300, 2-75 2:50 225 > 2-00 re = 3 ow fe NTs a w a Q 150 ° | we O is Ww WwW 5 1-00 w W ie “75 50 25 ie) vq Wi -00 N - ww www Www We esses ge<¢csaececec ec Cu eee ee Ot JeuzgezZ@ekrsaerogei zd? eS zZzGESSSESEZLPERC ZL SS BSS Zep uti eve eoeetSzeges ef get 5 7uv84 5 eo vu fe ZaZz%5 << $2640 Zs <5 *% 2 845 5 o Y% 8 s) x a 5 - 3 < Ss « Fic. 4. RELATIVE SUSCEPTIBILITY OF THE GENUS CRATAEGUS TO G. GLOBOSUM AS INDICATED BY OBSERVATIONS ON NATURAL INFECTION. The number within each column refers to the number of trees considered within the group. e) The relative degrees of susceptibility obtained from observations of natural infection, as previously stated, are presented in Table II, and have been plotted in Fig. 4. In regard to data obtained by serial inocula- tions, it is quite obvious from Table III that inoculations before the leaves unfurl, and again late in June, result in very little infection. « However, as the foregoing discussion on the duration of the period of 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 117 susceptibility demonstrates, such a phenomenon, while interesting, plays an insignificant role in determining the amount of infection that might take place on any tree. The two significant inoculations are those made TABLE Ill PRESENTING DATA ON THE RELATIVE SUSCEPTIBILITY OF THE GENUS CRATAEGUS TO G. GLOBOSUM AS INDICATED BY SERIAL INOCULATIONS Percentage of species within the various groups of Crataegus falling into the dif- ferent classes of susceptibility,as indicated by serial artificial inoculations (a) (bd) tifi 1 ulation A 23 Artificial inoculation May 7 and 8 No. % species within classes No. % species within classes Group species 0 Ee, 85, Se Group species OQ 1 2 3 lae 100.0 0.0 0.0 0.0 An ae . . 0? 14. Bracteatae 100.0 0.0 0.0 0.0 Bracteatae +O 100. ’ . Coccineae 3 0.0 100.0 0.0 0,0 Coccineae 0. -O 100. Crus-galli 43 100.0 0.0 0.0 0,0 Crue-galli 48 25.0 39 22.9 12. Dilatatae = a = = Dilatatae . rv) 50. 50. Douglasianae . . Douglasiana . 0.0 66. 33, Flavae 100. . Flavae 10 10. 50.0 40. i Intricatae 100. . Intricatae 11 18. 36. 36. . Macracanthae 1 92. 7 . . Macracanthae 20 15. 10.0 30. 45. Microcarpae 100. . Liicrocarpae 100. QO. . +0 Kolles 83.2 1 . Molles 1 «2 12.5 62. 18. 0 captha 50. 50. . Oxyacanthae . 66. 33. ° Pruinosae 96. ’ . . Pruinosae 3 el. 37. 24. 16.4 Pulcherrimae 100 . . . Pulcherrimae 1 100. Oo. ° ° Punctatae 1 85. 14. . . Punctatae 1 . 16. 38. 38.9 Rotund tfoliae 87 12.5 . . Rot el ° 15. 53. 23. Silvicolae - - - Silvicolae 25. 50. . 25. Tenuifoliae 40.0 60.0 0.0 Tenuifoliae 14 . 14. 21. 64.3 Triflorae 100.0 0.0 0.0 0.0 r rae 50.0 50. . . Uniflorae 2 100.0 0.0 0.0 0.0 Uniflorae 2 -O 100. . . Vir 70.0 20.0 10.0 0.0 Virides 13 - 30. el 46.1 (c) (da) Artificial inoculation May 2 2 and 33 Artificial inoculation June 28 No. % species within classes No. % species within classes Group species OQ 4 2 3 Group species O S33 aS e 0.0 0.0 0.0 100.0 nomala 100.0 0.0 0.0 0.0 acteatae v.0 100.0 0.0 0.0 Bracteatae - - - - Coccineae 0.0 0.0 100.0 0.0 Coccineae 100.0 0.0 0.0 0.0 Crus-galli 78 69.2 20.5 10.3 0.0 Crus-galli 22 100.0 0.0 0.0 0.0 Dilatatae - - = = Dilatatae = = = ad Douglasnae 66.7 3 oO. ° Doujlasianae 100. Flava 18.2 5 27. : Flavae 80. 20. Intricat 12.5 6 25. ( Intricatae 0. . . Macracanthae 16 37.5 3 25 Macracanthae 83. 16.7 Mi 8: bw. oO. Microcarpae 100. . . Molles cr ay pe © Ti. 8 olles 1 . : Oxyacanthae 11 27. 5 18. Oxyacanthae 100 . . Pruinosae 32.6 34.9 23 . Pruinosae 23 95 3 0. Pulcherrimae C. 100. . Pulcherriwae 0. 100.0 0. Punotatae 20 30. 5 10 1 Punctat 18 87. -O 12.5 . Rotundifoliae 12 25.0 5 16 tundifoliae 6 83. 16. 0 0. 1 1100. ie) S8ilvicolae - 7 = Tenuifoliae 14 22.2 4 llel = 82. Tenuifoliae L 100.0 0.0 0.0 0.0 Triflorae 21lu0. 0.0 ° Triflorae = = - a Uniflorae 3 wo. . 0.0 . Uniflorae ) = _ = = Virides 17 64. 17.6 17.6 ° Virides 5 100.0 0.0 06.0 0.0 in May, (b) and (c), and for fifteen of the major groups the percentage frequency of occurrence of inoculated representatives falling into the respective classes of susceptibility have been plotted in Fig. 5 (p. 118). In comparing these tables and figures to make a final classification of the groups according to their relative susceptibility, one must remem- ber that these results were obtained from two altogether different meth- 118 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XVI ods of approach. For those groups the representatives of which have a heavy cuticle, a much lower degree of susceptibility would be indicated by natural infection than by artificial inoculation where the amount of inoculum and the cultural environment are optimum. The number of representatives examined in each group, and especially for natural in- fection, must also be given consideration. ANOMALAE COCCINEAE CRUS-GALLI DILATATAE DOUGLASIANAE 100 100 100 100 50 50 so 50 ok. 0 ° ° ' O1re3 Or23 ‘Ores Ol oVre3 (¢) 2 ve3 ' O123 @ 10} @) (4) (48) (78) (2) (0) (3) @ FLAVAE INTRICATAE MACRACANTHAE MOLLES OXYACANTHAE 00 00 100 00 ce] . | Ol ! ! cio) ia) an (8) (20) (16) (16) (9) (6) ap FREQUENCY OF OCCURRENCE (%) WITHIN SUSCEPTIBILITY CLASSES wo ° ° } en = ° ny Ww - 2 ° ° 7) ° a = ° ° PRUINOSAE PUNCTATAE html TENUIFOLIAE VIRIDES 10 100 10 100 100 50 50 50 50 50 ° ° fe) ok ° : ?, gn (43) (18) (20) (13) (2) (4) (4) G3) an CLASSES OF SUSCEPTIBILITY Fic. 5. RELATIVE SUSCEPTIBILITY OF FIFTEEN GROUPS OF THE GENUS CRATAEGUS TO G. GLOBOSUM AS INDICATED BY SERIAL INOCULATIONS. The results of two inoculations, (b) and (c) respectively, are presented in each sub-graph. The numbers on the abscissae of the sub-graphs refer to the classes of susceptibility. The numbers in parentheses refer to the number of species (with the exceptions noted in text) inoculated in each group. By correlating the degree of susceptibility as indicated by natural infection, and the frequency of occurrence of inoculated representatives falling into the various classes of susceptibility, the groups may be classi- fied and arranged within each class according to susceptibility as follows:' Very susceptible—ANOMALAE, COCCINEAE, TENUIFOLIAE, DILATATAE. 1In classifying these groups according to their relative susceptibility, ring for minor groups, not included in the figures, were taken directly from the table 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 119 Moderately susceptible—Mo Lies, MACRACANTHAE, ROTUNDIFOLIAE, PUNCTATAE, DOUGLASIANAE, SILVICOLAE, PRUINOSAE, VIRIDES, FLAVAE, OXYACANTHAE, INTRICATAE, Resistant—CRUS-GALLI, BRACTEATAE, *AZAROLI, *MICROCARPAE, *NIGRAE, *PINNATIFIDAE, *SANGUINEAE, *TRIFLORAE, *UNI- FLORAE.' Immune—None. None of the groups examined proved to be wholly immune. No in- fection was obtained on the one representative of the MICROCARPAE, namely, C. Phaenopyrum (L. f.) Medic. (= C. cordata Ait.), but this species has been previously reported as a host to the rust from both Delaware and Tennessee. Of almost five hundred and fifty determined species and varieties studied, less than one percent of the artificial in- oculations gave negative results and it is indeed possible that, given optimum conditions for germ-tube penetration, not a single species could be considered totally immune. However, as previously stated, it must be remembered that the conditions favorable for infection set up by artificial inoculation far exceed any that might occur in nature, and many species that are now classed as suscepts would probably never exhibit infection under field conditions. E. SUGGESTIONS FOR THE SELECTION OF RESISTANT SPECIES AND VARIETIES The best guide in the selection of Crataegus trees to be planted on estates where G. globosum is in the vicinity would be the thickness of the foliar cuticle. A striking example of this was found on an estate at Canton, Massachusetts, where two Crataegus trees, one a COCCINEAE species and the other a CRUS-GALLI species, were planted side by side, surrounded by red cedars bearing heavy infections of G. globosum. These have been under observation for the past three years, and each season the foliage on the CoccrNEAE species has suffered very severe infection, resulting in more than eighty percent defoliation by the latter part of August. The tree is now in a very weakened condition. The CRUS-GALLI species, on the other hand, has been entirely unaffected by this rust. In choosing from species of American origin one should definitely avoid the ANOMALAE, CoccINEAE, TENUIFOLIAE and Divaratak if G. globosum be in the vicinity. Certain of the species within the groups 1The small number of representatives in the resistant groups indicated by asterisks made it impossible to arrange these groups within the class ‘‘Resistant” according to susceptibility and they Hee been arranged "alphabetically. 120 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI classed as moderately susceptible have considerable cuticle on the leaves and these may be planted with a relative degree of safety. The CRUS-GALLI, however, are very resistant, and offer a wide variety of species. They are, as Rehder (1927) states, handsome ornamentals with dense, dark green foliage which remains till late in autumn or early winter, and are very attractive in bloom, with decorative bright red fruits that are persistent during the winter. If one desires the Eurasian type, the PINNATIFIDAE offer a group with lustrous leaves and large showy fruit. Some varieties of these are cultivated in northern China for the edible fruit. The OxyACANTHAE will also withstand severe in- fection unless under abnormal proximity to Juniperus rusted by G. globosum, with C, Oxyacantha Jacq. including some of the most showy garden forms. This presentation has been confined to foliar lesions, and while infec- tion has been obtained on all parts of the flower as well as the fruit and young twigs, such instances were sufficiently rare that they were not worthy of consideration at this time and have been set aside for a sec- ond publication on the life history of G. globosum Farl. No consideration has been given to the possibility of variation in virulence within different strains of this rust. Practically all the inocu- lum was obtained from two adjacent red cedar trees at Waltham, Massa- chusetts. One must also bear in mind that the relative susceptibility of groups within the genus Crataegus to G. globosum is in no respect correlated with their susceptibility to other Gymnosporangium rusts. Crowell (unpublished) has found, for example, that the CRUS-GALLI, so resist- ant to G. globosum, are quite susceptible to G. clavipes Cke. & Pk. Pyrus — RELATIVE SUSCEPTIBILITY AS INDICATED BY SERIAL INOCU- TIONS Studies on relative susceptibility within the genus Pyrus were con- fined to the results obtained from serial inoculations made in 1934. The species represented in the Arboretum were artificially inoculated in a manner similar to that described for Crataegus: (a) on April 25, at which time the condition of the foliage varied from buds just bursting through the winter scales to leaves a quarter to a half inch long; (b) on May 9, when the leaves were fairly well expanded on all species; (c) on May 22 when the foliage was fully expanded; and (d) on June 28. Cer- tain of the species which had given negative results in the previous in- oculation were omitted in the June inoculation. In Table IV are given the species inoculated, their distribution, the degree of infection obtained on the respective dates of inoculation, the 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 121 stages of the rust produced on the foliage, and finally, a classification of their relative susceptibility. TABLE IV PRESENTING DATA ON THE RELATIVE SUSCEPTIBILITY OF SPECIES OF THE GENUS PYRUS TO G. GLOBOSUM, AS INDICATED BY SERIAL INOCULATIONS Deg. suscept. indicated by inoculations Stages Degree Species distrib. (a) (b) (c) (d) found suscept. | eae nsae Decne. Eurasian Oi.a25-0° 0 0&1 2 Pp. Seen Bge. Eurasian 3. S32 eo 0&1 3 P. Bretschneideri Rehd. Eurasian 0 1 0 - 0&1 1 P. communis L. urasian O INOS 0 1 P. elaeagrifolia Pall. Eurasian 0 OO O - fe) P. Korshinskyi Litv. Eurasian 1 0 O0O - 0 1 P. Michauxii Bosc.! — 0 1 0 -=- 0) 1 P. Lindleyi Rehd. Eurasian T6380 0) 1 P. nivalis Jac Eurasian 0 1 0 - 0) 1 Re See Rehd. Eurasian O: L Ad 0 0&1 1 P. salicifolia Pall. Eurasian 0°05 “OR 0 P. serotina Rehd. Eurasian 0 2 0 - 0&1 2 | ede ue Rehd. Eurasian 1 1 1 O 0 1 P. syriaca Boiss. Eurasian O° - tb 1 0 0&1 1 P. ussuriensis Maxim. Eurasian j eee ee Oye 0&1 1 1P. Michauxii is a hybrid (P. amygdaliformis * P. nivalis). No consistent correlation between the relative susceptibility of the various species and the type of leaf is evident; all species have consider- able cuticle on the foliage, and a few are somewhat tomentose. Nor can the differences in susceptibility be correlated with the distribution of the host. The lesions in general were found to be much smaller than those ex- hibited on Crataegus, and except in the case of P. betulaefolia rarely measured more than one to two millimeters in diameter. Certain spe- cies, designated in the table, showed spermogonia only; the lesions were extremely small, and died before any hypertrophy or aecial formation was evident. However, it is possible that with a different strain of the rust some of these might produce aecia; P. communis, for example, ex- hibited only spermogonia in my inoculations but has been reported pre- viously from seven different states. As in Crataegus, there is a definite duration to the period of suscepti- bility, the degree of which reaches its maximum during and immediately after the period of foliar growth and expansion, and then falls off gradu- ally so that by the end of June all species examined are immune. 122 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI Classified according to their relative susceptibility, the species exam- ined may be arranged (alphabetically) as follows: Very susceptible—P. betulaefolia Bge. Moderately susceptible—P. Balansae Decne., P. serotina Rehd. Resistant—P. Bretschneideri Rehd., P. communis L., P. Korshinskyi Litv., & P. Michauxii Bosc, P. Lindleyi Rehd., P. nivalis Jacq., P. phaeocarpa Rehd., P. serrulata Rehd., P. syriaca Boiss., P. ussuri- ensis Maxim. Immune—P, amygdaliformis Vill., P. elaeagrifolia Pall., P. salicifolia Pall. Previous reports of Pyrus suscepts are confirmed, for the most part, to P. communis, to the Kieffer Pear (P. communis & P. serotina) and other varieties used commercially in the orchard. Stevens and Hall (1910) report G. globosum as being particularly abundant on the Japa- nese strain of pear (P. serotina), while Stewart (1910) reports the Kief- fer pear as suffering infection from this rust at Long Island, New York. In particular he finds that both the fruit and leaves are attacked, and that the diseased fruits are very small and misshapen, usually exhibiting circular black areas devoid of aecia, although a few show aecia. On the other hand, Stewart (1910), and Hesler and Whetzel (1917) classify the Bartlett, Bosc, Duchess, and Worden varieties as being for the most part immune, although the fruit of the Worden variety is subject to infection. While little can be added to the knowledge of the relative suscepti- bility of the orchard varieties, one may conclude from the foregoing classification that, with the exception of P. betulaefolia, P. Balansae, P. serotina, and as indicated from previous reports, P. communis, the re- mainder of the species can be safely planted in vicinities where the rust is present. This conclusion holds true especially for P. amygdaliformis, P. elaeagrifolia, and P. salictfolta. Sorbus — Revative Susceptipi_ity As INDICATED BY SERIAL [NocU- -ATIONS Serial artificial inoculations were made in 1934 on species and vari- eties of Sorbus available in the Arnold Arboretum: (a) on April 25, at which time the foliar buds were just beginning to break open and the tiny leaves in many cases exhibited a heavy tomentose covering which was removed without injury to the leaf by rubbing the latter between the fingers, and the inoculum was placed on the exposed green tissue; (b) on May 9, at which time practically all the foliage was going through a period of rapid growth and expansion; (c) on May 24, at which time the leaves were fully expanded (blossoms where present were 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 123 also inoculated); (d) on June 28, at which time the leaves for all prac- tical purposes were mature. The results of these inoculations appear in Table V which presents, where positive results were obtained, the species and varieties inocu- lated, their native distribution, the degree of infection obtained from the respective inoculations, the stages of the rust exhibited, and finally the resultant classes of susceptibility. TABLE V PRESENTING DATA ON THE RELATIVE SUSCEPTIBILITY OF SPECIES AND VARIETIES OF THE GENUS SORBUS TO G. GLOBOSUM, AS INDICATED BY SERIAL INOCULATIONS Deg. suscept. indicated by Native inoculations Stages Degree Species and varieties: distrib. (a) (b) (c) (d) found suscept. S. americana “Mar sh. American ir 3 2 eo O&l 3 S. americana var. fructu albo* Hort. merican 1 1 0 - 0&1 1 S. americana var. nana Hort. American 1 0 oO - 0 1 S. arnoldiana Rehd.! Eurasian tat oO = 0 1 S. Aucuparia var. Backhousei Eurasian 1 0 0 - 0 1 S. dumosa Greene American Te MOMs Or = 0 1 S. japonica var. calocarpa ehd. Eurasian in w0 PEO: w= 0 1 S. thuringiaca Fritsch? Eurasian 5 ieee ee 0 0&1 1 1§. arnoldiana is a hybrid (S. Aucuparia < S. discolor). 2S. thuringiaca is a hybrid (S. Aucuparia X S. Aria). No infection was obtained on the following (alphabetically arranged ) species and varieties, which are all of Eurasian origin: Sorbus alnifolia K. Koch, S. amurensis Koehne, S. Aria Crantz, S. Aria var. angustifolia Hort., S. Aria var. Decaisneana Rehd., S. Aria var. longifolia Pers., S. Aria var. lutescens Hartwig, S. Aria var. magnifica Hort., S. Aria var. salicifolia Myrin, S. Aria var. sulphurea Hort., S. Aria var. theophrasta rt., S. Aucuparia L., S. Aucuparia var. Dirkenti aurea Port... Aucuparia var. edulis Dieck, S. Aucuparia var. nana Hort., S. Aucu- paria var. xanthocarpa Hartw. & Ruempl., S. commixta Hedl., S. com- mixta var. rufo-ferruginea Schneid., S. discolor Hedl., x S. hybrida L., S. intermedia Pers., X S. latifolia Pers., S. latifolia var. atrovirens Hort., S. Matsumurana Koehne, < S. Meinichii Hedl., S. pohuashanensis Hedl., S. Zahlbruckneri Schneid. 124 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI All species of American origin that were inoculated proved to be susceptible, with S. americana as the only species on which the foliage was materially injured by the rust. Of the thirty-one inoculated Eura- sian types, infection was obtained on only four, and these proved to be quite resistant. 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 sorus. Those species on which spermogonia only were obtained (see Table V) exhibited bright yellow lesions until the spermo- gonia were mature, following which time no further development took place and the infections died. An interesting type of natural infection was observed on Mt. Monadnock in New Hampshire; the lesions were as large as any ever obtained on Crataegus, some being as much as ten to twelve millimeters long, each bearing abundant aecial horns. Whether this type of infection results from a more susceptible variety of S. americana, or another strain of G. globosum, is not known. With the exception of S. americana, no infection was obtained on any of the species after the second inoculation, while practically all the sus- cepts exhibited some infection from the initial inoculation. It would seem, therefore, that the resistant forms at least are most susceptible during, and immediately after, the period when the foliar buds are unfurling; S. americana, however, reached its maximum degree of sus- ceptibility immediately after the leaves had expanded. It is extremely doubtful that, with the exception of S. americana and its varieties within the American types, and possibly the hybrid Eura- sian type, S. thuringiaca, any representative of the genus Sorbus would be seriously affected by G. globosum regardless of proximity to the rust. This is certainly true for the species of Eurasian origin. Malus— Revative Susceptipitity as INDICATED BY SERIAL INocu- LATIONS Serial artificial inoculations were made in 1934, similar to those de- scribed for the preceding genera: (a) on April 24, at which time the leaves had already unfurled and were undergoing the period of rapid expansion; (b) on May 9, at which time the foliage was almost mature size, and most of the blossoms were in the pink stage; (c) on May 22, at which time most of the petals had dropped. No inoculation was made in June. Table VI presents the species on which positive results were obtained, the origin of the various species, the results obtained from the respective serial inoculations, the stages of the rust obtained, and finally the relative degree of susceptibility. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 125 TABLE VI PRESENTING DATA ON THE RELATIVE SUSCEPTIBILITY OF SPECIES AND VARIETIES OF THE GENUS MALUS TO G. GLOBOSUM, AS INDICATED BY SERIAL INOCULATIONS Deg. suscept. indicated by Native inoculations Stages Species and varieties distrib. a) (b) (c) ound _ suscept. M. astracanica Dum. -Cours.! 1 Eurasian Oo 1 0 O&l 1 M. baccata Borkh. Eurasian 1-2-0 0&1 2 M. coronaria Mill. American OP 10 0) if M. Dawsoniana Rehd.? Hybrid es Cae | 0&1 1 M. glabrata Rehd. American Oo 1 C O&l 1 M. ioensis var. plena Rehd. American i. 32> ak O&l 2 M. magdeburgensis Schoch? Eurasian Oo 1: 0 0 1 M. Soulardi Britt.4 Hybrid - 2 0 0&1 2 M. sublobata Rehd.® Eurasian Oo 1 0 0&1 1 1M. astracanica is a hybrid (M. prunifolia x M. ne °M. Dawsoniana is a hybrid (M. fusca * M. pumila). 3M. magdeburgensis is a hybrid (M. pumila X we spectabilis). 4M. Soulardi is a hybrid (M. ioensis * M. pumila). 5M. sublobata is a hybrid (M. prunifolia *« M. Sieboldii). The following species, alphabetically arranged according to distri- bution, gave negative results: American distribution: Malus angustifolia Michx., M. bracteata Rehd., M. fusca Schneid., M. glaucescens Rehd., M. ioensis Britt., M. lancifolia Rehd., M. platycarpa Rehd. Eurasian distribution: & Malus arnoldiana Sarg., M. asiatica Nakai, < M. atrosanguinea Schneid., M. brevipes Rehd., M. floren- tina Schneid., M. floribunda Sieb., M. Halliana var. Park- manitt Rehd., «K M. Hartwigiti Koehne, M. honanensis Rehd., M. kansuensis Schneid., M. micromalus Mak., M. hupehensis (Pamp.) Rehd. (= M. theifera Rehd.), M. pumila Mill., M. prunifolia Borkh., < M. purpurea var. Eleyi Rehd., «* M. robusta Rehd., M. Sargenti Rehd., M. Sieboldii Rehd., M. sikkimensis Koehne, M. spectabilis Borkh., M. sylvestris Mill., M. toringoides Hughes, M. Tschonoskii Schneid., M. yunnanensis var. Veitchii Rehd., * M. zumi Rehd. A variety of an American species, M. ioensis var. plena, and the hy- brid M. Soulardi proved to be moderately susceptible to G. globosum, while two species, M. coronaria, and M. glabrata, and the hybrid M. Dawsoniana, may be classed as mildly susceptible. On the remainder of the American species inoculated no infection could be observed; never- theless, Thaxter (1889) obtained aecia on M. pumila Mill. (= M 126 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XVI Malus Britt.). Of all the Eurasian species inoculated only one proved to be moderately susceptible, namely M. baccata, and three hybrids between Eurasian species, M. astracanica, M. magdeburgensis and M. sublobata, proved to be mildly susceptible. Although a higher percentage of the American species proved to be susceptible, no outstanding correlation could be observed between rela- tive susceptibility and geographic distribution. Nor can susceptibility be correlated with the type of leaf or type of infection produced. In all cases the lesions were small; they were rarely more than one to two millimeters in diameter. The serial inoculations indicated a definite duration to the period of susceptibility which reaches a maximum about the time the blossoms are in the pink stage, and falls off to almost zero within a period of two weeks. Excluding the species found to be susceptible it is very doubtful that any of the remaining species considered would suffer from the rust re- gardless of proximity to red cedars infected by G. globosum. Previous reports would indicate that the commercial varieties of apple are more susceptible than the above ornamental types. Bliss (1931) using telial material from Iowa culturally obtained flecking on the vari- eties Baldwin, Delicious, Fameuse, Greening, McIntosh, Tolman, Wealthy, Yellow Transparent, and York Imperial. From reports of Clinton (1934), Thomas and Mills (1930), Sherbakoff (1932), Miller, Stevens and Wood (1933), and others, the relative susceptibility of the commercial varieties of apple may be classified as follows Varieties on which moderate to severe infection has been observed: Fallawater, Fameuse, Hubbardston, Northwestern Greening, Rhode Island Greening, and Wealthy. Varieties reported susceptible: Baldwin, Cortland, Esopus, Spitzen- burg, Fall Pippin, Gano, Golden Delicious, Jonathan, McIntosh, New- ton, Northern Spy, Pewaukee, Rome Beauty, Russett, Stark, Tolman Sweet, Tompkins King, Wagener, Winesap, and York Imperial. Resistant variety: Ben Davis. Amelanchier' Farlow (1885) obtained spermogonia on leaves of Amelanchier cana- densis Med. and Harshberger (1902) lists the same species as a suscept to G. globosum, exhibiting both spermogonia and aecia. Stone (1908) lists A. alnifolia’ as a suscept from Alabama. The following species and 1Relative atime in this and the following genera was determined by non- serial inocula °This tae refers to A. canadensis or A. laevis, since A. alnifolia is not native ama. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 127 varieties of Amelanchier were inoculated early in May, 1933: Amelan- chier amabilis Wieg., A. asiatica Endl., A. Bartramiana Roem., A. Bartramiana < A. laevis, A. canadensis Med., A. florida Lindl., X A. grandiflora Rehd., A. humilis Wieg., A. humilis X A. sanguinea, A. intermedia Spach, A. laevis Wieg., A. oblongifolia Roem., A. ovalis Med., A. sanguinea DC., A. sera Ashe, A. spicata K. Koch, A. stolonifera Wieg. All the inoculations gave negative results. No reports can be found indicating that any of the species and vari- eties of Amelanchier are very susceptible to G. globosum. Cydonia Thaxter (1888) by culture obtained spermogonia on Cydonia oblonga Mill. (= C. vulgaris Pers.). Cook (1913) reports G. globosum as being of common occurrence on quince in New Jersey. Harshberger (1902), Clinton (1904), and Giissow (1915) report this rust on quince from two other states and from the Niagara Peninsula. Cydonia oblonga, inoculated by the writer in early May, 1933, proved to be moderately susceptible to G. globosum, producing both spermogonia and aecia. None of the varieties of Cydonia oblonga was inoculated, and no in- formation can be given with respect to their relative susceptibility. The remaining smaller genera were artificially inoculated and the results from these inoculations may be summarized and tabulated as follows: Comptonia Comptonia aspleniifolia Ait—immune. Crataegomespilus Crataegomespilus grandiflora Bean (Crataegus Oxyacantha X Me- spilus germanica)—very suceptible; both spermogonia and aecia ob- tained: severe leaf killing resulted. Natural infection was also observed. Mespilus Mespilus germanica L.—moderately susceptible; both spermogonia and aecia obtained. Myrica Myvrica caroliniensis Mill—immune. M. Gale L.—immune. Photinia Photinia villosa DC.—immune. Sorbaronia Sorbaronia alpina Schneid. f. superaria Zabel (Aronia arbutifolia X 128 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XVI Sorbus Aria)—resistant; exhibited spermogonia only. Aronia flori- bunda * Sorbus Aucuparia—no infection obtained. Sorbopyrus Sorbopyrus auricularis Schneid. (Pyrus communis < Sorbus Aria )— resistant; exhibited spermogonia only. III. RELATIVE SUSCEPTIBILITY OF HOSTS WITHIN HE GENUS JUNIPERUS To our present knowledge of the relative susceptibility of Juniperus little can be added by the writer. From previous reports, including those of Adams (1919), Arthur (1926) (1927), Bliss (1933), Claassen (1897), Connors (1934), Hunt (1926), Kern (1929), Martin (1922) (1925), Stone (1909), and others, and from an examination of the material in the Farlow Herbarium and the herbarium of Professor J. H. Faull, the host list includes at least six species of Juniperus, and at least four varieties of Juniperus virginiana. These have been presented in the subsequent host list. It may be added here that Martin (1922) lists Larix species as hosts to G. globosum from nine states. No infection by this rust has ever been observed on Larix in the Arnold Arboretum. Juniperus virginiana is the most common telial host throughout the eastern and central part of North America, having been reported from twenty-five states and from Ontario. Severe infection may occur, as exemplified at the Morton Arboretum, Lisle, Illinois and from many estates and nurseries surrounding Boston. The writer has observed trees that were killed by the abundance of galls present. Other trees, while not killed, were disfigured to such an extent that they were no longer of ornamental value and had to be removed. Juniperus scopu- lorum has also been reported as suffering from infection by G. globosum at the Morton Arboretum. As far as the eastern and central part of North America are concerned no information to date would indicate that any species other than Juni- perus virginiana and Juniperus scopulorum and their varieties would suffer to any extent from infection by G. globosum. IV. THE HOSTS OF GYMNOSPORANGIUM GLOBOSUM FARL. The following list includes as far as can be ascertained all the known hosts of G. globosum. The hosts have been arranged alphabetically by genera and their included species. Within the genus Crataegus the species and varieties have been arranged within their respective groups. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 129 Following each host name in parentheses are symbols which may be defined as follows: a—as obtained by inoculations made by the writer; the inclusion of an author’s name and reference indicates that this host has been determined previously by inoculation. n—as determined by observations of natural infection made by the writer. The inclusion of the abbreviated name of a State implies that this species has been reported previously as a host from that State. All new hosts submitted would necessarily be records for the State of Massachusetts, as all studies were made in the Arnold Arboretum, Boston. HOSTS FOR THE 0 & 1 STAGE An asterisk preceding a host indicates that the 0 stage only was found. AMELANCHIER: Amelanchier alnifolia Nutt. (Ala.),1 A. canadensis Med. (Thaxter [1885]; Penn.). CRATAEGOMESPILUS: Crataegomes pilus grandiflora Bean (a; n). CRATAEGUS (by groups): ANOMALAE: Crataegus affinis Sarg. (a; n), C. asperifolia Sarg. (a; n; Vt.), C. Brockwayae Sarg. (a; n), C. Coleae Sarg. (n), C. cyclophylla Sarg. (a; n; Vt.), C. Dunbari Sarg. (a; n), C. Egglestonti Sarg. (a; n; N. Y., Vt.), C. errata Sarg. (a; n), C. honesta Sarg. (n), C. Jdeae Sarg. (n), C. improvisa Sarg. (n), C. misella Sarg. (n), C. pinguts Sarg. (n; Mich.), C. putata Sarg. (n), C. repulsans Sarg. (n), i Saundersiana Sarg. (n), C. scabrida Sarg. (a; n; Vt.), C. shirley- ensis Sarg. (a; n), C. urbana Sarg. (n). AZAROLI: Crataegus Heldreichii Boiss. (a), C. tanacetifolia Pers. (N. Y.). BRACTEATAE: Crataegus Ashei Beadle (a; n), C. Harbisonii Beadle (a; Tenn.). COCCINEAE: Crataegus acclivis Sarg. (n), C. arcuata Ashe (n; Penn.), C. assur- gens Sarg. (a; n), C. aulica Sarg. (n), C. caesa Ashe (n), C. chip- pewaensis Sarg. (n), C. confinis Sarg. (n), C. conspecta Sarg. (n), 1See foot-note on page 126. 130 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI C. contigua Sarg. (n), C. cristata Ashe (n), C. Dayana Sarg. (n), C. delecta Sarg. (n; Ill.), C. densiflora Sarg. (n), C. Eamesii Sarg. (n; Conn.), C. elongata Sarg. (n), C. fluviatilis Sarg. (a; n), C. fretalis Sarg. (n; Conn.), C. Hillii Sarg. (n), C. Holmesiana Ashe (a; n; Conn., N. Y., Vt.), C. Holmesiana var. tardipes Sarg. (n), C. Holmesiana var. villipes Ashe (n), C. irrasa Sarg. (n), C. lenta Ashe (n), C. lobudata Sarg. (n), C. Macounii Sarg. (n), C. miranda Sarg. (n), C. neolondinensis Sarg. (n; Conn.), C. pedicellata Sarg. (a; n), C. pedicellata var. gloriosa Sarg. (n), C. perrara Sarg. (n), C. polita Sarg. (n; previously reported, state not given), C. polita var. Tatnalliana (Sarg.) Eggl. (Mo., N. Y.), C. Pringlei Sarg. (a, Arthur [1907]; n; Conn., Ind., N. Y.), C. pura Sarg. (n), C. sejuncta Sarg. (n), C. sertata Sarg. (n), C. Thayeri Sarg. (n), C. uticaensis Sarg. (n), C. vivida Sarg. (n). CRUS-GALLI: Crataegus algens Beadle (a; n), C. arborea Beadle (a; n), C. ardu- ennae Sarg. (a; n; Ind.), C. armata Beadle (a), C. arta Beadle (a), C. attenuata Ashe (a; n), C. barbata Sarg. (a), C. barrettiana Sarg. (a), C. Bartramiana Sarg. (a), C. bellica Sarg. (a), C. calo- phylla Sarg. (a), C. Canbyi Sarg. (a; n), C. cerasina Sarg. (n), C. consueta Sarg. (a; Mo.), C. crus-galli L. (a, Thaxter [1891]: n; Ind., Ky., Maine, Mass., Miss., Mo., N. Car., Ohio, Penn., Tenn., Va.), C. crus-galli var. arbutifolia Hort. (a), C. crus-galli var. exigua (Sarg.) Eggl. (n), C. crus-galli var. pyracanthifolia Ait. (a; n), C. crus-galli var. rubens Sarg. (a), C. efferta Sarg. (a), C. effulgens Sarg. (a), C. Engelmannii Sarg. (a; n; Mo.), C. erecta Sarg. (a; n), C. Farwellii Sarg. (a; n), C. fecunda Sarg. (n), C. Fontanesiana (Spach) Steud. (a; n), C. geneseensis Sarg. (a), C. hamata Sarg. (a), C. hirtella Sarg. (a), C. infesta Sarg. (a: n), C. insignis Sarg. (a), C. jasperensis Sarg. (a), & C. Lavallei Herincq (a; n), C. lawrencensis Sarg. (a), C. leptophylla Sarg. (a; n), C. livoniana Sarg. (a; n), C. macra Beadle (a), C. Mohrii Beadle (a; n; Ga.), C. munita Sarg. (a), C. pachyphylla Sarg. (a), C. Palmeri Sarg. (a; n), C. paradoxa Sarg. (a), C. parciflora Sarg. (a; n), C. Parkae Sarg. (a), C. Pennypackeri Sarg. (a; n), C. peoriensis Sarg. (n), C. permera Sarg. (a; n), C. persimilis Sarg. (n), C. persistens Sarg. (a; n), C. phlebodia Sarg. (a; n), C. pili- fera Sarg. (a), C. polyclada Sarg. (a), C. regalis Beadle (a; n), C. Reverchonii Sarg. (Tex.), C. rivalis Sarg. (a; n), C. robusta Sarg. (a; n), C. rotunda Sarg. (a), C. rubrifolia Sarg. (a; n), C. rudis Sarg. (a), C..setosa Sarg. (a), C. severa Sarg. (a), C. signata 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 131 Beadle (a), C. sinistra Beadle (a), C. sublobulata Sarg. (a; n), C. tardiflora Sarg. (a), C. tetrica Beadle (a; Tenn.), C. trium- phalis Sarg. (a; n), C. uniqua Sarg. (a), C. vallicola Sarg. (a; n), C. villiflora Sarg. (a), C. Wilkinsoni Ashe (a). DILATATAE: Crataegus coccinioides Ashe (a; n; Mo.), C. dilatata Sarg. (= C coccinioides var. dilatata |Sarg.] Eggl.) (a; Mass., N. Y., Penn., Vt.), C. durobrivensis Sarg. (n), C. hudsonica Sarg. (n). DOUGLASIANAE: Crataegus colorado Ashe (n), C. columbiana Howell (a), C. Doug- lasii Lindl. (a, Farlow [1885]; n), C. Douglasti {. badia Sarg. (n), C. Douglasii var. Suksdorfii Sarg. (n), C. erythropoda Ashe (n), C. Piperi Britt. (a), C. rivularis Nutt. (n). FLAVAE: Crataegus arrogans Beadle (a), C. colonica Beadle (a), C. dispar Beadle (a; S. Car.), C. elliptica Ait. (a), C. frugiferens Beadle (a), C. ignava Beadle (a; n), C. impar Beadle (a), C. insidiosa Beadle (a), C. limata Beadle (a), C. visenda Beadle (a). INTRICATAE: Crataegus apposita var. Bissellii (Sarg.) Eggl. (a; Conn.), C. bilt- moreana Beadle (Mo.), C. Boyntonii Beadle (N. Car.), C. Buck- leyi Beadle (a; N. Car.), C. Delosii Sarg. (a), C. foetida Ashe (a), C. fortunata Sarg. (a), C. laetifica Sarg. (a; n), C. macilenta Beadle (Ala.), C. modesta Sarg. (a), C. neobushii Sarg. (n), C Painteriana Sarg. (a; n), C. rubella Beadle (a), C. Sargentiu Beadle (a), C. scabra Sarg. (a; n), C. Schweinitziana Sarg. (Penn.), C. straminea Beadle (Penn.), C. tecta Beadle (Ala.), C. villicarpa Sarg. MACRACANTHAE: Crataegus ambrosia Sarg. (n), C. aqguilonaris Sarg. (n), C. ardua Sarg. (n), C. baccata Sarg. (n), C. Balkwillii Sarg. (n), C. Becki- ana Sarg. (n), C. bristolensis Sarg. (n), C. calpodendron (Ehrh.) Medic. (Penn.), C. chadfordiana Sarg. (n), C. Chapmanii (Beadle) Ashe (a; n; N. Car.), C. conspecta Sarg. (n), C. conspicua Sarg. (n; Vt.), C. corporea Sarg. (n), C. delectabilis Sarg. (Ont.), C. Deweyana Sarg. (a; n), C. divida Sarg. (n), C. dumicola Sarg. (n), C. Emersoniana Sarg. (a; n), C. ferentaria Sarg. (a; n), C. ferta Sarg. (n), C. fertilis Sarg. (a; n), C. finitima Sarg. (a; n), C. flagrans Sarg. (n), C. flammea Sarg. (n), C. frutescens Sarg. (n), C. fulgens Sarg. (a; n), C. fulgida Sarg. (n), C. Gaulti Sarg. 132 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XvI (a; n), C. gemmosa Sarg. (n), C. glabrata Sarg. (n), C. globosa Sarg. (a; n), C. Halliana Sarg. (n), C. Aystricina Ashe (n), C. illi- noiensis Ashe (n), C. integriloba Sarg. (n), C. Laneyi Sarg. (a; n), C. laurentiana Sarg. (n), C. macracantha Lodd. (a; n; Conn., N. Y., S. Dak., W. Va., Wis.), C. macracantha var. succulenta Rehd. (= C. succulenta Schrad.) (n; Penn., Wis.), C. mem- branacea Sarg. (n; Vt.), C. michiganensis Ashe (n), C. micro- sperma Sarg. (n), C. mtssouriensis Ashe (a; n), C. neofluvialis Ashe (n; Penn.), C. nuda Sarg. (n), C. ogdensburgensis Sarg. (n), C. Peckietta Sarg. (N. Y.), C. pellucidula Sarg. (n), C. peramoena Sarg. (n), C. pertomentosa Ashe (Iowa, Kansas), C. pisifera Sarg. (n; Vt.), C. praeclara Sarg. (a), C. propixa Sarg. (a), C. pruni- folia (Marsh.) Pers. (a; n), C. pudens Sarg. (a; n), C. rhombifolia Sarg. (n; Conn., N. Y., Mass., Vt.), C. Robinsonii Sarg. (n), C. rupicola Sarg. (a), C. saeva Sarg. (n), C. Searsit Sarg. (n), C. simulata Sarg. (n), C. spatiosa Sarg. (n), C. spinulosa Sarg. (a; n), C. structilis Ashe (n), C. tomentosa L. (a, Thaxter [1880]; n; IIL, Iowa, Ky., Maine, Miss., Mo., Ohio, Ont., Que., Wis.) , C. truculenta Sarg. (n), C. vaga Sarg. (a; n), C. vegeta Sarg. (a; n), C. venu- losa Sarg. (a; n), C. venustula Sarg. (n), C. Wilsonii Sarg. (n). MACROSPERMAE: Crataegus Handyae Sarg. (n). MICROCARPAE: Crataegus Phaenopyrum (L. {.) Medic. (= C. cordata Ait.) (Del., Tenn.). MOLLEs: Crataegus anomala Sarg. (n; Conn., N. Y.), C. arnoldiana Sarg. (a; n), C. Berlandieri Sarg. (n), C. canadensis Sarg. (n), C. cham- plainensis Sarg. (a; n; N. Y.), C. contortifolia Sarg. (n), C. corusca Sarg. (Ill.), C. digna Sarg. (n), C. dispessa Ashe (a; Mo.), C. dume- tosa Sarg. (a; Mo.), C. Ellwangeriana Sarg. (a; n), C. exclusa Sarg. (n), C. Fulleriana Sarg. (n), C. Greggiana Eggl. (a), C. induta Sarg. (a), C. invisa Sarg. (n), C. lanigera Sarg. (n), C. lanuginosa Sarg. (a; n), C. lasiantha Sarg. (a; n; Mo.), C. lauta Sarg. (n), C. limaria Sarg. (a; n), C. macrophylla Sarg. (n), C. meridionalis Sarg. (n), C. mollipes Sarg. (n), C. mollis (Torr. & Gr.) Scheele (a, Bliss [1931]; n; Ill., Ind., Iowa, Kan., Ky., Mass., Mo., Nebr., Ohio), C. noelensis Sarg. (n), C. nutans Sarg. (n), C. pennsylvanica Ashe (n), C. peregrina Sarg. (a; n), C. Robesoni- ana Sarg. (n), C. sera Sarg. (a; n), C. submollis Sarg. (a; n; Vt.), 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 133 C. Tatnalliana Sarg. (n), C. Tracyi Ashe (a), C. transmississippi- ensis Sarg. (n), C. Treleasei Sarg. (Mo.), C. umbrosa Sarg. (n), C. urbica Sarg. (n). NIGRAE: < Crataegus hiemalis Lge. (n), C. nigra Kit. (n). OXYACANTHAE: Crataegus monogyna Jacq. (a; n; Mass.), C. monogyna var. in- ermis Rehd. (a), C. monogyna var. laciniata (Stev.) Regel (a; n), C. monogyna var. pteridifolia Rehd. (a; n), C. Oxyacantha L. a, Farlow [1885]; n; Maine, Mass., Ont.), C. Oxyacantha var. Gireoudii Bean (a), C. Oxyacantha var. leucocarpa Loudon (a), C. Oxyacantha var. rubra Hort. (a), X C. sorbifolia Lge. (a; n). PINNATIFIDAE: Crataegus pinnatifida Bge. (n), C. pinnatifida var. major N. E. Br. (n), PRUINOSAE: Crataegus alacris Sarg. (a), C. amoena Sarg. (a), C. arcana Beadle (n), C. aridula Sarg. (a), C. aspera Sarg. (a; n), C. ater Ashe (a), C. beata Sarg. (n), C. bellula Sarg. (n), C. bracteata Sarg. (a), C. caerulescens Sarg. (n), C. cestrica Sarg. (a), C. Clintoniana Sarg. (n), C. cognata Sarg. (n), C. comata Sarg. (n), C. comparata Sarg. (n), C. confragosa Sarg. (n), C. conjuncta Sarg. (a; n; Conn., Mass.), C. delawarensis Sarg. (a), C. deltoides Ashe (a; n), Gc, disjuncta Sarg. (a; Mo.), C. divisifolia Sarg. (n), C. exornata Sarg. (n), C. Ferrissii Ashe (n), C. festiva Sarg. (Conn., Wiis C. formosa Sarg. (a; n), C. fusca Sarg. (a), C. georgiana Sarg. (a; n), C. glareosa Ashe (n), C. horridula Sarg. (a; n), C. incisa Sarg. (a; n), C. inusitula Sarg. (a; n), C. iracunda Beadle (a; n), C. Jesupii Sarg. (Penn.), C. Kellermanii Sarg. (a), C. latifrons Sarg. (n), C. latisepala Ashe (a; n), C. leiophylla Sarg. (a; n; N. Y.), C. levis Sarg. (a; n), C. littoralis Sarg. (a), €. locuples Sarg. (a; n), C. numerosa Sarg. (a; n), C. oblita Sarg. (a; n), C. Pequotorum Sarg. (a; n; Conn.), C. perampla Sarg. (a; n), C. perjucunda Sarg. (a), C. philadelphica Sarg. (a; n), C. pilosa Sarg. (n), C. platycarpa Sarg. (a), C. Porteri Britt. (n), C. pro- cera Sarg. (a; n), C. pruinosa (Wendl.) K. Koch (a; n; Conn., Mo.,N. Y., Ohio, S. Car., Penn.), C. pruinosa var. latise pala (Ashe) Eggl. (Mass., Mich.), C. pulchra Sarg. (a; n), C.. quinebaugensis Sarg. (Conn.), C. radiata Sarg. (a; n), C. relicta Sarg. (n), Cs remota Sarg. (n), C. rubicundula Sarg. (a; n), C. scitula Sarg. 134 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI (n), C. sicca Sarg. (n), C. sitiens Ashe (a; n), C. tribulosa Sarg. (n), C. uplandia Sarg. (n), C. virella Ashe (a). PRUNIFOLIAE: Crataegus decorata Sarg. (n; Mo.). PULCHERRIMAE: Crataegus ancisa Beadle (Ala.), C. illustris Beadle (a). PUNCTATAE: Crataegus amnicola Beadle (a; n), C. angustata Sarg. (a), C. bar- bara Sarg. (a; n), C. Brownietta Sarg. (n), C. calvescens Sarg. (n), C. celsa Sarg. (n), C. collina Chapm. (Ga., Va.), C. com- pacta Sarg. (n), C. Dewingii Sarg. (n), C. Eatoniana Sarg. (n), C, Eastmaniana Sarg. (a; n), C. florifera Sarg. (a; n), C. glabri- folia Sarg. (a; n), C. incerta Sarg. (n), C. Lettermanii Sarg. (a), C. macropoda Sarg. (a; n), C. notabilis Sarg. (n), C. pausiaca Ashe (a; n), C. porrecta Ashe (n), C. praestans Sarg. (a; n), C. pratensis Sarg. (a; n), C. punctata Jacq. (a; n; IIl., Ind., Iowa, Maine, Mass., Mich., Mo., N. Y., N. Car., Ohio, Ont. Penn., Vt., W. Va.), C. punctata var. aurea Ait. (a; n), C. punctata var. canescens Britt. (n), C. punctata var. maliformis ? (n), C. punc- tata mutabilis Gruber (a; n), C. secta Sarg. (a; n), C. sordida Sarg. (a), C. suborbiculata Sarg. (a; n), C. succincta Sarg. (a), C. sucida Sarg. (Mo.), C. swanensis Sarg. (a; n), C. tenax Ashe (a; n), C. umbratilis Sarg. (a; n), C. verruculosa Sarg. (n), C. vicina Sarg. (a). ROTUNDIFOLIAE: Crataegus Bicknellii Eggl. (n), C. Blanchardii Sarg. (n), C. Brai- nerdu Sarg. (a; n; Vt.), C. Brunetiana Sarg. (a), C. caliciglabra Schuette (a), C. chrysocarpa Ashe (N. Y.), C. coccinata Sarg. (n), C. crassifolia Sarg. (n), C. cupulifera Sarg. (n), C. divergens (Peck) Sarg. (a), C. Dodgei Ashe (n), C. Evansiana Sarg. (a; n), C. Faxonii Sarg. (n), C. illuminata Sarg. (n), C. inaudita Sarg. (a), C. insolens Sarg. (n), C. Jackii Sarg. (n), C. Jonesae Sarg. (a; n), C. Keepii Sarg. (n), C. Kennedyi Sarg. (n), C. kingstonensis Sarg. (n), C. lemingtonensis Sarg. (n), C. maligna Sarg. (n), C. mans field- ensis Sarg. (n), C. Margaretta Ashe (n; Iowa, Mo.), C. Margaretta f. xanthocarpa Sarg. (n),C. Maribella Sarg. (n), C. Oakesiana Eggl. (a), C. praecoqua Sarg. (= C. praecox Sarg.) (n; N. Y.), C. Proc- toriana Sarg. (n), C. propria Sarg. (n), C. rotundata Sarg. (n), C. rotundifolia Moench (= C. coccinea L. p. p.) (a, Thaxter [1889]; n; Iowa, Mo., N. Y., Ont., Vt.), C. rotundifolia var. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 135 aboriginum Sarg. (n), C. rotundifolia var. pubera Sarg. (n), C. rotundifolia {. rubescens Sarg. (n), C. varians Sarg. (n), C. Web- stert Sarg. (n), C. Williamsti Eggl. (n). SANGUINEAE: Crataegus altaica Lange (n), C. dsungarica Zab. (n), * C. Lam- bertiana Lge. (n), C. Maximowiczii Schneid. (n), C. sanguinea Pall. (Ont.). SILVICOLAE: Crataegus aemula Beadle (n), C. allecta Sarg. (n), C. Barryana Sarg. (n), C. blairensis Sarg. (n), C. congestiflora Sarg. (a; n), C. cruda Sarg. (n), C. delectata Sarg. (n), C. diffusa Sarg. (= C. silvicola var. Beckwithae |Sarg.| Eggl.) (n; Conn., Vt.), C. dissona Sarg. (n; Mass., N. H., N. Y.), C. effera Sarg. (n), C. filipes Ashe (n), C. foliata Sarg. (n), C. Fretzii Sarg. (n), C. gravis Ashe (n), C. iterata Sarg. (n), C. laetans Sarg. (n), C. Livingstoniana Sarg. (n), C. luxuriosa Sarg. (n), C. macera Sarg. (n), C. Matneana Sarg. (n), C. medioxima Sarg. (n), C. opulens Sarg. (n), C. promissa Sarg. (a; n), C. prona Ashe (n), C. puta Sarg. (n), oF radina Sarg. (n), C. recordabilis Sarg. (n), C. Robbinsiana Sarg. (Vt.), C. ruricola Sarg. (n), C. stolonifera Sarg. (n), C. strigosa Sarg. (n), C. tortuosa Sarg. (n), C. xanthophylla Sarg. (a; n). TENUIFOLIAE: Crataegus acuminata Sarg. (a; n), C. acutiloba Sarg. (a; n; N.Y., Vt.), C. alnorum Sarg. (n), C. apiomorpha Sarg. (n), C. ascendens Sarg. (n), C. asperata Sarg. (n), C. basilica Beadle (a), C. bella Sarg. (a; n), C. benigna Sarg. (a; n), C. blandita Sarg. (n), C. Boothiana Sarg. (n), C. colorata Sarg. (a; n; Ont.), C. con- ferta Sarg. (n), C. crudelis Sarg. (n), C. cyanophylla Sarg. (a; n), C. Damei Sarg. (n), C. delucida Sarg. (n; Vt.), C. demissa Sarg. (n; Mass., Vt.), C. dissimilis Sarg. (a; n; Conn., Mass., Vt.), C. Edsoni Sarg. (n; N. H., Vt.), C. Eganii Ashe (n), C. firma Sarg. (n), C. flabellata (Bosc.) K. Koch (a; n), C. florea Sarg. (n), C. Forbesae Sarg. (a; n; Conn.), C. fucosa Sarg. (n), c. genialis Sarg. (a; n; Vt.), C. glaucophylla Sarg. (a; n; Conn., N. Y.), C. gracilipes Sarg. (n), C. Gruberi Ashe (n), C. Habereri Sarg. (n), C. Hadleyana Sarg. (n), C. heidelbergensis Sarg. (n), C. insolita Sarg. (n), C. leptopoda Sarg. (n), C. lucorum Sarg. (n), C. luminosa Sarg. (n), C. macrosperma Ashe (n; N. Be Penn.), C. marcida Ashe (n), C. matura Sarg. (n), C. media Sarg. (n), C. merita Sarg. (n), C. miniata Ashe (n), C. modica Sarg. (n), C. monstrata Sarg. (n), C. Napaea Sarg. (n), C. mescia Sarg. 136 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI (n), C. otiosa Ashe (n), C. Paddockeae Sarg. (n), C. Paineana Sarg. (n), C. pallidula Sarg. (n), C. parviflora Sarg. (n), C. pas- torum Sarg. (a; n), C. paucispina Sarg. (a), C. pentandra Sarg. (a; n; Vt.), C. perlevis Ashe (n), C. populnea Ashe (n), C. pumila Sarg. (n), C. retrusa Ashe (n), C. roanensis Ashe (Ky., Vt.), C. rubicunda Sarg. (n), C. rubrocarnea Sarg. (n), C. rufipes Ashe (n), C. sarniensis Sarg. (n), C. saturata Sarg. (n), C. serena Sarg. (n), C. sextilis Sarg. (n), C. siderea Sarg. (n), C. Slavini Sarg. (n), C. Streeterae Sarg. (n), C. suavis Sarg. (n), C. taetrica Sarg. (n), C. tarda Sarg. (n), C. tenella Ashe (n; Conn.), C. tenera Ashe (n), C. tenutloba Sarg. (n), C. trachyphylla Sarg. (n), C. uber Ashe (n), C. viridimontana Sarg. (n), C. vittata Ashe (a). TRIFLORAE: Crataegus austromontana Beadle (a). UNIFLORAE: Crataegus armentalis Beadle (a), C. Brittonii Eggl. (a). VIRIDES: Crataegus abbreviata Sarg. (a; n), C. atrorubens Ashe (a; n), C. blanda Sarg. (a), C. enucleata Sarg. (a; n), C. lanceolata Sarg. (a; n), C. larga Sarg. (a), C. lutensis Sarg. (a), C. nitens Sarg. (a), C. nitida (Engelm.) Sarg. (a; n), C. ovata Sarg. (a; n), C. penita Beadle (a), C. poliophylla Sarg. (a), C. uvaldensis Sarg. (a), C. velutina Sarg. (a), C. viridis L. (a; n; Okla.), C. vulsa Beadle (a; n). CYDONIA: Cydonia oblonga Mill. (= C. vulgaris Pers.) (a; Thaxter [1889]; Conn., Niagara Peninsula, N. J., Penn.). MALUS: Malus angustifolia Michx. (S. Car.), * M. astranica Dum.-Cours. (a), M. baccata Borkh. (a), M. coronaria Mill. (a, Arthur [1907] ), x M. Dawsoniana Rehd. (a), M. glabrata Rehd. (a), M. glau- cescens Rehd. (Ind.), M. ioensis var. plena Rehd. (a), « *M. magdeburgensis Schoch (a), M. pumila Mill. (= M. Malus ie Britt.) (Thaxter [1886]; Conn., Maine, Mass., Mo., N. H.,N. J., N. Y., Vt.), & M. Soulardi Britt. (a), x M. ahaa Rehd. (a). MESPILUS: Mespilus germanica L. (a). PYRUS: Pyrus Balansae Decne. (a), P. betwaefolia Bge. (a; n), P. Bret- schneideri Rehd. (a), P. communis L. (a; Conn., Ind., Iowa, Mass., 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 137 N. Car., N. Y., Penn., R. I.), P. elaeagrifolia Pall. (a), *P. Kor- shinskyi Litv. (a), *P. Michauxii Bosc (a), *P. Lindleyi Rehd. (a), *P. nivalis Jacq. (a), P. phaeocarpa Rehd. (a), P. salicifolia Pall. (a), P. serotina Rehd. (a), *P. serrulata Rehd. (a), P. syriaca Boiss. (a), P. ussuriensis Maxim. (a). SORBARONIA: x *Sorbaronia alpina Schneid. f. superaria Zabel (a). SORBOPYRUS: x *Sorbopyrus auricularis Schneid. (a). SORBUS: Sorbus americana Marsh. (a; Thaxter [1887 and 1891]; Maine, Mass., N. Y., Penn., Vt.), S. americana var. fructu albo Hort. (a), *S. americana var. nana Hort. (a), * *S. arnoldiana Rehd. (a), *S. Aucuparia L. var. Backhousei Hort. (a), *S. dumosa Greene (a), *S. japonica var. calocarpa Rehd. (a), & S. thuringiaca Fritsch (a). HOSTS FOR THE III STAGE JUNIPERUS: Juniperus lucayana Britt. (= J. barbadensis Auth., not L.) (Ala.), J. communis L. (Penn.), J. fragrans Hort. (Ont.), J. horizontalis Moench (= J. prostrata Pers.) (N. Dak.), J. scopulorum Sarg. (Colo., Ill., lowa, N. Dak.), J. virginiana L. (Ala., Conn., Ill., Ind., Iowa, Kansas, Ky., La., Mass., Mich., Minn., Miss., Mo., N. H., N. Y., N.Car., N. Dak., Ohio, Okla., Ont., Penn., S. Car., Tex., Vt., W. Va., Wis.), J. virginiana var. Burkii Hort. (Ill.), J. virgint- ana var. Canaertii Sénécl. (Ill.), J. virginiana var. elegantissima Hochst. (Ill.), J. virginiana var. glauca Carr. (Ill.). LARIX: Larix sp. (Conn., Kan., Minn., Miss., N. Y., Okla., Tex., Va., Va.). V. SUMMARY 1. At least ten genera, all within the Pomoideae, include hosts on which the aecial phase of Gymnosporangium globosum may occur. One genus only, Juniperus, is known with certainty to include hosts for the telial phase. 2. Relative susceptibility to G. globosum within the respective host genera has been studied by the writer to determine: (1) immune species; (2) resistant species which suffer no material harm from this rust; (3) moderately susceptible species which may be infected but not to the extent of defoliation; and (4) very susceptible species whose foliage can be ruined by G. globosum. 138 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI 3. These investigations were carried out by means of artificial inocu- lations, substantiated by observations of natural infection where pres- ent, in the Arnold Arboretum of Harvard University. 4. The results of these investigations on relative susceptibility, added to those of previous writers, may be summarized as follows: A. On host genera for the aecial phase of G. globosum. (a) On the genera on which serial inoculations were made. Crataegus. A marked variation in susceptibility was found within the genus, the degree of which is dependent primarily on the thickness and the rapidity of deposition of the foliar cuticle. Due to the large number of species and the unstable condition of taxonomy within the genus, the classification according to susceptibility to G. globosum was made by groups rather than by species. The observations on natural infection substantiated the results obtained by artificial inoculation. Suggestions have been made for the selection of resistant species and varieties within the respective groups. Pyrus. Of seventeen species inoculated, one proved to be very susceptible, two moderately susceptible, ten resistant, and three immune. Certain of the commercial varieties are classified from pre- vious reports according to their susceptibility to G. globosum. Sorbus. Infection was obtained on all the species and varieties of American origin inoculated. Of thirty-one species and varieties of Eurasian origin inoculated four are resistant, the remainder are immune. Malus. Of seven American species inoculated three proved to be susceptible, while infection was obtained on only one species and three hybrids of the twenty-seven Eurasian types considered. Infection was obtained also on two hybrids between Eurasian and American species. Certain of the commercial varieties are classified from previous reports according to their susceptibility to G. globosum. (b) On the genera otherwise inoculated. Amelanchier, Seventeen species and varieties were inoculated; all inoculations gave negative results. Nevertheless, the rust has been reported on two species, 4. canadensis and A. alnifolia.' It is not prob- able that any species in this genus would suffer severely from infection Cydonia. Gymnosporangium globosum has been reported as occur- ring commonly on quince in New Jersey. Cydonia oblonga by culture proved to be moderately susceptible to G. globosum. Crataegomes pilus, Mespilus, Sorbaronia and Sorbopyrus. The re- 1See foot-note on page 53. 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 139 sults obtained by inoculations on representatives of these more or less susceptible genera have been tabulated on page 127. Comptonia, Myrica and Photinia. These genera were found by in- oculation to be immune. B. Host genera for the telial phase of G. globosum. Juniperus. No information to date would indicate that any species other than J. virginiana and J. scopulorum and their varieties would suffer to any extent from infection by G. globosum. 5. In the genera Crataegus, Malus, Pyrus and Sorbus there is a defi- nite duration to the period of susceptibility reaching a maximum during or immediately after foliar expansion. 6. In selecting ornamentals to plant in vicinities where Gymno- sporangium rusts are present, it must be remembered that the relative susceptibility of any host to G. globosum is not necessarily correlated with its susceptibility to other Gymnosporangium rusts. 7. No consideration has been given to the possibility of variation in virulence within different strains of G. globosum. Such may very well occur. 8. A complete list of all the known hosts of G. globosum is recorded in this paper. VI. ACKNOWLEDGMENTS To Professor J. H. Faull for his generous assistance in making this study possible and for his guidance, supervision and other expressions of personal interest the writer acknowledges deep obligation. To the Arnold Arboretum for permission to use its facilities; to Pro- fessor A. Rehder and Mr. E. J. Palmer for their invaluable assistance in the taxonomic treatment of the host genera; to Dr. A. E. Navez for his careful analysis of data and for helpful advice; and to Dr. Ivan H. Crowell for his cooperation and his help in field work the writer also expresses gratitude. VII. BIBLIOGRAPHY ApaMs, J. F. (1919). Rusts on conifers in Pennsylvania, I. (Penn. Agric. Exper. Sta. Bull. 160:25. Axoxyo vas (1930). Apple rust diseases in New York. (Plant Dis. Rep. a 7 ee (1907). Cultures of Uredineae in 1906. (Jour. Mycol. “(1909). Cultures of Uredineae in 1908. (Mycologia, 1:239.) (1910). Cultures of Uredineae in 1909. (Mycologia, 2:229.) (1921). Cultures of Uredineae 1899-1917. (Mycologia, 13: —— (1924). The Soe (Rusts) of Iowa. (Proc. Amer. Acad. Arts and Sci. 31 .) 140 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI (1926). Additions and corrections. (N. Amer. Flora, 7:741.) (1927). Additions and corrections. (N. Amer. Flora, 7:826.) Buiiss, D. E. (1931). Physiologic specialization in Gymnosporangium (abstr. ). re 21:111.) ———— (| . The betas tive and seasonal development of 6 03 nosporangivn in low (lowa State Coll. Agric. and Mech. Arts Res Bull. URNHAM, , S. H. and LatHaM, R. A. (1917). The flora of the town of Southold, Long Island and Gardiner’s wae (Torreya, 17:116. CLAASSEN, E, (1897). List of the Uredineae of Cuyahoga and other counties of northern Ohio, eee with the names of their host-plants. (Ann. Rep. Ohio ei Acad. Sci. CLinTON, G. P. (1904). Diseases i plants cultivated in Connecticut. (Ann. Rep. Conn. Agric. Exper. Sta. 27:313, 343, 352.) a 4). Plant pest Poiback for Connecticut Diseases and injuries. (Conn. Agric. Exper. Sta. Bull. 358: 167, 506, Bed; ou) a I. L. (1934). VI. Diseases of ornamental plants. (Ann. Rep. lant Dis. Survey for 1933. 72. ee M. T. (1913). Report of the plant pathologist. (Rep. N. J. Agr. Coll. Exper. Sta. for 1912, ao Crowe tL, I. H. (1934). The hosts, life history and control of the cedar- apple rust fungus Gymnosporangium Juniperi-virginianae Schw. (Jour. Arn. Arb. 15:165-232.) Faritow, W. G. (1 880). The oe or hyping apples of the United States. (Anniv. Mem. Bost . Nat. Hist. pp. 34-35.) ee Notes on some spec cies oO ymnosporangium and re ae of the United States. (Proc. Amer. Acad. Arts and Sci. 0:312-320. Pichon H. T. (1915). Report from the Division of Botany for the fiscal year ending March 31, 1914. (Dom. Can. Dept. Agric. Sessional paper 16:847.) HIARSHBERGER, J. W. (1902). Two fungous diseases of the white cedar. Hester, L. R. Liga Une, HB. .H. 17). Manual of fruit diseases. (MacMillan ae Co., New York, pp. 341-344. ) unT, W. R. (1926). The Uredinales or rusts of Connecticut and the ( ot the rer New England States. (Conn. State Geol. and Nat. Hist. Survey 6:73-78 JACKSON, H. S. (1921). The Uredinales of Indiana III. (Proc. Ind. cad. Sci. for 1920, p. 172.) Kern, F. D. (1911). A biologic and taxonomic study of the genus Gym- nosporangium. (Bull. N. Y. Bot. Gard. 7:468-469. ) KERN, F. D., THurston, H. W., Jr., Orton, C. R. and Apams, J. F. (1929). The rusts of Pennsylvania. (Penn. State Coll. School Agric. and Exper. Sta. Bull. 239:20. Martin, G. H., Jr. (1922). Diseases of forest and shade trees, ornamental and miscellaneous as in the United States in 1921. (Plant Dis. Bull. Suppl. 23:428, 431, 4 (192 Diseases of forest and shade t trees, ornamentals and miscellaneous Jb in the United States in 1923. (Plant Dis. Rep. Suppl. 37:350, 361 Miter, P. R., STE N. and Woop, J. I. (1933). Diseases of plants i in the’ United Grains in on (Plant Dis. Rep. Suppl. 84:50.) 1935] MacLACHLAN, HOSTS OF GYMNOSPORANGIUM GLOBOSUM 141 eerie A. (1927). Manual of cultivated trees and shrubs. (MacMillan Co., New York, pp. Sede ae BF (193 V2 “The more important diseases of apples in ee (Univ. of Tenn. Agric. Exper. Sta. Bull. 145:34-37. STEVEN and Hatt, J. G. (1910). Diseases of economic plants. (MacMillan aa Co., New York, pp. 106-107.) STEWaRT, F. C. (1910). Notes on New York plant diseases, I. (N. Y. Agric. Exper. Sta. Bull. 328:376. STONE, R. E. (1908). Cedar apples and apple leaf rust. (Ala. Agric. Pxper> Sta ir, -2:5, ——— (1909). Species of Gymnosporangium in southern Alabama. (Torreya, 9:116.) THAXxTER, R. (1887). VII. On certain cultures of Gymnosporangium with notes on their Roesteliae. (Proc. Amer. Acad. of Arts and Sci. ——_—. 89). Notes on cultures of Gymnosporangium made in 1887 and 1888. te Gaz. 14:167-168. — (1891). The Connecticut species of Gymnosporangium (cedar apples). (Conn. Agric. ee Sta. Bull. 107:4 HOMAS, H. E. and Mitts, W. D. (1930). Rust diseases of the apple. (Plant Dis. Ree 14:214-215.) VIII. EXPLANATION OF PLATES PiaTE 125 Illustrations of the tendency of the mycelium to follow along the veins of Crataegus leaves: Fig. 1. A series of lesions obtained by inoculation on a waxy-type of leaf Sd toed tes giving the appearance of systemic infection along t Fig. 2. A single lesion a the Sennen stage on Crataegus suavis. The rust mycelium concentrates along the vascular strands caus- ing the latter to show as bright yellow lines within the lesion. Fig. 3. A single lesion extends along a lateral vein, forking at the junc- tion with a sub-lateral vein. Fig. 4. : typical vein infection; the long axis of the lesion correspond- ng with that of the vein PLATE 126 Types of infections and their ska effects on Crataegus leaves (ex- planations in text): Fig. 1. Illustrates oe relative amount of leaf killing caused by vein in- fectio by infections not primarily associated with the main veins. Fig. 2. A single Wier on the mid vein resulting in the death of over one-half o Fig. 3. A very small type of lesion, exhibiting no hypertrophy and pro- ducing a single aecial horn. Fig. 4. A single vein infection (indicated by the black spot on the plate), killing the leaf behind the mae alo ong the vein; suggesting a toxic agent on the part of the ru Fig. 5. A large single lesion which died shortly after spermogonia appeared; suggesting hypersensitivity on the part of the host. 142 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Pate 127 Figs. 1, 2, : on 4, illustrate the relative degree of susceptibility of Cratae- inglei, as indicated by serial inoculations on April 25, May Nay 23 and June 28, 1934, respectively. Fig. 5. The type of chamber used in all the ata en (Explanations in the text.) PLATE 128 Serial inoculations on eign fe Jonesae to illustrate the period of sus- ceptibility (explanations in tex Fig. 1. Inoculated May 7, at which time the two upper leaves were very small, while the five basal leaves were well expanded. As indi- cated by the number of lesions the latter are the more susceptible. Fig. 2. Inoculated June 8, at which time all leaves were fully expanded ; e two upper (youngest) leaves are now the more susceptible.: LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. LATE 125 VoL. XVI. ARB. 4 ARNOLD JOUR. l, Far 30SUM NGIUM GLO! 1rY MNOSPORA Hosts or ( Tue GLOBOSI M [ GY MNOSPORANGI Jour. ARNOLD Arp. VoL. XVI. THE — LosTs OF GYMNOSPORANGIUM GLOBOSUM Farl. PLATE Jour. Arnotp Ars. VoL. XVI, PLATE 128 arl, Tue Hosrs oF GYMNOSPORANGLUM GLOBOSUM 1935] HUNTER, STUDIES OF EUROPEAN SPECIES OF MILESIA 143 A PRELIMINARY NOTE ON LIFE HISTORY STUDIES OF ROPEAN SPECIES OF MILESIA LILLIAN M. HuNTER Although eleven species of Milesia are known to occur in Europe (FAULL, J. H. Taxonomy and Geographical Distribution of the Genus Milesia. Contr. Arnold Arb. Harvard Univ. II. 1932) up to the pres- ent the life histories of two only of them have been worked out, namely, M. Blechni (Syd.) Arth. (KLEBAHN, H. Kulturversuche mit Rost- pilzen. In Zeitsch. Pflanzenkr. 26: 257-277. 1916) and M. Kriegeri- ana (Magn.) Arth. from Dryopteris Filix mas (L.) Schott (Mayor, Euc. Notes Mycologiques VIII. In Bull. Soc. Neuchat. Sci. Nat. 58: 23-26. 1933). Recently it was my privilege to make certain investigations on life histories of Milesia rusts in England. ‘Teliosporic material of several species was assembled and inoculation experiments were made on vari- ous firs with the results that spermogonia and aecia of the following species of Milesia have been obtained for the first time— ‘(1) Milesia Scolopendrii (Fuckel) Arth. (from Scolopendrium vul- gare Smith) on Abies alba Mill., and A. concolor Lindl. and Gord. (2) Milesia Polypodii B. White (from Polypodium vulgare L.) on Agies alba and A. concolor. (3) Milesia vogestaca (Syd.) Faull (from Polystichum angulare Presl) on Abies alba. (4) Milesia Kriegeriana (Magn.) Arth. (from Dryopteris spinulosa [O. F. Miller] Kuntze) on Abies alba, A. concolor and A. grandis Lindl. Spermogonia and aecia were also obtained for Milesia Kriegeriana (from Dryopteris Filix mas) on Abies alba and on two new hosts, namely, A. concolor and A. grandis. Aeciospores thus obtained by cultures were used in inoculating vari- ous ferns, and uredospores were obtained for the following species— (1) Milesia Scolopendrii on Scolopendrium vulgare. (2) Milesia Polypodii on Polypodium vulgare. (3a) Milesia Kriegeriana (from Dryopteris spinulosa) on Dryop- teris Filix mas, D, spinulosa and D. spinulosa var. intermedia (Muhl.) Underw. (3b) Milesia Kriegeriana (from Dryopteris Filix mas) on Drvyop- teris Filix mas and D. spinulosa var. dilatata (Hoffm.) Underw. LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. : 7 JOURNAL OF THE ARNOLD ARBORETUM VOLUME XVI APRIL, 1935 NUMBER 2 STUDIES IN THE BORAGINACEAE, XI IvAN M. JOHNSTON CONTENTS 1. The Species of Tournefortia and Messerschmidia in the Old PE le gee ee a Bax ale ae bk eee ene 2. Notes on Brand’s Treatment of Cryptantha. ..............0000. 168 3. New or otherwise Noteworthy Species. ...........e.0eeseee0s ifs 1. THE SPECIES OF TOURNEFORTIA AND MESSERSCHMIDIA IN THE OLD WORLD THE SPECIES treated here have, in the past, all been referred to the genus Tournefortia. I am, however, suggesting that certain of them be segregated to form the redefined genus Messerschmidia. During the work on this paper I have been privileged to examine almost all the type- specimens concerned. This has permitted me to place definitely a large number of poorly understood old species that have troubled workers in the past. The work has been undertaken as part of a projected study of the Boraginaceae-Heliotropioideae. It is the first attempt to treat all the Old World species of Tournefortia since the presentation by DeCandolle in the ninth volume of the Prodromus in 1845. Tournefortia Linnaeus, Sp. Pl. 140 (1753) and Gen. Pl. ed. 5, 68 (1754). The species of Tournefortia found in the Old World all belong to the following: Section EuUToOURNEFORTIA Johnston, Contr. Gray Herb. 92: 66 (1930). —type-species, T. hirsutissima L. Tournefortia — Pittoniae Humboldt, Bonpland & Kunth, Nov. Gen. et Sp. 3: 80 (1818). — type- 146 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI species, T. hirsutissima L. Tournefortia sect. Pittonia Don, Gen. Syst. 4: 366 (1837). — type-species, T. hirsutissima L. Pittonia Plumier ex Adanson, Fam. Pl. 2: 177 (1763).— type-species, T. hirsutissima L. Oskampia Rafinesque, Sylva Tellur. 123 (1838).— type-species, O. scandens Raf. & O. hirsuta Raf. Tournefortia sect. Tetrandra DeCan- dolle, Prodr. 9: 527 (1845).— type-species, 7. tetrandra Blume. Tetrandra (DC.) Miquel, Fl. Nederl. Ind. 2: 928 (1858).— type- species, Tournefortia tetrandra Blume. The species of Eutournefortia found in the Old World are remarkable for their parallelism of variation. Most of them have corollas with the tube either long or short, herbage with the pubescence present or absent as well as leaf-blades that are broad or elongate. The combinations of these variations produce forms very diverse in gross appearance so that it is not at all surprising that botanists have been impressed by them and misled into giving specific names to many of them. A considera- tion of all the Old World Eutournefortiae and observation of the recur- rent pattern of variation among them, however, lead one to a proper estimate of the surprisingly diverse phases which they present. Like- wise, a consideration of the facts of distribution leads to a similar end. When the variations mentioned are given recognition it is found that the resulting numerous ill defined “species” grow together over most of a common area of dispersal. When the variations mentioned are dis- counted, species may be defined that have a credible geographic range— a range that is distinct from that of the closely related species and one quite similar and familiar among species of other genera within the region. I am accordingly of the opinion that the variations noted deserve at best no more than mere formal recognition. Since, however, I do not believe that obscure tropical plants should be burdened with numerous subspecific names until some evident use for them arises, I have refrained from any attempt at formally naming the reoccurring combinations of the paralleling intraspecific variations described. KeEY TO THE SPECIES mae fruit breaking up into four equal single-seeded nutlets, se prominently ribbed on their inner surface ....1. 7. sarmentosa. Ripened fruit breaking up into two carpels which are each com- posed of two seminiferous cells and an intervening empty Flowers 4-merous Flowers 5-m Guat ae from southern Asia (including the Anda- man Islands). ee ank ee eiaidiece Grae aime a wt ean eae aad ee 2. T. tetrandra. 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 147 Calyx-lobes 3-4 mm. long at anthesis, usually sub- ulate; leaves drying more or less golden-brown ETEathS SIKMIMNG 6co.0. 52 95,6's 5d oS RE WOES es T. Hookeri. Calyx-lobes 1-2 mm. long at anthesis, linear or lanceolate: Flowers with evident pedicels 1-2 mm. long; Madras BONS erie ho ha Xi ook Cee a . T. Heyneana. Flowers sessile or subsessile. Leaves abruptly long acuminate, blade more or less oval; flowers and fruit usually shortly ae) late; Southern Burma and the Andamans.... 5. T. ovata. Leaves short-acuminate and usually not nae so; blades oblong to lanceolate; flowers and fruit Ree at a ee oe aR eRe 6. T. montana. Insular plants Western Pacific Ocean. Leaves opposite, flowers sessile; Philippines.....7. T. luzonica. Leaves alternate; flowers short-pedicellate; Australia WE Fata ee Oahg aes ds eke eee Meee ees 8. T. Muelleri. Western Indian Ocean. Leaves obtuse or rounded at base, 4-11 cm. long...9. T. puberula. Leaves acute at base, 10-20 cm. long. Stems with minute short closely appressed brownish or golden hairs or quite glabrous; calyx very sparsely strigose, the lobes cuneate, more or less 1 erec Min ON: aca ca nn eae a ae eee 0. T. acuminata. Stems with evident abundant loosely ee hairs (usually more or less velvety); calyx distinctly hairy with the lobes more or less sais Sepals SC ReUgION hoses ba evn nee eus 11. 7. arborescens. Sepals ae WiaTIHOS 4.244 eee regen ie, ojert. 1. Tournefortia sarmentosa Lamarck, Tab. Encyc. 1: 416 (1791): Poiret, Encyc. 5: 357 (1804). Tournefortia orientalis R. Brown, Prod. 497 (1810); Banks & Solander, Bot. Cook’s Voy. 2: 64, tab. 210 (1901). Tournefortia tetrandra var. hirsuta Blume, Bijdrag. Fl. Nederl. Ind. 845 (1826). Tournefortia sarmentosa var. hirsuta Blume ex Miquel, FI. Ind. Batav. 2: 927 (1858), lapsus. Tournefortia hirsuta Reinwardt ex Boerlage, Hand. Fl. Nederl. Ind. 2?: 487 (1899). Tournefortia Urvil- leana Chamisso, Linnaea 4: 465 (1829). Tournefortia frangulaefolia Zippel ex Spanoghe, Linnaea 15: 334 (1841 ?), in synon. Tourne fortia Horsfieldii Miquel, Fl. Ind. Batav. 2: 927 (1858). Tournefortia acclinis F. v. Mueller, Frag. 4: 95 (1864). Tournefortia macrophylla K. Schumann & Lauterbach, Fl. Deutsch. Schutzgeb. Siidsee 520 (1901). Tournefortia sarmentosa var. magnifolia Domin, Bibl. Bot. 22(Heft 894): 1097 (1928). Tournefortia glabrifolia Domin, Bibl. Bot. 22 (Heft 894): 1098 (1928). 148 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Java to New Guinea, southward into northern Queensland and north- ward through the Celebes, Moluccas and Philippines to Formosa. A variable plant but readily recognized, even in its most diverse forms, by its characteristic fruit. At maturity this breaks up into four equal single-seeded nutlets. All the other Old World Tournefortiae have fruits with two 2-seeded carpels. Tournefortia sarmentosa has been repeatedly confused with continental species and has been the victim of numerous unsuccessful attempts at segregation. In 7. sarmentosa the corollas may be either long or short, the herbage either glabrous or pubescent and the leaf-blades either small or large. These characters in various combinations have produced a host of forms that are superficially very diverse in appearance. These forms, however, agree in fundamental fruiting structures. None of them shows any evident geographical cor- relation. Grouped together to constitute 7. sarmentosa, as here accepted, they appear as mere phases in a species which has a range that is natural and is of a type quite familiar to any student of the Malaysian flora. The type of 7. sarmentosa, in the Lamarck Herbarium at Paris, is labeled ‘“‘colitur in horto regio insulae Franciae” and “de M. Sonnerat.” It is a good specimen showing leaves and flowers but no fruit. The corolla has a tube ca. 2 mm. long and a limb ca. 2.5 mm. in diameter. The calyx is 1.5 mm. long and has broad hairy lobes. The inflorescence is velvety with a dense short but somewhat shaggy, tan-colored indument. The stems and under surface of the leaves have abundant gray hairs. The upper surface of the leaves are’ green and only sparsely strigose. The petiole is ca. 1 cm. long. The blade is rounded at the base, acute at the apex, and is 7.5-10 cm. long and 2.8—4 cm. broad. The plant is evidently the small-flowered hairy form of the common Tournefortia of the East Indian islands. It is certainly not a native of the Mascarenes! Gagnepain, Not. Syst. 3: 32-33 (1914), has discussed this species. His notes, except those referring to collections by Spire, Thorel, and Watt, all refer to the species as I have taken it. The excluded collections are from the Asiatic continent. The species is restricted to the islands and is not to be expected from the mainland. The type of 7. orientalis, at the British Museum, is labeled as col- lected in 1770 by Banks and Solander at Endeavor Bay in northern Queensland. It is a glabrous plant with ovate to oblong leaves, 7—9 cm. long and 3.5—6.5 cm. broad. The corollas are large with a tube ca. 8-9 mm. long and a limb 3—4 mm. broad. The type of 7. Urvilleana was collected by Chamisso in Luzon. It has corollas 8 mm. long and a limb 3 mm. broad. The leaves are slightly less pubescent but otherwise are as in the type of 7. sarmentosa. 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 149 Blume’s T. tetrandra var. hirsuta is given as from the Moluccas and described as follows: ‘“‘ramis foliis pedunculisque hirsutissimis.” At Leiden there is a specimen labeled: “Variet.; Tournefortia hirsuta; Manado; T. tetrandra Bl. Variet.” The first and the last items are in Blume’s script. The specimen is a form with elongate corollas and with more or less hairy leaves suggesting those of T. Horsfieldii. Menado is near the northern tip of the Celebes. Another specimen at Leiden has the following label: “1531 Tournefortia hirsuta R.; Habitat in insula Celebes ad viam inter Kema et Menado; Oct. 1821.” This is associated with a printed label reading “Herbarium Reinwardtianum; in Acad. Lugduno-Batavia.” This second specimen is similar to that first men- tioned and both are probably collections made by Reinwardt. They are, I believe, the types of 7. tetrandra var. hirsuta and T. hirsuta. The name T. frangulaefolia Zippel has appeared only in synonymy. At Leiden this name appears on two sheets having a printed label bearing: “Herb. Lugd. Batav.; Timor” and one in script reading: “176 Tournefortia frangulaefolia; Zp.” Miquel based his 7. Horsfieldii upon material cited: “Java, in Pat- jitan, Kelak (Horsf.).” I have examined specimens from Horsfield’s personal herbarium at the British Museum and those from the set he made for the East India Company (now kept as a unit) at Kew. He made two collections referable to T. sarmentosa, 1: Pajittan (Kalak) Horsfield (borage 6) no. 275; and 2: Blambangan, Horsfield (borage 7) no. 309. The former is evidently the type collection of 7. Hors fieldii. It is a plant with very large leaves that are grayish velvety beneath. The blade becomes 10-14 cm. long and 7-9 cm. broad. The corolla-tube is 7-8 mm. long and the limb is 3-4 mm. broad. Tournefortia acclinis is based upon material from Queensland col- lected by Bowman at Broad Sound and Amity Creek, and by Dallachy at Edgecombe Bay. A study of the original description and of a dupli- cate of Dallachy’s material at Kew shows this species to have moderately sized leaves (5-10 cm. long and 3.5-6 cm. broad), a coarse appressed pubescence, a corolla with a tube 3-5 mm. long, and a corolla-limb 3-4 mm. broad. It is very similar to T. Hors fieldii, except in leaf-size. The type of 7. macrophylla was collected by Lauterbach (no. 2003) at Erima in eastern New Guinea. It is in fruit. The leaves are similar in size and shape to those of typical T. Horsfieldii. In fact the plant differs from the type of that species only in the practical absence of pubescence. The leaves have only a few weak scattered inconspicuous hairs along the nerves. Domin’s T. sarmentosa var. magnifolia from northern Queensland 150 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI (Dietrich 724), to judge from description, seems to be merely a form of TL. sarmentosa with very large (12-15 cm. long, 6—6.5 cm. broad) hairy leaves, and small corollas (corollae tubo breviore). His 7. glabrifolia is another large-leaved (10-13 cm. long and 5—5.5 cm. broad) plant. The leaves are glabrous. The corolla-tube is ca. 3-4 mm. long and the limb is ca, 2mm. broad. The plant comes from Harvey’s Creek in northeastern Queensland. It appears to differ from the type of 7. orientalis only in its larger leaves and smaller corollas. 2. Tournefortia tetrandra Blume, Bijdrag. Fl. Nederl. Ind. 845 (1826). Tournefortia tetragona Blume ex Steudal, Nomencl. ed. 2, 2: 694 (1841). (?)Heliotropium scandens Norona, Verh. Bat. Genootsch 5: 78 (1827); Hasskarl, Cat. Hort. Bogor. 137 (1844), nomen. Tournefortia tetrandra var. glabra Hasskarl, Flora 257: Beibl., p. 27 (1842); Hasskarl, Cat. Hort. Bogor. 137 (1844); Hasskarl, Pl. Javan. Rariores 492 (1848). Tournefortia glabra (Hassk.) Zollinger & Moritzi ex Zollinger, Natuur- en Geneeskundig Archief v. Nederl. Ind. 2:5 (1845). Tetrandra glabra (Hassk.) Miquel, Fl. Nederl. Ind. 2:929 (1858). Tournefortia tetrandra var. longiflora Hasskarl, Cat. Hort. Bogor. 137 (1844), nomen; Hasskarl, Pl. Javan. Rariores 492 (1848). Tournefortia Wallichii DeCandolle, Prodr. 9: 527 (1845): Ridley, Fl. Malay Penin. 2: 441, fig. 115 (1923). Tetrandra Wallichii (DC.) Miquel, Fl. Nederl. Ind. 2: 928 (1858). Tetrandra Zollingeri Miquel, Fl. Nederl, Ind. 2: 928 (1858). Nicobar Islands, Malay Peninsula, Sumatra, Java, Borneo and Celebes. This is apparently the most common and best known of the Javan species of Eutournefortia. The Javan plant has received the following basic names, Tournefortia tetrandra Blume, Tournefortia tetrandra var. glabra Hassk., Tournefortia tetrandra var. longiflora Hassk., and Tetrandra Zollingeri Miquel. The differences between these named forms are minor and variable ones of corolla-size and of distribution of pubescence on the foliage. This variable plant of Java I am quite unable to distinguish from Tournefortia Wallichii DC., a species based upon material from Singapore and Penang. I have accordingly accepted Tournefortia tetrandra as ranging from the Nicobar Islands eastward to Java and the Celebes. The leaves of this species are ovate-acuminate or lance-ovate and are glabrous or sparsely strigose. The fruit is usually subglobose and 4—6 mm. in diameter. The only notable departure from this is found among material from northern Borneo where the fruit, of several different collections, is narrowly ovoid, 7 mm. long and 4—5 mm. 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 151 thick. This form may deserve some nomenclatorial recognition. There are, however, variations of Tournefortia tetrandra which I believe do merit recognition at this time. The characters which set these off from typical T. tetrandra may be organized as follows: Leaves 1%4-2% times as long as broad, ovate acuminate or FARCE VA ins oa 954 beets Oe eR eeeee ournefortia tetrandra Leaves 2%4-3% times as long as broad, more or less lanceolate by org a0 | et or ae var. sigs i dai Leaves somewhat glossy, Ceylon .............-0-se06- . Walkerae. 2A. Tournefortia tetrandra Blume var. sae Moritzi, Syst. Verzeich. 52 (1845-46). Known only from the type-collection in eastern Java. This variety is a peculiar plant with very dull thickish leaves that have only 3—4 pairs of primary veins evident. The secondary nervation is not discernible. I know it only from the type-collection by Zollinger (no. 939), made Dec. 17, 1842, ‘auf den Kalkfelsen von Kuripan.” 2B. Tournefortia tetrandra Blume var. Walkerae (Clarke), comb. nov. Tournefortia Walkerae Clarke in Hooker, Fl. Brit. India 4: 147 (1883); Trimen, Fl. Ceylon 3: 198 (1895). Known only from Ceylon. This plant is simply a narrow-leaved form of the species that is con- fined to Ceylon. The blades are lanceolate but are quite similar to those of the species in texture, nervation, etc. The fruit and flowers are simi- lar to the common Malaysian plant. 3. Tournefortia Hookeri Clarke in Hooker, FI. Brit. India 4: 147 (1883). Tournefortia Hookeri var. subtropica Clarke in Hooker, FI. Brit. India 4: 147 (1883). Known only from the base and lower valleys of the Sikkim Himalayas. SPECIMENS EXAMINED: Rangit, May 15, 1876, Clarke 27953 (K) ; Great Rangit, April 1850, Hooker & Thompson (K, Tyre of var. subtropica) ; Mangpu, 900 m. alt., May 1905, Meebold 4243 (BD); Rangbi, 1500 m., May 31, 1870, Claes 11790 (K, BM); Chunbati, 600 m., June 12, 1870, Clarke 12024 (K, BM); Chunbati, 600 m., April 1876, Gamble 579 (K) ; below Punkabaree, Hooker (BD ); Pancheni, 1875, Gamble 3370 (K); Siliguri, Jan. 1873, Gamble 3369 (K); Dalgaon, mixed forest, April 9, 1893, Haines 358 (K); Sikkim, 1862, Andersson 270 (BD); Sikkim, March 1871, Clarke 16774 (K); Sikkim Terai, Clarke (K); Sikkim, Griffith 5928 (K). Characterized by its slender well developed calyx-lobes and by the golden or golden-brown under surfaces of the leaves. These latter are nearly glabrous or have only scattered hairs along the dark-colored 152 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI nerves and veins. The corolla is usually 3-5 mm. long with the tube forming half (or even more) of this total length. In the var. subtropica, which is merely the large-flowered form of the species, the corollas become ca. 8 mm. long and the calyx-lobes only about a third as long as the tube. The species is a local one and probably worthy of recognition. It is most closely related to the form of 7. montana described as T. Rhasiana, 4. Tournefortia Heyneana Wallich, Num. List no. 910! (1828-29), nomen; Don, Gen. Syst. 4: 369 (1837); Clarke in Hooker, FI. Brit. India 4: 145 (1883); Gamble, Fl. Madras 893 (1923). Tournefortia reticosa Wight, Icones 4°: 16, tab. 1386 (1848); Wight, Spicileg. Neil- gherrense 2: 83, tab. 189 (1851); Gamble, Fl. Madras 893 (1923). Hills of southern peninsular India, about lat. 11°-13° N. and long. 76°-77° E SPECIMENS EXAMINED: Nilgiri Hills, April 1852, herb. Wight 2057 (K); Devala, Nilgiris, 900 m. alt., Nov. 1884, Gamble 15588 (K); S. E. Wynaad, Nilgiris, 900 m. alt., Nov. 1884, Gamble 15497 (K); Wynaad, Beddome 5437 (BM); Nadooputtah, June 1846, herb. Wight (K); Ana- malais, Beddome 5438 (BM); Carcoor-ghat, Nilgiris, Aug. 1887, flowers varying from '—-'% inch according to age, Lawson (Oxford); Coorg, White (Oxford); Peermade Reav (? spelling), 1350 m. alt., Dec. 1910, Meebold 12920 (BD); without data, herb. Wight, probable basis of Wight’s plate and the type of T. reticosa (K); without data, ex herb Heyne, Wallich 910° (herb. Wallich at Kew). A study of Wallich’s herbarium, now at Kew, shows his number 910 to consist of two different species from opposite ends of India. The label reads: “910 Tournef. Heyneana, Wall. — 1. Herb. Heyn. — 2. Pundua F.deS.” The Heyne plant represents the species from the Deccan with pedicellate flowers, which is the one treated here. The plant from Pundua, collected by de Sylva, is accompanied by a large special label indicating that is was found in the “‘Pundouh Hills” in Jan. 1824. Clarke describes the flowers of T. Heyneana as 1/8—1/6 inches (3-4 mm.) long. These measurements are evidently from the duplicate of the Wallich collections now in the general herbarium at Kew. ‘The Heyne material in the Wallich Herbarium at Kew has corollas 9-10 mm. long. The specimens, except for flower-size, are otherwise very similar and I believe they represent minor forms of the species. Significant in this connection is the note made by Lawson on one of his specimens cited above. He states that the corolla varies from 3-12 mm. according to age! Though Don makes no mention of the corolla-size in his descrip- tion, the first given to the species, we may suppose that it was the large- flowered phase since the Wallich Herbarium, then in charge of the 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 153 Linnean Society, was no doubt consulted by him. In any case, since the corolla-size is variable even within the type-collection, the chief character whereby Clarke distinguished 7. reticosa now disappears. The two species, T. Heyneana and T. reticosa, are, I believe, trivial forms of one species and quite synonymous. DeCandolle, Prodr. 9: 516 (1845), received only the second part of Wallich no. 910, and described this as 7. Heyneana. His description in the Prodromus, consequently, is based upon de Sylva’s specimens from Pundua. Clarke pointed out this mistake, gave 2 new name (T. Can- dollii) to the de Sylva collection described by DeCandolle, and properly restricted the name, 7. Heyneana, to the peninsular species collected by Heyne. 5. Tournefortia ovata Wallich, Num. List no. 908 (1828); Don, Gen. Syst. 4: 369 (1837); DeCandolle, Prodr. 9: 516 (1845); Clarke in Hooker, Fl. Brit. India 4: 147 (1883). Southern Burma and the Andaman Islands. SPECIMENS EXAMINED: Rangoon, Aug. 1826, Wallich (no. 15) 908 (herb. Wallich, Type); Rangoon, McClelland (K, three collections) ; Andamans, April 1891, Prain (Cambridge) ; Middle Andaman, Homfray Straits, climber, 1915, Parkinson 297 (K); Aberden, South Andaman, Kurz (K, parasitized; Delessert, normal) ; Chauldare, South Andaman, Characterized by its elliptical abruptly acuminate leaves, its sub- pedicellate flowers and its southern occurrence. The corolla becomes 8mm. long. The calyx is only 1.5 mm. long at anthesis. The leaves are mostly rather firm in texture and are usually brown and glabrous beneath. One of McClelland’s collections is consequently quite atypical in having the leaves not only thin in texture but golden-brown beneath as well. Another one of his collections is quite hairy on the lower leaf- surface. The pedicels in 7. ovata are usually at most only 1 mm. long, though in Parkinson’s material cited the pedicels become fully 2 mm. long and are quite evident. 6. Tournefortia montana Loureiro, Fl. Cochinch. 1: 122 (1790). Messerschmidia montana (Lour.) Roemer & Schultes, Syst. 4: 544 (1819). Lithospermum viridiflorum Roxburgh, Hort. Bengal. 13 (1814), nomen; Lehmann, Asperif. 1: 30 (1818), in synon.; Roxburgh, Fl. Indica 2: 4 (1824), description; Roxburgh, Icones ined. Kew. tab. 2120. Heliotropium viridiflorum (Roxb.) Lehmann, Asperif. 1: 30 (1818). Tournefortia viridiflora (Roxb.) Wallich, Num. List no. 907 (1828); Clarke in Hooker, Fl. Brit. India 4: 146 (1883). Tournefortia Sampsoni Hance, Jour. Bot. 6: 330 (1868). Tournefortia Wightii Clarke in Hooker, Fl. Brit. India 4: 146 (1883). Tournefortia Rox- 154 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI burghii Clarke in Hooker, FI. Brit. India 4: 146 (1883). Tournefortia viridiflora var. Griffithii Clarke in Hooker, FI. Brit. India 4: 146 (1883). Tournefortia Candollii Clarke in Hooker, Fl. Brit. India 4: 146 (1883). Tournefortia khasiana Clarke in Hooker, FI. Brit. India 4: 147 (1883). Tournefortia Boniana Gagnepain, Not. Syst. 3:33 (1914) and in Lecomte, Fl. Gén Indo-Chine 4: 217 (1914). | Tournefortia Gaudi- chaudii Gagnepain, Not. Syst. 3: 34 (1914) and in Lecomte, Fl. Gén. Indo-Chine, 4: 217 (1914). | Tournefortia Heyneana sensu DeCan- dolle, Prodr. 9: 516 (1845). In the hills, up to 1500 m. alt., in Assam, Upper Burma, northern Siam (Payap and Maharat), middle and northern Indo-China (Anam, Laos and Tonkin) and southern-most China (Yunnan, Kwangsi and Kwangtung). This species presents a number of diverse phases resulting from com- binations of variations in leaf-size, abundance and distribution of pubescence, and size of corolla-tube. These phases have been treated as “species” but their variability, their erratic distribution, and their occurrence together in various localities lead me to believe they are merely further manifestations of the surprising intraspecific variability of these structures among the Old World Tournefortiae. After dis- counting these variations as mere phases, I am struck with the natural- ness of the distribution of the resulting aggregate species. The distri- bution is of the pattern found in numerous species of other genera and families inhabiting this part of Asia. The type of T. montana has not been examined. Its source is not given, but the probabilities are that it came from Anam. Dr. E. D Merrill, who has devoted much time to the consideration of Loureiro’s writings, informs me that he knows no reason for doubting Loureiro’s generic attribution in the present case. After a study of the description I am perfectly content to accept Loureiro’s name for this species. The leaves are given as ovate-lanceolate and glabrous. Unfortunately, how- ever, no information is given as to the shape or size of the corolla-tube. The second name applied to our species is Lithospermum viridiflorum. It first appears in 1814 as a name in a list of the Calcutta Garden and is given as collected by Roxburgh at Chittagong. It was no doubt this same garden material that was described in 1824 by Wallich in Rox- burgh’s Flora and is now represented in Wallich’s herbarium (no. 907). It is also the plant represented in Roxburgh’s unpublished plates (no. 2120) now preserved at Kew. The first description of the plant, as Heliotropium viridiflorum, is that by Lehmann in 1818. His material also seems to have come from the Calcutta Garden. Hence, there is 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 155 every reason for taking the material grown at Calcutta as typical. This is a form characterized by distinctly lanceolate leaves that are velvety all over beneath and by small strigose corollas. The corolla-tube is 2-3 mm. long, usually densely strigose and commonly only twice the length of the calyx or less. This form has not been collected about Chittagong. As Clarke has indicated, I. c. 146, the common form of T. montana about Chittagong, particularly in the region in which Rox- burgh is known to have collected, is the plant with long corolla-tubes described by Clarke as T. Roxburghii. As matters stand, therefore, we may either believe that Roxburgh did not collect his plant at Chitta- gong, or that having collected the common long-tubed Tournefortia there it subsequently became a short-tubed form under garden conditions. I have seen material of the type-form of T. viridiflora from Assam, Burma and Siam. The type of T. viridiflora var. Griffithii is a collection made in the Khasia Hills by Griffith. It differs from the type-form of 7. viridiflora in having the leaves much less hairy or nearly glabrous beneath and corollas that are possibly a trifle larger. The type of 7. Boniana col- lected by Bon (no, 1932) at O-cach, on the mountain Ma-dong in Indo- China, is quite similar. I have seen this glabrescent small-flowered form from Assam, Burma and Indo-China. In publishing T. Wightii, Clarke gave its source as “Deccan Penin- sula, Wight.” The type is Wight no. 2056 and is accompanied with one of the old printed labels indicating that it was part of the Wight materials handled at Kew in 1866-67. The label proper is headed “Peninsula Indiae Orientalis.” We may accept that no. 2056 was part of the Wight Herbarium but as to the collector of the specimen and its original source doubt must remain. Since the plant agrees closely with plants from Burma I suspect that perhaps it came from that general region and may represent material received by Wight from Roxburgh or some other collector of that period. Gamble, Fl. Madras, 894 (1923), reports the species from the Anamalai Hills, Madras. The only Tourne- fortia I have seen from that general region is T. Heyneana! Until un- doubted material from Southern India is forthcoming I believe that T. Wightii should be accepted as clearly applying to the material east of the Ganges here discussed. In the type-form of T. Wightii the leaf- surface is velvety beneath much as in typical T. viridiflora. The corolla is much larger, however, with the tube 2—4 times as long as the calyx. Tournefortia Roxburghii is a form of T. Wightii which has lanceolate rather than ovate leaf-blades. It is a rather common form. I have seen plants similar to the type-form of T. Wightii and T. Roxburghii from throughout the range of T. montana. 156 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Tournefortia Candollii is based upon “T. Heyneana, DC. Prodr. ix. 516; Wall. Cat. 910, as to the Khasia examples.” In the Wallich Cata- logue no. 910 consists of two parts, 1. material from Heyne, the type of T. Heyneana Wall. and 2. material collected by de Silva at Pundua. DeCandolle’s specimen of Wallich 910 consists only of the second part of the Wallich number, that is to say, the material from Pundua by de Silva. This specimen was described by DeCandolle as T. Heyneana. Clarke, l. c. 145, recognizing that the name 7. Heyneana was obviously to be associated with Heyne’s material from southern India, gave a new name, 7. Candollii, to the plant improperly described as T.. Heyneana by DeCandolle. The type of 7. Candollii is accordingly de Silva’s material in the DeCandollean Herbarium. The specimen at Geneva is broken and poor but has good corollas. These are somewhat constricted at the throat and very similar to those found in the type of T. khastana. The leaves are lanceolate, dried brown beneath and nearly black above. They are very sparsely strigose above and have only scattered hairs along the principal veins beneath. I consider T. Candollii to be the form of T. montana with elongate corolla-tubes and glabrescent leaves. Belonging with it are several fur- ther synonymous forms. The type of 7. Sampsoni is from Sai-chu-shan caverns in the province of Kwangtung and is now deposited at the British Museum. There is some interesting variation within this collection. The corolla-tube is medium to long (5-8 mm.) and the lanceolate leaves are either distinctly appressed hairy or are quite glabrous beneath. The type of T. khasiana was collected by Clarke (no. 15227) at Nonpriang in the Khasia Hills. It is a form of 7. Candollii in which the corolla- tube is contracted upward toward the throat so that the throat is at times almost half the diameter of the base of the tube. The type of T. Gaudichaudii is a glabrescent plant with elongate corollas and broadly lanceolate leaves. It was collected in Anam (Tourane) by Gaudichaud. 7. Tournefortia luzonicasp. nov., scandens grisea; ramulis obscure tetragonis 2—4 mm. crassis pilis numerosis brevibus divergentibus ves- titis; foliis oppositis vel suboppositis; petiolis 5-14 mm. longis; lamina folii ovata vel late lanceolata 5-13 cm. longa 2—7 cm. lata apice breviter acuminata basi rotunda vel (1-4 mm. profunde) cordata, supra pilis rigidulis brevibus ascendentibus plus minusve numerosis vestita, subtus pallidiore pilis gracilibus falcatis saepe numerosis vestita nervis 6—9- jugatis ornata; inflorescentia hispidula; calycibus sessilibus 1—2.5 mm. altis, lobis anguste lanceolatis vel linearibus erectis; corolla virescenti- bus, tubo 2-4(—8) mm. longo, limbo 2—2.5 mm. lato; fructu globoso 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 157 3—4 mm. diametro albo glaberrimo succoso; nuculis 2 biovulatis laevibus. Endemic to the Philippines where it is confined to the mountainous regions of northern, east-central and southern Luzon. SPECIMENS EXAMINED: vicinity of Penablanca, Cagayan Prov., a vine on hillside, fl. green, fruit white, May 3, 1917, M. Adduru 237 (type, herb. Arnold Arboretum; isotype, Kew); Pefiablanca, 1926, Ramos & Edano 46663 (BM); Bangui, Prov. Ilocos Norte, Ramos 27563 (BM); Burgos, Prov. Ilocos Norte, Ramos 4799 (BD); Bocana del Abra, Prov. Ilocos Sur, Micholitzs (K); Mt. Pulog, Mountain Prov., Jan. 1909, Curran, Merrill & Zschokke 16103 (BD); Benguet, Loher 1541, 1542 (K); dist. of Lepanto, Mountain Prov., Vidal 3326 (K); Baguio, Benguet, Elmer 8467 (AA, K); Mt. Maquilong, Prov. Batangas, Vidal 3327 (K); Prov. Albay, Cuming 1215 (K, BM, BD). 7A. Tournefortia luzonica var. sublucens,var. nov., a forma typica speciei differt foliis sparse inconspicueque pubescentibus, supra vix griseis sed sublucentibus. Confined to the mountains of west-central Luzon. SPECIMENS EXAMINED: Anuling, Zambales Prov., 1924, Ramos & Edano 44553 (TYPE, herb. Arnold Arboretum; isotypes, Kew, Brit. Mus.) ; Zam- bales, 1907, Ramos 4799 (BD); Lamao, Mt. Mariveles, Bataan Prov., Meyer 2844 (K, BD); Lamao River, Mt. Mariveles, 350 ft. alt., slender vine growing over trees for many yards, Williams 525 (K). Among all the Old World species of Tournefortia this species is unique in the possession of opposite or subopposite leaves. In the treatments of the Philippine Boraginaceae by Robinson, Philip. Journ. Sci., Bot. 4: 694 (1909), and by Merrill, Enum. Philip. Pl. 3: 376 (1923), this plant has generally passed as T. Horsfieldit Miquel. That species, however, with its alternate leaves and a fruit composed of four uniovu- late nutlets is one of the forms of the widely ranging T. sarmentosa. The var. sublucens is confined to the mountainous country of west central Luzon, prov. Bataan and Zambales, and seems to have a range quite distinct from the typical form of T. luzonica which ranges in the other parts of the island of Luzon. Essentially a glabrate form of T. luzonica, with the upper leaf-surfaces more or less glossy, it is signifi- cant and worthy of nomenclatorial recognition only if it has a range apart, and is geographically correlated. 8. Tournefortia Muelleri, nom. nov. Tournefortia mollis F. v. Mueller, Frag. 1: 59 (1858); Bentham, Fl. Austral. 4: 390 (1869) ; Bailey, Queensland Fl. 4: 1041 (1901); not T. mollis Bertol. (1852). Northern Australia and Papua. SPECIMENS EXAMINED: Edgecombe Bay, Queensland, Dallachy (K); along Burdekin River, Mueller (K, isotype); Herbert River, Dallachy 158 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XVI (K); Cape York Peninsula Exped., Hann 146 (K); shores of Montague Sound, W. Australia, 1820, Cunningham 182 (K, BM) and 324 (BM); erect shrub 1.5-2 m. tall, fringing tidal areas, Kapa Kapa, Papua, Brass 505 (AA, K); Port Moresby, Papua, 1918, White 6 (AA). The carpels seem to be more bony than in other Old World species of this section. The leaves are usually lanate. 9. Tournefortia puberula Baker, Jour. Linn. Soc. London, 20: 211 (1883). Tournefortia Mocquerysi A. DeCandolle, Bull. Herb. Boiss. ser. 2, 1: 581 (1901). Forests of eastern Madagascar and the Seychelles. Possibly intro- duced in the latter archipelago. SPECIMENS EXAMINED: MADAGASCAR: forests east of Ivohibé, 1000 m. alt., fl. white, Nov. 3, 1924, Humbert 3163 (P); high valley of the Rienana, drainage of the Matitana, 1000—4000 m. alt., fl. white, Nov. 1924, Humbert 2523 (P); Central Madagascar, Baron 1957 (Kew, Type of T. puberula; BM, BD, isotypes), a (K, BM, P), 3106 (K, P) and 6991 (K); forest of Ivohimanitra, Nov. 8, 1894, Forsyth Major 64 (K, BM, BD, P); forest of Analamazaotra near d’Amboasary, ca. 950 m. alt., sinvidte with white flowers, Oct. 23, 1912, Viguierro & Humbert 978 (P); forest at head of Antongil Bay, a liana with white flowers, Mocquerys 161 (Deles., TYPE of T. branched yst). SEYCHELLES: Mahé, common shrubby climber in hills nez eams, Sept. 1871, Horne 247 (K); Mahé, hope (K); Mahé, yaaa 10 (BM); Mahé, 1867, Wright (BM); Terné, Mahé, 1908, Gardiner (K); Silhoutte, common in cultivation, 1908, Gardiner 113 (KK); indefinite, a twining shrub generally on rocks near rivers, May 1902, Thomassett 22 (K, BM). The types of T. puberula and T. Mocquerysi are quite indistinguish- able. The species is a readily recognizable one. The leaves are firm, apparently glabrous and the stems are covered with a minute brownish puberulence. There is both a short- and a long-corolla form. The plant of the Seychelles is certainly identical with that of Madagascar. Pos- sibly it represents a horticultural introduction to the islands. In accounts of the Seychelles flora, Baker, Fl. Mauritius and Seychelles 202 (1877), and Summerhayes, Trans. Linn. Soc. London, Zodlogy, 19: 284 (1931), the species has consistently been misdetermined as 7. sar- mentosa. 9A, Tournefortia puberula var. Kirkii, var. nov., a varietate genu- ina differt pilis brevibus pallidis adpressis ornatis. Islands off the northwest coast of Madagascar. SPECIMENS EXAMINED: . Island, Comoro Archipelago, April 1861, J. Kirk as “Tournefortia (3)” (TyPE, herb. Kew) ; Nossi-bé, June 1847, Boivin 2086 (P); Nossi-bé, ed Perrille (P). This variety comes from a much more arid region than typical T. 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 159 puberula and may be only a hairy xerophytic form of that species. In typical T. puberula the plant is provided with a minute, frequently some- what golden puberulence. In the var. Kirkii the stems have a sparse pale short strigosity that tends to disappear with age. The petioles are sparsely strigose. The lower surface of the leaf-blades has short white closely appressed hairs scattered along the rib and veins. The upper surface is somewhat strigose but less abundantly so than below. The inflorescence has numerous short ascending pale hairs. 10. Tournefortia acuminata DeCandolle, Prodr. 9: 520 (1845); Cordemoy, Fl]. Réunion 479 (1895). Endemic to the Island of Reunion (Bourbon). SPECIMENS EXAMINED: les hauts du Boucan Launay, Boivin 1241 (K, BD, DC, Boiss, P); Bébour au dessus de la plantation de Quinquinas, July 28, 1875, G. de l’Isle 499 bis (K, Coss.) ; Bourbon, arbor, [ ?Commer- son] (herb. Smith); chemin que conduit de Sante Rose a Sain Joseph avaunt la descente qui conduit au Volcan, 1812, Commerson (P) ; “I’Te de France au bourbon,” ex Mus. Paris (type, herb. DC). The type of 7. acuminata at Geneva is given as distributed from Paris in 1821 and as from either Reunion or Mauritius. It represents a form in which the stems, petioles and inflorescence are glabrous or only very scantily strigose. It is obviously a duplicate of the collection at Paris which is labeled as collected by Commerson on the road between Ste. Rose and St. Joseph on Reunion. The material which I have cited from Boivin, which is widely distributed in European herbaria, is a form in which the stems, petioles and inflorescence have a short and evident, though not very abundant nor very conspicuous strigosity that becomes more or less brownish or golden. This I believe is the common form of the species. The leaves in T. acuminata are 12-17 cm. long and 3.5-7 cm. broad, are acute at both ends, and have 10-15 pairs of nerves. The calyx is 1.5-2 mm. long at anthesis and has erect, cuneate or more or less lanceolate lobes. The corolla-tube is 3.5-7 mm. long. The limb is ca. 4 mm. broad. 11. Tournefortia arborescens Lamarck, Tab. Encyc. 1: 417 (1791); Poiret, Encyc. 5: 357 (1804). Tournefortia velutina Smith in Rees, Cyclop. 36: sp. no. 13 (Aug. 1817!), not T. velutina HBK. (1818). Tournefortia Bojeri sensu Cordemoy, Fl. Réunion 479 (1895). Endemic to the Island of Reunion (Bourbon). SPECIMENS EXAMINED: Grand Bassin, Aug. 6, 1875, it de l'Isle 454 P); Gauteuron (spelling ?) du Gol, woods, fl. white, Commerson (herb. Smith, Type of T. velutina) ; Reunion, Commerson bes. Smith, r 160 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI second sheet of T. velutina), Bory (Deles) Boivin (BD) and Guyot 431 (BD); “in Mauritius,” herb Bojer as T. bifida (BM); “de Vinde” [? Sonnerat| (Paris, Tyee of T. arborescens). The type material of T. arborescens is accompanied by a small label reading: ‘“Tournefortia d inde.” The collector is not indicated but both Lamarck and Poiret attribute it to Sonnerat who visited the Masca- renes during his voyage to India and Malaysia. The material consists of two sheets, one bearing a sterile shoot with entire oblanceolate leaves more or less tomentose beneath in the manner common in the spicate Cordia species of the section Varronia. The second sheet contains a Tournefortia in flower. The latter is a form of the species as here defined, having the leaves only very sparsely and obscurely strigose, particularly above. The stems bear numerous but not very abundant short appressed pale hairs. The calyx-lobes are ovate, acute and sparsely pale strigose. The specimen evidently represents the sparsely hairy form of the endemic species of Reunion. The type of T. velutina is the form of the species with very abundant long hairs. It has the leaves pale and silky with a dense indument of slender very pale hairs. The calyx lobes are ovate, densely hairy and more or less golden tawny. Smith mentioned atypical material of his T. velutina from Mauritius, but this, in fact, represents a form of T. Bojeri. DeCandolle, Prodr. 9: 514 (1845), incorrectly cited 7. velutina as a possible synonym of 7. argentea. I have cited above a specimen given as collected on Mauritius by Bojer. I doubt the accuracy of the geographical data and believe that the specimen is really from Reunion. Its broad calyx-lobes are ovate or orbicular-ovate and hence similar to those found in all material indubitably from that island. 12. Tournefortia Bojeri A. DeCandolle, Prodr. 9: 516 (1845); Baker, Fl. Mauritius, 202 (1877). Tournefortia bifida sensu Bojer, Hort. Maurit. 234 (1837). Endemic to the Island of Mauritius (Ile de France). SPECIMENS EXAMINED: Mauritius, woods, 1837, Bojer as T. bifida (type, herb. DC); without locality, 1839, Bouton as T. bifida (DC, co- type); Mauritius, mountains and forest, Bouton (K); Mauritius 1854, Boivin (K); Mauritius, 1811, Hardwick (BM); Mauritius, | Commer- son] (herb. Smith) ; Mauritius, Sieber 98 (BD) ; Mauritius, herb. Labillar- dire (Deles) ; “Bourbon,” 1853, Boivin (Boiss): Boivin’s collection which it cited above and attributed to Reunion is, I believe, mislabeled. Indubitable collections of T. Bo jeri come only from Mauritius. The species is very closely related to T. arborescens of Reunion, differing chiefly in the narrower calyx-lobes. In the DeCan- dolle Herbarium there is a branch of 7. Bojeri, mounted on a sheet with 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 161 isotypic material of the Philippine T. Urvilleana. The only label accom- panying this mixed sheet is in the script of Chamisso and belongs to the Philippine species. This mixed sheet makes comprehensible DeCan- dolle’s, Prodr. 9: 515, adnot. (1845), strange comparison of T. Bojert and T. Urvilleana. Since Chamisso never visited the Mascarenes it is evident that the spray of 7. Bojeri has somehow become divorced from its proper label. The name “7. cymosa Heyne” seems to be based upon material from Mauritius. For a discussion of this nomen see my list of doubtful and excluded species on p. 166. In T. Bojeri the stems, petioles and inflorescence are more or less velvety with a pale ascending or spreading (or very rarely appressed) usually abundant hairs. The leaf blade is acute at both ends, more or less strigose on both surfaces though usually less so above. It has 10-12 pairs of veins and is 11-17(—24) cm. long and 2—5(—6) cm. broad. The calyx is 2—2.5 mm. long and has the lobes cut at least 34 way to base. It is more or less silky strigose. The lobes are lanceolate to broadly lanceo- late or cuneate-lanceolate. The corolla-tube is 2-8 mm, long, and 2-5 times the length of the calyx. The corolla-limb is 2-3 mm. broad. The fruit is ca. 3 mm. in diameter. Messerschmidia Linnaeus ex Hebenstreit, Nov. Comment. Acad. Sci. Imp. Petrop. 8: 315, tab. 11 (1763); Gmelin, Fl. Sibir. 4: 77 (1769) ; Murray, Syst. Nat. ed. 13, 161 (1774); Linnaeus fil. Suppl. Pl. 132 (1781).—type-species, Tournefortia sibirica Linn. Messersmidia Linnaeus, Hort. Upsal. 36 (1748); Linnaeus, Mant. 1: 5 and 42 (1767); Linnaeus, Syst. ed. 12, 149 (1767); Linnaeus, Mant. 2: 334 (1771). — a variant spelling of Messerschmidia, type-species, Tournefortia sibirica Linn. Tournefortia sect. Messerschmidia (Linn.) DeCandolle, Prodr. 9: 528 (1845); as to nomenclatorial type only, not as to the species of Heliotropium treated. Argusia Amman, Stirp. Rar. Ruth. 29 (1739). Arguzia [Amman] Rafinesque, Sylva Tellur. 167 (1838); Steven, Bull. Soc. Nat. Moscow 241: 558 (1851). — type-species, Tournefortia sibt- rica Linn. Tournefortia sect. Arguzia [Amman] DeCandolle, Prodr. 9: 514 (1845); Ledebour, Fl. Ross. 3:97 (1847-49). — type-species, Tournefortia sibirica Linn. Tournefortia sect. Mallota A. DeCandolle, Prodr. 9: 514 (1845). — type-species, 7. argentea Linn. Tournefortia sect. Mallotonia Grisebach, Fl. W. Ind. 483 (1861).— type-species, Tournefortia gnaphalodes R. Br. ex R. & S. Mallotonia (Griseb.) Britton, Ann. Mo. Bot. Gard. 2: 47 (1915). — type-species, Tourne- fortia gnaphalodes R. Br. Segregated here, as the emended genus Messerschmidia, are three 162 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI remarkable species that depart widely in appearance from the numerous and habitually very uniform species formerly associated with them in Tournefortia, As I have redefined and amplified Messerschmidia it con- sists of the original Asiatic herb, Tournefortia sibirica, the strand-shrub of the Antilles, 7. gnaphalodes, and the well known strand-tree of the Indian and Pacific oceans, T. argentea. All these species differ widely not only in their habit of growth and in their selection of habitat from all the other species that have been traditionally placed with them in Tour- nefortia, but also in their pronounced development of a corky exocarp which sets them off not only from all species of Tournefortia but from all other Boraginaceae as well. All three of the species show a marked preference for saline conditions. Two of them are tropical strand-plants. The third species grows along the ocean in temperate eastern Asia, in more or less saline soils along streams and about inland seas in Central Asia and eastern Europe. The corky exocarp evidently adapts the three species for water dispersal. The nature of the hairy covering of these three species is of an essentially similar type, consisting of slender silky hairs rather different in texture and appearance from that predominat- ing among the species of true Tourne fortia. The generic name Messerschmidia (also spelled Messersmidia and Messerschmidtia) is based upon Tournefortia sibirica Linn., and is a synonym of Argusia (or Arguzia). The type-species was first described by Amman in 1739 who applied to it the mononomial, Argusia, and gave a lengthy description of it based upon notes and specimens made by D. G. Messerschmid in 1724 along what is now the northwestern frontier of Manchuria. The source of this material is given as “Locus in glareosis aridisque apricis Argun fluuii et Iike Dalai Noor in Dauria.” Although Amman’s mononomial was formed from the “loco natali” of the plant, i.e. the Argun River, its author deliberately and repeatedly spelled it “Argusia”! Amman states that seeds from Messerschmid’s collection germinated and grew in the gardens at St. Petersburg. These same cul- tures are probably those described and illustrated by Hebenstreit in 1763. The plants growing in the Upsala Garden in 1748 and described by Linnaeus as Messersmidia were probably derived from those grown by Amman. In the Correspondence of Linnaeus, ed. Smith 2: 200 (1821), there is a letter from Amman, dated Nov. 18, 1740, in which questions by Linnaeus concerning Argusia are answered and in which it is stated that dried specimens of Argusia were being sent him. When he proposed the mononomial “Argusia,” Amman justified his use of a geographic appellation in forming the name, but added that he had no objection if the genus was named after Messerschmid, its original col- 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 163 lector. Linnaeus seems to have preferred the latter. The collector’s name was spelled ‘Messerschmid” by his contemporaries. Linnaeus latinized it, “Messersmidia,” and was consistent in this usage in all his writings. Other writers of the last half of the 18th century, however, spelled it “Messerschmidia” and it is so spelled in the paper by Heben- streit who was the first to use the generic name subsequent to 1753. Writers of the past century tended to spell the generic name “Messer- schmidtia.” The generic name Messerschmidia has variant spellings in “Messersmidia” and “Messerschmidtia.’ Although clearly based upon, in fact named after the original collector of Tournefortia sibirica, the generic name Messerschmidia (variously spelled) eventually became associated with two other very diverse groups of Boraginaceae. A study of the facts here presented, however, makes it evident that the name “Messerschmidia” is only very improperly applied either to the Ameri- can species of Tournefortia sect. Cyphocyema, or to Canary Island and South African species of Heliotropium as has been done in some large works. In another paper, Contr. Gray Herb. 92:73 (1930), I have given many facts concerning the misuse of the name “Messerschmidia.” The name was based upon Tournefortia sibirica and was originally applied solely to that plant. The type-species of Messerschmidia is obviously and logically the original Siberian species. KEY TO THE SPECIES Plant a low herb from rhizomes; inflorescence a loose open corymbose cyme; calyx pedicellate, lobes cuneate; anthers several times as long as broad; fruit pubescent, sunken in at apex ; carpels embedded in the center of the corky exocarp; femmerate HGTARia (ii sis iss cs ices er ivea aeetigeenes 1. M. sibirica. Plant a tree or shrub; inflorescence of scorpioid cymes; calyx sessile, lobes orbicular or oblong; anthers about two times as long as broad; fruit glabrous, apex conic or rounded ; carpels occupying the apical half of the fruit, the lower half com- posed entirely of corky exocarp; tropical strand plants. A tree 1-5 m. tall; leaves broadly oblanceolate or obovate, 5-9 cm. broad; inflorescence a conspicuous stiff panicle of loosely flowered elongating (up to 8 cm.) scorpioid cymes; corolla-lobes merely imbricate (not plicate) in the bud: anthers partially exserted from the short corolla- tube; fruit dull, breaking in half; apex and dorsal surface of carpels covered with corky exocarpial tissue ; tropics of the Old Wold é S. Oliveri Britt. cai Ria pubescent; leaves less than 3 cm. broad, broadest r below the middle. Be broadest near the middle, beneath silky strigose, lus- trous, blade 4-7.5 cm. long; fertile branches 3-5 cm. long, bearing about 5 ee old branches brown, with evident pale 1énticels..axcasce oe eee eee nae Ewa S. brasiliense n. sp. 182 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Leaves broadest near base, beneath velvety or somewhat tomentose, dull, blade 5-12 cm. long; fertile branches 10-20 cm. long, bearing about 10 leaves; old branches gray or only rarely brown, without evident lenticels.... er er a ee ee Te ere S. lanceolatum H. & B. Saccellium Oliverii Britton ex Rusby, Bull. Torr. Bot. Cl. 26: 147 (1899). This species is known only from the type-collection made by Rusby, no. 2535, in May 1886 at 600 m. alt at Guanai (or Huanay), Bolivia. The locality is in the department of La Paz at the confluence of the Rio Mapiri and Rio Tipuani at about lat. 15°30’ S. and long. 68° W. in Amazonian Bolivia. Only fruiting specimens of the species are known. In its slender brownish branches, rather evident pale lenticels, and general leaf-outline, the species resembles S. brasiliense more than it does S. lanceolatum. The leaves are glabrous except for a few short inconspicuous ascending hairs along the midrib and principal veins. A similar scanty inconspicuous indument is also found in the inflorescence. Saccellium lanceolatum Humboldt & Bonpland, Pl. Aequin. 1: 47, tab. 13 (1806); Humboldt, Bonpland & Kunth, Nov. Gen. 7: 209 (1825); Miers, Trans. Linn. Soc. London, Bot. 1: 25, tab. 6 (1875). In two widely separated areas, 1. Northern Peru in northern parts (prov. Jaen) of the Dept. Cajamarca, lat. 5°-6° S., in the Amazonian drainage; 2. mountains of southern Bolivia (prov. Chuquisaca and Tarija) southward along the mountains of northern Argentina to Tucu- man, ca. lat. 27 Peru: between Jaen and Bellavista, prov. Jaen, 600-700 m. alt., shrub or small tree, common, April 29, 1912, Weberbauer 6209a (BD); Valley of the Maranon between Bellavista and ~ mouth of the Rio Chinchipe, prov. Jaen, 500 m. alt., small tree 4 m. tall, flowers white; accrescent calyces yellow-green, May 1, 1912, W ise 6226 (G, BD); Valley of the Maranon at the mouth of the Rio Chinchipe, prov. Jaen, 400-500 m. alt., tree 6 m. tall, flowers white, accrescent calyces yellowish green, April 30, 1912, Weberbauer 6217 (G, BD); near Rio carnhpechri Bonpland (Type, Paris; fragments, DC, Lindl., Gray). Bortvia: between Atajado and ov 700 m. alt., small tree, Dec. 1910, baa 1192 (BD); south o Pilcomayo, prov. Tarija, Feb. 18, 1916, Steinbach 1776 (BD); Bolivia, Pampas, evergreen tree 4.5-6 m. tall, woods, May 1864, Pearce (BM). ArcentinaA: Islota, Sierra Sta. Barbara, Jujuy, dry open place, tree 15-20 m. tall, July 5, 1901, Fries 260 (Munich) ; Sierra Sta. Barbara, Salta, Schuel 38 (G); Rio Blanco, dept. Oran, Salta, 650 m. alt., flowers yellowish, tree 10 m. tall, trunk 5 dm. thick, in high forest, Nov. 19, 1927, Venturt 5546 (AA, G, K, BM); Abra Grande, dept. Oran, March 1927, O m. alt., tree 5 m. high, flowers yellowish, Venturi 6780 (AA); Rio Piedras, dept. Oran, Nov. 15, 1911, Rodrigues 85 (G); Campo Duran, 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 183 dept. Oran, a tree common on higher slopes, “‘Guayabil,’ Jan. 28, 1930, Parodi 9269 (G); Tartagal, Salta, a tree, Feb. 1923, Hauman (G); hills near the crossing of the Rio Juramente, Salta, tree or shrub up to 6 m. tall, Feb. 21, 1873, Hieronymus & Lorentz 295 (BD, Deles) ; Alemania, dept. Guachipas, Salta, 1100 m. alt., flowers white, tree 6 m. tall, in high forest, trunk 2 dm. thick, Dec. 22, 1929, Venturi 10005 (G, K, BM); El Cadillal, dept. Burruyacu, Tucuman, Dec. 20, 1909, Lillo 9823 Steet Tucuman, dept. Capital, alt. 450 m., tree 10 m. tall, flowers white, eho 1907, Lillo 7234 (G); Estate of Protessor Lillo, dept. Capital, eee 460 m. alt., March 1925, Venturi 3816 (AA); Tucuman, Dec. 12, 1907, Stuckert 18375 (Deles) ; Tucuman, Feb. 10, 1910, Lillo (G). The distinctly lanceolate leaves and the leafy, elongate stiffish branches readily characterize this species. The range of the tree is peculiar for it occurs in two far-separated regions in Peru and Argen- tina. Though this behavior suggests that two species or that a species and a variety is involved, a careful comparison of copious material has failed to produce any differences that would justify the proposal of even a new variety. The Peruvian plant differs from that of Argentina only in its perhaps somewhat sparser and slightly more slender pubescence on the herbage and in its somewhat darker stems. In the Plantae Aequinoctiales 1: 47 (1806), the source of the origi- nal Humboldt & Bonpland collection of Sa@ccellium is given as ‘“‘ad rivos fluvii Guancabamba.” Similar data are on the type-specimen at Paris. In the Nova Genera, 7: 208 (1825), the locality is given in more detail as follows: “inter Loxam et Tomependam Bracamorensium, ad ripas fluminis Guancabambae.”” The locality, Loja, of course, is in southern Ecuador. Tomependa is a ruined village near the junction of the Rio Chinchipe and the Rio Maranon. The Rio Huancabamba joins the Maranon about 50 km. above Tomependa. In all probability the type was collected in or near the province Jaen, in the region of northern Peru in which it has been collected by Weberbauer, Bot. Jahrb. 50: suppl. p. 92 (1914). Coldenia conspicua, sp. nov., prostrata ut videtur annua; caulibus articulatis laxe ramosis 2-15 cm. longis, juventate dense graciliterque hispidulis et plus minusve glanduliferis; foliis aggregatis numerosis, lamina late lanceolata vel elliptica 5-13 mm. longa 2—5 mm. lata, subtus prominenter costata et nervosa (nervis 2—3-jugatis vix conspicuis) pilis gracilibus brevibus numerosis erectis asperata, supra pustulosa pilis robustioribus longioribus rigidioribus numerosis ascendentibus asperata, margine laxe revoluta integra vel obscurissime sparsissimeque crenata; petiolis gracilibus 2-9 mm. longis glanduliferis pilis abundantibus longis gracilibus erectis setosis; calyce 5-partito, lobis gracilibus basim versus 184 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI subinduratis et subnavicularibus praeterea linearibus hispidis glandu- liferis ad anthesin ca. 9 mm. longis fructiferis ad 15 mm. longis; corolla conspicua coerulea, tubo ca. 9 mm. longo 2.5-3 mm. crasso lobis calycis subaequilongo intus glaberrimo, limbo 10-12 mm. lato patenti, lobis 4—5 mm. diametro, faucibus haud appendiculatis, filamentis 4-5 mm. longis glabris apicem versus tubi affixis ca. 2 mm. longe extrusis, antheris oblongis medio-affixis 1-1.4 mm. longis; stylo filiformi glabro 15 mm. longo 2 mm. profunde bilobato, stigmatibus 2 minutis obscure bilobu- latis; nuculis 4 globosis 1.5—-2 mm. diametro dense minuteque tessellato- tuberculatis per carunculas 1 mm. longas et crassas in apice receptaculi basaliter affixis; receptaculo ad anthesin cylindrico, fructifero turbinato. Peru: sand flat near Mejia, Dept. Arequipa, 40 m. alt., flowers blue, Oct. 26, 1923, Guenther & Buchtien 155 (tTypE, Inst. Bot. Hamburg) ; Mejia, July 21, 1923, Guenther & pica 156 (Hamburg); Mollendo, Dept. oes. Miss D. Stafford K60 (Kew). A very distinct and remarkable species i engi to the Chilean and southern Peruvian section Sphaerocarya, Johnston, Contr. Gray Herb. 70: 57 (1924). The nutlets of the new species are quite similar to those of this section in size, shape and markings. From the previously de- scribed species of the section, however, C. conspicua differs in its ex- tremely large corollas, its protruding stamens and its remarkable nutlet- attachment. The corollas are at least twice the size of those of any other species of Coldenia. The nutlet-attachment is also unique in the genus. In the known species of the section Sphaerocarya the immature nutlets are attached laterally at the middle of the sides of an erect subcylindrical gynobase. This is distorted somewhat by the crowding of the growing nutlets and tends to become constricted medially. After the nutlets have fallen away it is consequently more or less spool-shaped. In the proposed species the immature nutlets are borne laterally, not about the middle, but about the summit of the subcylindrical gynobase. By growth and the consequent pressure of crowding, the nutlets at ma- turity come to be attached basally in the expanded summit of the now turbinate gynobase. What is most peculiar is that each nutlet has a well developed strophiolate basal plug which is immersed in the gynobasal tissue. At maturity the strophioles loosen from the gynobase and with their attached nutlets fall away leaving 4 deep more or less united sock- ets in the much broadened apex of the gynobase. The mature gyno- base, hence, becomes more or less cupulate. The species is known only from along the coast in southern Peru in the general region of the port of Mollendo. The type has been reported, Bruns, Mitt. Inst. Allgem. Bot. Hamburg 8: 67 (1929), as C. dicho- 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 185 toma, but that species, of course, has small corollas and utterly different fruit-structures. The other species of the section Sphaerocarya are poorly understood. Since publishing, 1. c., on the South American species of Coldenia I have seen the types of Philippi’s species. I have been unable to separate his C. litoralis, C. atacamensis and C. parviflora, though from geographic considerations one would expect that the plant from the coastal region (C. litoralis) would be distinct from that of the high Puno de Atacama (C. atacamensis and C. parviflora). The type of C. parviflora is quite distinct from the Peruvian plants of the Arequipa region, which I cited under that name in my synopsis of the South American species of Coldenia. The correct name for this species is C. elongata Rusby! Its elongate leaf-blades, woolly petioles and calyces, and usually evidently crenate leaf-margins serve to distinguish it from Philippi’s species. Coldenia elongata is known only from middle alti- tudes east of the coastal deserts of southern Peru and northern-most Chile. In Peru only two species of the section Sphkaerocarya are known. These are C. conspicua which grows along the coast and C. elongata which grows along the cordilleras in the interior. Coldenia Nuttallii Hooker, Kew Jour. Bot. 3: 296 (1851); John- ston, Contr. Gray Herb. 75: 43 (1925). Coldenia decumbens Hauman, Apuntes Hist. Nat. Buenos Aires 1: 55 (1909) and Anal. Soc. Cient. Argentina 86: 301 (1918). This species so wide-spread in the intermontane area of the western United States has been known only from two small areas in the high cordilleras of Argentina, in northwestern San Juan, Johnston, Physis 9: 316 (1929), and in the Uspallata Pass region in Mendoza, Hauman, l.c. The plant was collected around 3000 m. alt. in San Juan and about 2300 m. alt. in Mendoza. A third locality for the species in South America, one much further south and so, not surprisingly, at lower alti- tudes, may now be put on record. I have seen a collection of C. Nuttalli in the herbarium at Munich which was obtained by Erik Ammann (no. 5) in Oct.-Nov. 1927, at 700 m. alt. near Cobunco, Neuquen, Argentina. Tournefortia brasiliensis Poiret, Encyc. 5: 357 (1804); Johnston, Contr. Gray Herb. 92: 89 (1930). I have studied the type of this doubtful species in the Lamarck Herbarium at Paris. It represents a specimen of Vernonia scorpiotdes (Lam.) Pers., with the flowers just beginning to develop. It is remark- ably like, and probably a part of the collections by Commerson made at Rio Janeiro (‘de l’ile aux chats”) in July, 1767. Consequently it may be a part of the same material as the type of Conyza scorpioides Lamarck, Encyc. 2: 88 (1790). 186 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI Heliotropium transalpinum Vellozo, Fl. Flum. 68 (1825) and Icones, 2: tab. 40 (1827). Heliotropium tiaridioides var. schizocarpum Johnston, Contr. Gray Herb. 81: 7 (1928), where other synonyms are cited. Vellozo in describing and illustrating his species gave no indication as to whether the carpels were dorsally sulcate or not. Suspecting that the carpels were sulcate, however, since only plants with such develop- ments were known about Rio Janeiro, I provisionally cited the name H.. transalpinum among the synonyms of my H. tiaridioides var. schizo- carpum. Vellozo’s name, unhappily, is several years older than H. tiari- dioides Cham., the species I then accepted. Subsequent study and con- sideration of much South American material of Heliotropium, not avail- able when my monograph was written, has left no reasonable doubt as to the identity of the plant described and illustrated by Vellozo. The scores of specimens examined from Sao Paulo, Rio Janeiro, Minas Geraes and northward in Brazil, uniformly have sulcate nutlets, and there seems every reason for believing that Velloza’s plant had them also. I am accordingly taking up H. transalpinum as the correct appellation for the plant formerly treated by me as H. tiaridioides var. schizocarpum. The southern plant with non-sulcate nutlets, which I treated as H. tiaridioides var. genuina must have the new name I am publishing below. The type of H. transalpinum was collected in the state of Rio Janeiro near Boa Vista, ca. 9 km. up the Rio Parahyba from the town of Para- hyba do Sul and beyond the coastal mountains (whence the specific name) from the city of Rio Janeiro. Heliotropium transandinum var. tiaridioides (Cham.) comb. nov. Heliotropium tiaridioides Chamisso, Linnaea 4: 453 (1829). Helio- tropium tiaridioides var. genuina Johnston, Contr. Gray Herb. 81: 6 (1928), where other synonyms are cited. Heliotropium angiospermum Murray, Prodr. Stirp. Gottingen 217 (1770); Johnston, Contr. Gray Herb. 81: 10 (1928). Heliotropium humile Lamarck, Tab. Encyc. 1: 393 (1791). In my treatment of the South American species of Heliotropium, Contr. Gray Herb. 81: 66 (1928), I cited H. humile Lam. as a doubtful synonym of H. fruticosum L. This I now find is incorrect. In the La- marck Herbarium at Paris there is only one specimen determined by Lamarck as H. humile, this bears a label in his script reading: ‘‘heliotr. humile lam, illustr.” The specimen is small but represents good H. angiospermum. The original description of Lamarck’s species reads: “1757 HELIOTROPIUM humile. H. foliis ovato-lanceolatis villosis; 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 187 spicis solitariis lateralibus. Ex ins. Carib, Annuum. H. Dict. no. 6 Quoad escr.” The reference is apparently to Lamarck’s earlier account of Heliotropium in vol. 3 of the Encyclopédie, pp. 92-95 (1789), but no mention of H. humile is to be found there. Species no. 6 in the work is H. fruticosum, described as having linear-lanceolate leaves. Poiret, Encyc. Suppl. 3: 25 (1813), was evidently puzzled by Lamarck’s de- scription of H. humile. He mentions that species under H. ternatum but suggests that it might be H. fruticosum. I am content, however, to place H. humile among the synonyms of H. angiospermum, for the named specimen in Lamarck’s herbarium seems authentic and agrees with the few words in the original description. Lasiarrhenum pinetorum, sp. nov., herba; caulibus erectis simplici- bus 10-15 cm. altis gracilibus strigosis foliosis; foliis lineari-subulatis 1—3 cm. longis 1—-1.5 mm. latis sessilibus medio-costatis sed vix nervatis apicem versus caulis gradatim reductis margine valde revolutis supra sparse strigosis; floribus cymas terminalis 3—7-floris lineari-bracteatas aggregatis; pedicellis 2-3 mm. longis strictis strigosis; calycibus 5-lobatis ca. 4 mm. longis, lobis lineari-lanceolatis strigosis; corolla flava ca. 10 mm. longa extus strigosa, tubo ca. 4 mm. longo ca. 1.5 mm. crasso in fauces 3.5-4 mm. longas ca. 3 mm. crassas abrupte transmutato intus glaberrimo, lobis erectis oblongis 2.5 mm. longis 2 mm. latis apice rotundis; antheris glaberrimis 2 mm. longis oblongis basi sagittatis erec- tis (vix versatilibus) sub medium affixis; filamentis 4 mm. supra basim corollae affixis inclusis 1-1.3 mm. longis late alatis (in ambitu obovatis) apicem versus ca. 0.7 mm. latis; stylo 12 mm. longo filiformi longe (ca. 5 mm. i. exsertis; fructu ignoto. MEXICO: growing in the mountains in pine-forest, very rare, September, Gin ecch: Sil (TYPE, Paris). This is a remarkable species which is placed in Lasiarrhenum chiefly because of its broadly winged filaments. From L. strigosum, formerly the only known member of its genus, it differs in its very small size, its uninerved leaves, its glabrous anthers and its precociously long-exserted style. The rounded corolla-lobes and the expanded filaments separate L. pinetorum from the genus Onosmodium, while the long-exserted style, the erect corolla-lobes, the obovate filament and the sagittate anthers distinguish it from Lithospermum. No locality is given for this inter- esting plant. Ghiesbreght, however, collected chiefly in southern Mexico and mostly in the state of Oaxaca. Lithospermum Muelleri, sp. nov., perenne; caulibus erectis gracili- bus foliosis simplicibus vel rariter stricte et simpliciter longeque ramosis 188 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI 2—5 dm. altis e radice crasso dense multicepite rumpentibus strigosis vel basim versus breviter hispidis; foliis strictis firmis costatis sed vix nervatis vel rarissime perinconspicue sparseque nervatis, inferioribus oblongo-ellipticis, aliter lanceolatis, sessilibus, apicem versus caulis gradatim reductis 1-4 cm. longis 3-8 mm. latis, apice acutis, supra minute strigosis et pustulatis, subtus in margine et costa strigosis sed ceterum glabris; inflorescentia bracteata scorpioidea terminali solitari vel gemi- nata vel ternata 3-10 cm. longa; calyce ad anthesin ca. 6 mm. longo, lobis inaequalibus cuneatis, pedicellis 1-3 mm. longis strigosis; corolla subcylindrica 15-19 mm. longa ca. 3 mm. crasso ut videtur flavescente intus glaberrima extus adpresse pubescente, lobis minutis ascendentibus suborbicularibus ca. 1 mm. diametro, faucibus inconspicue plicato- appendiculatis saepe plus minusve constrictis; staminibus 2 mm. sub apice tubi corollae affixis, filamentis ca. 1 mm. longis, antheris oblongis ca. 2 mm. longis inclusis; stylo filiformi ad anthesin 1-3 mm. longe extrusis; fructu ignoto. Mexico: common in pine belt above Mesa de la Camisa on the north slope of Sierra Tronconal between Cafion de los Charcos and Canon de San Miguel, Sierra Madre Oriental, ca. 25 km. s. w. of Galeana, Nuevo Leon, 1800-2700 m. alt., June 4, 1934, C. H. & M. T. Mueller 739 (TYPE, Gray Herb.). A very distinct species of uncertain affinities. Its subtubular corolla, frequently with a narrowly constricted ring about the throat, and its extremely small round ascending lobes, separate it from L. strictum, the only species I am inclined to believe which possibly may be a close relative of it. Macromeria leontis, sp. nov., perennis erecta ca. 5 dm. alta e radice crasso profunde oriens; caulibus subsimplicibus pilis brevibus gracilibus erectis vel subretrorsis dense vestitis; foliis lanceolatis medium versus caulis grandioribus 4-10 cm. longis 1—2 cm. latis utroque acutis sessilibus evidenter nervatis, subtus pilis gracilibus brevibus erectis abundantibus vestitis vix pustulatis, supra viridis pustulatis et breviter hispidis; in- florescentia terminali evidenter bracteata; calyce 1.5—2 cm. longo, lobis linearibus, pedicello 1-5 mm. longo; corolla 5—7 cm. longo intus glaber- rimo extus breviter pubescenti, tubo 2—3 cm. longo 1.5—-2 mm. crasso supra in fauces 2 cm. longos 7-8 mm. latos gradatim ampliato, lobis triangularibus ca. 9 mm. longis et 6 mm. latis non rariter plus minusve recurvatis; antheris elongatis ca. 3.5 mm. longis; filamentis ca. 4 mm. infra apicem faucium corollae affixis 10-15 mm. longe exsertis; stylo filiformi tarde exsertis; fructu ignoto. Mexico: scattered in dense oak-woods on the ascent into Taray, Sierra 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 189 Madre Oriental, ca. 25 km. s. w. of Galeana, Nuevo Leon, ca. 2400 m. alt., June 5, 1934, C. H. & M. T. Mueller 754 (type, Gray Herb.) ; scat- tered in dense pine and oak woods along the descent into Alamar, Sierra Madre Oriental, May 29, 1934, C. H. & M. T. Mueller 594 (G). Probably a relative of M. Pringlei, but differing in having a fine slen- der spreading indument throughout. In M. Pringle: the more rigid, somewhat longer sparser hairs are closely appressed and the upper leaf- surfaces are a much clearer green than in M. leontis. The latter species has leaves noticeably grayer and duller in color. Macromeria barbigera, sp. nov., perennis, setosa, robusta; caulibus erectis 5—8 dm. altis saepe simplicibus; foliis lanceolatis vel ovatis evi- denter nervatis, inferioribus parvis vix persistentibus, ceteris latioribus 3—5 cm. latis 5-11 cm. longis subsessilibus basi plus minusve rotundis, superioribus elongatioribus et minoribus; floribus terminalibus in cymas racemosas bracteatas aggregatis; bracteis foliaceis 2-7 cm. longis 1—4 cm, latis; pedicellis ca. 5 mm. longis; calyce ad anthesin ca. 18 mm. longo, lobis inaequalibus subulato-linearibus erectis; corolla ut videtur flavescenti intus glaberrima 5—6 cm. longa recta vel plus minusve curvata, tubo 1.5—2 cm. longo 1.5—-2 mm. crasso lobis calycis paullo longiore, faucibus e tubo abrupte ampliatis ca. 2 cm. longis 5—6 mm. crassis cylin- draceis in alabastro paullo asymmetricis, limbo abrupte dilatato 12-15 mm. diametro, lobis 5—6 mm. longis acutis ascendentibus apicem versus recurvatis; filamentis in faucibus ca. 8 mm. infra sinibus loborum affixis inaequalibus 12-15 mm. longis glabris filiformibus exsertis; antheris oblongis medio-affixis; stylo filiformi breviter tardeque extruso; stigmato minimo bilobulato; fructu ignoto. EXICO: common in dense oak wood beyond the pine and fir belt, north slope of Sierra Tronconal between Cafion de San Miguel and Cafon de los Charcos, 1800-2700 m. alt., Sierra Madre Oriental about 25 km. s. w. of Galeana, Nuevo Leon, June 4, 1934, C. H. & M. T. Mueller 741 (TYPE, Gray Herb.). Related to M. Thurberi but quickly separable by its more robust habit, larger broader leaves and very different pubescence. The foliage of M. Thurberi is copiously and finely strigose with an admixture of coarse more or less spreading hairs. In the proposed species the strigosity is lacking and the spreading hairs much longer and very conspicuous. The corollas of M. Thurberi have a much more abundant and paler indu- ment than do those of M. barbigera. The range of the new species is to the southeast of the most easterly station of its relative. Evidently to be identified with M. barbigera are collections made by Mueller in 1933. These specimens have been kindly sent to me from 190 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI the Field Museum by Mr. P. C. Standley. One of these collections, no. 174 from the “Trail to Puerto,’ Nuevo Leon, has leaves becoming 17 cm. long and 7 cm. broad. Its flowers are immature. The second col- lection, no. 173 from Diente Canyon, 21 km. south of Monterey, is evi- dently from a very mature plant and consists of the elongated inflorescence showing mature bracts and the old pedicels and calyces. Among his collections of 1934 Mueller obtained one which may also represent a form of M. barbigera. This specimen, no. 830, was collected on Sierra Infernillo, about 25 km. s. e. of Galeana, Nuevo Leon, where it was common over small areas just below the crest, 2700-3000 m. alt. In leaf-outline and in general habit the plant suggests M. Thurberi, but differs in its lack of strigosity and in its very much less hairy flowers. The corollas differ from those of M. barbigera. They are somewhat smaller. The tube is gradually expanded towards the lobes and not abruptly expanded into a well developed cylindrical throat as I have indicated in my formal description above. In addition the corolla is slightly less hairy and the lobes not so acute. The plant is evidently related to M. barbigera and chiefly because of geographical considera- tions I am tentatively, at least, referring it to that species as a possible ecological form. Havilandia opaca, sp. nov., procumbens; caulibus foliosis abundan- ter ascendenterque ramosis 1—-1.5 mm. crassis in nodis radiculas graciles gerentibus pilis brevibus rigidis appressis dense vestitis, internodiis 3-10 mm. longis; foliis firmis subcoriaceis costatis sed enervatis numerosis, apice rotundis vel obtusis, supra glaberrimis sparsissime pustulatis in costa sulcatis, subtus supra medium pustulatis in costa prominente strigosis ceteris glabris vel sparsissime strigosis, margine strigoso-ciliatis vel basim versus sparse ciliatis; foliis ramorum fertilium ellipticis 4-10 mm. longis 3—5 mm. latis, basi rotundis et oblique 1-2 mm. lateque sessilibus; foliis ramorum sterilium plus minusve oblanceolatis 8-12 mm. longis paullo sub apicem basim versus in petiolum 1 mm, latum ca. 2 mm. longum gradatim attenuatis; floribus solitariis numerosis axillaribus; corolla alba 4 mm. diametro, tubo ca. 1.2 mm. longo 1 mm. crasso intus glaberrimo, limbo patenti, lobis suborbicularibus ca. 1.5 mm. diametro, appendiculis faucium 5 intrusis trapeziformibus; antheris oblongis inclusis ca. 0.4 mm. longis, filamentis perbrevibus paullo supra medium tubi affixis; calyci ad anthesin 2 mm. longo, lobis 5 ciliatis latis, pedi- cello 0.5-1 mm. longo; nuculis 4 erectis angulate ovoideis 1 mm. longis opacis dense minutissimeque papillatis, dorso convexis, ventre angulatis, imam ad basim anguli ventralis ad gynobasim planum affixis. 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 191 British NEw as common in open grassland, Murray Pass, Whar- ton Range, 2840 m. alt., prostrate herb forming masses 3 dm. broad or more, flowers iti June 12, 1933, Brass 4178 (type, Gray Herb.; 1so- TYPE, NY). A species evidently related to H. papuana Hemsl., from which it differs in its stout somewhat ovate, gray, dull, minutely papillate, rather than elongate, somewhat lance-lunate, black, lustrous, smooth nutlets. The margins of the leaves in H. papuana are evidently ciliate. In H. opaca the marginal hairs of the leaves, similar in size, number and posi- tion, are not spreading, but antrorsely appressed along the leaf-margin. The habit of growth in H. papuana is quite similar to that of H. opaca. Havilandia robusta, sp. nov., procumbens; caulibus elongatis sparse ramosis; foliis coriaceis oblanceolatis 2—4.5 cm. longis 5-9 mm. latis paullo sub apicem basim versus in petiolum vaginatum gradatim attenu- atis, apice rotundis vel subemarginatis, margine sparsissime strigosis, supra sparse strigosis, subtus glaberrimis vix nervosis, costa prominente sparsissime strigosa; floribus axillaribus; calycibus ad anthesin ca. 4 mm. longis, lobis lanceolatis margine sparsissime strigosis, pedicellis 5—7 mm. longis; calycibus maturitate ca. 6 mm. longis pedicellis 8—12 mm. longis; corolla 8-10 mm. diametro; nuculis 4 angulato-ovoideis opacis ca. 2 mm. longis dense minutissimeque papillatis, dorse convexis, ventre angulatis. BritisH NEw GUINEA: common about forest borders, Mt. Albert Ed- ward, 3680 m. alt., June 1933, Brass 5681 (type, N. Y. Bot. Gard.). Evidently related to H. opaca, also of southeastern New Guinea, from which it differs only in being much larger in all its parts, and in having well developed pedicels and more elongate leaves. The upper surface of the leaves is lustrous and distinctly strigose. Havilandia papuana Hemsley, Kew Bull. 1899: 107 (1899). British New GuINea: thickly massed on shallow soil over rock in grasslands, Mt. Albert Edward, ee m. alt., flowers white with yellow throat, June 18, 1933, Brass 4245 (G, NY). This species was briefly, aouah es described by Hemsley from material obtained on Mt. Scratchley, 3660 m. alt., and in the Wharton Range, 3330 m. alt. It is known only from the high mountains of eastern British New Guinea. The genus Havilandia is confined to high altitudes and consists of the three above enumerated species from British New Guinea, and H. borneensis Stapf from Mt. Kinabalu in British North Borneo. It is possible, in addition, that Lithospermum minutum Wernh., described from the Mt. Carstensz region in Dutch New Guinea, may also belong 192 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI to Havilandia. Unfortunately the type and only known collection of this puzzling species is so scanty and inadequate that it must remain an obscure, troublesome element in the flora of New Guinea until some- one recollects it. The type consists of two minuscule snips in flower only, a ridiculously inadequate basis for the proposal of any species of Boraginaceae and certainly for one whose acquaintance with the genera of that family may be judged by his selection of the genus under which he essayed to publish the imperfect specimen from Dutch New Guinea. Plagiobothrys Scouleri (H. & A.) Johnston, Contr. Gray Herb. 68:75 (1923) and Contr. Arnold Arb. 3:51 (1932). Mvyosotts Scouleri Hooker & Arnott, Bot. Beechey Voy. 370 (1840). Allocarya media Piper, Contr. U. S. Nat. Herb. 22: 107 (1920). Plagiobothrys medius (Piper) Johnston, Contr. Arnold Arb. 3: 58 (1932). Allocarya divaricata Piper, Contr. U.S. Nat. Herb, 22: 107 (1920). The original and only mention of Myosotis Scouleri in the writings of Hooker & Arnott appears in the Botany of Capt. Beechey’s Voyage in a note on a collection of Plagiobothrys Chorisianus from California. The note is as follows: ‘“‘The flowers here are on pretty long pedicels, while the Columbia plant has them shortly pedicellate; the latter presents, besides, a different aspect, and may be called M. Scouleri; it appears very Closely allied to M. californica, Fisch. et Meyer, but the corolla is longer than the calyx.’’ Gray, who apparently never studied the type of M. Scouleri, applied the name to an erect-growing plant with geminate spikes of conspicuous corollas, which is widely distributed in western Oregon and Washington, and all subsequent writers have followed him in that identification. A study of the type, however, shows this usage to be quite incorrect. The specimens evidently the type of M. Scouleri are to be found on a mixed sheet, formerly in the Hooker Herbarium, now at Kew. This sheet bears three different collections: (1) the specimen of P. Chorisianus mentioned in the Botany of Beechey’s Voyage, (2) specimens of P. scopulorum (?) or P. cognatus (?) collected by Nuttall, and (3) three plants labeled: “N. W. Coast, Dr. Scouler.” The latter evidently constitute the type of Myosotis Scouleri H. & A. Duplicates of this Scouler collection are to be found on a sheet from Bentham’s herbarium, at Kew, labeled: “Am. bor. occ. Scouleri, 1828,” and in the herbarium at Edinburgh labeled: “Columbia, Scouler, 1827, (932). These collec- tions appear to represent a form of the plant I have treated in my mono- graph as Plagiobothrys medius (Piper) Johnston. They have the rufous calyx-lobes, evident corollas, and the general habit of that species. 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 193 Scouler is known to have collected about the mouth of the Columbia and at many small ports along the coast of Washington and Vancouver Island. Plagiobothrys medius is the common species near the coast in northwestern Washington and on Vancouver Island, and there is every reason that Scouler should have encountered it. Though the nutlets of Scouler’s collection show certain peculiarities not matched in the avail- able material of P. medius, I believe that they can be accommodated in that concept. The nutlets of the type of M. Scouleri have the rather bony pericarp common in P. medius, but the ridge attending the lateral scar is very closely appressed to the latter and encloses an areole (entirely filled by the scar) scarcely, if at all, broader than long. The nutlets of P. medius are, however, very variable and I believe the nutlet- variations of M. Scouleri can be admitted without destroying the natural- ness of the concept. Plagiobothrys hirtus(Greene), comb. nov. Allocarya hirta Greene, Pittonia 1: 161 (1888). Allocarya Scouleri var. hirta (Greene) Nelson & Macbride, Bot. Gaz. 61:36 (1916). Plagiobotkrys Scouleri var. hirtus (Greene) Johnston, Contr. Arnold Arb. 3: 52 (1932). Allocarya calycosa Piper, Contr. U. S. Nat. Herb. 22: 101 (1920). I have indicated above that the type of Myosotis Scouleri H. & A. has been misinterpreted. The earliest correct name for the plant that has been called Krynitzkia, Allocarya and Plagiobothrys Scouleri is Allo- carya hirta Greene. It is, however, strictly applied only to a local plant of the Umpqua Valley, Oregon, which has evidently spreading rather than appressed pubescence. The common form of this species must bear the following name: Plagiobothrys hirtus var. figuratus (Piper), comb. nov. Allo- carya figurata Piper, Contr. U. S. Nat. Herb. 22: 101 (1920). This strigose form ranging from Oregon to Vancouver Islands is common. Plagiobothrys hirtus var. corallicarpus (Piper), comb. nov. Allo- carya corallicarpa Piper, Proc. Biol. Soc. Wash. 37: 93 (1924). Plagio- bothrys Scouleri var. corallicarpus (Piper) Johnston, Contr. Arnold Arb. 3: 52 (1932). A local form of southern Oregon characterized by its deeply alveolate nutlets. Plagiobothrys calandrinioides (Phil.) Johnston, Contr. Gray Herb. 78:91 (1927). Allocarya alternifolia Brand in Fedde, Repert. 26: 169 (1929). 194 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI The type of Brand’s species has been examined. The lowermost leaves are weathered and crowded and so account for the very misleading spe- cific name. The plant is the common Patagonian P. calandrinioides. Thaumatocaryon dasyanthum var. Sellowianum (Cham.), comb. nov. Anchusa Sellowiana Chamisso, Linnaea 8: 115 (1833). Moritzia Sellowiana (Cham.) Fresenius in Martius, Fl. Bras. 8': 63 (1857). Thaumatocaryon Sellowianum (Cham.) Johnston, Contr. Gray Herb. 70: 13 (1924) and 78: 16 (1927). Moritzia dasyantha var. Sellowiana (Cham.) Brand in Fedde, Repert. 27: 148 (1929). This plant differs from typical T. das yantha only in its smaller corollas and appressed pubescence. Difficulty with connecting forms has con- vinced me that Brand might best be followed in treating Sellowianum as a mere variety. A collection of this variety from the state of Rio Janeiro, by Glaziou (no. 8731), supplies the basis for Glaziou’s astonishing report of Cyphomattia lanata in Brazil, Bull. Soc. Bot. France 57: Mém. 3: 480 (1910). I have examined the specimen at Paris so determined by Glaziou. Hackelia patens (Nutt.), comb. nov. —-Rochelia patens Nuttall, Jour. Acad. Phila. 7: 44 (1834). Lappula coerulescens Rydberg, Mem. N. Y. Bot. Gard. 1: 328 (1900). Luppula subdecumbens coeru- lescens (Rydb.) Garrett, Fl. Wasatch Reg. 78 (1911). Hackelia diffusa var. caerulescens (Rydb.) Johnston, Contr. Gray Herb. 68: 48 (1923). Hackelia caerulescens (Rydb.) Brand, Pflanzenr. [Heft 97] IV. 2522: 130 (1931). Hackelia Nelsonii Brand in Fedde, Repert. 26: 170 (1929). Lappula decumbens Nels. ex Brand, Pflanzenr. [Heft 97] IV. 252°: 126 (1931), lapsus calami. I have examined Nuttall’s type of Rochelia patens at the British Museum. The specimen was collected “near the Flat-Head River” on June 8, 1833, by N. B. Wyeth. The specimen is a good one and is evi- dently sonspeuite with Lappula coerulescens, a species also based upon material from western Montana. The species is known from western Montana and Wyoming and westward into Idaho, northern Utah and northern Nevada. Hackelia grisea (Woot. & Standl.), comb. nov. Lappula grisea Wooton & Standley, Contr, U. S. Nat. Herb. 16: 164 (1913). A readily recognizable species of New Mexico and adjacent Texas. Its relatively small corollas, with ascending lobes, quickly distinguish it among the west American annual and biennial species of this genus. Lappula echinata Gilibert, FI. Lituanica, 1: 25 (1781). Cryp- 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 195 tantha Lappula Brand in Fedde, Repert. 24: 56 (1928) and Pflanzenr. [Heft 97] IV. 252°: 147 (1931). In the Pflanzenreich Brand placed his Cryptantha Lappula among the synonyms of Lappula Redowskii (Hornem.) Greene. I believe, however, that the species belongs under L. echinata Gilib. The evident corollas and the gross aspect of the type are of that species. A microscopic study of the (immature) nutlets of C. Lappula seems to show a double row of lateral prickles. Finally the type is given as from Concepcion, Chile, a locality at which L. Redowskii is certainly not to be expected to grow naturally, though a busy port at which an aggressive weed, such as L. echinata, might be introduced without any cause for surprise. Lappula echinata is generally accepted as introduced into North America. This seems probable, though it is to be noted that the plant was collected in the New World at a very early date. A specimen in the DuBois collection at Oxford is labeled “brot from Maryland by Mr. Wm. Vernon, 1698.” Among Michaux’s collections at Paris there is one of this species labeled as “Dans ville de Montreal, 1792.” The Smith collections in London contain a specimen labeled: “North America, 1817, F. Booth.” In the British Museum there is a collection made by Douglas, during 1826, “In the valleys of the Rocky Mts.,” most likely in northeastern Washington. It seems to have been again collected in the latter region only within the past ten years, though it has been well known in the southern parts of western Canada for at least a generation. There are reasons to believe that the railroads may have much aided in the distribution north of the International Boundary. The plant has exhibited an evident, progressive increase and migration westward across the more northern of the western United States. It is now rapidly increasing in eastern Washington where it gives every evidence of being a recent immigrant. Cryptantha circumscissa (H. & A.) Johnston, Contr. Gray Herb. 68: 55 (1923). A few years ago, l. c. 81: 75 (1928), I reported this characteristic plant of western United States from near Zapala, Neuquen, Argentina. A second station in Neuquen may now be recorded. At Munich I have seen specimens labeled as collected by Erik Ammann (no. 7) at Cerro Mesa between Sept. and Nov. 1927. The new station is nearly 90 km. southeast of Zapala. Cryptantha clandestina (Trev.), comb. nov. Lithospermum clan- destinum Treviranus, Del. sem. a 1832 in hort. Bonnensi collect. p. 2 (1832-3). Cryptantha glomerata Lehmann, Del. Sem. Hamb. 1832: 4 (1832), nomen nudum; Fischer & Meyer, Ind. Sem. Hort. Petrop. 2:8 196 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XVI and 35 (1836); Johnston, Contr. Gray Herb. 78: 58 (1927). Cryp- tantha microcarpa Fischer & Meyer, Ind. Sem. Hort. Petrop. 2: 8 and 35 (1836). A study of the original description of Lithospermum clandestinum, and of old garden material representing it, has made it clear that it is that well known cleistogamic species of Chile, the two forms of which have passed as Cryptantha glomerata and C. microcarpa. Fischer & Meyer, when describing C. microcarpa, in fact, actually cited L. clan- destinum as a synonym. In the Bonn seed-list for 1832, published in Dec. 1832 or Jan. 1833, the name Lithospermum clandestinum appears in the alphabetic list on the second of the pages of that quarto catalogue. A reference leads to a footnote which reads as follows: ‘Diffusum his- pidum; fol. lanceolatis amplexicaulibus; calycibus subsessilibus ven- tricosis corollam excedentibus; seminn. granulatis. Annuum. Corolla alba, tubo ventricoso, limbo conniventi. Semina duo plerumque abortiunt. T[reviranus].” The name, L. clandestinum, appears again in the Bonn list for 1833, but not in those for 1834 or 1835. Cryptantha glomerata Lehm. is the type-species of Cryptantha. Re- cently I had the privilege of consulting the extensive collections of old seed-catalogues at Berlin and Geneva. I now find it possible to record several important references in the history of that genus and species which were either unknown or unavailable to me at the time of my work on the group. The first mention of Cryptantha glomerata Lehm. and of the generic name appears in Lehmann’s seed-list of the Hamburg Garden for the year 1832. The binomial appears as a mere name on page 4, thus: “Cryptantha glomerata Lehm.” No description or ex- planation of the name is given! The list is dated 1832 and was prob- ably published, as was customary with such lists, around the close of the year. No mention of the binomial is found in the Hamburg lists for 1830, 1831 or for 1833 or 1834. In 1835, p. 4, again without descrip- tion, appears: “Cryptantha glomerata Lehm. (Del. Sem. 1832). In 1836, p. 4, the following two names appear bare of description: “Cryp- tantha glomerata Lehm.” and “Cryptantha microcarpa F. & M.” These are repeated in the list for 1837, p. 4. In the list for 1838, p. 4, there is merely the name, “Cryptantha microcarpa F. & M.” Fischer & Meyer, in their St. Petersburg seed-list for 1835, supplied the first descriptions of Cryptantha glomerata Lehm. and C: microcarpa F. & M. This Rus- sian list bears a censor’s date, Dec. 25, 1835, the equivalent of Jan. 5, 1836 of our present calendar. There is no mention of Cryptantha in the St. Petersburg list for 1834! Fischer & Meyer, when publishing and describing “C. glomerata Lehm.” in their list for 1835, attribute the 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 197 name to “Bernhardi in litt.” A study of Bernhardi’s seed-lists, Sel. sem. hort. Erfurt., shows that the name “C. glomerata Lehm.” appears as a mere binomial in those for 1833 (Jan. 18, 1834), 1834 (Feb. 24, 1835) and 1835. There is no mention of Cryptantha in the Erfurt list for 1832! From the facts I have given it becomes evident that Cryptantha glomerata was in cultivation at Bonn and Hamburg in 1832. Treviranus immediately described the Bonn cultures as Lithospermum clandes- tinum. Lehmann applied to his Hamburg cultures the name Cryptantha glomerata, but did not describe it, that being done for him three years later by Fischer & Meyer who based their description on plants grown at St. Petersburg. There is no information as to the channels by which the species was introduced into cultivation. I suspect, however, that the original seed may have been obtained by Bertero, who col- lected the plant near the Rio Quillota, Chile, as early as 1828, and that seeds from this source may have been distributed from Turin. Amsinckia intermedia Fischer & Meyer, Ind. Sem. Hort. Petrop. 2: 2 and 26 (1836). This name appears bare in the alphabetic list on page 2 of the seed- list cited above. On page 26 (p. 1 of reprint) the following description is found, “A. INTERMEDIA, A. corolla fauce glabra nuda, limbo tubo sub- breviore; staminibus ad faucem insertis.— Corollae tubus 1% lin. longus, limbus fere 3 lin. in diametro, saturate aurantiacus maculisque 5 saturatioribus pictus. — Species intermedia A. /ycopsioidem inter et A. spectabilem; a priore dignoscitur insertione staminum, a posteriore corollis longe minoribus et praesertim corollae tubo non (ut in illa) ad faucem plicis intrusis semiclausa. — Hab. cum sequente specie [A. spec- tabilis| circa coloniam ruthenorum Ross in portu Bodega Novae Cali- forniae. Annua.”’ The seed-list in which this description occurs bears the printed censor’s date, Dec. 25, 1835. This equals Jan. 5, 1836 of the present calendar. Through the kindness of Prof. B. A. Keller, Director of the Institute and Botanic Garden at Leningrad, I have received authentic material of Amsinckia intermedia. This consists of an authentic fragment of the species, from the herbarium of Meyer, one of the co-authors of the species, and a fine specimen from the plantings in the St. Petersburg Garden in 1836. The specimens agree with the interpretation of A, intermedia given by Suksdorf, Werdenda 1: 88 (1931). The plant is a member of that variable and bewildering island species that Mac- bride, Contr. Gray Herb. 49: 12 (1917), and Jepson, Man. FI. Pl. Calif. 844 (1925), have incorrectly called “A. Douglasiana.”’ Greene, Bot. S. 198 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Francisco Bay, 262 (1894), and Jepson, Fl. W. Mid. Calif. ed. 2, 350 (1911), earlier treated it, partly as A. intermedia and partly as A. spec- tabilis, The name, A. intermedia F. & M., is properly applied to the polymorphous species which is common in California in the interior valleys and on hillsides back from the immediate vicinity of the ocean. Amsinckia spectabilis Fischer & Meyer, Ind. Sem. Hort. Petrop. 2: 2 and 26 (1836). This species appears on page 2 of the above publication as a bare name in an alphabetic list of seeds. On page 26 (p. 1 of reprint) the following description is found: ‘A. spEcTABILIS. A. corolla fauce glabra plicis intrusis semiclausa, limbo longitudine tubi; staminibus ad faucem insertis. Species pulchritudine florum insignis atque distinctissima. Corolla aurea, limbo 6 lin. in diametro, ad faucem plicis 5, squamulas simulantibus, aucta. Annua.” In 1925 through the kindness of Prof. Boris Fedtchenko, with the assistance of Miss Olga Enden, I received two generous fragments of authentic specimens of this species. The specimens were grown in the St. Petersburg botanic garden in 1835-36. They are given as grown from seeds collected at Fort Ross, California, by Wiedemann. These fragments were examined by Suksdorf, Werdenda 1: 96 (1931). He correctly identified them with the coastal plant that Brand, in Fedde’s Repert. 20: 319 (1924), has described as A. nigricans. Brand’s plant, Heller 5614, is from the type-locality of A. spectabilis. I have seen many specimens of this plant in various herbaria from numerous garden- cultures. While evidently conspecific, these specimens rarely have the corollas as well developed as that found in the original culture at St. Petersburg in 1835. This is not surprising. I have grown Amsinckia in a botanic garden and under glass and have in most cases discovered remarkable differences in habit of growth and corolla-size between my cultures and the wild specimen from which the seed was obtained. Macbride, Contr. Gray Herb. 49: 7 (1917), in his monograph of the genus, has treated the coastal plant (the true 4. spectabilis F. & M.) under the name “A. intermedia.” Jepson, Man. FI. Pl. Calif. 844 (1925), attempted to follow him and has described the coastal plant as “A. inter- media.” His illustration, however, is the inland species, which just happens to be the true A. intermedia F. & M. Previous to Macbride’s paper in the writings of Gray, of Greene, and of Jepson, the coastal plant appears as “A. lycopsoides.” Macbride, |. c. 5, of course, was quite incorrect in applying the name “A. spectabilis” to the smooth-fruited A. grandiflora Kleeb ex Gray. The name A. spectabilis F. & M. properly 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 199 belongs to the strictly coastal plant of California that has small dark nutlets, acute more or less denticulate leaves, and a pair of the calyx- lobes frequently more or less united. Abrams, Fl. Los Angeles, 335 (1904), seems to have been the only author who has properly applied the names A. spectabilis and A. intermedia. Amsinckia lycopsoides Lehmann, Del. Sem. Hort. Hamburg 1831: 1 and 7 (1831). On the first page of the Hamburg seed-list for 1831 appears the name “Amsinckia lycopsoides Lehm.'” The exponent refers to a note on page 7 where the following is found, ‘“‘'Genus novum e familia Borraginearum, praeter alias notas cotyledonibus 4 distinctissimum. Benthamia Lindl. in litteris (non Richard Monog. des Orchidees iles de France et de Bour- bon pg. 43, t. 7, fig. 2). In the seed-list for 1833, p. 3, and 1834, p. 3, the binomial appears perfectly bare. In 1835, p. 3, it is listed in com- pany with A. angustifolia Lehm. In 1836, p. 3, and 1837, p. 3, it is listed as one of four species, A. angustifolia, A. intermedia, A. lycopsoides and A, spectabilis. In the writings of Fischer & Meyer the binomial, A. lycopsioides Lehm., appears as a bare name in company of A. angustifolia Lehm., in the first St. Petersburg list, Ind. Sem. Hort. Petrop. 1: 2 (1835). In the next list, 2: 2 (Jan. 1836), it appears with A. angustifolia, A. inter- media and A. spectabilis and on page 26 (p. 1 of reprint) has the follow- ing note concerning it: “AMSINCKIA LycopsiomDEs. A. corolla fauce barbata, limbo tubo triplo breviore; staminibus corollae tubo paulo supra basin insertis. — A. lycopsioides Lehm. delect. sem. h. Hamburg. 1831.— Tubus corollae 3% lin. longus; limbus 2 lin. in diametro, vix latior.” The species, Amsinckia lycopsoides Lehmann, is the type of the genus Amsinckia. The Hamburg seed-list in which it was first published is dated 1831. That it was actually published that year is proved by the review of this publication in the Litteratur-Bericht zur Linnaea (vol. 6) which bears the title-page date of 1831. The description of the species, Amsinckia lycopsoides Lehm., by Fischer & Meyer, appears in a seed- list for the year 1835 but this pamphlet bears a printed censor’s date, Dec. 25, 1835 which is the equivalent of Jan. 5, 1836 in our present calendar. It is to be noted that when, in 1831, Lehmann published his generic name, Amsinckia, that he cefinitely associates it with Benthamia of Lindley. This latter veneric name was published by Lindley, in the same year, but only as a nomen nudum, Lindley, Nat. Syst. 241 (1831). 200 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XvI It was undoubtedly based upon material collected by Douglas along the Columbia River. This is clearly indicated by specimens in herbaria at Cambridge, Kew, London and Geneva. In the Lindley Herbarium at Cambridge there is only one sheet that has been determined as Benthamia by Lindley. This contains Cuming’s no. 512 from Valparaiso and a specimen labeled “North West Amer. H. H. G. 1827, Douglas.” Lindley has written in the corner of the sheet “Benthamia lycopsoides Mihi.” There are various strong reasons for believing that this sheet in Lind- ley’s herbarium formerly bore only the material from Douglas and that the Cuming material was later added to it, probably after Lindley’s annotation. At Kew there are two significant specimens. One from the Bentham Herbarium is labeled “Benthamia lycopsoides Lindl. M. S. sem. ex Amer. occid. ex Douglas, Hort. Soc. Hort. London, 6-6-28.” A similar sheet from the Hooker Herbarium is labeled “Anchusa, fl. yellow, Benthamia Lindl. mss. N. W. Am. Douglas, cult.” Lindley published only the genus name, Benthamia. The binomial “Benthamia lycop- soides” seems to have been published first by DeCandolle, Prodr. 10: 118 (1846). This reference is clearly based upon a specimen at Geneva bearing the following data: “Benthamia lycopsoides Lindl. ined., Hort. Sociét. horticult. in Chiswick 6 jun. 1828.” The name on the label is in the script of Lindley. The source is written by DeCandolle. The date given is the same as that found on the sheet in Bentham’s herbarium and falls within the period when A. DeCandolle visited London for work on his Campanulaceae. The herbarium of the Horticultural Society was sold to the British Museum. There is a specimen from this source at South Kensington labeled: “sandy plains of the Columbia, 1825 (accord- ing to Lindley a new genus).” Lindley was in charge of identifying the plants grown in the gardens of the Horticultural Society at Chiswick. All the specimens mentioned are probably from seeds grown at Chis- wick. They all represent the plant recently described as A. simplex Suksdorf, Werdenda 1: 33 and 53 (1927 and 1931). There are a number of good reasons for believing that Lehmann’s genus Amsinckia and his species A. lycopsoides are based upon Lind- ley’s genus Benthamia and B. lycopsoides. In the first place shortly before 1830 Lehmann travelled in England and met various botanists there. He was a well known student of the Boraginaceae. Lehmann, in any Case, was later in correspondence with Lindley, for he cites his authority for Benthamia as “Lindl. in litteris,” and we may well believe that he received seed or specimens of Douglas’s curious borage from Lindley and grew it in the Hamburg garden. Lehmann devotes about half of his short description of Amsinckia to citing Lindley’s unpub- 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 201 lished Benthamia and its earlier published homonym. The specific name used by Lehmann is that selected by Lindley. What is most im- portant, however, is that an Amsinckia conspecific with Douglas’s plant was in cultivation in various European botanic gardens under the name “Amsinckia lycopsoides.” There is a specimen at Kew collected by J. Gay in the Jardin des Plantes at Paris in June 1833, only a year and a half after Lehmann published Amsinckia. This plant was grown under Lehmann’s binomial and represents the species collected on the Columbia by Douglas. In conclusion it may be noted that the short descriptive notes concerning A. lycopsoides, given in 1835 by Fischer & Meyer, apply to the plant collected by Douglas. A study of Douglas’s Journal, p. 116 (1914), fortunately reveals some information as to the original source of Amsinckia lycopsoides. The plant is evidently that mentioned under the date of May 2, 1825, in an enumeration of collections made on ‘‘Menzies Island, in the Columbia river, opposite the Hudson Bay Company’s establishment at Point Van- couver.” According to Piper, Contr. U. S. Nat. Herb. 11: 620 (1906), Menzies Island is that now known as ‘“‘Haydens Island.” The notes by Douglas are as follows: “(151) Myosotis sp., annual; hirsute, branch- ing; leaves long, entire; linear-lanceolate; flowers bright yellow; tube long; mouth of the corolla spreading, with a dark spot opposite teeth; seeds not yet known; this very interesting species was found on Menzies Island in company with Mr. Scouler, who agreed with me to call it Myosotis Hookeri [not Myosotis Hookeri Clarke (1883)]| after Dr. Hooker of Glasgow; scarce, only three specimens of it were found, two of which are in my possession. — I have since found it in abundance near all the Indian lodges above the Rapids of the Columbia. S|eeds].” From these notes it is evident that seeds were not obtained on Menzies Island and that, later, they were obtained somewhere above the Colum- bia Rapids. Amsinckia simplex Suksd. is known only from the general vicinity of Portland, Oregon (just south of Menzies Island). It is scarcely separable from A. arenaria Suksd. which is reported from the Columbia Gorge and in eastern Washington. The name Amsincktia lycopsoides (Lindley) Lehmann is properly applicable to these concepts. It has been shown that Amsinckia lycopsoides Lehm. is based eventu- ally upon material collected by Douglas along the Columbia River. In subsequent paragraphs I have shown that Lithospermum lycopsoides Lehm. (1830) is based upon collections made by Scouler on the north- western coast of Washington. In the writings of A. DeCandolle, Prodr. 10: 118, adnot. (1846), Gray, Synop. Fl. 2: 198 (1878), Macbride, 202 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI Contr. Gray Herb. 49: 7 (1917), Suksdorf, Werdenda 1: 101 (1931), etc., the binomial Amsinckia lycopsoides has been considered as merely a nomenclatorial transfer and as based upon Lithospermum lycopsoides. The similarity of the specific epithet is a mere coincidence. There are no reasons at all for supposing that these two species are identical. I have shown that Amsinckia lycopsoides is a plant from along the Colum- bia. Lithospermum lycopsoides is an earlier binomial, but since the specific name is preoccupied under Amsinckia it can not be legitimately transferred to that genus. A new name for the coastal plant of north- western Washington is accordingly needed. Lithospermum lycopsoides Lehmann, Pugil. 2: 28 (1830); Leh- mann in Hooker, Fl. Bor. Am. 2: 89 (1838). As was his custom in the Pugillus, Lehmann cited no specimens when he originally described L. /ycopsoides. In the Flora Boreali-Americana, in which he contributed the Boraginaceae, however, he repeated his original description verbatim and cited the basic specimen. This latter is given as ‘Straits of de Fuca, N. W. America, Dr. Scouler.” At Kew, from the herbarium of Hooker, there is a specimen that agrees perfectly with Lehmann’s description and is labelled “Lith. lycopsioides Lehm. De Fuca, N. W. Am. Scouler.” I agree with E. L. Greene, who has written on this sheet that ‘‘This, along with fragments in Herb. Benth. constitutes the type of Lithospermum lycopsoides Lehm. It has never been in cultivation.” The plant is undoubtedly conspecific with that of northwestern Washington and adjacent Vancouver Island which has passed as “Amsinckia lycopsoides” in Piper’s Flora of Washington, Contr. U. S. Nat. Herb. 11: 480 (1906), and in the monographs by Macbrice, Contr. Gray Herb. 49: 7 (1917) and Suksdorf, Werdenda 1: 101 (1931). It is not the same species as Amsinckia lycopsoides Lehm., which is based upon specimens collected by Douglas near the Columbia. The present plant, a coastal species related to true A. spec- tabilis F. & M. of California, strangely has no synonyms. Since the specific name is preoccupied under Amsinckia a new name is needed. The plant may be called: Amsinckia Scouleri, nom. nov. Lithospermum lycopsoides Leh- mann, Pugil. 2: 28 (1830) not A. lycopsoides Lehmann (1831). Amsinckia Douglasiana A. DeCandolle, Prodr. 10: 118 (1846). I have examined the type of this species in the DeCandollean Her- barium at Geneva. It is clearly a species with tessellate nutlets and large showy corollas. I consider it conspecific with A. Lemmonti Mac- 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 203 bride, Contr. Gray Herb. 48: 50 (1916). Suksdorf, Werdenda, 1: 102 (1931), who has examined authentic material of A. Douglasiana, pre- served at the Gray Herbarium, has considered it closely related to A. Lemmoniu but separable from it. He places these two species together in his monograph. Gray erroneously cited the name A. Douglasiana in the synonymy of the common inland species of California. Not having seen the type of A. Douglasiana, Macbride, Contr. Gray Herb. 49: 12 (1917), was misled by Gray’s erroneous citation and applied it to the common inland species of California. The plant treated as A. Doug- lasiana by Macbride, and by Jepson, Man. Fl. Pl. Calif. 844 (1925), who followed him, is properly identified as true A. intermedia F. & M. Amsinckia Douglasiana A. DC. is a relatively rare plant of the South Coast Ranges of California and was probably originally collected by Douglas in San Luis Obispo or southern Monterey counties during his journey from Monterey to Santa Barbara and return. Amsinckia parviflora Bernhardi, Selec. Sem. Hort. Erfurt. 1833: 1 and 4 (Jan. 1834). On the first page of the Erfurt seed-list for 1833 two Amsinckias appear in the alphabetic list of names, Amsinckia lycopsoides Lehm. and A, parviflora Bernh. A reference to the last, fourth but unnumbered page of the seed-list gives the following note concerning A. parviflora Bernh.: ‘‘(1) Lithospermum calycinum Moris, cui cotyledones 4, s. potius 2 bipartitae, speciem Amsinckiae sistit, quam A. parvifloram vocarem. An A. angustifolia Lehm. eodem planta ?”” The list bears a printed date, Jan. 18, 1834. No species of Amsinckia are mentioned in the Erfurt seed-lists for the year 1832. Amsinckia parviflora Bernh. appears to be no more than a mere renaming of Lithospermum calycinum Moris. The two names are, accordingly, exact synonyms and apply to Bertero’s plant from Rancagua, Chile, described and figured by Moris, Mem. Accad. Toririo 37: 98 tab. 22 (1834). In 1834 Lehmann cited Bernhardi’s binomial as a synonym of A. angustifolia Lehm. I am inclined to believe this is correct, for as I shall discuss, I suspect that Lehmann’s species is also based upon Chilean material. Amsinckia angustifolia Lehmann, Del. Sem. Hort. Hamburg 1832: 3 (1832), nomen; Fischer & Meyer, Ind. Sem. Hort. Petrop. 2: 26 (1836), description. The above binomial appeared as a bare name in the seed-list of the Hamburg garden for 1832. It appeared again as a bare name in the list for 1833, p. 3, was omitted in that for 1834, and in the list for 1835, p. 3, was cited as follows: ‘“Amsinckia angustifolia Lehm. (A. parvifolia 204 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Bernh. Sel. sem. h. Erf. 1833). This reference was repeated in 1836, p. 3. In the list for 1837, p. 3, it again appears as a bare name. In the first list from the garden at St. Petersburg, 1: 2 (1835) the name Amsinckia angustifolia also appears bare. In the next list from St. Petersburg, 2: 2 and 26 (1836), the name appears in the list of seeds and on page 26 (p. 1 of reprint) the following description and refer- ences are published: ‘‘A. ancustiFotia. A. corolla fauce glabra nuda, limbo tubo duplo breviore; staminibus ad faucem insertis. A. angusti- folia Lehm, delect. sem. h. Hamburg. 1832. A. parviflora Bernhardi select. sem. h. Erfurt. 1833. Lithospermum calycinum Moris, Enum. sem, h. r. bot. Taurinens, 1831 et in Mem. della Acad. d. Scienze di Torino Tom, XXXVII, p. 108. tab. XXII. — Corollae tubus vix 2 lin. longus, limbus vix 2 lin. in diametro. — A praecedente | A. lycopsioides Lehm.] floribus parvulis et praesertim staminum insertione diversis- sima.” All the references cited by Fischer & Meyer trace back to material, collected by Bertero in central Chile. What is more all the garden material, under the name A. angustifolia, seems best referred to the Chilean forms of the genus. All authors have applied A. angustifolia to the austral plant. There seems every reason for continuing to do so. I suspect that the cultures in European gardens were originally from seeds obtained by Bertero at Quillota or Rancagua, Chile, and subse- quently distributed from Turin by Morris or Colla, Omphalodes erecta, sp. nov., herba perennis erecta e caudice laxe ramoso oriens pilis mollibus gracillimis subcinerea; caulibus foliosis simplicibus vel supra medium sparsissime fertiliterque stricto-ramosis 3-6 dm. altis partibus maturis plus minusve glabrescentibus brunnescen- tibus 2-4 mm. crassis; foliis lanceolatis vel late lanceolatis 5-11 cm. longis 15-30 mm. latis (superioribus non-conspicue reductis) sub medium apicem versus in acuminem 1-3 mm. longam gracilem gradatim attenuatis, margine integerrimis basi angulatis vel subrotundis 3—6 mm. longe petiolatis, supra viridis sparse inconspicueque pubescentibus non rariter minute pustulatis, subtus pallidis saepe pilis abundantibus longioribus subcinereis; inflorescentia gracili laxe racemosa simplice vel basaliter furcata ebracteata 5-15 cm. longa 0-1 cm. longe pedunculata; pedicellis ad anthesin 3—6 mm. longis ascendentibus, fructiferis ad 2 cm. longis saepe decurvatis vel subcontortis; calyx ad anthesin pallide dense- que strigoso, lobis 5 inaequalibus lanceolatis ca. 4 mm. longis; corolla coerulea vel medium versus violacea, tubo ca. 3 mm. longo, appendiculis faueium 5 trapeziformibus ca. 1.3 mm. longis et latis apice subemargina- tis, margine pubescentibus, limbo ca. 13 mm. diametro patente ultra 1935] JOHNSTON, STUDIES IN BORAGINACEAE, XI 205 medium lobato, lobis 4-5 mm. longis rotundis, sinibus loborum incon- spicue plicatulis; antheris oblongis inclusis medium versus bubo corollae affixis; filamentis perbrevis; stylo ad anthesin 2 mm. longo, fructifero conspicuis 9 mm. longo; stigmato disciformi; fructu 4-ovulato; nuculo solitario (3 abortis) minute appresseque strigoso depresse lateque ovoideo, (cum alis) ca. 8 mm. diametro, margine evidenter 1—1.3 mm. late alato, ala plana patenti leviter denticulata, dorso nuculae convexo. Mexico: common in dense oak-wood along an arroyo near Santa Ana, between Alamar and Taray, Sierra Madre Oriental, ca. 25 km. s. w. of Galeana, Nuevo Leon, corolla blue with a light violet center, July 3, 1934, .H.& M. T. Mueller 992 (type, Gray Herb.) ; scattered in fields and wate places in canyon above Alamar, Sierra Madre Oriental, 1500-1800 m. alt., June 2, 1934, C. H. & M. T. Mueller 680 (G). A remarkable species differing from all its congeners in its coarse erect habit of growth. The general habit and appearance of the plant, indeed, is more suggestive of Cynoglossum than of Omphalodes. From the American species of its genus it is further distinguished by its large solitary nutlets which possess a weakly denticulate and spreading wing, rather than a strongly toothed upcurved one. The foliage of O. erecta is very distinctive. All the American species of the genus have long- petioled more or less cordate leaves. The new species has them very short-petioled and lanceolate. Only one Mexican species, O. aliena, has a similar bractless inflorescence. The plant is a remarkable addition to the list of Mexican Boraginaceae. HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 206 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XVI LORANTHACEAE COLLECTED IN THE SOLOMON ISLANDS BY L. J. BRASS AND S. F. KAJEWSKI, ON THE ARNOLD ARBORETUM EXPEDITION, 1930-1932 B. H. DANSER With plate 129 Amylotheca sp. San Crist6dval Island: Waimamura, Brass 2849, Sept. 9, 1932, “common, parasitic on rain forest trees, stout shrub, branches smooth and glaucous, leaves very thick and fleshy, perianth of unopened flowers lower half pale red, upper greenish-yellow.” Indeterminable for lack of open flowers, but closely allied to Amylo- theca Versteegii (Lauterb.) Danser from New Guinea and New Ire- land, differing, however, by longer-pedicelled lateral flowers of the triads, more thickish inflorescences and flowers, obtuse bracts and more distinct calyx lobes. Dactyliophora salomonia, n. sp. Plate 129 a-b Glabra, inflorescentiis floribusque iuventute forte parce tomentellis exceptis. Rami robusti; internodia foliifera teretia, iuventute apicem versus applanata nonnunquam ancipita, nodis dilatatis, postea teretia, 3-5 mm. crassa, nodis ad sesquiplo crassioribus. Folia opposita vel subopposita; petioli 14-24 mm. longi, basi teretes, laminam versus facie superiore profundius canaliculati; laminae triangulari-ovatae, plerumque 10-15 cm. longae, 5.5—7.5 cm, latae, basi rotundatae vel leviter cordatae, in petiolum contractae, apicem obtusum versus gradatim attenuatae, crassiusculae, fragiles, penninerves, costa facie inferiore basin versus magis prominente, costa cetera et nervis lateralibus primariis valde incurvatis utrinque distinctis paulum prominentibus. Inflorescentiae singulae in axillis foliorum et plures circum nodos defoliatos; pedunculi teretes, apice basique paulum incrassati, 15-20 mm. longi, 1—1.25 mm. crassi; axes ex internodiis 2 vel 1 compositi, quorum inferius 1-5 mm. longum, superius brevissimum; nodi deinceps circiter 8, 6, paucas triades ferentes; pedicelli triadum inferiores 8-10 mm. longi, c. 0.3 mm. crassi, Superiores paucis mm. breviores; pedicelli florum lateralium 2-3 mm. longi; bracteae bracteolaeque ovatae obtusae 1.25—-1.5 mm. longae. Calycis tubus obovato-campanulatus, 3.5-4 mm. longus, 1.5—2 mm. latus, limbus brevissimus erectus vel nonnihil inflexus. Corolla statu 1935] DANSER, LORANTHACEAE IN THE SOLOMON ISLANDS 207 alabastri adulti 25 mm. longa, tertia parte inferiore inflata ad 3 mm. lata, tertia parte media gradatim ad 1 mm. attenuata, tertia parte superiore in clavam 6-angulam obtusam 1.5—2 mm. crassam incrassata, postea divisa in petala 6 sublinearia, a basi c. 1 mm. lata in duabus tertiis inferioribus gradatim ad 0.5 mm. angustata, in tertia parte Superiore angustissime spathulata, apice acutiuscula crassiuscula, latere interiore c. 2 mm. supra basin squamula minima. Filamentorum pars libera c. 2 mm. longa; antherae c. 4 mm. longae, obtusae. Stylus a basi c. 0.6 mm. crassa gradatim angustatus, sub stigmate c. 0.25 mm. crassus; stigma obovatum, styli apice vix crassius. Fructus obovato-ellipsoides, ad 9 mm. longi, 6 mm. crassi, superea collo solido 2 mm. longo et lato coronati, calycis rudimento vix ullo. Guadalcanal Island: Vulolo, Tutuve Mt., 1200 m. alt., Kajewski 2497, April 14, 1931; ‘common, a loranthus growing on trees; there are ‘two different coloured flowers, yellow and yellow pink, but there is no specific difference; fruit green when ripe, length 1.1 cm., diameter 6 mm., with a white flesh inside; fruit with a blunt point at end; the leaves are heated and rubbed on sore legs.” — Vernacular name “Bitorchi.” The genus Dactyliophora was known, before now, only from New Guinea and New Ireland. Dactyliophora salomonia is closely allied to the New Guinea D. verticillata (Scheffer) Van Tieghem, and perhaps only a variety of it, but it differs by less cordate, longer-petioled leaves, axillary inflorescences, longer and less thick peduncles and pedicels, more slender calyces and shorter corollas. The peculiar beak on the fruit is not described for any other species, but as in most Dactyliophorae the fruit are unknown, and those of D. Novae-Guineae (Bailey) Danser are figured with a similar prolongation (cfr. Bull. Jard. Bot. Buitenz., sér. 3, 11: 359, fig. 14, h), it probably is not a characteristic of D. salomonia only. Sogerianthe versicolor, n. sp. Plate 129c-e Omnis glabra (vel pedicellis, bracteis, calycibusque minute puberulis). Rami graciles, parce dichotome ramosi, internodiis foliiferis 1-2.5 (—3) mm. crassis, plerumque 3—9 cm. longis, nodis valde incrassatis, duplo vel fere triplo crassioribus. Folia opposita; petioli 5-10 mm. longi, supra applanati praesertim laminam versus, subtus rotundati; laminae oblongo- ovatae, plerumque 6-10 cm. longae, 2.5—4.5 cm. latae, sub basi rotundata in petiolum contractae vel magis attenuatae, apicem obtusum versus magis minusve acuminatae, crassiusculae, fragiles, utrinque opacae (vel facie superiore lucidulae), penninerves, costa facie inferiore prope basin valde prominente, ceterum costa nervisque lateralibus primariis incurva- 208 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI tis utraque facie visibilibus vix prominentibus. Flores singuli vel in umbellis paucifloris omnino sessilibus in scrobiculis corticis inserti, pauci in axillis foliorum vel plures circum nodos foliatos et defoliatos; pedi- celli teretes, basi saepe paulum clavati, 2-4 mm. longi, 0.3(—0.5) mm. crassi; bracteae bracteolaeque triangulares vel ovatae, 0.5-1 mm. plerumque 0.75 mm. longae, obtusae vel acutae, basi nonnihil connatae. Calycis tubus campanulatus, subcylindricus, (1.5—)2—2.5 mm. longus, 1.25 mm. latus, limbus erectus paulum cupuliformis, 1(-1.25) mm. longus, ore (1.75—)2 mm. lato, integerrimus (vel irregulariter incisus). Corolla sympetala, statu alabastri adulti (24-)30-31 mm. longa, in duabus tertiis inferioribus fusiformiter inflata ad 3 mm. lata, in tertia parte superiore subcylindrica, c. 1.5 mm. lata, apice obtusa, postea divisa usque ad dimidiam longitudinem in lacinias 6 lineares reflexas et volutas apice crassiusculas acutiusculas, facie interiore prope basin squamulis nullis. Filamentorum pars libera 3.5-4 mm. longa; anthera basifixa, 3.5 mm. longa, obtusa. Stylus corollae aequilongus (vel paulo longior), a basi ad apicem aequicrassus; stigma globosum, styli apice circiter sesquiplo crassius. Fructus ellipsoides, ad 8 mm. longus, 4 mm. diametro, calyce integro erecto paulum aucto coronatus. San Crist6val Island: Waimamura, lowlands, on rain forest trees, Brass 2676 (type), Aug. 18, 1932, “plentiful, leaves very pale, corolla-tube white, segments pink, filaments pink, style green.” Malaita Island: Quoimonapu, 200 m. alt., rain forest, Kajewski 2355, Dec. 12, 1930, “common, a large loranthus growing on the rain forest trees, base of corolla pink, ends of petals white cream.” Ysabel Island: Tiratona, 600 m. alt., Brass 3227, Nov. 26, 1932, “common, leaves stiff, margins incurved; perianth reddish, with brown lobes.’”’— Vernacular names “Oong” (under no. 2355), and “Buraronu” (under no. 3227). Description after the type Brass 2676, the dimensions between brackets after Kajewski 2355, which mainly differs by shorter corollas 24-26 mm. long, and puberulous pedicels, bracts and calyces. The number Brass 3227 is much like the type, but the leaves are much smaller, including the petioles 2—5.5 cm. long and 1—1.6 cm. broad. The new species does not show the articulation in the pedicel nor the scales at the inside of the corolla tube considered characteristic for the genus Sogerianthe till now. As, however, it has a 6-merous long-tubed sympetalous corolla and 3 bracts at the base of the flower, and moreover agrees with the species already known in general appearance, I do not hesitate to place it in the same genus with S. sogerensis (S. Moore) Danser and S. sessiliflora Danser. ———_—S A ee : NaN A as ee | ANS SS IN : . . a _ : 7 oe _ ee oo On re 7 1935] DANSER, LORANTHACEAE IN THE SOLOMON ISLANDS 209 Dendrophthoé falcata (Linn. fil.) Ettingshausen in Denkschr. Akad. Wiss. Wien, Math.-Naturwiss. Cl. 32: 53 (1872). — Danser in Bull. Jard. Bot. Buitenz. sér. 3, 11: 403 (1931). Loranthus falcatus Linn. fil., Suppl. 211 (1781). Guadalcanal Island: Berande River, sea level, Kajew- ski 2415, Jan. 6, 1931, “common, a loranthus growing on rain forest trees, petals green-cream with orange edges, very showy, the largest fruit on specimens, the natives say are pretty full growth, length 1 cm., diameter 4 mm.; the natives use this plant superstitiously to stop rain by placing twigs upright in the ground.” — Vernacular name “Ti-nu- issi.”” Distribution: from tropical southeastern Asia all over the Malay Archipelago to tropical Australia, but before now not collected farther eastward than the Bismarck Archipelago. For the very numerous synonyms cfr. Verh. Kon. Akad. Wetensch. Amsterdam, Afd. Natuurk., sect. 2, 29, 6: 44 (1933). Notothixos leiophyllus K. Schumann in Schumann & Lauterbach, Nachtr. Fl. Deutsch. Schutzgeb. Siidsee, 260 (1905).— Danser in Bull. Jard. Bot. Buitenzorg, sér. 3, 11: 456 (1931). Ysabel Island: Tataba, 50 m. alt., parasitic on kranches of tall rain forest trees, Brass 3432, Jan. 4, 1933; “plentiful, small much branched shrub, leaves brittle, underside glaucous in old leaves, indu- mentum golden yellow.” Distribution: Philippine Islands, eastern part of the Malay Archipelago, Queensland, but before now not farther eastward than New Britain EXPLANATION OF PLATE 129 Fig. a and b: Dactyliophora salomonia (type, na ski 2497); a, twig with leaves and inflorescences in bud, X %; b, fruit, * 1, Fig. c-e: wires ‘ersicolor (type, Brass 2676); c, ae with flowers in bud, xX 44;d, flower, X 1; e, fruit, x 1 GRONINGEN UNIVERSITY, GRONINGEN, HOoLLaAnp. 210 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI CHROMOSOME NUMBERS IN THE HAMAMELIDACEAE AND THEIR PHYLOGENETIC SIGNIFICANCE EpGAR ANDERSON AND KARL SAX With three text figures ReEINScH! introduced his morphological survey of the Hamamelida- ceae by the observation that they form one of those natural families as to whose precise delimitation and relationships there has been much difference of opinion. The forty years which have elapsed since the appearance of his paper have merely provided further illustrations of the justice of his remarks. Though universally conceded to be a natural group, the only general agreement as to its phylogenetic position seems to be the opinion that it occupies an important one. Because of this uniformly recognized phylogenetic significance an effort has been made to make as complete a survey of chromosome numbers as possible. The living collections of the Arnold Arboretum fortunately include several genera such as Sinowilsonia and Parrotiopsis which are very rare in cultivation but the work has been hindered by the very great technical difficulties involved. The chromosomes are small, there is much sec- ondary pairing, the cytoplasm is murky and the chromosomes do not stain sharply. In most of these details the family shows cytologically a strong resemblance to the Rosaceae, paralleling the morphological resemblances which have been commented on by most students of the group. The following chromosome counts have been made. The genera are arranged according to the classification of Harms in Engler and Prantl. In each case the counts were obtained from aceto-carmine smears. Typi- cal meiotic plates are illustrated in Figure 1. SuUB-FAMILY HAMAMELIDOIDEAE CHROMOSOME NUMBER Tribe 1 Hamamelis vernalis 12 Tribe 3 Corylopsis pauciflora 12 Corylopsis spicata 36 Corylopsis Veitchiana 36 Tribe 4 Parrotiopsis Jacquemontiana LZ Fothergilla major 36 Fothergilla monticola 24 Tribe 5 Sinowilsonia Henryi 12 SuB-FAMILY LIQUIDAMBAROIDEAE Liquidambar Styraciflua is 1Engler in Bot. Jahrb. 11: 347 (1890). 1935] ANDERSON AND SAX, CHROMOSOMES IN HAMAMELIDACEAE 211 Meiotic irregularities, accompanied by a high percentage of pollen sterility were encountered in Liguidambar Styraciflua. This is some- what puzzling since this species exhibits none of the morphological peculiarities which are usually associated with irregular meioses. It is a “‘good”’ species with no closely related forms occurring within the same area. Its behavior is more probably to be explained as due to climatic influences. It is a southern species and at the Arboretum is being culti- @ * @ ® %e - = e 8 me : - y bg 4 é ¢ e ee te 8 A @ e.**. @ & a A <7 e® e = e e 1 2 3 @% 0% a * 3°e @ : a "3 @ bg ® pew @ 3 ® g? & i) ee a e %e, 4 6 q © ee Ficure 1. Camera a drawings (X 3000) of pollen mother cells: 1. Liquidambar Styrac ua. — 2. Parrotiopsis Jacquemontiana. — 3. Fothergilla major. — 4. Corylopsis pauciflora. — 5. Sinowilsonia ryi. — 6. Hamamelis vernalis. vated somewhat north of its natural range. Whitaker’ has demon- strated the effect of abnormal temperatures upon meiosis in Cypho- mandra. It is possible that the irregular chromosome behavior and con- sequent pollen sterility of Liguidambar Styraciflua at the Arnold Arbo- retum may have a similar explanation. It would be interesting to know if L. Styraciflua is characterized by low percentages of fertile pollen in its native home. The cytological studies present a number of facts of taxonomic sig- nificance. 1. The Hamamelidoideae are a coherent group with a com- 1Jour. Arnold Arb. 15: 113-117 (1933). 212 JOURNAL OF THE ARNOLD ARBORETUM LVOL. XVI mon base number. 2. The count on Liquidambar suggests that the Liquidambaroideae may possibly be derived from a different stock than the Hamamelidoideae since they apparently have a different base chro- mosome number. If this difference in base number should be found to persist in the other genus of that sub-family it would indicate that the divergence between the two sub-families occurred before the differentia- tion of the family as a whole. This is in accordance with the views expressed by Harms.! Summarizing the anatomical evidence he states” that the Hamamelidoideae are a unified group anatomically while the other subfamilies, particularly the Liquidambaroideae, have many dis- tinctive peculiarities. Reviewing the entire evidence of relationship" he suggests that the Liquidambaroideae are so distinctive that they might well be considered a separate family. 3. Polyploid series have been found in Fothergilla and Corylo psis and are not to be unexpected in other genera of the family when these are investigated more extensively. This discovery is of some taxonomic consequence since it indicates that in such genera we may expect phylo- genetic relationships between species which will be, in part at least, reticulate. That is to say that a complete phylogenetic tree of the genus Fothergilla or Corylopsis would show anastamosing branches. It will be noted that Fothergilla monticola has 24 pairs of chromosomes and is therefore a tetraploid and that F. major with 36 pairs is a hexaploid. The phylogenetic relationships within and between these two species, as indicated by these chromosome counts, must be intricate. These two species are so similar that it is very doubtful if F. monticola deserves more than varietal rank. The cytological evidence would suggest that F. monticola is merely a tetraploid variety which arose spontaneously from the hexaploid species F. major. Such relationships are not un- known in other genera of the flowering plants. Erlanson for instance has shown’ that Rosa acicularis var. Sayi (Schw.) Rehder is an octoploid race (2n = 56) of the hexaploid species Rosa acicularis (2n = 42). To the larger problem of the phylogenetic position of the family itself this cytological survey contributes important evidence, though unfor- tunately not as decisive as the obscurity of the case requires. Before going into details it may be said that on the whole the cytological evi- dence favors Hutchinson’s interpretation of the phylogenetic position of “loc. cit. p loc. cit. p. 316 4Anderson in Arnold Arb. Bull. Pop. Inform. ser. IV, 1: 61-64 (1933). ‘Bot. Gaz. 96: p. 231 (1935). 1Engler Prantl Nat. Pflanzenf. 2. Aufl. 18a: 303-345 (1930). . 307. 1935] ANDERSON AND SAX, CHROMOSOMES IN HAMAMELIDACEAE 213 the Hamamelidaceae.t_ Comparisons of two treatments of the family are presented in Figure 2. The numbers in the diagram are the base chromosome numbers so far as they have been determined.’ * By inference and by actual experimental test two kinds of change of chromosome number have been established as occurring in the higher plants; (1) the addition of whole sets of chromosomes, that is of two Carurmaheys ~ le 21a (for?) - /F fagaier-!/, 8, 1F Wales — Garryaer -// SaUucaes -/9 Legummasae -6 Fiyacede - 7, 89 an apenaa“es 12,73: Saxagacede -§ M13, (EVE Fosaef HUMMNENTICEDE AL, A 5,7, 0,7 Fratanacese -£/, Cynon aes 1, 13/4, 4/8 Nichhifon Linger Ficure 2. The phylogenetic position of the Hamamelidaceae, accord- ing to Hutchinson and Engler. The numbers are the basic chromosome numbers, so far as is anne sets of six to make a 12 or the addition of an eight and a nine to make a 17; (2) the gradual stepping up or stepping down of the chromosome number by fusion and fraction of one or two pairs of the chromosomes in the previous set, that is, the derivation of an 11 chromosomed species from one with 12, etc. The whole subject is still in the experimental stage but it is at least far enough advanced to indicate that these two processes are among the main forces involved in the separation of genera in the higher plants. It will be seen that higher numbers may be derived a Families of Flowering Plants. I. Dicotyledons. Macmillan and Co., London nan L. O. in Genetic, 12: 161-320 (1930). 8Sax, K. Published and unpublished work on chromosome numbers. 214 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI from lower by either process or by both, but that lower numbers can be derived from higher ones only by the second. Everything else being equal, therefore, those genera with high base numbers will be farther out towards the tips of the phylogenetic net-tree than those with much lower numbers. In the present case the cytological evidence favors the view that the Hamamelidaceae with their base numbers of 12 and 15 are derived from the Rosales stock where base numbers of 6, 7, and 8 are characteristic. Another cytological fact points in the same direction. Hamameiidaceae pr ek seed arnamen dadeae Posordeae JSprrordeae. Liguidambarage ve Prunoideae Fapcidece Soxfragaceae 7 3 Fosades Stock ((» port) Ficure 3. Phylogenetic seuss of the Hamamelidaceae as sug- gested by the cytological evidence. Numbers represent basic chromosome numbers. Further explanation in the text There’ is very strong secondary pairing throughout the family. As shown in figure 1, the chromosomes are not scattered equally over the plate but tend to be more or less grouped. This phenomenon which was first described by Darlmgton and Moffett! and which has been exten- sively studied by Lawrence® indicates that the chromosome complement under observation arose ultimately from the duplication of separate complements. These facts, together with such other information as bears upon the subject, have been utilized in constructing the diagram in Figure 3. It ‘Jour. Gen. 22: 129-151 (1930). 2Cytologia, 2: 352-384 (1931). 1935] ANDERSON AND SAX, CHROMOSOMES IN HAMAMELIDACEAE 215 should be emphasized that the diagram is purely speculative. It has been worked out for those morphologists who would be interested in knowing how a cytologist with such information as is available would speculate as to the relationships of the groups involved. It might well be used as one of a set of possible working hypotheses by students of phylogeny. While the anastamoses of the main trunks of the Rosales stock represent supposed true-breeding allo-polyploid hybridizations, they do not necessarily indicate a cross between families as such. On any evolutionary hypothesis, related families derive, ultimately, from forms no more differentiated than present day genera or species. All that need be hypothecated for these hybrids is that they are between forms as diverse morphologically as certain hybrids which have been experimentally obtained, those between Zea and Tripsacum, for in- stance." The diagram is based upon the evidence from chromosome number, secondary association and, in the case of the Pomoideae, from breeding experiments. It is much more speculative for the Saxifragaceae than for the Rosaceae. The Saxifragaceae, with base numbers of 8, 11, 13, 14, and 16 show a cytological complexity” paralleling their morpho- logical diversity. Only a few of the fossil “dead branches” have been indicated. There must certainly have been many more. In this respect as in several others the actual details of the diagram are probably in- correct. The general conception, however, of a more or less webbed net-tree for the Rosales is strongly supported by the cytological evi- dence. In some other groups of the flowering plants (the Tubiflorae, for instance) the webbing would be so much more complex that one would scarcely use the word tree in describing it. In the Cyperaceae, on the other hand, there would be few if any anastamosing branches. The cyto-genetic evidence shows with increasing force that the actual pattern of evolutionary progress has been different in different groups of plants. The main point of the diagram in figure 3 is to suggest the general nature of the evolutionary pattern of the Rosales. SUMMARY 1. Chromosome counts are given for nine species and six genera of the Hamamelidaceae. 2. The phylogenetic position of 'the family is discussed in the light of these results. ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1Mangelsdorf & Reeves in Jour. Hered. 22: 329-343 (1931). *Sax, K. in Jour. Arnold Arb. 12: 198-206 (1931). 216 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI CHROMOSOME STRUCTURE IN THE MEIOTIC CHROMOSOMES OF RHOEO DISCOLOR HANCE KARL SAX With plates 130 and 131 COILED CHROMONEMATA have been observed in both mitotic and meiotic chromosomes in various species of plants. The degree of coiling in mitotic chromosomes may vary from an irregular corrugation or loose spiral to a rather compact regular coil. At meiosis in certain plants the coils are much larger and can be analyzed in more detail. The behavior of the meiotic chromosomes of Rkoeo discolor provides additional information regarding the nature of the coiled chromonemata. The observations are based on both aceto-carmine preparations and per- manent smears fixed in Flemming’s solution. In both cases the micro- sporocytes, after smearing, were usually pretreated with alcohol and ammonia before fixing. Immersion in tap water for a few seconds and a brief exposure to ammonia fumes also gave good results. This is essentially the method used by Kuwada. The ammonia vapor seems to dissolve the chromosome matrix and permit the spiral chromonemata to expand. The chromosomes of Rhoeo are arranged in a ring or one or more chains at the first meiotic division. At early metaphase each chromo- some contains a coiled chromonema consisting of two closely associated chromatids. The general appearance of these rings is shown in the photographs from aceto-carmine preparations (Figs. 1 and 2), but the finer details are best observed in permanent smears (Figs. 3 and 4). The number of coils is four or five per chromosome. At this stage the two chromatids are so closely associated that the coils appear to be single, but their double nature can be observed at certain loci. The diameter of the coil decreases gradually before the separation of the chromatids. At late metaphase the coiled chromatids separate. The number of coils in each chromatid remains the same, but they are much smaller, even though there is little or no elongation of the chromosome as a whole (Figs. 5 and 6). At this stage both the terminal chiasmata and the fiber constrictions are very conspicuous, so that the order of the indi- vidual chromosomes in the ring can be determined. As observed earlier (Sax 1931), the order of the twelve chromosomes is always the same, thus supporting Belling’s suggestion that such rings are the result of 1935] SAX, CHROMOSOME STRUCTURE IN RHOEO DISCOLOR 217 segmental interchange. Six of the twelve chromosomes are distinctly heterobrachial, and the short arms are paired. The meiotic chromosomes are always paired by terminal “chiasmata” and, in the chromatids at late metaphase, most of the chiasmata appear to be symmetrical. The chromatids are often parallel in several suc- cessive chromosomes or even in most of the chromosomes in the ring. As the chromosomes pass to the poles at the first meiotic division, the chromatids separate except at the fiber constriction and become shorter (Fig. 7). During interphase the chromatids elongate considerably. At the second meiotic division they are about twice as long as they were at the first division. At this time the minor or somatic type of coiling can be observed. The coils are only about half as wide as the major coils found in the first division, and the number of coils is about 12 per chromatid (Fig. 8). These coils appear to be single at metaphase. There is some evidence of a split at late anaphase, as indicated by narrow regions at certain loci. If the chromatids are split at second metaphase, the two halves must be coiled together. At late anaphase they may tend to separate, but the slipping apart of the coils is difficult to detect except where there is a twist which appears to constrict the chromosome at such loci. More definite evidence of split chromatids has been obtained from microsporocytes which were subjected to low temperatures during development. Under normal conditions the microspores receive six chromosomes, but occasionally there are seven, owing to irregularities in the first meiotic division. About 80 per cent of the microspores fail to develop, owing to segmental non-disjunction. The normal fertile microspores undergo a single nuclear division, followed by the differentiation of the daughter nuclei into the large and more or less degenerate tube nucleus and the compact elongated generative nucleus. When the plants are kept at a temperature of about 6° C., two kinds of abnormalities appear. The chromosomes of the one nucleate micro- spore may divide but do not form daughter nuclei. The 12 chromosomes pass back into the resting stage and divide regularly at the next division to form diploid gametes. The other type of abnormal development begins when the low temperature inhibits chromosome pairing at meiosis. The twelve univalent chromosomes pass into the resting stage without nuclear division. They come out of the resting stage, divide without nuclear division and form a giant nucleus with 24 chromosomes, each of which consists of two coiled chromatids held together only at the fiber constriction (Fig. 9). Throughout this process the chromosomes never pass through the contraction characteristic of normal telophase stages. 218 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI The chromatid spirals are much looser than those found in the chroma- tids of the normal chromosomes at the second meiotic division, and it is perfectly clear that many of these chromatids are split. The splits are especially clear near the ends of the chromatids. At a somewhat later stage the two chromatids become completely separated, the chromatids elongate, and their structure is very clear. In the same chromatid one can observe the transition from a single coil to two parallel finer coils (Fig. 10). The mechanism of separation of daughter chromatids in these chromosomes, which are essentially of the somatic type, is similar to that found in the coiled chromatids in the first meiotic division. The structure of the meiotic chromosomes in the permanent prepara- tions was clear enough to permit an analysis of direction of coiling in the spiral chromonemata. We were able to determine the direction of coiling at all loci in each of the twelve chromosomes in 14 cells with com- plete chromosome rings. The classification of right- or left-handed spirals is purely arbitrary, since the direction depends on the sequence of determination in the ring. There is a strong tendency for the direc- tion of coiling to be in the same direction in both arms of a chromosome. Of the 168 chromosomes examined, 50 had a right-handed spiral in both arms, 52 a left-handed spiral, and 66 showed a reversal of coiling, pre- sumably at the fiber attachment. In only two chromosomes was there a second change of direction of coiling. Individual chromosomes could not be identified consistently, so that the direction of coiling could not be established for any one chromosome in all the different cells, but there is good evidence that direction of coiling is not a stable character. The number of chromosomes with left-handed coils ranged from 1 to 5, with reversed coils from 2 to 8, and with right-handed coils from 1 to 7, in different cells. The direction of coiling of chromonemata of paired adjacent chromosome arms is at random, with 85 coiling in the same direction and 83 coiling in reverse directions. Both rings and chains of chromosomes are found at meiosis. If one or more chains are formed, the breaks may occur between either the long segments or the short segments. Three of the terminal chiasmata are formed between short segments, and eight of the terminal chiasmata are formed between long segments. The extra chiasma is between a long and a relatively short segment. The position of the breaks, or failure of chiasma formation, was obtained for 20 cells containing one or more chains. There were 9 breaks between the short segments and 16 between the long segments. These results suggest that chiasma formation is some- what less likely to occur, or less likely to persist until late metaphase, between the short segments. The latter possibility is more probable 1935] SAX, CHROMOSOME STRUCTURE IN RHOEO DISCOLOR 219 because we have found chromosome rings in practically all cells in some preparations. The lengths of the chromosomes at different stages in meiosis were obtained to aid in the analysis of factors involved in chromosome con- traction. It was not possible to get an accurate measurement of the length of the pachytene spireme in Rhoeo, but the total length is approximately 700 microns or an average of somewhat more than 100 microns per chromosome. The approximate average chromosome length at meiotic metaphase is 5—6 microns, and is about 9 microns at the second meiotic division. THe MECHANISM OF CHROMOSOME CONTRACTION The great contraction in chromosome length between pachytene and meiotic metaphase stages in Rhoeo is associated with the coiling of the chromonema. This coiling may not be the only factor involved. Belling (1928) believed that the approximation of chromosomes caused about one-third of the contraction in the chromosomes of Lilium, and that the coiling of chromonema effected the final shortening to give approximately a 10 to 1 reduction in length of the meiotic chromosome. Bridges (Alexander, 1928), on the other hand, assumed that coiling is the primary factor in chromosome contraction, and that the gene string maintains approximately the same length at all stages in the chromosome cycle. Another factor in chromosome contraction is the secondary or minor coils within the primary or major coils as described in Tradescan- tia by Fujii, Kuwada and Nakamura (1933) and found in Sagittaria by Shinke (1934). We believe that three factors are involved in the great decrease in length of the meiotic chromosomes of Rhoeo; first, a linear contraction of the gene string; second, the major coiling of the chromonema; and third, the formation of minor spirals within the major spiral. The minor coils are not clearly differentiated at the first meiotic division in Rhoeo, but there is some evidence of loose coiling. The contraction of the chromonema and reduction in the width of the major coils between early and late metaphase are attributed to the further coiling of the minor spirals in each chromatid. A similar reduction in the major coils with no increase in chromosome length is found in Secale (Sax, 1930). The coiled chromonema at early metaphase consists of two chroma- tids coiled together so that the chromonema often appears as a single coil, as is the case in Tradescantia, Secale, Lilium, and Vicia. The free separation of coiled chromatids has been explained by Kuwada (1927). 220 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI The reverse twists postulated by Kuwada can be observed in Trade- scantia (Sax and Humphrey, 1934) and in Trillium (Huskins and Smith, 1935). The direction of coiling in the chromonema spirals is more or less at random in Rhoeo and in Tradescantia (Nebel, 1932, Sax and Humphrey, 1934). According to Huskins and Smith, the paired meiotic chromo- somes of Trillium usually coil in opposite directions, although no statis- tical evidence is presented, nor is there any adequate explanation for such behavior. Within a single meiotic chromosome the direction of coiling may change at the spindle fiber point, but is seldom reversed at other loci in Rhoeo, Secale, Gasteria (Taylor, 1931), Tradescantia, and Sagittaria (Shinke, 1934). Huskins and Smith find frequent changes in direction of coiling of anaphase chromosomes in Trillium. These changes in direction of coiling are usually associated with chiasmata. We have found changes in direction of coiling at chiasmata in the meiotic chromosomes of Vicia. If the chromonema coiling is caused by a con- traction of the matrix, as suggested by Kuwada, the fiber attachment points and the chiasmata would tend to break any continuity of stress on the chromonema and changes in direction of coiling would be expected to be more or less at random at these points. There is a strong tendency, in both Rhoco and Tradescantia, for the direction of coiling to be the same on both sides of the spindle fiber attachment, and only about one-third of the chromosomes show reversal of coiling at this locus. Huskins and Smith find that the reversals in direction of coiling between the fiber attachment and the distal ends of the chromatids at first anaphase is about twice the chiasma frequency at metaphase. This relation would be expected if the direction of coiling in homologous chromosomes is at random, and if reversals in coiling occur at random at the chiasmata. In Rhoco all chiasmata are terminal, and most of them appear to be symmetrical. The short chromosome arms are paired almost as fre- quently as the long arms. Changes in direction of coiling are rarely observed between the fiber and the distal end of the chromosome. These observations seem to indicate that chromosome pairing in Rhoeo is not dependent on the formation of interstitial chiasmata, but is dependent on a terminal association of homologous chromosome segments. This terminal association in certain rod bivalents in Tradescantia seems to involve the chromosome pellicle or matrix, but in Rhoeo there is evi- dence of fine chromatic connecting fibers. During interphase the meiotic chromosomes of Rfoeo elongate but maintain some evidence of loose coils during the resting stage. At the 1935] SAX, CHROMOSOME STRUCTURE IN RHOEO DISCOLOR 221 second meiotic division, a new spiral appears which has finer and more numerous spirals than the major coils at the first meiotic division. These minor spirals have been described in Lilium, Rhoeo, Allium, Tricyrtis, Najas, and Hosta by Shinke (1930), and in Tradescantia by Nebel (1932), Kuwada and Nakamura (1933) and Sax and Humphrey (1934). These coils in Rhoeo are wider than the minor spiral’ within the major spiral at the first meiotic division and presumably are formed inde- pendently during the prophase of the second meiotic division. These spirals are much like those found in certain somatic chromosomes. The transition from about five major spirals to 20-25 minor spirals in the successive meiotic divisions in Tradescantia and a similar behavior in Rhoeo is difficult to reconcile with the ‘“‘heterogonic growth” hypothesis of spiralization suggested by Huskins and Smith. The anaphase chromosomes at the second meiotic division have been described as two-parted in a number of plants (Gasteria, Taylor, 1931; Galtonia, Smith, 1932; Tradescantia, Nebel, 1932; Trillium, Huskins and Smith, 1935; et al.; Cf. Sharp, 1934). Both Kuwada and Naka- mura, and Sax and Humphrey found only single coils in the second ana- phase chromosomes of Tradescantia. In Rhoeo there is evidence that the anaphase coil is double, but that the two half-chromatids are coiled together so intimately that they appear as a single coil at early anaphase. As the coils begin to separate the gyres are matched so closely that the dual structure is not clear, but where a twist occurs there is a narrow region in the chromosome. The abnormal “microspores” of Rhoeo show the chromatid splits clearly in various stages of separation. The minor coils, characteristic of somatic chromosomes, are similar in struc- ture to the major coils; the two chromatids (or half-chromatids) are coiled together in parallel in such a manner that they can separate freely without entangling. If the split occurs while the chromosome is coiled, there must be some lateral polarity so that the division occurs in only one plane parallel to the axis of the chromosome, as Nebel (1933) has suggested. If there is a chromatid split in the anaphase chromosomes of the second meiotic division in Tradescantia, the chromatid must behave as a single unit until midprophase of the microspore division. Tradescan- tia microspores subjected to x-rays show chromatid breaks for about two days after raying, but after three or four days only chromosome breaks are observed at metaphase (Riley—unpublished ). We find that either abnormally low or high temperatures will cause nuclear irregularities. These include failure of chromosome pairing at 222 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI meiosis, chromosome division without nuclear division, and failure of normal differentiation of nuclei. Since these temperatures are within the range occasionally experienced in nature, it is probable that tem- perature fluctuations have played an important part in chromosome changes in nature (Cf. Randolph, 1932). This work with temperature effects has been aided by a grant from the American Academy of Arts and Sciences. SUMMARY The twelve chromosomes of Rhoeo discolor are arranged in a seg- mental interchange ring at the first meiotic division. Each chromosome contains a spiral chromonema consisting of two chromatids coiled together. There is some evidence, both direct and indirect, that there is a minor spiral within the chromatids of the major spiral. During metaphase the major spirals become smaller, and the two chromatids separate. The chromosomes elongate greatly during interphase, but there is evidence of a spiral structure during the resting stage. At the second meiotic division, new minor spirals are formed which are smaller and more numerous than the major spirals of the first division. The chromatids at anaphase of the second meiotic division are split, but the two half-chromatids are so closely coiled together that they are not easily observed. They can be differentiated easily in cells where the normal chromosome cycle is disturbed by subjection to low temperatures. The reduction in the length of the meiotic chromosomes of Rhoeo, between prophase and first metaphase is attributed to three factors: — a linear contraction of the gene string, the coiling of the chromonema into major coils, and the development of minor coils in the chromatids of the major spirals. The direction of coiling in the major spirals seems to be at random. In a single chromosome the direction of coiling may change at the fiber attachment point, but it is seldom reversed at other loci. The meiotic chromosomes are paired at the ends, apparently without the formation of interstitial chiasmata. LITERATURE CITED ALEXANDER, J. (1928). oo Chemistry. Vol. u, p. 1029. (The Chemical Catalogue Co., New York. BeLuinG, J. (1928). Za eos of chromosomes during pease Cal. Pu divisions in Lilium and other plants. (Univ b. Bot. 14:335-343. ) Husk C. L., and S. G. SmitH (1935). " Meiotie chromosome struc- nds in » Tyilliaa erectum L. (Ann. Bot. 49:119-150.) 1935] SAX, CHROMOSOME STRUCTURE IN RHOEO DISCOLOR 223 Kuwapa, Y. (1927). On the spiral structure of chromosomes. (Tokyo Bot. Mag. 41:100-109. ———— and Nakamura (1933). Behavior of chromonemata in mitosis. I. Observation of pollen aes cells in Tradescantia reflexa. (Mem. rt Sci. Kyoto Imp. Univ. 9: ae — AKAMURA (193 Behavior of chromonemata in mitosis. aot Artificial unravelling of coiled aaa: (Cytologia, 5:244-24 NEBEL, B. R (193 2). Chromosome structure in Tradescantiae. I. Methods and morphology. (Zeit. f. Zellf. u. mik. Anat. 1 :251- ———._ (1932). direction of coiling oe e chromonema in Tradescantia reflexa Raf., T. virginiana L., na “pendula Schnizl., and Rhoeo discolor Hance. 16 :285-304. (Zeit. f£. Zellf. uw ce pe ——— (l . Chromosome sractis re in Tradescantiae. IV. The history of the chromonemata in mitosis of Tradescantia reflexa Raf. (Cytolog ia, a 1-14.) RANDOLPH, L. F . (1932). Some ee of high temperature on polyploidy 222-229. and other variations in maize. (Proc. Nat. Acad. 18: Sax, Karu (1930). romosome ae and the mechanism of cross- ing a, A he Arnold Arb. 11:193-220. —_——— (1931). Chromosome ring formation in Rhoeo discolor. (Cyto- logia, 3: 36-88 1 L. M. HUMPHREY (1934). Structure of meiotic chromo- somes in 1 mnicrosporogenes is of Tradescantia. (Bot. Gaz. 96:353-361.) Snarp, L. W. (1934). Introduction to Cytology. (McGraw-Hill Book ew York.) SmitH, F. H. (1932). The structure of the somatic and meiotic chromo- somes of om candicans. (La Cellule, 41:243-263. ) SHINKE, N. (193 n the spiral structure of chromosomes in some higher plants. Mem Coll. Sci. Kyoto Imp. Univ. 5:239-245. a 34). Spiral structure of chromosomes in meiosis in Sagit- taria Aginashi. ie m. Coll. Sci. Kyoto Imp. Univ. 9:367-392 Taytor, W. R. (1931). Chromosome studies on Gasteria. ITI. Chromo- some structure during microsporogenesis and the postmeiotic mitosis. (Amer. Jour. Bot. 18:367-386.) EXPLANATION OF PLATES PLATE 130 pd fia ieee of Rhoeo discolor. Figs. 3 and 4 from permanent All others from aceto-carmine preparations. X 2000. Figs. ce ve 2. The coiled chromonemata in the chromosome ring at early m Figs. 3 and 4. Coiled chromonemata showing reversal of coiling and re- duction in width of coiling at metaphase. Figs. 5 and 6. Separation of coiled chromatids at late metaphase. The me number of coils are found in the coiled chromatids as in the coiled chromonemata, but the “spiral are smaller. Fig. . Telophase of first meiotic divisio Fig. 8. Chromosomes at the s econd TEE division showing minor spirals which appear to be single. N 224 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XvI PLATE 131 Chromosomes from sie “microspores” produced by chromosome multiplication induced by cold treatment. Magnification xX 2000. Fig. 9. A giant cell eau ee a microsporocyte by chromosome divi- sion without nuclear division. Each of the 24 chromosomes consists of two chromatids held together at the fiber attachment. Many of the chromatids are split. Fig. 10. Chromatids at a somewhat nee stage showing the transition from a single coiled chromatid to two coiled daughter chromatids. ARNOLD ARBORETUM, Harvarp UNIVERSITY. Jour. ArNoLtp Ars. Vor. XVI PLATE 130 CHROMOSOME STRUCTURE IN MEIOTIC CHROMOSOMES OF RHOEO DISCOLOR Jour. ArNotp Ars. Vor. XVI PLATE 131 Sa ee CHROMOSOME STRUCTURE IN MEIOTIC CHROMOSOMES OF RHOEO DISCOLOR 1935] FAULL, ELAIOPLASTS IN IRIS 225 ELAIOPLASTS IN IRIS: A MORPHOLOGICAL STUDY ANNA F. FAULL With plates 132 to 137 INTRODUCTION ELAIOPLASTS are a heterogeneous group of intracellular bodies pre- senting the characteristics of fatty substances to a marked degree but not recognizable as ordinary types of plastids, chondriosomes or vacuoles. There is no general agreement in the literature regarding their structure, origin, development, classification or chemical compo- sition. They have been variously described as aggregations of lipoid globules, as modified or unusual types of plastids or vacuoles, as nuclear derivatives, as aggregations of mitochondria-like bodies or as inde- pendent structures. They have been linked with various physiological processes such as assimilation, excretion or degeneration. Much of the confusion regarding elaioplasts is due to the use of inadequate techniques and to a consequent lack of accurate information about the early developmental stages of these bodies. Many of the discrepancies are also due to failure to visualize and interpret correctly the full range of morphological variability of chondriosomes and plastids. The investigations described in this paper were undertaken to clarify our conception of the elaioplasts in /ris and to compare these bodies with those in other monocotyledons and in liverworts. By using an improved technique critical evidence has been obtained to show the early stages in the development of the elaioplasts in /ris and the changes which these bodies undergo in different tissues and at different seasons. In addition to the morphological study an extraction and preliminary analysis of the so-called oil in the elaioplasts of /ris and some physio- logical experiments on the metabolism of two types of Iris rhizome, one of which contains abundant elaioplasts, have been made. ‘The results of these investigations are being published elsewhere. HISTORICAL RESUME Since the middle of the last century papers have appeared from time to time describing cytoplasmic bodies associated with oil. The writers have used various names for these structures which through usage have become more or less interchangeable. Thus they are termed elaioplasts, 226 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI oléoplasts, oléosomes, Olk6rper, oil bodies, Zellenblaschen, Zellenkérper, fatty bodies, elaiosferer, oelplastids, oléoleucites, éléments oléiféres and systeme oléifére. Sometimes the terms are restricted in their applica- tion. For example 6lk6rper is used only for oily bodies in the liver- worts, and elaioplast is kept for those in the monocotyledons. But recently with an increasing tendency to consider all of these oily bodies essentially similar, one name is often used to designate all of them. The earliest references to elaioplasts are found in the writings of Mirbel (35) in 1835, of Gottsche (13) in 1843, of Holle (24) in 1857, of Hofmeister (23) in 1867 and of Ward (49) in 1883. But the first adequate descriptions of oil bodies were published in papers by Pfeffer (38) in 1874 and by Wakker (47) in 1888. These, together with a paper by Lidforss (32) in 1893, provide a description of the three main types of mature oil bodies; from this later authors have diverged little. Although often resembling one another, the three main types present certain distinct features which are further emphasized by their restric- tion to a given group of plants. Pfeffer (38) described oil bodies characteristic of the liverworts. In common with such bodies in general, they are highly refractive struc- tures which stain brilliantly in “fat” dyes, such as alkannin, and which are more or less soluble in 95‘¢ alcohol and in fat solvents such as ether. They are distinguished from other oil bodies by their almost complete solubility in alcohol, by a characteristic residual ring left after treat- ment with alcohol, by their location commonly in the peripheral cyto- plasm but within the chloroplast-bearing layer, by their presence in practically every species of the group, by their appearance commonly in every cell of a plant and by their permanency as cell structures. The Marchantiales present a contrast to other elaioplast-bearing hepatics in the restriction of the oil bodies to special cells scattered throughout the thallus and in the location of a single large oil body in the center of each of these cells. Oil bodies in the liverworts vary in shape from round to spindle-shaped as a rule, though some are irregular in form. They vary in color from colorless to dark brown and in appearance from granular to segmented or homogeneous. The oil bodies described by Wakker (47) differ from those in the liverworts in their location near the nucleus, in their invariably granular appearance, in their often irregularly lobed shape, and in their charac- ter of being more or less temporary cell structures. Elaioplasts of this type are often yellowish in color and are marked by their reaction with some reagents which cause an extrusion of the oil and leave a charac- teristic net-like structure. Although reported from most tissues, they 1935] FAULL, ELAIOPLASTS IN IRIS 227 are often restricted to certain ones. Raciborski (41) and Beer (4) found them only in flower or fruit tissues, while Politis (39) described them in these tissues and in those of bulbs. Oil bodies of this type are further restricted to a few groups of flowering plants. Lists published by Zimmermann (51) and by Politis (39) record them in groups of species in the Orchidaceae, Liliaceae, Amaryllidaceae, Iridaceae and Malvaceae, while Beer (4) found them in one of the Compositae. The third type of oil body described by Lidforss (32) is character- ized by its homogeneous appearance, by its spherical shape and by its unrestricted location in the cell. It is reported from leaf tissues of flowering plants and is of common occurrence in this group. Besides these three classes of oil bodies there are isolated descriptions of elaioplasts that are not included in any of the types described. Such are the reticulate, highly refractive structures saturated with an amber- colored oil described by Keene (26, 27) in two molds. Such also are the yellow, green or black oil bodies near the nucleus found by Hierony- mus (22) in some algae. In 1888 Wakker (47) demonstrated by abnormal plasmolysis that the oil bodies in the monocotyledons and in the liverworts are located in the cytoplasm. He showed that, although these structures often protrude into the vacuole, they are never located in it as Pfeffer (38) and Rattray (42) had thought. Later investigations have substan- tiated Wakker’s observations and extended them to include all types of elaioplasts. There is no general consensus of opinion on the structure of the non- homogeneous oil bodies. Pfeffer (38) described them as aggregations of homogeneous oil globules, a view expressed in modern times by Guilliermond (20), by Meyer (34) and by Kozlowsky (28). Other students have described a stroma with embedded oil globules. This view was first expressed by Wakker (47). It was elaborated upon by Zimmermann (51), who pointed out less refractive inclusions which he termed vacuoles or portions not producing oil. Later Beer (4) and Politis (39) described the elaioplasts in Gaillardia and in the mono- cotyledons as aggregations of smaller bodies, each composed of a stroma with included oil globules. A more elaborate structure was postulated by Woycicki (50) and by Keene (26, 27). Woycicki described elaio- plasts in Vanilla with central oily drops surrounded by a mucilaginous layer which in turn was covered by a granular layer. Keene described a somewhat similar structure in the oil bodies of Sporodinia which showed a denser reticulate center and a coarser reticulate outer portion. The presence of an unfixable stroma in the oil bodies of the liverworts in 228 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI contrast to the fixable one in those of the monocotyledons was pointed out by Kuster (29), Later Gargeanne (9) and Dombray (7) attempted to show that this unfixable stroma was a fluid or a semi-fluid. The question of an enveloping membrane has been raised with refer- ence to the oil bodies in the liverworts. Pfeffer (38) inferred the pres- ence of a membrane from the characteristic ring left after treatment with alcohol. Kuster (29) demonstrated in 1894 that this ring is an artefact. Gargeanne (9) repeated the demonstration but maintained that, although the ring is an artefact, the bodies possess a true mem- brane homologous with the tonoplast of the vacuole. Later writers have not agreed with Gargeanne in recognizing a membrane. The pres- ence of a membrane about elaioplasts in the monocotyledons has been described only by Raciborski (41) who considered the stroma at times to be reduced to a surrounding layer. The development of the oil bodies is also a disputed point in the literature. Pfeffer (38), Rivett (43), Lidforss (32), Chalaud (5), Meyer (34) and Guilliermond et al. (20) have considered the formation of elaioplasts to be a process of aggregation of small drops in the cyto- plasm with more or less fusion. Kozlowsky (28) has further stated that the drops are first extruded from the chloroplasts. A second theory has been postulated by Wakker (47), by Kiister (29) and by Harper (21). They consider that a stroma appears first as a shadowy, wrinkled structure in which refractive oil drops appear later. Gargeanne (9° stated that the oil drops are secreted by a surrounding membrane, while Dombray (7) described the deposition of substances from the cytoplasm and their transformation by the cell sap as a catalyser. An- other theory is expressed by Hieronymus (22) and by Beer (4) who described elaioplasts formed by the aggregation of degenerating plastids with the production of oil. Somewhat similar is Woycicki’s (50) theory of the aggregation, partial degeneration and fusion of mitochondria-like bodies forming oil. Keene (26, 27) postulates the formation of a reticu- late structure in homogeneous bodies with the later fusion of several of these bodies. Still another theory by Politis (39) and by Raciborski (41) describes the development of elaioplasts by the growth of refrac- tive drops and the subsequent fusion of the bodies so formed. The division of elaioplasts has been noted in a few instances. Raci- borski (41) in 1893 described a fragmenting of the bodies after they had passed maturity and a breaking off of bud-like protrusions. Again in 1914 Politis (39) described division of the elaioplasts. Politis con- sidered division not merely an incidental or degeneration phenomenon, but a method of increasing the number of these bodies. Besides the 1935] FAULL, ELAIOPLASTS IN IRIS 229 budding already described by Raciborski, Politis described passive division of the body by the cell wall during cell division. The history of oil bodies after they have reached maturity has been studied. In the liverworts they are generally thought to remain un- changed even after the death of the cell, although Dombray (7) noted a decrease in size, fusion of the oily globules and aggregation of the oil bodies before death. Elaioplasts in the monocotyledons are generally thought to degenerate some time after reaching maturity. Wakker (47) described their disappearance in older tissues of Vanilla. Beer (4) and Woycicki (50) described a resolution of the oil bodies into scattered oily spheres. Politis (39), on the other hand, described the disappear- ance of the oil first, leaving a vacuolated protein mass which might later disappear also. Movement has been noted in connection with elaioplasts. In 1893 Zimmermann (51) first recorded the rotation at times of oil bodies in the monocotyledons, a phenomenon observed also by later investigators. A second type of motion consisting of Brownian movement of the globules within the oil bodies appears in oily structures in the hepatics. Gargeanne (9) described this as an injury phenomenon, but recently it has been noted by Dombray (7) as a normal condition in the elaioplasts of some species. The chemical composition of the elaioplasts and particularly of the oily portion has received much attention. The theories advanced are based chiefly upon microchemical reactions. Dombray (7) has inter- preted microscopical observations in the light of analyses of extraction products. Two opposing theories regarding the composition of the oil have been formulated. In one the oil is said to be chiefly a mixture of essential oils. This is the view recently expressed by Popovici (40) and by Rivett (4) in her description of the oil as a mixture of essential oils with small amounts of protein and fatty oils. Dombray (7) stated that the oily substance was a mixture of essential oil and “tannoides.” The opposing theory considers the oil to be composed chiefly of fatty oils. This is the opinion of most investigators. Pfeffer (38) described the oil as a mixture of fatty oil with some water and protein and with traces of wax and resins. Later Kiister (29) designated the oil in the elaioplasts of liverworts as a fatty oil resembling castor oil. Lidforss (32) identified the oil in the homogeneous oil bodies of flowering plants as a non-drying oil containing fatty acids of the type C® Hon-2 O02. The stroma, if present, is generally considered to be a protein, a view first expressed by Zimmermann (51). There is little agreement among investigators concerning the origin 230 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI and identity of oil bodies. Raciborski (41), Kuster (29) and Gavaudan (10, 11, 12) have considered them to be cell systems independent of the vacuome, chondriome or plastidome and originating more or less de novo in the cytoplasm. Wakker (47) considered them to be special plastids, while Beer (43) and Hieronymus (22) described them as degenerating plastids. Kozlowsky (28) has stated that they are products of the chloroplasts. That they are special or transformed vacuoles has been postulated by Keene (26, 27), by Dombray (7), by Gargeanne (9) and by Rivett (43). Woycicki (50) and Lundstrom (33) have described oil bodies originating from mitochondria-like bodies. Politis (39) has ascribed a nuclear origin to them. A relationship between elaioplasts in the monocotyledons and crystal formation has been suggested. Wakker (47), Politis (39) and Monte- verde (36) stated that there is no connection between the oil bodies and the calcium oxalate crystals found in the same plants. But Warlich (48) considered them to be interdependent structures, while Woycicki (50) in 1929 described crystals forming in some of the elaioplasts in Ornithogalum. Many writers have ventured theories on the physiological and bio- logical significance of the oil bodies. In general they have considered those in the liverworts and also the homogeneous ones in the flowering plants to be excretions. But those in the monocotyledons they term assimilation products, although Raciborski (41) stated them to be excretions. Various other theories have been offered. Beer (4) in 1909 demonstrated that the bodies in Gaillardia are degeneration prod- ucts of the plastids with the secondary function of producing color. Hieronymus (22) and Lundstrom (33) suggested that the bodies are protective in function, a theory opposed by Dombray (7). In concluding the summary of the literature on oil bodies it should be noted that these structures do not include the dlplasma described by Leiner (31) and by others of Tschirch’s school, nor do they include the oil cells described by Lehmann (30) and others. The former (6lplasma) deals with oil in the cytoplasm — chiefly of fatty seeds. The subject is well summarized in the account by Tunman and Rosenthaler (46). The phenomenon of the appearance of oil in special oil cells involves the transformation of large portions of the cytoplasm or secretion from the modified cell wall rather than the appearance of oily bodies in the cytoplasm. In addition to the literature on oil bodies, some reference should be made to the literature on the structure of Iris cells. The most recent and complete studies are those by Guilliermond (15-20) and by Dan- 1935] FAULL, ELAIOPLASTS IN IRIS 231 geard (6). They have developed a method of vital observation espe- cially adapted to this type of study. They have pointed out the presence in Iris cells of vacuoles and their inclusions, of cytoplasm, of oil globules, of chondriosomes of various types and of plastids. In par- ticular Guilliermond has described the chondriosomes and plastids and their developmental stages. He has noted the presence of oil globules in most plastids and chondrioconts in Jris. These globules which he has found associated more often with young or degenerating types of plastids he considers to be lipoids separating out from the plastid sub- stance. He has described the development of plastids from mitochon- drial types differentiated from other mitochondria by their potentiality for plastid formation. He has described the formation of chloroplasts from an intermediate chondriocont stage by budding and fragmenting. Other phases of studies carried out on Jris include the action of hypo- and hypertonic solutions on chondriosomes, observations of the amoe- boid movements of chondrioconts, and the identification of an oily body in the vacuoles of certain cells as a phenol compound. MATERIAL The plants used in my studies of elaioplasts included numerous irises, some liverworts and a few representative flowering plants. They were obtained from several sources. The major part of the study was made on colonies of Jris versicolor and of an Iris pallida of hybrid origin which grew in abundance near the laboratory. For work on living tissues it was desirable to have the plants as close at hand as possible. It was also desirable to locate single colonies in a natural habitat for the basic study of variations. In this way differences due to season, development, etc., were less likely to be confused with those due to location, to abnormal habitat or to individual variations. As a rule the material was used as soon as it was collected. But in some instances it was kept in water or in wet sand in the greenhouse for later observations, or it was transplanted to garden beds. In the early part of the study a few plants of J. pallida and of J. versicolor were transplanted to pots in the greenhouse to supplement the outdoor material. Although some interesting observations were made on these plants, they grew so poorly that this method of providing material was abandoned. Fortunately, it was not necessary to rely on greenhouse or garden material at any period. The /ris versicolor was taken from a swampy field at the corner of Weston St. on the Cambridge-Concord turnpike about an hour’s drive from the laboratory. The Jris pallida hybrid, a garden plant, grew in 232 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI beds within a few rods of the laboratory. Both of these species were sufficiently near at hand to be obtained as they were needed for examination. The other irises used as supplementary material and for a general survey of the genus were obtained from several places. A group of native West Coast species was studied in California. Three of these, I. macrosiphon, I, Douglasiana and J. longipetala, grew naturally within a few hours’ drive of the laboratory. But J. Hartwegii and J. missouri- ensis had to be brought to Palo Alto from the eastern part of the state. A large number of other species were obtained from the Missouri Botani- cal Garden, from the Brooklyn Botanical Garden and from the New York Botanical Garden. Those at the Missouri Garden were examined in situ, but the ones from New York were brought to Boston for examination. For a list of the species of Jris studied see the table on page 246. In addition to the Iris plants, a group of rhizomatous plants and a number of liverworts were obtained. The former were studied at the Missouri Botanical Garden for the most part, although a few were collected around Boston. Two species of Vanilla, the plant used by Wakker (47) in his classical studies on elaioplasts, were obtained from Panama. ‘The liverworts were collected in the New England woods for study in Boston or they were sent from Oregon to the California labora- tory for use there. The hepatics were kept in the laboratory in moist glass containers over a period of weeks. A list of the flowering plants studied is given on page 248 and one of the hepatics on page 254. TECHNIQUE In choosing a method for the morphological examination of the elaio- plasts one fundamental requirement was kept in mind. It was desir- able to observe the bodies in as unaltered a condition as possible in order to-discover their normal development and variations due to seasonal, environmental or specific differences in the plants examined. At the present time there are two methods used in the study of cyto- plasmic bodies. The first of these is the fixation technique introduced in the later decades of the nineteenth century and developed to the highest degree in the complicated “mitochondrial techniques” and “silver or osmic impregnation methods.” Essentially it consists of kill- ing and fixing blocks of tissue in reagents that solidify proteins and fats, rendering them insoluble in specific fluids, and then staining sections differentially. Incidentally the technique involves a rather complicated process of embedding and one or more dehydrations. 1935] FAULL, ELAIOPLASTS IN IRIS 233 The other method is that of examining untreated tissue either with or without the aid of vital dyes. Although untreated tissue was used before the introduction of fixatives, it was superseded by them. Recently the so-called vital technique has been revived and developed, notably by the Dangeards and by Guilliermond in France and by Bailey in America. Guilliermond has described a technique for vital staining in his studies of the vacuome and has contributed data on various aspects of injury and death in his studies of the chondriome. Bailey (1) in his investigations of the cambium has tabulated criteria that can be used in distinguishing living from dying or dead cells. Bailey and Zirkle (2, 3) have clarified the vital staining technique by their investi- gation of the toxicity of a large number of dyes, of the most suitable media in which to use the stains, of the staining properties of different dyes and of the varying reaction of vacuoles to given stains. Both of these methods were tried in the study of /ris, but that of fixa- tion was eventually discarded because of the difficulties involved. The vacuoles in the rhizome were found to contain large quantities of a sub- stance that precipitated with fixatives and stained deeply, obscuring the sections, while the elaioplasts in the rhizomes of Iris versicolor con- tained quantities of “oil” that either was dissolved or was extruded in large masses obscuring the cell structure. In the one or two instances where this did not occur, a good fixation was obtained in mature but not in meristematic cells. The fixation images in sections of rhizome meri- stem were not comparable with those obtained in root-tip meristems, nor could they, as in the case of the root-tips, be identified with structures clearly seen in similar “living” cells. A third difficulty, that might in time have been overcome, lay in the persistent plasmolysis of cells in the rhizome meristem and in leaf tissue. For these reasons it was felt that the fixed material did not give an image of unaltered cells, nor could it be relied upon for comparative work. Better results were obtained with the “vital” technique where dead and dying cells could be observed and where those that survived for some hours without undergoing lethal changes seemed to present a more reliable picture of an unaltered condi- tion. Consequently after some months of unsuccessful experimenting with fixatives and dehydrating reagents and with different hydrogen ion concentrations of single fixatives, the method was entirely aban- doned and the “vital” technique alone retained. Although fixation methods were finally discarded, it should be noted that in certain instances satisfactory results were obtained in this way. Thus the mitochondrial fixatives and stains proved successful for root- tips where they apparently produced little or no alteration in the cell 234 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI structure. Likewise, since chromic and osmic acids fixed the elaioplast “oil,” occasional slides were obtained of mature rhizome tissue quite comparable with that examined “vitally.” Other fixations, although they did not give exact images of the cytoplasmic contents of the cell, proved useful in determining the structure of the oil bodies. The fixa- tives that proved most successful for the occasional rhizome slides were 0.5% osmic acid solution and Flemming’s weak solution followed by Flemming’s triple stain. The most satisfactory of the mitochondrial fixatives was ammonium Erliki solution (25 cc. each of 1% solutions of ammonium and potassium bichromates plus 25 cc. of an 8% solu- tion of formaldehyde) followed by Milovidov’s modification! of Vol- konsky’s stain. With these fixatives the usual dehydrating and paraffin embedding schedules were satisfactory. A third instance of useful fixation was found for the mitochondrial fixatives. These, although not entirely successful except for root-tips, did fix mitochondria throughout the plant sufficiently well for a rough survey of the distribution of these elements. The “vital” method was preferred and finally used exclusively because it presented a more reliable picture of unaltered cell structures. Although this was the main consideration, there were other factors that made the ‘“‘vital” technique especially favorable for the study of devel- opmental and other changes within the cell. Of primary importance was the possibility of observing fluctuating changes of a moment’s dura- tion, as well as those more permanent ones associated with age or season. This was possible only with a technique which left the more or less fluid contents of the cell unchanged. The ‘vital’? method provided such a technique. Another factor favoring the “vital” method was its practical simplicity. Although some skill was required in sectioning, after this was obtained the actual preparation required but a few seconds. Not only was this a saving of time but it was possible to examine material as it was brought in, a method that enabled one to proceed quickly with the study. A third factor of importance was the applicability of the method without modification to all kinds of material. In a comparative study of tissues and plants this was an essential requirement for the technique. As used in this study the “vital” technique was essentially that de- veloped by Bailey (1) for the study of cambium. The material was solution in.; stain gentian violet; differentiate in alcohol. [Milovidov, P.F. Sur les méthodes de double coloration du chondriome et des grains d’ami- don. — Archiv. Anat. Micro. (24), 1:9. 19 1935] FAULL, ELAIOPLASTS IN IRIS 235 sectioned, placed in appropriate solutions and examined immediately and at intervals. For distinguishing the living from injured or dead cells criteria were established based upon comparisons between obvi- ously injured cells and those that survived for some hours before show- ing signs of injury. The only differences in the technique for Jris lay in the details of sectioning and of preparing solutions and in the possi- bility of more firmly establishing criteria for living cells by comparisons with mounts of thin, unsectioned tissue. The sectioning was done with a ‘““Gem” razor blade freehand, or, for some rhizomes, with a Thomson-Spence sliding microtome. Although the microtome sections were more uniform in thickness and more con- venient for mature rhizomes, they were less satisfactory with the other tissues. Apparently a thinner blade produced less injury in rhizome meristems, while it was the simplest means for sectioning leaf, flower or root tissue. The razor blade was used for mature rhizomes also when a microtome was not. available. In either case, sections were obtained varying from one to several cells in thickness. Measured by the micro- tome, sections of mature rhizome varied from 30 yp to 50 wy or more, while those of the smaller-celled meristem were 15 to 20 i or less. The solutions in which these sections were immersed consisted of a basic solution plus one or more of the “‘vital” dyes, or merely of the basic solution alone. Of the three fluids tried, water, nujol and sucrose solution, the sucrose solution in a five to ten per cent concentration, proved most satisfactory. The dyes most commonly used were Shes Red, Janus Green BB, Chrysoidin Y and Benzene-azo-alpha Although Chry- soidin Y is the only one of these dyes tic: stains the elaioplasts, the light staining of the vacuole with Neutral Red throws the cytoplasm into relief and makes its structures more clearly visible. The other dyes in combination with Neutral Red and Chrysoidin Y have a clari- fying effect. None of these dyes stain the immature oil bodies, while the staining of the mature oil bodies by Chrysoidin Y is but temporary. Almost all dyes will stain dead, mature oil bodies. In practice, only traces of the dyes were used (one drop of a concentrated aqueous solu- tion to 25 cc. of sugar solution). Staining is better and more rapid when the sucrose solution is made alkaline with Clark’s buffers (pH 8.2 to 8.6) which shorten the staining period from an hour or more to fifteen minutes or less. Since most stains, even in small amounts, are toxic after a time, sections that it was desired to keep were removed to pure sucrose solu- tions. In this way cells were kept “living” for twelve hours or more. An essential part of the technique was the establishment of criteria 236 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI for distinguishing living from dead or dying cells. By comparing obviously injured cells with those which survive sectioning for some hours without sign of injury, such criteria have been established for cambial tissue. By the same method criteria were found for Iris cells. In addition unsectioned roots, bracts and flower parts of Jris and the thin leaves of a Potamogeton were examined. Living cells in Jris, like those in cambium, are marked by the following characters: regular cyclosis, absence of Brownian movement in the cytoplasm and a stain- ing of the vacuole in the presente of Neutral Red. Two additional criteria were found for living cells of Jris, namely, a pulsation of the cytoplasm in isodiametric cells and the amoeboid movement of the chondriosomes. Both of these phenomena are essentially a swelling or contraction of parts of the structure involved. The pulsation, for example, is the swelling of one part of a protoplasmic thread at the expense of another, a phenomenon involved in changes in the concen- tration of the substance at a given point. The pulsation of the cyto- plasm occurs principally in isodiametric cells where there is no streaming. The amoeboid movement of the chondriosomes may occur in any cell. Both criteria proved valuable as indications of the condition of the cells. Dead cells of Jris, as of the cambium, show one or more of the following characteristics: coagulation of the protoplasm, a general formation of granules in the cytoplasm, staining of the nucleus and cytoplasm in the presence of dyes, increasing opacity of the whole cell and Brownian movement in the cytoplasm. Dying cells in Jris were found to show the following characters: jerky or irregular streaming and Brownian movement within the plastids. Parallel phenomena were found in the irregular streaming and degenerating plastids of some epi- dermal, bract and flower tissues. The validity of these criteria for distinguishing living from dying or dead cells should be considered. The possibility of injury lies in the sectioning, in the action of the solutions in which the sections are placed, in the pressure of the cover glass used in mounting sections and in the strong light used for microscopic observations. In establishing criteria, the use of unsectioned material eliminated the possibility of injury due to sectioning, while the examination of water plants in the water of their natural habitat provided a check upon the effects of the solutions used in the study of Jris. A similar check upon the effects of pressure from the cover glass was provided by removing it. The possibility of injury due to strong light alone remains. That strong light will pro- duce injury and death is clear, but the effects are slow in appearing and, if the light is removed in time, they are temporary. They can be taken 1935] FAULL, ELAIOPLASTS IN IRIS 237 into account in establishing criteria for distinguishing living from dying or dead cells. That there are undetected, instantaneous changes is im- probable, for the reactions in plants are in general slow. The effect of the light appears chiefly in the slowing down of streaming, and, if exposure is continued, unmistakable signs of death such as coagulation of the cytoplasm finally are observed. It should be noted that the observation of minute details of cyto- plasmic structures can be carried on only with the aid of the best high- powered microscopic equipment. For the observation of sections mounted in aqueous media a water immersion objective is essential. Without such equipment, many of the details of structure described in the following section cannot be seen. OBSERVATIONS DESCRIPTION OF ELAIOPLASTS IN RHIZOMES OF IRIS VERSICOLOR Elaioplasts occur typically in the parenchyma of the rhizomes of Iris versicolor. They appear in every cell as granular, highly refrac- tive masses with a decidedly yellowish cast (Fig. 1). The individual elaioplasts are almost spherical in shape and seem to be composed of closely compacted globules approximately one micron in diameter (Fig. 2). They are relatively constant in size within a given rhizome, gen- erally averaging 10 to 13 microns in diameter. Although in some material they may be twice this size, they are never as large as the nucleus which has a diameter of the order of 40 to 50 microns. Often a hundred or more of these elaioplasts will be found in a single cell, aggregated for the most part into one large mass. Sometimes there are as few as twenty to a cell, but often they more than half fill the cell lumen, obscuring the nucleus and protruding into the huge vacuole. All evidence shows that the elaioplasts are located in the cytoplasm. Although they protrude into the vacuole, protoplasmic threads are often observed to spread at their surface as if to include them (Fig. 1). Occa- sionally one is seen moving in the streaming protoplasm. The study of similar bodies in the root, where they obviously are included in the cyto- plasm, substantiates these observations. Microchemical tests indicate that the bodies are mainly lipoid in character. They stain brilliantly in ‘fat’? dyes such as Sudan III, alkannin and nascent indophenol blue.1. They are almost completely 1For this technique see Zweibaum, J. Sur la coloration = graisses dans la cellule vivante. Comp. Rend. Soc. Biol. 1923. — Zweibaum, J. and G. Mangenot. Appli- une m mise a pega la Mae vitale et post vitale des graisses de la cellule végétale. Comp. Rend. Soc. Biol. 1923. 238 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI soluble in lipoid solvents such as ether, chloroform and carbon tetra- chloride. They also dissolve largely in 95% alcohol, a solvent for some oils. They are insoluble in hydrochloric acid, sulphuric acid and potas- sium hydroxide, although they are more or less structurally disorganized by these reagents. They are not volatile at 100°C., which indicates the presence of lipoids rather than essential oils.. The reaction of the elaioplasts to heat and to many reagents in which they are insoluble is marked by the extrusion of the lipoid in drops (Fig. 5). The reaction occurs relatively slowly so that it can be watched. A net-like residue remains which is not distinctly lipoid in character. The drops characteristically remain in contact with the net and are flattened on their attached side. The reaction occurs with heat, picric acid, dilute sulphuric acid, Gram’s solution, etc. Injury to the cell typically produces active Brownian movement of the globules within the limits of the elaioplast which eventually bursts, freeing the globules within the cell lumen. A similar phenomenon occurs in elaioplasts which escape from cut cells. It can be induced by mechanical pressure. STRUCTURE OF ELAIOPLASTS IN RHIZOMES OF IRIS VERSICOLOR The structure of the elaioplasts in the mature rhizome is that of a matrix with embedded globules. This is best shown in sections of fixed material, for the globules in fresh material are so refractive and so closely packed that it is difficult to distinguish any structure clearly. With osmic acid and some of the chrom-osmic fixatives the globules are preserved in situ (Fig. 3b). They clearly show a network of a differ- ent substance between them. With other fixatives the globules are never preserved, but a net-like structure with lacunae of the approxi- mate size of the globules remains (Fig. 3a). This can be seen best by the use of mitochondrial or plastid fixatives such as ammonium Erliki and an appropriate stain. It is well shown, too, by Wakker’s method for double staining elaioplasts with anilin blue and alkannin after fixa- tion of sections in picric acid. In this case, the extruded globules are stained red and the matrix appears as a purple network with blue interstices. The behavior of the bodies in fresh material supports the observa- tions on their structure as seen after fixation. The globules show no tendency to fuse, a fact which indicates a separation by the presence of at least a surface film. In injured material, they move apparently un- changed in a liquid portion of the intact body. Further proof of a matrix is found in developmental forms and in homologous oil-bearing 1935] FAULL, ELAIOPLASTS IN IRIS 239 bodies in other species. Here the matrix is often so abundant as to be clearly visible in untreated material. Such is the case in very young cells of the rhizome, in some cells of the root-tip, in rhizomes of Jris pallida and of Iris Hartwegii, etc. The matrix is also clearly shown in the root during the degeneration of elaioplasts. Here before death the refractive globules disappear leaving only a net with lacunae. This net is very similar in structural appearance, although not in shape, to the net-like image of rhizome elaioplasts in fixed material. The globules were identified as the material which gives the elaioplast as a whole its lipoid characters. They show the reactions previously described for the elaioplasts and additional ones equally characteristic of lipoids. They stain in the “fat” dyes. This is apparent in intact bodies, but it is more clearly seen with the moving globules in disinte- grating ones., They are highly refractive, a property seen in both intact and disintegrated elaioplasts. They disappear from sections treated with “fat” solvents such as carbon tetrachloride, ether, etc., but they may be preserved in sections treated with “fat” fixatives such as osmic acid and chrom-osmic mixtures. They are completely soluble in alcohol. This was demonstrated with globules in suspension in alkaline water. Upon the addition of 95% alcohol a homogeneous fluid resulted indicating the complete solution of the globules. The matrix was shown to be of a different substance from the globules. It appears to be more like the cytoplasm in composition. Unlike the globules it requires no special fixative for its preservation. At least a portion of it is insoluble in alcohol and lipoid solvents such as carbon tetrachloride, for it sometimes remains intact after the use of these reagents. It is not refractive, for this character can be seen in young tissue and in injured cells to be a property of the globules only. Nor is it stained to any extent by the “fat” dyes such as Sudan III, etc. This is evident in elaioplasts with globules in Brownian movement where the stain is largely confined to the globules. That the matrix is of a plasma substance was suggested by the difficulty of staining it differentially from the protoplasm. This view was substantiated later by the identi- fication of the elaioplasts with the plastidome and chondriome. No evidence of a differentiated membrane about the elaioplasts could be found. None could be seen in fresh material, nor has any been brought out by reagents or fixation techniques. The only observation that might be interpreted as indicative of a membrane was the “burst- ing” of injured elaioplasts already described. But no fragments of membrane remained. It is more probable that the sudden freeing of the globules depended upon changes in the matrix which made it miscible with the surrounding medium. 240 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI SEASONAL VARIATIONS IN ELAIOPLASTS IN [RIS VERSICOLOR Certain variations in the form and structure of the elaioplasts are due to the seasonal appearance of starch. In New England the elaioplasts are without starch from early November through June. By July or August the starch begins to appear, while by September or early October a maximum development has been reached. The disappearance of the starch then begins and proceeds rapidly. By early November no traces of it can be found. The starch can be identified with Gram’s solution and polarized light. In the former the grains stain a blue to a bluish-black, a reaction typical of starch in the presence of iodine. In polarized light they appear as bright grains with a black maltese cross on each. The type of starch formation in rhizomes of /ris versicolor is charac- teristic and constant. Each elaioplast develops several included grains (Figs. 4b and c). Counts made in early October showed commonly from 8 to 12 grains, with a recorded range of 1 to 16 per elaioplast. Although the grains are always grouped more or less centrally within the globule-filled portion, they form bulges in the otherwise rounded contour of each elaioplast (Fig. 4). The individual starch grains are approximately isodiametric. They show the central hilum character- istic of this shape of grain when it is included within the plastid. In size they are small, generally 6 to 7 microns in diameter as measured in material collected in early October. Climatic differences in the disappearance of the starch from the grow- ing point of the rhizome are indicated. In material from the vicinity of Boston and of New York the starch disappears completely in the winter. But in plants grown in the Missouri Botanical Garden it may be found about the growing point in March, although completely absent from the rest of the rhizome. The disappearance of the refractive globules of the elaioplasts has not been observed. Numerous observations have been made from Sep- tember to May, during which time they remain in abundance. They are likewise present in the rhizome during June, July and August, although a less thorough study has been made of their behavior during those months. DEVELOPMENT OF ELAIOPLASTS IN RHIZOMES OF IRIS VERSICOLOR By tracing back to the meristem, the elaioplasts in the rhizome were found to develop from mitochondria-like primordia by increase in size, in visibility and in the number of contained globules (Fig. 19). In the youngest cells there are small, irregular, shadowy proplastids with two or three included non-refractive globules. In increasingly older cells 1935] FAULL, ELAIOPLASTS IN IRIS 241 these bodies become more distinct and larger with a greater number of included globules. At the same time the globules become refractive and the whole body even more irregular in contour. Later with further increase in size and in the number of included globules, the irregular contour is lost. The cells then contain the granular, smoothly rounded, mature elaioplasts characteristic of the rhizome. The young elaioplasts are distinguished by the following characters. They have more matrix in proportion to the number of globules than the mature forms. They do not stain after death to any degree in Sudan III nor in any other anilin dyes in contrast to the brilliant stain- ing of the mature elaioplasts. They are restricted to a small region about the growing ‘point, while the youngest stages are found only in the cells of the growing point. They are all irregular in contour, but this irregularity is emphasized in the intermediate forms which are almost nodulose. The youngest stages show characters ordinarily associated with mitochondria. They are about the size of Iris mitochondria, ranging from this up to several times their size. They are indistinctly visible like much of the chondriome with a peculiar fading and reappearing quality. Thus a period of clear definition of these shadowy forms will be followed by a fading and disappearing. This, in turn, after a few minutes or after several hours may be succeeded by another period of clear definition, and so on. In general, although not always, these forms show included non-refractive globules. This is a character shared by the rod-shaped mitochondria of the species. In the young elaioplasts there is no definite arrangement of the globules which in the rod-like mitochondria always form a single row. The formation of starch occurs in any of the young or mature forms of plastids. It was found during the season of its formation in all of them. In the young forms the starch grains ordinarily protrude from the globule-filled mass of the elaioplast, in contrast to the completely included grains of the mature elaioplast. No evidence of increase by division was found in mature or develop- mental stages. No division was seen at any time, although material was collected from September to June and kept under observation for hours at a time. In the rhizome tissue even the ‘“‘dumb-bell”’ figures so often cited as evidence of division were absent. DEGENERATION OF ELAIOPLASTS IN RHIZOMES OF IRIS No evidence of degeneration was found in the rhizomes of Jris versi- color. Elaioplasts are found unchanged and in abundance even in the oldest living cells. 242 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Two isolated cases of degeneration of elaioplasts similar to those in Iris versicolor have been noted in the rhizomes of other species. One of these is an abnormal condition produced in a slowly dying plant. The second is a normal phenomenon in otherwise morphologically un- changed cells. It is apparently unassociated with the death of the cells, for no other signs of degeneration appear. This phenomenon occurs consistently in the cortex of the rhizomes and in the epidermis and sub-epidermis of the roots of /ris macrosiphon var. californica. In /ris macrosiphon, elaioplasts in the mature cells of the cortex of the rhizome appear as large lipoid spheres (Fig. 14c). These spheres are marked by their large size, by their distinct yellow color and by their brilliant staining in “fat” dyes, Sudan III, etc. They may be demon- strated to be in the cytoplasm by coagulating the surrounding proto- plasm with fixatives (Fig. 15). The study of developmental forms which can be seen to be carried in the streaming protoplasm offers fur- ther proof of their inclusion in the cytoplasm. Stages in the formation of the lipoid spheres from mature elaioplasts can be seen in cells not far from the growing point. The process con- sists of the formation of homogeneous spheres by the fusion of the globules and the disintegration of the matrix (Fig. 14). A single elaio- plast resolves itself into one or more of these spheres. In older cells still further fusion occurs for the spheres in them are larger and fewer. In these cells each sphere probably includes the substance of more than one elaioplast. A similar formation of lipoid spheres can be observed in epidermis and sub-epidermis of the root-tip (Fig. 18). The phenomenon is identical with that in the rhizome, although starch is present in the root elaio- plasts. It shows more clearly than in the rhizome the steps in the resolution of the elaioplasts. The fusion of the globules proceeds for some time before the apparent structure of the elaioplast is lost. The final degeneration products include starch grains as well as lipoid spheres. The grains and spheres remain distinct in the cytoplasm, although the starch is indiscriminately scattered among the lipoid spheres. Proof that the formation of lipoid spheres in Jris macrosiphon is a degeneration phenomenon is based on two points. First, the structure of the elaioplast characteristic of the functioning body is lost. There is no evidence that the lipoid spheres can produce starch as the elaioplasts do, or function actively in any way. Secondly, the phenomenon occurs in tissue which tends to die and slough away. In the root, the epidermal cells in which the spheres form are short-lived. This is less evident in the cortex of the rhizome where the cells may live for a season or more 1935] FAULL, ELAIOPLASTS IN IRIS 243 after the formation of the spheres. But even in this tissue the outer cells die and the formation of lipoid spheres is more marked in the outer cells. It is not found in the inner cells of the cortex or elsewhere in the rhizome. A second case of degeneration was found in rhizome cells of /ris tectorum (Fig. 7). In a slowly dying plant the elaioplasts appeared closely compacted in each cell into one or two masses. The rounded contour of each elaioplast was lost, while the matrix seemed to have become more plastic. The identity of each elaioplast was lost in the mass which appeared as a single granular body with indistinct partitions within it (Fig. 7b). Where starch was present the grains were included in the composite mass. This condition has never been found in healthy plants. OIL-BEARING PLASTIDS IN RHIZOMES OF OTHER SPECIES OF [RIS Oil-bearing plastids are found in the rhizomes of practically all species of Jris. They show the same fundamental structure and devel- opment as those in /ris versicolor just described. But they differ from one another in their formation of starch. Two clearly marked types based on the mode and time of starch production occur. The first type is that found typically in /ris versicolor (Figs. 1, 2 and 4). It has already been described. In contrast to the second type, it is marked by the disappearance of starch during the winter dormant season, by the formation of several starch grains in each plastid and by the inclusion of the starch within the plastid. Plastids of this type vary considerably, but they usually show at least two of the general characters. In some species of /ris the starch persists more or less throughout the winter; in others it may persist one season and not the next, and in still others, such as /. versicolor, it always disappears. The inclusion of the starch in mature plastids, although not in the younger forms, is com- plete in most instances. But in some cases the starch grains tend to protrude slightly. This is more often the case in the cortex, although it may characterize the whole rhizome. An extreme case accompanied by an unusually reduced number of lipoid globules in the plastids (Fig. 12d) was found in one of two collections of /ris Hartwegii. The second type is characterized by the persistence of the starch through the dormant season, by the formation of one, large, asymmetric starch grain or sometimes two in each plastid and by a conspicuous pro- trusion of the grain from the globule-filled portion of the plastid (Fig. 9). Caplike elaioplasts attached to one end or side of the large starch grains are typical of these plastids. Often the lipoid globules are larger than 244 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI in the plastids of J. versicolor, while the matrix is abundant enough to be clearly seen between them. A typical example of this type of plastid is found in rhizomes of /ris pallida (Fig. 8). Although the disappearance of starch is not general in this second type, it has been noted in one or two instances. The starch disappears from the main part of the rhizome of Jris pumila during flowering (Fig. 11), although it persists around the growing point. The solution of the starch leaves peculiarly cup-shaped elaioplasts (Fig. llc). A second case of the disappearance of starch may occur under abnormal condi- tions. It was induced in the rhizomes of /ris pallida placed in the green- house during the winter. It is accompanied by a lack of vigor and the disappearance from the rhizome cell vacuoles of substances ordinarily present at that season. The change in the vacuoles is apparent in fixed material in a lack of the precipitate characteristically produced in them by reagents during the winter. The plastids in the rhizomes of the greenhouse plants resembled the spherical ones of winter material of Iris versicolor except for their smaller size and fewer numbers. The distribution within the genus of the two types of rhizome plastids has been found to follow closely the recognized taxonomic grouping. The homogeneous and closely related groups show the same type of plastid, while a heterogeneous group such as the Apogon shows both types. In the latter case aberrancies from the prevailing type in the group are often correlated with anomalous taxonomic characters. Sufficient material has been examined to show definitely the condition in the two largest groups, Pogoniris and Apogon, and in several of the smaller groups, Evansia, Regelia and the Pardanthopsis and Gynandiris species. An indication of the prevailing type in each of the other groups may be found in the notes made on a few representative species. Rhizomes have been examined largely during the late winter. In late winter the pallida type of plastid shows its characters clearly, while the versicolor type is generally without starch at that season. An indi- cation of the mode of starch formation in the latter type of species can often be obtained from the persistent starch in the plastids about the growing point. In this study, such observations have been supple- mented by notes made during the starch-forming season. The pallida type of rhizome plastid occurs characteristically in Pogon- iris, Regelia and Oncocyclus. These are homogeneous groups which together with Pseudoregelia form a unit of closely related species. The same type has been found in a Juno Iris and in a Xyphium Iris. In the latter case it occurs only in tissue about the vascular bundles but not in the large parenchyma cells which are filled with starch. It is also found 1935] FAULL, ELAIOPLASTS IN IRIS 245 in Pardanthopsis, in the closely related hexagona sub-group of the Apogons, in the anomalous Apogon, /. verna, and in one of the variable Apogon spuria group, J. spuria ochroleuca. The versicolor type of plastid appears in the Evansia group and in the majority of the Apogons. In the latter it characterizes the following sub-groups: Sibirica, Laevigata, Longipetala, Californian, Tripetalous, Spuria and Ensata. Its distribution is more limited than that of the pallida type, for the Evansia and Apogon sub-groups include but one- third of the species. The pallida type appears to characterize the other two-thirds of the genus. Since the Evansia and Apogon sections include all of the American irises, the versicolor type is predominantly the type found in North American species. The only exceptions are the anoma- lous Apogons cited in the preceding paragraph and one to be described later. A similar predominance of the versicolor type is to be found in the Asiatic species. The American and Asiatic species contrast in this respect with the more strictly European and Mediterranean species, which belong chiefly to groups showing a pallida type of plastid, notably Pogoniris and related groups and the bulbous forms. The absence of oil-bearing plastids has been noted in five irises. In these cases the rhizome cells are filled with starch. The starch is of two types paralleling in distinguishing characters and in distribution the two kinds formed by the oil-bearing plastids. One is present as large, single, asymmetrical grains similar to those in the pallida type of oil- bearing plastid. They characterize the anomalous Apogon, /ris ungut- cularis, the Reticulata Iris and Gynandiris. The last of these is not always included in the genus; in the formation of starch and no lipoid in its corms it resembles the closely related genus, Moraea. The Reticulata is a group closely related to the Xyphium, which shows simi- lar starch grains and a few oil-bearing plastids of the pallida type. The second starch grain type resembles the starch grains of the versicolor plastids in their small size, in their isodiametric shape and in their forma- tion in groups within a single leucoplast. Like the versicolor type of plastid they are found in Apogon Irises, the Japanese Iris and /ris Sin- tenisii. The former is a hybrid of J. laevigata, one of a group charac- terized by the versicolor type of plastid, and another member of the same group. The other /ris belongs to the Spuria sub-section, a group of intergrading and variable forms, for which no single characteristic type of plastid was found. For the type of plastid found in individual species, the reader is referred to the table on page 246 and also Figs. 12 and 13. The table also includes data on the material, its source, the season of examination, etc. TABLE OF IRISES EXAMINED saa THE TYPE OF RHIZOME PLASTID AND THE SOURCE OF THE ERIAL FOR EACH SPECIES! SEASON SEE PL. 133, SPECIES SECTION Group TYPE SOURCE OF MATERIAL ParTs COLLECTED Fic. 13 I. pallida Lam. X? Pogoniris allid 2 og ners All All Qa I. pallida variety Pogoniris Pallid 2 Mo. sard r, | March — I. Cengialti Amb. ogoniris Pallida 2 Mo. a r, 1] Aarch Of I. pumila Pogoniris Pumila 2 M neg Gard r, | March Qe I. pumila variety Pogoni Pumila 2 Mo. . Gard. r, | March -— I. Korolkowi Regel Regelia 2 Brooklyn Gard. r April Pb I. se 0 Dyk Regelia 2 Brooklyn Gard. r April Pa I.s aL. Oncocyclus 2 Brooklyn Gard. r April O xl. “Zeoannenbuts” Hort. Oncocyclus 2 Mo. Bot. Gard. 1 March — I. alata Poir. Juno 2 Brooklyn Gard. bulb April N I. Xyphium L, Xyphium 2 ee a tenet M I. dichotoma Pardanthopsis 2 Mo. Bot. Gard. r, 1 March R I. foliosa sag & Bush Apogon Hexagona 2 Mo. Bot. Gard. r, | March Ic Mo. Bot. Gard. r, 1 March la I. fulva Ker Apogon Hexagona 2 ‘Beokin Gard, 7 eal x XI. hexagona Walt. X? Apogon Hexagona 2 Mo. . Gard. r, | March — I. vinicolor Small Apogon Hex ; 2 peokiva ae r April Ib I. verna Apogon Ver 2 Virginia r May H I. spuria Pall. var. ochroleuca Apogon Spuria 2 Mo. Bot. Gard rd March Ee I. halophila Pall. Apogon Spuria I Brooklyn Gard r April Ea I. ensata Thunb Apog Ensata I Brooklyn Gard r April & I. setosa Pall. Apogon Tripetalous I Brooklyn d r April Fb I. setosa var. canadensis Foster Apog Tripetalous I Brooklyn Gard r April Fa . . Brooklyn Gard r April Da I. Douglasiana Herb. Apogon Californian I feaeee oF ue. a 5 |: ; 7 Califo r, lo Aug. — dtuiecienana Apaoe Aneraian ' oeetig Gard. r April De I. paca Torr. var. Apogon Californian I California r, l,o Aug. ai I. ten Apogon Californian I ial Gard. r April Db I. aaa ola a rb. Apog Longipetala I Califor r Sept —_ I. missouriensis Nutt. Apogon Longipetala I California r July = WOALAYOdUV AIONYV AHL AO IVNYNOL IAX “TOA ] TABLE (Continued) SEASON SEE Pt. 133, SPECIES SECTION Group TYPE Source OF MATERIAL Parts COLLECTED Fic. 13 I. virginica L. Apogon . Laevigata I Mo. Bot. Gard. r, | March Bb : . {Mo. Bot. Gard. El March Ba I. versicolor L. Apogon Laevigata I \ Boston All All = XI. robusta E. Anders. Apogon Laevigata I Mo. Bot. Gard. Tel March — (Bussey garden i Be I. pseudacorus L. Apogon Laevigata I {Arnold Arbor. r — | Mo. Bot. Gard | March — I. Kaempferi Sie Apogon Laevigata I Brooklyn Gar r April Bd XI. Wilsont W nN x? Apogon Sibirica I Brooklyn Gard r April Ad I. stbirica L. Apogon Sibirica I Mo. Bot. Gard rel March Aa I. prismatica Pursh Apogon Sibirica I eae ne os April BE I. orientalis Mill. Apogon Sibirica I Brooklyn Gard r April Ac XI. “Quest” Hort. pog Sibirica I Mo. Bot. Gard ri March — I. Clarkei Baker Apogon Sibiri I Mo. Bot. Gard r;, March —- I. chrysographes Dykes Apogon Sibirica I Brooklyn Gard r April Ab ‘T. cristata Ait. Evansia I hae l r ae Ja I. gracilipes A. Gray Evansia I N. Y. Bot. Gard r April Jb I. tectorum Maxim. Evansia pce is My I aan I. lacustris Nutt. Evansia I Brooklyn Gard r April Je I. japonica Thun Evansia I Brooklyn Gard r April Jd I. unguicularis Poir Apo 3 Brooklyn Gard. r April G I. reticulata Bie Reticulat 3 any. Gard. bulb April ip i sisyrinchium “a Ceaatice 3 Brooklyn Gard. corm April K I. Sintenisii Jan Apogon Spuria 4 Brooklyn Gard. r April Eb I. laevigata ies "? Apogon Laevigata 4 Mo. Bot. Gard. r March Be —versicolor type of rhizome plastid 2— pallida type of rhizome plastid 3--rhizome plastid er plastid _ iid — starch vera erne he Mil —rhizo t 1 Same iia ern fe "Dy ae pe Genus Iris, SIMI NI SISVIdOIV1S “TINVA [Sol LVS 248 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI OIL-BEARING PLAsTIps IN RHIZOMES, BULBS ETC. OF OTHER PLANTS No oil-bearing plastids have been found in any other rhizomes or bulbs examined. None are present in either of the species of Moraea examined, a closely related genus replacing /ris in the southern hemi- sphere. Nor are there any in the many Araceae, Bromeliaceae, Com- melinaceae, Liliaceae and Scitamineae examined. Rather, all of these plants contain large asymmetric starch grains in their storage organs. The chloroplasts in all of these plants characteristically contain more or less refractive granules. In general, such appears to be the condi- tion in all of the monocotyledons and in many of the dicotyledons. Indeed it seems to be true even of some of the lower forms such as the liverworts and mosses, although in these the granules are often not refractive. The following is a list of the species of monocotyledons examined. The species are grouped according to families. ARACEAE: Acorus Calamus L., Aglaonema sp., Arisaema triphyllum (L.) Schott, Dieffenbachia sp., Nephthytis sp., Philodendron Selloum C. Koch, Philodendron cordatum Kunth, Schismatoglottis crispata Hook. f., Schismatoglottis rupestris Zoll. and Mor., Spathiphyllum sp. BROMELIACEAE: Ananas macrodontes E. Morr., Billbergia sp., Cryptan- thus sp. CoMMELINACEAE: Palisota sp. IRIDACEAE: Moraea iridioides L., Moraea sp. LiviaceaE: Allium sp., Hemerocallis sp., Ornithogalum umbellatum L., Yucca filamentosa L. ORCHIDACEAE: Vanilla planifolia Andr., Vanilla pompona Schiede. Musaceat: Strelitzia sp. ZINGIBERACEAE: Alpinia nutans Rosc., Amonum sp., Hedychium sp. MARANTACEAE: Calathea sp. OIL-BEARING PLASTIDS IN OTHER PARTS OF THE PLANT OF IRIS SPECIES The observations in this section apply to any species of /ris unless otherwise stated. A careful study has been made of the conditions in Iris pallida and in Iris versicolor. Additional notes have been made on other species. The formation of oil globules is characteristic of plastids throughout the tissues of plants of the genus /ris. The globules are not always so numerous as those in the rhizome plastids of Jris versicolor where they are developed to an unusual degree. An extreme example of a limited formation of globules is found in the chloroplasts of the guard cells where 1935] FAULL, ELAIOPLASTS IN IRIS 249 the matrix of the plastids is relatively abundant and clearly visible. Nor are the lipoid globules usually the only observable product of the plas- tids. Ordinarily starch is also present, while in some plastids chloro- phyll or a yellow pigment is formed. The elaioplast condition described for rhizome plastids of the versi- color type may occur in any of the uncolored tissues. It is dependent upon the absence of starch and pigment and upon a large production of oil. Such conditions are found at times in the rhizome, in the root and in uncolored leaf and flower tissues. In the rhizome and root elaioplasts occur generally throughout the tissues of these organs. They are restricted to certain species and, at least for the rhizome, to certain seasons. There is no connection between their presence in the rhizome of a species and their appearance in the root of the same species. For example elaioplasts were found in the rhizome of Jris versicolor but not in the root (Figs. 2 and 35). On the other hand, they were found in the root of /ris pallida but not normally in the rhizome (Figs. 9 and 30). An example of their formation as a seasonal phase of the leucoplasts in the rhizome has already been de- scribed for Jris versicolor. Whether or not they also form a seasonal phase for leucoplasts in the root has not been investigated. In the leaf and flower the elaioplasts are restricted to a few cells. Often they are but transitional forms appearing for a very brief time. Such is the case in the flower where they may occur in the course of the develop- ment of the chromatophores. - Because of their limited occurrence in a few cells it is usually easy to identify them with the leucoplasts or chromatophores in neighboring cells. In these tissues they do not de- velop the brownish color so characteristic of the rhizome elaioplasts in Iris versicolor, Instead they remain entirely colorless. The development of elaioplasts can be induced under unfavorable conditions. An example of this has already been cited in their formation in rhizomes of Jris pallida grown in the greenhouse (p. 244). In this case they were formed by the dissolution of the starch leaving only the oil-bearing plastid. By growing plants in semi-darkness chloroplasts can be prevented from forming pigment or starch. They then appear as elaioplasts. In neither of these cases is an increase in the number of oil globules involved. Nor has the formation of unusual numbers of elaioplasts been observed as a result of abnormal conditions. The oil-bearing plastids in other parts of the plant show essentially the same features as those described for the rhizome. They differ from those in the rhizome in minor characters, also in.the absence of a general elaioplast phase except in the root and in the formation of pigments. In 250 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI addition they show in some cases chondriome types as an intermediate stage in their development from the proplastids. In some tissues, not- ably in the flower and in epidermal tissues, the mature plastids often show further changes involving chondriome types not found in the rhizome. These points will be taken up separately in the succeeding paragraphs. Minor differences between the plastids in other parts of the plant and the type found in the rhizome and root appear in the lesser production of lipoid globules and in their complete lack of color in colorless tissue. Correlated with the smaller number of globules is a greater stability. This is shown in their greater resistance to injury by mechanical pressure and to distortion or destruction by reagents. Their lack of color when not pigmented can be seen in colorless leaf tissue in marked contrast to the strongly yellowish cast in the equally unpigmented rhi- zome plastids. This is particularly well shown in Jris versicolor. The formation of pigment, chlorophyll or yellow pigment, occurs ordinarily in the young plastids. But there is no specific stage at which it is developed. In the leaf chloroplasts it often forms shortly after the appearance of starch in the young plastids, although it may not develop for some time. In the chromatophores of the root it sometimes appears before the formation of starch, for example in the rootcap of /ris versi- color. At other times yellow pigment appears in plastids which do not form starch, for example in the chromatophores of roots of Iris pallida. In many cases yellow pigment is found in chondriocont types of plastids, but its formation is quite unconnected with the phenomena producing these forms. Proof of this is found in its formation in the approximately spherical plastids of the root-tip before they pass into a chondriocont State and in those of the rootcap of /Jris versicolor where the mature plastids retain a more or less spherical state. The location of pigment in the refractive globules and also in the matrix and its greater solubility in the former was demonstrated. In the guard cells of Jris pallida, where there is little chlorophyll, the green color can be seen to be confined to the globules, while the matrix remains colorless. That it may also be dissolved in the matrix is shown in degen- erating chromatophores of the root where color remains in the matrix after the disappearance of the refractive globules. Intermediate developmental forms of the plastids are found in the root-tip (Figs. 20-24). They differ from the small plastids in other differentiating tissue by the retention for a longer period of the shadowy visibility of the proplastids and by a plasticity amounting in the younger stages to an almost fluid character. They resemble other young plastids 1935] FAULL, ELAIOPLASTS IN IRIS 251 in their origin in the proplastids, in their development by increase in size, in visibility, in the number of included globules and in their final development in many cells into the same type of plastid. In their often elongated shape they resemble the chondrioconts of many authors. The shadowy character of the younger intermediate forms is evident in the peculiar fading and reappearing already described for the proplas- tids (p. 241). With the differentiation of the tissue this shadowy quality is lost (Figs. 20-24), but the bodies do not become refractive until a late stage (Figs. 24 and 25). Often the more or less indistinct forms persist for long periods. The plastic quality of the intermediate forms is shown in their more or less elongated shape and in their movement in the streaming proto- plasm. The movement consists of a continuous changing of form (Figs. 20-23). Both movement and elongation are more marked in the younger stages, some of which are almost fluid. In older cells the plastids be- come less and less elongated with increasing viscosity until they are more or less spindle- or tadpole-shaped. At the same time the motion of the plastid becomes reduced to a moving about of the ends. In the fully differentiated plastid the shape is roughly spherical and there is no movement. Often the plastids remain in the spindle- or tadpole-stage for some time. The continuous motion of the intermediate types is essentially an amoeboid movement of the plastid (Figs. 20-23). This appears to some degree in all of the intermediate types. In its most exaggerated expression in the youngest stages, it consists of a change in form from a filament, through intermediate stages, to a sphere. Another example characteristic of the plastids before the globules have become refractive is the formation of two swollen ends connected by a thread. In some cases the thread becomes invisible, but it always reappears and shortens to reunite the two ends. In its least pronounced form in the older spindle- and tadpole-shaped forms, the movement is confined to a turn- ing from one side to another of the tapered ends. That the movement is-not wholly connected with cyclosis, although probably aggravated by it, appears likely. In cells where there is no cyclosis, the intermediate forms customarily show a pulsating movement associated with changes in thickness. An example of this is seen in young plastids in the isodiametric cells of the rootcap. It is worthy of note that in none of these forms has division of the plastids been seen. Many observations have been made at different times and over periods of an hour or more. But even plastids which appear to be divided are seen shortly to be connected by a thread which after a time thickens to reunite the two parts. 252 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI The liquid character of the youngest forms is shown in the movement of the globules within the plastid. This consists of a sloshing about of the globules. In older forms this movement does not appear. Any rearrangement of the globules in them is due to the amoeboid movements of the plastid. The intermediate plastid types develop into leucoplasts in the cortex of the root or into chromatophores in the rootcap. But in other regions of the root and in the elongated cells of the fibrovascular bundles throughout the plant they persist as chondriocont types. The shadowy, very plastic forms are found in the central cells of the root. In other parts of the central cylinder more differentiated forms are found. Simi- lar ones appear in the fibrovascular bundles throughout the plant. In the inner cortex the tadpole- and spindle-shaped forms with refractive globules often remain. The chromatophores of the rootcap often retain their chondriocont-like shape and plasticity after the formation of pig- ment and oil (Fig. 29b). Associated with the persistent developmental forms are shadowy leucoplasts and mitochondrial types not ordinarily linked with plastids. In /ris, the latter are marked by their gradation into the plastid forms. In the same cells with the persistent chondriome-like plastids, they appear as long filaments with a single row of globules and a twisting or wavy motion in the streaming protoplasm (Figs. 26-27). Some of the filaments are shadowy, while the granules in others are refractive. The filaments with refractive granules sometimes show thickened ends which contain more than one row of granules. It may be noted here that in addition to the proplastids and fila- mentous chondriosomes, the more usual types of mitochondria, that is globular or rod-shaped forms, are found in /ris, They appear in all cells but are concentrated about the apical meristems and in tissues of the leaf and root. They are less evident in the cells of the rhizome. Two observations are worth recording. The so-called spherical mitochondria were observed to be more or less fluctuating in form and only approxi- mately spherical. The rod-shaped ones were seen to contain globules. Changes in plastids after ‘‘maturity” are found in epidermal cells and in the flower. The changes in both cases are marked by an increasing fluidity of the matrix and by the disappearance of the refractive globules. In the epidermis of the root and in the flower the changes culminate in the death of the cell. But in the leaf epidermis the plastids remain as more or less shade wy chondrioconts (Fig. 37). These are quite similar in appearance and in motion to the developmental types in the root. The changes in the plastids in the epidermal cells of the root have 1935] FAULL, ELAIOPLASTS IN IRIS 253 been carefully studied. The more or less spherical leucoplasts or chro- matophores first become somewhat elongated (Fig. 31). This is fol- lowed in older cells by increasing fluidity of the matrix and the gradual disappearance of refractive globules (Figs. 31-33). Where the plastids are pigmented some of the pigment remains after the globules are gone, but this also tends to disappear. Where these changes progress far, the plastids become shadowy nets much elongated in shape (Fig. 33). In the streaming protoplasm, they are often partially drawn out into long filaments (Fig. 32). In some cases the attenuated portions are bent back upon the rest of the plastid so as to include a small amount of protoplasm (Figs. 30, 33, 34). When an oil globule is included with protoplasm it is often in Brownian movement. The attenuated forms persist until the death of the cell. During flowering the leucoplasts of the floral tissue were seen to undergo a series of changes similar to those described as occurring in the root. These have not been studied in detail, but the following general changes have been noted. The leucoplasts become pigmented and chon- driocont-like in shape. As the flower opens and fades the chromato- phores become more and more fluid. At the same time the refractive globules disappear. An advanced stage shows them partially drawn out into long filaments (Fig. 34). Unlike the chromatophores of the root the pigment is retained. These forms remain until the death of the cell. Similar changes occur in the leucoplasts of the bracts. In Iris Xyphium the formation of refractive bosses on the pollen grains from oil-bearing chromatophores has been demonstrated (Fig. 17). In unripened anthers the pollen shows no markings except the small refractive dots forming a part of the wall structure. The grains are surrounded by the tapetal fluid in which are numerous oil-bearing chro- matophores. In the shed pollen the grains show not only the refractive dots but the closely appressed chromatophores which appear as granular, yellowish, refractive bosses. No similar observations have been made on pollen grains of other species. Although the latter show refractive spines or a network of refractive structures, these are in every case asso- ciated with wall formation. They are in no way connected with the plastidome or chondriome. All of the plastids described are found to show the following general characters. They tend to aggregate about the nucleus, a character also shown by mitochondria. Unless degeneration is involved they retain the ability of the proplastids to form the pigments and other products differentiating the different types of plastids. They also retain the ability to change from the plastid shape into a chondriocont form and vice versa. 254 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XVI ELAIOPLASTS IN PLANTS DESCRIBED IN THE LITERATURE Of the plants recorded in the literature as forming elaioplasts the following have been examined: Vanilla planifolia Andr., V. Pompona Schiede., Marchantia polymorpha L., Lunularia cruciata (L.) Dum., Pellia epiphylla (L.) Corda, Porella sp., Bazzania trilobata (L.) S. F. Gray, Scapania nemorosa (L.) Dum., Cephalozia sp,, Trichocolea tomentella (Ehrh.) Dum., Plagiochila asplenioides (L.) Dum., Lopho- colea heterophylla (Schrad.) Dum., two thallose species of the Junger- manniales from Oregon and two leafy species of the Jungermanniales from Oregon. The two classes of elaioplasts described by Pfeffer, Wakker and later writers were examined. These are the oil bodies characterizing the liverworts and those in Vanilla, a classical example of elaioplast-bearing monocotyledons. In both cases certain of the observations of previous writers have been verified and some additional notes made. The following observations made by earlier writers for Vanilla have been verified. The elaioplasts are present as highly refractive granular bodies near the nucleus in cells which contain also leucoplasts and chloroplasts. Structurally they consist of globules of refractive oil in a protein or plasma matrix. They are marked by their brilliant stain- ing in “fat” dyes and by the extrusion of large globules of oil after treat- ment with various reagents. In addition it has been noted that the elaioplasts are generally dis- tributed in all the cells of leaf, stem and root tissues rather than re- stricted to particular tissues in certain parts of the plants. It has also been observed that the single large elaioplasts are aggre- gates of smaller granular bodies (Fig. 45). The aggregation is more or less compact. In some cells it is difficult to distinguish the individual bodies, while in other cells they are but loosely grouped or freely circu- lating in the streaming cytoplasm (Fig. 46). In some cells the smaller bodies could be observed to aggregate into one or more groups from which individuals were carried away from time to time by the streaming protoplasm. The development of the smaller bodies from non-refractive granular ones can be observed in younger cells of leaf and root. In successively older cells the included globules gradually become more and more refrac- tive until the bodies assume the highly refractive condition typical of mature cells. In the less refractive stages the bodies seldom form com- pact aggregations. No specific stage has been noted in which aggrega- tion becomes the rule. The formation of compact groups appears possible at any time, although more characteristic of mature tissue. 1935] FAULL, ELAIOPLASTS IN IRIS 255 The rotary movement of elaioplasts described by Zimmermann and others as characteristic of these bodies has been shown to be an injury phenomenon. It is observed in cells which soon show unmistakable signs of injury followed by death. It is not seen in any cells which remain normal in appearance and actively streaming for a period of hours. The movement consists of rotation within a liquid vacuole. It is followed by Brownian movement of cytoplasmic inclusions and a general coagulation or disintegration of the cellular structure, that is by unmistakable signs of death. In the liverworts the following observations of earlier writers have been verified. Bodies included within the cytoplasm and marked by their refractivity, by their staining in “fat” dyes and by their solubility in alcohol appear generally throughout the group. They are located within the ring of chloroplasts, but, unlike those in Vanilla, show no particular affinity for the nucleus. They all characteristically leave a residual ring in solution with alcohol, etc. They vary in color from colorless to dark brown. Two or three classes are distinguishable. The first appears as a single large granular mass almost filling the cell lumen (Fig. 38). It is located in scattered cells throughout the plant body and is characteristic of the Marchantiales. The second and third types are found in the Jungermanniales which they characterize. They are smaller than those in the Marchantiales and are round, spindle- or disc-shaped in form (Figs. 40-44). They grade from a homogeneous type to a very granular one. Commonly there are from one to twenty in a single cell, located more or less characteristically in the peripheral cytoplasm. In this group they are not restricted to particular cells but are found in every cell. Unlike the bodies in Vanilla there is little or no tendency for them to aggregate. The development of the bodies has been observed in the Junger- manniales (Fig. 42). In the younger cells the oil-bodies appear as shadowy, wrinkled, granular bodies. They develop into the mature bodies of older cells by an increase in substance and in the refractivity of the granules. By the time the cells are fully mature, the bodies have become plump and refractive. There is no indication of a vacuolar. origin postulated by some writers. In addition the following new observations were made. The homo- geneous type found in the Jungermanniales are sometimes seen with attached granular bodies (Fig. 44). These appear in the younger cells. The single bodies in the Marchantiales can be shown to be aggre- gations of smaller ones. This is apparent in younger cells where they are less refractive and less highly colored (Fig. 39). In older cells the 256 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI structure is obscured by the dark color. Likewise in older cells the bodies appear to be more closely compacted. The Brownian movement described by some as characteristic of the bodies in certain species has been shown to be associated with older bodies or with injury. It is never seen in younger tissue, even in cells with mature oil bodies. It appears in some of the older cells of the Marchantiaceae and can be induced in any cell by injury. DISCUSSION It has been shown in the preceding observations that the oil bodies in /ris are a phase of ordinary plastids. In studying the development and variations of these plastids, many interesting observations have been made which have a bearing upon the status of elaioplasts and upon various problems concerning plastids. In particular the observations provide further evidence of the plastid character of elaioplasts and of a relationship between the various types of oil bodies described in the literature. They also clarify our conception of the interrelationships of plastids and chondriosomes. 1. SIGNIFICANCE OF PRESENT STUDY IN THE INTERPRETATION oF ELAIOPLASTS To identify the anomalous bodies in /ris as a seasonal elaioplast phase of plastids adds another instance to the accumulating evidence of the plastid character of oil bodies. This substantiates the theories of Wakker (47), of Beer (4), of Hieronymus (22) and of Kozlowsky (28) who postulate a relationship with plastids rather than with vacuoles or with the nucleus. There is no evidence in any of the observations de- scribed in this paper of a vacuolar origin or identity. On the contrary, the structural, developmental and chemical similarities between vacuoles and oil bodies recorded by some authors were not observed in any of the material examined. Nor was there any evidence of a nuclear deri- vation of the elaioplasts, a theory based upon the similarity in the staining properties of the nucleolus and elaioplasts and in the aggrega- gation of the elaioplasts about the nucleus. Both of these phenomena have been found to be characteristic of plastids in general. The possi- bility remains that some elaioplasts may be more or less fused aggrega- tions of oil globules which bear no relationship to plastids. The phe- nomenon was not observed, but the possibility of its occurrence was not disproved. . It is probable that the granular elaioplasts of the monocotyledons and liverworts are types of plastids. They show the same structure as 1935] FAULL, ELAIOPLASTS IN IRIS 257 that of the plastids, that is a matrix with embedded globules. That the stroma in the liverworts is non-fixable is not significant morpho- logically, although it indicates a chemical difference between the oil bodies in the liverworts and plastids in general. Further evidence of the plastid character of the granular oil bodies in the monocotyledons is found in their similarity in appearance and in general characters to those found in Jris, A comparison between the elaioplasts in Vanilla, as a classic example of the type found in monocotyledons, and those in Iris shows the following characters common to both: presence of refrac- tive granules, brilliant staining in “fat” dyes, extrusion of oil with picric acid, etc., aggregation about the nucleus, yellowish color, plastid struc- ture and the absence of the more usual plastid products such as starch and pigment. That the homogeneous oil bodies in the liverworts may be classed like the granular types as plastids is suggested. Heretofore no distinc- tion has been made between the two types because of the intergradation occurring between the two extremes. The appearance of attached granular portions in the younger stages of the homogeneous forms sub- stantiates the view that they should be classed with the granular types which, as has been suggested, are plastids. That elaioplasts are sometimes a phase of functional plastids as well as degenerate forms has been brought out in these studies. Heretofore they have been considered to be degenerate forms or secretions of plas- tids. In /ris they are found as functional plastids, as evidenced in the formation of starch and their apparently continuous presence in indi- vidual cells from season to season. That elaioplasts sometimes form by degeneration of plastids involving the production of oil has been shown by Beer (4). There is no evidence that they are ever secretions from plastids. It is probable that the granular elaioplasts described in the literature are sometimes functional plastids and sometimes degenerate forms. Those found in such organs as leaves, roots and bulbs or those found widely distributed throughout the plant as is the case in Vanilla are doubtless active plastids, while those restricted to the more or less evanescent floral tissues are probably degenerate plastids. The interpretation of the homogeneous oil bodies in the liverworts is not clear. They might be degeneration products, but they might also be an accumulation of normal plastid products within a plastid. It has also been shown in the studies of Iris macrosiphon that elaio- plasts of the type described by Lidforss (32) as homogeneous oily spheres may form by the degeneration of oil-bearing plastids. A similar 258 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI phenomenon has been described by Beer (4) as a final step in the degen- eration of plastids in floral tissue of Gaillardia. That the spheres de- scribed by Lidforss (32) are likewise degeneration products of plastids can only be surmised. It is possible that they are more or less fused aggregations of oil globules unconnected with plastids. Evidence of a relationship between the various types of elaioplasts described in the literature has been found in these studies. A structural similarity is seen between the oil bodies in the Marchantiaceae and those in Vanilla in that they are both aggregations of plastid-like bodies. I have found no record in the literature of the aggregation of these bodies in the Marchantiaceae, although the phenomenon was noted for elaio- plasts in the monocotyledons as early as 1914. In addition to this direct evidence of structural similarity the observations upon the devel- opment and variations of elaioplasts in Jris have demonstrated that these oily plastids show, at one time or another, the widely varying phenomena which have heretofore been considered distinctive of differ- ent types of oil bodies. It has already been pointed out that elaioplasts of the homogeneous type described by Lidforss (32) sometimes result as a degeneration product of a granular type of elaioplast. It has also been found in /ris that the oily plastids show at one time or another the following phenomena described in the literature for oil bodies: aggrega- tion and fusion of homogeneous oil globules, aggregation and compact- ing of plastid-like bodies, aggregation about the nucleus, unrestricted position in the cell, degeneration involving the disappearance of the oil and degeneration involving the formation of oily spheres. In brief the morphological distinctions between the various classes of oil bodies appear to be breaking down, while it is evident that plastids can show widely varying phenomena which, considered separately, might be inter- preted as bases for the distinction of fundamentally different types. Further study on this subject is highly desirable. In particular further observations on oil bodies and plastids in Vanilla, Ornithogalum and the hepatics are needed, for much of the literature deals with elaioplasts found in them. An additional point which tends to reduce the number of recorded distinctions between the oil bodies in the liverworts and those in the monocotyledons appears in the permanent character of the elaioplasts in the rhizomes of /ris. Heretofore elaioplasts in monocotyledons have been described as temporary structures, while those in the hepatics have been thought to be more permanent. It may be noted here that my own limited studies made on Vanilla indicate that elaioplasts are not the temporary structures even in this classical plant that one would infer from the literature. 1935] FAULL, ELAIOPLASTS IN IRIS 259 That conditions producing oil bodies are more or less restricted in their occurrence in the monocotyledons has again been brought out in these studies. Elaioplasts do not appear generally throughout the group, although the appearance of oil-bearing chloroplasts is not uncommon. This study adds another genus and many species to the published lists of monocotyledons in which elaioplasts occur. Although oil-bearing plastids occur in the rhizomes of practically all species, it is noteworthy that the elaioplast condition is restricted for the most part to the Apogon irises of Asia and America. This is the first record that I find of the occurrence of oil bodies in rhizomes, although Politis (39) has described them in bulbs. Evidence of the function and significance of the oil bodies has been found. In Jris the bodies are clearly assimilative organs as shown by their formation of starch. That the oil itself is a reserve food supply is indicated. In certain species it replaces at least morphologically the starch stored in the rhizomes of other species. There is no evidence that the elaioplasts are ever excretions, although they may be at times degeneration products. There is no evidence of the division of elaioplasts recorded by a few writers. The fragmentation described by Raciborski (41) and Politis (39) is but the separating of the aggregated plastid-like bodies. This can be seen in Vanilla. That there is ever a passive division of an aggre- gated mass of oil bodies by the cell wall is improbable. Neither such aggregations nor a great development of oil was found in the meristems of Jris, Vanilla and the hepatics. 2. SIGNIFICANCE OF PRESENT STUDY IN INTERPRETATION OF PLASTIDS AND CHONDRIOSOMES With the recognition of elaioplasts as plastids, a study of their varia- tions became a study of the variations in plastids and chondriosomes. No new phenomena have been noted, but significant interpretations of those already recorded in the literature' have been made. Most striking of the phenomena observed was the development of large quantities of oil globules in plastids. The formation of oil globules in plastids has been known for a long time and has recently been empha- sized by Guilliermond’s (15-20) studies of /ris. But even Guillier- mond’s extensive investigations have not shown an accumulation of oil in plastids comparable to that found in Jris versicolor where the quantity is ‘A summary of the present status of plastids and chondriosomes may be found in books and papers by Schiirhoff (44), Sharp (45), Guilliermond et al. (20) and Mottier (37). 260 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI so great as to obscure the structure of the plastids and render them unrecognizable for months at a time. The association of oil globules with young or degenerating forms more frequently than with mature plastids has been suggested by Guilliermond et al. (20). But such is not the case in Jris where the largest formation of oil is in the functioning plastids of the rhizomes. A second phenomenon noted was the plastic quality of chondriosomes and of transitional types of plastids. As evidenced in amoeboid move- ments this has often been recorded in the literature, while it has been emphasized in the recent studies by Guilliermond and his associates. But I have found in the literature no reference to the extreme plasticity amounting to fluidity such as occurs in some young leucoplasts where the included globules are moved about at random within the plastid. The significance of the chondrioconts has been brought out clearly in the survey of the variations of plastids and chondriosomes made in this study. The chondrioconts are essentially plastids producing at times all of the visible products of plastids such as starch, chlorophyll or a yellow pigment. They share, too, the plastic qualities of plastids which they display to a much greater degree. They occur in restricted tissues as transitional stages in the formation of plastids from mito- chondria-like primordia or as more or less degenerating forms of plastids. Often in the rootcap and in floral tissue they are pigmented, although the formation of pigment is not confined to them. It should be noted that in some tissues chondrioconts persist without assuming the more usual plastid form. Chondrioconts should not be interpreted as invariably forming a stage in the development of chloroplasts [Guilliermond et al. (20) ]. On the contrary my studies show that the majority of chloroplasts and other plastids develop from mitochondria-like primordia without the intervention of a chondriocont stage. Where chondrioconts do form a stage in the development of plastids, the whole chondriocont develops into a plastid. There is no budding or fragmenting of the chondriocont involved. The appearances that have been interpreted as budding in chondrioconts or as evidence of fragmentation are but temporary shapes of the plastic chondrioconts. It may be noted here that the studies of chondrioconts emphasize Kassmann’s (25) observations that plastids do not divide under normal conditions. This is a much debated point in the literature upon plastids. There was no evidence of vacuole formation in degenerating plastids or chromatophores such as have been described in flowers [{Guilliermond et al. (20)]. The appearance which has been interpreted as a vacuole is 1935] FAULL, ELAIOPLASTS IN IRIS 261 rather the inclusion of a small amount of protoplasm as a result of the amoeboid movements of the plastid at this time. It is worth emphasizing here that the complete degeneration of the plastids may occur without involving the death of the cell. It has already been noted by Beer (4) that such a phenomenon occurs in some floral organs where the life of the mature cells is comparatively brief. I have found no record, however, of the degeneration of the plastids in cells which remain alive for months thereafter, a phenomenon found in the rhizomes of Jris macrosiphon. In general, it may be stated that there is no sharp line of demarcation between elaioplasts, plastids, chondrioconts and mitochondria. In /ris they have all been observed to form starch and, with the exception of mitochondria, chlorophyll, oil and a yellow pigment. In some instances several of these products may appear at once, or they may develop in succession, or none of them may form. Nor should any of the chon- drioconts and plastids be considered end products of a developmental series originating from mitochondria-like bodies, for until irreversible changes occur such as a resolution into structureless spheres of oil, the shapes assumed are reversible. In other words there is no clear dis- tinction between amyloplasts, leucoplasts, chloroplasts, chromoplasts and elaioplasts; nor is it possible to consider plastids, chondrioconts, pro- plastids and mitochondria as unrelated cell structures. Rather it appears that these are all forms of the same fundamental cell organ differing only in size and in the chemical products being formed at the time. 3. SIGNIFICANCE OF THE STUDY OF OIL-BEARING PLASTIDS IN IRIS ROM A TAXONOMIC VIEWPOINT The occurrence of two types of plastids in rhizomes of /ris each more or less restricted to certain groups of species appears to be of taxonomic significance. The consistent appearance of the same type in well defined species indicates a character that may be useful in separating species. In addition it should be noted that the substitution of compound starch grains for the elaioplasts in rhizomes of a known hybrid and in one or two questionable species, although not an invariable phenomenon, sug- gests a possible means of identifying some plants as of hybrid origin. CONCLUSIONS 1. The anomalous bodies in the rhizomes of /ris versicolor are an elaioplast phase of leucoplasts persisting throughout the resting season, but forming starch throughout the actively growing period. 262 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI 2. Some, if not all, of the so-called “elaioplasts” are plastids in some form or other. 3. Elaioplasts of the plastid type are not necessarily degeneration types: in /ris they are functional plastids. 4. The rotary movement of elaioplasts described in the literature is an artefact due to slow death or injury; the Brownian movement de- scribed as characteristic of globules in certain liverworts is a degenera- tion or injury phenomenon. 5. The elaioplasts in Lunularia and Vanilla are morphologically similar in that they are aggregations of small plastid-like bodies that form oil. This establishes another link between elaioplasts in the liver- worts and those in the monocotyledons. 6. There is no sharp line of demarcation between the different kinds of plastids and chondriosomes each of which is a more or less temporary form capable of changing to the other types. 7. At all times the plastids are more or less plastic but particularly so in young tissues, fibrovascular tissue or slowly dying cells. 8. Leucoplasts, chloroplasts and chromatophores do not go through a set series of changes in developing from plastid types characteristic of meristematic tissues. They may pass through various series depending upon the type of mature tissue involved, or they may merely increase in size with probable changes in their physico-chemical structure. They never form by budding of chondrioconts succeeded by separation of the buds so-formed. 9. Chondrioconts may form an intermediate developmental stage in the formation of “mature” plastids, although not necessarily; they may persist in some tissue; or they may be an intermediate stage in the degeneration of plastids. 10. Plastids and chondriosomes in /ris all show the structure of a matrix with embedded globules. Pigments are more soluble in the globules than in the matrix, although they are found in both. 11. Two types of degeneration of plastids occur involving (a) an increasing fluidity and a decreasing refractivity or (b) a complete breaking down into large homogeneous spheres of oil. Degeneration of the plastids does not necessarily involve the death of the cell. 12. The formation:of a vacuole with at times an included oil drop in degenerating chondrioconts is in reality an inclusion of protoplasm. Different species of /ris are characterized by distinct types of elaio-leucoplasts in their rhizomes. The distrjbution of types follows closely the taxonomic groupings and may be of significance in separating species. 1935] FAULL, ELAIOPLASTS IN IRIS 263 14. The occurrence of such elaioplasts as those in rhizomes of /ris versicolor is confined, so far as could be ascertained, to rhizomes of this genus. For the most part they are restricted to rhizomes of certain species, chiefly Apogons of Asia and America. 15. Refractive bosses on pollen grains of Iris Xyphium are oil- bearing chromatophores adhering from the tapetal fluid. Other mark- ings found on pollen grains were part of the wall structure. ACKNOWLEDGMENTS I am under deep obligations to Professor I. W. Bailey of Harvard University for his continuously helpful suggestions and interest through- out the course of the investigations recorded in this paper. I also wish to thank Dr. George M. Reed of the Brooklyn Botanical Garden and Dr. Edgar Anderson, recently of Washington University and the Missouri Botanical Garden for the material which they kindly placed at my disposal. Dr. Anderson also aided materially in connection with pertinent taxonomic matters. I wish to express my appreciation too of the advice and material offered by Dr. Conway Zirkle and others. Most of the work was done in Professor Bailey’s laboratory in Har- vard University. I spent one summer in Professor Bailey’s laboratory at the Stanford Laboratory of the Carnegie Institution of Washington, Division of Plant Biology, Palo Alto, California, and several weeks at the Missouri Botanical Garden in St. Louis. I am grateful to these institutions for the facilities that they afforded. LITERATURE CITED 1. Bartey, I. W. The cambium and its derivative tissu V.Ar naissance of the vacuome in living cells. (Zeitsche, Zell. Mikr. Anat. 10:651-682. 1930.) 2. ———— and C. ZIRKLE. ae cambium and its derivative tissues. VI. The effects of hydrogen ion concentration in vital staining. (Jour. Gen. Physiol. 14: 363- 383, 1931. 3. ————, C. Zirkte and A. Fautv. Investigations on the cam- bium and its derivative tissues. (Carnegie Inst. Year Book, 30: 1. 19 E (Ann. Bot. 23:63-72. 1909.) . CHALAUD, G e cycle évolutif de Fossombronia pusilla Dum. (Rev. én. Bot. 41 4. 95, 129, 213, 293, 353, 409, 474, 541, 606, 676. un ve) ie3] nw 73, FOO 5 e. oes ) 2. 0) wn 2g ,W . Dancearp, P. A. Recherches Aa la structure de la cellule dans les Iris. [Compt. Rend. Acad. Sci. (Paris) 174:1653-1659. 19 . Domsray, P. Contribution a |’ étude des corps Grane des hepat- iques des environs de Nancy. Nancy, 1926. . DyKes, W. nr The genus Iris. Cambridge, 1913. . GARGEANNE, A. J. W. Die Olkorper der Jungermanniales. (Flora, 92:457-482. 1903.) On N WN JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI . GAVAUDAN, P. Sur l’origine et les caractéres des éléments oléiféres 5 des Jungermanniales. [Compt. Rend. Acad. Sci. (Paris) 184: 1473-1475. 1927. ur les rapports du vacuome et du systéme oléifére des Jungermanniaceae. P Crbid, 186:163-166. 1928.) Sur la presence du systeme oléifére dans les organes de multiplication des Jungermanniaceae. (Ibid. 186:886-888. 1928.) . GOTTSCHE, Anatomisch-physiologische Untersuchungen tuber Haplomitrium “Hookeri. (Verh. Leop. Carol. Akad. 12(1) :287- 288. 3 . Grout, A. A.J. Mosses with a hand lens. 3rd Ed. New York, 1924. ; ce ae: A. Sur l'étude vitale du ee = Ws nated des petales d’Iris germanica et son évolution en t chro plastes. [Compt. Rend. Soc. Biol. (Paris), 7a: 12801283, 1913.] Nouvelles observations vitales sur le chrondiome des cellules épidermiques de la fleur d’Iris germanica, I. (Ibid. 78:241-245. 1 lles observations vitales sur le chondriome des cellules , bi ————. Nouve s ‘ épidermiques de la fleur d’Iris germanica, II. (Ibid. 78:245-249. ) 1915. r l’origine et la signification des oléoplastes. [Compt. 1922. ] Rend. eg "Biol. (Lyon), 86:435. ——-— Su ur la présence d’un corps d’aspect oléagineux dans les cellules épidermiques d’Iris germanica [Compt. Rend. Soc. Biol. (Paris), 109:1182. 1932. ] ———,, O. Mancenot et L. PLANTEFoL. Traité de cytologie végeé- tale. Paris, 1933. A. . Harper, R. Nature and functions of plastids, especially elaio- 316. plasts. [Proc. Internat. Congr. Plant Sci. (Ithaca), 1:311- 1916. ] . HreronyMus, O. Beitrage zur Morphologie und Biologie der Algen. 9.) (Cohn’s Beitr. 5:467-468. . Hormeister, W. Die Lehre von der Pflanzenzelle. Leipzig, 1867. . Horie, O. Uber die Zellenblaschen der Lebermoose. (Verh. Leop. Carol. Akad. Heidlb. 1857, p. 396. 1857. ) . KASSMANN, F. ie Entwicklung der Chondriosomen und Chloro- plasten von Cabomba aquatica und Cabomba caroliniana auf Grund von Dauerbeobachtungen an eas Zellen. (Planta, 1:624-656. 26.) . Keen : M. L. Zygospores in Sporodinia grandis. (Ann. Bot. 28: eene, M. 455-470. 1914.) Studies of zygospore formation in Phycomyces nitens Kunze. (Trans. Wis. Acad. Sci. Arts and Letters, ] 19:1195-1220. 1919.) feuilles. [ Compt. Rend. Acad. Sci. (Paris), 173:497-499. rok . Kuster, W. Die Olkorper der Lebermoose und ihr Verhaltnis zu den Elaioplasten. ae l, . LEHMANN, C. Studi n iiber den Bau und die Entwicklungsgeschichte von Olzellen. (Pian 1:343-373. 1925. . Lerner, B. Untersuchungen ttber das Olplasma und die Oleoplasten. 19 (Jahrb. Phil. Fak. II. Bern, 4:95-113. . Liprorss, B. Studier ofver Elaiosferer i Ortbladens Mesofyll och Epidermis. (Kgl. Fysiogr. Sallsk. i Lund Handl. 4:1-33. 1893.) 1935] FAULL, ELAIOPLASTS IN IRIS 265 wn _ : ai a A. N. Uber farblose Oelplastiden und die mcg ete (Bot Bedeutung der Oeltropfen gewisser Potamogeton-Arten. Centrbl. 35:177-181. 1888. A. Morphologische und physiologische Analyse der Zelle. 0-19 . MIRBEL, Recherches anatomiques et ede pepe hag sur la Marchantia polymorpha. (Mém. Acad. Roy. Sci. Inst. France, 13:345. Si . Monteverpe, N. A. Uber Ablagerung von Calcium- und Magnesium- Oxalate in der Pflanze. (Bot. Centralbl. 43:327. 1890. . Morrtier, D. Chondriosomes ] the primordia of leucoplasts. 18. (Ann. Bot. 3291-112. 1918.) . PFEFFER, W. Die Olkérper der Lebermoose. (Flora, 57:2-6, 17-27, 33-43 . Porrtis, I. Sugli elaioplasti nelle mono- et dicotiledoni. (Atti Inst. 5-361. 191 Bot. Pavia, ser. 2, 14: . Popovict, ie Quelques remarques sur le développement des elai- oplasts ‘des hepatics. [Compt. Rend. Acad. Sci. (Paris), 185:77-80. 1927 ; RACIBORSKI, M. Uber die Entwickelungsgeschichte der sealer ee 93 bei Liliaceen. (Bull, Acad. Sci. Cracovie Compt. Rend. 18 259-271. 1893.) . Rattray, J. Observations on the oil bodies of the Jungermanniaceae. 6. (Trans. and Proc. Bot. Soc. Edin. 16:123-128. . Rivett, M. F. The structure of the cy racy: in the cells of Alicularia 1918.) scalaris Cord. (Ann. Bot 2.:207-21 . ScutruorF, P. N. Die Po. at 1924. . Suarp, L. W. Introduction to cytology. 2nd ed. New York, 1926. ‘ TUNMANN, O., & L. RosentHALER. Pflanzenmicrochemie. 2nd ed. . Waxkker, J. H. St udien iiber die Inhaltskorper der Pflanzenzelle. :423-496. 1888. : ange H. Uber Calciumoxalat in den Pflanzen. Marburg, 1892. p, M. On some cell contents in coffee and other plants. (Nature, 28: 380, 1883.) Wovycick!, Z. Sur les crystalloides des noyaux et les el chez Ornithogalum caudatum. (Bull. Internat. olon. Sci. Math. et Nat. Ser. B. Sci. Nat. (1) (Bot.) io 38-99. 1929. ZIMMERMANN, A. Uber die Elaioplasten. (Beitr. Morph. u. Physiol. Zelle, 1:185-197. 1893.) DESCRIPTION OF PLATES s. 1-4, 6, 8-10, 38, 39, 45, 46 were made with a camera lucida; the Fig magnifications given for these figures are exact. The other figures were drawn free-hand; the magnifications given for them are approximate. Fig. PLaTE 132 1. Iris versicolor L. ee cells of the rhizome from material collected in December. 75. Individual elaioplasts ae cells shown in Fig. 1. Xx 1600. Individual elaioplasts from cells shown in Fig. 1 after treatment with (a) ammonium Erliki fixative, and erythrosin and cyanin; and (b) 0.5% osmic acid. xX 1600. W bo on NO Ko) _ _ —" bo _ w JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI . Iris versicolor L. Individual elaioplasts from rhizomes collected O in October: (a) surface view; (b) included starch grains; (c) diagram to illustrate position, size, number and shape of starch grains. 600. . Iris versicolor L. Individual elaioplasts treated with Gram’s solution: (a) material from Duxbury, Mass.; (b) material from Lincoln, Mass. These were drawn at the same magnifica- . Isolated starch grain from elaioplast shown in Fig. 4. x 1600. Iris tectorum Maxim. Elaioplasts from young cells of rhizomes collected in March: (a) normal plant; (b) dying plant. x 845. PLATE 133 Iris pallida Lam. X? Living cell of the rhizome from material collected in December. xX 475. Individual oil-bearing plastids from cell shown in Fig. 8 showing plastids: (a) without starch; (b) with starch; (c) is a diagram showing the relative positions of plastid and starch. 1600. : ia ag starch grains from plastids similar to those shown n Fig. 9, ; Iris pumila L. Oil- pen hag i from living seo ofa po zome — at St. Louis in March: .. ae ) and (c) ar plastids from successively ater cells. ‘ Oi cep ‘plastids ay living cells of piwecen of the follow- g California species of /ris: AS he missouriensis Nutt.; (b) J. Fay ae eens Herb.; (c) J. lor gipetala Herb.; (d) J. Hartwegii er. aterial collected in Califor nia in August. X 1270. . Diagram showing the types of oil-bearing die found in [ris species in March. See table p. 246 for names PLATE 134 . Iris macrosiphon Torr. Elaioplasts in sie cells from - ve cortex of a rhizome collected in July: ae nd (c) ar taken from successively older cells. . Tris Xyphium L. Untreated pollen grains in surface i Pp view: (a) in tapetal fluid of immature anther; (b) from a ripened anther. x 475. PLATE 135 » Iris oe Torr. Oil-bearing plastids in living epidermal cells of r Bee (a), (b), (c) and (d) are from successively pies aia 128 . lits pci a L. *Piaioplasts from living cells of the meristem of a rhizome: (a) from one of the youngest cells; (b) and (c) from successively older cells. x 1620. . Iris pallida Lam. X? Plastids from living, elongated, differen- tiating cells of root-tip: (a) successive observations on a single plastid to show fluctuating variations in form; (b), (c), (d) ae (e) similar observations on four additional plastids. X 1620. . Plastid similar to those in Fig: 20 but from an older cell. X 1620. . Plastid similar to that shown in Fig. 21. 1620. Plastid similar to that shown in Fig. 21. x 1620. . Plastid similar to that shown in Fig. 21. x 1620. PLATE 132 Jour. ArNotp Ars. VoL. XVI ELAIOPLASTS IN [RIS Jour. ARNoLtp Ars. VoL. XVI PLaTE 133 ] 2 * no ‘o¢, ese ve CG ELAIOPLASTS IN IRIS PLATE 134 Jour. ArRNotp Ars. VoL. XVI \ \ \, \3 } Vad \ \s5 \ \~ Sy ELAIOPLASTS IN IRIS PuaTeE 135 Jour. Arnotp Ars. Vor. XVI ELAIOPLASTS IN IRIS Jour. ArRNoLtp Ars. VoL. XVI PLATE 136 ELAIOPLASTS IN IRIS Jour. Arnotp Ars. VoL. XVI PiaTE 137 ELAIOPLASTS IN IRIS 1o) Ke) Ww WwW =. © a nN FAULL, ELAIOPLASTS IN IRIS 267 — pallida Lam. X? Elaioplasts from living cells of cortex root-tip. xX 1 ie pallida Lam. X!? 2 Elaioplasts from living cells of cortex of er root. xX ». Iris pallida Lam. X : ? Plastids and chondriosomes from a single 1620. living cell of the central cylinder of a root-tip. : Plastids and chondriosomes from another cell of the central 0. PAS pallida Lam. X? Ojil-bearing plastids from living cells of the meristem of a rhizome: (a), (b) and (c) are from suc- cessively older cells and show the appearance of starcl 1 Tri ida Lam Chromoplasts from living cells of the root-cap: (a) and (b) are from successively older cells. X 1620. PLATE 136 Iris pallida Lam. X? Elaioplasts from living cells of cortex of 165 Ts : . Iris pallida Lam. X ? Chr a from living cells of the i (a), , (c) and (d) from suc- . Iris versicolor L. Plastids from living cells of the epidermis 1650. of a root-tip. . Iris pallida Lam. X ? Chromatophores from living cells of the 650. epidermis of a root-tip. . Chromatophores from living cells of the epidermis of a flower 1650. of a Pogoniris, probably of J. variegata L. . Iris versicolor L. Oijl-bearing plastids from living cell of cortex of root. X 1650. . Ins oe L. Chloroplasts from living parenchyma cells a leaf: (a) von starch ; with starch. . Iris versicolor L. stids from ine cells of the epidermis of a lea 1650. PLATE 137 . Lunularia cructata (L.) Dum. Living elaioplast-bearing cell from a mature thallus. xX 10 Lunularia cruciata (L.) Dum. "Living elaioplast-bearing cell from the younger tissue of a mature thallus. a of the Jungermanniales. Living cell from a mature leaf. ‘ Gil bodies from living cells of leaves of three different species 50. One of the Jungermanniales. Oil bodies from living differen- tiating cells of stem: (a), (b) and (c) from successively older cells. 250. . One of the Jungermanniales. Oil bodies from living cells of 250. mature plant. . One of the Jungermanniales. Oil bodies from living cells of 1250. younger tissue. . Vanilla Pompona Schiede. Elaioplast and chloroplasts from is. living cell of a lea x9 . Vanilla Pompona Schiede. Elaioplasts, chloroplast and chondri- xX 1720. osomes in living cell of cortex of root-tip. ARNOLD ARBORETUM, HARVARD UNIVERSITY. 268 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XVI NOTES ON YUCCA SusAn DELANO MCKELVEY With plates 138 and 139 Yucca Thornberi, spec. nov. Plate 138 Trunci 0.75-1.75 m. alti, congesti, infra foliis siccis reflexo-patentibus arcte obtecti, supra comam magnam satis elongatam foliorum viridium gerentes. Folia lineari-lanceolata, 0.30-1 m. vel ad 1.20 m. longa, 1.53.5 cm. lata, a basi vel a medio paullo latiore apicem versus attenuata, acuminata, pungentia, concava, utrinque laevia, luteo-viridia, stricta vel leviter recurva, flexibilia, initio margine saepe evanescenter denticu- lata, mox filifera filis crassiusculis curvatis tarde deciduis, parte basali 2.5-7.5 cm. longa et 7-12.5 cm. lata. Inflorescentia scapo 22—45 cm. longo incluso 1-1.30 cm. alta, angusta, basi et apice attenuata, folia quarta parte vel dimidio superantes; ramuli circiter 25, initio erecto- ascendentes, demum patentes; bracteae magnae, late triangulares; flores campanulati, 7.5-12.5 cm. longa; pistillum 5.5-7.5 cm. longum, ovario plerumque oblongo rarius attenuato, 4! ad 6% longiore quam lato, stylo 5—6 mm. longo, stigmatibus sub anthesi erectis vel fere erectis; filamenta 46.5 cm. longa, apice clavato brevi pro parte inferiore longo gracili. Fructus 13-17.5 cm. longus, 3—4.5 cm. diam., baccatus, incrassatus, apicem versus attenuatus et 2.5—5 cm. sub apice subito constrictus parte constricta plus minusve recurvata. Arizona. Pima Co.: foothills of the Rincon Mts., slightly north of Rincon Creek, a tributary of Pantano Wash, elevation 3600 ft., S. D. McKelvey, no. 1627, March 23, 1930 (type; herb, Arnold Arbore- tum). Also from the same region are the author’s collections nos. 1585, 2123, 2556, 2557, 2558, 2559, 2561, 2562, 2682, 2684. Yucca Thornberi forms large and crowded clumps and produces many rather long stems 2-5 ft. in height which are covered below with a thick thatch of reflexed-spreading dead leaves and are crowned by large, some- what elongated heads of green foliage which are constricted near the base and spreading above; the slightly broadened, concave, not conspicu- ously angled leaves are commonly smooth on both surfaces, yellow-green in color, with acuminate apex and, when young, often evanescently denticulate margins which, when the foliage is more mature become abundantly filiferous; the fibres are late-deciduous, moderately coarse 1935] McKELVEY, NOTES ON YUCCA 269 and loosely curled. The inflorescence is for some time rather fleshy and brittle, 3-4 ft. in length overall, with a scape 9-18 in. in length; the inflorescence proper is long, narrow, tapered at both ends, and extends for %4—% its length above the leaves; its branchlets are about 25 in number, at first erect-ascending, eventually spreading; its bracts are large, fleshy to leathery, broad-triangular in form. The flowers are campanulate, large, 3-5 in. in length; the pistil is 2'’¢—3 in. long with a commonly oblong, only rarely tapered, ovary which is 4144—6™% times as long as broad, the short style is ;3;— in. long and, at anthesis, with erect, or nearly erect stigmas; the filaments vary from approximately 11%4-2% in. in length and reach anywhere from slightly below to slightly above the shoulders of the ovary; their clavate tip is short in proportion to the long, slender, lower portion. The baccate fruit is 5-7 in. in length, slightly enlarged and tapered upward for its major lower portion, for 1-2 in. below the tip much contracted and commonly somewhat recurved. Yucca Thornberi appears to be most closely related to Y. arizonica and to Y. baccata Torr., differing conspicuously from the latter in habit of growth, from both species it is distinct in form of inflorescence, in char- acter of foliage and, though less so, in fruit. It is a pleasure to give to this new species the name Yucca Thornberi in recognition of the fact that Dr. J. J. Thornber of the University of Arizona called the author’s attention to the plant and with her spent considerable time in its study. Yucca brevifolia Engelm. var. Jaegeriana, var. nov. Plate 139 A typo recedit habitu humiliore vix 3-4 m. excedente, trunco brevi circiter 75-90 cm. longo, ramis brevibus fere erectis arcte congestis, foliis circiter 10 cm. longis vel 20 cm. vix excedentibus comam conges- tam et symmetricam formantibus, inflorescentiis vix 30 cm. longis, scapo 2.5-5 cm. longo et 2.5—3 cm. crasso incluso, ramulis tantum 2.5—6.5 cm. longis. CALIFORNIA. San Bernardino Co-.: vicinity of the Shadow Mts., elevation approximately 4000 ft., S. D. McKelvey, no. 2732, April 30, 1932 (type; herb. Arnold Arboretum). In several of the broad basins and foothill areas of the eastern part of the Mohave Desert of California and also in southern Nevada, occurs a form of the Joshua-tree which deserves varietal recognition. The plant— in appearance a miniature Joshua-tree—was brought to the attention of the writer by Mr. Edmund C. Jaeger of Riverside Junior College, River- side, California, and is named in appreciation of this fact. 270 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI This variety is primarily distinguished from the better known plant by its dwarfer habit,—the plants not exceeding, except rarely and only in extremely old specimens, 10-12 ft. in height; the trunk is short, about 214-3 ft. in length, stout (although, proportionately to that of the type of the species, slender); the branches are short, nearly erect and form an extremely dense, compact crown; the clusters of green leaves are crowded and symmetrical, about 1-2 ft. in length; the leaves are short, including the base about 4 in. long, not exceeding 8 in.,—or about the length of the shorter leaves of the type of the species; the inflorescence scarcely reaches 1 ft. in length; the scape is 1-2 in. long, 1-1) in. in diameter at base, and the flowering portion 9-10 in. in length with stout branchlets only 1-2% in. long. In flower and in fruit characters it is very similar to the plant of taller habit. Mr. Jaeger states (in litt., Oct. 2, 1934) that the distribution of the variety “reaches its greatest density in the vicinity of the New York Mts. in California.” It has been collected by the writer in southern Nevada between the Colorado River and Searchlight (no. 4094), in the Spring Mts. (no. 4142) and on the eastern slopes of the Charleston Mts. (nos. 4097, 4098, 4099, 4100, 4132), in the first and last of which regions it occurs in abundance. Yucca arizonica, nom. nov. Yucca puberula sensu Torrey in Botany, Emory Report, 221 (1859), in part, not Haworth. Yucca brevifolia Schott ex Torrey, Botany, Emory Report, 221 (1859), in part, as synonym of Y. puberula Torrey, not Haworth. — Engelmann in Trans. Acad. St. Louis, 3:46 (1873), in part, as synonym of Y. Schottii Engelm. — Trelease in Rep. Mo. Bot. Gard. 13: 100, tt. 57-59, 96 (fig. 2, range map) (1902), appearance as a valid species. — Not Y. brevifolia Engelm. (18 Yucca Treleasei MacBride in Contrib. Gray Herb. ser. be no. 56: 15 (1918); not Y. Treleasii Sprenger (1906 The name Y. puberula Haw. was first sean aly applied by Torrey to specimens collected by Arthur Schott in regions adjacent to Nogales, Arizona. Because of its connotation it is in the main referable to Y. Schottii Engelm., the inflorescences of which species are commonly puberulous. Schott’s material represented a complex. For a certain portion of this material the name Y. brevifolia used by Schott in his notes was adopted by Trelease in 1902. This name, as pointed out by Mac- Bride in 1918 was antedated by the name Y. brevifolia used by Engel- mann in 1871 for the Joshua-tree; MacBride in consequence gave to Schott’s plant the new name Y. Treleasei, Jour. ARNOLD Ars. VoL. XVI PLATE 138 Yucca THORNBERI McKelvey A plant growing at the type locality. Jour. ARNOLD ARB. VoL. XVI PLATE 139 YUCCA BREVIFOLIA var. JAEGERIANA McKelvey A plant, 12 ft. in height, growing at the type locality. 1935] McKELVEY, NOTES ON YUCCA 271 Unfortunately the name Y. Treleasii was used by Carl Sprenger in 1906 for a hybrid Yucca (See Bull. Soc. Tose. Ort. 31: 134. 1906.— Molon, Yucche, 192, t. 6. 1914); the plant is without a name and the new name Yucca arizonica is here adopted for this species. ARNOLD ARBORETUM, HARVARD UNIVERSITY. 7 _ — a _ — a _ —_ oe _ a a a oe a —_ —_ OO a a 7 oe oO ee ne BS a Oo OO 7 a a — a JOURNAL OF THE ARNOLD ARBORETUM VoLuME XVI JULY, 1935 NuMBER 3 THE VISIBLE STRUCTURE OF THE SECONDARY WALL AND ITS SIGNIFICANCE IN PHYSICAL AND CHEMICAL INVESTIGATIONS OF TRACHEARY CELLS AND FIBERS I. W. BarLtey AND THOMAS KERR! With plates 140-149 INTRODUCTION THE SECONDARY WALL of plant cells has long been known to be a heterogeneous structure. That it is more or less conspicuously striated and laminated was shown by Mirbel, Von Mohl, Valentin, Meyen, Th. Hartig, and other pioneer anatomists who demonstrated, in addition, that it may be resolved by specific chemical and mechanical treatments into lamellae, fibrils, granules, and other visible units of fairly constant form and size. This led, during the second half of the last century, to prolonged discussions concerning the fundamental structure of cell walls in general, and to much speculation regarding the physiological processes involved in their formation. Although a voluminous literature developed between 1850 and 1900, no consensus of opinion was reached concerning the exact physical and chemical significance of the visible heterogeneity of the secondary wall. Nor is there a general agreement among different groups of investigators at the present time. It is true that the study of anisotropy, of rod double refraction, of various types of dichroism, and of X-ray diagrams has in recent years contributed much toward a clearer understanding of sub- microscopic structures, and regarding the orientation of such structures in the grosser layers of the secondary wall, but it has not afforded as yet an adequate explanation of the finer types of visible heterogeneity. 1Parts of these Sams were made by the junior author as a National Research Fellow in Bot 274 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI In view of such facts as these, it seemed desirable to the writers to undertake a detailed investigation of the secondary wall in an endeavor (1) to verify and, if possible, to amplify the observations of previous workers; (2) to correlate results obtained by different techniques .and by the study of divergent cell types; and (3) to interpret the visible heterogeneity of the secondary wall in terms of its sub-microscopic structure and of its chemical composition. In an investigation of this character one is faced, at the outset, by a serious difficulty, upon the solution of which success or failure clearly depends. The range of recorded cases in which the details of wall struc- ture are even vaguely visible — without resorting to the use of softening or hardening processes, of macerating or swelling agents, and of other more or less drastic chemical and mechanical treatments — is very lim- ited. Severe treatments are capable of yielding extremely useful and significant data, but are likely to produce distortions and other artifacts, and therefore must be checked by observations on untreated material. In other words, an adequate system of controls — or means of accu- rately visualizing the normal structure of the secondary wall — is indispensable. As indicated in the preceding paper of this series (18), it is possible to section dense woods and other hard tissues without resorting to the use of softening processes which might modify their structure and chemical composition. It seemed advisable, accordingly, to make an extensive survey of a wide range of gymnosperms and angiosperms in search of species that afford clearly defined images of cell wall struc- ture in untreated sections. More than 3000 species, representing 160 families and 40 orders, were examined. It was found that the large- celled woods of various tropical dicotyledons provide unusually favor- able material for microscopic investigations. These plants are not bizarre or unusual forms; nor are they confined to any restricted group or genus. They are widely distributed and of not uncommon occur- rence in such families as the Theaceae, Monimiaceae, Icacinaceae, Rhizophoraceae, Euphorbiaceae, Flacourtiaceae, etc. When thin (5-10 1), smoothly-cut sections of the wood are examined in liquids of the right index of refraction, using the best modern optical equipment, the relatively broad expanse of wall in the fiber-tracheids and libriform fibers of certain of these plants reveals finely laminated, striated, and reticulated structures in exquisite detail. By using untreated sections of such plants as controls, it is possible to determine the exact effects upon normal structures of varied chemical and mechanical treatments, and thus to extend the scope of investigation to cover a wide range of less favorable species and tissues. 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 275 The following discussion of tracheary cells and fibers is divided into two parts. The grosser and more conspicuous types of layering of the secondary wall are dealt with in Part I; structures which more nearly approach the limits of microscopic visibility, in Part II. As previously stated, considerable is known! concerning the physical factors involved in the differentiation of the former structures, which must be clearly visualized and accurately correlated before proceeding to a detailed consideration of the finer types of visible heterogeneity. TERMINOLOGY The terms middle lamella, primary wall, secondary wall, and tertiary wall have been employed in several fundamentally different senses and to designate entirely different structures. This has led to much con- fusion in the literature and to serious discrepancies, not only in descrip- tive morphological work, but also in physiological, biophysical, and bio- chemical investigations. As a result of our detailed study of the cam- bium and its derivatives and of our preliminary investigations of other meristems and their derivatives, we attempted, in a former paper (18), to clarify the situation by suggesting that (1) the term middle lamella be used synonymously with intercellular substance in referring to the truly isotropic material which separates the walls of adjoining cells; (2) the term primary wall should no longer be applied to the first-formed layer of secondary thickening, but should be reserved for the original wall of the cell which is formed in the meristematic region and is carried over in more or less modified form into the fully differentiated tissues; and (3) the term secondary wall be used in referring to the strongly anisotropic layers of secondary thickening which are formed after a cell has attained its final size and shape. The term tertiary wall is so vari- ously used and interpreted and so confusing that its use should be dis- continued. We propose to employ our revised terminology in this and succeeding papers. I. THE PRINCIPAL LAYERS OF THE SECONDARY WALL A. LAYERING DUE To PHysIcAL Factors The secondary wall of normal tracheids, fiber-tracheids, and libri- form fibers commonly consists of three layers of different refractive character; (1) a relatively narrow outer layer, (2) a narrow inner layer, and (3) an intervening layer of variable thickness. When thin, per- 1For comprehensive reviews of the literature relating to this subject, the reader is referred to Van Iterson (30, 31) and Frey-Wyssling (13 276 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI fectly transverse sections of such cells are examined in polarized light between crossed nicols, Fig. 3, the inner and outer layers exhibit strong double refraction and are brilliant — except in positions of extinction — whereas the central layer is dark or noticeably less birefringent. The conditions tend to be reversed in longitudinal sections, Fig. 2, in which the central layer shows intense double refraction, and the inner and outer layers are dark or less conspicuously birefringent. In other words, as shown long ago by Dippel (7) and others, the secondary wall con- sists of anisotropic layers which are dark or brilliant in polarized light depending upon the plane of sectioning of the cell or upon the angle from which the wall is viewed. Our extensive survey of gymnosperms and angiosperms has demon- strated that most tracheids, fiber-tracheids, and libriform fibers are pro- vided with a secondary wall of this 3-layered type. The narrow inner and outer layers are of relatively constant thickness, not only in differ- ent parts of a given plant but also in plants of different systematic affinities. Variations in thickness of the secondary wall are due, there- fore, primarily to fluctuations in the width of the central layer. When the secondary wall is thin, as in the tracheids of the early wood of many conifers, the inner and outer layers are so closely approximated that the tenuous intervening central layer is invisible in polarized light, except in very thin (3-7 1), perfectly transverse sections of straight-grained tissue. In thicker or obliquely cut sections, the width of the inner and outer layers is much exaggerated by the scattering of light from these intensely birefringent structures. This fogs and conceals the central layer, just as the closely approximated brilliant outer layers of adjacent cells commonly obscure the tenuous primary walls and middle lamella (compare Figs. 1 and 3). Deviations from the normal 3-layered type of secondary wall are of not infrequent occurrence. Thus, many thick-walled libriform fibers and fiber-tracheids have no clearly differentiated inner layer, whereas others have more than three layers of varying width and birefringence, Fig. 4. Walls of a multiple-layered, anisotropic type, which are of rela- tively sporadic occurrence in the fiber-tracheids and libriform fibers of dicotyledons, are characteristic features of the fibers of many mono- cotyledonous stems. In transverse sections of such fibers, Fig. 6, there are narrow brilliant zones in polarized light which alternate regularly with broader and conspicuously less birefringent ones. Variations in the thickness of the secondary wall of these cells are due largely to variations in the number of successively formed layers. The optical behavior of the anisotropic layers of the secondary wall 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 277 of tracheary cells and fibers is closely correlated with the orientation of striations and so-called fibrillar structures, which are visible in cells that have been subjected to various chemical and mechanical treatments. When the striations and fibrils are arranged parallel, or nearly parallel, to the long axis of a tracheary cell or fiber, a layer is dark in sections cut at right angles to this axis, but is brilliant in longitudinal sections and in surface view — except, of course, in the four positions of extinc- tion. The intensity of the birefringence varies in obliquely cut inter- vening sections, decreasing as the plane of section approaches that of a truly transverse section. On the contrary, where the striations and fibrils are arranged approximately at right angles to the long axis of a cell, a layer is brilliant in cross sections and in surface view, but is dark in thin longitudinal sections, Fig. 2, which transect the fibrillar struc- ture. When the striations and fibrils have a helical arrangement and, therefore, are obliquely oriented in relation to the major axis of the cell, a layer is brilliant in surface view and more or less birefringent in both transverse and longitudinal sections. If the helix has a pitch of approxi- mately 45°, an oblique section, which is cut parallel to the striations and fibrils on one side of the cell, will transect these structures on the opposite side of the cell. Thus, in such sections, the layer will exhibit both isotropy and strong double refraction; i.e., it will be dark on one side of the section and brilliant on the opposite side. Changing the fibrillar orientation from a left-handed to a right-handed helix or vice versa will not alter the birefringence in transverse or in longitudinal sections so long as the angle of obliquity remains constant. In the typical 3-layered secondary walls of tracheids, fiber-tracheids, and libriform fibers, the striations and fibrils of the central layer are oriented parallel to the long axis of the cell, or at angles which do not deviate excessively from that axis; whereas those of the inner and outer layers are arranged more nearly at right angles to the major axis of the cell. Thus, the central layer exhibits strong double refraction in longi- tudinal sections, Fig. 2, and isotropy or relatively feeble double refrac- tion in transverse sections, Figs. 1 and 3; whereas the conditions are reversed in the case of the inner and outer layers of the secondary wall. In multiple-layered walls of the type illustrated in Fig. 6, the orienta- tion alternates regularly from parallelism to the major axis of the cell in the broader layers to marked obliquity in the narrower ones. The former layers exhibit intense double refraction in longitudinal sections; the latter layers, in transverse sections. In the case of optical anisotropy, the so-called index-ellipsoid has, according to Frey-Wyssling (13), a major axis (Ny) which is oriented 278 JOURNAL OF THE ARNOLD ARBORETUM [vOL. Xvi parallel to the striations and fibrils, and two minor axes (Ng and NB) which are placed at right angles to these structures. On the contrary, in the case of swelling-anisotropy, the ellipsoid of expansion has two major axes which are oriented at right angles to the striations and fibrils, and a minor axis which is parallel to these structures. Therefore, the dark layers of Figs. 1, 3, and 6, which have longitudinal striations, expand laterally, increasing in both width and circumference; whereas the strongly birefringent layers, the striations of which are oriented more nearly at right angles to the long axis of the cell, are unable to do so and expand longitudinally. Where the dark layers are of con- siderable width, they tend, by their excessive lateral expansion, to disrupt the thin birefringent layers, as indicated in Fig. 7. The strongly anisotropic behavior of the secondary wall suggests that its layers are composed of sub-microscopic units which have definite planes of orientation, and that there is a close correlation between the orientation of these units and of such visible structures as striations and fibrils. It was in fact a consideration of these phenomena which led Nageli to formulate the Micellar Hypothesis. More recently, X-ray analyses and other physico-chemical investiga- tions have indicated that native cellulose consists of chains of anhydrous glucose residues which are bound together by secondary valences into a space lattice of definite dimensions. These chains are arranged parallel to each other, and, in the case of the secondary wall of fibers and of Valonia, are oriented parallel to the striations and fibrils — as shown by Katz (17) and by Astbury and his co-workers (1). Furthermore, there is much cumulative evidence! from detailed investigations of anisotropy, of rod double refraction, of various forms of dichroism, and of X-ray analyses which suggests that the cellulose chains are not uniformly dis- tributed throughout the secondary wall, but are aggregated into more or less vaguely defined anisotropic units the major axis of which is oriented parallel to that of the visible striations and fibrils. In view of such facts as these, it is evident that layering of the type discussed on preceding pages is not due fundamentally to differences in chemical composition, but rather to changes in the orientation of aniso- tropic units of cellulose in the successively formed layers of the sec- ondary wall. B. Layerinc Due to CHEMICAL FACTORS The broad central layers of normal fiber-tracheids and libriform fibers frequently have subsidiary layers of varying width which are much 'This evidence has recently been summarized and discussed by Frey-Wyssling (13). 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 279 intensified by differential staining, Fig. 8. These subsidiary layers, un- like those illustrated in Fig. 4, are not closely correlated with variations in the orientation of the anisotropic cellulose, but are due to differences in lignification or to variations in the distribution of non-cellulosic con- stituents. They may be eliminated by delignification and other standard treatments for the purification of cellulose. It should be emphasized in this connection that the anisotropic layers of normal tracheids, fiber-tracheids, and libriform fibers are coherent even in walls that have been treated to remove their non-cellulosic constituents. There are evident planes of weakness but no actual discontinuities in the cellulosic matrix. Conspicuous. discontinuities are, however, of not infrequent occur- rence in the peculiar tracheids of “compression wood,” in so-called gelatinous fibers, in certain types of bast fibers, and in sclereids. They are due to narrow layers of truly isotropic material which contain little, if any, cellulose. Thus, when sections of unlignified or delignified cells are treated with standard solvents of pectic compounds and hemicellu- loses, the layers dissolve and liberate the anisotropic layers of cellulose which may be slipped apart as shown in Fig. 26. These truly isotropic layers may be accentuated by differential staining and are clearly visible in ordinary light, Fig. 27. They present some difficulties, how- ever, when sections are examined in polarized light between crossed nicols. For example, the entire laminated structure in Fig. 21, with the exception of the narrow outer layer, is dark in polarized light, owing to the fact that the orientation of cellulose in the anisotropic layers is parallel to the long axis of the cell. Therefore, the truly isotropic layers are concealed in transverse sections, but they are clearly visible in radial longitudinal sections and appear as dark lines between the birefringent layers of cellulose. There are similar tenuous isotropic films in the fibers of Pandanus on the outside of each narrow anisotropic layer, Fig. 6. They are masked in both transverse and longitudinal sections, since the broader anisotropic layers of cellulose are dark in cross sections, and the narrower ones are dark in longitudinal sections. C. LAYERING IN SCLEREIDS AND NON-FIBROUS SCLERENCHYMA It should be noted, before passing to a detailed consideration of the finer types of visible structures, that sclereids and other types of non- fibrous sclerenchyma have a fundamentally different type of secondary wall. The anisotropic layers of such cells—at least in tissues of the higher plants that we have examined thus far—show no conspicuous striations or fibrillar structures, either in the untreated or in the swollen 280 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XVI condition of the cell wall. Furthermore, the anisotropic layers are bril- liant in polarized light in all planes of section of the secondary wall, but are dark in surface view. The birefringent layers alternate more or less regularly with others which are dark in all planes of view, Fig. 5. A detailed discussion of these cells and of other non-fibrous types is re- served for subsequent papers of this series. II. THE FINER VISIBLE STRUCTURES OF THE SECONDARY WALL A. NorMAavL 3-LAYERED TRACHEIDS, FIBER-TRACHEIDS, AND LIBRIFORM FIBERS As stated in Part I, variations in thickness of the secondary wall of normal tracheids, fiber-tracheids, and libriform fibers are due primarily to fluctuations in the width of the central layer, which may attain a radial breadth of more than 15 y in the large-celled woods of various tropical dicotyledons. Therefore, the central layer provides more favor- able material for sectioning and for study at high magnifications than either the inner or the outer layers which are so tenuous as to present serious optical difficulties. Figure 10 is a transverse section of the wood of Siparuna bifida (P. & E.) A. DC. cut without preliminary softening or other modifying treat- ments. The broad central layer of the secondary wall is strikingly heterogeneous and exhibits a complex pattern of anastomosing radial striations. The striations are clearly visible in unstained sections mounted in water and in other liquids of varying indexes of refraction; and, in white light, are optically of two types, i.e., light and dark. There are corresponding light and dark striations in tangential longitudinal sections, Fig. 13. It is evident, accordingly, that the central.layer of the secondary wall in these cells is composed of thin plates or lamellae which have a radio-longitudinal or radio-helical orientation. The lighter lamellae are strongly birefringent in polarized light, Fig. 13, except in positions of extinction and in sections cut at right angles to the longi- tudinal axis of the lamellae; whereas the alternating lamellae are dark, or at least comparatively isotropic, in all planes of view.! The birefringence of the lighter lamellae is not due entirely to rod double refraction, as may be determined by examining sections in a graded series of liquids of varying indexes of refraction. Nor is the 1Extremely thin, smoothly cut sections are essential for critical examination in polarized light. If the sections are too thick or are scratched or roughened in cutting, e tenuous dark lamellae will be completely masked by the pt of light from the ay birefringent lamellae. 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 281 apparent isotropy of the intervening lamellae due solely to the masking effects of lignification or to the presence of other non-cellulosic con- stituents. The walls of immature unlignified cells show identical pat- terns and a similar differentiation into lamellae of two distinct cate- gories of birefringence, as do delignified cells that are treated for the removal of hemicelluloses and other non-cellulosic constituents. By subjecting untreated sections to the action of such swelling agents as acids, alkalies, chloro-iodide of zinc or cuprammonium hydroxide, and by carefully controlling the reactions, it is possible to expand the cen- tral layer and its constituent lamellae without distorting or seriously modifying the original structural pattern (compare Figs. 10 and 11). As the central layer expands and enlarges under the microscope, suc- cessively finer details of structure become visible. The lamellae are not discrete homogeneous entities, and are resolved during the expansion of the central layer into aggregations of elongated heterogeneous com- plexes of varying degrees of fineness which grade down to the limits of microscopic visibility. The darker lamellae are compact sheets of rela- tively isotropic material which contain a low ratio of birefringent com- plexes. On the contrary, the lighter lamellae are aggregations contain- ing a high ratio of birefringent complexes and a low ratio of apparently isotropic ones. There are no discontinuities in the structural pattern which is firmly knit together by lateral anastomoses and interlocking complexes. After treatment for the removal of non-cellulosic constituents, the purified cellulose exhibits a similar structural pattern, which upon swell- ing, Fig. 14, is resolved into a complex and firmly coherent matrix, hav- ing elongated, intercommunicating interstices of varying degrees of fine- ness. The darker and more compact parts of the matrix, which corre- spond to the lighter lamellae of Fig. 10, are strongly birefringent in longitudinal sections and show conspicuous dichroism when carefully stained with congo red or chloro-iodide of zinc; whereas the lighter and more porous parts of the matrix, which correspond to the darker lamellae of Fig. 10, are so feebly birefringent that they appear to be compara- tively isotropic. Conversely, when the central layer is freed of cellulose by treatments with 72% sulphuric acid, the details of the swollen pattern are preserved in the so-called “lignin” residue, Fig. 11, which also is a complex and firmly coherent structure, having elongated, intercommunicating inter- stices of varying degrees of fineness. The lighter, finer residues of the originally birefringent lamellae exhibit well defined rod double refraction in longitudinal sections; whereas the darker, denser residues of the originally isotropic lamellae do not. 282 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI It is evident from a detailed comparison of Figs. 11 and 14, that the denser parts of the “lignin” residue correspond to the more porous parts of the matrix of purified cellulose, and that the “lignin” residue may be interpolated within the interstices of the swollen cellulose. Furthermore, the rod double refraction of the lighter lamellae of the “lignin’’ residue suggests that the two interpenetrating complexes grade downward in size far below the limits of microscopic visibility. In other words, each of the visible parts of the original structural pattern is heterogeneous and com- posed of optically different complexes. Removal of either the “lignin” or the cellulose leaves a coherent matrix of varying texture and porosity. It is possible to reconstruct the structural pattern of the swollen cellu- lose from the “lignin” residue or vice versa, since they are positive and negative images of the same pattern. Although swollen sections of purified cellulose afford excellent preparations for visual examination, they are difficult objects for photographic reproduction. Therefore, a majority of our photomicrographs were made from “lignin’’ residues. The structural pattern of the central layer is not a constant: it varies greatly not only in different groups of plants, but also at times in homologous cells of the same plant, and even within the wall of a single cell. For example, in Siparuna bifida, the two optically different com- plexes may be segregated into coarsely radial patterns which are clearly visible in untreated sections, Fig. 10, or they may be diffused in finer radio-reticulate patterns, Fig. 16, the finest of which are invisible in unswollen sections of the secondary wall. In such cells, conspicuous concentricities usually are due either to abrupt changes in the texture of the structural pattern, Fig. 11, or to zones of varying intensities of lignification, Fig. 9. The former persist in purified cellulose: the latter are eliminated during delignification. Structural patterns of a basically concentric type are, however, of common occurrence in the normal tracheids of conifers, Fig. 18, and in the fiber-tracheids or libriform fibers of such dicotyledons as Poraqueiba sericea Tul., Fig. 15. In the central layer of these cells, the optically different complexes are segregated into concentric lamellae of varying widths and spatial groupings. The lamellae are of two types, Le., strongly birefringent and comparatively isotropic. They are not dis- crete homogeneous entities, but may be resolved by treatment with swelling agents into complexes of varying degrees of fineness. As in the case of Siparuna bifida, the darker lamellae are compact aggregates of relatively isotropic material, Figs. 15 and 18, and contain a low ratio of birefringent cellulose; whereas the alternating lighter lamellae are composed largely of birefringent cellulose and contain a low ratio of 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 283 isotropic material. The structural pattern persists in delignified sec- tions which are treated with standard solvents of hemicelluloses and of other non-cellulosic constituents. When the purified cellulose is swollen, it appears as a complex and firmly coherent matrix, which exhibits a structural differentiation into compact, strongly birefringent and looser, comparatively isotropic lamellae. It should be emphasized, in this connection, that the concentric struc- ture of swollen cotton hairs —to which the work of Balls (2) has directed so much attention — appears to be due fundamentally to a similar structural pattern. When extremely thin, very smoothly cut sections of raw or purified cotton are treated with diluted Schweizer’s reagent and are examined in polarized light between crossed nicols, the lamellae are, during the early stages of swelling, clearly of two optically different types, i.e., strongly birefringent and comparatively isotropic. During subsequent apelin. Fig. 17, the central layer is resolved into a complex and firmly coherent, spongy structure, the conspicuously bi- refringent parts of which are denser and obviously contain a higher ratio of cellulose than the more porous, intervening parts. In other words, the structural patterns of the central layers of cotton hairs, Fig. 17, of coniferous tracheids, Fig. 18, and of the fiber-tracheids of Poraquetba sericea, Fig. 15, appear to be of a fundamentally similar type. In cotton hairs, as in tracheary cells, the width of the concentric lamellae is not a constant, but varies within relatively wide limits. The structural pattern of cotton can not be due to a segregation of cellulosic and non-cellulosic constituents, since the central layer of cotton is composed of practically pure cellulose — the low ratio of non-cellulosic constituents in cotton is confined chiefly to the so-called cuticle or pri- mary wall and to the lumen of the cell. Nor can the concentricities be due merely to inequalities in the penetration or modifying effects of the swelling agent, as may be demonstrated by cross-correlating the struc- tural patterns of different hairs from the same boll. For example, in Fig. 17, in passing outward from the lumen, there is the following sequence of lamellae: six narrow alternating light and dark zones, an unusually wide light zone, two broad dark zones separated by a narrower light zone, two narrow dark zones and three narrow light zones, and six broad dark zones separated by narrower light zones. The fact that this identical complex of varying concentricities occurs in other hairs from the same boll can not be due to purely fortuitous circumstances, but might be due, either directly or indirectly, to the modifying effects of environmental factors upon the developing hairs. or can the structural patterns of tracheids, fiber-tracheids, and libri- 284 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI form fibers be due to inequalities in the penetration and modifying effects of the swelling agents, since the patterns are visible under favorable conditions in untreated sections. Thus, the striking similarities in the finer visible structures of the central layer of unlignified and delignified cells and of “lignin” residues indicate that there are fundamental struc- tural differences in the underlying cellulose to which the pattern of lig- nification must more or less closely conform. Combinations of radial and concentric patterns of varying texture and complexity are of common occurrence in the fiber-tracheids and libri- form fibers of dicotyledons.' In such cells there may be abrupt transi- tions within the central layer from coarse to fine texture and from radio-reticulate to concentric arrangements and vice versa. Fig. 19 is a transverse section of the wood of Tetramerista glabra Miq., cut with- out preliminary softening or other drastic treatments. It illustrates a type of complex radio-concentric structure which is clearly visible in unstained sections mounted in water and other liquids of varying indexes of refraction. The pattern is complicated, however, as is so often the case in cells of this type, by the presence of zones of varying intensities of lignification. A radio-concentric pattern of much finer texture is illustrated in Fig. 20. In the case of the more heavily lignified zones of such central layers, Figs. 9 and 20, both the birefringent and the comparatively isotropic parts of the structural pattern persist in “‘lignin’’ residues; whereas, in the less intensely lignified zones, the birefringent parts leave no struc- tural residue. It is of interest, in view of the significance that has been attached to the work of Freudenberg and his co-workers (12), that in longitudinal sections the residues of heavily lignified parts exhibit con- spicuous rod double refraction; whereas the residues of the less intensely lignified parts do not.’ In other words, there appear to be submicro- scopic structural differences in the two optically different complexes of the structural pattern which are reflected in their “lignin” residues. Furthermore, as previously noted, when delignified sections are stained with chloro-iodide of zinc or congo red, the strongly birefringent parts of the structural pattern may become markedly dichroic; whereas the more nearly isotropic parts do not. The observational and experimental data that we have assembled in 1Concentric patterns with ee radial groupings are of not infrequent occurrence in the tracheids of conifer Rod double refraction is appar only in _ parts of the lignin — sere are iets arcane in the original material. Therefore, it can be s in sections which transect the so- ere Bbrilar structure, since all ro ie is dark in eval ial “Tight j in such planes of secti 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 285 our extensive survey of a wide range of gymnosperms and angiosperms indicate that the central layer of normal tracheids, fiber-tracheids, and libriform fibers is composed, in all cases, of a complex and firmly coherent matrix of cellulose with elongated, intercommunicating inter- stices. Within these interstices more or less “lignin” and other non- cellulosic constituents may be deposited. The denser and more porous parts of the cellulosic matrix exhibit striking contrasts in birefringence, which are accentuated by lignification. Where these optically different parts are diffused in various patterns of fine texture — as is usually the case in the tracheids of conifers and in the fiber-tracheids and libriform fibers of many dicotyledons — the structural complexes are invisible in untreated sections of the secondary wall, but may be swollen to micro- scopically visible dimensions, Figs. 9, 12, 15, 16,18, and 20. On the contrary, where the two optically different parts are segregated into coarser structural complexities, Figs. 10, 13, and 19, the patterns are clearly visible in unswollen sections. The cellulosic matrix of the central layer is composed, in all cases, of anastomosing elongated complexes which are oriented parallel to the long axis of the cell or in a helical arrangement. In fact, it is these elongated complexes of two optically different types, Fig. 13, which give a longitudinally or helically striated appearance to the central layer and determine its helical or longitudinal planes of cleavage into so-called fibrils. In other words, fibrils are heterogeneous shredded parts of an originally continuous and coherent matrix. Although there are serious optical difficulties in studying the tenuous inner and outer layers of the secondary wall in sectional view, the striated appearance of these layers in surface view strongly suggests that they have similar structural patterns, the elongated, strongly birefringent complexes of which are oriented more nearly at right angles to the longi- tudinal axis of the cell. B. MULTIPLE-LAYERED FIBERS The orientation of the elongated complexes of the structural pattern may be relatively uniform throughout the central layer of tracheids, fiber-tracheids, and libriform fibers, or it may deviate more or less in successively formed parts of this layer. Not infrequently, the changes in orientation are correlated with fluctuations in the texture of the struc- tural pattern. Where the deviations are of considerable magnitude, they may be detected in polarized light, as illustrated in Fig. 4. The brilliant internal zones resemble the inner and outer layers of the secondary wall in having their birefringent complexes oriented more 286 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI nearly at right angles to the longer axis of the cell, and therefore are bright in transverse sections. Although there is a superficial similarity between Fig. 4 and Fig. 6, the two cell walls are of a fundamentally different type. In the fibers of Pandanus, Fig. 6, as in the libriform fibers of various representatives of the Flacourtiaceae, Figs. 21 and 26, and in the bast fibers of ramie and of other dicotyledons, there are, as previously stated, actual discontinui- ties in the cellulosic matrix produced by narrow isotropic films of a non- cellulosic character. It should be emphasized, in this connection, how- ever, that the individual anisotropic zones of these multiple-layered fibers have complex structural patterns of the general types discussed on preceding pages. For example, Fig. 21 is a transverse section of the unswollen wall of Homalium luzoniense F. Villar. The layers of cellu- lose have a radio-reticulate pattern, the finer structural details of which are more Clearly visible in swollen sections or in “‘lignin”’ residues, Fig. 22, The elongated birefringent complexes of the structural pattern are oriented parallel to the long axis of the cell. Therefore the entire com- plex of layers is dark in polarized light in transverse sections. Coarsely radial patterns of the type illustrated in Fig. 10 are of not uncommon occurrence in the individual anisotropic layers of certain bast fibers; whereas in Pandanus, Fig. 7, the two optically different aggregates of cellulose are diffused in a pattern of unusually fine texture. Where the strongly birefringent complexes are oriented closely parallel to the longitudinal axis of the cell, the layer is dark in cross sections, Fig. 6, and merges with the truly isotropic film of non-cellulosic material; where they are oriented more nearly at right angles to the major axis of the cell, the layer is brilliant in transverse sections. Variations in the orientation of cellulose in successively formed parts of the secondary wall have a marked effect upon the swelling of tracheary cells and fibers. Owing to its specific anisotropy, the cellulose expands at right angles to the so-called fibrillar axis, and, during extensive lateral swelling produced by strong chemical reagents, actually contracts in a direction parallel to this axis. In the case of isolated, delignified tracheary cells and fibers having normal 3-layered secondary walls, the laterally expanding central layer frequently splits the tenuous, longi- tudinally expanding outer layer into a series of constricting rings and helical bands, Fig. 23, and bulges outward between these structures. This ring-bead type of swelling occurs in cotton and has received con- siderable attention in literature dealing with commercial fibers. Al- though the so-called cuticle or primary wall may aid at times in bead formation, the controlling factor in cotton hairs, as in tracheary cells 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 287 and fibers, appears to be differences in orientation of cellulose in the outer and central layers of the secondary wall. There are no trans- verse plates of non-cellulosic material in the secondary wall which are concerned in ring-bead formation as hypothesized by Ludtke (21). In the case of multiple-layered tracheary cells and fibers, it is possible to verify conclusions based upon the study of cells of the 3-layered type. We have shown that the concentric anisotropic layers of various repre- sentatives of the Flacourtiaceae, Figs. 21, 22, and 26, are separated by films of non-cellulosic material, and that the orientation of the cellulose is constant except in the outermost layer of the secondary wall, where it is more nearly at right angles to the longitudinal axis of the cell. When such cells are partly or completely delignified and are swollen in cupram- monium hydroxide, the internal complex of anisotropic layers expands laterally and disrupts the tenuous outer layer into constricting rings, Fig. 27, or helical bands, Fig. 24. The internal layers of cellulose — which may be slipped apart as shown in Fig. 26 — expand more or less in unison, Figs. 24 and 27, and no subsidiary internal constrictions are formed. On the contrary, in the multiple-layered fibers of Pandanus and of other monocotyledons — which have similar isotropic films of non-cellu- losic material, but where the orientation of the cellulose changes in the successively formed anisotropic lamellae — each of the narrow aniso- tropic layers, Fig. 6, may be disrupted by the lateral expansion of the broader layers, Fig. 7, and in the case of entire, delignified fibers, may give rise to constricting rings and helical bands, Fig. 25. In other words, the fiber behaves as if it were composed of several two-layered secondary walls, each of which swells in turn, forming similar ringed and beaded structures, Fig. 25. The two outermost layers swell first, the expansion working from the ends towards the center of the cell. The first formed ringlike constrictions commonly determine the position of subsequently formed internal constrictions. Multiple-layered fibers of the Pandanus type are of common occur- rence in the primary tissues of the stems of many monocotyledons. It is evident from Liidtke’s (20, 22) figures and descriptions that the fibers of bamboo are of this structural type, and that they exhibit similar phenomena during their expansion in such swelling agents as cupram- monium hydroxide. It is obvious, in addition, that purely physical phenomena of swelling have been misinterpreted by Ludtke as evidence for the existence of transverse plates (Querelemente) of non-cellulosic material. 288 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XvI DISCUSSION A. CONCENTRICITIES The secondary walls of tracheary cells and fibers are extremely com- plex and variable structures. Therefore, it is misleading and fruitless to attempt to homologize all types of fibers in a single structural model. For example, there are five different types of visible concentricities, due to: 1. The segregation of two optically different aggregates of cellulose into concentric patterns. A rupt changes in the form or texture of the structural pattern. 3. Changes in the orientation of the elongated birefringent com- plexes of the structural pattern. 4. Varying intensities of lignification or differences in the distri- bution of non-cellulosic constituents within the structural pattern. 5. Alternation of cellulosic and non-cellulosic layers. In so far as we are able to judge from a study of a wide range of gymnosperms and angiosperms, most, if not all, tracheary cells and fibers exhibit more or less conspicuous concentricities of the third type, i.e., those due to changes in the orientation of the elongated birefringent complexes of the structural pattern, but the number and magnitude of the deviations in orientation are variable. Inability to detect such con- centricities appears to be due to inadequate techniques or to errors of interpretation. In most cases, the third type of layering occurs in asso- ciation with one or more of the other four types of concentricities. Thus, in the secondary wall of cotton hairs, it occurs with the first type; in the fiber-tracheids of Siparuna bifida, with the second and fourth types; in the fiber-tracheids of Tetramerista glabra, with the first, second, and fourth types; in the fibers of Pandanus, with the first and fifth types, etc. Variations in the intensity of lignification or in the distribution of other non-cellulosic constituents may at times be closely correlated with changes in the orientation or the texture of the structural pattern. For example, the narrow inner and outer layers of the secondary wall may be more heavily lignified than the central layer or vice versa. Similarly, the coarser parts of the structural pattern of the central layer may be more heavily lignified or contain a higher ratio of hemicelluloses than the finer parts or vice versa. It is such fortuitous correlations as these which have led, in certain cases, to the unwarrantable conclusion that all types of visible heterogeneities in the secondary wall are due primarily to differences in chemical composition. There are investigators who believe that all fibers are composed of concentric lamellae of cellulose which are held together by non-cellulosic 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 289 material. Thus, Liidtke (21, 22), who has attempted to homologize all types of fibrous cells in a single structural model, is of the opinion that the lamellae are separated by a “‘Fremdsubstanz” which differs from both cellulose and lignin in its chemical composition. Ritter (26) argues that it is possible to dissect the secondary wall by chemical means into concentric lamellae which may be slipped apart as shown in Fig. 26. Liidtke’s conclusions appear to have been derived largely from a study of bamboo fibers; and Ritter’s, from investigations of the libriform fibers of elm. We have shown that the anisotropic lamellae of mono- cotyledonous fibers frequently are separated by films of non-cellulosic material. The libriform fibers of elm are commonly of the so-called gelatinous type, which also are characterized in many cases by having both cellulosic and non-cellulosic lamellae. In such fibers, where there are actual discontinuities in the structural pattern of cellulose, the anisotropic lamellae may readily be separated by chemical treatments and slipped apart. On the contrary, in cotton hairs and in normal tracheids, fiber-tracheids, and libriform fibers, the entire matrix of cellu- lose is firmly coherent, and can be dissected only by forcibly tearing or rupturing the structural pattern. In Siparuna bifida the more obvious planes of weakness in the cellulosic matrix are radio-longitudinal or radio-helical; whereas in cotton hairs or in Poraqueiba sericea they are concentric-longitudinal. B. “Frsrits” AND OTHER “UNITS” OF CELLULOSE Since the pioneer days of Von Mohl, Valentin, and Th. Hartig, a succession of investigators have visualized the secondary wall as com- posed of visible units of cellulose — elementary fibrils, dermatosomes, etc. — that are held together by non-cellulosic material. It is essential to understand the relationship between these units and the visible struc- tural patterns produced by different optical aggregates of cellulose. We have shown in Part I of this paper that the orientation of the cellulose is correlated with that of the so-called fibrillar structure, as has been demonstrated by analyses of X-ray diagrams, of anisotropy, of dichroism, and of other physical properties of the cell wall. How- ever, these physical correlations are concerned only with the orientation of the fibrillar structure and afford no conclusive evidence that fibrils obtained by chemical or mechanical treatments are discrete entities of constant length or cross sectional area. Ritter (27) has discussed the length of the so-called fibril and con- cludes that it is variable. He states that “although fibril segments of only 230 microns in length have been isolated, it seems that some may 290 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI be as long or longer than the fiber.” Liidtke (22), on the contrary, claims that the length of fibrils is determined by the presence and spac- ing of transverse plates of non-cellulosic material. Jancke, working with R. O. Herzog (15), measured the width of fibrils and obtained values of some 0.3—0.5 1. Balls and Hancock (3), proceeding upon the assumption that lamellae’ are composed of a single concentric row of fibrils, inferred that the width of both lamellae and fibrils in cotton is 0.4 p. Frey-Wyssling (13) tabulates the dimensions of fibrils as 0.4 « 0.4 & 100 LL. Fibrils may be dissected by relatively drastic treatments with oxidiz- ing agents or acids into short segments which are variously designated as dermatosomes, spherical units, ellipsoid bodies, etc. According to Frey-Wyssling (13), dermatosomes have dimensions of 0.4 * 0.4 x 0.5 4; whereas Farr and Sisson (11) state that ellipsoid bodies pre- pared from cotton have axes of 1.1 yp and 1.5 yp. Liidtke (20) believes that dermatosomes are held together by his ‘“‘Fremdsubstanz”; whereas Farr and Eckerson (9) maintain that the ellipsoid bodies of cotton are jacketed by a pectic cement. We have demonstrated in Part II that the central layer of tracheary cells and fibers is composed of an extremely complex and firmly coherent matrix of cellulose and that the details of the structural patterns of this matrix grade down to the limits of microscopic visibility. There is no evidence, either in untreated or in carefully swollen fibers, of discrete entities of cellulose, i.e., of fibrils or dermatosomes, which may be lib- erated simply by dissolving non-cellulosic constituents. The matrix of cellulose is shredded and disrupted during the production of fibrils and dermatosomes, which are heterogeneous fragments of larger size than the finer visible complexes of the structural pattern. In cotton, Fig. 17, as in Pinus, Fig. 18, Poraqueiba, Fig. 15, and Siparuna, Figs. 10 and 14, the lamellae obviously are not composed of a single row of adherent fibrils, but are alternating layers of varying width, porosity, and bire- fringence. The finer, visible, elongated complexes of the lamellae are 0.1 «1 or less in thickness. As indicated at (a) in Fig. 17, the cross sec- tional area of an ellipsoid body of the size postulated by Farr and Sisson covers more than four lamellae and a relatively large number of the finer visible complexes. ‘The form and size of the fragments which may be dissected from the secondary wall are clearly dependent upon the structural pattern of the matrix of cellulose, and upon the type and severity of the chemical and ‘Balls did not recognize two distinct categories of lamellae and evidently obtained the value of 0.4 u by dividing the total width of the wall by the number of denser, strongly birefringent lamellae. 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 291 mechanical treatments to which the material is subjected. Splits or cracks develop in the more porous and weaker parts of the matrix, thus liberating the denser parts which contain a higher ratio of birefringent cellulose. In addition, there are submicroscopic, transverse, or oblique planes of cleavage, i.e., ‘‘slip planes,” to which the work of Von Hohnel (16) and of Schwendener (29) has directed so much attention. It is these slip planes, rather than Liidtke’s hypothetical ‘“‘Querelemente,” which facilitate the dissection of the fiber and of the elongated com- plexes of its structural pattern into shorter segments. It is of interest, in this connection, that a fibrillar structure is visible after the action of 72% sulphuric acid upon longitudinal sections of fibers which yield coherent “lignin” residues. By the use of mechanical pressure during the initial stages of the action of the acid, the walls of tracheary cells, Fig. 12, and fibers may be resolved into long “lignin” threads, similar to fibrils. These shreds of the originally coherent frame- work of “lignin” may be dissected by more drastic chemical and mechan- ical treatments into nearly isodiametric fragments resembling dermato- somes. As previously stated, the amorphous non-cellulosic constitu- ents are deposited within the elongated, intercommunicating interstices of the cellulose matrix, resulting in two continuous, interpenetrating systems. Neither system is composed of discrete entities of visible dimensions, but each may be disrupted into fragments of varying size and form. If there are actual discontinuities in the systems, they must occur in the submicroscopic field, e.g., in the realm of micelles or of molecular chains. It should be emphasized, in addition, that so-called fibrillar structures are not visible in the secondary walls of parenchyma, of sclereids, or of other cells which exhibit statistical isotropy in surface view. The structural pattern of the cellulose matrix in such walls is ofa fundamentally different type from that which occurs in fibrous cells. Dermatosomes, spherical units, and ellipsoidal particles are difficult to homologize, either as regards their size or their form. They are obtained by the action of oxidizing agents or of acids which tend to modify the cellulose. Neale (24) has summarized the modifying effects of oxidation and hydrolysis as follows: ‘The loss of strength and fall in viscosity which accompany the hydrolysis or oxidation of cellulose are quite irreversible, and the general term degradation is applied to these changes. The degradation of cellulose is accompanied by the appear- ance of chemical properties foreign to the original material. The hydrolysis of the glucoside-oxygen bridge causes the appearance of reducing sugar properties which may be quantitatively, though arbi- trarily, expressed as ‘copper number’ or ‘iodine number.’ The reducing 292 JOURNAL OF THE ARNOLD ARBORETUM [VOL XvI sugar properties also arise ac a result of oxidation and may be accom- panied by the development of acidic properties, so that oxidized cellu- lose may retain traces of caustic alkali or absorb basic dyes. This latter property has been put on the quantitative basis so essential in the chemistry of cellulose in the form of the methylene blue absorption test.” Thus, it may be seen that the action of acids, which are supposed to dissolve some cementing substance and to liberate integral units of cellulose, may actually result in partial degradation of the cellulose. We have found that the staining of hydrocellulose and oxycellulose with ruthenium red is similar to the methylene blue absorption values, as listed by Dorée (8). Ruthenium red behaves, in some respects, as a basic dye, and the staining of ellipsoidal particles, obtained by treating cotton with relatively strong acid (10), may be interpreted as an indi- cation of the degradation of the cellulose rather than as evidence for believing that the particles are coated with a pectic cement. Ruthenium red is not a specific test for pectic compounds, as botanists have fre- quently assumed. It is removed from dilute aqueous solutions by coagu- lated protoplasm and other nitrogenous substances, by gums, mucilages, hemicelluloses, oxycelluloses, hydrocelluloses, and certain lipoids, as well as by pectic compounds. Any hypothesis concerning the visible structure of the secondary wall must account not only for the varying structural patterns of a wide range of cells, but also for well known facts regarding the physical and chemical properties of cellulose. In the case of the hairs of the cotton plant, the constituents which do not yield glucose upon hydrolysis are small in amount, and are confined chiefly to the so-called cuticle or primary wall and to the lumen of the cell. There obviously is not a sufficient volume of cutinlike substances or of pectic compounds in the secondary wall to serve as a cementing substance of the type postulated by Ludtke (21) or by Farr and Eckerson (9). Furthermore, when cotton is treated with solvents of such constituents, without degrading the cellulose, the structural pattern is not affected. It persists as a firmly coherent matrix of cellulose. It is now generally admitted that the cellulose molecule is a long chain of glucose residues bound together by oxygen bridges. Furthermore, there is evidence from X-ray analyses, from anisotropy, dichroism, etc., to indicate that cellulose is built up of submicroscopic, crystal-like aggregates of these chains. The length of the cellulose chain and its arrangement within the crystallite are still subjects of dispute. Thus, it is uncertain whether the chain is shorter or longer than the crystallite or of equivalent length, and whether micelles are discrete and separate entities, or merely parts of a continuous system of overlapping chains. 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 293 Estimates of the length of cellulose molecules range from 100-3500 glucose residues. The highest value of 3500 units, i.e., that of Kraemer and Lansing (19), is based upon measurements of viscosity. Such molecules would have a length of approximately 1.8 yp, and would be visible microscopically if they were of sufficient thickness, which they obviously are not. Since the cellulose chains are arranged parallel to the so-called fibrillar orientation, and since there are no visible struc- tures which transect this axis, it is possible to conceive of chains of the length postulated by Kraemer and Lansing arranged in an overlapping manner along the fiber axis. Our investigations indicate that the cellulosic matrix of the secondary wall is composed of complexes of cellulose of varying birefringence which grade down to the limits of microscopic visibility, and that the funda- mental units of cellulose are of submicroscopic dimensions. In the case of cotton, the available chemical and physical data make it appear im- probable that the variations in birefringence are due to differences in chemical composition. Correns (5) recognized, more than 40 years ago, that cellulose is heterogeneous and attempted to explain the visible striations and certain types of lamellae as due to differences in water content. This hypothesis originated with Nageli (23), who postulated water rich and water poor layers as a means of explaining concentrici- ties and still permitting growth by intussusception. Differences in water content apparently do exist, and may be a factor influencing the inten- sity of birefringence in different lamellae. However, it is difficult to evaluate such differences by a study of dried material. Drying the walls shrinks the cell so that structures just within the limits of micro- scopic visibility when the preparation is in water, may be contracted to invisible dimensions. Furthermore, differences in water content must be explained in terms of submicroscopic differences in the cellulose which permit varying degrees of hydration. The question whether the varia- tion in birefringence of different complexes of the cellulosic matrix is due to fluctuations in the size, number, or orientation of submicroscopic units of cellulose is one which must be attacked from the physical and chemical, rather than from the botanical, side. C. SIGNIFICANC E OF BIOLOGICAL VARIABLES IN PHYSICAL AND CHEMICAL INVESTIGATIONS Our survey of a wide range of gymnosperms and angiosperms indicates that the secondary wall is a very complex structure, and that the struc- tural pattern of the cellulose matrix varies greatly, not only in different groups of plants but also, at times, in homologous cells of the same plant, 294 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI and even in different parts of the same cell. There is a similar varia- bility in the distribution of “lignin” and of other non-cellulosic constitu- ents. Therefore, since all types of secondary walls can not be homolo- gized in a single structural model, there are grave dangers in generaliz- ing from intensive investigations of isolated species, e.g., cotton, spruce, bamboo, or ramie. Deductions concerning the structure of the cell wall based upon physi- cal or chemical analyses, should be checked by microscopic investigations and by accurate information concerning the numerous biological vari- ables. This is particularly necessary in the interpretation of X-ray diagrams, where the investigator of necessity must deal with complex aggregates of plant material. Van Iterson (31) has shown that certain misconceptions regarding Valonia might have been avoided by an acquaintance with the work of Correns (6) and others upon the visible structure of the walls of algae. Preston (25) undoubtedly errs in con- cluding, from an examination of X-ray diagrams, that there is a single plane of orientation of “fibrils” in the secondary wall of the’ tracheids of Sequoia and of other conifers. Accurate interpretations of X-ray diagrams of growing cells and of differentiating tissues are especially difficult, and such conclusions regarding structural changes as those of Clark and Farr (4) and Ritter and Stillwell (28) must be carefully verified from the histological side. Although “lignin” residues of thick sections exhibit rod double re- fraction, as demonstrated by Freudenberg and his co-workers (12), a careful study of the residues of thin sections shows that a considerable proportion of the secondary wall “lignin” is isotropic. Similarly, there are parts of the cellulosic matrix which do not exhibit a clearly defined dichroism when thin sections are stained with chloro-iodide of zinc or congo red. The woods of certain dicotyledons leave no coherent struc- tural residue when subjected to standard treatments with 72% sulphuric acid, as shown by Harlow (14); whereas others leave compact residues such as have been considered to be typical of conifers. In the wood of certain plants, the bulk of the “lignin” is confined to the so-called middle lamella, as Ritter (27) maintains; whereas in others, there is a relatively large proportion in the secondary wall. It should be emphasized, in conclusion, that most of our own data were obtained from a study of tracheary cells and fibers, and that many additional types of cells must be investigated before it will be possible to visualize the full range of structural variability of the secondary wall. In a succeeding paper, we shall discuss methods that have been perfected for studying the small-celled, lightly lignified woods of dicotyledons of temperate regions. 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 295 SUMMARY AND CONCLUSIONS 1. An extensive survey of a wide range of gymnosperms and angio- sperms has shown that the structural pattern of the secondary wall is clearly visible in the large fiber-tracheids and libriform fibers of various dicotyledons. 2. By using untreated sections of such cells as controls, it is possible to observe the exact effects of specific chemical and mechanical treat- ments upon normal structures, and thus to extend the scope of investi- gation to cover a wide range of less favorable material. he cellulosic matrix of the swollen secondary wall of cotton, as of normal tracheids, fiber-tracheids, and libriform fibers, is an extremely heterogeneous but firmly coherent structure, the finer details of which grade down to the limits of microscopic visibility. 4. There is no reliable evidence to indicate that the matrix is com- posed of discrete entities of visible size — e.g., elementary fibrils, derma- tosomes, ellipsoidal bodies, etc. — that are bound together by non- cellulosic material. On the contrary, our data demonstrate that such putative entities actually are heterogeneous fragments that are shredded or disrupted from an originally continuous and coherent matrix. If there are discontinuities in the structural pattern of the cellulose in normal tracheary cells, they are confined to the submicroscopic field, e.g., to the realm of micelles or molecular chains. The visible structural pattern of the cellulosic matrix varies greatly in form and texture, not only in different plants, but also in homologous cells of the same plant, and even in different parts of the same cell. 6. There are at least two optically different elongated complexes of cellulose which may be segregated into radio-helical, radio-longitudinal, or concentric-longitudinal lamellae, or into various radio-concentric patterns. 7. The orientation of the elongated complexes of the structural pat- tern fluctuates more or less in successively formed parts of the secondary wall. In the case of normal tracheids, fiber-tracheids, and libriform fibers, there are three layers due to varying orientations: narrow inner and outer layers, in which the orientation is more nearly at right angles to the longitudinal axis of the cell, and a central layer of varying width, in which the orientation is parallel to this axis or does not deviate ex- cessively from it. “Lignin” and other non-cellulosic constituents may be deposited in the elongated, intercommunicating interstices of the cellulosic ‘matrix, thus resulting in two continuous, interpenetrating systems. In heavily 296 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI lignified forms, either system may be dissolved without seriously modi- fying the structural pattern of the remaining system. The purified cellulose and the “lignin” residue reveal positive and negative images of the original structural pattern. 9. Deviations from the typical 3-layered type of secondary wall are of not infrequent occurrence. Thus, many thick-walled libriform fibers and fiber-tracheids have no clearly differentiated inner layer; whereas others have more than three layers of varying “‘fibrillar” orientation. 10. Conspicuous discontinuities in the structural pattern of the cellulose commonly occur in the multiple-layered walls of so-called gela- tinous fibers, in certain types of bast fibers, and in sclereids. They are due to narrow layers of truly isotropic material which contain little, if any, cellulose. 11. There are five different types of visible concentricities which occur in varying combinations, and may be associated at times with radio-helical or radio-longitudinal lamellae. Therefore, it is misleading and fruitless to attempt to homologize all types of fibers in a single struc- tural model. LITERATURE CITED 1. Astsury, W. T., Marwick, T. C., and Bernat, J. D. X-ray —— of the structure of the wall of Valonia ventricosa. (Proc Soc. London B, 109: 443-450, 1932. 2. BALLs, ; e existence of daily ae rings in the cell wall of cotton. (Proc. Roy. Soc. London B, 90: 542-555. 1919.) 3. Barus, W. L. and HaANcock, H. A. at observations on cell wall structure as seen in cotton hairs. (Proc. Roy. Soc. London B, 3: 426-440. 4. CrarKk, G. L. Cellulose as it is completely revealed by X-rays. (Ind. Eng. Chem. a 474-487. 1930.) 5. CorRENS, r Kenntnis der inn ia eae: der vegetabilischen og abiouen, (Jahrb. Wiss. Bot. 28: 254-338. 1892. 6. ———— Zur Kenntnis der inneren a einiger Algenmem- cate (Zimmermann’s Beitr. Morph. Physiol. ST areata : 260-305. 1893. ) 7. pelts L. Das Mikroskop und seine Anwendung. (Braunschweig. Ist ed. 1867-69. 8. Dorér, C. The methods of cellulose chemistry. (London, 1933.) 9. Farr, W. K. and Ecxerson, S. H. Formation of cellulose membranes by microscopic particles of uniform size in linear arrangement. (Contrib. — Thompson aed 6: 189-203. 1934.) 10. Farr, W. K. and Eckerson, S. H. Separation of cellulose particles in ene of cotton fibers by paces with hydrochloric acid. (Contrib. wig Thompson Inst. : 309-313. 1934. RR, W. SISSON ‘ “ee diffraction patterns of cellu- lose pe and interpretation of cellulose diffraction data. (Contrib, Boyce Thompson Inst. 6: 315-321. 1934.) — — 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 297 12. — ay — on FREUDENBERG, K., ZocHER, H., and DUrr, W. Weitere Versuche mit Lignin. XI. Mit tteilung iiber Lignin ne Cellulose. (Ber. Deutsch. Chem. Gesell. 62: ss gece 1 REY-WYSSLING, A. e Stoffausscheidung der hoheren Pflanzen. Hariow, W. M. Contributions to the chemistry of the plant cell wall. II. Lignification in the secondary and tertiary layers of the cell walls of wood. (Tech. Pub. N. Y. State Coll. Forestry, no. 24. : Herzoc, R. O. Der Feinbau der Zellulosefaser und seine Bedeutung 12 1 fiir die Aufschliessung. (Papier Fabr. 23: 121-122. 1925.) . Howunet, F. von. Ueber den Einfluss des Rindendruckes auf die Beschaffenheit der Bastfasern der Dikotylen, (Jahrb. Wiss. Bot. 15: 311-326. 1884. Katz, J. R. Micellartheorie und Quellung der Zellulose. (IN HEss, K. Die Chemie der Zellulose und ihrer Begleiter. Leipzig, 1928.) . Kerr, T. and Bar.ey, I. W. he cambium and its derivative tissues. X. Structure, optical properties, and chemical composition of the so-called middle lamella. (Jour. Arnold Arb. 15: 327-349. 1934.) . Kraemer, E. O. and Lansinc, W. D. The molecular weight of cellu- loses and cellulose derivatives. (Amer. Chem. Soc., ie oe ymposium on the nature of cellul . Ltprxke, M. Zur Kenntnis der pflanzlichen Zellmembran. a *(Liebig’s reas 466: 27 57. 1928.) Untersuchungen uber Aufbau und Bildung der 2 eae Zellmembran und ihrer stofflichen Komponenten. (Biochem Zeitschr. 233: 1-57. 1931.) erden und Serer g der pflanzlichen Zellmembran. fp eipelcian. 22: 457-488. 1934. . Nace, C. Ueber den inneren Bau der vegetabilischen Zellenmem- 8 oe 2: branen. (Sitzber. Bay. Akad. Wiss. Miinchen, 1: 282- 114-171. 1864. ) . Neate, S. M. The modification of natural cotton cellulose by swelling 1933.) and degradation. (Trans. Faraday Soc. 29: 228-238. . Preston, R. D. The organization of the cell wall of the conifer tracheid. (Phil. Trans. Roy. Soc. London B, 224: 131-174. Oey . Ritter, G. Composition ne structure of the cell wall of woo (Ind. Eng. ‘Chem. 20: 941-945. 1928.) ructure of the cell wall of wood fibers. (Paper Industry, ——— St 16: 178-183. 1934. . Rirrer, G. J. and Strttwett, C. W. Rate of formation of the c talline structure of wood fibers. (Paper Trade Jour. 98(22): 37-40, 1934.) . SCHWENDENER, S. Ueber die Verschiebungen der Bastfasern im Sinne Von Hohnel’s. (Ber. Deutsch. Bot. Gesell. 12: 234-248. 1894.) . Van Iterson, G. Die wording van den plantaardigen celwand. (Chem. Weekblad, 24: 166-187 1927.) Biologische inleidung tot het cellulose symposium. (Chem. 1933.) Leda, 30: oe x Wiesn_er, J. Untersuchungen iiber die Organisation der vegetabili- aes “Zelih’iute.. "(Sitzber. Akad. Wiss. Wien, 93(Abt. 1): 17-80. 1886. ) 298 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI DESCRIPTION OF PLATES Figs. 1-6 and 13 were made from unstained sections and were photo- graphed in ee light between crossed _ nicols. e remaining wae rier were made with an arc-light and Zettnow’s filter. Figs. 1 _ 10, and 21 were made from sections mounted in diaphane (N = PLATE 140 ; hh aa da simplicifolius Brong. & Gris. Transverse — f the xy ylem, showing a fiber-tracheid and parts of seven adjoin- a 08 es inner layer, and a wide intervening dark lay e, this thickness, 15 p, the isotropic ena substance and the or fe TQ bo ~~ ~ Q & x Q ia) s) 8 = 2s = 't Q ~~ aS _ 2) % = ron » oe a o =) 2 c cat = | a wn a ,a) sy ° 3 io) eh ct = a) liant. The intercellular substance, the feebly birefringent a walls, and the inner and outer layers of the secondary wall are dark. A bordered pit is shown in the center of the photomnicrogra h. endron — Sieb. & Zucc Transverse section of fe a we feebly birefringent primary walls and a truly isotropic aoe “of 1400 . Myodocarpus simplicifolius. Transverse section of the xylem, showing a fiber-tracheid and parts of seven adjoining ones. e thick secondary wall of the central cell consists of a series of a 750. a gg as Urandra corniculata. Thick secondary wall of a sclerenchyma- tous cell in sectional view, showing jee Ge brilliant and dark layers. X 1750. 2 a Sal PLATE 141 Fig. 6. Pandanus odoratissimus L. Transverse section of a group of lignified fibers, showing meats walls composed of regularly alternating ee and dark layers Fig. 7. The same. Transverse section of a fiber after standard treat- ment with 72% sulphuric acid, staining with Haidenhain’s 08 matoxylin, and mounting in balsam m, showing residue of s ondary wall. The brilliant layers of Fig. 6 are split and em- bossed. 1300. PLATE 142 . eee bifida (P. & E.) A. DC. Transverse section of a racheid and of parts of several adjoining cells, ante: with “Haidenhain’ s haematoxylin and safranin, showing zones of oo page aali of peewee s x 2000. . The Transverse section of a fiber-tracheid after standard nhitelains ‘with 72% et aie acid, staining with Haidenhain’s 52) " 00 = m9 Ne) 1935] BAILEY AND KERR, STRUCTURE OF THE SECONDARY WALL 299 — ios) — _ — ON haematoxylin and mounting in aniline oil, showing finely radio- reticulate pattern and zones due to varying intensities of ligni- fication. Dark zones heavily lignified, light zones less ened lignified. 1300. PLATE 143 : ee bifida. Transverse section of a fiber- ae aan of The ame. Oe diverse Section ofa fe eae after standard haematoxylin, and mounting in aniline oil, showing radially stri- ated and finely reticulated residue of the central layer of the secondary wall. In the outer part of the central ae a is a concentricity due to an ab orupt transition from co o fine texture. The inner concentricity is due to ane ee of lignification. x 1900. PLATE 144 Tetramerista glabra Miq. Tangential longitudinal section of the central layer of a fiber-tracheid after treatment with 72% sul- phuric acid, staining with Haidenhain’ S “haematoxylin, and anastomosing threadlike components. The longitudinal orienta- tion has been somewhat distorted during swelling. x 1900. Siparuna bifida. Tangential longitudinal section through i central layer of a fiber -tracheid mounted in water and phot graphed with polarized light between crossed _ nicols, oe alternating pe hiab and eee striae. X 1900. The same. nsverse section o elignified fiber- tracheid, after treatment with diluted peer hydroxide and stain- ing with congo red. The denser, darker radii of the purified cellulose correspond to the lighter radii of Figs. 10 and 11. 200. PLATE 145 . Poraqueiba sericea Tul. Transverse section of the secondary fibe of a fiber- hae ‘after standard treatment with 72% sul- phuric acid, staining with Haidenhain’s haematoxylin, and mounting in aniline oil, showing concentrically lamellated resi- due of the central layer. 200. . Stparuna bifida. Transverse section of the secondary wall of a fiber-tracheid after standard treatment with 72% sulphuric ons in line oil, showing finely radio- reticulate residue of the een layer. X 3200. ATE 146 Gossypium hirsutum L. Transverse section of a cotton hair width and porosity in the ‘nner part of the secondary wall. A particle 1 w in diameter in the untreated wall would expand to the size of the circle at (a). X 1200. Owing to swelling, the original width of the lamellae has been increased 7500 times in this photomicrograph. 300 Fig. bo So ae: bo Ww bo & oi bo N JOURNAL OF THE ARNOLD ARBORETUM (VOL, XvI . Pinus ponderosa Dougl. Transverse accion of the secondary wall of a tracheid after treatment with 72% sulphuric acid, stain- ing with Haidenhain’s haematoxylin, and mounting in aniline oil, showing concentrically Gasol residue of the central layer. 1900. PLate 147 pe il iaar glabra. Transverse section of a fiber-tracheid and p several adjoining cells, mounted in a dilute aqueous n Zeiss 70-water-immersion lens. The broad central layer has a coarsely radio-concentric pattern which is complicated by zones of varying intensities of lignification. ‘he same. Transverse section of a fiber-tracheid after treat- ment with 72% sulphuric acid, staining with Haidenhain’s haematoxylin, and mounting in balsam, showin nely radio- concentric pattern and broad zones due to varying intensities of lignification. 1300. PLATE 148 . Homalium foe liebe F. Villar. Transverse section of a libri- form fiber stained with Haidenhain’s haematoxylin and safranin, showing ea broad cellulosic and narrow non- cellulosic layers. e radio-reticulate structure of the former layers is vaguely visible. 3200. The same. Transverse section of a libriform fiber after standard treatment with 72% sulphuric acid, staining with Haidenhain’s haematoxylin, and mounting in aniline, showing residue of both the cellulosic and the non-cellulosic layers. x 1300. PLATE 149 . Rhizophora mangle L. Isolated, delignified, libriform fiber, swollen in diluted Schweizer’s reagent, showing beadlike swell- ing of the central layer of the secondary wall. The outer layer of parang | wall is resolved into a series of constricting rings and hel cal bands. X 6 ; Olmediella Betschleriana (Goepp.) Loes. ee delignified, he libriform fiber, swollen iluted Schweizer’s reagent. outer layer of the eng ee wall is i into constricting helical bands. KYAS . Pandanus odoratissimus. Isolated delignified fiber, swollen in diluted Schweizer’s reagent, showing that each of the internal brilliant layers in Fig. 6 may be resolved into constricting rings and helical bands. 650. Olmediella Betschleriana. Segment of a libriform ~ iso- ated from a thick transverse ie of the xylem a elig- nification and treatment with 50% na Liendeat acid. e con- e 50. . OlmedieHa Betschleriana. Isolat ed, delignified libriform fiber, swollen in diluted Schweizer’s reagent. The outer layer of the secondary wall is resolved into a series ‘of constricting rings and helical bands. 400. ARNOLD ARBORETUM, Harvarp UNIVERSITY, AND CARNEGIE INSTITUTION OF WASHINGTON. Jour. ArNotp Ars. VoL. XVI PriaTeE 140 VISIBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE-— MERIDEN Jour. Arnotp Ars. VoL. XVI Pate 141 ViIstBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE—~ MERIDEN Jour. ARNoLp Ars. VoL. XVI PLATE 142 VISIBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE—~ MERIDEN Jour. ArRNoLp Ars. VoL. XVI PLATE 143 VISIBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE—— MERIDEN Jour. ARNoLD Ars. VoL. XVI PLATE 144 et VISIBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE — MERIDEN Jour. ArNotp Ars. VoL. XVI Piate 145 a yee ae ea ‘ i ety " _— . VISIBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE — MERIDEN Jour. ARNoLp Ars. VoL. XVI PLATE 146 VISIBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE — MERIDEN Jour. ARNoLtp Ars. VoL. XVI PLATE 147 VISIBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE — MERIDEN Jour. ArNoLp Ars. VoL. XVI PLATE 148 ie VISIBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE —~ MERIDEN Jour. ARNoLp Ars, VoL. XVI PLate 149 VISIBLE STRUCTURE OF THE SECONDARY WALL FULL-TONE — MERIDEN 1935] SAX, TEMPERATURE AND NUCLEAR DIFFERENTIATION 301 THE EFFECT OF TEMPERATURE ON NUCLEAR DIFFERENTIATION IN MICROSPORE DEVELOPMENT KARL SAX With one text figure and plate 150 “HEREDITY is effected by the transmission of a nuclear preformation which, in the course of development, finds expression in a process of cytoplasmic epigenesis” (Wilson, 1925). The evidence from genetic and cytological investigations has proven conclusively that nuclear pre- formation is dependent upon the genic constitution of the chromosomes. The mechanism of expression in cytoplasmic epigenesis is more obscure. The problem is difficult because it is not subject to direct attack. A comparison of induced and hereditary effects has provided a method for studying certain developmental processes in Drosophila, An analysis of the effect of temperature on developing microspores has provided some information regarding nuclear cytoplasmic relations in differentiation and development, and has some bearing on the problem of genic ex- pression. Normal microspore development in Tradescantia has been described in detail by Sax and Edmonds (1933). The young microspore contains a centrally located nucleus surrounded by cytoplasmic granules. The granules disappear, and the nucleus migrates to the end of the oval- shaped microspore. The cytoplasm is massed around the nucleus, and at the other end of the cell there is a large vacuole. There is then a migration of cytoplasm and vacuole so that two vacuoles are formed, one at each end of the cell. Most of the cytoplasm lies between the vacu- oles so that the longer axis of the cytoplasmic mass lies in the short axis of the cell. The nucleus at this time lies toward the heavy or dorsal wall of the microspore, — originally the inner wall at the time of tetrad formation. When the nucleus divides, the daughter nucleus near the heavy wall of the microspore is enclosed by a thin temporary wall which includes little cytoplasm. This nucleus does not pass into the typical resting stage, but retains its chromaticity and finally elongates to form the generative nucleus. The other nucleus formed near the center of the cytoplasmic mass enlarges to form the inactive tube nucleus. Shortly after the division of the microspore nucleus, the vacuoles disappear, and the cytoplasm appears to be rather homogeneous. The normal development of the microspore of Pseudolarix amabilis 302 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI resembles that of Pinus, described in detail by Coulter and Chamberlain (1901). Soon after the release of the microspore from the wall of the microsporocyte, the wings develop rapidly on opposite ends of the spore towards the ventral side of the cell, while the dorsal side of the spore, which was formed during meiosis, becomes thickened. At the time of the first nuclear division, the nucleus lies near the dorsal wall surrounded by most of the cytoplasm, and the region towards the wings and the ventral side of the spore are more vacuolate. The first division occurs across the short axis of the microspore, and the nucleus near the dorsal wall is cut off and degenerates. The other nucleus divides again in the same axis, and another prothallial cell is cut off. The third division produces the generative nucleus and the tube nucleus of the mature pollen grain. The generative nucleus lies near the inner or dorsal wall of the pollen grain and is cut off by a thin wall which encloses little cytoplasm, while the large tube nucleus lies free near the center of the cell. THE EFFECT OF TEMPERATURE ON NUCLEAR DIFFERENTIATION Both high and low temperatures are effective in producing abnormal development in the microspores of Tradescantia. Plants were placed in a constant temperature chamber where the temperature was maintained at about 6°C. for the cold treatment, or at about 35°C. for the heat treatment. Three days’ treatment was sufficient to produce abnormal. development at either temperature range, and doubtless a shorter time would be effective at the higher temperature. The microspores were examined soon after exposure to abnormal temperatures, and for several subsequent days after they had been placed in the normal greenhouse environment. Three types of abnormalities were produced. Under normal condi- tions, only two nuclei are formed in the pollen grain,—the generative nucleus and the tube nucleus. The tube nucleus normally does not divide and is inactive in further development. The heat treatment occasionally causes the tube nucleus to divide in Tradescantia (Figs. 1 and 2). The subsequent fate of the daughter nuclei is not known. Apparently no great deviation from normal environment is required to produce a second division, since this behavior was observed by Mr. R. H. Goodwin in Tradescantia plants grown in the greenhouse at the Bio- logical Institute of Harvard University. The second type of abnormality is produced by either cold or heat treatment. The polarity of the microspore is disturbed so that the nuclear division is no longer oriented across the short axis of the cell. 1935] SAX, TEMPERATURE AND NUCLEAR DIFFERENTIATION 303 In extreme cases the division is at right angles to the normal axis, the daughter nuclei are not differentiated, and the temporary cell wall is formed across the center of the microspore (Fig. 3). In most cases there is partial differentiation of the daughter nuclei, but the more compact nucleus does not elongate to form the typical generative nucleus (Fig. 4). The degree of differentiation of the two nuclei is closely associated with the angle of division. With a smaller degree of variation from the normal axis of division there is increased differentiation of the genera- tive nucleus (Fig. 5). In a single anther all degrees of differentiation are found, including the normal condition (Fig. 6). In order to determine more accurately the relation between the angle of division and the differentiation of the nuclei, a statistical study was made. All microspores measured were from a single flower taken from a plant which had been kept at a temperature of about S0-( tor thréé days. Camera lucida drawings were made of 163 microspores selected at random among those which showed the two nuclei in the same focal plane. The angle between the normal axis of division across the short diameter of the microspore and the line drawn through the centers of the two nuclei was taken as the angle of division. The length of the outline of the “generative” nucleus was measured in millimeters. This work, as well as the tabulation and analysis of the data, was done by my wife, Dr. Hally Jolivette Sax. The relation between the angle of division and the length of the generative nucleus is shown in Table 1. The high correlation of —.81 + .02 shows that the degree of nuclear differentiation is closely associated with the angle of division. A third type of aberrant development was found in microspores which began to “germinate” before the division of the nucleus. In some cases one of the daughter nuclei was found in the original microspore and the other in the newly-formed outgrowth (Figs. 8 and 10). In these cases a thin cell wall divided the cytoplasm into approximately equal parts, and there was no indication of nuclear differentiation. The division may occur so that neither daughter nucleus remains in the original micro- spore (Figs. 7 and 9). If the division is oriented lengthwise of the cyto- plasmic mass, there is no nuclear differentiation, but if it is oriented across the short diameter of the outgrowth, so that one daughter nucleus is near the cell wall, there is a differentiation which resembles that fol- lowing normal diyision in a normal microspore (Fig. 7). The precocious growth of the microspores of Tradescantia is unlike normal pollen-tube growth. The pollen-tube usually grows from the end of the pollen grain adjacent to the heavy dorsal wall, while the aberrant outgrowth occurs at the ventral side of the microspore. Per- 304 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVE haps this abnormal growth is the first stage in the development of an embryo-sac-like structure such as Stow (1930, 1933) has found in the anthers of Hyacinthus. Unfortunately, a study of the further develop- ment of the abnormal growth in Tradescantia microspores could not be continued because the anthers dehisced and disintegrated so soon. Pos- sibly these peculiar microspores could be developed further in a nutrient solution. TABLE L THE — BETWEEN ANGLE OF DIVISION AND THE DIFFERENTIATION OF THE NUCLEI IN MICROSPORES OF TRADESCANTIA Angle of division 2 0 10 2 30 40 50 60 70 80 90 p 8-11 ) 1 2 a 2 & 3 3 =/ 11 2 1 5 3 8 3 9 vo! 14 i t 3 tf Y «# & 2 42 S| 17 5 4 3 3. 2 = | 20 4 7 7 2 2 2 4 4 o 23 1 4 5 1 1 1 1 hoa 9 3. 3 1 1 2 | 29 3 1 1 S| 32 3 1 |) 35 | 38-40 4 1 n = 163 r=—81 + 02 The development and differentiation of the microspores of Pseudo- larix amabilis is also affected by environmental conditions. Branches containing male flowers were placed in a warm corner of the greenhouse for about two weeks. During this time meiosis occurred, and the micro- spores developed to maturity. Most of the microspores were normal in their development (Figs. 11 and 12), but various types of abnormalities were observed. If the first division occurs lengthwise of the cell in the axis of the wings, there is no differentiation of the nuclei if each is an equal distance from the cell wall (Fig. 13). If, however, one nucleus lies near the cell wall, regardless of the orientation of the division spindle, this nucleus tends to remain small and form a prothallial cell, while the nucleus near the center of the cytoplasmic mass remains large and divides again (Figs. 14 and 16). The first division may occur in the normal position, but the second division may be aberrant in orienta- tion (Figs. 15 and 16). Several years ago Mrs. Sax found a mature pollen grain of Picea which contained four undifferentiated nuclei of 1935] SAX, TEMPERATURE AND NUCLEAR DIFFERENTIATION 305 approximately equal size. Apparently relatively slight changes in en- vironmental conditions can cause abnormal development of conifer microspores. Nuclear differentiation in these microspores appears to be entirely dependent upon the orientation of the division spindles and the position of the nuclei in relation to the cytoplasmic mass. Nuclear differentiation in the microspores of Tradescantia and Pseu- dolarix appears to be determined by the nuclear cytoplasmic relation- ships. In Tradescantia it is possible to observe the relations of nuclei, vacuoles, and cytoplasm in the living microspores. Observations at vari- ous stages of development show that normal development is dependent upon the synchronization of cytoplasmic and nuclear activities. In the microspores which develop under normal conditions, the cell contents show a gradual shifting in position before the nucleus divides. A large vacuole is formed at one end of the microspore, and the nucleus and most of the cytoplasm move to the opposite end. The vacuole then extends towards the opposite end, near the ventral side of the spore, and finally forms two vacuoles, one at either end of the cell. Meanwhile the cytoplasmic mass and the nucleus migrate towards the center of the cell. The cytoplasm extends between the ventral and dorsal walls so that the length of the cytoplasmic mass is across the short diameter of the cell. Some cytoplasm extends around the entire periphery of the cell. The nucleus lies near the dorsal wall at the time of division, After the division the nucleus near the dorsal wall is cut off by a thin temporary wall, and then develops into the elongated gen- erative nucleus. The other nucleus enlarges, loses its chromaticity, and becomes the inactive tube nucleus. The sequence of early development of the normal microspore is shown in text figures a, b, c, and d, which are camera lucida sketches drawn from living material. When the microspore develops at low temperatures, the same cycle of development begins, but the nucleus divides before the vacuole, cyto- plasm, and nucleus reach their normal positions. (Text figure e.) Since the cytoplasm migrates towards the center of the cell along the dorsal wall of the microspore, the length of the cytoplasmic mass is at an angle to the normal axis of division. The nuclear spindle is oriented in the long axis of the cytoplasmic mass, and the daughter nuclei lie towards one end of the microspore and do not undergo complete differentiation. The exposure to high temperatures for several days appears to accel- erate the cytoplasmic movement without causing a corresponding activity of the nucleus. At the time the cell constituents are in the position usually associated with nuclear division (Text figure c), the nucleus may remain inactive. The vacuoles then become smaller or may dis- 306 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI appear entirely before nuclear division. As a result, the long axis of the cytoplasmic mass is not oriented in the short axis of the cell, and the division may occur at various angles, depending in part upon the cyto- plasmic distribution (Text figures f, g, and h). In a single flower, following heat treatment, the angle of division may vary from O to 90 degrees. Camera lucida drawings were made from a random sample of these cells, and the angle of division was determined in relation to the distribution of cytoplasmic mass. The length of the tT Ficgure. Development of the microspore under normal and ab- normal conditions. All eek are from camera lucida drawings of living microspores. Figur b, c, and d show the movement of the cell contents during early ieee of the microspore under normal conditions. The vacuole becomes extended along the ventral wall and finally forms two vacuoles The nuc ne at the time of division becomes oriented near the dorsal or heavy w the microspore. Figure e yun the nuclei formed by nuclear division before normal Be Bi rth of the cell constituents is attained. This microspore developed at a low temperature, which seems to io cytoplasmic movement with- out retarding nuclear division. Figures f, g, and h are drawings of microspores pens eg been sub- ick to a high temperature. The vacuoles are small or ab at the time of nuclear division, and the axis of division tends to occur in ee long axis of the cytoplasmic mass. cytoplasmic mass was determined for the long axis of the cell, and the width was measured across the short axis of the cell. For example, in a normal microspore (fig. d) the length of the cytoplasmic mass is the distance between the vacuoles, and the width is the distance across the microspore in the axis of division, and the angle of division is very small. In this cell the length-width ratio is about 0.5. In figure e the length- width ratio is about 1.0, and the angle of division is about 45 degrees, while in figure h the length-width ratio of the cytoplasmic mass is about 2.0, and the angle of division is about 80 degrees. The length-width 1935] SAX, TEMPERATURE AND NUCLEAR DIFFERENTIATION 307 ratio of the cytoplasmic mass correlated with the angle of division gave a value of r = .66 + .03. It is evident that there is a strong tendency for the nucleus to divide in the long axis of the cytoplasmic mass, al- though as the volume of cytoplasm increases, there is not a correspond- ing tendency for the nucleus to divide in the longer axis. In general, however, the direction of division in the.microspore is controlled by the distribution of the cytoplasm in accord with Hertwig’s rule, and the nuclear differentiation is controlled by the position of the daughter nuclei in relation to the cytoplasmic mass. It is not possible to follow the cytoplasmic movements in the living cells of conifers, but judging from the description of normal develop- ment (Ferguson, 1904) and the behavior of the nuclei in abnormal micro- spores, the failure of normal differentiation is also based on the dis- turbed relations of nucleus, cytoplasm, and vacuoles. A comparison of nuclear differentiation in Tradescantia and Pseudo- larix microspores and in the embryo sacs derived from microspores in Hyacinthus shows a good deal of similarity in polarity. In both Tradescantia and Pseudolarix the center of activity in early microspore development is near the dorsal wall which was formed during micro- sporogenesis. If a second division occurs in the Tradescantia micro- spore, the nucleus nearer the center of the cell divides, as is the case in normal microspore development in the conifers. The vegetative nucleus is always the one nearer the ventral wall and is surrounded by a large amount of cytoplasm, while the generative or sexual nucleus lies near the dorsal wall and is enclosed by a thin temporary cell wall which in- cludes little cytoplasm. The “embryo sacs” which develop from micro- spores of Hyacinthus (Stow, 1933) show the exine of the microspore at the egg or sexual end of the embryo sac, while the polar or vegetative nuclei lie in the center of the embryo sac, apparently surrounded by a relatively large amount of cytoplasm. Stow was able to induce embryo sac-like structures in anthers of Hyacinthus by subjecting the bulbs to a temperature of 28°C. for 18 to 24 hours at the time of planting in the fall. The abnormal development observed in the following spring may have been induced either by the temperature treatment or by the effects produced by the large number of degenerating sterile microspores. At any rate the differentiation of the microspore to produce a normal pollen grain or an embryo sac appears to depend upon environmental conditions. We are inclined to believe that the precocious growth of Tradescantia microspores is the first stage in embryo sac formation, and that the complete structure could be de- veloped, under temperature control, if the cells could be kept alive over a long period of time, as is the case in Hyacinthus. 308 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI FACTORS IN DEVELOPMENT AND DIFFERENTIATION According to Osterhout (1921), life is dependent upon a series of reactions which normally proceed at rates which bear a definite relation to each other. If for life we substitute development and differentiation, we have an hypothesis which seems to explain development and differen- tiation. Certainly the differentiation of the microspore nuclei appears to be dependent upon the relative rates of cytoplasmic migration and nuclear activity. Less direct evidence indicates that the difference be- tween a pollen grain and an embryo sac may be dependent upon the same type of timing relationships. If sex can be determined by the timing relationships of different reactions, effected either by environmental con- ditions or hereditary factors, we have indirect evidence that genic ex- pression may be effected by differential reactions. More direct evidence regarding the mechanism of genic expression is found in the behavior of the chromosomes at meiosis. A failure of chromosome pairing, or asynapsis, may be caused by genetic factors or by environmental conditions. Genetic asynapsis has been found in Drosophila, Zea, Triticum, Rumex and Datura. Induced asynapsis can be effected in Rhkoeo and Datura by subjecting the plants to low tem- peratures for several days, and it has been obtained in Tradescantia following treatment at low and high temperatures. Both the hereditary and induced effects are similar in their expression. The chromosomes are unpaired at meiosis, the divisions are irregular, diploid gametes may be produced, and there is a high degree of pollen sterility in the asynaptic plants. Both types of asynapsis may be attributed to the same cause. If chromosome development is not synchronized with other cell activi- ties, the chromosomes may not be effectively paired before nuclear divi- sion is initiated. An exceptionally clear case of the timing factor in genic activity was found in Aquilegia by Anderson and Abbe (1933). The “compacta” mutant of Aquilegia is dependent on a single genetic factor. In the mutant type the branches are more erect and numerous, and the flowers are upright from the beginning. The dwarf type is caused by the preco- cious thickening of the cell walls, and the somatic expression is simply the result of disturbed timing relationships, — “the precocious initia- tion of a normal feature of normal development.” Further aspects of the relation between genes and development in Drosophila have been discussed by Schultz (1935) and by Goldschmidt (1935), and similar work is being done on the cucurbits by Sinnott (unpublished). The temperature chambers used in this work were paid for, in part, by a grant from the American Academy of Arts and. Sciences. 1935] SAX, TEMPERATURE AND NUCLEAR DIFFERENTIATION 309 SUMMARY The subjection of Tradescantia plants to low and high temperatures may produce three kinds of abnormalities in microspore development. (1) The tube nucleus, which in normal microspores is inactive and ulti- mately degenerates, may divide. (2) The polarity of the cell may be disturbed so that the division of the microspore nucleus is not oriented in the normal axis. The angle of deviation is closely correlated with the differentiation of the daughter nuclei. (3) The microspore may become greatly extended on the ventral side, and the nuclear division may occur in this new outgrowth. This abnormality may be the first step in the transformation of a microspore to an “embryo sac,” as found in Hyacinthus by Stow. When Pseudolarix microspores are developed at a relatively high temperature, there is a failure of normal differentiation of the nuclei. The differentiation of prothallial cells and generative and tube nuclei is dependent upon the nuclear cytoplasmic relationships in the developing microspore. The normal differentiation in Tradescantia microspores is dependent upon the synchronization of cytoplasmic movements and nuclear activity. There is some evidence that many differences in development and differentiation, induced either by genetic factors or by environ- mental conditions, are dependent upon differences in reaction rates of different processes. LITERATURE CITED Anperson, E. and L. B. Aspe (1933). A comparative anatomical study of a mutant ee aaa (Amer. Nat. 67: 380-384. ) CouLter, J. M. and C. J. oe ie Morphology of sperma- tophytes. (D- SRaleien and Fercuson, M. C. (1904). Contihitans to the knowledge of the life history of Pinus, with special reference to sporogenesis, the develop- ment of the gametophytes, and fertilization. (Proc. Wash. Acad. Sci. —202. GotpscuMipt, R. (1935). Gen und Ausseneigenschaft. (Zeit. f. Ind. bstam. u. Vererb 69: 38-131. OsterHout, W. J. V. (1921). The mechanism of injury and recovery of the cell. (Science N. S. 53: 352- 356. ) Sax, Kart and H. W. Epmonps (1933). Development of the male game- tophyte in Tradescantia. (Bot. Gaz. 94: 156-163. Scuuttz, J. (1935). Aspects of the relation between genes and develop- ment in reget mer. Nat. 69: 54. Stow, I. (1930). Experimental studies on the formation of the embryo- sac-like giant pollen grain in the anther of Hyacinthus orientalis. (Cyto- logia, amar 933). On the female tendencies of the gayi giant pollen hd of Hyacinthus orientalis. (Cytologia, 5: 88— Witson, E. B. (1925). The cell in development and oo (The Mac- millan Co. New York. ) 310 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XvI DESCRIPTION OF PLATE 150 Photographs of aceto-carmine preparations of abnormal microspores o Tradescantia and normal and abnormal microspores of Pseudolarix ama- bilis. Figures 1 to 6 inclusive, magnified xX 800. The other figures are magnified < 600. TRADESCANTIA diab land 2. Division of the tube nucleus of the microspore. Figure 3. The division of the microspore nucleus has occurred at right gates to the normal axis of division, and the daughter nuclei do not beco Figures 4,5, and 6. The nuclear divisions have occurred at various angles ollowed by a corresponding amount of nuclear differentiation. hese pramee nat were developed at a high temperature treatment PSEUDOLARIX Figures 11 and 12. Stages in the normal development of the microspore. All divisions are across the short axis of the cell, and the prothallial cells are always cut off near the heavy dorsal wall. Figure 13, 14, 15, and 16. Abnormal development induced by heat cee nt. The nuclei may divide at various angles followed by var ae of diferentiation of the daughter nuclei. The prothallial cells may be cut off at any point along the cell wall. CyTOLOGICAL LABORATORY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. Jour. ARNoLD Ars. VoL. XVI PiaTeE 150 i red ae TEMPERATURE AND NUCLEAR DIFFERENTIATION FULL-TONE — MERIDEN OS 7 7 Oe 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 311 NOTES ON THE LIGNEOUS PLANTS DESCRIBED BY LEVEILLE FROM EASTERN ASIA? ALFRED REHDER LABIATAE Leucosceptrum sinense Hemsley in Jour. Linn. Soc. Bot. 26: 310 (1890). — Léveillé, Fl. Kouy-Tchéou, 209 (1914).— Dunn in Not. Bot. Gard. Edinb. 8: 171 (1913); 6: 192 (1915). Elsholtzia Cavaleriei Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 8: 424 (1910). Leucosceptrum Bodinieri Léveillé in op. cit. 9: 224 (1911). CHINA. Kweichou: environs de Tsin-gay, au bord d’une riviere, E. Bodinier, no. 2709, Sept. 20, 1899 “sous -arbrisseau, fl. roses” (syntype of Elsholtzia Cavaleriei; photo. in A. A.); environs de Tou- chan, J. Cavalerie in herb. Bodinier, no. 2710, Sept. 1899 (syntype of Elsholtzia Cavaleriei; photo. in A. A.). The name Elsholtzia Cavaleriei was changed by Léveillé to Leuco- sceptrum Bodinieri and later the two type specimens were enumerated by him in his Flore du Kouy-Tschéou under L. sinense without citation of synonymy, the reduction being based on identifications made by Dunn, to whom Léveillé had sent material of his Labiatae for revision, as it appears from a note in Léveillé’s Flore du Kouy-Tchéou p. 203 under Labiacées which reads “(D. Dunn revisit ).” Leucosceptrum plectranthoideum (Lévl.) Marquand in Kew Bull. Misc. Inform. 1930: 207. Buddleia plectranthoidea Léveillé, Cat. Pl. Yun-Nan, 171 (1916). Cuina. Yunnan: patures des montagnes a Pé-long-tsin, 3200 m., E. E. Maire, Nov. 1912 (holotype of Buddleia plectranthoidea; merotype in A. A.). This species seems nearest to L. sinense Hemsl. but can be at once distinguished by the shorter inflorescence, the yellowish closer tomentum of the calyx and the bracts, and the shorter elliptic or ovate-elliptic to oblong-elliptic leaves reticulate beneath and tomentulose above. Colquhounia Seguini Vaniot in Bull. Acad. Intern. Géog. Bot. 14: 165 (1904).— Rehder in Sargent, Pl. Wilson. 3: 380 (1916).— Bese from Vol, 15: 326; for preceding he see Vol. 10: 108-132, 164-196; 12: 275-281; 13: 299-332; 14: 223-252; 15: 1- 117 312 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Léveillé, Cat. Il. Seu-Tchouen, 92, pl. 44 (1918). — P’ei, Verben. China in Mem. Sci. Soc. China, 1 (no. 3): 180 (1932). Colquhounita elegans Wall. var. pauciflora Prain in Jour. As. Soc. Beng. 62: 38 (1893).— Dunn in Not. Bot. Gard. Edinb. 6: 179 1915). nani fluminis Léveillé, Sert. Yunn. 3 (1916); Cat. Pl. Yun- Nan, 298 (1917). — P’ei, Verben. China in Mem. Sci. Soc. China, 1(no. 3): 180 (1932). Cuina. Kweichou: environs de Ou-la-gay (Tchin-lin), J. Seguin in herb. Bodinier, no. 2237, March 1898 “‘longues tiges sous-lig- neuses, lianeuses” (syntype of C. Seguini; photo. in A. A.); rives du fleuve Bleu, alt. 450 m., E. E. Maire, June 1912, “petit arbuste, feuilles persistantes, fleurs roses” (holotype of Caryopteris fluminis ; photo. in Yunnan: environs de My-tsao, Fr. Ducloux, no. 110, March 4, 1897, “long tiges s’enlagant aux arbres et buissons, fleurs coc- cinées, 4 graines ailées” (syntype of C. Seguini; photo. in A. A.). By Dunn (1. c.) this species was referred to C. elegans var. pauci- flora Prain, but as I pointed out in 1916 (1. c.) this variety is best con- sidered a distinct species which becomes C. Seguini Vaniot. Micromeria biflora Bentham, Labiat. 378 (1834).— Dunn in Not. Bot. Gard. Edinb. 6: 157 (1915). — Léveillé, Fl. Kouy-Tchéou, 210 (1914); Cat. Pl. Yun-Nan, 138 (1916). Thymus Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 11: 298 (1912). Cutna. Kweichou: Tin-lan, montagnes sablonneuses, J. Cavalerie, no. 3778, “blanc-rose” (syntype of Thymus Cavaleriei ; photo. inA.A.). Yunnan: Pan-pien-kai, paturages des coteaux calcaires, alt. 2550 m., E. E. Maire, Sept. 1911 “Thymus vivace, étalé, fl. roses” (syntype of Thymus Cavaleriei; photo. in A. A.). Elsholtzia rugulosa Hemsley in Jour. Linn. Soc. 26: 278 (1890). — Léveillé, Fl. Kouy-Tchéou, 208 (1914); Cat. Pl. Yun-Nan, 138 (1916).— Dunn in Not. Bot. Gard. Edinb. 6: 149 (1915). Elsholtzia Labordei Vaniot in Bull. Acad. Intern. Géog. Bot. 14: 177 (1904). Cuina. Kweichou: environs de Tsin-gay, 4 Tchao-see, abonde dans la mont. et bord de routes, J. Laborde in herb. Bodinier, no. 2711, Sept. 7, 1899 “fleurs d’un bleu trés pale” (holotype of E. Labordei; photo. in A. A.). Elsholtzia fruticosa (D. Don) Rehder in Sargent, Pl. Wilson. 3: 381 (1916). Elsholtzia polystachya Bentham, Labiat. 116 (1832).— Dunn in Not. 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 313 Bot. Gard. Edinb. 8: 161 (1913); 6: 149 (1915). —Léveillé, FI. Kouy-Tchéou, 208 (1914) ; Cat. Pl. Yun-nan, 138 (1916). Elsholtzia tristis Léveillé in Fedde, Rep. Spec. Nov. 8: 424 (1910). Elsholtzia Dielsit Léveillé in op. cit. 9: 441 (1911). Elsholtzia Souliet Léveillé in op. cit. 9: 248, non p. 218 (1911). Cuina. Szechuan: Ta-tsien-lu, J. A. Soulié, nos. 781 and 1023, in 1893 (syntypes of E. Dielsii [E. Souliei Lévl. p. 248, non p. 218]; photos. in A. A.) Kweichou: environs de Kouy-yang, mont du College, E. Bodinier, no. 1944, Nov. 3, 1897, “tige 1 m., fleurs blanches” (holotype of E. tristis, photo. in A. A.). Elsholtzia ochroleuca Dunn in Not. Bot. Gard. Edinb. 8: 161 (1913). Elsholtzia lampradena Léveillé in Bull. ek Bot. 25: 25 (1915) ; Cat. Pl. Yun-Nan, 137 (1916). — Synon. CuIna. Yunnan: paturages des Sine a Tong-tchouan, alt. 2600 m., FE. E. Maire, Sept. 1912 “arbrisseau rameux, haut 0.40 m., fleurs blanches en é€pis dressés” (holotype of EL. lampradena; photo. in A. A.). Elsholtzia lampradena has been identified with E. ochroleuca accord- ing to a note on the type specimen. Pogostemon glaber Bentham in Wallich, Pl. As. Rar. 1: 31 (1830). — Léveillé, Cat. Pl. Yun-Nan, 143 (1916). ae Esqutroliit Léveillé in Fedde, Rep. ee Nov. 9: 449 Sea ye as uy Tchéou, 440 (1915). — Synon Cuina. Kweichou: “YTchou-ly, alt. 900 m., = pun no. 2053, March 1, 1910, “fl. blanche, labelle rose” (holotype of Caryop- teris Eeciaraia: merotype in A. Ao Pogostemon glaber has not yet been recorded from Kweichou, as far as I know, but it is known from Yunnan. Plectranthus ternifolius Don, Prodr. Fl. Nepal. 117 (1825).— Léveillé, Fl. Kouy-Tchéou 214 (1914); Cat. Pl. Yun-Nan 143 (1916).— Dunn in Not. Bot. Gard. Edinb. 6: 138 (1915). Elsholtzia Lychnitis Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 8: 425 (1910). Teucrium Esquirolii Léveillé in Bull. Géog. Bot. 22: 236 (1912). Cutina. Kweichou: Tchen-lin-tchéou, route de Lo-pie a Ou- la-gay, L. Martin in herb. Bodinier, no. 1937, Oct. 9, 1899, “tige de 1.50 m. de haut, fleur blanches, ou blanc-bleuatre” (syntype of Elsholtzia Lychnitis ; photo. in A. A.); route de Tou-tchéou a Pien-yang, J. Cava- lerie, no. 2573, Nov. 1905 (syntype of E. Lychnitis ; photo. in A. A.) ; coteaux de Lo-fou, J. Esquirol, no. 2576, Nov. 1910 (holotype of Teucrium Esquiroli; ex Léveillé). 314 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Of Teucrium Esquirolii | have seen no specimen, but Léveillé enu- merates it in 1916 (Cat. Pl. Yun-Nan, 143) as a synonym of Plec- tranthus ternifolius, probably on identification by S. T. Dunn, though Dunn does not cite it in 1915. It does not appear in the Flore du Kouy- Tchéou. Plectranthus ternifolius should probably not be classed as a ligneous plant, though Hooker describes it in his Flora of British India 4: 621) asa bush 3-5 ft. high. Plectranthus coetsa Hamilton ex Don, Prodr. Fl. Nepal. 117 (1825). — Léveillé, Cat. Pl. Yun-Nan, 141 (1916). Plectranthus Mairei Léveillé, Cat. Pl. Yun-Nan, 141 (1916), pro synon. P. coetsa Ham. Cuina. Yunnan: paturages des mont. derriere Tong-tchouan, 2700 m., E. E. Maire, Oct. (1912-13), “plante annuelle tres-rameuse, fl. rouge vif” (type of P. Mairei; photo. in A. A.) ; haies, plaine de Tché- hay, 2550 m., E. E. Maire, Sept. [1912-13], “plante vivace, tumescente, en touffes, haut 1.40 m.” (in herb. Léveillé with P. Mairez; photo. in Plectranthus Mairei is apparently an unpublished name and is cited only as a synonym of P. coetsa (1. c.).. On Mairei’s specimen from Tong-tchouan the name Plectranthus Mairei appears in Léveillé’s hand- writing; the specimen from Tché-hay is without any name, but placed in the cover of P. Mairei. The majority of Labiatae described by Léveillé are herbaceous and most of them have been examined by S. T. Dunn. He published his identifications in his Notes on Chinese Labiatae (in Not. Bot. Gard. Edinb. 8: 153-171. 1913) and in his Key to the Labiatae of China (op. cit. 6: 127-208. 1915). The identifications and reductions made by Dunn were accepted by Léveillé and incorporated in his Flore du Kouy-Tchéou (p. 203-217) and in his Catalogue des plantes de Yun- Nan (p. 136-149). SOLANACEAE Solanum aculeatissimum Jacquin, Coll. 1: 100 (1786); Ic. Rar. 1: t. 41 (1781-86). — Merrill in Contr. Arnold Arb. 8: 149 (1934). Solanum Bodinieri Léveillé & Vaniot in Bull. Soc. Bot. France, 55: 206 (1908). Solanum Cavaleriei Léveillé & Vaniot, 1. c. 207 (1908). — Synon. nov. Cuina. Hongkong: plage sablonneuse de Vile Verte, E. Bodinier, July 31, 1895 (holotype of S. Bodinieri; photo. in A. A.). Kweichou: Ly-po-hien, J. Cavalerie in herb. Bodinier, no, 2722, Aug. 10, 1899 (holotype of S. Cavaleriei; photo. in A. A.). Solanum Bodinieri represents a glabrescent form of this very variable 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 315 species, while S. Cavaleriei is much more pubescent throughout. Solanum Bodinieri has been identified with S. aculeatissimum by Mer- rl (1e.), SCROPHULARIACEAE Brandisia racemosa Hemsley in Kew Bull. Misc. Inform. 1895: 114. Deutzia funebris age tele Yunn. 1 (1916). — Cat. Pl. Yun-Nan, 296 (1917).— Syno Cuina. Kweichou: rives du fleuve Bleu a Kiang-pien, alt. 350 m., E. E. Maire, Aug. 1913, “arbuste un peu grimpant; fl. roses” (holotype of Deutzia funebris ; merotype in A. A.). BIGNONIACEAE Incarvillea Delavayi Bureau & Franchet in Jour. de Bot. 5: 138. (1891) Tecoma Maret Léveillé, Cat. Pl. Yun-Nan, 20 (1916). — Synon. nov. Cuina. Yunnan: rochers sous brousse, mont. de Pe-long-tsin, alt. 3200 m., E. E. Maire, May 1911, “plante vivace, fl. roses grandes” (holotype of Tecoma Mairei; photo. in A. A.). Though this is an A orbacions species, I am including it in this enu- meration, because Léveillé has described it under the ligneous genus Tecoma. ACANTHACEAE Phlogacanthus pubinervis T. Anderson in Jour. Linn. Soc. Bot. 9: 508 (1867). — Léveillé, Cat. Pl. Yun-Nan, 6 (1915). Aeschynanthus Dunnii Léveillé in Fedde, Rep. peer Nov. 9: 453 11); Fl. Kouy-Tchéou, 180 (1914). — Synon. n Lonicera Menelii Léveillé, Fl. Kouy-Tchéou, 63 (1914). — Synon. Cae. Kweichou: without precise locality, J. Esquirol, no. 737, “fleur rougeatre” (syntype of Aeschynanthus Dunnii; photo. in A. A.); Lo-fou, J. Cavalerie, no. 3475, March 1909 “couleur jaunatre” (syntype of Aeschynanthus Dunnii, in fruit; photo. in A. A.); Thing- mei, 1100 m., J. Esquirol, no. 3540, Dec. 2, 1913 (holotype of Lonicera Meneltu; merotype in A. A.). Cystacanthus yangtsekiangensis (Lévl.), comb. nov. Strobilanthes yangtsekiangensis Léveillé, Cat. Pl. Yun-Nan, 7 (1915). Cuina. Yunnan: rives du fleuve Bleu, alt. 400 m., E. E. Maire, May 1912, “plante sous-ligneuse en touffes dressées, fl. bleues” (syntype of Strobilanthes yangtsekiangensis, photo. in A. A.); rives du fleuve Bleu a Ta-tchai, alt. 450 m., “plante vivace, sous-ligneuse, en touffes, fl. roses” (syntype of S. yangtsekiangensis ; photo. in A. A.). 316 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI This species is very near C. yunnanensis W. W. Sm., but is easily dis- tinguished by the closer and finer pubescence of the young branchlets, the inflorescence and the calyx, and by the numerous lateral 2—6- flowered inflorescences along last year’s branches. To C. yangtse- kiangensis apparently belong Rock 8049 from Yunnan, between Tang- yueh and Likiangfu, and Schneider 671 from southern Szechuan, between Mo-so-ying and Kung-mu-ying. RUBIACEAE Oldenlandia Bodinieri(Lévl.) Chun in Sunyatsenia, 1: 310 (1934). Hedyotis Bodinieri Léveillé in Fedde, Rep. Spec. Nov. 11: 64 (1912). Cuina. Kwangtung: Tay-mo-chan, sommet de la mont., 3500 ft., E. Bodinier, no. 1158, May 7, 1895, “petit sous-arbrisseau croissant dans les rocailles, fleurs blanches” (holotype; photo. in A. A.). I have not been able to identify this plant with any described species. Oldenlandia macrostemon (Hook. & Arn.) Kuntze, Rev. Gen. 1: 292 (1891). — Pitard in Lecomte, FI. Gén. Indo-Chine, 3: 138 (1922). Hedyotis macrostemon ake & Arnott, Bot. oe Voy. 192 (1841). — Léveillé, Cat. Pl. Yun-Nan, 245 (191 Hedyotis Esquirolii Léveillé in Fedde, Rep. Shee Sais 13: 176 (1914) ; Fl. Kouy-Tchéou, 367 (1915). — Syn Oldenlandia Esquirolit (Leévl.) Chun in Sdn — 310 (1934). Cumna. Kweichou: without precise locality, J. Esquirol (holotype of H. Esquirolii; photo. in A. A.). Ophiorrhiza japonica Blume, Bijdr. 978 (1826). Ophiorrhiza Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 13: 177 1914 alle allel ee 1. c. (1914); Fl. Kouy-Tchéou, 370 15).— Synon Cuina. Kweic be ou: district de Tsin-gay, rocailles, bois a Kao-tchay, J. Laborde in herb. Bodinier, May 15, 1898 (holotype of O. Labordei; photo. in A. A.); without locality, J. Cavalerie (holotype of O. Cavaleriei ; ex Léveillé). Ophiorrhiza Cavaleriei is cited by Léveillé in his Flore du Kouy- Tchéou as a synonym of O. Labordei, but Cavalerie’s specimen is not cited, only Laborde’s specimen from Tsin-gay. Cavalerie’s specimen I have not seen. Ophiorrhiza cantoniensis Hance in Ann. Sci. Nat. sér. 4, 18: 222 (1862). — Léveillé, Fl. Kouy-Tchéou, 370 (1915). Ophiorrhiza Seguini Léveillé in Fedde, Rep. Spec. Nov. 13: 177 (1914). — Synon. nov 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 317 Ophiorrhiza violaceo-flammea Léveillé in Bull. Géog. Bot. 25: 47 (1915) ; Cat. Pl. Yun-Nan, 247 (1917). — Synon. nov. Cuina. Kweichou: environs de Gan-pin, croissant dans les rocailles, les trous entre des rochers, E. Bodinier, no. 1549, April 29, 1897, “fleurs blanches” (holotype of O. Seguini, photo. in A. A.). Yunnan: vallée de Li-tse-pin, 2700 m., E. E. Maire, April 1912, “sous-arbrisseau toujours vert, fl. violettes” (holotype of O. violaceo- flammea ; photo. in A. A.). Ophiorrhiza Seguini is not mentioned by Léveillé in his Flore du Kouy-Tchéou, but its type is enumerated, together with another speci- men, under O. cantoniensis, which shows that Léveillé had reduced it to O. cantoniensis. The color of the flowers on the specimen of O. violaceo-flammea said to be violet does not look at all different from the color of flowers of O. japonica. Ophiorrhiza cantoniensis is closely related to O. japonica Bl., but it may be distinguished by its oblong leaves attenuate at the base and quite glabrous, broadest at or above the middle, while O. japonica has shorter generally ovate or oblong leaves less attenuate or even nearly rounded at base and puberulous on the midrib beneath. Wendlandia ligustrina Wallich, Num. List. 6272 (1832), in part, nom. nud. — Don, Gen. Syst. 2: 518 (1834). — Léveillé, Cat. Pl. Yun- Nan, 242 (1917). — Cowan in Not. Bot. Gard. Edinb. 16: 242 (1932); 18: 183 (1934). Luculia gratissima Sw. sensu Léveillé, Fl. Kouy-Tchéou, 368 (1915), non Sweet (1826). Cutna. Kweichou: bords du Hoa-kiang, L. Martin in herb. Bodinier, no. 2563, Feb. 18, 1899, “petit arbuste, fleurs blanches’’ (photo. in A. A.). This collection extends the range of W. ligustrina into Kweichou. Martin’s specimen was identified with W. ligustrina by J. M. Cowan according to a note on the specimen. Wendlandia salicifolia Franchet in herb. ex Castello in Jour. de Bot. 9: 208 (1895). — Cowan in Not. Bot. Gard. Edinb. 16: 244 (1932). Ligustrum Thea Léveillé & Dunn in Fedde, Rep. Spec. Nov. 10: 147 (1911). — Léveillé, Fl. Kouy-Tchéou, 295 (1914), Cuina. Kweichou: without precise locality, J. Esquirol, no. 327, Dec. 16, 1904, “sous-arbrisseau des bords du fleuve, submergé aux grandes eaux; les feuilles donnent une infusion theiforme assez em- ployée; fleur blanc-rose” (holotype of Ligustrum Thea; photo. in A. A.); without precise locality, J. Esquirol, no. 239 (cited in Fl. Kouy- Tchéou; photo. in A. A.). 318 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Ligustrum Thea was first referred to W. salicifolia by Cowan in his “The Genus Wendlandia” (op. cit. 233-316). Wendlandia Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 10: 434 (1912); Fl. Kouy-Tchéou, 373 (1915).— Cowan in Not. Bot. Gard. Edinb. 16: 263 (1932). Wendlandia Feddei Léveillé in Fedde, Rep. Spec. Nov. 10: 434 (1912); Fl. Kouy-Tchéou, 373 (1915 Cutna. Kweichou: Lo-fou, J. Cavalerie, no. 3297, April 1907, ‘“fleurs blanches” (holotype of W. Cavaleriei; merotype in A. A.) ; route de Pin-fa a Lo-fou, J. Cavalerie, no. 2732, April 4, 1906 (holotype of W. Feddei; photo. and merotype in A. A.). Wendlandia uvariifolia Hance subsp. Dunniana (Lévl.) Cowan in Not. Bot. Gard. Edinb. 16: 287 (1932); 18: 185 (1934). Wendlandia Dunniana Léveillé in Sea Rep. Spec. Nov. 10: 434 ; Fl. Kouy-Tchéou, 373 (1915). CHInA. Kweichou: Lo-fou, J. Cavalerie, no. 3476, March 1908 (holotype of W. Dunniana ; merotype in A. A.). Wendlandia longidens (Hance) Hutchinson in Sargent, Pl. Wilson. 3: 392 (1916).— Cowan in Not. Bot. Gard. Edinb. 16: 301 (1932). CuHInA. Yunnan: a mi-mont de Siao-ho, alt. 2800 m., E. E. Maire, |1911-14] “arbuste buissonant, fl. roses” (in herb. Léveillé sub Leptodermis Mairet; duplicate in A. A.). In the herbarium Léveillé there were in the cover of Leptodermis Mairet three specimens collected by Maire, of which two belonged to L, Mairet which was identified by Dr. H. Winkler as L. pilosa (Franch. ) Diels var. glabrescens H. Winkl., while the third was not a Leptodermis, but represents the species cited above. From the Szechuan and Hupeh specimens before me it differs slightly in the smaller leaves not exceeding 14 mm. and somewhat more densely pubescent on both sides; the flow- ers also are slightly smaller and are rose-colored according to the col- lector, while Wilson under his numbers 3756 (Veitch Coll.) and 2359 gives the color of the flowers as white. Possibly the specimen cited by Cowan (1. c.) as Maire, no. 326, without locality, is of the same collec- tion as the specimen cited above. Emmenopterys Henryi Oliver in Hooker’s Icon. 19: t. 1823 (1889). Mussaenda Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 13: 178 (1914); Fl. Kouy-Tchéou, 368 (1915).— Hutchinson in Sargent, Pl. Wilson. 3: 397 (1916).— Synon. nov. Mussaenda Mairei Léveillé in Bull. eh Bot. 25: 47 (1915); Cat. Pl. Yun-Nan, 247 (1917).— Synon. no 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 319 Cuina. Kweichou:. sur des rochers a Touan-po prés Pin-yue (Pin-ue), J. Cavalerie, no. 2481, Aug. 10, 1905, “petit arbre a fl. blanches” (holotype of Mussaenda Cavaleriei; merotype in A. A.). Yunnan: vallée de Long-ky, 700 m., E. E. Maire, July 1912, “arbre moyen, fl. blanches” (holotype of Mussaenda Mairei; merotype in A. A.) Adina racemosa Miquel, Cat. Mus. Bot. Lugd.-Bat. 1: 44 (FI. Jap.) (1870). Cornus Esquirolti Léveillé in Fedde, ee Spee. Nov. 13: 257 (1914) ; Fl. Kouy-Tchéou, 116 (1914).— CuHtInA. Kweichou: riviére, Sen J. Esquirol, no. 407, June 1905 (holotype of Cornus Esquirolii; photo. in A. A.). Uncaria scandens (Sm.) Hutchinson in Sargent, Pl. Wilson. 3: 406 (1916). — Léveillé, Cat. Pl. Yun-Nan, 248 (1917). Cephalanthus Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 10: 434 (1912) ; Fl. Kouy-Tchéou, 365 (1915).— Synon. nov. Cuina. Kweichou: Ma-jo, J. Cavalerie, no. 3015, May and Nov. 1908 (holotype of Cephalanthus Cavaleriei; merotype in A. A.). Besides the flowering specimen described by Léveillé there is a fruiting specimen under the same number in his herbarium which is probably the specimen collected in November. This specimen apparently represents U. rhynchophylla (Miq.) Miq. which is of wide distribution in Eastern Asia, but to my knowledge has not been previously collected in western China. Neonauclea Navillei (Lévl.), comb. nov. Cephalanthus Navillei Léveillé, Fl. Kouy-Tchéou, 365 (1915). CuiIna. Kweichou: ruisseau qui monte a Kiao-miay, alt. 800 m., J. Esquirol, no. 3631, June 5, 1913, “arbre, 6 m.” (holotype of Cephalanthus Navillei; merotype in A. A.). This species is very similar to NV. Griffithii (Hook. f.) Merr. but easily distinguished by ternate heads on a rather slender peduncle about 3 cm. long, the slender pedicels being 3—4 cm. long with scars of bractlets near the middle. The leaves are identical with those of specimens of JN. Griffithii from Yunnan (Henry, nos. 12676 and 12880) which are in flower while Esquirol’s specimen is in fruit. Mussaenda Esquirolii Léveillé, Fl. Kouy-Tchéou, 369 (1915). Mussaenda Woilsonii. Hutchinson in Sargent, Pl. Wilson. 3: 393 (1916).— Synon. nov. Cuina. Kweichou: foréts de Tong-tchéou, 1400 m., J. Es- quirol, no. 3264, June 22, 1912 “couleur jaune pale” (holotype of M. 320 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Esquiroli; photo. in A. A.). Hupeh: Chang-lo-hsien, ravines, 650 m., HE. H. Wilson, Arn. Arb. Exp. no. 3265, June 1907 “bracts white, flowers yellow” (holotype of M. Wilsoni in A. A.). The name Mussaenda Esquirolii does not appear on the original speci- men, but an unpublished name under another genus with the same specific epithet in Léveillé’s hand. The specimen differs from the type of M. Wilsoni in the more conspicuous pubescence of the veins and veinlets of the under side of the leaf, but this may be due to the younger state of the leaves. Mussaenda pubescens Aiton f., Hort. Kew ed. 2, 1: 372 (1810). Mussaenda Bodinicri Léveillé in Bull. Soc. Bot. France, 55: 59 (1908) ; Cat. Pl. Yun-Nan, 246 (1917). — Hutchinson in Sargent. Pl. Wilson. 3: 396 (1916).— Synon. nov. CuHina. Kwangtung: pied du Tay-mo-chan (Taiman-san), commun dans les haies pres des villages, EF. Bodinier, no. 1159, May 6, 1895, “arbrisseau a branches sarmenteuses, fleurs blanches; differs des autres Mussaenda de l’absence de grande bractée florale” (holotype of M. Bodinieri; photo. in A. A.). In Mussaenda pubescens the enlarged calyx-lobe is sometimes much reduced in size or entirely absent. I find it entirely absent in Hongkong Herb. 2784, J. B. Norton 1475, and H. H. Chung 3391 from Fukien, and in R. C. Ching 5433 and 5435 from Kwangsi. The locality given by Léveillé as Cay-mo-chan is apparently a misprint for Tay-mo-chan as it is clearly spelled on Bodinier’s label; the same locality appears on English maps as Taimau-san. Tarenna mollissima (Hook. & Arn.) Merrill in Philipp. Jour. Sci. Bot. 13: 160 (1918). — Metcalf in Jour. Arnold Arb. 13: 29 (1932). Ehretia Esquiroli Léveillé, Fl. Kouy-Tchéou, 54 (1914), non Léveillé (1913). Cuina. Kweichou: route de Tong-tchéou, 1200 m., J. Es- quirol, no. 3775, June 1912 (holotype of Ehretia Esquirolii of 1914: photo. in A. A.). The name Fhretia Esquirolii (in Fedde, Rep. Spec. Nov. 12: 335. 1913) had been given by Léveillé to another specimen collected by Esquirol at about the same time and at the same locality and numbered 3214. This specimen cannot be found in the Léveillé herbarium and was probably identified by Leéveillé with another plant and the name used again for Esquirol 3775. The number 3775 seems to have been a mixture, for Léveillé enumerates in his Flore du Kouy-Tchéou the same number under Ehretia Dunniana, E. Esquirolii and E. macrophylla; under the last named as 3775 p.p. and without locality. 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 321 Tarenna incerta Koorders & Valeton in Meded. Lands Plantent. 59: 268 (Bijdr. Boomsort. Java, 8) (1902).— Merrill in Philipp. Jour. Sci. 17: 469 (1920). Tarenna zeylanica Koorders & Valeton, |. c. 82 (1902) ; non Gaertn. ? Webera pallida Franchet ex Brandis, Ind. Trees, 378 (1906). Webera Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 9; 325 (1911): Fl. Kouy-Tchéou, 372 (1915). Webera Henryi Léveillé, Sert. Yunnan. 1 (May 1916) ; Cat. Pl. Yun- Nan, 296 (1917). Tarenna ee (Franch.) Hutchinson in Sargent, Pl. Wilson. 3: 410 (Aug. 1916). Cuina. Kweichou: Pin-fa, J. Cavalerie, no. 2342, June 8, 1905, “petit arbre” (holotype of Webera Cavaleriei; photo. in A. A.). unnan: Szemao, A. Henry, no. 11923a (holotype of Webera Henryi; photo. in A. A.). The type of Webera Cavaleriei consists only of a year-old branch with a few leaves and a small fragment of an inflorescence with very young fruits. It differs somewhat from the type of W. Henry: in the mani- festly truncate calyx without any indication of teeth, though in the latter specimen perfectly truncate calyces occasionally occur. The type of W. Henryi agrees exactly with Henry no. 10686 which was identified by Hutchinson with Tarenna pallida together with Henry, nos. 11923, 11923c and 11923F. Tarenna incerta seems to be somewhat variable in the number of ovules. Koorders & Valeton state that there are two or sometimes only one ovule in each locule and refer to a tree in the garden which had in all flowers only one ovule in each cell. In the one ovary of Webera Cavaleriei which I examined I also found only one ovule in each cell. Merrill (1. c.) states that the usual number of seeds in each fruit is apparently two. Brandis (1. c.), however, describes the fruit of W. pallida as having 4—6 seeds. Two fruits examined of Henry 11923F had 1 and 3 seeds each. It, therefore, seems somewhat doubtful if Webera pallida Franch. of which I have not seen the type really belongs to T. incerta. Gardenia jasminoides Ellis in Philos. Trans. 51(2): 935, t. 25 (1761). Gardenia florida Linnaeus, Spec. Pl. ed. 2, p. 305, 1679 (1762). — Léveillé, Fl. Kouy-Tchéou, 366 (1915); Cat. Pl. Yunnan, 245 (1917). Gardenia Schlechteri Léveillé in Fedde, Rep. Spec. Nov. 10: 146 (1911) ; Fl. Kouy-Tchéou, 366 (1915). — Synon. nov. Cut1na. Kweichou: without precise locality, J. Esquirol, no. 322 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI 777, April 1905, “arbre, fl. blanche” (holotype of G. Schlechteri; mero- type in A. A.). Esquirol no. 777 represents a rather small-flowered form, but other- wise it does not differ from G. jasminoides. Varneria augusta L. (in Amoen. Acad. 4: 136, 1759) upon which Merrill based the new combination G. augusta, isa nomen nudum. The oldest available specific epithet is jasminoides, though florida has been generally adopted. Ixora Henryi Léveillé in Fedde, Rep. Spec. Nov. 13: 178 (1914); Fl. Kouy-Tchéou, 367 (1915); Cat. Pl. Yun-Nan, 245 (1917). — Pitard in Lecomte, Fl. Gén. Indo-Chine, 3: 324 (1924).— Chun in Sun- yatsenia, 1: 306 (1934). Cuina. Kweichou: Lo-fou, J. Cavalerie, no. 3496, March 1909 (syntype; merotype in A. A.). Yunnan: Szemao, s. moun- tain forests, 5000 ft., 4. Henry, no. 11637a, “shrub 5 ft., red flowers” (syntype; photo. and isotype in A. A.). Here also belong Henry nos. 10407 and 10407 a-c from Mengtze and 11637 and 11637a-p from Szemao, Yunnan. Psychotria Henryi Léveillé in Fedde, Rep. Spec. Nov. 13: 179 (1914).— Hutchinson in Sargent, Pl. Wilson. 3: 415 (1916). Cuina. Yunnan: Szemao, s. e. mountains, 4000 ft., 4. Henry, no. 12146p, ‘shrub 4 ft., red fruit” (holotype; photo. and isotype in A.A.). Psychotria rubra (Lour.) Poiret, Encycl. Méth. Suppl. 4: 597 (1816). Psychotria elliptica Ker in Bot. Reg. 8: t. 607 (1822) ; non H. & B. ex Roem. & Schult. (1819). Psychotria Esquirolii Léveillé in Fedde, Rep. Spec. Nov. 10: 435 (1912) ; Fl. Kouy-Tchéou, 371 (1915).— Synon. nov. Cuina. Kweichou: Ouang-mou, J. Esquirol, no. 119, June 1904 “fl. blanchatre” (holotype of P. Esquirolii; photo. in A. A.). This species has apparently not yet been recorded from western China, but seems common in southeastern China west to Kwangsi. Psychotria Prainii Léveillé in Fedde, Rep. Spec. Nov. 9: 324 (1911): Fl. Kouy-Tchéou, 371 (1915). Cuina. Kweichou: Ouang-mou, J. Esquirol, 76 (holotype; photo. in A. A.); Héou-hay-tse, J. Esquirol, no. 860, June 1906, “fl. blanche” (enumerated in Fl. Kouy-tchéou; photo. and merotype in A. A.). 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 323 This species resembles in its capitate sessile or subsessile inflorescence P. morindoides Hutchins., but the inflorescence, branchlets and leaves beneath are covered with a ferrugineous pubescence similar to that of P. pilifera Hutchins., though shorter; besides it differs from P. pilifera in the sessile inflorescence, the smaller more coriaceous leaves glabrous above and in the short petioles. Lasianthus Hookeri Clarke ex Hooker, f., Fl. Brit. Ind. 3: 184 (1880). — Léveillé, Cat. Pl. Yun-Nan, 246 (1917). Lasianthus Dunniana Léveillé in Fedde, Rep. Spec. Nov. 11: 64 1912) ; Fl. Kouy-Tchéou, 368 (1915).— Synon. nov. Cuna. Kweichou: Lo-fou, J. Cavalerie, no. 3459, Oct. 1908 (holotype of L. Dunniana ; photo. and merotype in A. A.). Léveillé’s description of the species is taken from a note on the type specimen which reads “aff. L. trichophlebus Hemsley, sed margine foliorum dense ciliata distincta,” and is signed S. T. D(unn). It agrees exactly with Yunnan specimens referred by Hutchinson to L. Hookeri (in Sargent, Pl. Wilson. 3: 402. 1916). Lasianthus Biermanni King ex Hooker f., Fl. Brit. Ind. 3: 190 (1880). — Léveillé, Cat. Pl. Yun-Nan, 246 (1917). Lasianthus Esquirolii Léveillé in Fedde, Rep. Spec. Nov. 11: 295 (1912) ; Fl. Kouy-Tchéou, 368 (1915). — Synon. nov. Cuina. Kweichou: without ae locality, J. Esquirol, no. 648 (holotype of L. Esquirolii; photo. in A. Esquirol’s specimen agrees well with sen no. 11148, eee by Hutchinson with L. Biermanni (in Sargent, Pl. Wilson. 3: 402. 1916). Lasianthus Labordei (Lévl.) Rehder in Jour. Arnold Arb. 13: 340 (1932). arnt Labordei Léveillé in Pee Rep. Spec. Noy. 13: 178 - Fl. Kouy-Tchéou, 384 (191 ane . weichou: district de Tsin- -gay, mont. de Kao-tchay, penchant escarpé des montagnes, J. Laborde in herb. Bodinier, no. 2109, March 7, 1898, ‘“arbuste” (holotype of Canthium Labordei; photo. in A. A,). Lasianthus Hartii Franchet in Bull. Soc. Bot. France, 46: 209 (1899). intel Dunnianum Léveillé in Fedde, Rep. pp Nov. 9: 324 - Fl. Kouy-Tchéou, 364 (1915).— Synon. n Cuina. Kweichou: without precise ae J. Esquirol (holotype of Canthium Dunnianum; photo. in A. A.); Pin-fa, mon- 324 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI tagnes, J. Cavalerie, no. 3226, May 20, 1907, “1 m. de h., fl. blanches” (cited in Fl. Kouy-Tchéou; merotype in A. A.). In Flore du Kouy-Tchéou Léveillé cites only Cavalerie, no. 3226, which is in bloom, while the type, which is not cited, is a fruiting specimen. This species is apparently related to L. japonicus Miq. from which it differs in the glabrous branchlets, glabrous calyx and glabrous or nearly glabrous leaves, and to L. longicauda Hook. f. from which it differs in the more coriaceous leaves with the veinlets less prominent beneath and in the five corolla-lobes being densely bearded inside up to the tip. The only flowering specimen of L. longicauda I have seen is Henry no. 10633 which has a 4-lobed corolla; this agrees with Hooker’s original descrip- tion of the flowers as “usually 4-merous.”’ The flowers are also 4-merous in Lasianthus Labordei, Lasianthus spec. Canthium Cavalerici Léveillé in Fedde, Rep. Spec. Nov. 10: 434 (1912); Fl. Kouy-Tchéou, 364 (1915). Cuina. Kweichou: Ma-jo, J. Cavalerie, no. 3350 (holotype of Canthium Cavaleriei; photo. and merotype in A. A.). This species I am unable to identify with any species of Lasianthus and in the absence of flowers an exact determination is not possible. It can not be a Canthium, since the fruit is a several-seeded berry. Paederia scandens (Lour.) Merrill in Contr. Arnold Arb. 8: 163 (1934) Paederia eles Thunberg in Nov. Act. Soc. Sci. Upsal. 4: 32 (1783) ; Jap. 106 Sdn ee Fl. Kouy-Tchéou, 376 (1915). Bat Linnaeus (17 Paederia tomentosa Blume, ree 963 (1826). Paederia chinensis Hance in Jour. Bot. 16: 228 (1878). Paederia Esquirolii Léveillé in Fedde, Rep. Spec. Nov. 10: 146 19 ese Dunniana Léveillé, 1. c. (1911). Paederia Mairei Léveillé in Fedde, Rep. Spec. Nov. 18: 179 (1914). Paederia tomentosa Bl. var. Mairei (Lévl.) Léveillé, Cat. Pl. Yun- Nan, 247 (1917). — Synon. nov. Cutna. Kweichou: Ky-che-ten, J. Esquirol, no. 184, Aug. 1904 (holotype of P. Esquirolii; photo. in A. A.); without precise lo- cality, J. Esquirol, no. 775, April 1905, “fleur 4 gorge rouge” (holotype of P. Dunniana; photo. in A. A.). Yunnan: broussailles des col- lines & Siao-ou-long, 2550 m., E. E. Maire, July 1911, “arbuste grim- pant, odeur fétide, fl. roses” (holotype of P. Mairei; photo. in A. A.). The specimens cited above represent the typical glabrous form of 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 329 P. scandens ; Paederia Esquirolii and P. Dunniana have narrow generally oblong leaves cuneate at base, while P. Mairei has larger and broader generally ovate leaves rounded to truncate at base. This species which is the most widely distributed of the genus has been generally called P. tomentosa BI. which is described by Blume as having the leaves tomentose beneath. I have seen no specimens from the type region, but as long as I have no evidence to the contrary, I accept P. tomentosa Bl. as a synonym of P. scandens, representing the form with leaves pubescent beneath which occasionally occurs also in China. Paederia Wallichii naan Fl. Brit. Ind. 3: 196 (1881). Paederia tomentosa Bl. var. ee caerulea Léveillé & Vaniot in Bull. Soc. Bot. eee 55: 59 (1908). Paederia Bodinieri Léveillé, = Kouy-Tchéou, 371 (1915) ; Léveillé (1914). — Synon. n wana anon Leveillé Cat. Pl. Yun-Nan, 13 (1915).— Synon CHINA. =e weichou: environs de Hoang-ko-chou, rochers, etc., J. Seguin in herb. Bodinier, no. 2501, Sept. 8, 1898, “liane grimpante, fleurs, atro-pourpre, avec bordure bleu-rouge, inodore” (holotype of P. tomentosa var. purpureo-coerulea and P. Bodinieri ; photo. and merotype in A. A.). Yunnan: brousse des montagnes a Tcha-ho, alt. 2800 m., E. E. Maire, Nov. 1911, “arbuste grimpant, tomenteux, fl. violettes”’; rives du fleuve Bleu a Siao-ho, 400 m., E. E. Maire, July 1912 “arbuste grimpant, feuil. blanches, velues en dessous” (syntypes of Cynanchum yunnanense ; photos. in A. A.). Paederia Bodinieri was described by Léveillé without reference to his earlier P. tomentosa var. purpureo-caerulea, but it is based on the same specimen which bears only the name var. purpureo-caeruea in Léveillé’s hand. The name is a later homonym of his earlier P. Bodinieri (in Fedde, Rep. Spec. Nov. 13: 179. 1914) which he referred the same year to Marlea as M. Cavaleriei and which turns out to be identical with Gardneria multiflora Mak. (see Jour. Arnold Arb. 15: 309). Cynan- chum yunnanense is not different from Seguin’s specimen except that it has shorter inflorescences. The specimens enumerated above are identical with Henry’s nos. 9126 and 12442 and, judging from the description, seemed conspecific with P. Wallichii Hook. f. I am indebted to Sir Arthur W. Hill for a com- parison of the Henry numbers with the type specimens of P. Wallichii in the Kew Herbarium; he writes me that Mr. C. E. C. Fischer reports on these specimens as follows: “the only difference between these num- bers and the type of P. Wallichii Hook f. that I can see is that the basal 326 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI lobes of the leaves are rather more rounded and the sinus slightly deeper and also the buds rather shorter. I doubt that this would suffice for even a variety.” With the material at hand P. Wallichii Hook. f. may be characterized by cordate to subcordate leaves densely scabrid above with setulose appressed hairs somewhat bulbous at base and densely villous-pubescent or tomentose beneath, and by the lateral branches of the inflorescence bearing one or several capitate flower clusters. In the capitate or subcapitate flowers it resembles P. microcephala Pierre from which it differs in the short calyx-teeth and in the dense pubescence of the leaves. Paederia Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 13: 179 (1914), Paederia tomentosa Bl. ex Léveillé, Fl. Kouy-Tchéou, 371 (1915).— Non Blume (1826) Cuina. Kweichou: bois des montagnes, J. Cavalerie, no. 2058, Aug. 1904, “plante tres puante” (holotype of P. Cavaleriei ; photo. in A. A.). This species is similar to the preceding, but differs in the long hirsute ferrugineous pubescence of the branches, the petiole and the inflo- rescence, in the petiole being 7.5—8 cm. long and in the 1—1.5 mm, long fairly straight hairs thinly covering the underside of the leaves, but dense on the midrib and veins, also in the glabrous narrower calyx-teeth. According to the original description it resembles P. pilifera Hook. f. in the long pubescence, but Pitard (in Lecomte, Fl. Gén. Indo-Chine, 3: 412) describes the leaves as very tomentose beneath, the calyx-tube as very tomentose and the petiole as 1—2.5 cm. long. A specimen collected in Kweichou at Sanhoa by W. Y. Chun (no. 6301) agrees in the pubescence of the stem and the leaf and in the thin texture of the leaf very well with P. Cavaleriei, but the lateral branches of the inflorescence end in cincinnate cymes, as is the rule in P. scandens, and not in subcapitate or capitate cymes, characteristic of this and the preceding species. Leptodermis Potanini Batalin in Act. Hort. Petrop. 14: 319 (1898).— H. Winkler in Fedde, Rep. Spec. Nov. 18: 152 (1922). Leptodermis Esquirolti Léveillé in Fedde, Rep. Spec. Nov. 9: 324 (1911) ; 18: 179 (1914); FI. Kouy-Tchéou 368 (1915); Cat. FI. Yun-Nan, 246 (1917). CuinA. Kweichou: Hin-y-fou, J. Cavalerie, no. 3930, July 1912 (cited in Fl. Kouy-Tchéou; duplicate in A. A.). Yunnan: Ouan-tse, J. Esquirol, no. 1503, May 22, 1909 “blanche a l’interieur, rouge a l’exterieur” (holotype of L. Esquirolii; photo. in A. A.). 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 327 Léveillé published L. Esquirolii a second time in 1914, but with a briefer description, both based on Esquirol no. 1503 which is a flower- ing specimen. Cavalerie no. 3930 from Kweichou is a fruiting specimen and owing to its fully developed broader leaves looks somewhat different, but apparently belongs to this species. Leptodermis Esquirolii was first identified with L. Potanini by H. Winkler (1. c.). Leptodermis Potanini var. glauca (Diels) H. Winkler in Fedde, Rep. Spec. Nov. 18: 153 (1922). ee motsouensis Léveillé in Bull. Géog. Bot. 25: 47 (1915); Cat. Pl. Yun-Nan, 246 (1917). — Synon. nov. ee Yunnan: collines arides de Mo-tsou, 800 m., £. E. Maire, May 1912, “sous-arbrisseau en touffes, fl. blanches” (holotype of L. motsouensis ; merotype in A. A.). The identification of L. motsouensis with L. Potanini var. glauca was communicated to me by Dr. H. Winkler in a recent letter as were the identifications of the following species of Leptodermis. Leptodermis ae var. tomentosa H. Winkler in Fedde, Rep. Spec. Nov. 18: 153 (1922). Leptodermts ie ora Léveillé in Bull. Géog. Bot. 25:47 (1915); Cat. Pl. Yun-Nan, 246 (1917), “tongtchouanensis” — Synon. nov. Cuina. Yunnan: rochers des coteaux autour de Tong-tchouan, 2550 m., &. E. Maire, May 1912, ‘‘arbrisseau, feuilles velues blan- chatres, fl. blanches soyeuses” (holotype of L. tongchouanensis ; mero- type in A. A.). Leptodermis pilosa (Franch) Diels var. glabrescensH. Winkler in Fedde, Rep. Spec. Nov. 18: 160 (1922). Leptodermis Mairei Léveillé in Fedde, Rep. Spec. Nov. 13: 179 (1914) ; Cat. Pl. Yun-Nan, 246 (1917).— Synon. nov Cuina. Yunnan: plaine de Long-tang, 2500 m., E. E. Maire, Aug. 1912, “arbuste non grimpant, en touffes, fl. violet” (holotype of L. Matret; merotype in A. A.); haies, plaine de Tong-tchouan, 2500 m., E. E. Maire, in 1912 ‘“‘arbuste buissonant, écorce blanche, fl. roses” (in herb. Léveillé in cover of L. Mairei; duplicate in A. A.). Only the specimen from Long-tang bears the name L. Mairei in Léveillé’s hand. As Maire no. 21 Winkler (1. c.) enumerates under his L. pilosa var. glabrescens a specimen apparently of the same collection as Léveillé’s type of mi Mairet Another specimen ‘‘a mi-mont de Siao-ho, 2800 m.” in herb. Léveillé under L. Mairei belongs to Wendlandia longidens (Hance) Hutch. (see p 318.). ”) 328 JOURNAL OF THE ARNOLD ARBORETUM LVOL. XVI Leptodermis oblonga Bunge in Mém. Sav. Etr. Acad. Sci. St. Pétersb. 2: 108 (Enum. PI. Chin. Bor. 34) (1833). Leptodermis Chaneti Léveillé in Bull. Géog. Bot. 25: 47 (1915). Cuina. Hopei: montagnes de Ping-chan, L. Chanet, no. 538 bis, Aug. 1910; without precise locality, L. Chanet, no. 574, June 1904 (syn- types of L. Chaneti; merotypes in A. A.). Prismatomeris Henryi (Lév!.), comb. nov. Canthium Henryi Léveillé in Fedde, Rep. Spec. Nov. 13: 178 (1914) ; Cat. Pl. Yun-Nan, 245 (1917).— Synon. nov Prismatomerts brevipes Hutchinson in Sargent, Pl. Wilson. 3: 413 916). — Léveillé, Cat. Pl. Yun-Nan, 247 (1917). — Synon. nov. Cuina. Yunnan: Meng-tse, S. E. mountains, 5000 ft., A. Henry, no. 9040F, “shrub 6 ft., fruit red” (holotype of Canthium Henryi; photo. in A. A.; paratype of P. brevipes ; isotype in A. A.). There is also an isotype of the holotype of P. brevipes, Henry 9040k, in the herbarium of the Arnold Arboretum and of another paratype, Henry 9040p. CAPRIFOLIACEAE Sambucus javanica Bl. var. Argyi (Lévl.), var. Sambucus Argyi Léveillé in Bull. Géog. Bot. ae ante: 292 (1914); in Mem. Acad. Ci. Arts Barcelona, ser. 3, 12: 545 (Cat. Pl. Kiang-Sou, 5) (1916). CHINA. Kiangsu: Ka-se-dao, trouvé venant de Tou-ka-dou, Ch. d’Argy [1848-66] (holotype of S. Argyi; photo. in A. A.); Sé-hom, (jardin), Ch. d’Argy [1848-66] “fruit rouge” (with S. Argyi in herb. Léveillé; photo. in A. A.). This variety differs from the type in the shorter and comparatively broader more coarsely serrate leaflets, the lateral ones ovate-oblong to oblong (5 1.8), the terminal elliptic or elliptic-obovate (6 « 3), and slightly scaberulous on the veins. Sambucus Argyi was first referred to S. javanica by H. K. Airy-Shaw according to a note on the specimens. Viburnum erubescens Wallich, Pl. As. Rar. 2: 29, t. 143 (1830).— Léveillé, Cat. Pl. Yun-Nan, 28 (1916). Viburnum botryoideum Léveillé, Cat. Pl. Yun-Nan, 28 (1915).— Synon. nov. Cuina. Yunnan: rochers, brousse de Kiao-me-ti, 3100 m., E. E. Maire, May 1913, “arbuste, fleurs roses” (holotype of V. botryot- deum, merotype in A. A.). Viburnum oliganthum Batalin in Act. Hort. Petrop. 13: 372 (1894). 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 329 Viburnum Stapfianum Léveillé in Fedde, Rep. Spec. Nov. 9: 443 11); Fl. Kouy-Tchéou, 66 (1914). — Synon. nov Cuina. Kweichou: Ma-jo, J. Cavalerie, no. 3002, May 1908, “petit arbre, fleurs roses” (holotype of V. Stapfianum; merotype in A.A This species seems common in Szechuan; the specimen cited above is the first I have seen from outside of that province. Viburnum sympodiale Graebner in Bot. Jahrb. 29: an (1901). — Rehder in Sargent, Trees & Shrubs, 2: 83, 108, t. 139 (190 Viburnum Martini Léveillé in Fedde, Ren a Nov. + 443 (1911); ouy-T'chéou, 66 (1914). — Synon. nov Cumna. Kweichou: Pin-fa, bois, rare, Z Cavalerie, no. 2272, April 4, 1905, “fl. blanches” (holotype of V. Martini; merotype in ‘ae; ee Viburnum Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. 9: 442 (1911); Fl. Kouy-Tchéou, 66 (1914). Cutna. Kweichou: Pin-fa, montagnes, J. Cavalerie, no. 977, April 13, 1903, “h. 1 4 2 m., fl. blanches odorantes,” (holotype; photo. and merotype in A. A.). This species is closely related to V. fallax Graebn. and V. chinchanense Graebn. differing from the former in its coriaceous leaves rugulose and stellate pubescent above and from the latter in the glabrous ovaries and the leaves being sparingly stellate-pubescent and somewhat scabrid above. Both related species have been collected in Kweichou; V. fallax is represented in this herbarium from Kweichou by Steward, Chiao & Cheo 11, and V. chinchanense by Y. Tsiang 7419 and 9216 and also by another specimen, Tsiang 8424, which approaches V. Rosthornii Graebn. by its larger subcordate leaves sparingly stellate above. Viburnum congestum Rehder in Sargent, Trees & Shrubs, 2: 111 (1907). — Léveillé, Cat. Pl. Yun-Nan, 28 (1915).— P’ei in Mem. Sci. Soc. China 1(no. 3): 90 (Verben. China) (1932). Hedyotis Mairei Léveillé in Fedde, Rep. Eee Nov. 138: 176 (1914) ; Cat. Pl. Yun-Nan, 245 (1917).— Synon. Viburnum Mairei Léveillé, Cat. Pl. Yun- Nan, 28 (1915). — Synon. Eeeming Esquirolii Léveillé, Sert. Yunnan. 3 (1916); Cat. Pl. Yun- Nan 7). Oldenlandia Mairei (Lévl.) Chun in Sunyatsenia, 1: 314 (1934). Cutna. Yunnan: brousse du plateau de Ta-hai-tse, alt. 3200 m., E. E. Maire, May (1912 or 1913), “grande arbuste, feuilles cadu- ques, fl. blanches” (holotype of Hedyotis Mairei, named H. yunnanensis 330 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI on the label of the type specimen; photo. in A. A.); brousse au pied de Io-chan, 3200 m., E. E. Maire, May 1912, ‘“arbuste haut 1.30 m.,” (holo- type of V. Mairei; merotype in A. A.); brousse des montagnes a Mo- tsou, 800 m., FE. E. Maire, May 1912, “arbuste, feuilles caduques, fl. blanches” (holotype of Premna Esquirolii; merotype in A. A.). The three specimens cited above have the corolla tube somewhat shorter than in the type of V. congestum, the tube being only slightly longer than the lobes, but in the otherwise similar V. utile Hemsl. the tube is wide-campanulate and much shorter than the lobes. Viburnum congestum differs from V. utile also in the less dense grayish tomentum of the under side of the leaves, the individual hairs being distinguishable and usually have shorter rays, while in V. utile they are matted and the tomentum is whitish, brownish on the veins in young leaves. There are, however, specimens of V. congestum with a denser tomentum similar to that of V. utile. Geographically the two species seem to be well sepa- rated; V. congestum is common in Yunnan and extends into Kweichou and western Szechuan, while V. utile occurs in Hupeh and extends to eastern Szechuan. Viburnum cylindricum Ham. var. crassifolium (Rehd.) Schneider in Bot. Gaz. 64: 77 (1917). Viburnum crassifolium, Rehder in Sargent, Trees & Shrubs, 2: 112 1908). — Léveillé, Cat. Pl. Yun-Nan, 28 (1915). Viburnum pinfaense Léveillé in Fedde, Rep. Spec. Nov. 9: 442 1911); pro aig ig id Cavalerie no. 1483; Fl. Kouy-Tchéou, 66 (1914). — Synon Cuina. Kweichou: Pin-fa, bois, J. Cavalerie, no. 1483, Oct. 12, 1903 (syntype of V. pinfaense ; merotype in A. A.). This variety has been collected in Kweichou in three different locali- ties by Y. Tsiang (nos. 4121, 7558 and 9137). The leaves of Cavalerie’s specimen are unusually small and narrow. Under V. pinfaense Léveillé describes two different plants, of which the fruiting specimen belongs here, while the flowering one belongs to V. sempervirens K. Koc Viburnum ternatum Rehder in Sargent, Trees & Shrubs, 2: 37, 112 t, 147 (£1907), Viburnum Chaffanjoni Léveillé in Fedde, Rep. hay Nov. 9: 443 (1911); Fl. Kouy-Tchéou, 66 (1914).— Syn Ov. Cu1na. Kweichou: Pin-fa, bois presque a pic, J. Cavalerie, no. 3093, July 2, 1907 (holotype of V. Chaffanjoni; merotype in A. A.). This very distinct species, differing in its ternate leaves from all other species, was known to me before only from Szechuan: banks of Min 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 331 River (Wilson 3736), Mt. Omei (W. P. Fang 2461, 2631, 3309, 3355 and F. T. Wang 23138), and Kuan-hsien (W. P. Fang 2021). Viburnum Schneiderianum Handel-Mazzetti in Akad. Wiss. Wien Anzeig. 1925: 66 (Pl. Nov. Sin. Forts. 33: 4) (1925 Cuina. Yunnan: rochers de [o-chan, alt. 3200 m., E. E. Marre, May (1911-13), “arbrisseau rampant, toujours vert,” (in herb. Léveille sub Gaultheria crenulata; duplicate in A. A. The specimen cited above was referred by Léveillé to Gaultheria crenulata Kurz and represents, at least partly, the plant enumerated under that name in his Cat. Pl. Yun-Nan, 86 (1916). Viburnum sempervirens K. Koch, Hort. Dendr. 300 (1853). — Rehder in Sargent, Trees & Shrubs, 2: 95, 113, t. 145 (1908). Viburnum pinfaense Léveillé in Fedde, Rep. Spec. Nov. 9: 442 (1911), pro parte, quoad specim. no. 1056; Fl. Kouy-Tchéou, 66 (1914). — Synon. nov Cuina. Kweichou: Pin-fa, montagnes, J. Cavalerie, no. 1056, June 11, 1903 (syntype of V. pinfaense ; merotype in A. A.). This species seems to be rare in western China. I have seen it only from Pin-fa, Kweichou (Cavalerie 1056 and Y. Tsiang 6385), and from Szemao, Yunnan (Henry 12753). Viburnum foetidum Wallich, Pl. As. Rar. 1: 49, t. 61 (1830).— Léveillé, Cat. Pl. Yun-Nan, 28 (1915).— P’ei in Mem. Sci. Soc. China, 1, no. 3: 90 (Verben. China) (1932). Viburnum ajugifolium Léveillé in Fedde, Rep. ee Nov. 9: 441 (1911) ; Fl. Kouy-Tchéou, 65 (1914). — Synon. n Premna Valbrayi Léveillé, Sert. Yunnan. 4 (1916) + ‘Ch Pl. Yun- Nan, 299 (1917). Cutna. Kweichou: environs de Kouy-yang, mont du Collége, c. dans les haies, bords des ruisseaux, E. Bodinier, no. 2231, May 18, 1898, “fl. blanches” (holotype of V. ajugifolium,; merotype in A. A.). Yunnan: haies et brousses des montagnes, a Tong-tchouan, 2500- 2700 m., E. E. Maire, July 1912, “arbuste gréle, feuill. caduques” (holo- type of Premna Valbrayi; merotype in A. A.). The two specimens cited above are similar to the form described as V. ceanothoides C. H. Wright. Viburnum foetidum var. rectangulatum (Graebn.) Rehder in Sar- gent, Trees & Shrubs, 2: 114 (1908), “rectangulum.” Viburnum Touchanense Léveillé in Fedde, Rep. Spec. Nov. 9: 442 1911) ; Fl. Kouy-Tchéou, 66 (1914), — Synon. nov 332 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI Hedyotis — Léveillé in Fedde, Rep. 7. nad 13: 176 (1914) ; Cat. Pl. Yun-Nan, 245 (1917).— 8 Oldenlandia yunnanensis (Lévl.) Chun in Nin intrt 1: 310 (1934). Cuina. Kweichou: environs de Tou-chan, bord de la route, J. Cavalerie, no, 2192, July 5, 1897, ‘“‘arbrisseau aux fleurs odorantes” (holotype of V. Touchanense ; merotype in A. A.). Yunnan: brous- sailles des collines a Long-ky, alt. 700 m., E. E. Maire, June 1911, “arbuste a feuilles caduques, fl. blanches” (holotype of Hedyotis yunnan- ensis ; photo. in A. A.). Viburnum setigerum Hance in Jour. Bot. 20: 261 (1882). — Reh- der in Jour. Arnold Arb. 12: 77 (1931). Viburnum theiferum Rehder in Sargent, Trees & Shrubs, 2: 45, 113. t. 121 (1907). Viburnum Bodiniert Léveillé in Fedde, Rep. Spec. Nov. 9: 442 1911) ; Fl. Kouy-Tchéou, 65 (1914). Cuina. Kweichou: environs de Kouy-yang, bois de Kin-lin- chan, E. Bodinier, no. 2193, April 14, 1898, “arbuste, fl. blanches” (syn- type of V. Bodinieri; merotype in A. A.); Pin-fa, bois ombreux, J. Cava- lerie, no. 1285, May, 1903, “fl. blanches, odorantes” (syntype of V. Bodinieri ; photo. in A. A.). Viburnum Bodinieri was identified with V. setigerum by the writer and the identification published in 1931 (1. c.). The species has been collected in Kweichou also by Y. Tsiang near Tsunyi (no. 5318) and on the Yun-fu-shan near Pin-fa (no. 5510) and near Tuyun (5942). Viburnum corylifolium Hooker f. & Thomson in Jour. Linn. Soc. 2: 174 (1858). Viburnum egal Léveillé in Fedde, Rep. he Nov. 9: 442 (1911); Fl. Kouy-Tchéou, 66 (1914).— Synon. no Viburnum Sarna Léveillé, Fl. Kouy-Tchéou, 65 (1914). — ynon. nov. Cuina. Kweichou: route de Pin-yue a Kouy-yang, bords d’une riviére, L. Martin in herb. Bodinier, no. 2598, May 13, 1899, “grand arbuste, fl. blanches”; environs de Kouy-yang, mont du Collége, J. Chaffanjon, May 1, 1898, “arbuste, fl. blanches”; route de Pin-fa A Oug-lan, J. Cavalerie, Aug. 1908 “‘fruite rouge” (syntypes of V. Dunn- tanum,; photos. of Martin’s and Cavalerie’s specimens, merotype of Chaffanjon’s specimen in A. A.); Pin-fa, bois de hautes montagnes, J. Cavalerie, no. 1742, Aug. 1904 (holotype of V. barbigerum ; merotype in A. A.) Viburnum barbigerum agrees in all its characters with the other speci- mens cited, but the fruits are strikingly different in being densely cov- ered with long setose hairs. I suspect, however, that this development 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 333 of hairs is abnormal, since I have found a few other specimens of Viburnum namely R. C. Ching nos. 2826 and 2952 of V. ichangense (Hemsl.) Rehd. from Anhwei with some of the fruits densely covered with similar, though somewhat shorter, hairs, while the rest of the fruits was perfectly normal and glabrous. Viburnum corylifolium is perhaps only a variety of V. dilatatum Thbg. differing chiefly in the long spreading hairs of the young branchlets, in- florescence and petioles, while in V. dilatatum these parts are covered by a short and close stellate tomentum. Viburnum corylifolium has been collected in Kweichou also by Y. Tsiang (no. 5779) near Tu-yun and by Steward, Chiao and Cheo, (no. 583) on Niu-tu-shan; V. dilatatum was collected by Y. Tsiang (no. 6270) on Yao-ren-shan, Sanhoa. Viburnum erosum Thbg. var. Taquetii (Lévl.) Rehder in Sargent, Pl. Wilson. 1: 311 (1912).— Nakai in Nakai & Koidzumi, Trees & Shrubs Jap. ed. 2, 1: 609 (1927). — Makino & Nemoto, FI. Jap. ed. 2, p. 1146 (1931). Viburnum Taquetii Léveillé in Fedde, Rep. Spec. Nov. 9: 443 (1911). Viburnum erosum var. punctatum Franchet & Savatier ex Nakai, Fl. Sylv. Kor. 11: 42, t. 12 c. d (1921) quoad synon. V. Taquetii Lévl., vix Franch. & Sav. Cuina. Korea: Quelpaert, in silvis Yengsil, 1000 m., E. Taquet, no. 4281, Aug. 12, 1910 (holotype of V. Taqueti,; photo. and isotype in TO. 0) This peculiar variety chiefly characterized by the narrow leaves partly with two basal lobes near the base has been collected in Quelpaert also by E. H. Wilson (no. 9406). Nakai in 1921 (1. c.) referred it to V. erosum var. punctatum Franch. & Sav., but that variety represents apparently the plant with broader leaves densely stellate-pubescent above which seems to be the most widely distributed form. The gla- brous or glabrescent form, var. /aeve Franch. & Sav., which seems much rarer must be considered the typical form, since Thunberg (FI. Jap. ’ 124) describes the leaves as glabrous. Dipelta yunnanensis Franchet in Rev. Hort. 1891: 246, fig. 62. — Léveillé, Cat. Pl. Yunnan, 27 (1915). Cavaleriella Dunniana Léveillé, Fl. Kouy-Tchéou, 61 (1914).— Synon. nov. CuHina. Kweichou:_ hautes montagnes, Long-ly, J. Cavalerie, no. 3023, May 1908 (holotype of Cavaleriella Dunniana; merotype in A. A.) The leaves are pilose on the midrib and veins beneath, also the young branchlets and the inflorescence are pilose. 334 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI Abelia verticillata Léveillé, Fl. Kouy-Tchéou, 61 (1914). CHINA. Kweichou: Pin-fa, ruisseau du sud, J. Cavalerie, no. 497, Sept. 1912, “fl. blanche-violette-pourprée” (holotype; photo. in This species appears to be closely related to A. uniflora R. Br. and A. Graebneriana Rehd., but differs from both in the pilose branches and in the ovary being pilose with rather long white hairs; it also differs from the former in the ciliate leaves and from the latter in the subcoriaceous leaves not bearded in the axils and without hairs alongside the midrib and the base of the lateral veins. On one of the branches of the type specimen the leaves are in whorls of threes, from which the specific epithet is derived, but the other branch has opposite leaves; branches with ternate leaves are also occasionally found in A. uniflora and in other species, e. g. in Wilson no. 747 of A. parvifolia Hemsl. Abelia Schumannii (Graebn.) Rehder in Sargent, Pl. Wilson. 1: 121 (1911). — Léveillé, Cat. Pl. Yun-Nan, 26 (1915) Strobilanthopsis deutziaefolius Léveillé in Fedde, Rep. Spec. Nov. 12: 20 Abelia detsacolia (Lévl.) Léveillé, Fl. Kouy-Tchéou, 60 (1914). — Synon Strobilanthes deutziaefolia, Léveillé, 1. c. (1914), . synon, Abeliae deutziaefoliae. — Ind. Kew. Suppl. 4: 252 (1921). Abelia Mairei Léveillé, Cat. Pl. Yun-Nan, 26 (1915).— Synon. nov. Cuina. Kweichou: sous bois, J. Esquirol, no. 466, June 1905, “‘sous-arbrisseau, fl. blanches, panachées de rose,” (holotype of Stro- bilanthopsis deutziaefolius ; merotype in A. A.). Yunnan: rochers des collines a Siao-ou-long, 2250 m., E. E. Maire, June 1912, ‘“‘arbuste delicat, buissonant, haut 0.80 m., fl. roses (holotype of A. Mairei; mero- type in A, A.). The leaves of Esquirol’s no. 466 are rather large being up to 3.5 cm. long and resemble somewhat those of A. Graebneriana Rehd., though scarcely acuminate, but the branchlets are pubescent. Abelia myrtilloides Rehder in Sargent, Pl. Wilson. 1: 120 (1911). Strobilanthopsis ee Leveille in Fedde, Rep. Spec. Nov. 12: 20 (1913), “hypercifolia Abelia Bodinieri Léveillé, Fl Kouy-Tchéou, 61 (1914) pro synon. A. parvifoliae. Abelia parvifolia Hemsl. sec. Léveillé, Fl. Kouy-Tchéou, 61 (1914), saltem quoad specim. Bodinier, no. 1607, vix Hemsley. Strobilanthes hypericifolia Léveillé, 1. c. (1914), pro synon, A. parvifoliae. — Ind. Kew. Suppl. 5: 252 (1921). Cuina. Kweichou: mont de Lou-tsong-koan, 1500 m., Kien- 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 335 lin-shan ¢a at 1a dans les montagnes rocailleuses, E. Bodinier, no. 1607, June 1, 1897 and June 19, 1899, “petit arbuste a jolies fleurs roses” (syntypes of Strobilanthopsis hypericifolius; photo. and merotype in A.A Léveillé cites in his Flore du Kouy-Tchéou (1. c.) under A. parvifolia as synonyms besides Strobilanthes hypericifolia the unpublished name A. Bodinieri; both names appear on the labels of the type sheet of Bodinier no. 1607. The two specimens on the type sheet are somewhat intermediate between A. myrtilloides and A. parvifolia Hemsl., but in the oblong-elliptic or oblong-ovate shape of the leaves glabrous above and nearly so beneath they seem closer to the former, only in the glandular under surface they approach A. parvifolia which typically has ovate leaves of thicker texture pilose and glandular on both surfaces. A form very similar to Bodinier’s specimen was collected near Kwei-yang, Kweichou, by Handel-Mazzetti, (no. 10477) who determined it as “A. parvifolia Hemsl. trans. ad A. myrtilloidem Rehd.” Abelia Cavaleriei Léveillé, Fl. Kouy-Tchéou, 60 (1914). Cuina. K weichou:. sud de Tin-fan, mont. rocheuses, J. Cava- lerie, no. 1909, Oct. 1904, “fl. blanches” (holotype; merotype in A. A.). This is a very distinct species on account of its subcoriaceous leaves which recall those of Ligustrum strongylophyllum Hemsl. The species is apparently nearest A. chinensis R. Br., but is readily distinguished by the subcoriaceous perfectly glabrous quite entire leaves broadly ovate to orbicular-ovate, 1-2 cm. long, rounded or broadly cuneate at base, obtuse or rounded and apiculate at the apex. The branchlets and the many flowered terminal inflorescence are minutely puberulous. The specimen is in fruit but according to the collector the flowers are white. Lonicera tangutica Maximowicz in Bull. Acad. Sci. St. Pétersb. 24: 48 (1877); in Mél. Biol. 10: 75 (1877). Lonicera Rocheri Léveillé in Bull. Géog. Bot. 24(no. 301): 289 (1914) ; Cat. Pl. Yun-Nan, 27 (1915). — Synon. nov. Cuina. Yunnan: brousse de Lan-mou-kiao, 3000 m., E. E. Maire, May 1912, “arbuste en touffes; fl. jaunes” (holotype of L. Rocheri; merotype in A. A.). The species cited above differs somewhat from typical L. tangutica in the linear-lanceolate somewhat leafy bracts about twice as long as ovary, in the anthers being exserted about one-half and in the less slender corolla-tube, but in its other characters it agrees with this species. Lonicera ligustrina Wallich in Roxburgh, Fl. Ind. ed. 2, 2: 179 (1824), — Léveillé, Cat. Pl. Yun-Nan, 27 (1915). 336 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Lonicera missionis Léveillé, Fl. Kouy-Tchéou, 63 (1914), pro parte, quoad specim. “Esquirol (May 10, 1906), Chaffanjon no. 2215”; Cat. Pl. Yun-Nan, 3 (1915). CuIna. Kweichou: environs de Kouy-yang, mont. du Col- lege, a la cascade, J. Chaffanjon in herb. Bodinier, no. 2215, April 14, 1898, “arbuste” (syntype of L. missionis; photo. in A. A.); mont du College, grotte, J. Esquirol, May 10, 1906 “fl. blanches” (syntype of L. missionis ; merotype in A. A.). Yunnan: sous bois de couteaux a Long-ky, 700 m., E. E. Maire, “arbuste toujours vert, fl. blanches,” (in herb. Léveillé under L. missionis ; duplicate in A. A.) This species has been collected in Kweichou also by W. Tsiang (nos. 4580, 5972, 7640) and by Steward, Chiao & Cheo (no. 244). Lonicera pileata Oliver in Hooker, Icon. Pl. 16: t. 1585 (1887). — Léveillé, Cat. Pl. Yun-Nan, 27 (1915). Lonicera missionis Léveillé, Fl. Kouy-Tchéou, 63 (1914) he parte, quoad specimen “Laborde 2502”; Cat. Pl. Yun-Nan, 27 (1915).— Lonicera buxifolia Léveillé, Fl. Kouy- cana = (1914); Cat. Tl. Seu-Tchouen, t. 11 (1918) MS. — Syn CHINA. Kweichou: environs de ae a Kia-la-tchong, J. Laborde in herb. Bodinier, no. 2502, Nov. 1898 “‘les fruits sont des jolies perles bleu-tendre, pulpeuses’” (syntype of ZL. missionis; photo. in A. A.); environs de Kouy-yang, mont du Collége, rochers de la cascade, au bords de l’eau, J. Chaffonjon in herb. Bodinier, no. 2169, April 12, 1898, “fl. jaunatres” (syntype of L. buxifolia; photo. in A. A.); grotte du College, 1350 m., J. Esquirol, no. 2069, April 1910, “blanche” (syn- type of L. buxifolia; photo. in A. A.); Pin-fa, ruisseaux, J. Cavalerie, no. 1319, April 9, 1902, “fl. blanche” (in herb. Léveillé under L. buxi- folia; photo. in A. A.). This species has been collected in Kweichou also by Y. Tsiang (nos. 4529, 4562, 7937) and by Steward, Chiao & Cheo (no. 803); the last named specimen is approaching in the shape of its leaves f. linearis Rehd. The specimens named L. buxifolia by Léveillé differ from typical L. pileata in their rather small leaves. The two species, L. ligustrina and L. pileata, are closely related and connected by intermediate forms in regard to shape of the corolla and of the leaves and to pubescence. Without flowers L. ligustrina may be distinguished by the leaves being generally ovate, rounded at base, acuminate, and with the midrib more or less impressed above at least toward the base and strigose, while L. pileata has generally elliptic to oblong leaves, narrowed at base, obtuse to acute at apex, with the midrib distinctly elevated above and glabrous. In regard to shape and pubescence of the leaves L. nitida Wils. seems 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 337 intermediate between the two, but the leaves are much smaller and usually broader, generally ovate, but not acuminate. Lonicera virgul- torum W. W. Sm. is very close to L. ligustrina and chiefly distinguished by the shape of the corolla. Lonicera fragilis Léveillé in Fedde, Rep. Spec. Nov. 13: 337 (1914); Cat. Pl. Yun-Nan, 37 (1915). CuinaA. Yunnan: vallée de Li-tse-pin, 2800 m., E. E. Maire, April 1913, “arbuste cassant, haut de 1.20 m., fl. roses” (holotype; mero- type in A. A.). Frutex metralis ramis hornotinis sparse setosis vel glabris; gemma terminalis interdum evoluta perulis duabus exterioribus et 4—6 interiori- bus. Folia nondum plane evoluta, elliptico-oblonga, acuminata, basi cuneata, utrinque hirsuta, glandulis sparsis intermixtis, margine ciliata et stipitato-glandulosa. Flores praecoces in axillis bractearum ad basin ramulorum; pedunculi brevissimi glabri; bracteae late ovatae, 8-10 mm. longae, irregulariter eroso-denticulatae, basin versus ciliatae et sparsis- sime stipitato-glandulosae, apicem versus glabrae, ceterum extus intusque glaberrimae; ovaria subglobosa, glabra; calyx ovario circiter duplo longior, latus et plicatus, dentibus carnosulis inaequalibus 1.5—3 mm. longis, late ovatis apice rotundatis margine irregulariter erosulis glabris; corolla rosea (ex collectore), infundibuliformis, tubo 7-8 mm. longo basi manifeste gibboso supra paullo ampliato extus basi excepta sparse setoso-hirsuta, intus a medio ad faucem villoso-hirsuto, lobis late ovatis apice rotundatis 3 mm. longis glabris; stamina medio tubo affixa, an- theris 2.5 mm. longis faucem non attingentibus, filamentis glabris brevissimis; stylus medium tubum non superans, glaber. As Léveillé’s description is very brief and inaccurate particularly in regard to the calyx which he describes “‘calyce ciliato,” apparently taking the bracts for the calyx, I have given above a more complete description. The species seems nearest to L. nubigena Rehd., from which it chiefly differs in the bracts being quite glabrous except ciliate toward the base, in the large calyx, in the corolla being sparingly setose-hirsute outside, not short-pubescent and glandular, in the hirsute pubescence at the mouth with the anthers much below the mouth, not just reaching the mouth as in L. nubigena, and in the glabrous style. Lonicera lanceolata Wallich in Roxburgh, Fl. Ind. ed. 2, 2: 177 (1824). — Léveillé, Cat. Pl. Yun-Nan, 27 (1915). Lonicera acrophila Léveillé in Bull. Géog. Bot. 24(no. 301): 289 . (1914) ; Cat. Pl. Yun-Nan, 27 (1915). — Synon. nov. Cuina. Yunnan: haut plateau de Je-ma-tchouan, 3200 m., 338 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI E. E. Maire, July 1912, “arbre moyen, fl. roses” (holotype of L. acro- phila ; merotype in A. A.). According to Maire this is a medium-sized tree, but by most collectors it is described as a shrub, 4—8 ft. tall. Lonicera Koehneana Rehder in Sargent, Trees & Shrubs, 1: 41, t. 21 (1902). — Léveillé, Cat. Pl. Yun-Nan, 27 (1915 Lonicera gynopogon Léveillé in Bull. Géog. Bot. oe 301): 289 (1914) ; Cat. Pl. Yun-Nan, 27 (1915). — Synon. CHINA. Yunnan: brousse derriere uaa alt. 2550 m., E. E. Maire, May 1912, “Lonicera non grimpant, rameaux courts et gréles, fl. blanc-jaune” (holotype of ZL. gynopogon; photo. in A. A.); haies, plaine de Tong-tchouan, alt. 2500 m., FE. FE. Maire, May [1912], “fl. de Lonicera, mi-blanche, mi-jaune” (in herb. Léveillé under L. gyno- pogon; photo in A. A.); haies de Tchéou-kia-tse-tang, alt. 2500 m., E. E. Maire, “arbuste buissonante, haut 2 m., feuilles molles, velues et blanches en dessous, fl. mi-blanches, mi-jaunes, inodores” (in herb. Léveillé under L. gynopogon; duplicate in A. A.). Specimens from the same locality and partly apparently of the same collection have been distributed by the Arnold Arboretum under Maire, no. 142 and no. 286. Lonicera Pampaninii Léveillé in Fedde, Rep. Spec. Nov. 10: 145 (1911); Fl. Kouy-Tchéou, 64 (1914); Cat. Pl. Yun-Nan, 27 (1915). Lonicera Henryi var. setuligera W. W. Smith in Not. Bot. Gard. Edinb. 10: 47 (1917) CHInA. Kweichou: mont de Lou-tsong-koan, Tsin-gay, ro- cailles a Ché-tiou-tchay, Gan-pin, buissons et rochers de la montagne, L. Martin in herb. Bodinier, no. 1623, June 10, 1897 and June 27, 1899, ‘‘fleurs jaunes” (syntypes; merotype in A. A.). This species is similar to L. Henryi Hemsl., but is easily distinguished by the slenderer corolla-tube densely clothed with reflexed yellowish hairs; in the subsessile or sessile flowers with subulate pilose bracts exceeding the pilose calyx-lobes; the leaves which closely resemble those of L. Henryi are pilose on the midrib above and below otherwise glabrous even on the margin. This species has been collected in Kweichou also by Y. Tsiang near Tsun-yi and Pin-fa; nos. 5277 and 5377; also Steward, Chiao & Cheo no. 271 from Tsun-yi is probably the same, but it has no flowers. Lonicera macrantha Sprengel, Syst. Veg. 4°: 82 (1827). — Léveillé, Fl. Kouy-Tchéou, 63 (1915 Lonicera Guilloni Léveillé & Vaniot, in Bull. Soc. Bot. France, 51: exliv (1904). 1935] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 339 CHINA. Kweichou: Pin-fa, J. Cavalerie, no. 1015, May 28, 1903, “fl. blanches et jaunes au veillisant, sans odeur” (holotype of L. Guilloni; photo. in A. A.). Cavalerie no. 1015 is cited by Léveillé in his Flore du Kouy-Tchéou under L. macrantha (1. c.), but the name L. Guilloni is not mentioned. The specimen differs somewhat from typical L. macrantha in the shorter and slighter pubescence of the branches and in the scarcely ciliate leaves. Lonicera Esquirolii Léveillé, Fl. Kouy-Tchéou, 63 (1914). CHINA. Kweichou: without locality, J. Esquirol, no. 889, June 1903, “fl. jaunes apres floraison” (holotype; photo. and merotype ins Aas This species seems most nearly related to L. ferruginea Rehd., but is easily distinguished by the shorter, not hirsute pubescence and the glabrous ovary. From L. inodora W. W. Sm. it differs in the glabrous style, the glandular pubescence of the corolla, the sessile or subsessile inflorescence and in the setulose pubescence extending over the whole under surface of the leaf. Lonicera japonica Thunberg, Fl. Jap. 89 (1784). — Léveillé, Cat. Pl. Yun-Nan, 27 (1915); in Mem. Acad. Ci. Art. Barcelona, ser. 3, 12: 545 (Cat. Pl. Kiang-Sou, 5) (1916). Lonicera Fauriet Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 5: 100 (1908). — Synon. nov. Japan. Nippon: in littore Shiogama, U. Faurie, no. 6823, Oct. 1905 (holotype of L. Fauriei; photo. and merotype in A. A.). Léveillé compares his species with L. bracteolaris Boiss. & Buhse and describes the fruit as having 3 persistent hairy styles; he apparently mistook for styles the sepals which in one of the fruits appear to be only three, the other two not being clearly visible. Lonicera yunnanensis Franchet in Jour. de Bot. 10: 310 (1896). Lonicera Mairei Léveillé in Bull. Bot. Géog. 24: 289 (1914).— Synon. nov. CuIna. Yunnan: collines herbeuses autour de Tong-tchouan, alt. 2550 m., #. E. Maire, fl. blanc-jaunatre” (holotype of L. Mairei; merotype in A. A.). The branches of Maire’s specimen are apparently from different plants; one has the leaves quite glabrous beneath as in the type of L. yunnanensis, while in the other they are slightly pubescent beneath and are referable to var. tenuis Rehd., but there is no difference in the size of the leaves. 340 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XvI COMPOSITAE Pertya Bodinieri Vaniot in Bull. Acad. Intern. Géog. Bot. 12: 116 (1903).— Beauverd in Bull. Soc. Bot. Geneéve, sér. 2, 1: 386, fig. 6 (1909). — Léveillé, Cat. Pl. Yun-Nan, 47 (1915). Curna. Yunnan: environs de Yunnan-fou, dans les ravines de la montagne, EF. Bodinier, no. 10, Jan. 27, 1897, “‘tiges sous-ligneuses, de 0.6-1 m., fl. roses” (holotype; merotype [from herb. Léveillé] and photo. of isotype [in herb. Paris] in A. A.). Though this is not one of Léveillé’s species I have included it here, since the type is in the herb. Léveillé. Pertya Esquirolii Léveillé (in Bull. Géog. Bot. 24: 251 (1914); FI. Kouy-Tchéou, 100 (1914) from Kweichou, based on Esquirol no. 3633 is an herbaceous plant and belongs to Ainslea; it seems very near or identical with A. rubrifolia Franch. which I have not seen. (To be continued ) HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1935] REHDER, HUODENDRON 341 HUODENDRON, A NEW GENUS OF STYRACACEAE ALFRED REHDER With plates 151 and 152 and one text figure. Huodendron, gen. nov. Flores hermaphroditi, actinomorphi, pentameri; calycis tubus ovario adnatus, dentibus 5 triangularibus vel ovatis circiter dimidium tubum aequantibus; petala 5, initio basi coherentia, demum libera, lineari- oblonga, anguste imbricata vel valvata, sub anthesi revoluta; stamina 7-10, uniserialia, libera, petalis subaequilonga, sed ob petala revoluta valde exserta, filamentis complanatis linearibus, antheris anguste ob- longis introrsis, loculis distinctis, connectivo cum filamento continuo et supra antheras in appendicem conspicuum tri- vel rarius bidentatum elongato; ovarium inferum, triloculare; styli 3, triente inferiore vel fere ad apicem connati, stigmatibus capitellatis; ovula in quoque loculo numerosa, axi centrali affixa, erecta. Fructus capsularis, ovoideus, parva, triente infra apicem sepalis circumcincta, trilocularis, loculicide dehiscens, valvis interdum demum septicidis, endocarpio crustaceo, exo- carpio tenui; semina numerosa, scobiformia, minuta, oblonga vel ellip- tico-oblonga, leviter complanata, testa tenui reticulata, basi et apice fimbriata et saepius ad marginem sparse breviterque fimbriata, albumi- nosa, embryo centralis, rectus. — Arbor vel frutex ramis gracilibus, gemmis parvis nudis pubescentibus; folia decidua, alterna, petiolata estipulata, ovato-elliptica vel ovato-oblonga, acuminata, basi cuneata, integra vel remote minuteque denticulata, glabra vel fere glabra, penni- nervia, nervis curvatis anastomosantibus; inflorescentiae terminales et axillares, paniculatae vel subcorymbosae, ebracteatae et ebracteolatae, floribus satis parvis albis graciliter pedicellatis; capsula parva, pedicello recurvo. Ab aliis Styracacearum generibus, petiolis et staminibus liberis vel fere liberis, filamentis supra antheram in appendicem 3-vel 2-dentatum elongatis, stylo 3-fido, capsula valvis 3 dehiscente, seminibus scobi- ormibus numerosis bene distincta. Ob semina numerosa Alniphyllo affinis videtur, sed petalis et staminibus liberis, stylo trifido, connectiva appendiculato, capsula 3-loculari subinfera, seminibus scobiformibus circiter 1 mm. longis facile distinguitur. Type Species: Huodendron tibeticum (Anthony) Rehd. DISTRIBUTION: The genus is restricted to southern China and extends 342 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI northwest across the border into southeastern Tibet and northeastern Burma and south into northern Tonkin, where it occurs near Lao-kay, about 150 km. southeast of Mengtze. Within China it ranges from western Yunnan through southern Kweichou, to Kwangsi and Kwang- tung. Of the two species H. tibeticum is restricted to southeastern Tibet, about N. Lat. 29°, while H. biaristatum ranges from northeastern Burma to Kwangtung and extends south into Tonkin; it does not seem to occur north of N. Lat. 25°. The two species now known of the new genus were originally both referred to the genus Styrax to which the flowers bear a great resem- blance, but the fruit is entirely different. In Styrax the fruit is inde- hiscent or irregularly dehiscent and contains only one or two rather large subglobose or ellipsoid seeds, while the fruit of Huodendron resembles strongly that of some Saxifragaceae-Hydrangeae, as Deutzia and Hy- drangea, in shape and size and dehiscence of the capsule and in the numerous scobiform seeds; also the divided style recalls Saxifragaceae, and in some species of Deutzia the flattened filaments are elongated beyond the anther or are dentate at the apex. The petals and stamens fall off separately after anthesis, though in bud they are cohering at the very base; in Styracaceae free stamens and petals are very rare. Any doubt, however, one might have in regard to the affinity of Huodendron, is convincingly set at rest by the nodal structure of the stem, which shows the unilacunar nodes characteristic of all Ebenales, while the Rosales have trilacunar or quinquelacunar nodes, as pointed out by Dr. I. W. Bailey to whom I am indebted for the examination of the stem. As type of the genus I have selected Huodendron tibeticum, because this species represents the distinctive characters from Styrax and other allied genera in a more pronounced degree, particularly by the deeply divided style and by the absence of stellate or fascicled pubescence and also in the distinctly corymbose inflorescence. For the loan of additional specimens supplementing the material in the herbarium of the Arnold Arboretum (A. A.), I am indebted to Dr. E. D. Merrill of the New York Botanical Garden (N. Dady OT. Bs Mason of the University of California (U. Calif.) and to Sir William Wright Smith of the Royal Botanic Garden of Edinburgh (Edinb.). I take pleasure in associating with this new genus the name of Dr. H. H. Hu, director of the Fan Memorial Institute of Peiping, one of the foremost and active Chinese botanists, who has contributed and is still contributing extensively to our knowledge of the flora of China. Huodendron tibeticum (Anthony), comb. nov. Styrax tibeticus Anthony in Not. Bot. Gard. Edinb. 15: 245 (1927). 1935] REHDER, HUODENDRON 343 Arbor vel frutex 6-25 m. altus, ramis gracilibus teretibus vel apicem versus leviter complanatis glabris; folia alterna, sed interdum apicem ramulorum versus subopposita, decidua, papyracea, elliptico-ovata vel oblongo-ovata vel ovato-lanceolata, 6—11.5 cm. longa et 2.5—4 cm. lata, longe acuminata apice mucronulata, basi late cuneata, integra, nervis utrinsecus 5—9 utrinque leviter elevatis, costa apicem versus supra leviter Ficure 1. HvuopeNpRON TIBETICUM (Anth.) Rehd. 1. Flower. _—2. Stamens. X 8.—3. Flower with petals and stamens removed. x 8.—4. Cross-section of ovary. X 20.—5. Capsule. X 10.—6. Seed. x05 elevata basin versus plana, subtus manifeste elevata; petioli glabri, 5-10 mm. longi, supra leviter canaliculati. Inflorescentia glabra, corymboso- paniculata, terminalis 5—7 cm., lata, laterales cum pedunculo 1.5-3 cm. longo 4-8 cm. longa et 2.5—5 cm. lata; pedicelli graciles, 3-5 mm. longi, ut ramuli glanduloso-verruculosi; calycis tubus cupuliformis, glanduloso- verruculosus, 1 mm. longus, dentibus triangulari-ovatis dimidium tubum 344 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI subaequantibus ciliolatis; petala valvata, lineari-oblonga, 6-7 mm. longa et 1-1.5 mm., lata, obtusiuscula, extus tomentosula, intus fere glabra, sub anthesi revoluta; stamina petiolis subaequalonga, filamentis 4—5 mm. longis intus triente inferiore excepto villosis extus glabris, antherae 1.25-1.5 mm. longae, glabrae, apice tridentato circiter 1 mm. longo, dente medio lateralibus plerumque breviore, styli in triente inferiore vel ad medium connati, graciles, glabri; discus glaber. Capsula pedicello plus minusve recurvo suffulta, ovoidea, 3 mm. longa, fusco-brunnea, subinfera; semina brunnea, circiter 1 mm., longa. SOUTHEASTERN TIBET. Tsarong: Salween and Kiu-chiang divide, northwest of Si-chi-to, Lat. 28° 35’ N., Long. 98° 30’ E., alt. 10-11000 ft., G. Forrest, no. 21648, June 1922, “shrubby, 20-30 ft., flowers fragrant, white, in open thickets by streams” (holotype in herb. Edinb.) ; same locality, G. Forrest, no. 22882, Oct. 1922 (paratype in herb. Edinb.); Salween and Irrawaddi divide, near banks of Salween at Champutong, forests, alt. 7000 ft., J. F. Rock, no. 22020, May-July 1932, tree 70-80 ft. tall, flowers white (A. A., N. Y., U. Calif.); moun- tains west of Champutong, forests of upper Salween River, alt. 9000 ft., J. F. Rock, no, 22474, Oct. 1932 (A. A., N. Y., U. Calif.) This species has a very restricted distribution and is apparently con- fined to the mountains of extreme southeastern Tibet between the head- waters of the Irrawaddi and Salween Rivers. In some of its characters, particularly by the deeply divided style and by the absence of stellate or fascicled pubescence is it farther removed from other styraceous genera than the more widely distributed H. biaristatum. The fruiting branch of this species has some resemblance to certain species of Deutzia. Huodendron biaristatum (W. W. Sm.), comb. n Styrax rue W. W. Smith in Not. Bot. ee ‘Edinb. 1St250 1920 . C. Fischer in Kew Bull. Misc. Inform. 1933: 365. Frutex vel pos a 12 m. altus, ramis gracilibus hornotinis initio tomentosulis demum glabrescentibus, vetustioribus flavido-cinereis vel fusco-cinereis cortice demum rimoso vel fibroso vestitis. Folia alterna, papyracea, oblonga vel elliptico-oblonga vel obovato-oblonga, 8-17 cm. longa et 2.5—6 cm. lata, acuminata, basi cuneata, margine minute et remote denticulate vel integra, supra luteo-viridia, opaca, costa fascicu- lato-pilosula excepta glabra, subtus vix pallidiora, axillis saepe barbulatis exceptis glabra, costa supra leviter impressa subtus elevata, nervis utrin- secus 5—9 arcuatis margine anastomosantibus supra vix infra manifeste elevatis, venulis subtus elevatis; petioli 6-15 mm. longi, supra tantum vel undique fasciculato-pilosi. Inflorescentiae terminales et axillares, paniculatae, multiflorae, 3-10 cm. longae, ebracteolatae, cinereo- 1935] REHDER, HUODENDRON 345 tomentellae; pedicelli 2-5 mm. longi; calyx cupuliformis, tomentellus, tubo 1-1.5 mm. longus, dentibus late triangularibus acutiusculis tubo brevioribus; petala imbricata, anguste oblonga, 6—9 mm. longa et 2—2.5 mm. lata, utrinque tomentella; stamina petalis subaequilonga, filamen- tis compressis utrinque dense pilosulis circiter 3 mm. longis, antheris glabris 2 mm. longis connectivo dorso puberulo in appendicem triden- tatum vel rarius bidentatum elongata dentibus lanceolatis acutis medio plerumque minore; stylus staminibus paullo longior, crassus dense pilo- sulus, apice 3- esac: ovarium semisuperum. Capsula ovoidea, resu- pinata, 4-5 mm. ionaa: cinereo-tomentella, in triente superiore sepalis persistentibus cincta; semina 1—-1.25 mm. longa, flavo-fusca. Cuina. Yunnan: in thickets in ravines on the western flank of the Shweli-Salween divide, Lat. 25° 40’ N., alt. 9000 ft., G. Forrest, no. 18020, May 1919, “shrub 20-30 ft., flowers fragrant, creamy-yellow” (syntype in herb. Edinb.); side valleys of the Shweli-Salween divide, Lat. 25° N., alt. 8000 ft., G. Forrest, no. 17894, June 1919, “shrub 10-20 ft., flowers immature” (Edinb., A. A.) ; N’Maikha-Salween divide, at Ho-tou, in thickets and open forests, Lat. 25° 55’ N., alt. 7-8000 ft., G. Forrest, no. 18400, Aug. 1919, “shrub 12-18 ft., in fruit” (syntype in herb. Edinb.) ; same locality, G. Forrest, no. 18833, Nov. 1919 (syn- type in herb. Edinb.); Mengtze, S. E. mountain forests, 6000 ft., A. Henry, no. 10764 “tree 15 ft.’ (syntype in herb. Edinb., A. A., N. Y.); Mengtze, A. Henry, no. 136624, “shrub 10 ft.”’ (syntype in herb. Edinb.; A. A., N. Y.); south of Red River, A. Henry, no. 13662, “tree 40 ft.” (syntype in herb. Edinb.; A. A.); Shweli-Salween divide, Lat. 25° 10’ N., Long. 98° 50’ E., alt. 9000 ft., in open thickets and forests, G. For- rest, no. 26108, Dec. 1924, “tree 30—40 ft.” (Edinb., N. Y.); without precise locality, G. Forrest, no. 26108, 1924-25 (Edinb., N. Y.). Kweichou: Waichai, Tuh-shan, near border of Kwangsi, alt. 330 m., in densely shaded ravine, Y. Tsiang, no. 6686, Aug. 25, 1930, “tree 6 m., diam. of trunk 12 cm., bark pale gray” (A. A.). Kwangsi: Chin-fong, Lin-yuin-hsien, valley forest, alt. 1300 m., Steward & Cheo, no. 336, May 6, 1933, “tree 7 m., flowers white, fragrant” (A. A., N. Y.); Ta-tse-shan, Yung-hsien, forest, alt. 540 m., Steward & Cheo, no. 843, Aug. 21, 1933, “tree 9 m., fruit gray” (A. A., N. Y.). Burma: Myitkyina Distr., Htangan, 3100 ft., Sukoe per C. E. Parkin- son, no. 9197; Pyet Pass, 7200 ft., Sukoe per C. E. Parkinson, no. 10115 fee. b. (. bischer, 1.¢,); TONKIN: route de Lao-kay a Chapa, alt. 1500 m., A. Petelot, no. 3803, Aug. 1930 (N. Y.); massif du Fan-isi-pou, chemin du col de Lo-qui-ho, environs de Chapa, alt. 1400 m., A. Petelot, no. 4373, Sept. 1931 (N. Y.). 346 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI This species is readily distinguished from H. tibeticum by the pu- bescent inflorescence, the thicker texture of the leaves, the pubescent stout style 3-lobed only at the apex, the shorter filaments pubescent on both sides, the broader narrowly imbricate petals pubescent on both sides and the tomentulose capsules. The fact that the petals in one species of this genus are valvate and in the other imbricate is not un- usual in Styracaceae, for both kinds of aestivation are found in Styrax. The stamens are mostly 3-toothed at the apex, but the middle one is often shorter than the lateral ones; two teeth, as implied by the specific epithet, are only occasionally found. The leaves of H. biaristatum show some variation in dentation, tex- ture, pubescence and in the number of veins. The Forrest specimens have remotely denticulate leaves and are of rather thin texture, the leaves of the Henry specimens are occasionally furnished with minute denticulations reduced to a mucro, but are mostly entire like the other specimens and like those are of thicker chartaceous or subcoriaceous texture. The midrib is usually impressed and puberulous like the petiole, but in Petelot 4373 from Tonkin the midrib is glabrous except slightly puberulous toward the base and slightly elevated and quite glabrous toward the apex, also the lateral veins are slightly elevated and number about 5 pairs, while the leaves of the other specimens have mostly 6 or 9 pairs; by these characters this Petelot specimen approaches the following variety and connects it with the typical form. Huodendron biaristatum var. parviflorum (Merrill), comb. nov. Styrax parviflora Merrill in Jour. Arnold Arb. 8:15 (1927) A typo recedit praecipue ramulis foliis petiolisque glabris foliis magis coriaceis integris nervis utrinsecus 4-6, costa media nervisque supra glabris et elevatis, venulis subtus minus conspicuis. CuHInA. Kwangtung: Lung-t’au Mountain, near Iu, in forest, Canton Christian College, nos. 12070 (holotype in hb. N. Y.; A. A.) and 12349 (paratype in hb. N. Y.; A. A.). The flowers and fruits of the Kwangtung specimens, as far as can be judged from the rather poor material, are identical with those of typical H. biaristatum and the difference in the leaves does not seem sufficient to separate the Kwangtung form as a distinct species, considering the fact that the leaves of H. biaristatum show considerable variation and transi- tions to this variety. HERBARIUM, ARNOLD ARBORETUM, HaArvaArp UNIVERSITY. Jour. ARNOLD Ars. VoL. XVI. PLATE 151 Tent Cn TY OIN LOM PLANTAE ROCKIANAE Raped western Ch Plants of Southeast Tibet Meo dhecratren The h cvs» 4 ) Mt, Kenichuupo, eastern and western slopes, Salwin and Irrawady divide Joseph F, Rock, No, 2 May-July 19 os HUODENDRON TIBETICUM (Anth.) Rehd. Jour. ARNOLD Ars. VoL. XVI. PLATE 152 =e { WM .s me P| . E-i =e “% Nice owraleget barre Naber (hte. Sr) Fede, FLORA OF CHINA, Miran btere ba tat YM 5m Cour A HENRY HUODENDRON BIARISTATUM (W. W. Sm.) Rehd. 1935] KOBUSKI, STUDIES IN THEACEAE, I 347 STUDIES IN THEACEAE, I EURYA SUBGEN. TERNSTROEMIOPSIS CLARENCE E. KosuskKI With plate 153 Durinc the past year the author has been making a survey of the Old World Theaceae, starting with a critical study of the genus Eurya. The genus as a whole has presented some rather difficult problems in specific delimitation and in synonymy which cannot be settled until more ample material or photographs of types deposited in various for- eign herbaria can be had. In Szyszylowicz’s treatment,’ Eurya comprises three sections: Cleyera (DC.), Freziera (Sw.) and Proteurya Szysz. In 1896, Urban? separated E. sandwicensis from Proteurya and made it the type of the new genus, Ternstroemiopsis. The following year Engler’ united Ternstroemiopsis with Eurya as a new subgenus, elevating, at the same time, the three sections of Szyszylowicz to subgeneric rank. At present, however, Eurya is generally considered as containing but two subgenera, Proteurya and Ternstroemiopsis, while Cleyera and Freziera represent distinct genera, the former Asiatic and American, the latter exclusively American. In this paper, the subgenus Ternstroemiopsis is considered. This is distinguished from the subgenus Proteurya by the spiral arrangement of its leaves, the thick glandular sepals, fleshy petals and stamens whose anthers are twice as long as the filaments. Proteurya is characterized by two-ranked leaves, petals more or less membranous and anthers as long as or shorter than the filaments. Geographically also Ternstroemi- opsis is distinct being confined solely to the Hawaiian Islands while Pro- teurya, although found in nearly all the Pacific islands and Asia, does not invade the Hawaiian group with a single species. The institutions from which material for this study was borrowed along with the abbreviations used in this paper, are as follows: herbarium of the Arnold Arboretum of Harvard University (AA), herbarium of Otto Degener (D), Gray Herbarium of Harvard University (Gr), herbarium of the New York Botanical Garden (NY). 1$zyszylowicz in Engler & Prantl, Nat. ruieeas III. 6: 189 (1893). 2Urban in Ber. Deutsch. Bot. Gesell. 14: 49 (18 3Engler in Engler & Prantl, Nat. Pflanzenfam. Aare 1: 247 (1897). 348 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Key TO THE SPECIES AND VARIETIES A. Leaves subcordate or truncate at base, obtuse or rounded at apex B. Small trees or erect shrubs .............. 1. E. sandwicensis BB. Prostrate shrubs ......... la. E. sandwicensis var. prostrata AA. Leaves cuneate at base, acute at apex C. Leaves 3.54.7 cm. long, 1.2-1.7 cm. wide. .2. FE, Degeneri CC. Leaves either larger or smaller than C D. Leaves 5.3-8.7 cm. long, - 5- : 5 em. wide E. Degeneri {. grandifolia DD. Leaves 3.0-4.5 cm. long, 07 1.2 cm. wide 2b. E. Degeneri f. stenophylla 1. KEurya sandwicensis A. Gray, Bot. U. S. Expl. Exped. 1838- 1842, 1: 209 (1854).— H. Mann in Proc. Amer. Acad. Arts Sci. 7: 156 (Enum. Hawaiian Pl.) (1867); Mem. Boston Soc. Nat. Hist. 1: 534, 539 (1869).— Hillebrand, Fl. Haw. Isl. 41 (1888).— Drake del Castillo, Il. Fl. Ins. Maris Pacif. 117 (1890). — Szyszylowicz in Engler & Prantl, Nat. Pflanzenfam. 111, 6: 190 (1893).— A. A. Heller in Minn. Bot. Studies, 1: 856 (1897).— Rock, Indig. Trees Haw. Isl. 308 (1913).— Melchior in Engler & Prantl, Nat. Pflanzenfam. Ed. 2, 21: 147 (1925). Eurya sandwicensis A. Gray var. sessilifolia A. A. Heller in Minn. Bot. Studies, 1: 856 (1897), as a synonym. Ternstroemiopsis sandwicensis Urban in Ber. Deutsch. Bot. Ges. 14: 49 (189 Small trees, 5-6 m. in height, occasionally shrubby in higher altitudes, 2-3 m.; branches crowded with leaves, ultimate branchlets strigose ; leaves lone, elliptical or obovate, coriaceous, glabrous, occasionally strigosely hairy on midrib, 4.5-9.0 cm. long, 1.5-3.7 cm. wide, on short petioles 2-3 mm. long, sometimes subsessile, obtuse or rounded at the apex, more or less cordate, occasionally truncate at the base, closely serrulate with inflexed mucronulate teeth, veins and veinlets finely reticulate beneath, reddish brown in color; flowers solitary, occasionally two in axils, nodding, ebracteolate, pedicels approximately 5 mm. long: calyx purplish brown, quite coriaceous, subtended by two small unequal bracts; sepals five, unequal, 3—4 mm. long, persistent, suborbicular, thick in central portion, membranous, lighter in color and slightly glandular on margin, occasional strigose hairs on external surface; corolla pale yellow or cream-color, imbricated; petals five, obovate, 5—6 mm. long, united at base, somewhat fleshy in central portion; stamens in staminate flowers 10-15, slightly adnate to base of corolla, filaments distinct, half as long as the oblong mucronate anthers; staminodia in pistillate flowers, five sometimes six, 2-3 mm. long; pistil having three or occasionally 1935] KOBUSKI, STUDIES IN THEACEAE, I 349 four styles, sometimes connate nearly to stigma, usually divided; stig- mas three (or four); ovary glabrous, 3-celled, axial placentation; fruit a globose berry, 7-10 mm. across, dark blue-black, many-seeded; mature persistent, subcordate calyx-lobes 8 mm. long, 7.5 mm. across at widest portion, lobes at base lighter in color and more membranous. SPECIMENS EXAMINED: HawatltAn Istanps. Oahu: Nuuanu-Pali, U. Faurie, no. 284, October 1909 (AA); on mountains behind town of Honolulu, Wm. Rich, collected in 1840 (type) (Gr, NY); exact locality lacking, M. J. Remy, no. 562, collected 1851-1855 (Gr); data lacking, M. J. Remy (NY); exact locality lacking, H. Mann & W. T. Brigham, no. 524, col- lected 1864-1865 (Gr, NY); exact data lacking, W. Hillebrand (Gr); precise data lacking, C. Gaudichaud, collected probably 1836 (Gr); in rain-forest from Kahana church up ridge to summit of mountain south- east of Kahana Bay, O. Degener, no. 8680, July 3, 1932 (AA, D); Waipio, Waiawa Ridge, on Dicranopteris-covered ridge, O. Degener & Dr. C. L. Shear, no. 9838, March 5, 1928 (tree 15 ft.; fruit inky blue- black) (AA, D); open forest in Dicranopteris tangle, Manana Gulch ridge, O. Degener, W. Bush & K. K. Park, no. 8679, October 2, 1932 (AA, D); on and near the summit of Konahuanui, A. A. Heller, no. 2240, May 2, 1895 (NY); lower slopes of Konahuanui, above Manoa, A. A. Heller, no. 2311, May 13, 1895 (AA, NY, Gr); ridge west of Kalihi valley, C. N. Forbes, no. 1483.0, March 17, 1910 (NY); ridge between Pololo and Waialue iki, C. N. Forbes, no. 2408.0, January 30, 1917 (NY); Koolau Mts., Pumaluu, J. F. Rock, no. 627, December 3-10, 1908 (Gr); Pumaluu, J. F. Rock, no. 843, December 1908 (NY); pre- cise data lacking; J. F. Rock, collected 1910 (Gr). Kauai: Mt. Waialeale, alt. 5200 ft., J. F. Rock, no. 8864, October 20, 1911 (Gr, NY); along stream-beds, Kaholuamano, J. F. Rock, no. 5499, Septem- ber 1909 (Gr, NY); Kaholuamano, J. F. Rock, collected March 3-10, 1909 (NY); Hanapepe, U. Faurie, no. 286, December 1909 (AA). Mauai: MHonakahau Drainage Basin, C. N. Forbes, no. 421.M, September 25—October 17, 1917 (NY). Hawaii: Kilauea, near fern-forest, O. Degener, no. 8678, November 10, 1929 (AA, D). In the whole genus this species is probably the most outstanding. Although confined to the Hawaiian group, it has been found in nearly all the islands from which material has been collected. The flowers and fruit are nearly twice the size of any other species. Along with this size character can be mentioned the distinct reddish reticulate veining of the lower surface of the leaf and the subcordate or truncate base of the leaf. The collections of Otto Degener made during the last few years on 350 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI the various islands of the group had great influence in the decision finally to place the majority of Hawaiian specimens in this species. His specimens were so ample that it was possible to make two and even three sheets of each for the herbarium of the Arnold Arboretum. These sets of material show great gradation in leaf-size — a character which might cause some, especially in this genus, to describe new species. Faurie’s specimen, no. 286, according to the collector, was made at Hanapepe, Oahu. This probably is a mechanical error made in trans- ferring the field notes to the herbarium label. Hanapepe is on the island of Kauai and the collection date of no. 286 agrees with other material collected by Faurie on Kauai. la. Eurya sandwicensis A. Gray var. prostrata, var. nov. A typo recedit habitu prostrato et foliis remotis. SPECIMENS EXAMINED: HawanlAn Istanps. Molokai: At edge of windswept forested pali, Ohialele Pali, O. Degener, no. 8676 (type, AA) May 10, 1928 (more or less trailing along ground with branches sometimes eight feet long; flowers yellow, difficult to distinguish plant from Vaccinium) (AA, D); Pelekunu trail, C. N. Forbes, no. 249.Mo, July 1912 (NY). A quotation from a recent letter from Otto Degener, the collector of the cited type throws considerable light on this variety and the species E. sandwicensis : —“Eurya, as 1 have found it on Oahu, grows as an erect small tree with very dense foliage. It is rare, and where found, usually grows in openings in the lower forest, covered over with Glei- chenta, Rainfall would be moderate.” “The Molokai specimen I have found nowhere except in a typical dense extremely rainy rain-forest, and curiously enough, not anywhere in that region but only on the brink of a cliff extending for several miles. In short, it grows on the very “backbone” of Molokai where the rain and fog drive violently over the mountain crest. I collected five months on Molokai and do not remember seeing any Eurya except in that one type of locality. The rain-forest reaches up to this cliff and it is among the shrubs and small trees immediately overlooking the cliff that the Eurya is to be found. The plant sprawls rather than creeps, producing slender branches of unusual length—possibly 12 feet—with its leaves spaced far apart.”’ The collector remarked further that at first on seeing sterile plants, he thought this variety to be a low-growing Vaccinium. However, later on finding flowering material, he discovered it to be an Eurya. He suggested it as a variety or possible new species. The second cited specimen, Forbes no. 249.Mo, resembles the type 1935] KOBUSKI, STUDIES IN THEACEAE, I 351 in the remoteness of leaves. However the habit of the plant was not given by the collector, but it appears to be prostrate. 2. Eurya Degeneri, spec. nov.; a E. sandwicensi A. Gray foliis ellipticis 3.5—4.7 (2.5-6.5) cm. longis, 1.2—-1.7 (1.2—-2.2) cm. latis, apice acutatis et emarginatis, basi cuneatis recedit. Branches covered with leaves especially at ends, ultimate branchlets sparsely strigose; leaves elliptic, coriaceous, 3.5—4.7 (2.5—6.5) cm. long, 1.2-1.7 (1.2-2.2) cm. wide, acute at the apex, emarginate, cuneate at the base, closely serrulate with inflexed mucronulate teeth, conspicu- ously reticulate on under surface, veins and veinlets reddish brown in color, especially near base of leaf where color spreads into the leaf; petiole 3 mm. long; mature flowers unknown, bud resembling E. sand- wicensis sufficiently in coriaceous character and color of calyx with occasional strigose hairs, character and number of stamens and corolla to show it to be typical of the genus and closely related to E. sandwi- censis; berry (probably not mature) blue-black, 5 mm. across with the persistent styles separate to near the base, 3-celled, axial placentation, many-seeded. SPECIMENS EXAMINED: HawatlAN IsLANps. Kauai: open forest, Waineke swamp Kokee, O. Degener, no. 8675 (type AA) July 1, 1926 (AA, D); high plateau of Waimea, Halemanu to Kaholuamano, J. August Kusche, nos. 28, 139, 140, collected in 1919 (AA); Kilauea, U. Faurie, no. 285, January 1910 (AA); west side Waimea Drainage Basin, Kanaikinaua, C. N. Forbes, no. 1016.K, July 3—August 18 (1917) (AA, NY). This species is very closely allied to E. sandwicensis. The leaf char- acters are most distinctive between the two species. Eurya Degeneri has elliptical leaves, cuneate at the base, acute and emarginate at the apex. On the other hand, Eurya sandwicensis has leaves which are oblong or elliptic, subcordate or truncate at base and rounded or obtuse at the apex. Eurya Degeneri and its varieties are confined to the island of Kauai, while Eurya sandwicensis is found on nearly all the islands in- cluding Kauai. Otherwise these two species belonging to this distinctive section of the genus are very similar. This is especially true in flower and fruit characters. Although the mature flowers and fruit were not available in E. Degeneri, the material such as it is shows conclusively that there is a great resemblance. It is a pleasure to dedicate this species to Otto Degener of Hawaii, whose recent collections from the islands are extremely fine and whose material of Eurya aided tremendously in clearing up this section. 352 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI 2a. Eurya Degeneri Kobuski forma grandifolia (Wawra), comb. Eurya sandwicensis A. Gray 8 var. Hillebrand, Fl. Hawaiian Isl. 41 (1888).— Drake del Castillo, Il. Fl. Ins. Maris Pacif. 117 (1890). Eurya sandwicensis A. Gray var. grandifolia Wawra in Flora, 56: 168 (1873).— J. F. Rock, Indig. Trees Haw. Isl. 308 (1913). A typo recedit foliis amplioribus, 5.3-8.7 cm. longis, 2.5-3.5 cm. latis. SPECIMEN EXAMINED: HawalIAN ISLANDS: Kauai: Wainiha, U. Faurie, no, 298, Janu- ary 1910 (AA). This large-leaved form has seemed rather evasive to most collectors. Hillebrand, Del Castillo, Wawra and Rock have made reference to it in literature. The first three in their treatments were working with a single specimen, that of Wawra collected at Kealia on the island of Kauai. Incidentally we have only the single specimen collected by Faurie (no. 298) collected at Wainiha. The Faurie specimen was collected on the north coast of Kauai while Wawra made his collection on the west coast. Rock, although having collected considerably on the islands, never encountered this large-leaved form. However, he collected the narrow- leafed form cited next. These two forms like the species have distinctly cuneate leaf-bases and acute apices. Their variation from the species lies chiefly in the leaf size. Again, like the species, they are found only on the island of Kauai. 2b. Eurya Degeneri Kobuski forma stenophylla, forma nov. A typo recedit foliis minoribus angustioribusque, 3.0—4.5 cm. longis, 0.7-1.2 cm. latis. SPECIMEN EXAMINED: Hawaiian Istanps. Kauai: precise locality and date of col- lection lacking, J. F. Rock, no. 17274 (type) (AA). Unfortunately, the Rock specimen cited above is sterile and was placed in this genus under E. sandwicensis with some hesitation by an earlier student. At first, I was quite dismayed because I felt that it belonged to a species other than EF. sandwicensis, but because of the lack of flowers or fruit I hesitated to describe it as new. It was not until more material came to my attention that its true affmity with E. Degen- ert was discovered. HERBARIUM, ARNOLD ARBORETUM, HarvVARD UNIVERSITY. Jour. ARNOLD Ars. VoL. XVI. PLate 153 EuryaA DEGENERI Kobuski 1935] PALMER, TWO NEW SPECIES OF CRATAEGUS 353 TWO NEW SPECIES OF CRATAEGUS FROM MISSOURI ERNEsT J. PALMER With two text figures Crataegus hannibalensis, sp. nov. Arbor 6-8 m. alta vel frutex arborescens 4-6 m. altus, ramulis annotinis gracilibus vel paulo validis glabris fusco-viridibus, spinis numerosis. Folia ovata vel obovata vel oblongo-ovata, serrata, apice acuta, basi cuneata in petiolum 8-12 mm. longum attenuata, 2.5—4 cm. longa, 2—3 cm. lata, surculorum vali- dorum ad 5-6 cm. longa, 4-5 cm. lata, matura firma, crassa, glabra, dentata, nervis superne manifeste impressis. Inflorescentiae glabrae, laxae, plerumque 5—16-florae; flores 14-16 mm. lati, staminibus circiter 10, stylis 1-3 plerumque 2; sepalis lanceolatis integris vel sparse denta- tis. Fructus ovoideus, obovoidea-oblongus vel rare subglobosus, 8-10 mm. longus, 7--8 mm. latus, firmus, viridis denique rubro-luteus raro pruinosus, seminibus 2-3 plerumque 2 ovalibus dorso sulcatis. A tree 6-8 m. tall, or sometimes an arborescent shrub 4—6 m. tall, with intricate ascending or horizontal branches and slender to stoutish flexu- ous branchlets, glabrous and olive-green or olive-brown at the end of the first season, usually armed with numerous stout, straight or curved purplish thorns 3—6 cm. long. Bark gray or pale brown, slightly scaly. Leaves obovate, oblong-obovate or oval, acutely pointed, short-acumi- nate or rarely rounded at apex, cuneate at the base and attenuate into the short 8-12 mm. long petioles, sharply serrate usually nearly to the base, glabrous, firm to subcoriaceous at maturity, yellowish-green above and slightly paler beneath, with slender but prominent mid-rib and 5-7 pairs of parallel veins elevated on the under surface and conspicu- ously impressed above, those of the fruiting branches mostly 2.5—4 cm. long and 2—3 cm. broad, on vigorous sterile shoots often 5—6 cm. long and 4—5 cm. broad, and with margins coarsely serrate or dentate. Flow- ers in loose glabrous compound 5—16-flowered corymbs, 14-16 mm. in diameter; pedicels slender, often glandular, 1-2 cm. long; stamens about 10; anthers in specimens examined pale yellow; styles 1-3, usu- ally 2; calyx-lobes lanceolate, entire or slightly serrate towards the base, glabrous without and glabrous or slightly villous within. Fruit oval, oblong-obovoid or rarely nearly globose, 8-10 mm. long, 7-8 mm. thick, hard and green until late in the season, turning dull red or orange-red and becoming mellow when fully ripe late in September, rarely with a 354 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI slight pruinose bloom. Fruiting calyx sessile or slightly elevated, with a broad shallow cavity; calyx-lobes often persistent and appressed, flesh thin; nutlets 1-3 but usually 2, relatively large, oval or elliptic in outline, blunt or rounded at the ends, and with broad shallow ridges and grooves on the dorsal surface. W.B.Pa st. Figure 1. CRATAEGUS HANNIBALENSIS E. J. Palmer. x 2/3 Thickets and borders of woods, in fertile soil, on limestone hills or often along bluffs and banks of streams. Crataegus hannibalensis is rather abundant in northern Missouri and southeastern Iowa, and it is probably more widely distributed. A specimen collected in western Ohio seems to belong here. This species is conspicuous and easily distinguishable on account of its rather large (for the group) yellowish-green leaves with deeply im- pressed veins, the pale olive branchlets and comparatively small oval or oblong fruit. In their deeply impressed veins the leaves resemble some- what certain species of the Punctatae group, but the characters of the fruit and flowers and the entire absence of pubescence seem to place it 1935] PALMER, TWO NEW SPECIES OF CRATAEGUS 355 clearly in the Crus-galli group. The type specimen is in the herbarium of the Arnold Arboretum. Missouri: Hannibal (Marion Co.), John Davis, no. 177, Oct. 6, 1911, May 14, Oct. 10, 1912, Oct. 13, 1913; E. J. Palmer, no. 20381, Sept. 7, 1921; no. 20382 (type), Sept. 7, 1921; no. 20405, Sept. 8, 1921; no. 22337, Oct. 24, 1922; south of Hannibal (Ralls Co.), John Davis, no. 1645, Oct. 4, 1916; Eolia Pike Co., John Davis, no. 25, Sept. 30, 1912; no. 2147, Sept. 20, 1913; no. 2149, Sept. 22, 1913; no. 2153, Sept. 21, 1913; Dumas, Clark Co., B. F. Bush, no. 10139, July 28, 1923; be- tween Renick and Clark, Macon Co., E. J. Palmer, no. 35943, May 21, 1929: between Lancaster and Downing, Schuyler Co., Palmer & Steyer- mark, no. 40970, June 30, 1933; Mill Grove, Mercer Co., no. 41270, July 4, 1933; Shelbina, Shelby Co., no. 40865, June 28, 1933; Eagleville, Harrison Co., no. 41340, July 6, 1933; St. Francois Co., C. S. Sargent, Oct. 5, 1899. Ohio: Springfield, R. E. Horsey, no. 338, May 17, Oct. 25, 1915. Iowa: Keokuk, Lee Co., E. J. Palmer, no. 21829, Sept. 6, 1922; no. 21831, Sept. 6, 1922; no. 40595, June 25, 1933; Bur- lington, Des Moines Co., EF. J. Palmer, no. 21800, Sept. 6, 1922. Crataegus Danielsii, sp. nov. Arbor 6—7 m. alta vel frutex arbor- escens 4—6 m. altus. Folia oblongo-ovata, elliptica vel rhombica, grosse serrata, saepe supra medium obscure inaequaliter incisa, apice acuta vel acuminata, basi cuneata in petiolum gracillimum 8-15 mm. longum attenuata, matura papyracea sed firma, superne glabra, infra paulo villosa, 2.5—4 cm. longa, 1—-2.5 cm. lata, ramulorum sterilium ad 5—6 cm. longa 3-5 cm. lata. Inflorescentiae laxae, ramosae, paulo villosae, 6—15- florae, bracteis linearibus glanduloso-serratis. Flores 14-16 mm. lati, staminibus circa 12-15, antheris rubicundis, stylis 2—4, plerumque 3, sepalis lineari-lanceolatis integris vel paulo glanduloso-serratis. Fructus subglobosus, 8-12 mm. latus, maturus rubicundus; seminibus 2—3 dorso sulcatis. A tree 6-7 m. tall, or sometimes an arborescent shrub 4-6 m. tall, with erect or ascending intricate branches and slender branchlets, more or less villous when young in the typical form, and armed with slender thorns 2-3 cm. long. Leaves oblong-elliptic, oblong-obovate or narrowly rhombic in outline, sharply and irregularly serrate, often obscurely incised above the middle with one or more pairs of shallow lobes or unsymmetrical with one or more odd lobes, pointed or acuminate at the apex, attenuate at the base into the slender 1—1.5 cm. long petioles, usu- ally red as they unfold, and then villous on both surfaces, thin but firm at maturity, glabrous above and more or less villous along the veins beneath, those of the fruiting branches mostly 2.5—4 cm. long and 12.5 356 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI cm. broad, and up to 5—6 cm. long and 3—5 cm. wide on vigorous sterile shoots; petioles usually sparsely villous, sometimes with a few scat- tered glands. Flowers 14-16 mm. in diameter, in loose slightly villous compound corymbs; pedicels slender, 8-15 mm. long, glabrous or sparsely villous; bracts numerous and conspicuous, narrowly linear, finely glandular-serrate on the margins; stamens usually 12—15; anthers pink or rose-color in specimens examined; styles 3—4, usually 3; calyx- lobes linear-lanceolate, entire or somewhat glandular-serrate towards H.B.Rust. Ficure 2. Cratarecus Dantevsit E. J. Palmer. X 2/3 the base. Fruit subglobose, 8-12 mm. in diameter, pruinose, becoming dull crimson when ripe in late September or October, flesh thin and hard; nutlets 2—3, usually 3, oblong, rounded at the ends and with broad shallow grooves and ridges on the dorsal surface. Limestone glades and hillsides in the vicinity of Columbia, Missouri. Several trees referable to this species have been found, all so far as known within a few miles of the type locality. The extremely local distribution, the variable and often asymmetrical outline of the leaves and the sparse and variable pubescence, nearly or quite absent in some specimens, all suggest the possibility of a hybrid origin, and it may have 1935] PALMER, TWO NEW SPECIES OF CRATAEGUS 357 originated as a cross between Crataegus crus-galli and C. verruculosa, both of which are growing in the immediate vicinity. The specific name is for Dr. Francis Daniels, author of a Flora of Columbia, Missouri, and vicinity, who first collected it there. Missouri: near Columbia, Boone Co., Hawthorn glades, north of Columbia, Francis Daniels, Sept. 26, 1902; May 3, 1903; W. H. Rickett, no. 8 (Crat. #8), 50 yds. west of Balanced Rock, May 3, 1931; no. 36 (Crat. #35), north side of Walnut St., west of highway 63, May 6, 1931: Francis Drouet (Crat. #8, W. H. R. #107), 50 yds. west of Bal- anced Rock, Oct. 4, 1931; (Crat. #8, W. H. R. #69), west of Balanced Rock, Sept. 15, 1931; (Crat. #35, W. H. R. #71), north side of Walnut St., west of highway 63, Sept. 22, 1931; E. J. Palmer, no. 39265 (type), May 4, 1931; near Hinton, Boone Co., W. H. Rickett, no. 40 (Crat. #39), 4.4 miles north of Hinton, May 17, 1931; no. 43 (Crat. #42), 1.3 miles south of Hinton, May 17, 1931; 84 (Crat. #54), 3.6 miles north of Hinton, Sept. 31, 1931; no. 86 (Crat. #56), north of Hinton, Sept. 30, 1931; no. 88 (Crat. 42), 1.3 miles south of Hinton, Sept. 30, 1951: no. 91 (Crat. #43), 2.3 miles south of Hinton, Sept. 30, 1931. Type in the herbarium of the Arnold Arboretum. All other specimens examined are in the herbarium of the University of Missouri. In a few specimens examined the young foliage, branches, and in- florescence are quite glabrous and in others there is only the slightest trace of pubescence in the form of a few scattered hairs on either the pedicels, petioles or veins of the leaves. This may be distinguished as Crataegus Danielsii {. glabra, f. nov." Thickets, limestone hills and glades, Boone County, Mo. With the type. W. H. Rickett, no. 39 (Crat. #38), 4.4 miles north of Hinton, Mo., May 17, 1931, in the herbarium of the University of Missouri, may be taken as the type of this form. HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1A typo differt ramulis foliis inflorescentiis glabris vel raro leviter pilosis. 358 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI NEW HYBRIDS FROM THE ARNOLD ARBORETUM EpGarR ANDERSON AND ALFRED REHDER >< Akebia pentaphylla (Mak.) Makino in Tokyo Bot. Mag. 16: 30 (1902) = A. quinata Dene. X trifoliata (Thbg.) Koidz. Akebia trifoliata var. pentaphylla Makino in Tokyo Bot. Mag. 5: 329 (1891 Artificial hybrids between Akebia trifoliata Koidz. and A. quinata Dene. were produced at the Arnold Arboretum in 1932 by Dr. Karl Sax. Though they have not yet flowered the hybrid seedlings have now reached a stage where their intermediate character is clearly evident and is in close agreement with Makino’s description (1. c.) of & A. penta- phylla, a putative hybrid widely distributed in Japan. The hybrids, on the whole, resemble A. trifoliata somewhat more closely than they do A, quinata, As yet many of the leaves are three-foliolate, although leaves with four and five leaflets have been produced. It is of interest that Makino originally considered A. pentaphylla as a variety of A. trifoliata (1. c.). In this connection one might speculate as to the origin of A. trifoliata Koidz. var. australis (Diels) Rehd. Diels? in describing the variety commented on its extreme variability, and it occurs in a region where both A, ¢rifoliata and A. quinata are native. Further- more, it is intermediate between the two species in both leaf and flower, though resembling A. trifoliata more closely. It seems not impossible that it has resulted through extensive hybridization between A. trifoliata and A. quinata in a region where the former is relatively more abundant. It should be remembered that the consequences of hybridization be- tween two species may be quite different in different parts of their ranges, depending upon the relative frequency of the two species, the presence of polyploid races within either parent, the adaptability of the hybrid to local conditions, etc. Such matters are usually highly specu- lative. ‘The production of an artificial hybrid will make it possible, ulti- mately, to study such questions experimentally in the genus 4kebia. E.A Prunus Juddii FE. Anderson, hybr. nov. = P. Sargentii Rehd. 9 yedoensis Mats. 2. ‘Arnold Arb. Bull. ser. iv. 2: 17-20 (1934). — They are growing in the Arboretum under no. 624-32 and sfiecimens Neier June 3, 1935 are preserved in the 2Bot. Jahrb. 29: 344 (1900). 1935] ANDERSON AND REHDER, NEW HYBRIDS 359 Intermedia inter parentes, et ab utroque differt praecipue inflorescen- tiis 2—6-floris breviter racemosis et breviter pedunculatis, calycis lobis sparse et leviter glanduloso-serratis, stylo basi sparse piloso. Growing in the Arnold Arboretum under no. 22489 and type specimens collected May 5 and 10 and June 3, 1935, are preserved in the herbarium. An upright tree with spreading branches. Branchlets glabrous. Leaves ovate, acuminate, doubly serrate, dull brownish green when un- folding, glabrous throughout. Flowers before the leaves in very short- peduncled racemes of two to six, subtended by greenish bracts. Pedi- cels with weak scattered hairs at the base. Petals oblong, white or whitish, flushed with deep rose pink (Ridgway). Calyx-tube cylindric to sub-urceolate, glabrous. Calyx-lobes weakly and _ irregularly glandular-serrate. Style with scattered hairs at the base. Fruit black. Among the seedlings of Prunus Sargentii Rehd. (Prunus serrulata Lindl. var. sachalinensis [F. Schmidt] Mak.) which have been raised from the original trees at the Arnold Arboretum were certain plants which are evidently hybrids between that species and other cherries which were flowering at about the same time. In the case of one of these specimens the evidence for its exact parentage is so clear and the hybrid tree promises to be of such horticultural importance for New England that it seemed desirable to provide the hybrid with a scientific name. I take pleasure in naming the hybrid after the propagator for the Arnold Arboretum, Mr. W. H. Judd, whose precise record of the material which has passed through his department is of great scientific importance. The hybrid originated in 1914 at the Arnold Arboretum as a seedling of one of the original trees of Prunus Sargentii raised from seed sent from Japan by Dr. W. S. Bigelow in 1890. Prunus yedoensis was acquired in 1902 and for many years a large specimen stood adjacent to Prunus Sargentii, no. 5777. Since their flowering dates usually overlapped it is not at all surprising that cross-fertilization should have taken place. Mr. Edwin L. Hillier of the West Hill Nurseries, Winchester, England, writes me that he has obtained similar hybrids from seed sent him from the Arnold Arboretum. Since seed of both Prunus yedoensis and P. Sargentii have been distributed very widely for a number of years by the Arnold Arboretum, it is quite possible that the hybrid may have turned up in a number of nurseries and gardens. s< Prunus Juddii has proved hardy during the last two phenomenally cold winters though it is planted at the edge of one of the coldest spots in the Arnold Arboretum. It furthermore holds its flowers longer than does P. Sargentii and is a thrifty quick-growing tree. From P. Sargentit 360 JOURNAL OF THE ARNOLD ARBORETUM [vOL. xvI it can most easily be distinguished by the greener young leaves, by the scattered hairs at the bases of the style and the pedicel, and by the glandular serrations of the calyx. From P. yedoensis it can be distin- guished by its brighter flowers and by its glabrous calyces and leaves. A more complete comparison of the hybrid and the parental species is given in Table I. TABLE I. COMPARISON OF x PRUNUS JUDDII WITH ITS PARENTS P. yedoensis x P. Juddi P. Sargentii branches spreading to horizontal leaves greenish when unfolding flowers in 2—6-flowered short-peduncled racemes pedicels finely pubescent — apenes oblong, wee a urceolate- pr hareeg finely pubescent calyx-lobes strongly glandular-serrate style pubescent branches spreading leaves dull brownish green when eras flowers in 2-4 red very hii santa neler with weak scattered hairs at the base petals Milas flushed with r pink calyx tube ee urceolate, an ie calyx dee weakly and regularly glandular se stots with a hairs at the bas branches upright sat bright bronze n when u nfolding s in sessile or ati sessile clusters pedicels glabrous petals narrowly oblo typically apr rose we calyx tube c san bateiet aid ie calyx-lobes entire style glabrous Since it has not been found wild, Prunus yedoensis has itself been thought to be a hybrid between Prunus Lannesiana and Prunus sub- hirtella.* The fact that it comes true from seed? makes this hypothesis less likely, though such true-breeding hybrids are not unknown in the genus Prunus. EA < Viburnum Juddii Rehder, hybr. nov. = V. Carlesii Hemsl. 2 bitchiuense Mak. 6 A Viburno Carlesii praecipue differt foliis supra minus dense pilosis, petiolis paullo brevioribus, corymbo laxiore magis multifloro, corolla extus magis roseo suffusa graciliore, limbo paullo minore, lobis angusti- oribus filamentis quam antherae longioribus; A V. bitchivensi differt praecipue foliis supra magis pilosis, petiolis paullo longioribus, 5-7 mm. longis, corymbo 6—7 cm. diam. magis florifero, corollis majoribus tubo 1Wilson, E. H. The Cherries of Japan, p. 19. ae (1916). *Russell, Paul. The Oriental Flowering Cherries Washington (1934). 8C. D. Darlington in Jour. Genet. 19: 213-256 oa 1935] ANDERSON AND REHDER, NEW HYBRIDS 361 9-10 mm. longo, limbo 14-15 mm. diam., lobis paullo latioribus circiter 5 mm. latis, staminibus medio tubo affixis antheris faucem attingentibus. Growing in the Arnold Arboretum under no. 447—20; type specimens collected May 14, 1929, May 9, 1930, May 14, 1931 and May 14, 1935. This hybrid is in almost all characters intermediate between the parent species which are closely related and very similar, the chief difference being in the stamens which in V. bitchiuense are inserted in the lower fourth or third of the corolla-tube with the filaments about twice as long as the anthers and the tips of the latter 1.5-2 mm. below the mouth of the corolla-tube, while in V. Carlesii the stamens are inserted above the middle with the filaments as long or shorter than the anthers which reach the mouth of the corolla-tube. Table II shows the chief characters by which the hybrid may be distinguished from the parent. TABLE II. COMPARISON OF VIBURNUM JUDDII WITH ITS PARENTS V. bitchiuense x V. Juddi V. Carlesii Leaf broad ovate to ovate ovate to ovate-oblong ovate to oblong-ovate, or elliptic, sparingly or elliptic, furcate- rather densely furcate- furcate-pilo , pilose above, bright pilose and grayish green slightly lustrous above green, not rugose when young, not rugose and usually rugose Petiole 2-7 mm. long 4—9 mm. long 5-12 mm. long Inflor- 4-5 cm. across, rather 6-8 cm. across, rather 4.5-6 cm. across, com- escence loose, rays 7-12 loose, rays about 1.5 pact, rays 5-8 mm. long, slender cm. long, slender long, stout Corolla pink outside, tube 7-8 pink outside, tube 9-10 corolla faintly flushed m. long, limb 12-14 m. long, limb 15-16 pink outside, tube 7-8 mm. across, lobes 4-5 mm. across, lobes about . long, limb 15-16 mm. broad 5 mm. broad mm. across, lobes 5-6 mm. broa Filaments ee 2 the lower inserted about or slight- inserted above to near third the corolla ly below’ the middle, the middle as long or tube, Feaut twice as about 1-% as long as_ slightly longer than long as anth anthers anthers Anthers tips reaching the mouth _ tips reaching the mouth tips 1.5-2 mm. below the mouth As shown by the table above, the hybrid holds the middle between the two parent species except in the size of the inflorescence and the length of the corolla-tube, in which it exceeds both parents. In its general appearance it resembles more V. bitchiuense on account of its looser habit and the looser inflorescence and more brightly pink flowers. As an ornamental plant it is superior to either parent. Viburnum Juddii was raised in 1920 by Mr. William H. Judd of the Arnold Arboretum staff from seed of V. Carlesii. The largest plant of the hybrid is now 2 m. tall and flowered for the first time in 1929. Like the parent species it has stood the severe cold of the last two winters without injury to its flower-buds. AR 362 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI < Syringa diversifolia Rehder, hybr. nov. = Syringa pinnatifolia Hemsl. ° odlata Lindl. var. Giraldii (Lemoine) Rehd. ¢. A Syringa pinnatifolia differt praecipue foliis partim integris, partim basi pinnatifidis pinnis 1-4 ovato-oblongis vel anguste ovatis 2-3 cm longis acuminatis basi anguste decurrentibus leviter ciliolatis ceterum glabris, foliolo terminali ovato-oblongo sensim acuminato 3.5-5 cm. longo, foliis integris ovato-oblongis, 3.5—5 cm. longis, 1.4-2.2 cm. latis, basi rotundatis, inferioribus interdum fere ovatis, inflorescentia ad 12 cm. longa et laxiore, corolla coeruleo-lilacina, tubo circiter 8 mm. longo, limbo circ. 1 cm. diam., lobis apice leviter cucullatis, antheris faucem paene attingentibus; a S. oblata var. Giraldii recedit praecipue foliis partim pinnatifidis minoribus et angustioribus, gemma _ terminali ramorum evoluta et ramum foliiferum emittente, inflorescentia minore, corollae tubo breviore et limbo angustiore, antheris faucem attigentibus. Growing in the Arnold Arboretum under no. 148-30; type specimens collected May 17 and 21, 1935, preserved in the herbarium. A comparison of the chief characters by which the hybrid differs from its parents are given in Table ITI. COMPARISON OF SYRINGA DIVERSIFOLIA WITH ITS PARENTS TABLE III. S. pinnatifolia S. diversifolia S. oblata var. Giraldii Leaf arog with 7-11 leaf- partly entire and partly always entire broad lets, 3-6 cm. long, leaf- pinnatifid with 3-5 ovate 4-10 cm. long, lets’ 4-10 mm. broad, leaflets 4-6 cm. long, and 3-6 cm. broad, finely ciliate when lateral leaflets ie 14mm. — glabrous young sere entire leaves 2.5 cm. ey glabrous Branches with terminal bud with or without ter- without terminal bud minal bud Panicles 4-7 cm. long, usually to 11 cm. long, usually to 15 cm. long: usually several pairs along the on ir at end of one pair at end of branches, sessile branches, sessile branches, ‘cline Corolla white, usually ing whitish or bluish lilac, lilac or purple lilac, pale lilac, - be 5-6 m a to _ —_* tube tube 15-18 mm. long, lon g, lim ong, limb out 1.5 cm. pein pie oval-ovate, limb ont ic cm. "across, across, lobes oblong, not cucullate lobes ies slightly cu- strongly cucullate cullat Anthers © slightly exserted anther just reaching anthers about 1.5 mm. the uth below the mouth This hybrid was raised in 1929 from seed collected in 1929 from S. pinnatifolia Hemsl. the flowers of which were apparently pollinated by a plant of S. oblata var. Giraldii (Lemoine) Rehd. not very far from S. pinnatifolia. In the same year, Dr. K. Sax fertilized S. pinnatifolia 1935] ANDERSON AND REHDER, NEW HYBRIDS 363 with pollen of S. oblata var. Giraldii and plants were raised from this pollination; these plants have not yet flowered, but in their vegetative characters agree with the plant described above. The pollen of S. pin- natifolia is defective, at least that of our plant, and self-pollinated flow- ers produce no seeds. The hybrid is clearly intermediate between these two species, readily distinguished from both species by the partly pinna- tifid and partly entire leaves. In the partly entire and partly pinnatifid leaves the hybrid resembles S. persica L. var. laciniata, which can be distinguished by the broadly decurrent often obtusish lobes of the leaves and by the narrower and generally smaller entire leaves, by the smaller panicles usually in several to many pairs along the branches, the absence of the terminal leaf-bud, and by the anthers not reaching the mouth. A. R. HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 364 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XvI HYPODERMELLA HIRATSUKAE, A NEW SPECIES OF HYPODERMATACEAE FROM JAPAN? GRANT D. DARKER With plate 154 FIve species of Hypodermataceae have been reported on conifers in Japan by Shirai and Hara (1927). Only one species, Lophodermium pinastri (Schrad. ex Fr.) Chev., was listed as occurring on pines. The present paper describes a new Japanese species of Hypodermataceae of unusual interest which was encountered during a hasty examination of the Hypodermataceae in the Mycological and Pathological Herbarium of the United States Department of Agriculture in Washington, D. C. Grateful acknowledgment is made to Dr. C. L. Shear and Mr. John A. Stevenson for the privileges extended to the writer. Hypodermella Hiratsukae, sp. nov. Hysterotheciis in uno ordine epiphyllis oblongis ellipticisque atroni- tidis, 0.54-1.30 0.26—-0.34 mm., longitudinali incisura aperientibus; hysterotheciis in transversali sectione in medio subcuticularibus sed ad marginem subepidermalibus, 0.16—0.22 mm. profundis; basilari plecten- chymate achrod 20-35 crasso; tegente strato atri pseudoparenchy- matis 28-34 «1 crasso; hymenio 100-110 y: crasso. Ascis latis fusi- formibusque octosporis 87-102 « 18-24 yy). Paraphysibus 100-110 x 1 » simplicibus filiformibus muco involutis. Ascosporis clavatis fusi- formibusque ad basin attenuatis hyalinis 35-56 « 3.5-5.0 , muco 8 J crasso involutis. In foliis Pini pumilae Regel, in monte Kuro-dake, provinciae Ishikari Japoniae, mense Augusto, 1927, Naohide Hiratsuka legit. Hysterothecia in a more or jess ree row, epiphyllous, oblong and elliptical, shining black, 0.54-1.30 « 0.26-0.34 mm., opening by a longitudinal fissure; hysterothecia in cross section sibeuticwlae in the middle but subepidermal at margins, 0.16—0.22 mm. deep (closed) ; basal layer colorless, plectenchymatous, 20-35 «1 thick; covering layer of dark pseudoparenchyma 28-34 uy thick; hymenium 100-110 thick. Asci broad, somewhat fusiform, truncate to rounded at maturity at tip, 8-spored, 87-102 * 18-24 yp. Paraphyses 100-110 X 1 un, simple, 1CONTRIBUTION FROM THE ne LABORATORIES AND THE FARLOW HERBARIUM, Harvarp University, No. 1 PLATE 154 Jour. ARNoLD Ars. VoL. XVI. ») ) om J yas 4, s sl wy . eee “ev A AM *« A +¥ e a* ) HypopERMELLA HIRATSUKAE Darker. 1935] DARKER, HYPODERMELLA HIRATSUKAE 365 filiform, surrounded by a delicate gelatinous sheath. Ascospores clavate fusiform, tapering towards the base, hyaline, 36-56 * 3.5—5.0 u, sur- rounded by a conspicuous gelatinous sheath up to 8 i thick. On needles of Pinus pumila Regel, Mt. Kuro-dake, Province Ishikari, Japan, August 12, 1927, collected by Naoside Hiratsuka. Hypodermella Hiratsukae is of special interest because of certain morphological resemblances to Hypodermella Laricis v. Tub., the type species of the genus. As previously pointed out by the writer in 1932, the species of Hypodermella fall readily into four easily recognized groups named after the first described species in each as follows: (a) H. Laricis group, (b) H. ampla group, (c) H. nervisequia group, and (d) H. sulcigena group. Of nineteen species recognized in the genus, H. Hirat- sukae approaches most closely H. Laricis, hitherto the only species in that group. The linear arrangement of the hysterothecia, the broad clavate asci and ascospores and the absence of a slit band along which the hysterothecium ruptures are common to both species. The position of the fruiting body of H. Laricis in the host tissue is difficult to deter- mine even in microtome sections but is considered to be subcuticular by the writer. In the new species the hysterothecia are subcuticular in the centre and subepidermal at the margins as in Lophodermium pinastri. A prominently developed slit band, however, is characteristic of L. pinastri. Pycnidia with spores of the microconidial or spermatial type which are conspicuous and abundantly formed in the life cycle of H. Laricis are unknown in the case of H. Hiratsukae although in the material examined there are present certain small blister-like areas between the hysterothecia which may represent the remains of pycnidia. BIBLIOGRAPHY Surat, M. & K. Hara (1927). A list of Japanese fungi hitherto known. (3rd edition. pp. 448. DarRKER, GRANT eras (1932). The Hypodermataceae of conifers. (Contr. Arnold Arb, 1: 1-131). DESCRIPTION OF PLATE 154 Fig. i. Hypodermella Hiratsukae, sp. nov., on Pinus pumila Regel. Portion of needle with hysterothecia (x 17). 00). WN > DW QO io) own E>] eo) = om wn _~ x Hysterothecium in cross-sectional view (X 270 approx.). FarLow HERBARIUM, HARVARD UNIVERSITY, CAMBRIDGE, MASS. JOURNAL OF THE ARNOLD ARBORETUM VoLuME XVI. OCTOBER, 1935 NUMBER 4 THE HOSTS, LIFE HISTORY AND CONTROL OF GYMNOSPORANGIUM CLAVIPES C. AND P. Ivan H. CroweELu With plates 155-160 TABLE OF CONTENTS ile: TEN TERODUGTION © copcgestcc eek oS. c/o sale: oD eos aise o) otalesapeyiese avarov eve 368 Tt: Tue Pomaceous Hosts oF GyMNOSPORANGIUM CLAVIPES C. AND P., THerr TAXONOMIC PosITION AND THEIR GEO- GRAPHIC RUANGE! cictaieie oie ie oreie oa Oa aNe ae oi ole ater anemia ete ateratie 369 III. PRELIMINARY STUDIES ON THE PERIOD OF SUSCEPTIBILITY OF Pomaceous Hosts oF GYMNOSPORANGIUM CLAVIPES ..... 78 IV. Tue JuNrPERUsS Hosts oF GYMNOSPORANGIUM CLAVIPES C. AND P., THEIR TAXONOMIC PoSsITION AND THEIR GEO- GRAB HTCHRCAIN GE oxic cnet etet as on eee oi arisretaicl. oc oacaraee aie 379 Vie SYMPTOMATOLOGY OF THE DISEASES CAUSED BY GYMNOSPO- RANGIUM CLAVIPES 2.4 $0545 00s esa vaaiee Gan oee se edes 380 1. On Pomaceous Hosts. Morphological Symptomatology ............-. 380 b. Histological Symptomatology ................ 384 2. On Juniperus Hosts. a. Morphological Symptomatology .............. 386 b. Histological Symptomatology ................ 387 VI. Lire History STUDIES OF GYMNOSPORANGIUM CLAVIPES .... 388 VII. Controt MeEAsuRES APPLICABLE TO GYMNOSPORANGIUM CLAVIPES ON POMACEOUS AND ON JUNIPERUS HostTs ...... Willie . RECOMMENDATIONS. «-cciccvsretererteeetnretereieieteieteiala «(a © se ese @ateve 401 IX. SIMA RVG% face cee cin eee ee ee erat cles crsceiees « 405 IBTBETOGRARELY: dere cvevarchevecsve Ge GAOT OER eae Oia, aia Sp iovlas iatsgiiny arrestee 406 EXPLANATION OF THE ELATES. ncaa no ia ecto eta 408 368 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XvI I. INTRODUCTION GYMNOSPORANGIUM CLAVIPES C. and P., commonly referred to as the quince rust fungus, was one of the earliest species of the genus to be defined in North America. Schweinitz (1832) described it from the aecial phase occurring on Crataegus sp. as Caeoma (Peridermium) germinale. Later, Cooke and Peck (1871) described the telial phase of a rust on Juniperus virginiana L. as Podisoma (Gymnosporangium) clavipes. It was not until 1886, however, that Thaxter (1887) demon- strated by means of controlled cultures that these were but two phases of the same organism. Throughout the literature both specific names have been used. Some authors have followed Kern (1911) in calling the species G. germinale (Schw.) Kern. But Sydow (1915), Arthur (1934) and other mycologists adhere to the International Rules of Nomenclature and so designate the species G. clavipes C. & P. The prevalence and destructiveness of the diseases caused by G. clavipes on ornamental pomaceous hosts and on many orchard varieties of apples, as well as on red cedars, have occasioned numerous inquiries regarding the pathogenicity and control of this rust. The information pertainable to these matters has been so meagre that comprehensive studies on the causal organism and the diseases produced by it were begun four years ago. The results obtained are presented in this paper. Certain phases of the investigations not yet completed are being continued. The main lines of my investigations are as follows: 1. A determination by means of cultures and field observations of the pomaceous hosts of G. clavipes together with a discussion of their taxo- nomic position and geographic range. 2. Similar determinations and discussions of the Juniperus hosts of G. clavipes. 3. The symptomatology of the diseases caused by G. clavipes. 4. A thorough examination of the life history of G. clavipes on poma- ceous and Juniperus hosts. 5. The practicability of fungicidal and of eradicative control measures of the diseases caused by G. clavipes. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 369 Il. THE POMACEOUS HOSTS OF GYMNOSPORANGIUM CLAVIPES C. AND P., THEIR TAXONOMIC POSITIONS AND THEIR GEOGRAPHIC RANGE Contributions to our knowledge of the pomaceous hosts of G. clavipes as determined by artificial cultures are due chiefly to the work of Thax- ter (1887), Arthur (1910, 1912), Thomas and Mills (1929) and Miller (1932). Numerous other investigators, by their field observations, have added several species to the list of pomaceous hosts. The total number previously reported is about thirty-six species in seven genera distributed as follows: Amelanchier (8), Aronia (3), Chaenomeles (1), Crataegus (20), Cydonia (1), Malus (2), Pyrus (1). In my own studies on the pomaceous hosts, 701 species and varieties in 15 genera were inoculated, following essentially the same procedure described in a previous article (1934). These genera and the number of species of each inoculated are — Amelanchier (18), Amelosorbus (1), Aronia (4), Chaenomeles (2), Crataegomespilus (1), Crataegus (588), Cydonia (1), Malus (44), Photinia (1), Pyrus (19), Sorbaronia (1), Sorbopyrus (1), and Sorbus (17). One species of Comptonia and two of Myrica were also inoculated. From the results obtained the inoculated plants were placed in two groups. Those plants which de- veloped no evidence of infection were classed as immune; while those plants on which infection was evident were classed as susceptible. The number of hosts determined by artificial inoculations was augmented by field observations in the Arnold Arboretum, on private estates about Boston and from the reports of former investigators. The results showed that more than 480 species and varieties of poma- ceous plants (including 48 varieties of orchard apples) scattered among eleven genera are susceptible. These hosts are presented in table 1. TABLE I. POMACEOUS HOSTS OF GYMNOSPORANGIUM CLAVIPES CUAND) Pa Amelanchier alnifolia Nutt., A. Bartramiana Roem. (A. oligocarpa [| Michx.] Roem.), A. Bartramiana X laevis, A. Bartramiana x oblongi- folia, A. canadensis Med., A. canadensis nana, A. erecta Blanch., A. florida Lindl., A. intermedia Spach., A. laevis Wieg., A. laevis * humilis, A, oblongifolia (Torr. and Gray) Roem., A. sanguinea DC., A. spicata K. Koch, A. stolonifera Wiec. Amelosorbus Jacku Rehd. the taxonomy of the genus Crataegus, Palmer (1925) was used. Mr. Palmer has kindly checked the hosts of G. vipe s in the genus Crataegus against his revised but unpublished catalogue of Crataegi. For the Se of the other genera, Rehder’s Hane (1927) was followeg as far as possi 370 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI Aronia arbutifolia (L.) Ell., A. floribunda Spach (A. arbuttfolia atro- purpurea |Britt.| B. L. Robinson), A. melanocarpa Ell., A. melanocarpa elata Rehder, A. monstrosa Zabel. Chaenomeles japonica Lindl., C. lagenaria Koidz., C. lagenaria marmo- rata, C. lagenaria foliis rubris, C. lagenaria sanguinea semiplena. Crataegomes pilus grandiflora Bean. Crataegus ANOMALAE: C. affinis Sarg., C. asperifolia Sarg., C. Brockwayae Sarg., C. Coleae Sarg., C. cyclophylla Sarg., C. Dunbari Sarg., C. Eggle- stonii Sarg., C. e: ‘ata Sarg., C. honesta Sarg., C. Ideae Sarg., C. impro- visa Sarg., C. misella Sarg., C. pinguis Sarg., C. putata Sarg., C. repul- sans Sarg., C. Saundersiana Sarg., C. scabrida Sarg., C. shirleyensis Sarg., C. urbana Sarg. CoccrnEAE: C. acclivis Sarg., C. arcuata Ashe, C. assurgens Sarg., C. aulica Sarg., C. caesa Ashe, C. chippewaensis Sarg., C. confinis Sarg., C. conspecta Sarg., C. cristata Ashe, C. delecta Sarg., C. densiflora Sarg., C. elongata Sarg., C. fluviatilis Sarg., C. fretalis Sarg., C. Hillii Sarg., C. Holmesiana Ashe, C. Holmesiana tardipes Sarg., C. lenta Ashe, C. Macounii Sarg., C. miranda Sarg., C. neolondinensis Sarg., C. pedicellata Sarg., C. perrara Sarg., C. polita Sarg., C. Pringlei Sarg., C. pura Sarg., C. sejuncta Sarg., C. sertata Sarg., C. Thayeri Sarg., C. vivida Sarg. Crus-GALLI: C. arborea Beadle, C. arduennae Sarg., C. armata Sarg., C. attenuata Sarg., C. barrettiana Sarg., C. Bartramiana Sarg., C. bellica Sarg., C. calophylla Sarg., C. Canbyi Sarg., C. cerasina Sarg., C. crus-galli L., C. crus-galli arbutifolia Hort. ex Nicholson, C. crus-gatli exigua Sarg., C. crus-galli pyracanthifolia Ait., C. crus-galli rubens Sarg., C. crus-galli splendens Ait., C. effulgens Sarg., C. erecta Sarg., C. Far- wellii Sarg., C. Fontanesiana (Spach) Steudel, C. geneseensis Sarg., C. insignis Sarg., C. Lavallei Herincq, C. lawrencensis Sarg., C. leptophylla Sarg., C. livoniana Sarg., C. macra Beadle, C. pachyphylla Sarg., C. Palmeri Sarg., C. parciflora Sarg., C. Pennypackeri Sarg., C. peoriensis Sarg., C. persimilis Sarg., C. phlebodia Sarg., C. Reverchonii Sarg., C. rivalis Sarg., C. robusta Sarg., C. rubrifolia Sarg., C. rudis Sarg., C. trium phalis Sarg. DivataTaE: C. coccinoides Ashe, C. dilatata Sarg. DouctasIANaE: C. colorado Ashe, C. columbiana Howell, C. Douglasii Lindl., C. Douglasii {. badia Sarg., C. Douglasii Suksdor fu Sarg., C. erythropoda Ashe, C. Piperi Britt.,.C. rivularis Nutt. apud Torr. & Gray. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 371 FLavAE: C. colonica Beadle, C. dispar Beadle, C. flava Ait., C. ignava Beadle. INTRICATAE: C. Delosti Sarg., C. flavida Sarg., C. modesta Sarg., C. nemoralis Sarg., C. neobushii Sarg., C. scabra Sarg., C. straminea Beadle. MACRACANTHAE: C. admiranda Sarg., C. aquilonaris Sarg., C. ardua Sarg., C. baccata Sarg., C. Beckiana Sarg., C. bristolensis Sarg., C. calpodendron (Ehrh.) Med., C. Calvinii Sarg., C. chadfordiana Sarg., C. corporea Sarg., C. Deweyana Sarg., C. divida Sarg., C. Emersoniana Sarg., C. ferentaria Sarg., C. ferta Sarg., C. fertilis Sarg., C. finitima Sarg., C. flagrans Sarg., C. fulgens Sarg., C. gemmosa Sarg., C. glabrata Sarg., C. Handyae Sarg., C. hystricina Ashe, C. laxiflora Sarg., C. macra- cantha Lodd., C. membranacea Sarg., C. microsperma Sarg., C. missouri- ensis Ashe, C. neofluvialis Ashe, C. occidentalis Britt., C. ogdenburgensis Sarg., C. peramoena Sarg., C. prunifolia (Marsh.) Pers., C. radiosa Sarg., C. rhombifolia Sarg., C. Robinsonii Sarg., C. Searsti Sarg., C. spatiosa Sarg., C. spinulosa Sarg., C. structilis Ashe, C. succulenta Schrader, C. tomentosa L., C. truculenta Sarg., C. vegeta Sarg. MricrocarpaeE: C. Phaenopyrum (L. f.) Med., C. spathulata Michx. Mottes: C. anomala Sarg., C. arnoldiana Sarg., C. champlainensis Sarg., C. contortifolia Sarg., C. digna Sarg., C. dispessa Ashe, C. Ell- wangeriana Sarg., C. exclusa Sarg., C. Fulleriana Sarg., C. induta Sarg., C. invisa Sarg., C. lanuginosa Sarg., C. lasiantha Sarg., C. lauta Sarg., C. limaria Sarg., C. mollis (Torr. & Gray) Scheele, C. nutans Sarg., C. pennsylvanica Ashe, C. peregrina Sarg., C. Robesoniana Sarg., C. sera Sarg., C. submollis Sarg.; C. Tatnalliana Sarg., C. urbica Sarg. OxYACANTHAE: C. altaica Lange, C. hiemalis Lange, C. Heldreichii Boiss., C. intermedia, C. Maximowiczti Schneider, C. monogyna Jacq., C. monogyna albo-plena Schneider, C. monogyna inermis Rehd., C. monogyna laciniata Loud., C. monogyna pteridifolia Rehd., C. mono- gyna spectabilis, C. monogyna stricta Loud., C. monogyna versicolor, C. Oxyacantha L., C. Oxyacantha Gireoudi Bean, C. Oxyacantha alba West., C. Oxyacantha rubra Schneider, C. pinnatifida Bunge, C. sorbi- folia Lange, C. Wilsonii Sarg. PRUINOSAE: C. arcana Beadle, C. aridula Sarg., C. aspera Sarg., C. austera Sarg., C. beata Sarg., C. bellula Sarg., C. brachypoda Sarg., C. bracteata Sarg., C. cestrica Sarg., C. cognata Sarg., C. comparata Sarg., C. confragosa Sarg., C. delawarensis Sarg., C. disjuncta Sarg., C. dissona Sarg., C. divisifolia Sarg., C. Ferrissti Ashe, C. formosa Sarg., 372 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI C. fusca Sarg., C. georgiana Sarg., C. glareosa Ashe, C. horridula Sarg., C. incisa Sarg., C. Jesupii Sarg., C. Kellermanii Sarg., C. latifrons Sarg., C. levis Sarg., C. macrocalyx Sarg., C. numerosa Sarg., C. patrum Sarg., C. pequotorum Sarg., C. perampla Sarg., C. perjucunda Sarg., C. phila- delphica Sarg., C. platycarpa Sarg., C. pruinosa (Wendl.) K. Koch, C. pulchra Sarg., C. relicta Sarg., C. remota Sarg., C. rubicundula Sarg., C. sicca Sarg., C. tribulosa Sarg. PRUNIFOLIAE: C. decorata Sarg. PunctTATAE: C. amnicola Beadle, C. angustata Sarg., C. barbara Sarg., C. Brownietta Sarg., C. calvescens Sarg., C. celsa Sarg., C. com- pacta Sarg., C. desueta Sarg., C. florifera Sarg., C. glabrifolia Sarg., C. incaedua Sarg., C. Lettermanii Sarg., C. notabilis Sarg., C. pausiaca Ashe, C: porrecta Ashe, C. praestans Sarg., C. pratensis Sarg., C. punc- tata Jacq., C. punctata aurea Ait., C. punctata canescens Britt., C. rigens Beadle, C. suborbiculata Sarg., C. succincta Sarg., C. tenax Ashe, C. verruculosa Sarg. RotuNDIFOLIAE: C. Bicknellii Eggl., C. Blanchardii Sarg., C. Brainerdii Sarg., C. Brunetiana Sarg., C. chrysocarpa Ashe, C. coccinata Sarg., C. cupilifera Sarg., C. Dodgei Ashe, C. inaudita Sarg., C. Jonesae Sarg., C. Kennedyi Sarg., C. kingstonensis Sarg., C. maligna Sarg., C. Margaretta Ashe, C. Margaretta xanthocarpa Sarg., C. Maribella Sarg., C. Oakesiana Eggl., C. praecoqua Sarg., C. Proctoriana Sarg., C. propria Sarg., C. rotundata Sarg., C. rotundifolia Moench, C. rotundifolia pubera Sarg., C. rotundifolia {. rubescens Sarg., C. varians Sarg., C. Websteri Sarg. SILVICOLAE: C. allecta Sarg., C. Barryana Sarg., C. blairensis Sarg., C. compta Sarg., C. delectata Sarg., C. diffusa Sarg., C. dissona Sarg., C. effera Sarg., C. filipes Ashe, C. foliata Sarg., C. Fretzti Sarg., C. tra- cunda Beadle, C. iterata Sarg., C. Livingstoniana Sarg., C. luxurtosa Sarg., C. medioxima Sarg., C. opulens Sarg., C. promissa Sarg., C. prona Ashe, C. radina Sarg., C. recordabilis Sarg., C. ruricola Sarg., C. stoloni- fera Sarg., C. strigosa Sarg., C. tortuosa Sarg., C. xanthocar pa Sarg. TENUIFOLIAE: C. acuminata Sarg., C. acutiloba Sarg., C. alnorum Sarg., C. apiomorpha Sarg., C. ascendens Sarg., C. asperata Sarg., C. bella Sarg., C. benigna Sarg., C. blandita Sarg., C. Boothiana Sarg., C. colorata Sarg., C. conferta Sarg., C. crudelis Sarg., C. cyanophylla Sarg., C. Damei Sarg., C. delucida Sarg., C. demissa Sarg., C. Edsoni Sarg., C. Eganii Ashe, C. firma Sarg., C. flabellata (Bosc.) K. Koch, 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 373 C. florea Sarg., C. Forbesae Sarg., C. fucosa Sarg., C. gentalis Sarg., C. glaucophylla Sarg.. C. gracilipes Sarg., C. Gruberi Ashe, C. Habereri Sarg., C. Hadleyana Sarg., C. heidelburgensis Sarg., C. insolita Sarg., C. leptopoda Sarg., C. lucorum Sarg., C. luminosa Sarg., C. macrosperma Ashe, C. marcida Ashe, C. matura Sarg., C. media Sarg., C. merita Sarg., C. miniata Ashe, C. modica Sarg., C. monstrata Sarg., C. Napaea Sarg., C. nescia Sarg., C. otiosa Ashe, C. Paddockeae Sarg., C. Paineana Sarg., C. pallidula Sarg., C. parviflora Sarg., C. pastorum Sarg., C. paucispina Sarg., C. pentandra Sarg., C. perlevis Ashe, C. populnea Ashe, C. pumila Sarg., C. retrusa Ashe, C. roanensis Ashe, C. rubicunda Sarg., C. rubro- carnea Sarg., C. rufipes Ashe, C. sarniensis Sarg., C. serena Sarg., C. sex- tilis Sarg., C. Slavini Sarg., C. Streeterae Sarg., C. suavis Sarg., C. taetrica Sarg., C. tarda Sarg., C. tenella Sarg., C. tenera Sarg., C. tenui- loba Sarg., C. trachyphylla Sarg., C. viridimontana Sarg., C. vittata Ashe. TRIFLORAE: C. conjungens Sarg. UNIFLORAE: C, uniflora Moench. ViripEs: C. abbreviata Sarg., C. blanda Sarg., C. penita Beadle, C. velutina Sarg., C. viridis L., C. vulsa Beadle. Cydonia oblonga Mill. Malus angustifolia Michx., M. floribunda Sieb., M. ioensis plena Rehd., M. pumila Mill., M. spectabilis Borkh., M. sylvestris Mill. Photinia villosa DC. Pyrus communis L., P. sinensis Lindl. Sorbus americana Marsh, S. dumosa Greene. A complete enumeration of all species and varieties of inoculated plants on which the results were negative is as follows: Amelanchier amabilis Wieg., A. asiatica Endl., A. grandiflora Rehd., A. humilis Wieg., A. humilis & sanguinea, A. ovalis Med., A. sera Ashe. Malus arnoldiana Sarg., M. asiatica Nakai, M. atrosanguinea Schneid., M. baccata Borkh., M. baccata costata Hort., M. baccata gracilis Rehd., M. baccata Jackii Rehd., M. baccata mandshurica Schneid., M. baccata microcarpa Regel., M. baccata pendula Hort., M. brevipes Rehd., M. flexilis Hort., M. florentina Schneid., M. Halliana Parkmanii Rehd., M. Halliana spontanea Rehd., M. Hartwigii Koehne, M. hona- nensis Rehd., M. hupehensis Rehd., M. kansuensis Schneid., M. spec. (Pyrus Malus laurifolia Gibbs), M. spec. (Pyrus Lemoinei Hort.), M. magdeburgensis Schoch, M. micromalus Mak., M. orthocarpa Lavall., M. Prattij Schneid., M. pumila Niedzwetzkyana Schneid., M. purpurea Rehd., M. purpurea aldenhamensis Rehd., M. purpurea Eleyi Rehd., 374 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI M. robusta persicifolia Rehd., M. Sargenti Rehd., M. Scheideckeri Zabel, M. Scheideckeri “Excellenz Thiel,’ M. sikkimensis Koehne, M. spectabilis Riversii Nash, M. sublobata Rehd., M. toringoides Hughes, M. trilobata Schneid., M. Tschonoskii Schneid., M. yunnanensis Schneid., M. yunnanensis Veitchii Rehd., M. zumi Rehd., M. zumi calo- carpa Rehd. Pyrus amygdaliformis Vill., P. Balansae Decne., P. betulifolia Bge., P. Bretschneideri Rehd., P. denticulata Hort. Angl. ex Dum.-Cours., P. elaeagrifolia Pall., P. Korshinskyi Litv., P. Lindleyi Rehd., P. longipes Coss. & Dur., P. Michauxii Bosc., P. nivalis Jacq., P. pashia Buch.-Ham., P. phaeocarpa Rehd., P. salvifolia DC., P. serotina Rehd., P. serrulata Rehd., P. syriaca Boiss., P. ussuriensis Maxim. Sorbaronia alpina superaria, S, spec. Sorbopyrus auricularis bulbiformis Schneid. Sorbus Aria Crantz, S. arnoldiana Rehd., S. Aucuparia L., S. com- mixta Hedl., S. discolor Hedl., S. Aybrida L., S. intermedia Pers., S. japonica calocarpa Rehd., S. Matsumurana Koehne, S. Meini- chii Hedl., S. pohuashanensis Hedl., S. rotundifolia Hedl., S. subpinnata Hedl., S. thuringiaca Fritsch (S. decurrens Hedl.). Other plants inoculated were Comptonia asplenifolia Ait., Myrica carolinensis Mill. and M. Gale L. In addition to the pomaceous hosts just reported, several varieties of orchard apples have been found by former investigators to be susceptible. A compilation of these is presented in table 2. Mills (1929) gives the following account of the occurrence of G. clavipes on orchard apples in New York: “Counts in 14 orchards in 4 counties showed fruit infection on Delicious (3 counts) 60 per cent; Fameuse (1 count) 21 per cent; Hubbardston (1 count) 28 per cent; McIntosh (15 counts) 18 per cent average; Winesap (2 counts) 74 per cent; Yellow Transparent (1 count) 84 per cent. Specimens were from 6 or 7 counties. Not found on foliage or twigs.” This account and the fact that many other varieties of orchard apples are susceptible serve to stress the economic importance of G. clavipes to orchardists. In several orchards visited in Massachu- setts the disease caused by G. clavipes was found to be particularly abun- dant on the Delicious variety. As high as 90 per cent of the fruits were attacked. This disease, one of the most severe on the Delicious apple, is of much concern to orchardists because this variety is being grown in greater quantities to meet the steadily increasing demands for it on the market. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 375 TABLE II. THE RELATIVE SUSCEPTIBILITY OF ORCHARD APPLES TO G. CLAVIPES Variety Susceptible Immune Alexander N.Y. Baldwin Ind., Me., N. Y., N.S. Bechtel’s Crab Mass. Bellflower Me. Ben Davis N. Y Ind. Bishop N.S. Black Twig Va., W. Va Cortland Me., N. Y Crimson Beauty eS Delicious Ind., Me., Mass., N. Y., Tenn., Va., W. Va Duchess Me., N. Y Early Red McIntosh Me. Fameuse Yes So Family N.S Gideon Ind. Golden Delicious Me. Gravenstein Me., N. Y., N.S Grimes Ind., Tenn Hubbardston N.Y. Jonathan Ind., Me., N. Y. Tenn. King David Ind Maiden Blush Ind. McIntosh Me., N. Y., N.S Northern Spy 1 ie Northwestern Greening Ind., N. Y. Red Delicious Me., N. Y., Tenn. Red Winesap Tenn. R. I. Greening N. Y.,N.S Ribston S Rome Md., N. Y., Tenn., Vt., Va Ind. Roxbury 1 ee Russett Nia Yale Stark N.S: Starkey Me., N.S Starking Me Stayman Ind., N. Y., Tenn., Va., W. Va. 376 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI TABLE II. (Continued) Variety Susceptible Immune Stayman Winesap Tenn. Sweet Winesap N. Y. Tolman Me., N. Y. Tompkin’s King N.S. Twenty Ounce Me., Mass., N. Y. Wagener Me., N.S. Wealthy Ind., Mass., Me., N. Y., Tenn. Winesap Ind., Me., N. Y., Tenn., Vt., Va., W. Va. Winter Banana Ind., Me., N. Y. Wolf River . Yellow Bellflower N.S. Yellow Transparent Md., N. Y., Tenn., N. S. Va. York —- Va. The species and genera of pomaceous hosts recorded above show no simple relationship or correlation by which one can formulate a rule to encompass them and set them apart from related non-susceptible plants. Nevertheless, they possess several distinctive features chiefly with respect to their taxonomy and their geographic range. Pomaceous hosts of G. clavipes are found in eleven genera. So far as I am able to learn no other species of Gymnosporangium is known to have hosts in so large a range of genera. All of the host genera of G. clavipes are closely related, however, and confined within the family Rosaceae. The native geo- graphic range of these hosts is more extensive than for those of any other species of Gymnosporangium known to me. Pomaceous hosts of G. clavipes are found throughout the whole of the temperate portion of the northern hemisphere. One of the most outstanding of the introduced foreign host species to become parasitised is Cydonia oblonga, the quince. This plant is native over the greater portion of Asia but has been introduced into North America over a portion only of the range of the rust. In the genus Chaenomeles, also native to Asia, two of the three species listed in Rehder’s Manual are attacked. Several varieties and forms of the Japa- nese quince (CAaenomeles spp.) are also parasitised. One European and one Asiatic species of the genus Pyrus are attacked, while all other species (Eurasian) have so far proved to be immune. Some native and some foreign species of Malus as. well as many orchard varieties of apples are 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 377 hosts to G. clavipes. In the genus Sorbus the two North American species are susceptible while all of the foreign ones inoculated remained immune. The hybrid genus Amelosorbus has but one species susceptible to G. clavipes. All of the species of the genus Aronia, native to North America, are hosts. With the exception of three or four species the genus Amelanchier, as represented in North America, is attacked by G. clavipes. A single species of the genus Photinia, the Eurasian species P. villosa, is liable to infection. 7 sPn74 £ as 4 5 ilvicolas 4 ai ™Mlatet m= Rotundi £011 2 ¢ cme 3 =Vacaracantha ¢ qe -Bracteatae M4 an '™ + 74 TInt #7 ai % suscept. 100 100 100 100 96 94 88 84 83 Noe Pe wtea «21:19 «8 124 7832 3150 5255 59 9 5 5 6514 2 212 1 | o ae rn cep) ao fo] | a =I Q “ a a £3] =) ts] o » iw) oo Noe hostse 21 #19 8 123 73 28 2625 3926 40 6 22 40 7 11 5 0 Approxe no e ae an os 27 11 161 120 51 151 59 122 62143 37 4 412579 35186 2 FI 1. Data on Species and Varieties in Groups of the Genus iene Susceptible to Ae "s Cae acti clavipes C. and P, The susceptibility of each group of the genus Crataegus is presented as a graph in figure 1. The data for this graph were obtained by determin- ing the percentage relationship between the number of species and vari- eties tested and the number that proved to be susceptible. While little significance can be attached to the results obtained from groups with a small number of species and varieties, nevertheless, reliable deductions can be made from those with large numbers as well as from the genus as a whole. Susceptible species were found in all groups except the Bracteatae. Unfortunately this group was represented in the Arnold Arboretum by but a single tree. In these investigations, 79 per cent of the species that produced fruit when the tests were made proved to be susceptible. Many species and varieties have not yet been tested, either because they were not available or because they did not produce fruits in the years of my 378 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI investigations. Therefore, the host list is far from complete. All of the foreign species and varieties available for testing (placed in the group Oxyacanthae) proved to be susceptible. Of the native Crataegi, the percentage of species susceptible in the several groups varied widely. In some groups, for example, the Anomalae and Douglasianae, all of the species and varieties tested were susceptible while in other groups, for example, the Molles, Virides and Flavae, the percentage of susceptible members was less. It is interesting to note that the relative susceptibility of species and varieties and of the groups of the genus is not the same for hosts of G. clavipes as MacLachlan (1935) found for hosts of G. globosum, though hosts for both of these fungi were in most cases deter- mined from the same individual trees. III. PRELIMINARY STUDIES ON THE PERIOD OF SUSCEPTIBILITY OF POMACEOUS HOSTS OF GYMNOSPORANGIUM CLAVIPES It has been shown by several investigators that the leaves of poma- ceous hosts of certain species of Gymnos porangium are susceptible during a limited period in their youth only. This interim is known as the peri of susceptibility. Thomas (1933) showed that leaves of Crataegus spp. were susceptible in their youth only to the attack of G. clavipes. My pre- liminary investigations, with respect to this phenomenon on fruits of pomaceous hosts of G. clavipes, also indicate that their period of suscep- tibility is brief. In this connection inoculations were made every two days for a period of four weeks on different fruit clusters of two species of Crataegus, C. tomentosa and C. fertilis. Experimentation began when the flower buds of each species were opening. On C. tomentosa the ovaries, calyces, petals, pedicels, peduncles and twigs were attacked before the flowers opened. All but the ovaries and young fruit became immune within ten days. The fruits became decreasingly susceptible and by the time the petals fell they could no longer be infected. On C. fertilis, the young fruits only became infected. The flowers were immune up to the time the buds were opening and the petals began to expand; they then entered a brief period of susceptibility extending until the time when the petals began to fall, after which they again became immune. A measure of susceptibility of the hosts of G. clavipes is here suggested. Those hosts that are susceptible for a longer period may be considered to be more susceptible than those that can be infected for a shorter period. No consistent difference was observed with respect to the abundance of aecia produced on susceptible as compared with resistant hosts. The possibility of any relationship between immunity and fertilization was not investigated. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 379 IV. THE JUNIPERUS HOSTS OF GYMNOSPORANGIUM CLAVIPES C. AND P., THEIR TAXONOMIC POSITIONS AND THEIR GEOGRAPHIC RANGE Cultural studies for the purpose of determining the Juniperus hosts of G. clavipes have been very limited indeed. Arthur (1912) cultured G. clavipes on J. communis depressa Pursh (J. sibirica Burgsd.) and Dodge (1918) obtained heavy infection from sowing aeciospores of G. clavipes on J. virginiana L. Several other investigators have substantiated these reports and added other hosts from their field observations. In my own work I have repeatedly cultured this rust on the red cedar, J. virginiana L. The technique for this work followed essentially the same procedure as for culture work on pomaceous plants. A complete enumeration (to 1934) of the Juniperus hosts of G. clavipes together with data on their taxonomic position, as given by Rehder (1927), and their geographic range are presented in table 3. I have examined the rust on all of these species and varieties. TABLE III. JUNIPERUS HOSTS OF G. CLAVIPES Juniperus hosts gang ee J. communis L. Oxycedrus North Amer., Eurasia var. depressa Pursh . ve - var. hibernica Gord. . Euro var. montana Ait. i North "Amer, Eurasia J. horizontalis Moench Sabina J. Sabina L urope Is scopulorum Sarg. “ Western North Amer. J. virginiana L. se Eastern North Amer. In 1933 each individual cedar in the collections at the Arnold Arbore- tum of Harvard University was examined for infection and the species and varieties were recorded in immune and susceptible groups. The species and varieties of Juniperus on which no infection was observed are as follows: Juniperus chinensis L., J. chinensis globosa Hornibrook, J. chinensis japonica Lav., J. chinensis mas Gord., J. chinensis pendula Franch., J. chinensis Pfitzeriana Spaeth, J. chinensis plumosa Hornibrook, J. chi- nensis plumosa aurea Hornibrook, J. chinensis pyramidalis Beiss., J. chinensis Sargentii Rehd., J. chinensis Watereri Hort., J. communis Ashfordii Hort., J. communis aurea Nichols., J. communis aureo-spica Rehd., J. communis compressa Carr., J. communis cracovica Hort., J. communis oblongo-pendula Sudw., J. communis oblonga Loud., J. com- 380 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI munis pyramidalis Hort., J. communis suecica Ait., J. conferta Parl., J. formosana Hayata, J. horizontalts alpina Rehd., J. horizontalis Doug- lasti Rehd., J. Aorizontalis glomerata Rehd., J. horizontalis plumosa Rehd., J. Aorizontalis variegata Hort., J. procumbens Sieb., J. rigida Sieb. & Zucc., J. Sabina cupressifolia Ait., J. Sabina pyramidalis Hort., J. Sabina tamariscifolia Ait., J. Sabina variegata Carr., J. scopulorum hori- zontalis D. Hill, J. scopwlorum viridifolia Hort., J. squamata Buch.- Ham., J. squamata Fargesii Rehd. & Wils., J. squamata Meyeri Rehd., J. squamata Wilsonit Rehd., J. virginiana aurea Hort., J. virginiana Burkii, J. virginiana Canaertii Senecl., J. virginiana Chamberlaynii Carr., J. virginiana cinerascens Hort., J. virginiana elegantissima Hochst., J. vir- giniana fastigiata Hort., J. virginiana filifera D. Hill, J. virginiana glauca Carr., J. virginiana globosa Beiss., J. virginiana Hillit Hort., J. virginiana Kosteri Beiss., J. virginiana pendula Carr., J. virginiana polymorpha Hort., J. virginiana pyramidalis Carr., J. virginiana pyramidalis glauca Hort., J. virginiana plumosa Rehd., J. virginiana reptans Beiss., J. virginiana Schottt Gord., J. virginiana tripartita R. Smith and J. vir- giniana venusta Rehd. Usually the Juniperus hosts of species of Gymnosporangium are con- fined to a single section of the genus. Those of G. clavipes are excep- tional in that they are classified in two sections of the genus, Sabina and Oxycedrus. Geographically the Juniperus hosts of G. clavipes are found throughout the greater portion of the temperate region of the northern hemisphere, an unusually wide distribution for telial hosts of any one species of Gymnos porangium. V. SYMPTOMATOLOGY OF THE DISEASES CAUSED BY GYMNOSPORANGIUM CLAVIPES 1. On Pomaceous Hosts (A) Morphological symptomatology Morphological symptoms of disease caused by G. clavipes were first described from infected fruits on pomaceous hosts. Schweinitz (1832) gave a brief description of the gross morphological symptoms of the disease on the fruit of Crataegus sp. Farlow (1880) stated that G. clavipes (Roestelia aurantiaca) was “by far the most beautiful species of the genus which we have, at once attracting the popular eye by its bril- liant orange or almost cinnabar colored spores and shining white peridium. It is generally found on young fruit, though it is occasionally found on the stems and petioles, but I do not recollect having seen aecidia on the leaves. — One sometimes sees a quince two inches in diameter 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 381 more than half covered by the bright orange aecidia and occasionally small apples are affected in a similar way. Roestelia aurantiaca is gen- erally found in midsummer. I have, however, seen it on C. crus-galli as late as October.” Weimer (1917) stated that “the veins alone (of quince leaves) are attacked and often become swollen to double their normal size.’ The swellings, he noted, caused the leaves to curl but the infected areas were not discolored. No aecia were found on the quince leaves. Adams (1921) reported that branches and buds of hawthorns were very severely injured by this rust. He stated that “the aecia on the branches always precede the appearance of aecia on the fruit of hawthorns.”’ Thomas (1933) reported that symptoms on apple foliage appeared 10 to 18 days after inoculating and that on apple fruits the symptoms were predomi- nately necrotic or hypoplastic. On artificially inoculated leaves of Crataegus he obtained numerous spermogonia “‘but aecia were produced only sparingly along the larger veins.” In the present study, data for the symptomatology of the disease caused by the aecial phase of G. clavipes were obtained from observations made on more than 400 pomaceous hosts. Data on the relationship between elapsed time and progressive stages of development of symptoms and signs were recorded from plants inoculated at various stages of development of the flowers, twigs and fruits. Comparisons of the symp- toms and signs resulting from artificial inoculations made possible a more thorough understanding of the phenomena that occur in nature. Gymnosporangium clavipes attacks primarily the fruits, less fre- quently the twigs and buds and rarely the leaves of its pomaceous hosts. The earliest observable symptom of disease on ovaries and young fruits is a pale yellowish green discoloration. This symptom was seen on cer- tain species even before the petals fell. Occasionally the petals become infected and when they do they usually remain attached for the greater part of the season. In my experience, fruits are susceptible when in early stages of development only. After the petals drop the fruits of most species are no longer subject to infection. On the average, from 6 to 10 days after inoculating the diseased fruits begin to show evidence of infection by slightly pale swellings. From 4 to 5 days later the hyper- trophied zone becomes dotted in its central portion with numerous, tiny, deep-reddish points — the developing spermogonia of the fungus. Within 1 to 5 more days, the first formed spermogonia begin to exude a pale-red, sweetish liquid. During further development the diseased area con- tinues to increase in all dimensions and finally involves but a portion of, or, in many cases, the whole fruit. Occasionally, the infection spreads to the pedicels and even extends into the peduncles and twigs. Spermo- 382 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI gonia are matured progressively over a large portion of the diseased tissue, and oozing of the young spermogonia continues for several days. The older spermogonia die and turn black. Several irregularities or anomalies have been observed in the sympto- matology of the disease during the life span of the spermogonia. Very frequently the infected tissues do not become hypertrophied, but quite cease development. During this time the adjacent tissues continue to expand, resulting in invaginated areas. This is particularly common in orchard apples where infection occurs most frequently at the blossom end. In many instances the spermogonia are few in number and do not reach full maturity. Tissues with this type of infection are usually green and firm. Fruits of hawthorns, shadbushes and chokeberries are found in which the swelling and discoloration involve the whole fruit, but in such cases no or few spermogonia are produced. Many of the fructi- fications exhibit various stages of abortion. Symptoms of this type were obtained from inoculation tests on Crataegus spp. and Amelanchier oblongifolia toward the close of the period of susceptibility of the fruits. Of course, in some instances it is possible that more than one kind of parasite is involved and that as a result the normal course of the Gymno- sporangium disease is altered. In 20 to 40 days after inoculating the second fructifications of the fungus, the aecia, begin to make their appearance within the diseased area. They are usually produced peripheral to, or to one side of, the spermogonia; but they are often found among them. Aecia, in pro- gressive stages of development, are easily observed during their early appearance as is shown in plate 155, fig. 1. Often no aecia penetrate the surface of the infected fruits, yet internal ones are frequently formed. This is particularly true of orchard apples but has also been observed in many other hosts. Fruits that become infected late in the period of susceptibility are commonly observed to exhibit this phenomenon. = Fail- ure to fully develop aecia may be physiological — possibly a type of hypersensitivity — or it may be due to the development of a cuticle so tough as to prevent aecia from breaking through. Many fruits were observed in which the lesions occupied by G. clavipes were browned or blackened and quite rotted. From such decayed areas imperfect fungi were repeatedly obtained by culture. It seems, therefore, that these are the real cause of the discoloration and decay noted. Some have ascribed such phenomena to G. clavipes, but it seems erroneously so. The fact that relatively few areas infected by G. clavipes become decayed strengthens this conclusion. A photograph of a rust-infected area that was parasitised by an imperfect fungus is in plate 155, fig. 2. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 383 An anomolous symptom was frequently observed on fruits of Cratae- gus monogyna infected with G. clavipes. The whole of the fruits and the pedicels were infected, but from the blossom ends of these fruits numerous petals and abortive structures resembling stamens developed. Here again, more than one parasitic organism may have been active in causing these symptoms. Twigs, including thorns, infected with G. clavipes are commonly found on certain species of hosts (plate 156). In my experience twigs of the current season only have become infected. The early symptoms of disease on twigs are pale, yellowish green, elongated areas occurring on the young bark. The infection spreads rapidly, however, usually girdling the stem and extending up to 3 inches longitudinally. The infected por- tion of the twig becomes hypertrophied. Frequently a fusiform swelling results, but irregularly swollen and cankered twigs are also common. Rarely, however, rotund galls are formed on the twigs. Thorns are also subject to infection and manifest similar symptoms. Spermogonia on infected twigs usually do not reach maturity until 9 to 12 days after inoculation. They follow the same course of development as on fruits. Aecia of the rust are produced among the spermogonia as well as outside the area occupied by them. Very often they are sparsely produced, but, on the other hand, twigs are sometimes found in which the aecia are very abundant. The first aecia reach maturity on the twigs and thorns in 30 to 40 days; others are produced progressively, as was described in the case of infected fruits, for a period of about a week. In the longitudinal advancement of the fungus in a twig, it frequently encounters a terminal or lateral bud, which in turn usually becomes infected. Subsequently, the buds are forced to develop beyond the resting stage normal for the current season. Similar phenomena were observed on buds of ornamental apples infected with G. Juniperi-vir- ginianae (Crowell, 1934). Forced growth of buds caused by G. clavipes has been observed on a large number of hosts. The early stages of in- fection, evidenced by a yellow discoloration and swellings of the buds, are not observable until late in June. The abnormal development of stunted twigs and leaves results. The photographs in plate 156 show some of these symptoms in forced buds. Spermogonia in all stages of development are produced along the deformed petioles and veins of the leaves. Rarely, however, are aecia produced on the forced buds. On C. Phaenopyrum several twigs were found in which the infection devel- oped systematically, as shown in plate 156, fig. 5. It is possible that infection occurred on these twigs while they were in the stage of rapid elongation. That infection was confined largely to these malformations and did not extend into the main twig substantiates this supposition. 384 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Many twigs infected with G. clavipes in the spring of 1934 were ex- amined in the spring of 1935 to determine whether or not the fungus over- wintered in its aecial hosts. No instance of overwintering was found. Thomas (1933), however, reported having observed the overwintering of this rust in pomaceous hosts. Dodge (1918) reported that the aecial phase of G. biseptatum Ellis (G. Botryapites [Schw.| Kern) and of G. fraternum Kern (G. transformans |{Ellis] Kern) were observed by him to be perennial. Tubeuf (1906, 1907) also noted this phenomenon on pears infected with G. Sabinae (Dicks) Wint. The leaves of the pomaceous hosts of G. clavipes are rarely infected. When such occurs the lesions are always small and few spermogonia develop on them. The spermogonia on leaf lesions are late in appearing, and on many hosts they never reach the oozing stage. Spermogonia have been observed on leaves of Cydonia oblonga, Amelanchier spp., Craetegus spp., and Aronia spp. Amelanchier oblongifolia is the only host, however, in which I have seen aecia of G. clavipes on affected leaves. On its more susceptible pomaceous hosts G. clavipes is very destruc- tive. The diseased fruits become misshapen and discolored, and often fall prematurely. In the case of quince and apples the loss is primarily a commercial one and frequently very great. Indeed, growing of quinces was impossible in many sections of eastern North America due largely to the ravages of this rust. Many varieties of orchard apples are also very susceptible. Thomas and Mills (1930) report instances in which as high as 95% of the fruits of the Delicious variety in New York were attacked. Ornamental plants, such as certain species of Crataegus, Amelanchier, etc., whose beauty and usefulness depend in a large measure upon an abundant production of colored fruits that persist long into the winter, sometime becomes worthless because so many of their fruits are spoiled by G. clavipes. Twigs killed beyond the infected portion are also un- sightly and they tend to materially deform the trees and shrubs because of sequent prolific sprout growth. Hawthorns have been seen in which so many twigs were killed by the rust that death resulted. (B) Histological symptomatology Tissues of pomaceous hosts infected with G. clavipes are usually hypertrophied and the diseased portions of fruits and twigs are often greatly enlarged; infected parts of leaves, however, are changed but little. In the fruits chiefly the outer or cortical tissues are affected. They are often greatly enlarged and many of them are more or less filled with a deep-staining material while others, fully as hypertrophied, are often quite devoid of contents. Cells towards the center of the fruit, although they show no evident hypertrophy, contain much of the deeply staining 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 385 material observed in the cortical cells. Exceptions to the hypertrophy of the fruit cells as noted have been observed in the infections of many orchard varieties of apples. In such infections the earliest stages of the disease are manifested as an hypertrophy of a small number of cells. The disease soon ceases to be an enlargement; indeed, the infected region becomes inhibited in its development. The surrounding uninfected tissues, however, continue to expand in their development while the diseased area remains stunted, depressed and usually greenish. In twigs, mainly the cortical tissues are affected. The effects are simi- lar to those in fruits. Occasionally, however, and always in the greatly enlarged twigs, the phloem and xylem are hypertrophied. Many of their cells contain much material that stains intensely with haematoxylin. The mycelium of the aecial phase of G. clavipes is found in the cortical tissues of the fruits and twigs and in the palisade and mesophyll tissues of the leaves. Haustoria are abundantly formed in the cells of the fruits and twigs but are seldom seen in cells of the leaves. Frequently, haus- toria are found in close association with the nucleus of its host cell. The peridial cells and the aeciospores of G. clavipes are salient fea- tures for determination of the species. Gymnosporangium clavipes can also readily be identified from median longitudinal sections of the spermo- gonium. In my investigations of the spermogonia of the genus (to be reported in a separate article) three dimensions were averaged, namely, total width, total height and the depth to which the fructification is sunken in the host tissue. The measurements for G. clavipes were found to be 203 uy & 207 uy & 163 1 respectively. The spermogonia of G. clavipes are conspicuous because of their large size and their rotund form and they are almost completely sunken in the mesophyll. A photomicrograph of spermogonia of G. clavipes is shown in plate 156, figure 7. Aecia of G. clavipes are found most abundantly in the fruits and twigs and are seldom produced in the leaves. Aecia develop in the outer cortex of fruits and twigs and in the mesophyll of leaves. They are sunken to a depth of approximately one millimeter and are of greater diameter near the base than at the apex. Fresh aeciospores of G. clavipes vary in color from bittersweet-orange to flame-scarlet (Ridgway, 1912). They are irregularly rotund, verruculose and measure 32.2 p * 34.4 uy. Peridial cells are almost white in color; the inner wall is maked by coarse ridges forming an irregular mosaic pattern. The form of peridial cells varies widely, but on an average they measure 14.5 yp in width & 55.6 y in length. They usually adhere in large numbers in water mounts. 386 JOURNAL OF THE ARNOLD ARBORETUM [VvOL. XVI 2. On JuNiIpERUS Hosts (A) Morphological symptomatology Cooke and Peck (1873) gave a description of the symptoms of G. clavipes on Juniperus virginiana; they stated: “the younger branches are slightly swollen where attacked by this fungus and the bark is scaly.” For many years, however, the fungus and the disease caused on its Juniperus hosts seem not to have been understood. Thaxter (1891) after an extensive series of cultural experiments clearly identified two species that were formerly confused with G. clavipes, namely G. clavipes proper and a new species, G. Nidus-avis. Kern (1911), in an account of the symptomatology of G. clavipes, stated that the telia were “caulicolus, appearing on slightly fusiform swellings, usually aggregated, roundish, one to four millimeters, often confulent, hemispheric — teliospores two- celled ellipsoidal 18 to 26 uy & 35 to 51 p — pedicels carotiform. Dodge (1918) stated that small witches-brooms are sometimes formed on twigs of red cedars infected with G. clavipes. Dodge (1922) stated that needles as well as the main trunk are also infected, and that the bark over the infected portions of the trunk becomes much thickened and blackened. The pertinent observations on the symptomatology of G. clavipes re- viewed in the foregoing give a clear foundation for an appreciation of the disease caused by G. clavipes on its Juniperus hosts. My own findings and interpretations, while they add little to what is already known will, nevertheless, trace in sequence the development of the disease and its relative importance in the various organs attacked. On its Juniperus hosts, particularly J. virginiana L., the disease in- duced by G. clavipes is perhaps one of the most destructive caused by any species of the genus Gymnosporangium. The needles, twigs, branches and even the main trunk are attacked. Very frequently the disease occurs on the needles but infected needles are relatively incon- spicuous and are often overlooked. Usually but a single crop of telio- spores is produced on needles, after which they die. From the needles, however, the fungus often migrates to the twigs and it is on these that the disease is most frequently found. Twigs may also become infected directly. On twigs the disease appears as slightly fusiform swellings covered by a flaky, darkened bark. Usually by the end of four to six years most of the infected twigs die. Occasionally, however, the infected twigs survive for a longer period and the larger limbs are distorted and covered with a heavy, cracked and blackened bark. If the diseased por- tion of a branch is near the main trunk the latter is liable to infection by the fungus advancing along the cortex. Infections of G. clavipes on the 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 387 main trunk of red cedars are easily mistaken for one of the diseases caused by other species of the genus, because vertically elongated, often irregu- lar, blackened, heavy-barked bulges are typical for diseases caused by several species of Gymnosporangium. Fructifications are, in my ex- perience, necessary to identify the causal organism on trunk lesions. That the disease caused by G. clavipes is very destructive to its telial host is evident since all organs attacked are killed usually in a few years. Trunk infections of many years standing are not uncommon, to be sure; but trees bearing such burdens show evident symptoms of poor health. Recently, I came upon a very striking demonstration of destruction of red cedars infected with G. clavipes. Several years ago, an estate owner cleared a natural grove of cedars of all other trees and shrubs. Gymmno- sporangium clavipes was very abundant on the twigs, branches and main trunks of certain of these cedars. Following the unusual cold of the winter of 1933-34 every tree that was heavily infected with G. clavipes died entirely or in large part, while all of the trees in this group that were not infected survived. An examination of the cedars in the surrounding uncleared lands revealed complete destruction of trees that were heavily infected with G. clavipes. As the only variable seemed to be the relative abundance of infections the loss seemed clearly attributable to the infec- tion of G. clavipes. (B) Histological symptomatology Dodge (1922) gave a very complete account of the histological symp- tomatology of G. clavipes. A review of his paper is presented here. Dodge stated that “infection takes place on the proaxial side of the young leaf, or directly on the young stem at the base of the leaf — after entering the leaf, the mycelium invades the region between the cuticle and the cellulose walls of the epidermis on the proaxial side. The effect of the fungus on this part of the cell wall is usually marked by consider- able swelling and the disorganized substances take the stains very readily — the fungus explores and feeds in the cuticularized layer, but it may go deeper and invade the palisade-mesophyll tissue. Haustoria are found in the epidermal cells, sometimes even in the guard cells of the stomata.” Young stems are “susceptible to infection, either directly or through the invasion of the fungus by way of the leaf axils.” The mycelium occu- pies a cancellate portion of the periphery of the young cortex. The mycelial “strands actually interlace, weaving in and out around the veins [leaf traces] of the leaves and forming a closed network system in the cortical region of the young stem.” After the leaves are shed the fungus, closing the gaps, may be found in the cortex around the entire circum- 388 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI ference of every section. In the main trunk, infections are confined to a portion only of the circumference. My findings with respect to the histological symptomatology agreed in detail with those of Dodge. My examination included the leaves and twigs of Juniperus virginiana as well as infected twigs of all other Juni- perus hosts. Certain additional remarks with respect to the fructifica- tions are also recorded. The telia of G. clavipes arise at irregular intervals from aggregated masses of the mycelium on the phellogen. The telia first appear early in April in Massachusetts. In youth they are a deep-reddish color, expand- ing upon gelatinization to regular pulvinate sori. After two or three gelatinizations, however, the sori change their deep red color to a yellow- ish red and their shape is very irregular. Five to six gelatinizations occur during the season, after which the telia drop from their host. VI. LIFE HISTORY STUDIES OF GYMNOSPORANGIUM CLAVIPES C. AND P The life history of Gymnosporangium clavipes is essentially the same as that of other species of the genus. It differs in certain details only from the life history of the more generally known species G. Juni peri- virginianae. A review of the extensive literature dealing with the life- history of G. clavipes shall be confined to the more pertinent reports of former investigators. Schweinitz (1832) described the aecial phase of a rust which he found occurring “‘rarissime in germinibus Rosae” as Caeoma (Peridermium) germinale, Kern (1911) reported the determination of the host of this rust to be an error for 2 species of Crataegus. Cooke and Peck (1873) gave an account of the telial phase of a rust which they called G. clavipes occurring on Juniperus virginiana L. These two rusts were considered to be distinct species until Thaxter (1887) showed from the results of controlled cultural experiments that the spermogonial and aecial stages on pomaceous hosts were in reality genetically connected with the telia stage on red cedars. Following this basic step many observations have been reported with respect to details in the life history of G. clavipes. In the development of the rust on pomaceous hosts Farlow (1886) and Thaxter (1887) found that 10 and 11 days respectively elapsed between the date of inoculating and the first appearance of spermogonia. Thomas (1933) stated that the first symptoms of disease on “relatively resistant” apple foliage were observed 10-18 days after inoculating but that spermo- gonia were never formed on the lesions. In contrast he found that symp- toms appeared in 4-6 days and spermogonia in 13-15 days after 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 389 inoculating “susceptible” foliage of Crataegus sp. and further that aecia were formed sparingly along the larger veins only of the infected leaves, Thaxter (1887) determined from his cultures that approximately 30 days were required for the maturation of aecia on young shoots of Amelanchier canadensis. Miller (1932) and other authors observed that spermogonia were frequently abortive in orchard apples and that aecia rarely matured in the fruits. These observations while they differ greatly when considered sepa- rately are nevertheless in harmony when reviewed in the light of more complete knowledge of the behavior of the rust. From my cultural ex- periments on hosts in several genera, the same variations as reported in the foregoing were observed. These variations seemed to be correlated with two phenomena, — first, the relative susceptibility of the host plants and second, the stage of development of the diseased parts at the time of inoculating. The rust developed more rapidly on the more susceptible hosts and more slowly on less susceptible ones. In fact, on very resistant hosts, as certain orchard apples or on resistant organs as, for example, the leaves of nearly all hosts, longer time was required for the maturation of the fructifications; not infrequently these reached an imperfectly developed stage only. The average time required for the maturation of spermogonia of G. clavipes on fruits and twigs was from 7 to 10 days. The beginning of spermogonial exudation was taken as the criterion of maturity. On indi- vidual lesions spermogonia continued to be produced for a period up to 3 weeks. The production of exudate by any one spermogonium contin- ued up to about 7 days, after which the spermogonium became filled with long filaments, and soon died, usually turning black. Others in the lesion followed these developments. Spermogonia of G. clavipes were sub-epidermal in origin and position; they were reddish in color and among the largest of the genus. Details of average measurements of their size have been given on page 385. The aecial primordium of G. clavipes is located deep in the cortical tissues of fruits and twigs and in the mesophyll of leaves. It is typical for the genus as is also the mycelium and haustoria. From thirty to forty days after inoculating, aecia of G. clavipes reached maturity and pene- trated the epidermis of the host among or close by the spermogonia. Aecia were usually developed progressively over a period of a week or more. Upon their first appearance, about June 1 in Massachusetts, they had the form of short, blunt, white cylinders. One or two days later the white peridia were ruptured irregularly, usually with the loss of the cap cells, exposing and releasing the enclosed reddish aeciospores. Vari- 390 JOURNAL OF THE ARNOLD ARBORETUM [voL. xvI ous stages of this phenomenon are shown in plate 155, fig. 1. The aecium of G. clavipes is broader at about midway between the hymenium and surface of the host than at the surface of the host itself, It is evident, therefore, that a crowding of the aeciospores occurs at the zone of con- striction. Dodge (1924) associates these phenomena with a mechanism for the forcible discharge of aeciospores. The peridial cells of G. clavipes are broadly rectangular and measure 14.5 uy X 55.6 y with extremes of 33.4 1 to 83.5 u & 10.4 i to 234 a: Their inner and thicker walls are ornamented with low interconnecting prominences forming irregular mosaic patterns. Peridial cells are of much diagnostic value as Fischer (1891) and Kern (1910) demon- strated. Those of G. clavipes may be easily identified by their markings, their size and their form. Peridial cells of G. clavipes usually remain flat in water mounts and adhere forming a large sheet of cells. Aeciospores of G. clavipes are exceptional for the genus in their re- markably intense color. By comparisons of fresh aeciospores the color was determined as varying from bittersweet-orange to flame-scarlet according to Ridgway’s (1912) color standards. With increase of age of the mature spores, however, their color gradually changes to orange and yellows. In old herbarium material they are frequently almost colorless. In many of the better preserved specimens the reddish color still is conspicuous. Aeciospores of G. clavipes are among the largest of the genus. They measure 32.2 1 34.4 y with extremes of 28.4 u to 42.3 4 X 26.7 4 to 37.4 y. Their outer surfaces are ornamented with numerous, tiny, low papillae. The problem of germinating aeciospores of species of Gymnos poran- gium has been given a great deal of attention. Difficulty has been ex- perienced in germination of aeciospores of many species. Several investigators have shown that a period of rest at low temperature con- tributed greatly to the germinability of aeciospores of certain species of Gymnosporangium (Fukushi, 1925; Miller, 1932). Thomas and Mills (1929) reported moderate germination of aeciospores of G. clavipes stored for twelve weeks at 3° C. Thomas (1933) tested the germina- bility of aeciospores of G. clavipes that were precooled for various lengths of time at various temperatures as well as aeciospores that were not pre- cooled. The highest germination was obtained from spores that “were mounted at 18° C. without precooling.” Thomas also demonstrated that aeciospores kept dry at 3° C. rapidly lost the property of germination. It was found, however, that a small number of aeciospores remained viable in aecia in fruits throughout the winter. It should also be re- marked that Professor J. C. Arthur observed internal aecia of G. clavipes 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 391 in fruits of orchard apples. Dr. Arthur stated in a letter to Dr. Stein- metz, who forwarded the material to him, that aeciospores removed from internal sori germinated in the usual manner. n my studies of the germination of aeciospores of G. clavipes it was found that they germinated readily in a moist atmosphere at room tem- 85 a Cermination Percentage uw w 8G nN Y r) 3 6 9 l2 15 18 al 24 27 30 53 Degrees Centigrade. Fic Graph showing the Germination of Aeciospores of Gymno- sporangium clavipes C. and P. perature at any time during the summer and fall seasons. Careful studies of germination revealed that they germinated over essentially the same range as did most species of the genus studied to date. Optimum con- ditions for germination in distilled water on glass slides were reached at about 15° C. The results of the tests are shown in figure 2 and in table 4. 392 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XVI TABLE IV. GERMINATION OF AECIOSPORES OF G. CLAVIPES No. of spores No. of spores Percentage of Temp. poets poe i 00°C 1000 0 0 3 1000 53 §.3 6 1000 224 22.4 9 1000 506 50.6 12 1000 740 74.0 15 1000 807 80.7 18 1000 750 75.0 21 1000 698 69.8 24 1000 625 62.5 27 1000 413 41.3 30 1000 152 15.2 oo 1000 0 0 Apart from the irregularities in the aecial phase that were discussed under the heading of symptomatology, an unusual development was ob- served in the aecia on fruits of Crataegus sp. collected by Prof. J. H. Faull in Pennsylvania. Some of these aecia and aeciospores appeared normal in all respects. Certain of the aecia were internal with the hymenia oriented in various directions. All of these latter and many of the aecia that developed in a normal position were filled with irregularly produced aeciospores and aeciospore chains. Cytologically either one or two nuclei were present in the cells. The cell contents were irregularly inter- spersed with vacuoles and deep-staining materials. Camera lucida draw- ings of two aeciospore chains are shown in plate 157, figure 2. Aeciospores of G. clavipes are primarily wind borne. They are dis- tributed throughout the growing seasons and probably germinate and infect the alternate hosts shortly after inoculation. Previous to 1910, however, little investigative work had been con- ducted on the telial phase of G. clavipes. Arthur (1912) sowed aecio- spores of G. clavipes on the common juniper and reported successful cul- tures. Dodge (1918) traced the life cycle of G. clavipes under controlled conditions from the telial phase on red cedars to the aecial phase on Crataegus Oxyacantha, thence back to the telial phase on red cedars. Dodge found that few telia were produced in the spring following inocu- lation but that many developed in the second spring. Inoculations made during the present investigations substantiated these findings. In addi- tion it was shown that inoculations made on July 1, August 10 and Octo- ber 3 all resulted in abundant infection, indicating that in nature aecio- spores are a menace throughout their entire pei.od of production. Infection of the telial host was first described by Dodge (1922). 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 393 Dodge stated that entrance of the germ tube was gained through the adaxial surface of the leaves and tender epidermis of the twigs. In the leaves of its Juniperus hosts the mycelium of G. clavipes is confined al- most exclusively to the epidermal cells. In twigs it is restricted to the phellogen cells — a most unusual limitation for the mycelium of a species of Gymnosporangium. Within the infected cells characteristic, binucle- ate, sac-like haustoria are formed. They are abundant, usually occurring singly but frequently in twos or threes. Telial sori of G. clavipes arise from masses of the mycelium. In leaves the mycelium usually dies after the production of one crop of spores, but in twigs the fungus is perennial for several years. After each successive crop of teliospores a new phello- gen layer is formed immediately beneath the sorus. The surrounding vegetative mycelium then grows over this new tissue. It is from the mycelium on the older phellogen that the telium for the ensuing spring is produced. In certain microscopic sections it is possible to observe progressively (a) the dead mycelium on partially sloughed-off phello- gens, (b) sori of the present season as well as (c) the primordia of sori for two seasons to come. The camera lucida drawing in plate 160, fig. 1 was made from such a section. Telia of G. clavipes are produced on leaves and on the bark of various- sized branches and even the main trunk of red cedars. They were never observed on branches of red cedars more than one-half inch in diameter. Upon their early appearance telia of G. clavipes are aggre- gated, pulvinate in form, 2 to 5 mm. across and are distinctly bright reddish in color. During rains in the spring the telia swell to regular gelatinous forms as is shown in plate 159. After three or four gelatiniza- tions the telia lose their regular form and deep-red color, becoming shapeless yellowish-red masses. After 6-8 gelatinizations the telia drop from the infected parts. The development of teliospores of G. clavipes is essentially the same as Dodge (1918, 1922) reported for this and other species of the genus. Camera lucida drawings made during these investigations of teliospores in various stages of development are shown in plate 160. Teliospores of G. clavipes are at once distinguished by the swollen pedicels near their bases. No other species of the genus in eastern North America has this characteristic.. Certain data in regard to teliospores of G. clavipes are of interest. Both one- and two-celled teliospores are produced. Ina count of 1000 spores, 94.8% were found to be two-celled while 5.2% were one-celled. Two-celled teliospores have one germ pore in each cell. In the upper cell the germ pore is apical while in the basal cell the germ pore is located near the pedicel. In other respects one- 394 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI celled teliospores resemble the two-celled ones in all but the septum. One-celled teliospores have the germ pore at the apex. In the telium of G. clavipes both thick- and thin-walled spores were found. Thick-walled spores are more numerous and are always produced on the outer surface of the telium. Almost invariably one-celled teliospores are thick-walled. Thin-walled teliospores are located within or beneath the layer of thick- walled spores. Measurements were made of the lengths of the upper and basal cells, the total length and the width of teliospores of G. clavipes. The upper cell measured 23.0 ,) with extremes of 16.5 to 33.0 1, the basal cell 21.2 yy with extremes of 14.9 to 33.0 yp, the total length was 44.5 with extremes of 33.0 to 57.8 y and the width 22.7 y with extremes of 16.5 to 33.0 pp. Single-celled spores measured 19.7 : in width by 33.4 in length with extremes of 26.4 to 51.2 u & 14.9 to 29.7 x. When teliospores of G. clavipes first break through the cortical cover- ing layer of the host, they expand but little when wetted and the telio- spores do not germinate. From 1—3 weeks after their first appearance, however, the telia expands fully to a regular pulvinate form and the teliospores germinate in great abundance. Several workers have reported the results of their investigations on germination tests of teliospores of G. clavipes. Weimer (1917) formu- lated a general curve of germination percentages obtained at various temperatures and stated that it applied to the germination of teliospores of G. clavipes as well as teliospores of other species of Gymnos porangium. The extreme temperatures found by Weimer were 7° C. and 29° C. and the optimum temperature was between 22° C. and 25° C. Miller (1932) found the extreme temperatures of germination to be 4° C. and 32° C. and the optimum germination at 25° C. He also investigated the phe- nomenon of the maximum rate of germination and found that when telia of G. clavipes were immersed in water for 25 minutes and removed to a moist atmosphere they discharged basidiospores in considerable num- bers within two hours. It was also found that when telia were mounted in moist cotton and kept at 25° C. an abundant germination of the telio- spores and the beginning of basidiospore formation occurred after an interval of eight hours. Thomas also investigated the time required for and the rate of germi- nation of basidiospores. He stated that basidiospores submerged in water at room temperature developed germ tubes equal in length to the spore in two hours and from four to six times their diameter in ten hours. Basidiospores were also mounted on a moistened leaf of Crataegus and held at 25° C. with the result that germ tubes reached a length of five to seven times their diameter in eight hours. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 395 Farlow (1886) reported the production of secondary basidiospores from primary basidiospores of G. clavipes, a phenomenon that has fre- quently been reported for other species of the genus. Beyond observing the usual germination of the teliospores of G. clavipes, a process common to other species, no further studies were made on the phenomenon. It was observed, however, that germination occurred in the field when the telia remained gelatinized for periods of es ah seer teres as | pssst e ow r ae ap Scesisees 9 — Wagsstes: He eases “nh Cong Aigeesse: UA rn Ficure 3. Geographical Distribution in North America of G. clavipes and its Hosts. The pomaceous hosts of G. clavipes shown as vertical lines. The Juniperus hosts of G. clavipes shown as horizontal lines. Stations for G. clavipes shown as dots. Note the extreme northern limit in northern Alberta, Canada, and the extreme eons limit near Mexico City, Mexico. two hours or longer. Four to six gelatinizations were usual during the season and basidiospore dissemination occurred at most, if not at all, of these. It is essential to the production of infection that the pomaceous hosts are within their period of susceptibility at the time of inoculation. As the period is very short in many species, the failure for the coincidence of basidiospore dissemination within this period may be held responsible for the variable abundance of spermogonia and aecia on certain poma- ceous hosts in different years. Crowell (1935) reported unusual de- 396 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI velopments of the aecial phase of this and other species of Gymnospo- rangium in the very dry spring of 1935. In the spring of 1935, no rains sufficient to cause gelatinization of telia of G. clavipes occurred during the flowering period of most pomaceous hosts; most fruits becoming immune before inoculation took place. This resulted in a very limited number of infected fruits and the late appearance of the rust generally. Although the host relationships for the perpetuation of G. clavipes are found over the greater part of the temperate northern hemisphere, never- theless, so far as I am able to learn, this rust is not known outside of North America. Reported stations for this rust are most abundant in the eastern part of this continent as shown by dots on the outline map of North America in fig. 3. The rust has been collected, however, in widely separated stations outside the region of greatest concentration, an un- usual feature in the distribution of any species of this genus. Gymnosporangium clavipes has been reported from southern New- foundland, from all the provinces of Canada except Prince Edward Island, from all the states of the United States except Arizona, Cali- fornia, Idaho, Kansas, Minnesota, Nevada, New Mexico, North Dakota, Oregon, South Dakota and Washington. It is also reported from Mexico State in Mexico. No other species of Gymnosporangium is known to have a longitudinal range extending from central Mexico State to northern Alberta in Canada. An even greater range is possible on the basis of host distribution. It would be very interesting as well as of much practical value to determine the phenomena responsible for the distribu- tion of a rust within a portion of the territory occupied by both host groups. VI. CONTROL MEASURES APPLICABLE TO GYMNO- SPORANGIUM CLAVIPES ON POMACEOUS AND ON JUNIPERUS HOSTS Prophylactic measures to control G. clavipes have been largely an adaptation of those practiced for the control of the cedar-apple rust diseases caused by G. Juniperi-virginianae. The results obtained in control work have been essentially parallel for both of these diseases, namely, fungicidal control as practiced was found unsatisfactory, while eradication of Juniperus hosts gave excellent protection to pomaceous hosts. Halsted (1893) reported the destruction of quince, apple, hawthorn and shadbush fruit by G. clavipes in New Jersey and recommended remedial measures. Concerning the disease Halsted wrote: ‘‘an enemy is beyond the fence; therefore, go out and slay him with an axe.” This 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 397 was the earliest record that I found regarding the pathenogenicity and control of this rust. Bailey (1894) also recommended the destruction of red cedars as a control measure. He gave evidence to show that spraying thoroughly (with Bordeaux ?) was of considerable value. Many other authors stated that eradication of red cedars offered the most satisfactory solution to the control of this rust. y own investigations on control measures applicable to G. clavipes were carried out simultaneously with those for the control of the cedar- apple rust fungus, G. Juniperi-virginianae (Crowell, 1934). Studies in this problem included an exploratory investigation of numerous fungi- cides with respect to their control value both on pomaceous hosts and on red cedars. The most promising of these fungicides were then tested on an extensive experimental scale on numerous trees under various weather conditions. The fungicides were applied to red cedars (a) to prevent germination of the teliospores, (b) to protect them from infection by aeciospores, and (c) to pomaceous hosts as a protection against basidio- spore infection. (A) Fungicidal applications on red cedars to prevent germination of teliospores On the Lyman estate, Canton, Massachusetts, Gymnosporangium clavipes was in very great abundance and exploratory tests with several fungicides were made there. In table V are enumerated the sprays and dusts that were used in these tests. SPRAY AND DUST MATERIALS USED IN EXPLORATORY TESTS Bordeaux 3:3 :50, 4:4:50, 6:6:50 Linco colloidal sulfur? ae Y%, oe 2%, 3% Lime-sulfur 1:30, 1:40, 1:50, 1:70 Soluble palustrex? oh ‘2%, 3%, 4% “é “é A. “cc e “ec se “ce “cc B. iz “ce “ce “é “ce ce Cc. ce “ce “cc “ee Sunoco oil? 2%, 4% 50% Sunoco oil and 50% soluble palustrex 2%; 4% 50% Sunoco oil and 50% soluble palustrex B. 2%, 4% 80% Sunoco oil and 20% copper resinate 2%, 4% 80% Sunoco oil and 20% soluble palustrex 2%, 4% Kolo base Kolo dust Pomo green Sulfur dust 1Obtained from Linder and Co., 296 North Beacon St., Boston, Mass. “Obtained from E. W. Coolidge, Jacksonville, Florida. 8Obtained from Sun Oil Co., Boston, Mass 398 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI All of these spray and dust materials were first applied to potted red cedars in the greenhouse in the spring. None of them caused burning of the young foliage. In the field the sprays were applied to twig lesions as follows: (1) before the telia had emerged, (2) just after the telia had emerged, and (3) after one, two or three gelatinizations of the telia, but always when the telia were dry. Telia to which the sprays were applied were brought into the laboratory for examination; smear slides were made of the spores and examination was completed shortly after their arrival. It was not the purpose of this examination to determine the relative value of each spray, rather the purpose was to determine which ones would prevent germination of the teliospores. The most satisfac- tory sprays were soluble palustrex B. at 4% and Linco colloidal sulfur at 1%, 2% and 3%. In testing these sprays further it was found that this colloidal sulfur at 2% and 3% was the most constant in its reactions. Colloidal sulfur was, therefore, chosen for the experimental work that followed. It should be added, however, that colloidal sulfur at 1% greatly reduced the amount of germination; colloidal sulfur at 2% com- pletely prevented germination of the teliospores of G. clavipes. Colloidal sulfur at 3%, therefore, was unnecessary. In April 1933, many telia of G. clavipes were sprayed upon their first appearance with colloidal sulfur at 1% and 2%. The effect of the spray was determined by gathering several sprayed telia three days after the application and also unsprayed telia at the same time for controls, thoroughly wetting and keeping them in a moist chamber over night. When germination was abundant a spore print resulted, but when no spore print was formed a smear slide was made and the teliospores ex- amined under the microscope. The unsprayed teliospores always germi- nated in abundance. Teliospores sprayed with colloidal sulfur at 1% strength germinated to some extent but few basidiospores germinated. Telia sprayed with 2% colloidal sulfur showed no germination of the teliospores. After the rain following each spray application, telia were again gathered and immediately tested. The controls germinated in abun- dance. The telia sprayed with colloidal sulfur at 1% and 2% germi- nated to a slight extent. Some change, therefore, was called for in order to control this small amount of germination. Certain substances were used to lower the surface tension of the spray material which might aid its penetration into the telium. Spreaders were tried but with no suc- cess. Calcium casienate, a combined spreader and sticker, was next used in a series of experiments. A new lot of telia was sprayed with colloidal sulfur at the strength of 2% plus 2 pounds of calcium casienate 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 399 per hundred gallons of water. Telia tested in the laboratory showed a high percentage of germination for those that were unsprayed and no germination for those that were sprayed. In subsequent rains none of the teliospores were observed to have germinated nor did the telia at any time regain the property of gelatinizing fully, although they did gela- tinize to a slight extent. (B) Fungicidal applications on red cedars as a protection against infection by aeciospores Fungicidal protection to red cedars against infection by aeciospores of G. clavipes has been demonstrated on potted red cedars in greenhouse trials. In these tests Linco colloidal sulfur at the strength of 12% or 6 pounds per 100 gallons of water only was used. Twenty-five red cedars twelve to eighteen inches high were sprayed with the fungicide. These and twenty-five unsprayed red cedars were thoroughly wetted with a strong stream from a hose the following day. They were then heavily inoculated with fresh aeciospores of G. clavipes and kept in a moist chamber for five days. In the spring of the second year after inoculating, each of the unsprayed plants produced an abundance of sori — a total of more than one thousand separate infections — while but a single infec- tion was found on the sprayed plants. In field experimentation, three applications (one each in July, August and September) were made to a group of red cedar trees near heavily infected hawthorns. Examination of the twigs the second spring after spraying showed a very marked reduction in the number of lesions pro- duced. It was estimated after comparing the amount of infection on unsprayed red cedars in the vicinity that about 75% control was obtained. The use of fungicidal means of protecting red cedars in practice should be guided, along with other considerations, by two important factors, namely, (1) the date of maturity of the aecia and (2) the duration of aeciospore production. The first of these will vary with the season and with the time of blos- soming of the host. In Massachusetts aecia reach maturity early in June. The second and the more variable of the two is the period of active aecio- spore production. The duration of this period is dependent upon the date of maturity of the fruit of the associated pomaceous hosts. As the fruit of Amelanchier ripen and drop about the middle of July in Massa- chusetts and aeciospore production ceases at this time, protective sprays, therefore, need be applied for but a brief period. Twig infections, if present on the shadbushes, should be carefully removed, because aecio- spore production will continue on them throughout the entire growing 400 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XVI season. On the other hand fruits of Crataegi and most other pomaceous hosts do not reach maturity until late fall and aeciospore production usually continues during this period. (C) Fungicidal applications on pomaceous hosts as protection against infection by basidiospores Fungicidal protection of pomaceous hosts from infection by G. clavipes has been conducted through field experiments only. Individuals of Amelanchier oblongifolia were sprayed with Linco colloidal sulfur at the strength of 12% or 6 pounds per 100 gallons of water after each of three rains during the early development of the flowers in the spring. Spraying was begun just before the first rain after the flower clusters began to unfold; the last application was made when about three-fourths of the petals had fallen. The protection afforded these plants was ex- cellent. Counts of infected fruits showed that 98% of them remained free from infection, while 95% of the fruits on unsprayed plants nearby were infected. Certain other experiments with this same material on hawthorns and apples have not met with the same success. Excellent protection was afforded certain of the tested trees while practically none was obtained on others. It is believed that this irregularity can be over- come. It is not yet known just when infection occurs nor the limits of the period of susceptibility of many of the host species. A knowledge of these is necessary to satisfactory control. The value of thorough spraying cannot be over-emphasized. Gymnosporangium clavipes is largely a fruit parasite on its pomaceous hosts, and a difficulty arises with respect to spraying while the hosts are in flower. Sprays may be applied at any time except for a few days imme- diately after the unfolding of the petals. It is at this time that pollination occurs. The problem of controlling the disease caused by G. clavipes on orchard apples has not as yet been given attention. A project to deter- mine the limits of the period of susceptibility and modifications, if any, of the spray schedule now generally used in apple orchards is planned. In addition to protective spray applications for the control of G. clavipes, other means of attack may be employed. In ornamental plant- ings of red or common cedars judicious pruning of infected branches and twigs is very effective. Not only will pruning remove the disease from infected trees but will afford a degree of protection to neighboring poma- ceous hosts as well. The work is best done in the spring when infected parts are clearly marked by the presence of telial sori. Small twigs may be cut off below the nearest uninfected shoot, but it seems advisable that larger branches be removed well below visible lesions; in many instances this may be back to the main trunk. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 401 The possibility of freeing infected trunks of red cedars from disease has been the subject of an, as yet, incomplete investigation. It will be recalled that the mycelium of G. clavipes is localized on the phellogen layer only of red cedars. Experimentation was conducted for the pur- pose of investigating the possibility of removal of the fungus by remov- ing the outer bark. For this purpose a coarse wood rasp was used, and the bark, over and for about an inch around the lesion, was scraped off and painted with shellac and later with an antiseptic tree dressing. The work was done in March and no telia developed in the spring nor did the trees show visible symptoms of injury the following growing season. Only one season has passed since the undertaking and thus far the operation seems successful. Eradication of Juniperus and of pomaceous hosts, while often limited in practice, is; nevertheless, a very effective adjunct or under certain conditions the most effective means of controlling this rust. The low growing types of junipers, such as J. communis and varieties and J. hori- zontalis, are frequently weed plants. In some localities in the Annapolis valley of Nova Scotia wild junipers have been eradicated for one-quarter to one-half mile around commercial orchards with the result of almost complete protection to the susceptible varieties of apples. From their observations in an apple orchard in Maine, Steinmetz and Hilborn (1934) state: “the approximate shortest distance between the infected junipers and infected apple trees is 3900 feet. The extreme distance between infected hosts is over 4500 feet.” Various environmental fac- tors, however, may influence the distance that the contagion will travel as I have already discussed in another paper (Crowell, 1934 pp. 202— 209) and these should be considered in plans for eradication. As an expedi- ent to the protection of either pomaceous or Juniperus hosts, all nearby hosts that can be dispensed with should be removed. Elaborate protec- tive measures may be largely upset by a single weed-host plant located near valued plantings. VIII. RECOMMENDATIONS The problem of controlling the diseases caused by G. clavipes on poma- ceous and on Juniperus hosts usually varies with the relative value of the infected plants and with the interest of the owner. The problem, how- ever, merits wider attention. The moral obligation of consideration for a neighbor’s earnest endeavors to improve a serious situation is too often thought of very lightly. Several practical control measures have been demonstrated. These are as follows: 402 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI 1. Selection of immune or highly resistant species and varieties for planting. Attention should be given not only to avoiding hosts of G. clavipes but to avoiding hosts of other species of Gymnosporangium native in the vicinity. Host lists for all of the species of Gymnosporan- gium in eastern North America are now rapidly nearing completion. MacLachlan (1935) has shown that practically the whole of the genus Crataegus is more or less susceptible to attack by G. globosum. Many species and varieties of hawthorn are also susceptible to G. clavipes as shown in the present paper and to another widespread species common in this region, namely, G. clavariaeforme. (The presentation of results of investigations on this species is planned for an early publication. ) Species and varieties of Crataegus, either native or foreign, cannot be planted in the vicinity of Juniperus hosts in eastern North America with expectation of their remaining entirely free from infection by one or more species of Gymnosporangium. In the genus Malus, Crowell (1934) has shown that all of the native species and varieties and one foreign species (/. sylvestris) can harbor and reproduce the cedar-apple rust fungus G. Juniperi-virginianae. The present paper and that of MacLachlan (1935) give but few additional species and varieties of Malus that are hosts to G. clavipes and G. globosum. All other Eurasian species and varieties do not harbor and reproduce the rusts native to this region. In the genus Amelanchier one species only, namely, 4. ama- bilis, has proved to be highly resistant to all of the Gymnosporangia in this region. Other species and varieties are subject to infection by one or more species of Gymnos porangium. Few hosts for the native Gymnosporangia were found in the genera Pyrus and Sorbus while practically all of the species and varieties tested in the. genera Aronia, Crataegomes pilus, Cydonia and Photinia were susceptible. Information as to the hosts in other pomaceous genera is too meagre for general recommendations at this time. For specific host lists together with the relative data of various hosts attention is directed to the present and the following publications now available. For hosts of G. globosum, (MacLachlan 1935). For hosts of G. Juniperi-virgini- anae, on ornamental apples (Crowell 1934); on orchard apples (Crowell 1935). Of the telial hosts, with the exception of Juniperus virginiana, J. hori- zontalis, J. scopulorum and several of their varieties among the native species and J. communis and J. Sabina and several of their varieties among the foreign species, all other species and varieties tested (see p. 379) may be considered as desirable for plantings in this locality. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 403 2. Planting a screen of tall non-susceptible trees about groups of alternate host plants. Groups of alternate host plants in close proximity may be effectively protected from infection by surrounding them with a screen of tall trees. Densely branching trees, as many of the conifers, are particularly effective. The corollary is also true. Groups of alter- nate host plants may be grown in close proximity if planted among taller non-susceptible trees. These phenomena have frequently been observed in nature. A fuller discussion is given on page 204 of an account of the cedar-apple rust disease by Crowell (1934). 3. Eradication of pomaceous or Juniperus hosts. To be most effec- tive eradication of either pomaceous or Juniperus host plants should be complete over a radius of at least one-half mile. Even though eradication is carried out over this area, complete protection is not assured. Thus, the direction of winds during the time of spore production, continued humidity and the location of the source of inoculum are factors that may tend to offset or vary the results. A high degree of protection may be expected from eradication, however. If eradication cannot be complete, partial eradication of host plants will reduce the amount of inoculum and will therefore aid in controlling the rust. All wild or scrubby pomaceous as well as Juniperus hosts that can be dispensed with should be removed. 4. Removal of infected parts of host plants. As infections in Juni- perus hosts are perennial for several years removal of diseased twigs and branches will contribute materially to control measures. Diseased twigs should be removed well below visible lesions. In the case of infected branches it may be necessary to remove them back to the main trunk. Similarly, on pomaceous hosts diseased fruits and twigs may be removed as an aid to control measures. Unless diseased parts are few in number and can be easily and thoroughly picked by hand, the undertaking is not recommended. 5. Removal of the fungus from infected trunks of red cedars. In- complete experimentation has shown that infections on the main trunk of red cedars may be satisfactorily removed. This is done by rasping off the outer bark down to the living outer tissues, painting the wound with shellac and later with an antiseptic tree dressing. The practice is limited to trunk lesions and is not generally recommended. Removal of diseased parts or the disease from parts of Juniperus hosts is best accomplished when telia are present on infected areas to guide one in the work. 6. Protective spray applications. Protective spray applications to pomaceous plants for protection against infection by basidiospores of G. clavipes are of value during the early stages of development of flowers 404 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI and fruits. Linco colloidal sulfur was the only fungicide tested for this purpose. Applications at 2% strength or 6 lbs. per 100 gallons of water plus a casein sticker are recommended. The first application should be made before the first expected rain as the cluster buds are breaking. Sub- sequent applications of the same strength should be made at 7-10 day intervals until most of the petals have dropped. More frequent applica- tions may be necessary if rain is unusually heavy or prolonged, or if flowers have expanded with unusual rapidity. Spray applications to Juniperus hosts should be confined to the period of aeciospore production on neighboring pomaceous hosts, and should be made after each two or three rains. In Massachusetts aeciospores are first liberated about June first. In the case of most species of Amelan- chier the fruits are ripened and dropped by the middle of July. There- fore, spray applications need not be continued longer than this time, pro- vided that Amelanchier species are the only hosts in the vicinity. Care should be taken, however, to remove all infected twigs from these plants for aeciospore production on them continues throughout the entire grow- ing season. On most other hosts, aeciospores continue to be liberated during the entire growing season. Under such conditions spray applica- tions should be continued until the end of the growing season. The telia of G. clavipes are fully exposed one to two weeks before susceptible parts of pomaceous hosts are released from their buds. This affords an opportunity to attack the telia before it is possible for them to cause infection. Two spray applications of colloidal sulfur at the strength of one percent or 10-12 pounds per one hundred gallons of water are recommended. The applications should be made before buds of pomaceous hosts burst. Each application should be made just after a rain in which the telia are fully expanded and are beginning to dry. At this time they expose the greatest surface and absorb water with much avidity. It should not be expected that the fungus in the twigs and branches will be killed by this means, the telia of the present season only will be destroyed. The relative merits of each of the foregoing means of control for G. clavipes will vary with individual situations. Single or a combination of methods of control may be employed. Selective planting methods will doubtless give the most permanent results but are limited in their adapta- tion. Methods of eradication (3, 4 and 5) may be employed where the plants are of such high value, or are few, or the rust sufficient sparse as to make hand labor practical. Spray applications are perhaps most wide in applicability since the practice of spraying is so general. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 405 IX. SUMMARY 1. Inoculations with Gymnosporangium clavipes C. and P. and ex- aminations for infection were made on approximately seven hundred species and varieties in thirteen genera of pomaceous hosts. These gen- era were Amelanchier, Amelosorbus, Aronia, Chaenomeles, Crataego- mespilus, Crataegus, Cydonia, Malus, Photinia, Pyrus, Sorbaronia, Sorbopyrus and Sorbus. The genera Comptonia and M yrica as repre- sented in the Arnold Arboretum were also inoculated. The results show that hosts were distributed in eleven of these genera, namely, Amelanchier, Amelosorbus, Aronia, Chaenomeles, Crataegomes pilus, Crataegus, Cydonia, Malus, Photinia, Pyrus and Sorbus. Although pomaceous hosts of G. clavipes are found over the entire temperate region of the northern hemisphere, the fungus is confined to North America. 2. Investigations made on the period of susceptibility of flowers and fruits of certain pomaceous hosts showed that the flowers and fruits were susceptible after they were released from their buds for a brief period only. The more susceptible hosts were susceptible for a longer period than less susceptible hosts. 3. Inoculations and examinations for infection on the genus Juni- perus in the Arnold Arboretum and accounts in the literature showed that a total of eight species and varieties were susceptible to G. clavipes. Hosts were found in two sections of the genus Juniperus. These hosts occur over essentially the same geographical range as do the pomaceous hosts. 4. The disease caused by G. clavipes on pomaceous hosts was found to occur most frequently on fruits, less frequently on twigs and buds and but rarely on leaves. It was most severe on fruits, twigs and buds, usu- ally causing marked hyperplastic distortion. Infected buds were not only swollen but were forced to develop beyond the usual for the current season. On certain fruits, particularly varieties of orchard apples the disease produced was limited to small hypoplastic lesions usually at the blossom end. On leaves the disease was limited to small, usually partially necrotic, spots. 5. On its Juniperus hosts the disease was most abundant on twigs from one to five years old but was also found on leaves, branches and the main trunk. Diseased leaves were discolored and slightly swollen. They were usually killed in one or two years. The disease was perennial for several years on twigs and branches. They were usually girdled and cov- ered with a thick, flaky or furrowed blackened bark. On the main trunk the disease lived for many years but was usually confined to elongated swollen patches covered with deeply furrowed and blackened bark. 406 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI 6. The life history of the aecial phase of G. clavipes was essentially the same on fruits and twigs. It was slower in its development in leaves and in fruits of very resistant hosts, in fact in the latter it was often aborted. In forced buds the mycelium was essentially systemic and developed spermogonia, rarely aecia, progressively as the buds elongated. 7. The mycelium of the telial phase of G. clavipes was confined to the epidermis of leaves and to the phellogen of twigs, branches and the main trunk of its Juniperus hosts. It remained in leaves for but one, occasionally for two years. It was perennial for several years in twigs, branches and the main trunk. Telia were produced annually on infected organs. 8. Several means have been demonstrated for the control of G. clavipes on pomaceous and on Juniperus hosts. Especial attention has been given to finding satisfactory fungicides and formulating practical spray programs. Of the fungicides tested Linco colloidal sulfur gave very promising results. It was the only one used in field experimentation. 9. Recommendations with respect to the control of this rust have been discussed under the headings of: selective plantings, eradication of hosts, removing infected parts from pomaceous and Juniperus hosts, re- moving infections from trunks of red cedars and spray applications on pomaceous and Juniperus hosts. BIBLIOGRAPHY Apams, J. F. (1916). Internal Uredineae. (Mycologia, 8: 181-182.) a 19). Rusts on conifers in Pennsylvania —I. (Pennsylvania Agric. Exper. Sta. Bull. 160: 1-30. ——— (1921). Observations on the infection of Crataegus by Gymno- ihe (Mycologia, 13: 45-49. ArtuHour, J. C. (1901). Generic nomenclature of cedar apples. (Proc. Indiana Acad Sci. 1900: 131-136.) —— (1909). Cultures of Uredineae in 1908. (Mycologia, 1: (1929). Plant rusts. (John Wiley and Sons, New York, Manual of the rusts in United States and Canada -) Beacu, S. A. (1905). The apples of New York. (New York Agric. Exper. Sta. iy 1903, Vols. I and IT.) Cooke, M. C. and C. H. PrcK (1871). Note on Podisoma. (Jour. Quekett Micr. Club, 2: 255-268.) Coons, G. H. (1912). Some investigations of the cedar rust fungus, Gymnosporangium Juniperi-virginianae Schw. (Ann. Rep. Nebraska Agric. Exper. Sta. 25: 217-245.) 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 407 CraBiLL, C. H. (1913). Production of secondary sporidia by Gymno- sporangium. ielea 3: 282-284. CrowELL, I. H. (1934). The hosts, life history and control of the al apple rust fungus Gymnosporangium Juniperi-virginianae Schw. (Jou Arnold Arb. 15: 163-232. a (1935) Compilation of reports on the relative susceptibility of orchard varieties of apples to the cedar-apple rust disease. (Proc. Amer. Soc. Hort. Sci. for 1934. 32: 261-272.) —— i9 . Unusual developments of cedar rusts on pomaceous hosts in 1935. (Plant Disease ee 19: 197-198. DoncgE, B. O. (1918). Studies in the genus Gymnosporangiu t on the distribution of the mycelium, buffer oe and the Seca of the aecidiospores. (Brooklyn Bot. Gar d. Mem. 1: 128-140. ——— (191 Studies in the gen -one sporangium — IT. Rep on cultures made in 1915 and 1916, (Bull. Torrey Bot. Club, vee 287-300. ) (1918). Studies in the genus Gymnosporangium — III. The origin of the teleutospor re. (Mycologia, 10: 182-193. a Studies in the genus Gym mnospo rangium — IV. Distri- bution of the mycelium and the ae origin of the telium in G. clavipes. mer. Jour. Bot. 9: 35 ———. (19 Aecidiospore discharge as related to the character of the spore wall. (Jour. Agric. Res. 27: 749-756. ) —— (1924). Expulsion of aecidiospores by the May apple-rust. Puccinia podophylli Schw. (Jour. Agric. Res. 28: Sale — (1931 Studies in the genus Gymnosporangium A de- structive red cedar rust eee (Jour. New York Bee. ona 32: 101- ae 33). The orange rust of hawthorn and quince invades the trunk of red cedar. (Jour. New York Bot. Gard. 34: 233-237.) Faritow, W. 1880). The Gymnosporangia or cedar-apples of the United States. (Anniversary Memoirs Boston Soc. Nat. Hist. [Sci. Papers, 4], 2pl.) ee 1). Notes on Gymnosporangia. (Bull. Torrey Bot. Club, 8: 85-87.) (1885). Notes on some species of Gymnosporangium and are somyxa of the United States. (Proc. Amer. Acad. Arts and Sci. 20: 311-323.) Fuxusul, T. (192 5). Studies on the apple rust caused by haar oe eae gium Yamadae Miyabe. (Jour. Col. Agric. Hokkaido Imp. Univ : 269-307. Hockey, J. F. (1926). Report of the Dominion field laboratory of plant pathology, Kentville, Nova Scotia. (Rep. Dominion Botanist 1925, 3 Kern, F. D. (1910). The morphology of the peridial cells of the Roes- teliae. caer Gaz. 49: 445-452. 1911). A sae and taxonomic study of the genus Gymno- sporangium. (Bull. New York Bot. Gard. 7: 391-483.) MacLacuian, J. D. (1935). The hosts of Gymnosporangium globosum Farl. and their ‘relative susceptibility. (Jour. Arnold Arb. 16: 98-142.) Miter, P. R. (1932). The pathogenicity of three red cedar rusts that occur on apple. (Phytopath. 22: 723-740. Mitts, W. D. (1929). Cedar rusts, Gymnosporangium Juniperi-virgini- anae, G. globosum and G. germinale. (Plant Disease Reporter, Supp. 79: 188. ) 408 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Pappock, W. (1902). Quince rust. (Plant diseases of 1901. Colorado gric. Exper. Sta. Bull. 69: 18-20. PautMer, E. J. (1925). Synopsis of North American Crataegi. (Jour. PAMMEL, L. H. (1905). a apple fungi and apple rust in lowa. (lowa Agric. Exper. Sta. Bull. 84: 1-36. Fecr, C, H. (1873). ree of the oe (Ann. Rep. New York State Mus. Nat. Hist. 25[1871]: 57-123. Racan, W. H. (1905). Nomenclature of the apple. (U.S. Dept. Agric. Bull. 56, issued 1926, 1-390. ) Renper, A. (1927). Manual of cultivated trees and shrubs. (Macmillan an o., N ScHwWEINITz, L. D. (1832). ‘Synopsis fungorum in America Boreali media degentium. fpr a Phil. Soc. II. 4: 141-316. STEINMETZ, F. . T. HILBoRN (1 4). Observations on Gymno- sporangium germinal, (Phytopath. 24: 833-834. Stone, G. E. and R. E. Smirx (1898). The quince rust. (Rep. Massa- chusetts Agric. Exper. Sta. 10: 61-63.) STONE, R. E. (1909). Studies of Gymnosporangium in southern Alabama. , P. and H. (1915). Monographia papery aie ILI. TuHaxter, R. (1887). On certain cultures gee Nab pr vk with notes on their Roesteliae. (Proc. Amer. hee Arts and Sci. 22: 259-269. ) — (1889). Notes on cultures of Gymnosporangium made in 1887 and 1888. (Bot. Gaz. 14: 163-172. ———. (1891). The Connecticut species of See aa (cedar- apples). (Connecticut Agric. Ex xper. Sta. Bull. 107: Tuomas, H. 1933). The quince rust ee caused by Gymno- gation) germinale. (Phytopath. 23: 546-5 —_—_—_—_ MILLs (1929), Three rust diseases of the apple. (Cornell pene Exper. Sta. Mem. 123: 1-21. ————— ——— (1930). Rust ae of the apple. (Plant Disease Reporter, 14: 214-215.) Tuseur, K. von (1906). Uberwinterung des Birnenrostes auf dem Birn- baum. (Nat. Zeits. Land Forstw. 4: 150-1 ——— (1907). Perennieren des Aecidien mycels vom Birnenrostpilz. (Nat. Zeits. Land Forstw. 5: 217-219. Weimer, J. L. (1917). Three cedar rust fungi, their life histories and diseases they produce. (Cornell Agric. Exper. Sta. Bull. 390: 507-549. ) EXPLANATION OF THE PLATES PLATE 155! Fig. 1. Progressive stages are shown in the exposure and rupturing of aecial fructifications of Gymnosporangium clavipes on an orchar apple. Fig. 2. The dark colored lesion was caused by an imperfect fungus. Aecial Se ps aero of G. clavipes are shown on the upper portion of the les Fig. 3. Infected twig a fruit of Cydonia oblonga, the quince. 1Figures 1, 2 and 3 were obtained through the courtesy of Mr. K. A. Harrison of the Dominion Experimental Farms, Kentville, N. S. 1935] CROWELL, GYMNOSPORANGIUM CLAVIPES 409 Fig. 4. This shoot of Amelanchier oblongifolia was inoculated with teliospores of G. clavipes. Infection occurred, however, on the fruits and pedicels only. Fruits of Coe goa infected with G. clavipes. Fig. 6. Fruits of Malus floribunda infected with G. clavipes. PLaTE 156 Fig. 1. Twig and thorns of Crataegus sp. infected with G. clavipes. = ™ eal Fig. 2. Forced growth of infected buds of Crataegus mollis. The diseased buds on the twigs of this species have enlarged greatly. Compare with the foal tude on the twig shown on the extreme rig Fig. 3. Another eae of the forced growth of infected buds on Crataegus sp. On this species the buds and stems have enlarged but little vas the ee. though small, have taken on features of normal matur Fig. 4. Another ike atid of the forced growth of infected buds on Crataegus sp. On this host the buds have swollen considerably and the young stems have elongated but the leaves have remained quite stunted in their growt Fig. 5. Infected shoot of Crataegus Phaenopyrum. This type of symp- the tom was also found on t ‘english hawthorn, C. Oxyacantha. Fig. 6. A globose gall- ae swelling of an infected Gee of C. Oxyacantha. Fig. 7. mt Sata G. clavipes. Note its deep, sunken location and rotund forn PLATE 157 Fig. 1. Normal aeciospore chain of G. clavipes. gee the regular occur- Fig. 2. Abnormal aeciospore chain. The aeci Se are irregular in shape as well as in their arrangement with respect to the inter- calar ig. 3, utline drawings of aeciospores of G. clavipes Fig. 4. Camera lucida drawings of germinating aeciospores of G. clavipes. Fig. 5. Face and side views of ace cells of G. clavipes. Fig. 6. Outline drawings of peridial cells of G. clavipes Fig. 7. Camera lucida drawings of genre of G. clavipes as seen in the cells of fruits and twigs of various pomaceous hosts. All forms shown may be found in the same organ of any specific host. PLATE 158 Fully gelatinized telia of G. clavipes on branches ee red cedar. The bright red sori are very conspicuous during spring rain PLaTE 159 Fig. 1. ices lesion caused by G. clavipes on red cedar. Lesions are cally oval in Sine and the bark over them is darker in color than ee oer bark. ee the former location of a branch near the of the lesio Bie 2. Se trunk a een on a red cedar tree. Note that the trunk lesions are found in conn oe with lateral branches; many of these have died and been remo Figs. 3: ae of G. clavipes in their hee stages of gelatinization. At this time they are almost shapeless masses a ou — i) Ww nm + nN JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI . A slight swelling of the twig and a darker color of the bark are f re of the lesions on twigs of red cedars. These photographs ere taken in mid-winter. ; cca section of a telium of G. clavipes on J. virginiana. Note the pulvinate form of the sorus. PLATE 160 . Shows diagrammatically the location of the telia and the course of extension of the mycelium of G. c a over the phellogen of its host. This is a camera lucida drawin The mycelium of G. clavipes in an early stage of extension over the new phellogen that has recently a Aer with the existing one. Note that the mycelium is found on the phellogen only of its hos eee of the telial sorus. Buffer cells are in various stages of development. It will be observed ou also that the mycelium does not penetrate peanath the phellogen. Early stages in the development of corre iia The teliospore ne are elongating into the buffer ce . Stages in the maturation of teliospores. These e stages are se Ju to those or the maturation of teliospores of G. virginian pete in the phellogen cells of red cedar. They stand out clearly in prepared sections. Two and three haustoria are com- monly found in a single host cell. LABORATORY OF PLANT PATHOLOGY ARNOLD ARBORETUM, HARVARD i Jour. ARNoLD Ars. VoL. XVI Pirate 155 GY MNOSPORANGIUM CLAVIPES C. AND P. FULL-TONE~~ MERIDEN Jour. ARNOLD Ars. VoL. XVI PLATE 156 GY MNOSPORANGIUM CLAVIPES C. AND FP, FULL-TONE — MERIDEN Jour. ARNOLD Ars. VoL. XVI PLATE 157 GY MNOSPORANGIUM CLAVIPES C. AND P. FULL-TONE— MERIDEN Jour. Arwotp Ars. VoL. XVI Pirate 158 GYMNOSPORANGIUM CLAVIPES C. AND P. FULL-TONE — MERIDEN PLATE 159 Tour. ARNOLD ArB. VoL. XVI * t* Fhe. e idege / ‘et af 2. ® $s GYMNOSPORANGIUM CLAVIPES C. AND P. FULL-TONE — MERIDEN PLATE 160 Jour. ARNOLD Ars. VoL. XVI a mrrerenatic wy, +4 = Bas — om le 2 =. te we 4~ NA) iH, PN, re nr ’ . yeas f ie eto rR bse CLAVIPES C. AND P. FULL-TONE — MERIDEN GY MNOSPORANGIUM 1935] MacLACHLAN, SPORE DISPERSAL OF GYMNOSPORANGIUM 411 THE DISPERSAL OF VIABLE BASIDIOSPORES OF THE GYMNOSPORANGIUM RUSTS J. D. MacLacHLan With two text-figures INTRODUCTION THE BASIDIOSPORES of Gymnosporangium rusts are thin-walled and so are subject to rapid dessication. It would seem then that their effective range would be limited to a mile or less but instances have come to the attention of the writer which indicate that, under certain circumstances, the range may be as great as seven or eight miles. Three of such in- stances may be cited. The first was an observation made by Professor Roland Thaxter (1887) on basidiospores of G. biseptatum Ellis. He found infection on an island off the coast of Maine, eight miles distant from the only known source of inoculum. The second instance occurred at Lake Wentworth, New Hampshire, and was brought to the attention of the writer in the summer of 1933 by Mr. L. S. Mayo. In this locality infection by G. clavipes Cke. & Pk. and G. globosum Farl. on Amelan- chier and Crataegus, respectively, was found on the northeastern side of the lake, while the nearest source of inoculum was from a stand of cedars on the opposite side of the same lake, more than six miles distant. The prevailing winds were from the southwest and had presumably carried the basidiospores across the lake. The third instance occurred in the Arnold Arboretum, where infection by G. Juniperi-virginianae Schw. and G. globosum Farl. on species of Malus and Crataegus, respectively, has been observed during the past three years. On both hosts the infection, while slight, was markedly uniform over the entire plantations, indicat- ing that the sources of inoculum were considerably removed. A detailed scrutiny of the surrounding country revealed that no source of inoculum sufficient to produce this infection existed within a radius of more than six or seven miles. In an effort to explain the occurrence of instances such as those that have been described, an airplane collection of basidiospores was made at different altitudes over infected cedar areas immediately following a rainy period in May, 1934; this was followed by laboratory tests on the duration of the viability of basidiospores of G. Juniperi-virginianae when subjected to varying temperature and humidity conditions. 412 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI AERIAL DISTRIBUTION OF BASIDIOSPORES AS INDICATED BY AN AIRPLANE COLLECTION Two areas were selected in which an abundance of basidiospores were being discharged, namely, Waltham, Massachusetts, where heavy but restricted infections by G. Juniperi-virginianae and G. globosum existed, the other about seven miles distant from Waltham at Cochituate where there were approximately three acres of red cedars among which were scattered apple trees highly susceptible to G. Juniperi-virginianae ; prac- tically every cedar tree in the latter area was heavily loaded with galls of G. Juniperi-virginianae. On May 5, 1934, ideal conditions prevailed for making such a collec- tion. Prior to this date there had been intermittent rain, fog, and sat- urated humidity for approximately sixty hours; examination of the in- fected cedars showed that basidiospores were being released almost con- tinually during this period and no high winds had occurred that would scatter the spores beyond the possibility of collection. As the plates re- vealed after the flight, the air had been washed relatively free of dust and smoke so that little trouble was experienced from such contaminations. At the time of flight, namely, 8.30 A. M., the clouds had just broken but the air was still saturated with moisture, evidenced by the fact that as we left the airport water was precipitated out of the air by the propellor and thrown on the windshield of the plane. It happened that a south- west wind was blowing directly over the area at Cochituate towards the Waltham area. Thus, optimum conditions existed for making the collection. The spores were collected on petri-plates, each plate containing a thin layer of 2% pure bacto-agar. To prevent growths by other fungi no nutrients were added. Previous tests showed that the basidiospore would germinate freely on the agar alone and since they do not grow on artificial media, germination was taken as sufficient evidence that the basidiospores were viable. Crude but efficient holders for the plates were made by using circular lids, whose diameters were slightly greater than that of a petri-plate, attached to the ends of wooden rods two and one- half feet in length. Wire clamps that could be opened readily served to hold the plates in place while the exposures were being made. Continuous exposures of one minute each, one overlapping the next by fifteen seconds, were made by removing the cover, clamping the open plate in the holder and immediately thrusting the latter out through the window the length of the handle so that the holder rested on a wind strut. Alternate exposures were made from each side of the plane on numbered plates. At the same time the course of the flight was outlined on a map 1935] MacLACHLAN, SPORE DISPERSAL OF GYMNOSPORANGIUM 413 and by marking the position of the plane at the end of each minute it was not difficult to determine afterwards the exact area over which the plate had been exposed. Exposures were made from within about one mile of the first source of spores (the Waltham area) three times at an altitude of 200 feet, then directly into the wind at an altitude of 500 feet for seven miles to the Cochituate area. Over the latter area four flights were made at 100 ft. altitude, then two flights each at altitudes of 500 ft., 1000 ft., 1500 ft., and 2000 ft., respectively, finally directly with the wind at an altitude of 2000 ft. for ten miles, bringing us back again over the Waltham area. The average speed during the flight was about 85 miles per hour. Figure 1 illustrates the course of the flight. 2000 FT. 43 2000FT.- — a +s att a O) 1SOO FT. SCALE 2000 FT. SOO FT. 2ai2) 7 9 8 7 6 200 FT 1707) 2216) 10 (8) i) O) a) fa) D—— 100 FT. 420 13 .--SEA LEVEL COcHITUATE LIGHT SCATTERED INFECTION WALTHAM INFECTION INFECTION SCALE SMILES Ficure 1. Illustration of the course of the flight over the areas of infected cedars : Each line represents a single exposure; the chp? in ioimpewniy below the fie number, indicates the number of spores caught on 20 sq. of the plate. The topography of the land is shown in profile (from U. S. Ccslned ESinees nei, Five hours after the flight the plates were placed in the refrigerator at 0° C. and spore counts begun. An area enclosing twenty square centi- meters was ruled off on the lower side of each plate prior to making the count. Within the area fine parallel lines were ruled, the distance be- tween any two adjacent lines being slightly less than twice the diameter of the field covered by the low power of the microscope. The plates were examined from the upper side (lids removed). By moving the plate on the microscope stage so that a line (visible though out of focus) was just perceptible on one side of the field, and returning across the plate so that the next line was just in view on the other side of the field, the area enclosed was completely covered without the possibility of either overlapping or omitting any of the plate surface to be examined. The basidiospores were distinguished from other spores caught on the plates by (1) the characteristic yellowish color of their protoplasm as seen under strong light, (2) the size and shape of the spores and (3) the characteristic germ tubes when present. As a means of comparison 414 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI basidiospores of G. Juniperi-virginianae and of G. globosum were allowed to drop on similarly prepared plates from gelatinized telial masses; these two preparations were used for constant reference during the examina- tion of the plates. Table I gives, for the respective plates, the region over which each collection was made, the plate number, the number of spores found on twenty square centimeters of the plate and finally the number of spores exhibiting germ tubes. TABLE I DATA OBTAINED BY MEANS OF AN AIRPLANE COLLECTION OF BASIDIOSPORES OF THE GYMNOSPORANGIUM RUSTS Altitude Plate Total No. spores Region in ft. No spores germinated Over Waltham 200 1 7 4 infection 2 4 1 3 2 1 Between Waltham 500 4 0 0 and Cochituate 5 4 2 infections 6 0 0 rg 9 4 8 3 0 9 8 2 10 10 5 11 18 8 Over Cochituate 100 12 12 3 infection 13 29 9 15 20 10 17 i? 5 500 22 8 3 24 12 5 1000 27 2 0) 29 2 0) 1500 33 (0) 0 35 0 0 2000 40 0 0 42 0 0 Between 2000 43 0 0 Cochituate 44 3 1 and Waltham 45 4 1 infections 46 1 0 47 2 0 48 0 0 49 0 0 50 i 1 1 0 0 : Total spores identified — 182 Total spores germinated — 65 1935] MacLACHLAN, SPORE DISPERSAL OF GYMNOSPORANGIUM 415 The count included only those basidiospores of unquestionable iden- tity. Spores that were embedded in the agar from the force of impact and could not be identified accurately due to the air spaces around them, as well as the spores which had burst when they hit the agar, were not considered. Certain plates exposed during turning or gaining altitude over the Cochituate area were omitted due to the difficulty in determin- ing the exact location over which these plates were exposed (see Fig. 1). As may be seen from Table I or Fig. 1, basidiospores were picked up almost continuously during the entire flight. Approximately one-third of them exhibited germ tubes at the time of examination. A sharp in- crease in the number of spores collected was evident as the large area at Cochituate was approached. Over this area there was a rapid decrease in the number of spores collected with increase in altitude, no spores being collected at 1500 ft. altitude; this may be accounted for by the fact that the wind was carrying the spores away from the immediate cedar area below a ceiling of 500 to 1000 ft. It is interesting to note, however, that spores were again picked up on the returr flight (going with the wind) at an altitude of 2000 ft., a mile or so from this area (Fig. 1). The number of spores picked up over the Waltham area was much smaller than that over the Cochituate area; this coincided with a smaller amount of infection in the former area. Before drawing any conclusion from the data obtained as to how far viable basidiospores may travel, the area over which the flight was taken was carefully examined to determine how much infection existed between the two main areas of infection. The survey revealed that scattered in- fection loci existed within less than one-half mile of each other over the entire area. Thus, although viable basidiospores were found over the entire flight it was impossible to determine the exact location of their source of dissemination and consequently the distance they had travelled. DURATION OF THE VIABILITY OF BASIDIOSPORES OF G. JUNIPERI-VIRGINIANAE WITH RELATION TO THE FACTORS OF TIME, TEMPERATURE AND A knowledge of the length of time during which the basidiospores will live when they are subjected to varying temperature and humidity con- ditions should, indirectly, give an indication of the distance that the basidiospores may travel and still have the potential ability to produce infection. Reed and Crabill (1915) state that five to ten days is the life limit of basidiospores of G. Juniperi-virginianae in an air-dry condition. They also state that in direct sunlight the spores are killed within two to five hours; sunlight, then, may be a limiting factor in determining the 416 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XVI longevity of the basidiospores. However, as has been stated previously, dispersal takes place during rainy weather, under which circumstances the basidiospores might travel many miles under the cloud line. Laboratory tests were conducted in which fresh basidiospores were subjected to temperatures ranging from 0° to 35° C. at five degree inter- vals and humidities ranging from 0% to 100% at twenty-five percent intervals. After varying lengths of time samples of the spores were re- moved from the respective environmental combinations and tested for germination. ) FIGURE of the viability of basidiospores. (Explanation in text. ) The type of humidity chamber used in testing the duration Such an experiment required eight constant temperature chambers. Refrigerators kept in a warm room and equipped with thermostats that kept the temperature constant within 0.5° C. served for the O, 5, and 10 degree chambers; temperatures of 15, 20, 25, 30 and 35 degrees were maintained in De Khotinsky ovens kept in a cold room at 4° For humidity chambers Kolle culture flasks proved to be very ethiadi ; they were of convenient size, exposed a large surface of the humidity con- trolling agent with respect to the volume of the flask and could be opened without altering the humidity within the flask to any great extent. Figure 2 illustrates a typical chamber. 1935] MacLACHLAN, SPORE DISPERSAL OF GYMNOSPORANGIUM 417 Zero percent humidity was obtained by placing a layer of dry CaCl, in the bottom of the flask. Humidities of 25%, 50% and 75% were obtained by using aqueous solutions containing, respectively, 55.1%, 42.8% and 29.9% of H.SO,. These values were taken from data given by Wilson (1921) and are approximately half way between the values given for the respective humidities at temperatures of 0° C. and ri eget Oe The greatest difference in percentage of H,SO, to maintain any of these humidities over a temperature range of 0° C. to 25° C. is 1.6, which would be insignificant when considering such wide humidity intervals. Enough H,SO, solution was poured into each flask to cover the bottom to a depth of about one centimeter. The lip between the flask proper and the neck served to prevent the H,SO, solution from running out. The mouths of the flasks were closed with covers made by cutting gum- rubber tubing of the appropriate diameter into four-inch lengths and sealing one end with rubber cement. Such flasks could not be used for 100% humidity as water of condensation would form inside the flask on the flat upper surface and drop on the spores. To eliminate this diffi- culty, small dome-shaped bell-jars were placed inside moist chambers containing distilled water; this type of chamber allowed any water of condensation to run down the sides of the dome. Small trays of a size that two would conveniently slip into each flask and hold the cover glasses bearing the spores well above the humidity controlling agents were made out of wire screening. These were coated with paraffin (melting point 56° to 58° C.) to prevent the water or H,SO, from coming in contact with the spores as well as prevent corrosion of the trays by the H,SQ,. A complete set of humidity chambers from 0% to 100% was placed in each of the respective temperature chambers twenty-four hours before the experiment started. In this way the temperature and the humidity within the flasks were brought to a definite equilibrium before the spores were introduced. Telial material of G. Juniperi-virginianae was collected in the field, re- moved from the galls and soaked in water to form a thick gelatinous paste. This was smeared over the tops of the inside of large glass moist chambers, which were then set over clean cover glasses arranged in solid squares. Approximately five hundred cover glasses were placed under each chamber. The teliospores germinated and released an abundance of basidiospores, which fell on the cover glasses in a very uniform layer. Excess water was avoided in preparing the telial smears to prevent any condensation on the cover glasses; otherwise the basidiospores would germinate immediately. Ten hours after this experiment was set up 418 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. Xvi the cover glasses bearing the basidiospores were removed and immedi- ately placed on the trays within the series of humidity flasks in the re- spective temperature chambers. Fifteen cover glasses with the spores on the upper surfaces were placed in each unit. It is essential that the cover glasses are not inverted, as water of condensation in the 100% humidity chambers would form on the spore surfaces; no trouble was experienced from such condensation when the cover glasses were arranged so that the spores were on the upper surfaces. At the same time three cover glasses bearing fresh basidiospores were inverted over Van Tieghem cells which were placed in a petri-plate lined with wet filter paper. Sufficient water of condensation formed on the lower (spore) surfaces of the cover glasses for optimum germination; the addition of more water causes irregularity in the rate of germination. This culture chamber was then placed in a temperature chamber main- tained at 18° C. and the percentage germination was determined twenty- four hours later by counting five hundred spores on each cover glass. This count gave the percentage germination when the basidiospores were fresh and was used as the basis of comparison for all succeeding counts. At intervals, given in Table II, one cover glass was removed from each unit of the complete series, set up for germination as described above, and the percentage germination of five hundred spores on each cover glass determined twenty-four hours later. By this experiment the relative effect of the three factors, temperature, humidity and time on the potential viability of the basidiospores of G. Juniperi-virginianae was determined. Table II presents the data ob- tained. Analysis of these data reveals certain significant facts: (1) The basidiospores are killed at 0% humidity at all temperatures within at least twenty-two hours time. Practically speaking this is of no importance because such an environment is never attained in the field. (2) Above a temperature of 30° C. the basidiospores died within twenty-two hours under all humidity conditions. This is also of little practical significance as such temperatures are rarely if ever attained during wet periods in early May. (3) Above a humidity of 25% and below a temperature of 25° C. the spores could live a sufficient length of time to blow many miles and still remain viable; such humidity and temperature conditions prevail during the normal dispersal of basidiospores of the Gymnosporangium rusts. Below a temperature of 10° C., as far as humidity and tempera- ture alone are concerned, the spores can remain viable for more than a 1935] MacLACHLAN, SPORE DISPERSAL OF GYMNOSPORANGIUM 419 TABLE II DATA ON THE PERCENTAGE GERMINATION OF BASIDIOSPORES OF G: Sen Ra nae ca THAT MAY ES HAVE EN SUBJECTED FOR VARYING LEN BE OBTAINED AFTER THE GTHS OF OT ME TO DIFFERENT TEMPERATURE AND HUMIDITY CONDITIONS TEMPERATURE of OC. % ea desi after penereotaon o humiditie of 25% 50k 75k 1008 0.0 73 .C 82.2 81.6 70.6 0.0 64.€ 78.0 78.0 57.4 0.0 53. 73.4 74.2 74,2 - 50. 72.4 59.8 60,0 - 47. 53.6 56.0 42.4 = 44, 51.4 50.6 44.2 - oO. 53.0 42.4 42.8 - . 0 37.4 lost - 0. 20 0.0 0.9 - oO. 0 0.0 0.0 TEMPERATURE of 10 C. % bE shy cei after peaiesysen o humiditie of 25k sok 75% 008 0.0 83.6 88. 88.2 88. 0.0 72.8 78. 82.4 19, 0.0 lost 66. 78.4 80. - lost 56. §2.2 70. ~ 58.2 62, 60. 13. - 25.2 55. 44.4 71.6 ° Fie) 7. 1.6 53. - 0.0 10. 5.2 59. - 0.0 1. 8 22. - - 0. 0.0 0. TEMPERATURE of 20 C. % germination after subjection to humidities of og 25% sox 75% 100g 0.0 82.0 79.6 76.4 90.6 0.0 46.8 60,2 72,2 84.6 0.0 43.4 3.0 682.2 62.4 - 0.0 0.0 0.0 1.6 - 0.0 0.0 0.0 ie) - 0.0 0.0 0.0 2.0 - - - - -0 Lal - 7. = 0 - - - - 0 TEMPERATURE of 30 C. % germination after msec to humiditie of 2sf sok 78K 00g 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0,0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 TEMPERATURE of 5 C. % aermoeta after appdacnion o humidities of 25% 50g 75% 008 0.0 82.6 88. 89, 81.2 0.0 80,2 86. 84, 80.4 0.0 73.4 85.6 719. 79.6 - 64.0 60.€ 6l. 45.8 - 61.6 Ble 63.% 50.2 - lost 59. 61.€ 52.8 - 0.0 52 59.€ 52.4 - 0.0 26. 36, 53.0 - 0,0 52.4 41,4 lost - - One 28 .€ 45.0 TEMPERATURE of 15 C. % coratevacD after Ean section umiditie of 25g = SOR 75K 1008 0.0 lost lost 76.6 80. 0.0 3.2 21042 76.8 83. 0.0 0.0 0,0 4.6 2. - 0.90 0,0 0.0 §5. - 0.0 0.0 0.0 4. - - - 0.0 24 - - - - O. - - - - QO. ad = - @ ¢ e TEMPERATURE of 25 C. % Rereyneticn after ee eaenet pe to miditie O% 25% 50% 75 100% 0.0 5.2 4.8 4.8 0.0 0.0 0.0 O20-8-.76.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 TEMPERATURE of 35 C. % germination after paubyection to humiditie 7.3 25% 50% 75% 100% 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 420 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI week; while at high humidities and a temperature of 5° C. they can live for more than twenty-five days. Whether the basidiospores of other Gymnosporangium rusts have the same potential viability exhibited by G. Juniperi-virginianae in this ex- periment is not known. Such may very well occur. CONCLUSION The data obtained from the airplane collection on the aérial distri- bution of the basidiospores and from the laboratory tests on the longevity of basidiospores of G. Juniperi-virginianae afford a possible explanation of the occurrences, described at the beginning of this presentation, of infection several miles from the source of inoculum. The infection that occurred in the first two instances described may have been facilitated by the fact that the basidiospores had an unobstructed passage over water where the humidity was relatively high and the temperature low- ered to some extent. BASIDIOSPORE DISPERSAL AND THE PRACTICE OF CEDAR ERADICATION A tremendous number of basidiospores may be released from a large stand of infected cedars; moreover, as has been shown in this presenta- tion the basidiospores, when subjected to the environmental conditions that are met during their dispersal, have the ability to live for a suffi- cient length of time to be carried many miles. The question then arises as to why effective control of the cedar-apple rust has been repeatedly obtained by the removal of the red cedar within a radius of one to two miles from the pomaceous host. An explanation of this question may be afforded by certain factors that are necessarily involved before injurious infection of the pomaceous host can take place: (1) The rapidity with which the area of any circle described around a locus of infection increases with increase in radius would dilute the spore population correspondingly; the area enclosed within a radius of one mile involves an area of more than eighty-seven and one-half million square feet, while a radius of ten miles would involve an area of one hundred times that or between eight and one-half and nine billion square feet. (2) The infection resulting from a basidiospore is not systemic in the host but is restricted to a small area on a single leaf, fruit or twig. Several lesions per leaf are required to materially injure the tree and 1935] MacLACHLAN, SPORE DISPERSAL OF GYMNOSPORANGIUM 421 when one considers the hundreds of thousands of leaves on a single large tree, a tremendous number of basidiospores must fall within the area occupied by that tree and successfully produce infection before the foliage is materially injured. (3) So far as is known, aeciospores cannot reinfect the same host; consequently, the amount of infection on any one tree is limited to the number of lesions originating from the basidiospore infections in the spring. It is very fortunate that such a short cycle does not exist; other- wise it would be very difficult to grow a susceptible host in any region where these rusts are present. (4) One cannot assume that every basidiospore that alights on a host tree will produce infection; many must die before the germ tubes can penetrate the proper host tissue. When the factors outlined above are fitted into the picture one can readily see that, in spite of the tremendous output of basidiospores and the length of time that they are able to live under the conditions that are met during their dispersal, eradication of the red cedar within a radius of one to two miles would ordinarily be sufficient to amply protect the alter- nate host from any injurious infection. SUMMARY Instances have been recorded to show that the basidiospores of at least certain of the Gymnosporangium rusts can produce infection on alter- nate hosts that are removed from the source of inoculum by several miles. Attempts were made to explain the occurrence of such instances; the investigations included an airplane collection of basidiospores over in- fected cedar areas as well as laboratory tests on the duration of the via- bility of basidiospores of G. Juniperi-virginianae Schw. with respect to the factors of time, temperature and humidity. The results of these investigations revealed that viable basidiospores are present in the air during rainy periods in early May at altitudes of at least 2000 feet and that basidiospores of G. Juniperi-virginianae can live for many days under the environmental conditions that prevail dur- ing their normal dispersal. These results give a possible explanation of the occurrence of infection on pomaceous hosts that are removed by several miles from the source of inoculum. Experience has shown that eradication of the red cedar within radii of one to two miles will, ordinarily, amply protect the pomaceous host from injurious infection by the Gymnosporangium rusts. Certain of the fac- tors which make possible the efficiency of this means of control have been presented. 422 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI ACKNOWLEDGMENTS To Professor J. H. Faull for guidance and supervision during the pursuance of this investigation; to Dr. A. E. Navez for helpful sugges- tions in the compilation of data and to the Department of Biology, Har- vard University, for laboratory facilities and financial assistance, the writer expresses sincere gratitude. The writer also wishes to thank Mr. L. S. Mayo, Assistant Dean of the Graduate School of Arts and Sciences, Harvard University, for bringing to attention the locus of infection at Lake Wentworth, and Drs. Ivan H. Crowell and P. A. Vestal for assistance in the airplane collection of spores. BIBLIOGRAPHY Reep, H. S. and C. H. Crasity (1915). The cedar rust diseases of apples caused by Gymnosporangium Juniperi-virginianae Schw. (Va. Agric. Exper. Sta. Bull. 9: 32, 33. THAXTER, R. (1887). VII. On certain coulda of Gymnosporangium with notes on their hae (Proc. Amer. Acad. Sci. 22: 259-269.) Witson, R. E. (1921). Humidity ae ‘by means of sulfuric acid solu- tions, with critical ae ena of vapor pressure data. (Jour. Indust. Engin. Chem. 13: 326-331.) LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1935] SAX AND SAX, CHROMOSOME STRUCTURE AND BEHAVIOR 423 CHROMOSOME STRUCTURE AND BEHAVIOR IN MITOSIS AND MEIOSIS HALLy JoLIVETTE SAX AND Karu SAx With plates 161-164 A stupy of chromosome structure and behavior at mitosis and meiosis has been made in order to compare the two types of divisions and to aid in the analysis of the mechanism of meiosis. This work is based on a comparison of chromosome lengths at different stages in the mitotic and meiotic cycles, and the relation of these changes to the internal structure of the chromosomes. The chromosome cycle in mitosis and meiosis has been studied in Tradescantia paludosa, Vicia faba, Lilium regale, and in Allium Cepa. The length of the chromosomes at various stages was also obtained in somatic cells of Trillium grandiflorum, and some work was done on the meiotic divisions in Secale cereale and in Zea mays. Recent advances in cytological technique have made possible a fairly accurate study of the length and structure of the chromosomes at various stages in the mitotic and meiotic cycles. The meiotic figures were obtained from microsporocytes which were smeared on a dry slide, pretreated with 30 percent alcohol containing about six drops of ammonia water per 50 cc., and fixed and stained with aceto-carmine, or fixed in Flemming’s solutions and stained with crystal violet iodine. The best preparations of somatic divisions were obtained from young microspores. After smearing on a dry slide they were pre- treated with the alcohol ammonia for about a minute and then fixed either in aceto-carmine or Flemming’s solution. Root tips were fixed for 12 to 15 hours in a mixture of absolute alcohol (70 cc.) and glacial acetic acid (30 cc.) and then macerated in a drop of aceto-carmine. In all cases the aceto-carmine smears were heated to clear the cytoplasm and the cover glass pressed to flatten the cells. The preparations were then sealed or made permanent by McClintock’s method. The aceto- carmine preparations showed almost as much detail of structure as those fixed in Flemming’s solutign, and since the cells fixed in aceto-carmine could be flattened, these preparations were used in measuring chromo- some lengths and were photographed to illustrate the various stages in mitotic and meiotic cycles. 424 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI THE MITOTIC CYCLE IN TRADESCANTIA MICROSPORES At anaphase of the second meiotic division each chromosome consists of two spiral chromatids. When the chromosomes pass into the resting stage the chromonemata tend to uncoil and form a loose spiral structure which completely fills the nucleus. Well fixed preparations show distinct chromatic threads loosely coiled in the resting nucleus rather than a reticulate network (Plate 161, photo 1). At early pro- phase the nucleus enlarges and the spiral chromosomes are more easily observed (Plate 161, photo 2). The chromatids of each chromosome are so clearly associated that the doubleness is hardly discernible, a con- dition also observed by Kuwada and Nakamura (1935). As the pro- phase continues the coils tend to straighten out and at the same time there is evidence of a new coiling in the closely associated chromatids (Plate 161, photo 3). At this stage there is little evidence of the double thread structure of the chromosomes even though they appear to be two-parted at the preceding anaphase. When the old coils are straight- ened out so that the spireme thread can be followed at all loci, there is clear evidence of a longitudinal split and each chromatid is independently coiled (Plate 161, photo. 4). These new coils apparently shorten the chromosomes and draw out the old coils persisting from the previous anaphase of the second meiotic division. The chromosomes continue to shorten during anaphase and the chromatids become thicker and more clearly separated (Plate 161, photo. 5). There is some tendency for the two chromatids of a chromosome to be twisted about each other, but at most only two or three twists occur. These are usually eliminated by metaphase although overlaps and an occasional twist is found at this stage (Plate 161, photo. 7). In preparations fixed without pretreatment, there is little or no evi- dence of the coiled chromosome structure at metaphase (Plate 161, photo. 6). The two chromatids of each chromosome usually can be identified although with certain types of fixation and staining the meta- phase chromosomes appear as single rods. After effective pretreatment of the microspores the coiled chromatids can be seen at early metaphase and at anaphase (Plate 161, photos. 7, 8,9). The diameter of the chro- matid is so near the limit of microscopic resolution that it has not been possible to determine the direction of coiling, nor can the number of coils be determined accurately, but there appear to be about twenty-five coils in each chromatid (Plate 161, photo. 8). When the chromatids are twisted about each other at early metaphase the chromosome appears to be constricted at the point of overlap as is shown in the chromosome at the right in photo. 7 of Plate 161. 1935] SAX AND SAX, CHROMOSOME STRUCTURE AND BEHAVIOR 425 Each chromatid at metaphase and at anaphase usually appears to con- sist of a single coil, but there is some evidence that these chromatids con- tain two threads which are coiled together. In the first meiotic division the major coils are so closely associated that they appear as a single coil unless lightly stained, but at late metaphase the two coils separate and lie parallel. In the somatic chromosomes the two coiled threads in the anaphase chromosome do not separate enough to appear as two parallel coils, and the diameter of the chromatids is too small to permit the direct observation of two chromatic threads coiled together, but there is appar- ently a tendency for the two coils to separate so that when a twist occurs in an anaphase chromosome there is a constricted locus at the twist (Plate 161, photo. 9). Such constrictions may be observed even in the chroma- tids at early metaphase (Plate 161, photo. 8). The chromosomes at late telophase appear so compact that little detail in structure can be observed, but as they elongate at later stages the coiled chromonemata expand and irregular coils and corrugations may be observed. The chromaticity of the chromosomes is reduced so that it is not possible to follow the coiling in any single chromosome and the entire nucleus is filled with loosely coiled chromonemata in the resting stage. The somatic divisions observed in aceto-carmine preparations of root tips did not show the detail of the structure found in the microspores, but the general behavior is the same, except that the root tip chromo- somes are longer than those of the microspore at the metaphase stage of division. THE MITOTIC AND MEIOTIC CYCLES IN VICIA FABA The early prophase stages in root tips of Vicia faba show the irregular spiral chromonemata. At this stage the chromatic threads appear to be single at most loci. As the spireme threads tend to straighten their dual nature is easily observed at all points (Plate 162, photo. 1). The chroma- tids are twisted about each other to a greater extent than is found in Tradescantia, and as many as five or six twists may be observed in a single chromosome. ‘The chromatids appear to be independently coiled in small loose spirals at this stage. During later development the chro- matids thicken and shorten until metaphase, but we have been unable to observe the internal structure at this stage. The anaphase chromo- somes seem to show a double spiral structure (Plate 162, photo. 2), but not as clearly as in the figures published by Sharp (1929). The prophase stages in the microspore nucleus are more difficult to follow, presumably because of the rather thick wall of the microspore, 426 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI and the metaphase and anaphase stages show little or no detail of struc- ture (Plate 162, photos. 6, 7,8). The lengths of the prophase spireme and of the metaphase and anaphase chromosomes can be obtained, and it was found that the microspore chromosomes are shorter than those of the root tip cells at metaphase and anaphase. The early meiotic stages were not studied in detail, but measurements were made of the pachytene spireme. There is a great reduction in chro- mosome length between pachytene and the first meiotic metaphase. The meiotic chromosomes at the first meiotic metaphase are even shorter than the metaphase chromosomes of the microspore mitosis. The long ‘‘m”’ chromosome has an average chiasma frequency of about 6 while the average chiasma frequency of the short “mM” chromosomes is about 3 (Plate 162, photo. 3). The chromonemata are coiled in major spirals at metaphase and at anaphase (Plate 162, photo. 4). A description of these coils will be presented in a later paper. At late anaphase the meiotic chromosomes contract considerably (Plate 162, photo. 5). During the second meiotic division the major coils may persist, but frequently they are completely eliminated at this time and the chromosomes at anaphase appear as straight rods. THE MITOTIC AND MEIOTIC CYCLE IN LILIUM REGALE The root tip preparations of L. regale showed only the more general features of chromosome behavior. The prophase and metaphase stages were clear enough to provide measurements of the mitotic chromosomes (Plate 163, photos. 3 and 4). The pachytene stages of meiosis showed the association of chromo- meres as described by Belling and others. The pachytene chromosomes are much thinner and longer than the chromosomes of the ‘‘spireme”’ in root tip cells (Plate 163, photo. 1). The chromosomes of the first meiotic division are much shorter than the somatic chromosomes (Plate 163, photo. 2). There is clear evidence of major coils in these meiotic chromo- somes and the average chiasma frequency is about 3 per bivalent. The microspores did not provide good preparations for prophase stages, but Dr. W. S. Flory obtained metaphase figures in another species which could be measured. OBSERVATIONS OF CHROMOSOMES IN TRILLIUM, ZEA, ALLIUM AND SECALE Root tip preparations of Trillium grandiflorum provided prophase and metaphase figures which could be measured for comparison with cor- responding stages in other genera. The prophase spireme in somatic cells is not so clearly split as is the case in Tradescantia and Vicia (Plate 1935] SAX AND SAX, CHROMOSOME STRUCTURE AND BEHAVIOR 427 163, photo. 5). The contraction of the chromosomes from the prophase spireme to metaphase is less than it is in the other genera which have been studied, and the metaphase chromosomes are very long (Plate 163, photo. 6). There is some twisting of sister chromatids about each other even at metaphase. We have made no detailed study of Zea chromosomes, but McClin- tock’s figures (1933) show about an 11 to 1 reduction in length between pachytene and the first meiotic metaphase, and according to McClintock (personal communication) the ratio may be as great as 15 to 1. The meiotic cycle in Allium Cepa is especially clear for a study of chromosome contraction from pachytene to metaphase (Plate 164, photos. 3-6). he association of chromomeres can be observed at pachytene and the number of nodes is greatly reduced from early diplo- tene to metaphase. Most of these points of contact seem to be twists or overlaps. A few measurements of mitotic and meiotic chromosomes were ob- tained from Secale cereale. The structure of the meiotic chromosomes has been described in some detail in an earlier paper (Sax, 1930). CHROMOSOME LENGTH AT VARIOUS STAGES IN MITOSIS AND MEIOSIS We have obtained measurements of chromosome lengths at prophase and metaphase in mitotic and meiotic cells of the various species exam- ined. The prophase measurements of root tip cells were made after the old coils were straightened out and the new coils were started, the so- called spireme stage of mitosis. The cells were flattened so that most of the spireme could be drawn in two focal levels. The measurements of the chromosomes were made from camera lucida drawings, and no attempt was made to determine the additional length caused by fore- shortening of threads passing through several focal levels. The lengths of the pachytene chromosomes were easier to obtain, but even these are only approximate. The meiotic chromosomes at metaphase form loops between chiasmata and we have tried to include these in our measure- ments of Vicia and Lilium chromosomes. The anaphase chromosomes in both mitosis and meiosis are essentially the same length as the meta- phase chromosomes in some species so that anaphase figures were occa- sionally included in determinations of metaphase lengths. In view of the technical difficulties involved in determining comparable stages and in obtaining the prophase measurements, the results are only approxi- mate, but the differences in chromosome contraction in mitosis and meiosis are so consistent that they must be of some significance. The data obtained are shown in Table I. 428 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI TABLE I Average chromosome length in microns at prophase (P) and metaphase (M) in meiotic and somatic divisions. The number (n) of cells measured is indicated. Root tip Meiotic Microspore Species n P nM n PnM n P nM Vicia faba 4 48 5 13 2 98 2 9 4 36 2 Ill Tradescantia sp. 5 56 8 21 3 81 9 9 8 61 8 12 Lilium regale 3 35.2. 22 2 8 4 1 1 15 Trillium grandiflorum 1 91 2 40 Secale cereale 1 37 1 14 1 61 1 8 Allium Cepa 4 69 1 9 In every case where meiotic and mitotic prophases are compared the meiotic pachytene chromosomes are much longer than those at the somatic prophase spireme. The ratios range from about 1.4:1 in Trade- scantia to about 2.4:1 in Lilium and the average ratio for all species examined is about 2:1. The reduction in chromosome length from prophase to metaphase is much greater in meiosis than in mitosis. The chromosomes at pachy- tene are from 7 to 11 times as long as the chromosomes at meiotic meta- phase and the ratio may be even more extreme in certain species. In root tip cells the prophase chromosomes are shorter, but the metaphase chromosomes are longer than the corresponding stages of meiosis. Con- sequently the reduction in chromosome length from prophase to meta- phase is much less in root tip cells, ranging from less than 2:1 in Lilium to about 4:1 in Vicia. The metaphase chromosomes of the microspore are shorter than those of the root tip cells, but longer than the meiotic chromosomes at first metaphase. The technical difficulties in measuring microspore chromo- somes probably is responsible for the shorter prophase measurements in microspores of Vicia as compared with corresponding stages in root tip cells. The outstanding feature of these comparisons in chromosome length is the consistent and striking difference in the degree of chromosome contraction in mitosis and meiosis. For the species examined the average degree of chromosome contraction between pachytene and the first meiotic metaphase is about 8.6:1, while for comparable stages in mitosis in root tip cells the ratio is about 2.6:1. In view of the method of calcu- lating chromosome lengths and the greater difficulty in measuring somatic prophases, these average ratios may be considered approximately as 9:1 and 3:1 respectively. Another bit of evidence should be considered before discussing the possible significance of these observations. In general it is well known 1935] SAX AND SAX, CHROMOSOME STRUCTURE AND BEHAVIOR 429 that the meiotic cycle is a leisurely process. The resting stage of the sporocyte may be of short duration in certain species, but the early prophase is prolonged, and in certain conifers the microsporocytes may remain in the early prophase stage for several months (Sax and Sax, 1933). The pachytene stage is prolonged in most species of plants and animals judging by the ease and frequency with which this stage is found. The development from pachytene to metaphase may be rather rapid, but the first metaphase stages, interphase, and second meiotic division are more prolonged. The meiotic cycle from early prophase to tetrad formation in the microsporocytes of Tradescantia requires about six days (Sax and Edmonds, 1933). The somatic cycle in the microspore of Tradescantia is much more rapid. The prophase stage is not initiated until vacuolation of the microspore cytoplasm, and the development from this stage to the formation of the daughter nuclei occurs in about three days at most. We have no data regarding the time required for the mitotic cycle in root tips of Tradescantia, but the duration of mitosis in stamen hairs of Tradescantia is less than two hours at normal tempera- tures (Tischler, 1922). Laughlin (1919) found that the entire mitotic cycle requires only four hours in Allium Cepa, at a temperature of 20 degrees C. The duration of the development from the time that a definite spireme can be observed until the separation of sister chroma- tids appears to be much longer in meiosis than it is in mitosis, and it seems probable that the mitotic cycle is more rapid in root tips than in microspores. THE MECHANISM OF CHROMOSOME CONTRACTION IN MITOSIS AND MEIOSIS The chromonemata of mitotic chromosomes in Tradescantia are in the form of minor spirals at anaphase in the second meiotic division, the division of the microspore nucleus, and presumably in all other mitotic divisions (Cf. Sharp, 1934). As the chromosomes pass into the resting Stage the spirals tend to uncoil and fill the nucleus with loosely and irregularly coiled chromonemata. These old coils are never straightened out before the new coiling is initiated in the prophase for the next divi- sion. The new coils contract the chromonemata and apparently aid in drawing out most of the old spirals persisting from the previous division. At this point the chromosomes are in the typical “spireme” stage and their lengths can be measured approximately. The new coiling can be observed during the later prophase stages, and at metaphase there are about 20 to 25 minor spirals in each chromosome. The microspore chromonemata are compactly coiled at metaphase. Judging from the 430 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI relative lengths of the metaphase chromosomes in microspores and root tips, the chromonemata of the latter are not so tightly coiled (Cf. Sharp, 1934). In general the development of the minor spirals reduces the chromosome length about 70 percent between the prophase spireme and metaphase in root tip cells, and even more in microspore cells. At no period in the mitotic cycle are the chromosomes uncoiled completely ; the spirals from the preceding division persist until the new coiling begins at early prophase. Observations and measurements in several different genera seem to indicate that this behavior is of general occurrence in the somatic cycle of cell division (Cf. Kuwada and Nakamura, 1935). The chromosomes of the meiotic prophase appear to be free from spirals persisting from the previous mitotic anaphase, and if remnant spirals occur they are so nearly straightened out that they can hardly be recognized as spiral structures at pachytene. There is a great amount of chromosome contraction between pachytene and the first meiotic meta- phase in Vicia, Tradescantia, Lilium, Trillium, Secale, Allium and other genera. Belling (1928) found a 10:1 reduction in chromosome length in Lilium and Dark (1934) found an 11:1 reduction in Bellevalia. The average reduction in chromosome length between pachytene and meta- phase in the genera which we have examined is about 9:1. An examina- tion of published drawings of these stages in other genera of plants with large chromosomes indicates that a similar degree of chromosome con- traction is of general occurrence. The paired chromosomes at early pachytene are very slender, and even at late pachytene the diameter of the chromonemata is much less than it is at corresponding stages in mitosis. During the contraction between pachytene and metaphase the chromonemata become coiled in major spirals. The two chromatids of each chromosome are coiled together in single spirals at early metaphase in Tradescantia (Sax and Humphrey, 1934), Secale (Sax, 1930), Rhoeo (Sax, 1935) and Vicia, but two parallel coiled chromatids are found at this stage in Gasteria (Taylor, 1931), Trillium (Huskins and Smith, 1935) and Fritillaria (Darlington, 1935). These major coils are much wider and fewer in number than the minor spirals of the somatic chromosomes. In Tradescantia there are 5 to 6 major coils in each chromosome at meiosis, as compared with 20-25 coils in somatic chromosomes, and the gyres of the major coils are about twice as wide as those of the minor coils. Darlington (1935) finds from 8 to 15 major spirals in the meiotic chromosomes of Fritilaria and about 80 minor coils in the somatic chromosomes. Minor spirals within the major coils have been observed in Trade- scantia (Fujii, 1926; Ishii, 1931; Kuwada, 1932; Kuwada and Naka- 1935] SAX AND SAX, CHROMOSOME STRUCTURE AND BEHAVIOR 431 mura, 1933; Kato, 1934); in Hosta (Ishii, 1931); in Sagittaria and Lilium (Shinke, 1934), and in Trillium (Matsuura, 1934). (For these literature citations see Kato and Iwata (1935), who also describe the spiral within spiral structure of the meiotic chromosomes of Lilium.) Our observations and measurements of chromosome length in mitosis and meiosis seem to show that the relations of the major and minor coils differ in different genera. In Tradescantia the minor coils seem to be well established at early metaphase so that the separation of the major coils at late metaphase is associated with little change in the coiled chromatids. There is, however, some reduction in the width of gyres between early metaphase and anaphase, which can be attributed to the continuation of minor coiling. If the minor coils are well developed at metaphase the length of the meiotic chromonema of the major coil should be about the same at the somatic metaphase chromosomes. We have made wire models simulating the major coils in order to estimate the degree of contraction caused by the major spirals. The coiling is re- sponsible for about a two-thirds reduction in length, so that the meiotic chromonemata of Tradescantia, including only the major coils, are about 27 microns long, as compared with an average length of 21 microns for the somatic chromosomes. If only coiling is responsible for chromosome contraction in Tradescantia the minor coils at meiotic metaphase are nearly as well developed as they are in root tip chromosomes. In Secale we find a different relation between the major and minor coils at meiosis (Sax, 1930). The two chromatids of each homologue are coiled together in a single spiral at early metaphase, but at late meta- phase the major spirals tend to straighten out and the chromatids sepa- rate with no elongation of the meiotic chromosomes. The average length of these coiled chromonemata at early metaphase is estimated to be about 24 microns, but after the major coils are reduced at late metaphase the chromosome length is about 8 microns. The somatic chromosomes at metaphase have an average length of about 14 microns. Apparently the minor coils are not well developed at early metaphase, but are formed during metaphase and are effective in reducing the major spirals. The relations of the major and minor spirals in Vicia, Lilium and Rhoeo are more or less intermediate as compared with the conditions found in Tradescantia and Secale. If chromosome contraction is effected only by coiling of the chromo- nemata we would expect that the degree of reduction in length between pachytene and metaphase would be correlated with the relation between major and minor coiling. There does seem to be some correlation in certain genera. In Tradescantia where both major and minor coils occur 432 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI together at meiotic metaphase, the ratio of chromosome length at pachy- tene and meiotic metaphase is about 9:1. All of this chromosome con- traction might be attributed to major and minor coiling. In Rhoeo and Zea, however, the pachytene metaphase ratio may exceed 15:1, although there are no apparent major spirals in the meiotic chromosomes of Zea, and in Rhkoeo the minor spirals appear to be most effective in chromo- some contraction during late metaphase as the major spirals are reduced (Sax, 1935). Apparently there may be some linear contraction of the chromonema without coiling as Belling (1928) has suggested. The major spirals may be observed at the second meiotic division, as is usually or always the case in Gasteria, Trillium, Sagittaria and Fritil- laria, or only minor spirals may occur as is the case in Tradescantia, Rhoeo, and presumably in all genera in which the second meiotic division chromosomes resemble those of mitosis. In Lilium and in Vicia some chromosomes show major spirals and others only minor spirals at the second meiotic division. The nature of the coiling at this division appears to be associated to some degree with the length of the interphase. The outstanding features of meiosis in relation to chromosome con- traction are, (1) the almost complete elimination of the coils of the pre- ceding anaphase chromosome at the pachytene spireme, (2) the great reduction in chromosome length between pachytene and metaphase, and (3) the occurrence of major spirals. THE TIME OF THE CHROMOSOME DUPLICATION The anaphase chromosomes at mitosis have been described as two parted in Tradescantia (Kaufmann, 1926); Trillium, Allium, Trade- scantia, Vicia, Podophyllum (Sharp, 1929; Telezynski, 1930); Galtonia (Smith, 1932); Scilla (Hoare, 1934); Narcissus (Hedayetullah, 1931) ; Drosophila (Kaufmann, 1934) while four parted chromosomes at late anaphase or telophase have been described in Tradescantia (Nebel, 1932) and Paeonia, Allium and Tulipa (Stebbins, 1935). Two coiled chromatids in the anaphase chromosomes of the second meiotic division have been described in Gasteria (Taylor, 1929, 1931), Galtonia (Smith, 1932), Allium (Koshy, 1934), Scilla (Hoare, 1934), Trillium (Huskins and Smith, 1935), Rhoeo (Sax, 1935) and Tradescantia (Nebel, 1932; Kuwada and Nakamura, 1935). The anaphase chromonemata at mitosis are described as longitudinally single structures by Belar (1928), Dar- lington (1932, 1935) and Belling (1933), although Belar shows clearly the longitudinal split at late anaphase in Auwlacantha. The anaphase chromosomes of the second meiotic division in Trade- scantia have been described as single structures (Sax and Humphrey, 1935] SAX AND SAX, CHROMOSOME STRUCTURE AND BEHAVIOR 433 1934), but further study tends to confirm the interpretation of Kuwada and Nakamura, who present photographic illustrations which show the dual nature of these anaphase chromosomes. We can observe both in the second meiotic anaphase and in the anaphase chromosomes of the microspore division, some evidence of two closely associated coiled chro- matids and the constrictions apparently produced by the twisting of the partially separated spirals. The experiments conducted by Riley (Cyto- logia, in press) indicate that the split chromosome behaves as a unit in response to X-ray treatment, since the microspore nuclei rayed during the resting stage show only chromosome breaks at metaphase. X-ray treatment at the resting stage of microspore nuclei in Trillium produces only chromatid breaks at metaphase, indicating that the chromosomes are effectively split when they go into the resting stage (Huskins and Hunter, 1935). The differences in the response to X-ray treatment in these two genera may be caused by the degree of separation of the sister chromatids. In Trillium the chromatids are well separated at diplotene and form more or less independent major spirals at early metaphase, while in Tradescantia the two chromatids of each homologue are closely associated at early metaphase of the first meiotic division. A correspond- ing difference in the relations of the chromatids may exist during the later stages of meiosis and early microspore development. The available evidence seems to show that the chromonemata are longitudinally split when they enter the resting stage, and they may be four parted as indicated by the careful work of Nebel and Stebbins. If they are four parted at this stage the split chromatids may behave as units during the next mitotic cycle, so that the chromosomes may be con- sidered as composed of two chromatids when they enter the resting stage (cf. Nebel, 1933). The time at which chromosome duplication is initiated is a question on which there is considerable difference of opinion. Some investigators believe that it occurs at very early prophase while others find the split at late prophase, metaphase or even at anaphase or telophase. We are inclined to associate chromonema coiling with the longitudinal split of the chromosome. The chromosomes are considered as two parted when they enter the resting stage. In mitosis the two chromatids are coiled together in loose spirals at the beginning of prophase. At early prophase each of the two chromatids is longitudinally split. This split causes each split chromatid to coil independently. This coiling pulls out the remnant coils of the previous anaphase and -causes the chromatids to separate so that two more or less parallel strands are observed at the spireme stage. These shorten by coiling, separate at anaphase, and each anaphase chromosome contains two closely associated spiral threads. 434 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI The leptotene threads at meiosis are also split, as observed by McClung (1928) and Robertson (1931), although they usually appear to be single. The threads pair at pachytene, and at late pachytene each chromatid splits and the coiling of the chromonemata begins as in mitosis. At diplotene the homologous chromosomes separate except at chiasmata or where twists occur. The minor spirals at meiosis involve only two strands or half-chromatids, as in mitosis. The minor spirals appear to begin development before the major spirals, but they may develop so slowly that they continue to coil, or at least contract, after the major coils are established at early metaphase in certain species. The meiotic bivalent is an eight parted structure as described in Tradescantia (Nebel, 1932), in Trillium (Huskins and Smith, 1935) and as we have observed in Tradescantia (from a preparation made by Dr. Dermen). The chromosomes separate at anaphase, the two chromatids of each chromosome appear as parallel coils, but the ‘tertiary split” can not be seen. The interphase is brief and even the major coils may persist at the second division. There is apparently no true resting stage at interphase and the chromosomes pass to the second metaphase as four parted struc- tures. At second anaphase each chromosome contains two coiled chro- matids as a result of the split which occurred at late pachytene or early diplotene. The number of strands and time of splitting in any one of the 2n chromosomes is essentially the same in both the mitotic and meiotic cycles. THE MECHANISM OF MEIOSIS According to Darlington (1932 “meiosis differs from mitosis in the nucleus entering prophase before the chromosomes divide instead of after they divide. The “precocity theory” is based on the assumption that there is a curtailed resting stage or earlier prophase and that the leptotene chromosomes are single. Evidence from many sources indi- cates that the chromosomes contain at least two coiled chromatids when they enter the resting stage in the mitotic cycle and at the completion of meiosis. There is little reason for assuming that the last premeiotic division differs from other mitoses. There is good evidence that the meiotic cycle is a much more leisurely process than the mitotic cycle. This evidence, together with the observa- tions on chromosome length at prophase in mitosis and meiosis, seems to indicate that meiosis is associated with a retardation in cellular activity rather than precocity in development (Cf. Stebbins, 1935). The pro- longation of prophase at meiosis is associated with the straightening out of the old spirals of the preceding anaphase before the new coiling begins. The two chromatids of each leptotene chromosome are so closely asso- 1935] SAX AND SAX, CHROMOSOME STRUCTURE AND BEHAVIOR 435 ciated that they appear as a single thread. At mitosis there often is a tendency for homologous chromosomes to be associated in pairs, but intimate gene by gene pairing is inhibited by the coiled structure of the chromonemata. As the remnant coils begin to straighten out the chro- matids are split and the new coiling begins so that the chromonemata are always coiled during the mitotic cycle. In meiosis, however, no coils, or at least only very loose remnant spirals, are found at prophase, and an intimate association of homologous chromosomes is effected. The new split occurs in each chromatid at late pachytene, coiling begins, and homologous chromosomes begin to separate. At this stage of chromo- some development, in both mitosis and meiosis, each chromosome (2n) contains four chromatids and there is no longer any strong affinity be- tween homologues. In meiosis the homologous chromosomes usually appear to be held together by chiasmata, although other factors appear to be effective in the meiotic association of chromosomes in certain cases. At interphase the pairs of “sister” chromatids separate except at the fiber attachment, and at second metaphase they again become closely associated. The anaphase chromosomes pass into the resting stage as double spirals as in mitosis. The prolongation of the mitotic cycle in the microspore suggests that the retardation of meiosis tends to persist at the subsequent mitosis so that the microspore chromosomes are more compactly coiled than in root tip cells, but the retardation at early pro- phase is not sufficient to effect chromosome pairing even in auto- tetraploids. The retardation theory of meiosis is in accord with the numerous observations that the anaphase chromosomes pass into the resting stage as two parted (or four parted) structures, and with the fact that the meiotic prophase chromosomes are much longer than those of the mitotic spireme. The primary difference between mitosis and meiosis is the longer prophase in meiosis which enables the residual coils of the chro- monemata to straighten out and permit the homologous chromosomes to become intimately associated in pairs before the chromatids split and coil. SUMMARY A study of chromosome structure and behavior at mitosis and meiosis has been made in Tradescantia, Vicia, Lilium, Secale, and other genera. The somatic chromosomes at the resting stage are in the form of loose spirals. At prophase the chromonemata form new coils which appear as the remnant coils are straightened out. The contraction of the chro- mosomes between prophase and metaphase is effected by coiling of the chromonemata. The average reduction in length of the chromosomes 436 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI between the ‘“‘spireme” stage and metaphase is about 3:1 in the root tip cells of the species examined, and may be greater in microspores. As the chromosomes enter the resting stage the chromonemata tend to uncoil filling the resting nucleus with loose, irregular spirals. Thus the chromo- nemata are coiled at all stages in the mitotic cycle. The chromosomes at meiotic prophase are practically free from rem- nant coils and the new coils do not appear until late pachytene. The chromosomes at meiotic prophase are about twice as long as those of the mitotic prophase. The average reduction in chromosome length between pachytene and meiotic metaphase is about 9:1. This reduction in length may be effected by linear contraction of the gene string, and by major and minor coiling of the chromonemata. The relation of these factors in chromosome contraction may differ in different genera. A theory of the mechanism of meiosis has been proposed, based on the comparison of chromosome behavior in mitosis and meiosis, and the comparative duration of the mitotic and meiotic cycles. The chromo- nemata of mitotic chromosomes are in the form of spirals at all periods of the chromosome cycle and this coiling prevents any intimate associa- tion of homologous chromosomes. At prophase of meiosis the chromo- nemata are relatively free from coils and homologous chromosomes can become closely paired before the new coiling is initiated. The retardation theory of meiosis is in accord with the recent evidence regarding the time of chromosome duplication. LITERATURE CITED Bexar, Kart (1928). Die cytologischen Grundlagen der Vererbung. 412 pp. Bo ara Berlin. BELLING, J. (1928). Contraction of ae a race during maturation divisions in Lilium and other plants. (Univ. Cal. Pub. Bot. 14: 335-343.) 1933). Crossing over and gene baavbeeete in flowering plants. (Genetics, Lf 338-413. Da S. (1934). Chromosome studies in Scilleae II (Jour. 98. e. ARLINGTON, C. D. (1932). Recent Advances in Cytology. pp. 559. Blakiston Sons and Co., Philadelphia ee The internal ees of the nd a eg I. The nuclear cycle in n Fritillaria. (Proc. Roy. Soc. 118: 33- —— (1935). The internal ‘mechanics of the chromosomes. II. Pro- Ci. 1485 i ). The internal mechanics of the chromosomes. III. Re- lational coiling and crossing-over in Fritillaria. (Proc _ Roy. Soc. 118: 6. HEDAYETULLAH, S. (1931). On the grey and division of the somatic chr eat? in Narcissus. (Jour . Soc. 51 Hoare, G. (1934). A comparative study ‘of the cheers mes of Scilla nonscripta during somatic and meiotic mitosis. (La Cellule, 43: 7-41.) 1935] SAX AND SAX, CHROMOSOME STRUCTURE AND BEHAVIOR 437 Huskins, C. L. and A. W. S. Hunter (1935). The effects of X-radiation on chromosomes in the microspores of Trillium erectum Linn. (Proc. Roy Soc. 117: 22-33.) Huskins, C. L. and S. G. Smiru (1935). Meiotic chromosome structure in Trillium erectum L. (Ann. Bot. 49: 119-150.) Kato, K. and J. Iwata (1938). Spiral structure of chromosomes of Lilium. (Mem. Coll. Sci. Kyoto Imp. Univ. 10: 263-273.) KAUFMANN, BERwIND P. (1926). Chromosome structure and its relation to the chromosome cycle. I. Somatic mitoses in Tradescantia pilosa. (Am. Wer Bot. 13: 59-80. ——— (1934). Somatic mitosis of Drosophila melanogaster. (Jour. of Morphol. 56: 125-154.) Kosuy, T. K. (1934). Chromosome studies in Allium. II. The meiotic chromosomes. (Jour. Roy. Micr. Soc. 54: 104-120.) Kuwapa, Y, and T. Nakamura (1935). Behavior of chromonemata in mitosis. VI. Metaphasic and anaphasic longitudinal split of chromo- somes in the heterotype division in pollen mother cells in Tradescantia reflexa. (Cytologia, 6: 314-319. Laucuuiin, H. H. (1919). Duration of the several mitotic stages in the dividing root- tip cells of the common onion. (Carnegie Inst. of Wash- ington. Pub. 265: McCurntTock, B. (1933). The association of non-homologous parts of chromosomes i in the mid-prophase of meiosis in Zea Mays. (Zeit. Zellf. oe oe 19: 191-237. McCiun E. (1928). Differential chromosomes of Mecostethus gra ns (dait Zellf. Microsk. Anat. 7: 756-778.) NeBEL, B. R. (1932). Chromosome structure in Tradescantiae. I. Meth- ods and morphology. (Zeit. Zellf. Microsk. Anat. 16: 251-284.) ——._ (1933). romosome structure in Tradescantiae. IV. The history of the chromonemata in mitosis of Tradescantia reflexa Raf. (Cytologia, 5: 1-14, Rosertson, W. R. B. (1931).) Chromosome studies. II. Synapsis in ve Tettigidae with special reference to the spk ate split. Jou 9-146 Sax, Kart (1930). Chromosome structure and the mechanism of crossing over. (Jour. Arnold pen 11: 193-220.) —— ; romosome structure in the meiotic chromosomes of Rhoeo oe Hance. (Jour. Arnold es 16: 216-222. ) Sax, Karu . EpmMonps (1933). Say of the male game- tophyte i in Tadeo ane (Bot. Gaz. 94: 156— Sax, Karr and L. M. Humpnrey (1934). ee of meiotic chromo- somes in microsporogenesis of Tradescantia. (Bot. Gaz. 96: 353-362.) Sax, Kar and H. Sax (1933). Chromosome number and morphology in the conifers. (Jour. Arnold Arb. 14: 356— Suarp, Lester W. (1929). Structure of large somatic chromosomes. (Bot. Gaz. 88: 349-382. ) 934). Introduction to Cytology. ed. 3. pp. 567 McGraw-Hill. ondon SMITH, FRANK H. °(1932). The structure of the somatic and meiotic chromosomes of Galtonia candicans. (La Cellule, 41: 243-263.) 193 STEBBINS, G. romosome structure and the mechanism of meiosis in plants. am Nat. 69: 81. TAYLOR, Wo. R. . Chromosome structure in mitosis and meiosis. Proc. Internat. Can Plant Sci. (Ithaca), 1: 265-270.) 438 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI (1 . Chromosome studies in Gasteria. III. Chromosome dag be during microsporogenesis and the post- -meiotic mitosis. (Am. r. Bot. 18: 367- poe eee ak H. (1930). Le cycle du chromosome somatique I. (Ob- rg bag on vitales sur les poils staminaux de Tradescantia virginiana L. ) (Act. Soc. Bot. Polon. 7: 381: (1931). Cycle évolutif du chromosome somatique II. Observa- ren sur le materiel fixé (Racines d’Haemanthus Katharinae Back.) (Act. Soc. Bot. Polon. 8: 109-132. ae G. Sieg Allgemeine Pflanzenkaryologie. pp. 897. Born- traeger, Berlin DESCRIPTION OF PLATES PLATE 161 Photographs of Tradescantia microspores fixed and stained in aceto- carmine after ssa ey with ammonia alcohol. Magnification * 1200 except photo. 8 which i 000. Photo. 1. Resting ae nucleus filled with loose spiral chromonemata. Photo, 2 ee be Early prophase with remnant spirals from preceding S Photo. 4. Prophase spireme with most of the remnant coils removed and ye new spirals appearing in each chromati Photo. 5. Late prophase after the Poaniio) » have contracted and more clearly separated. Photo. 6. The chromosomes at metaphase fixed without effective pre- treatment. Photo. 7. Early metaphase showing the coiled chromonemata in each clad and partial twisting of chromosomes about each Photo. 8. Fay nila tae approximate number of coils, and vists in chrom Photo. 9. Anapha ase Taomene ce with two coiled chromatids in each romosome PLATE 162 Photographs of mitotic and meiotic chromosomes of Vicia faba. Aceto- carmine preparations. Magnification x 1200. Photo. 1. Prophase spireme of root ti Photo. 2. Anaphase from root tip ainiaekaee showing secondary co bere in the ‘‘m’” chromosomes, and some evidence of in- rnal spirals. Photo. 3. Meioti oe ase showing a of chiasmata. Photo. 4. Meioti papal showing major coils. Photo. 5. Meiotic tele ophase showin ng extreme chromosome contraction. Photo. 6. Prophase of microspore div Photos. 7 and 8. Anaphase ee of the microspore division. PLATE 163 Photographs of mitotic and meiotic cells from aceto-carmine prepara- tions. Magnification x 800. Photo. 1. Meiotic pachytene in Lilium regale. Photo. 2. Meiotic metaphase in Lilium regale. Jour. ARNOLD ARB, VOL. XVI. PLATE 161 CHROMOSOME STRUCTURE AND BEHAVIOR THE HELIOTYPE CORP. BOSTON Jour. ARNOLD ARB, VOL. XVI. PLATE 162 CHROMOSOME STRUCTURE AND BEHAVIOR THE HELIOTYPE CORP. BOSTON JouR. ARNOLD ARB, VoL. XVI. PLATE 163 CHROMOSOME STRUCTURE AND BEHAVIOR THE HELIOTYPE CORP. BOSTON Jour. ARNotD ARB, VOL. XVI. PLATE 164 CHROMOSOME STRUCTURE AND BEHAVIOR THE HELIOTYPE CORP. BOSTON 1935] SAX AND SAX, CHROMOSOME STRUCTURE AND BEHAVIOR 439 . Somatic prophase in Lilium regale. a a omatic metaphase in Trillium grandiflorum. PLATE 164 Photographs of meiotic chromosomes of Zea and Allium photographed at X 1200 and reduced in reproduction. Photos. 1 and 2. Meiotic chromosomes of Zea Mays showing reduction in ee romosome length between very late diakinesis or early diplo- e and diakinesis Photos. 36 eoue chromosomes of Allium Cepa showing eo in chromosome length between pachytene and metaphas CyToLoGicAL LABORATORY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 440 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI THE FLORA OF SAN FELIX ISLAND Ivan M. JOHNSTON With plate 165 THE ISLAND of San Felix, lat. 26° 16’S., long. 80° 0’ W., lies over 800 kilometers off the north coast of Chile and about an equal distance north of the islands of Juan Fernandez. It is volcanic, apparently a portion of a disrupted crater, and is surrounded by depths of about 4000 m. The island is about 2.5 km. long. At the narrow west end there is an abruptly elevated hill of yellowish tuff called Cerro Amarillo, 183 meters high. The principal part of the island, however, is broad and flattish and is composed of a series of black lavas that form a platform which gently slopes towards the north and broadens towards the east end to a maxi- mum width of nearly 1.5 km. It is bounded by sea-cliffs, 50-70 m. high on the south and 15-20 m. on the north side. This broad flattish part of the island is overlaid here and there with a thin covering of fine dry earth, and is strewn all over with fragments of lava of no great size. In appearance the island is extremely barren and desolate. Most of the surface of San Felix is devoid of plants. Dr. Chapin in- forms me that on the flattish major part of the island the flora consists of three evident species, of which the scattered, depressed growths of the shrubby Suaeda nesophila are the most conspicuous. The other two species, members of Eragrostis and Cristaria, are ephemeral annual herbs of scattered occurrence. Two species can be attributed to Cerro Amarillo. On its lower slopes Atriplex Chapinii grows in the soft yellow- ish volcanic rock. The only erect shrub on the island, Thamnoseris lobata, grows in a few sheltered ravines on its upper slopes, where some adiabatic fogs supplement the extremely scanty rains of this desert island. The flora of San Felix is hence very poor in the number of its species, as well as in the number of individual plants. It is, however, high in endemism. In the same small archipelago, about 18 km. east of San Felix, is the island of San Ambrosio. This is a slightly smaller island but is uni- formly lofty, being surrounded by imposing cliffs and reaching an alti- tude of 450 m. Its high fog-bathed cliffs and crests are very much more favorable for plants than the low arid flats and slopes of San Felix. The very different environmental conditions have given it a flora almost completely different from that of its neighbor. Though San Ambrosio 1935] JOHNSTON, FLORA OF SAN FELIX ISLAND 441 evidently has a much richer and more interesting flora, the difficulties of landing on the island and of climbing its precipitous sides have hindered its proper exploration. What is known of its flora today rests almost exclusively upon a few fragmentary specimens collected on its green crests by Simpson in 1869. These fragments, carried off the island in his hat (!), are all that has been collected of a number of very distinct endemic genera and species. The flora of San Ambrosio is obviously a remarkable one, high in endemism and still a promising source for new genera and species. No island off the west coast of America is in greater need of exploration. The present paper is concerned only with the relatively small flora of the more accessible San Felix and is an account of the two collections from the archipelago preserved in North American herbaria. Its prime purpose is to put on record certain new species and new names for use in research growing out of the recent visit to the island by the yacht, Zaca. The first botanical collections from San Felix and San Ambrosio appear to have been made by Enrique Simpson in August, 1869. The eight species represented were enumerated by R. A. Philippi, Bot. Zeitung, 28: 496-502, tab. 8a (1870). Simpson had only one species from San Felix, a Parietaria, which has not since been collected in the archipelago. The second and best existing account of the flora of San Felix and San Ambrosio is by Frederico Philippi, Anal. Univ. Chile, 47: 185-194, cum tab. (1875). This paper reviews the collections of Simpson and discusses those made by Ramon Vidal in September, 1874. Simpson reached the crest of San Ambrosio, whereas Vidal got only the few plants he could obtain from the sides of that island. Vidal, however, did collect more carefully on San Felix. In publishing, the younger Philippi, un- fortunately, treated the archipelago as a whole and gave only rare indi- cations as to the particular island upon which Vidal made his several collections. From some notes which I made in the Philippi Herbarium at Santiago in 1926, from internal evidence within Philippi’s report, and from mention of collections in Reiche’s Flora de Chile, it is possible to state that Vidal obtained on San Felix specimens of Suaeda, Cristaria, Lycapsus and Thamnoseris. It is just possible that a Tetragonia and a Frankenia were also obtained. In his official report of his “Esploracion de las islas San Félix i San Ambrosio,” Anal. Univ. Chile, 45: 735-756 (1874), Vidal antedated the report of the younger Philippi and gave an atrociously misspelled list of 9 angiospermous species stated to repre- sent the flora of San Felix. The determinations were attributed to 442 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI Philippi. The list evidently contains the species obtained by Vidal on San Ambrosio as well as San Felix. Among the species attributed to San Felix by Vidal, the following, to judge by the Vidal specimens in the Philippi Herbarium at Santiago, were collected on San Ambrosio only, — Sicyos, Atriplex foliolosum, Heliotropium (Nesocaryum), and Frankenia, Subsequent to the report of Frederico Philippi accounts of the flora of San Felix and San Ambrosio have been largely compilation. Hemsley, Report Voy. Challenger, Bot. 3: 97-100 (1884), apparently unaware of the report by the younger Philippi, translated and abridged the earlier and less complete report of the elder Philippi and recorded collections of Lycapsus and Thamnoseris made by Coppinger from rocks just south of San Ambrosio. Reiche, in Engler & Drude, Veg. Erde, 8: 269 (1907), compiled a few general notes on the flora of the islands and mentioned the visit to the island by Johow. There is only the most general informa- tion recorded concerning the work accomplished by this latter botanist. According to the brief reports, Deutscher Wiss. Ver. Santiago, Verhandl. 3: 525 and 529 (1898), Johow visited the island in October 1896. His collections, so long unenumerated, have only recently been placed in the capable hands of Prof. Carl Skottsberg for critical study and report. The only other botanical collections from the islands known to me are those specially treated in this paper. Prof. Bailey Willis visited San Felix in May 1923 for geological observations and there obtained five specimens now preserved in the herbarium of Stanford University. An account of his visit to the islands, with numerous photographs, is to be found in the Publications of the Carnegie Institution, vol. 382: 120— 124, tab. 64-68 and 74-75 (1929). Dr. James P. Chapin visited San Felix for ornithological work on Feb. 18, 1935. He was a member of the “Templeton Crocker Pacific Expedition” of which there is an account in the Scientific Monthly, 41: 281-285 (1935). His botanical collec- tions consisting of nine numbers represent four species. The first set of them is in the New York Botanical Garden and a set of duplicates is in the Gray Herbarium. The collections represent plants from various parts of the island and are uncommonly ample. They are the best that have been made on San Felix. Through the courtesy of Dr. Chapin I am able to reproduce two of the photographs he made on the island. In my list of the flora of San Felix there are seven species accredited to the island. Of this number two, Eragrostis and Parietaria, are non- endemic, occurring also on the arid coastal region of northern Chile and southern Peru. Of the remaining five, Lycapsus is known from San Ambrosio as well as San Felix. This is the only species accredited to 1935] JOHNSTON, FLORA OF SAN FELIX ISLAND 443 both of the islands. The endemic species of San Felix belonging to non- endemic genera, Atriplex, Suaeda and Cristaria, are rather well marked but have their closest affinities with plants of the coastal hills of south- ern Peru and northern Chile. The archipelago has two genera of the Compositae, each of which has a species on San Felix. These genera are endemic and are so distinct that their natural position in their re- spective subfamilies is yet open to question. The genus Thamnoseris perhaps is nearest to Dendroseris of Juan Fernandez. The tribal position of Lycapsus is still undetermined. If we may judge from the relation- ship evident in all other members of the flora of San Felix and San Ambrosio, we may perhaps surmise that these two genera of Compositae had relationships, now lost or obscured, in western South America. CATALOGUE OF THE SPECIES GRAMINEAE Eragrostis peruviana (Jacq.) Trinius, Mem. Acad. St. Petersb. sér. 6, 1: 396 (1831). Collected in a mature condition by Chapin (1108). Not only a new species for the known flora of San Felix but also the first monocot to be reported from the archipelago. The species has heretofore been known only in the coastal hills from central Peru south to the Taltal region in northern Chile. A close comparison of Chapin’s collections with abun- dant material from the continent has revealed no characters or even tendencies whereby it might be distinguished. URTICACEAE Parietaria debilis Forster, Prodr. 73 (1786). — Parietaria feliciana Philippi, Bot. Zeit. 28: 501 (1870); F. Philippi, Anal. Univ. Chile, 47: 192 (1875). This genus is known from San Felix only through a collection from Simpson. It was the only plant he obtained on that island. I have compared a fragment of his collection with the common and variable plant of western South America passing as P. debilis and can find no characters to separate them. (CCHENOPODIACEAE Atriplex Chapinii, sp. nov., perennis monoica fruticosa e caudice crasso lignoso erumpens depressa pallida 1—3 dm. alta, 3-12 dm. lata; caulibus prostratis vel decumbentibus ramosis, juventate summum ad apicem inconspicue evanescenter pubescentibus mox glabratis; foliis concoloribus glabris lanceolatis vel oblanceolatis numerosis confertis evidenter costatis sed inconspicue nervatis 8-15 mm. longis 2-7 mm. 444 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XVI latis apice subacutis basi in petiolum ca. 1 mm. longum gradatim attenu- atis margine integerrimis; floribus staminatis in spicas terminales 2—3 cm. longas infra medium plus minusve interruptas et bracteatas flaves- centas (maturitate plus minusve fuscas) aggregatis: floribus pistillatis in axillis foliorum superiorum dispositis; bracteis fructiferis ultra medium connatis in ambitu angulatis 6-7 mm. longis 5-6 mm. latis, corpore crassis induratis subobovoideis 2-3 mm. longis 2 mm. latis plus minusve verrucosis, margine prominente herbaceis planis sparse dentatis plus minusve trilobatis; seminibus erectis 1-1.5 mm. diametro, testa brunnescente, radicula verticali. SAN FELtx: low bush about 2.5 dm. tall and 3-9 dm. broad, Feb. 18, 1935, J. P. Chapin 1104 (Gray Herb., typr; NY); low bush, leaves grayish green, forming circular or oval clumps 3-12 dm. in diameter, 2.5-3 dm. high, Chapin 1105 (NY); with male flowers, Chapin 1106 (G, NY) and 1109 (G); a flat-growing plant keeping close to soil and rocks, May 2, 1933, Bailey Willis 4 (Stanf.). The material collected by Chapin and by Willis is quite similar and evidently conspecific. The island plant is most closely related to the poorly understood group of spreading monoecious perennials of the Chilean coastal region. Although collected by a busy geologist and by a busy ornithologist the species is curiously lacking in the collections of Vidal, who seems to have been the most energetic botanizer on the island, unless the report of Tetragonia maritima, by the younger Philippi, Anal. Univ. Chile, 47: 88 (1875), may have been based upon a sterile speci- men of it misidentified. Atriplex Chapinii is evidently distinct from A. foliolosum Phil. which is known only from sterile specimens collected on the adjacent island of San Ambrosio. The latter endemic has crowded sessile ovate-triangular leaves only 2.5 mm. long and 2 mm. broad. Suaeda nesophila, nom. nov. — Suaeda divaricata Mog. var. micro- phylla F. Philippi, Anal. Univ. Chile, 47: 193 (1875): Reiche, FI. Chile, 6: 175 (1911), not S. microphylla Pallas. This shrub was collected both by Willis (no. 2) and by Chapin (1107 and 1110). According to Dr. Willis it is the common bush growing on the flatter parts of the island in rounded masses up to 9 dm. in diameter and 5 dm. in height. Dr. Chapin notes that the leaves are at first a light grayish green which turns finally to a dull purplish red. The distal leaves are almost always reddish. He adds that the rounded clumps are 2.5—3.5 dm. tall and 9-12 dm. broad. This endemic species is evidently related to S. foliosa Mog. of the coastal hills of northern Chile and southern Peru, from which it is quickly distinguished by its very much more 1935] JOHNSTON, FLORA OF SAN FELIX ISLAND 445 slender and more branched habit and very much smaller clavate leaves. It is certainly not closely related to S. divaricata Mogq., which is a large bush confined to Argentina. MALVACEAE Cristaria insularis F. Philippi, Anal. Univ. Chile, 47: 186 (1875); Reiche, Anal. Univ. Chile, 91: 405 (1895) and FI. Chile, 1: 257 (1896). Collected on San Felix by Willis (no. 3 a—b) and by Chapin (nos. 388, 1111). Their collections are very mature with the leaves mostly dried and weathered. There are some flowers and much good fruit. The plant is endemic though related to a group of small-flowered annuals occurring in the coastal hills from central Chile to central Peru. COMPOSITAE Lycapsus tenuifolius Philippi, Bot. Zeit. 28: 499, tab. 8a, fig. 1-5 (1870); Philippi, Anal. Univ. Chile, 43: 484 (1873), locality incorrect; F. Philippi, Anal. Univ. Chile, 47: 188 (1875). — Alomia tenuifolia (Phil.) Benth. & Hook. ex Reiche, Anal. Univ. Chile, 109: 10 (1901) and Fl. Chile, 3: 260 (1902); Robinson, Proc. Amer. Acad. 49: 439 and 453 (1913). There are photographs and fragments in the Gray Herbarium of the original collections at Santiago made by Simpson and by Vidal. Simp- son’s collection is labeled as from San Ambrosio. Vidal’s collection is given as from San Felix. Dissections of this authentic material shows conclusively that this endemic genus is not a Eupatorioid as has been supposed. The plant has fertile pistillate marginal florets with a 3- toothed ligule about once and a half the length of the tube. The tubular inner florets appear to be hermaphroditic and sterile. The style-branches are linear, flattened and abruptly contracted into a short triangular apex. The receptacle bears conspicuous slender scales which seem to separate the marginal florets from the inner ones. Except for the brac- teate receptacle the plant is very suggestive of some of the Helenioids or even certain Asterioids. The bracteate receptacle suggests the Helian- thoids but none of the other structures suggest that group of the Com- positae. The same may be also said for the Madineae. Lycapsus has relations even more vague than Thamnoseris, the other endemic genus of the Compositae. Thamnoseris lobata, sp. nov. — Thamnoseris laceratus sensu F. Philippi, Anal. Univ. Chile, 47: 190, cum tab. (1875); Reiche, Anal. Univ. Chile, 116: 580 (1905) and Fl. Chile, 5: 6 (1910), as to shrub of San Felix. The specimens from Prof. Willis (no. 1) which consist of leaves, flow- 446 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI ering inflorescence and parts of stem, agree well with the collections from San Felix by Vidal which were described at length and illustrated by the younger Philippi. According to Willis’ notes the plant grew in a sheltered ravine on Cerro Amarillo, 150 m. alt., and formed a low, abun- dantly lactiferous shrub with thick woody trunk and branches. The bark was light gray, smooth and “swollen looking.” Neither the collection by Vidal nor that by Willis matches the single leaf of the lactiferous shrub of San Ambrosio upon which the elder Philippi, Bot. Zeit. 28: 499, tab. 8, fig. A6 (1873), based his Rea ? lacerata. This leaf from San Ambrosio is triangular in outline, broadest at the base, cut 9/10 of the way to the rachis and has the well spaced 3—4 pairs of lobes prominently and narrowly lobulate. In the San Felix plant, as given by the younger Philippi and later by Reiche, and as shown by the material collected by Bailey Willis, the leaves are lanceolate, broad- est above the middle, gradually contracted towards the base, cut 4—%4 to the rachis, and the crowded 3-5 pairs of lobes sparsely lobulate- toothed. The material from San Felix is consequently quite different in appearance from the scanty specimen originally described from San Ambrosio and so seems to merit a new name. The lengthy descriptions given by the younger Philippi and by Reiche are based upon the collec- tion of Vidal illustrated by the former author. This collection and the mentioned descriptions and plate amply characterize Thamnoseris lo- bata of San Felix. The imperfectly known plant that is correctly known as Rea lacerata Phil., Thamnoseris laceratus (Phil.) F. Phil. or Dendroseris lacerata (Phil.) Hemsley, is consequently known only from San Ambrosio where Simpson reported it as common on the crests and Hemsley, Report Chal- lenger Voy. Bot. 3: 99 (1884), reported it as occurring on a rock just south of that island. Until the flowers and inflorescence of this plant of San Ambrosio are described, its exact relationship with T. /obata of San Felix must remain a matter of surmise. The lactiferous shrub of San Felix is evidently a member of the Cichorioideae. It has naturally been compared with Dendroseris, an endemic genus of Juan Fernandez having similar habit, for Thamnoseris and Dendroseris not only occur in adjacent archipelagos, but are unique among the Cichorioids in having a pronounced woody habit. Floral structures of these two genera, however, show many differences. I am inclined to the opinion that we must await the judgment of some future student who has mastered the complexities of the classification and inter- relation of the Cichorioids, before we definitely select similarities of habit as indicative of direct relationship between the two insular endemic Jour. ARNoLD Ars. VoL. XVI PLaTE 165 WEST END OF SAN FELIX FROM THE LAVA CLIFFS ON THE SOUTHSIDE OF THE ISLAND; THE HIGH HILL OF TUFF IS CERRO AMARILLO. SoUTHEAST PORTION OF SAN FELIX SHOWING THE TOPOGRAPHY OF THE PRINCIPAL PART OF THE ISLAND; THE ISLET OF GONZALES IS ON THE RIGHT AND TO THE LEFT, EASTWARD IN THE DISTANCE, IS THE ISLAND OF SAN AMBROSIO FULL-TONE— MERIDEN OO - a Oe an OS oe re 1935] JOHNSTON, FLORA OF SAN FELIX ISLAND 447 genera. The similarities of habit may be simply parallel evolution, the similar responses of two different stocks isolated under equitable insular climates. Many groups of angiosperms, prevailingly low and herbaceous on the continents, have produced woody forms on oceanic islands. The woody habits of Dendroseris and Thamnoseris, accordingly may be simp- ly ecological and not indicative of immediate relationship. The exact relationship of the two insular genera with each other and with other Cichorioids is still uncertain. Their relationship does not seem to be Old Pacific, for the Cichorioids are very scantily represented in New Zealand and Australia and northward in the Pacific, and none of them in this region have structures suggestive of close affinity with our insular genera. In the past our genera have been associated with the genus Fitchia, a woody group of Polynesia, but as suggested by Drake del Castillo, Jour. de Bot. 12: 176 (1898), that genus now proves to be a Mutisioid. Consequently the old hypothesis as to an Old Pacific rela- tionship of our shrubs must seek new justification. Though I can find no evident relatives of them in South America, I suspect that, like other members of our insular florulas, the insular Cichorioids were probably derived from South American ancestors. The best development of the Cichorioid Compositae in the Southern Hemisphere is to be found in western South America. Our insular Cichorioid shrub may be merely aspects of the evolutionary activity centering on the adjacent continent. ARNOLD ARBORETUM, HARVARD UNIVERSITY. 448 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI SOME NEW TREES AND SHRUBS FROM MEXICO ALFRED REHDER Carya mexicana Engelm. forma polyneura, f. nov. A typo recedit foliolis plerisque elliptico-lanceolatis margine peni- cillato-ciliolato excepto glabris vel fere glabris, nervis utrinsecus 15-20 tantum 3—6 mm. distantibus, exocarpio tenui 1.5—3 mm. crasso. Nuevo Leon: Sierra Madre Oriental, Puerto Blanco to Taray, about 15 m.s.w. of Galeana, C. H. and M. T. Mueller, no. 1226, July 23, 1934, “shrub or small tree very abundant in more or less moist woods, prac- tically all over the mountain” (type). This form, at the first glance, looks very distinct with its closely and conspicuously veined leaflets, glabrous except the ciliate margin and a few hairs on the under side near the base, but among the numerous speci- mens under no. 1226 there are larger detached leaves, apparently from sterile vigorous branches, or from different trees, with more distant veins and thinly pilose over the whole under surface; these leaves do not differ at all from those of typical C. mexicana. Also the difference in the thickness of the husk does not seem to segregate the tree of the Sierra Madre Oriental from that of the type locality of the species, Alvarez, San Luis Potosi. A fruit of Palmer’s no. 835% which represents the type of C. mexicana has the husk in a dry state up to 6 mm. thick, while most specimens collected in the same locality by Dr. A. R. Goodman in 1910 have the husk only 2—3 mm. thick and a smaller nut resembling in shape that of C. ovalis var. obcordata, but one lot of seeds has the husk as thick as in the type. It thus appears that C. mexicana varies as much in the thickness of its husk and in the pubescence of its leaves as C. ovata (Mill.) K. Koch, to which it is closely related: in fact, I cannot find any strong character to separate the two, and I believe that C. mexicana is only a southern cli- matic form of C. ovata. The buds are generally smaller, but in a speci- men collected by Dr. E. Palmer in 1902 (no. 71) the terminal winter-bud is 1.5 cm. high. Nothing seems to be known of the character of the bark, and until we have more complete material, C. mexicana may be kept distinct. I also have before me a flowering specimen from La Jolla Ranch, Tamaulipas (Robert Runyon 1019) with nearly glabrous leaves. Thus the species seems to be restricted to three adjoining departments in northeastern Mexico, namely Nuevo Leon, San Luis Potosi and Tamaulipas. 1935] REHDER, NEW TREES AND SHRUBS FROM MEXICO 449 Litsea Muelleri, spec. nov. Frutex ramis gracilibus hornotinis ut petioli dense fulvo-villosis tertio vel quarto anno tarde glabrescentibus fuscis vel partim flavidis. Folia persistentia, coriacea, ovata, 2—4.5 cm. longa et 1.2—3.2 cm. lata, breviter acuminata vel acuta, basi rotundata vel subcordata, supra minute reticu- lata et tenuiter villosa demum glabrescentia, luteo-viridia, subtus tomen- tosa, glauca, nervis utrinsecus circiter 5 ut costa supra leviter vel vix elevatis, subtus distincte elevatis, costa flavescente; petioli 2-4 mm. longi, tomentosi. Inflorescentiae pleraeque 3-florae, axillares, solitariae vel 2—3 fasciculatae; pedunculi 1.5-3 mm. longi, fusco-pilosi ut pedi- celli; pedicellus medius 1-1.5 mm. longus, laterales breviores; bracteae caducae, extus fusco-pilosae; perianthium tubo nullo, segmentis ovalibus 2.5-3 mm. longis basi extus strigosa exceptis glabris; stamina 9, fila- mentis glabris vix dimidias antheras subrectangulares truncatas aequan- tibus, ea seriei tertiae glandulis binis filamenta subaequantibus aucta,; ovarium parvum, ovoideum, in stylum brevem attentuatum. Fructus tantum immaturi circ. 5 mm. diam. visi. Nuevo Leon: Sierra Madre Oriental, east side of divide between San Francisco Canyon and Pablillo, 15 miles s.w. of Galeana; alt. 8500 ft., common in dense woods, C. H. and M. T. Mueller, no. 379, May 14, 1934 (type); Sierra Madre Oriental, caflon above Alamar, about 15 miles s.w. of Galeana, alt. 5000-6000 ft., common in open oak wood, C. H. and M. T. Mueller, no. 689, June 2, 1934 (with immature fruit). This species seems to be most closely related to Litsea Neesiana (Schau.) Hemsl. which is easily distinguished by its larger and narrower leaves, cuneate or broadly cuneate at base, by the larger inflorescence, longer peduncles and filaments longer than the anthers; in the shape of the leaves it resembles L. parvifolia (Hemsl.) Mez, but that species is quite glabrous and the pedicels are long and slender. Amelanchier paniculata, spec. nov. (§ Malacomeles).* Frutex ramulis satis validis initio dense albo-tomentosi ut petioli et facies inferior foliorum, annotini tarde glabrescentes rubro-fusci, vetu- stiores griseo-fusci. Folia persistentia, chartacea vel subcoriacea, 1Amelanchier sect. Melecrnan (Dene.), comb. nov. Nagelia Lindley in Bot. Spe ena le 40 (184 5). Nagelia Lindley, Veg. King, 560 (1846). — Wenzig in Linnaea, se 80 (1880) “Naegelia.” — Non Naegelia Rabenhorst, E nioe. 1: 85 (18 Cotoneaster se Macnee Decaisne in Nouv. Arch. Mus. Hist. Nat. Paris, 10: 177 (1874). ae seep A. Naegelia (Lindl.) Wenzig in Jahrb. Bot. Gart. Mus. Berlin, 2: 304 3) ea sect. Nagelia [Lind].] owe Gatt. Pomac. (in Wiss. Beil. Progr. Falk- pee) nie Berlin, p. 25. 1900). — Schneider, Ill. Handb. Laubholzk. 1: 742 (1906). 450 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XVI oblongo-elliptica vel oblonga vel interdum obovato-oblonga, 3.5—7.5 cm. longa et 1.5-3 cm. lata, acutiuscula vel obtusiuscula, mucronulata, basi cuneata vel late cuneata, rarius fere rotundata, integra vel interdum remote et minute denticulata denticulis ad mucronem reductis, supra costa tenuiter villosa et margine dense villosula exceptis ab initio glabra, laete viridia, subtus dense albo-tomentosa, nervis utrinsecus 15—20 supra ut costa media leviter impressis, subtus prominulis et costa media mani- feste elevata; petiolis 5-10 mm. longi, dense tomentosi. Inflorescentiae terminales, paniculatae vel simpliciter racemosae, pleraeque pe- dunculis inferioribus trifloris superioribus unifloris, rarius simpliciter racemosae, rarissime pedunculis inferioribus 7-floris et superioribus tri- floris, 3-5 cm. longae, albo-tomentosae; pedicelli 3-10 mm. longi; calycis tubus 2.5 mm. longus, extus ut lobi leviter floccoso-tomentosus mox glaber, lobi triangulari-ovati, acuti, mucronulati; petala late ovalia, vel suborbicularia, basi abrupte contracta, circiter 6 mm. longa, glabra; stamina circiter 20, petalis dimidio breviora, antheris cordato-ovatis 1 mm. longis; carpidiis 3-5, dorso apice conico libero longe villoso excepto calycis tubo adnatis ventre fere ad basin liberis villosis; styli basi excepta glabri, staminibus paullo longiores. Fructus immaturus subglobosus, circiter 8 mm. longus, calycis lobis reflexis intus extusque glabris margine villoso-ciliolatis coronatus. Nuevo Leon: Sierra Madre Oriental, San Francisco Cafion, about 15 miles s.w. of Galeana, alt. 7500-8000 ft., scattered on rocky soil in open or scrub wood, C, H. and M. T. Mueller, no. 282, May 9, 1934 (type). This new species belongs to the section Malacomeles and seems most closely related to A. nervosa Dene., but readily distinguished by the much larger, generally oblong and entire or nearly entire leaves and the usually paniculate inflorescence. As Malacomeles is the earliest sectional name of this group, it must be maintained, when the section is transferred to Amelanchier. More- over, Nagelia is an illegitimate name being a later homonym of Naegelia Rabenhorst. Nagelia, Ndgelia and Naegelia must be considered ortho- graphic variants of the same name, since both genera are named in honor of Karl Naegeli. Arctostaphylos novoleontis, spec. nov. (§ Comarostaphylis). Frutex ramulis hornotinis puberulis, secundo vel tertio anno glabre- scentibus fuscis decorticantibus. Folia persistentia, lanceolata, 3—6 cm. longa et 6-12 mm. lata, interdum minora, acuta vel obtusiuscula, mucro- nulata, basi cuneata, margine integro cartilagineo et praesertim basin versus leviter revoluto, supra luteo-viridia, lucidula, tenuiter villosula, demum glabrescentia, subtus glauca, breviter villosula, costa media supra 1935] REHDER, NEW TREES AND SHRUBS FROM MEXICO 451 leviter impressa subtus elevata, nervis utrinque obsoletis; petioli 4-6 mm. longi, tenuiter villosuli. Panicula 3—6 cm. longa, interdum ad race- mum reducta, minute canescenti-puberula, laxiflora; pedicelli graciles, 4-10 mm. longi; sepala 5, triangulari-ovata, acuminulata, 1.75 mm longi, extus parce villosula, ciliolata, rubra; corolla cylindrico-urceolata, 8-10 mm. longa, extus glabra, intus sparse villosula; lobis latis rotundatis recurvatis; stamina dimidiam corollam aequantia, filamentis glabris ima basi dilatatis, antheris bi-aristatis; stylus corollam subaequans ut ovari- um 5-loculare glaber. Fructus immaturus 4 mm. diam., granulosus, obscure fusco-ruber. Nuevo LeEon:. Sierra Madre Oriental, San Francisco Canon, about 15 miles s.w. of Galeana, alt. 7500-8000 ft.; sparse on top of hill above the upper canon, C. H. and M. T. Mueller, no. 319, May 12, 1934, (type) ; Sierra Madre Oriental, last hill on west side of lower San Fran- cisco Canon, about 15 miles s.w. of Galeana, common in shrub zones on slopes and tops, C. H. and M. T. Mueller, no. 1032, July 15, 1934, “fruit dark brownish red.” This species seems to be most closely related to A. polifolia H.B.K., A. glaucescens H.B.K. and A. minor (Small) Standl.; from the first two species it is readily distinguished by the glabrous ovary and from A. poli- folia also by the glabrous filaments, from A. minor it differs in the larger and broader more pubescent leaves, the smaller corolla and the pubescent filaments. The fruiting specimen, no. 1032, differs somewhat in the smaller leaves 1.5-4 cm. long and 4-7 mm. broad and less densely pubescent. Menodora Muellerae, spec. nov. Suffrutex humilis, decumbens, ramosissimus, 15 cm. vix excedens, caulibus subteretibus viridibus, rima basibus foliorum decurrentibus formata hispidula excepta glabris. Folia opposita, crassiuscula, vix distincte petiolata, 4-10 mm. longa, lineari-oblonga vel anguste oblanceo- lata, inferiora interdum lineari-subulato, acuta et mucronata, basin versus sensim attenuata, basibus foliorum oppositorum contiguis rimam hispidulam ad par foliorum inferius decurrentem formantibus, margine setis brevibus leviter reflexis setoso-ciliolata, ceterum glabri, costa media subtus elevata. Flores in apice ramulorum solitaria; pedicellus 2-4 mm. longus, pilosulus; calyx campanulatus, tubo circ. 1.5 mm. longo glabro vel basin versus sparse pilosulo, lobis plerumque 10 subulato-linearibus 5—6 mm. longis ciliolatis; corolla hypocraterimorpha, tubo gracili apicem versus leviter ampliato 12-14 mm. longo, lobis oblongis 6-7 mm. longis acuminulatis; stamina exserta, dimidios lobos aequantia, filamentis glabris 5 mm. longis circiter 3 mm. infra faucem tubo affixis, antheris 452 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI oblongis 2.5 mm. longis; stylus staminibus paullo longior; stigmate capitato lato. Fructus circ. 7 mm. diam.; pedicellis recurvatis, coccis circumcissilibus; semina plerumque 2 vel 3, rarius 4 vel 1, ovoidea vel oblongo-ovoidea, plus minusve compressa, circ. 4 mm. longa, atrofusca. Nuevo LrEon: Sierra Madre Oriental, cedar savannah above Encinal wheatfields, about 15 miles s.w. of Galeana, alt. about 7500-8000 ft., abundant over small areas, C. H. and M. T. Mueller, no. 463, May 19, 1934 (type). This species seems to be most closely related to M. longiflora A. Gray on account of the salver-shaped corolla with a long slender tube, though not as long as in M. longiflora, but it is easily distinguished by the decum- bent habit, smaller, much narrower leaves, exserted stamens and soli- tary flowers with the pedicels recurved in fruit. The species is named in honor of Mrs. Mary Taylor Mueller who accompanied her husband on his Mexican expedition. HEBRARIUM, ARNOLD ARBORETUM, Harvarp UNIVERSITY. 1935] CREAGER, CEPHALOSPORIUM WILT OF ELMS 453 NEW FACTS CONCERNING CEPHALOSPORIUM WILT D. B. CREAGER ONE RESULT of the experience of having to wage a campaign against the Dutch elm disease in America has been the calling of attention to the native wilt diseases of elms. Of these there are two; one is tenta- tively designated ‘“‘Cephalosporium wilt” (or “Cephalosporium die- back’’), the other ‘‘Verticillium wilt.” They are widely spread and the first named is relatively frequent in nursery stock and on older trees of all ages. Both are being studied intensively at the Arnold Arboretum of Harvard University with the object of more clearly defining symptoms, discovering means of spread, determining the course of the infection and testing methods of control. My own investigations on the former are now sufficiently advanced to warrant the publication of certain findings which appear to point the way to control. 1. Two kinds of reproductive bodies, as recently noted by Dr. Curtis May, are produced by the fungus that causes Cephalosporium wilt. These are (a) naked spore heads such as are characteristic for the genus Cephalosporium and (b) pycnidia, a type of fructification not known for that genus. Both make their appearance in laboratory cul- tures and both occur in nature. The pycnidia are of special importance but no explicit statement of their occurrence in nature could be found in the literature. The significant feature to be emphasized, one hitherto unrecorded, is their natural abundance and their importance. They form profusely in the bark of infected twigs and branches as the bark tissues gradually die during the summer. These pycnidia contain myri- ads of small spores which remain viable over winter. The spores ooze out in a gelatinous matrix through ostioles to the surface of the bark. Apparently wind and rain play an important part in spore dissemination. Also, since spores are exposed to the outside, such agencies as insects and birds may serve an important role in transmission of the pathogene. 2. Wounds in the leaves provide the most common infection court. Large numbers of leaves injured by canker worms have been found to be infected in the vicinity of killed branches on the same and nearby trees. Any insect causing open wounds in leaves or stems such as the canker worms (Alsophila pometaria Harr. and Paleacrita vernata Peck), the spiny elm caterpillar (Hamadryas antiopa Linn.), the elm leaf beetle 454 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI (Gallerucella luteola Miill.). the small European elm bark beetle (Scoly- tus multistriatus Marsh) and numerous others, provide suitable wounds through which the fungus may enter the suscept. 3. An early symptom of the typically infected leaf is a yellowing near the infection court or wound. The yellowed area increases and the portion near the wound dies and becomes brown. Usually a brownish discoloration of the veins occurs in advance of the necrotic area in the mesophyll. Finally the discoloration extends through the vascular strands of the petiole into the stem, and the leaf drops off. The patho- gene has been isolated from a large number of such leaves and from the stem in the vicinity of the leaf base. 4. Artificial inoculations made on seedlings in the greenhouse sub- stantiate these facts observed in the field. Whether mycelium, spores from spore-heads, or spores from pycnidia were used as inoculum no infection resulted when unwounded leaves were inoculated. On the other hand, any one of these kinds of inoculum brought about infection when leaves were wounded. In stained serial sections of artificially inoculated leaves, the pathogene has been traced from the leaf wound into the stem through the vascular strands of lateral veins, midrib and petiole. The fungus has also been recovered from various regions of wounded leaves and stems of inoculated seedlings by culture methods. 5. Based on the foregoing facts, a number of experiments on control of the disease have been initiated. The first important step is to prune out all dead and infected branches in order to eradicate the disease from individual cases and to destroy the source of future inoculum. Spray tests are also in progress in which insecticides and fungicides are being used to prevent attacks of insects and fungi and to reduce possibilities for infection through wounds already present. Results of these prelim- inary tests are encouraging; they indicate that a combined pruning and spraying program offers promising possibilities for controlling Cepha- losporium wilt. It is with pleasure that acknowledgment is made of the excellent facili- ties afforded for this investigation by the pathological laboratory at the Arnold Arboretum and the adjunct laboratory of the North Country Garden Club at the Pratt Estate Oval, Glen Cove, Long Island, New York. LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1935] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 455 THE ARNOLD ARBORETUM DURING THE FISCAL YEAR ENDED JUNE 30, 1935 THE WINTER of 1934-1935 was nearly as severe as that of the previ- ous year.' Very little snow fell up to December 31, but a heavy fall of snow took place on January 23 and much of it remained all winter. In January temperatures varied, but low temperatures prevailed. This persistent cold, after the previous severe winter, caused considerable damage to trees and shrubs in the Arboretum. Some of this damage was not apparent until early summer. In the fall of 1934 several tem- perature stations were established in various parts of the Arboretum which recorded variations of as much as 12 degrees between different stations. Temperature records from these stations will be of much assistance in selecting favorable localities for new plantings. Early in the spring, an extensive spraying program was carried through to combat a bad infestation of canker-worm. This was successful and prevented defoliation. Other successful spraying programs were carried out for leaf miner in hawthorns and for various other insect pests. To facilitate the best growth in several groups of shrubs, it was neces- sary to move and rearrange a large number of plants. More of this work is to be carried out this fall; many large specimens which were badly damaged by winter conditions will have to be replaced. The extensive collection of lilac varieties flowered exceedingly well last spring and attracted thousands of visitors, also the crab apples and later the mountain laurels flowered profusely, while most of the cherries and particularly the Japanese cherries showed the effects of the two preceding severe winters, as did many of the azaleas and rhododendrons. During the year 1258 packages of seed were sent out, 782 in the United States, 476 to 12 foreign countries, also 2187 plants and 545 varieties of scions and cuttings. There were received from the United States and other countries 454 packages of seed, 1894 plants, and 592 cuttings and scions. Five hundred and ninety-five plants were added to the Arboretum collections. — L. V. S. Pathological Laboratory.—The laboratory of plant pathology has now completed the first seven years of its existence. It was established through the efforts of the Supervisor, Professor Oakes Ames, in fulfill- 1Records of the effects of the low temperatures during the winter of 1933-34 on the trees and shrubs in the Arnold Arboretum have been published in the Arnold Arbo- retum Bulletin of Popular Information (Ser. 4, Vol. II, nos. 7 and 456 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XVI ment of the expressed wish of the late Director, Charles Sprague Sar- gent, and as part of Dr. Sargent’s conception of the Arboretum as an institution for the study of woody plants in all fundamental aspects. Professor Ames wisely housed the laboratory in quarters that facilitate cooperative effort and interchange of ideas with the Arboretum’s propa- gator, superintendent and geneticist. The functions of the laboratory were defined as comprising interest in the Arboretum’s living collec- tions, extension services, instruction and research. Naturally collections so extensive, so varied and of such diverse ori- gins and requirements as are those of the Arboretum constantly present health problems of concern to the superintendent and much material of interest to the pathologist. Likewise from time to time pathological problems call for consultation or cooperative undertaking between the pathologist and the propagator or the geneticist. It should also be stated that control measures evolved in the laboratory are tested or applied in the Arboretum whenever possible. The Arboretum has always exercised a generous attitude with respect to extension services; its plant stores and its knowledge are freely open to all. As might be expected many inquiries are referred to the labora- tory of plant pathology and the number has increased yearly. They come from private individuals, nurserymen, arborists, city and town tree wardens, foresters, landscape architects, etc. An account in some detail of this phase of the laboratory’s activities formed a part of the report for 1933-4. Constituted as a unit of an educational institution, the Arboretum has felt that it should bear some instructional responsibility. So from the first the laboratory of plant pathology has assumed its share. A course in the pathology of the forest and of woody plants open to quali- fied undergraduates and graduates has been offered in Harvard Univer- sity and from two to six appropriate research students each year have been directed in their special investigations. The Biological Laboratories and the Bussey Institution have provided suitable laboratory space for this work. Of the ability and the research accomplishments of these students I can speak highly. Five of them received travelling fellow- ships, including three Sheldon awards, at the conclusion of their under- takings at the Arboretum; and seven of them are now filling responsible positions in pathology, mostly research, here or abroad. While this participation in instruction has taken considerable time, most of it has been in connection with research and one result has been the enriching of the research activities of the Arboretum. This is of significance be- cause research is the foremost function of the laboratory of plant pathology. 1935] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 457 Statements covering the research accomplished by the laboratory during its first seven years have been included in the preceding annual reports. It will suffice here to report for the past year and merely to add that in the seven year period many problems have received attention and that many of these have been fruitfully solved. For the year 1934-5 the laboratory reports progress with or comple- tion of studies on the following topics: coniferous rusts, Gymnosporan- gium diseases, mycotrophy in Pinus and wilt diseases of elms. With respect to coniferous rusts particular mention should be made of the elucidation of life history connections, involving firs and spruces, in the genera Chrysomyxa and Milesia, and the working out of features of taxonomic value based on the morphology of spermogonia. The investi- gations of I. H. Crowell and J. D. MacLachlan on Gymnosporangium diseases, begun four years ago, are finished. From the results obtained lists have been compiled showing the immunity or grade of suscepti- bility of hundreds of pomaceous and Juniperus species to the more important species of Gymnosporangium found in eastern North America. Moreover, a new means of practical control has been demon- strated that obviates the necessity of eradication of either host, the only method practised up to this time. Asa by-product of the research, about two thousand packets of culture materials have been added to the Arboretum’s herbarium —a unique contribution. Dr. MacLachlan will spend the year 1935-6 on a Sheldon fellowship in Jamaica, B. W. I., working in a cooperative undertaking between the Arboretum and the Jamaican government on a devastating new rust disease of Pimenta, a genus of the economically important family Myrtaceae. This calls to mind the fact that so far no work has been done by our laboratory of plant pathology at the Arboretum’s tropical branch in Cuba — almost certainly a fertile field. One of the most valuable results of the year’s program has been the demonstration of the role of mycorrhizae in Pinus as collectors of certain of the tree’s mineral requirements; thus an answer of far-reaching significance is afforded to long unanswered questions as to whether they are important and, if so, in what way. The fate of America’s elms, threatened as they are by the Dutch elm disease, is now the most important tree problem on this side of the Atlantic. The Arboretum’s program in the campaign against the Dutch elm disease, as outlined in the report of 1933-4, has been continued vigorously. Distinct progress has also been made in our biological and control studies on a native wilt disease of elms the symptoms of which are almost the same as those of the Dutch elm disease. (See article by D. B. Creager in this number of the Journal of the Arnold Arboretum. ) 458 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XvI An essential part of this work was done at our field laboratory on Long Island. That laboratory should be maintained for at least one more summer. Towards the end of 1934, when there was an alarming amount of wavering and pessimism over the possibility of successfully com- batting the Dutch elm disease in America, the Arboretum took a strong public stand in support of the policy of complete eradication. It is gratifying to know that federal and state efforts towards eradication during 1935 afford justification for our optimism. With more reason than ever, we again express the belief that our elms can be saved if public support be maintained. — J. H. F Cytogenetic Laboratory.—The work on cytotaxonomic problems during the past year includes an analysis of Robinia by Dr. Whitaker, a cytological analysis of Verbena by Dr. Dermen, and a cytotaxonomic study of Tradescantia with the cooperation of Dr. Anderson. A comparison of chromosome structure and behavior in mitosis and meiosis seems to explain the mechanism of meiosis. Other work on chromosome structure has been continued, especially the spiral structure of the chromonemata in meiotic chromosomes in Tradescantia and Vicia. The mechanism of development and differentiation has been studied in relation to environmental factors. The breeding work has been continued with lilacs, roses, magnolias and azaleas. A plant breeding nursery has been established in which there are about a hundred first generation hybrids. A few of these hybrids flowered this year for the first time. — K. S The Herbarium.— During the past fiscal year, 16896 specimens have been added to the herbarium bringing the total number up to 408699 specimens. Of these accessions about 3900 came from the United States and Canada, 4600 from Central and South America inclusive of Mexico, 1150 from Europe and western Asia, 2350 from China, 390 from For- mosa, 1150 from India and Malaysia, 1800 from Australasia, and 530 were cultivated plants. Among the more important collections received during the year may be mentioned: about 8000 specimens representing 1100 numbers from Mexico collected by C. H. and M. T. Mueller, 1350 numbers with many duplicates from Guatemala collected by A. F. Skutch, 290 specimens from Ecuador collected by H. J. F. Schimpff, 293 specimens from Peru collected by G. Klug, 1157 specimens from Brazil collected by A. B. Krukoff, 3100 specimens from Kwangtung and 701 specimens from Hainan received from Lingnan University, 2660 plants from western 1935] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 459 China collected by J. F. Rock, 705 specimens from Hupeh collected by H. C. Chow and received from Wu-han University, 1296 specimens from Borneo collected by J. and M. S. Clemens, 1476 specimens from Sumatra collected by Rahmat Si Torroes, 215 specimens from Hawaii and 124 specimens from Samoa received from the Bishop Museum in Honolulu; 140 specimens from South Africa collected by Ecklon and Zeyher. To the fruit collection, 395 specimens were added, bringing the total number up to 8379. Additions to the wood collections consisted of 155 numbers, bringing the total up to 3786. The collection of negatives of types and critical specimens, chiefly Chinese, amounts now to 3012 negatives, 268 having been added during the year. For study outside of the Arboretum herbarium, 847 specimens were sent out on loan to 14 institutions and individuals in this country and in Europe. The distribution of duplicates amounted to 19059 specimens sent to 44 institutions in ‘the United States, Canada, Europe, Asia, Australasia and Africa. Botanical exploration by members of the staff and by expeditions wholly or partly financed by the Arnold Arboretum, has been carried on in America, eastern Asia and Africa. During the summer of 1934, Dr. H. M. Raup continued the study of the flora of the Harvard forest and its neighborhood started the year before, and collected herbarium material. Mr. C. H. Mueller and Mrs. Mary T. Mueller returned in August, 1934, from their collecting tour to Mexico referred to already in last year’s report. They collected chiefly in the Sierra Madre Oriental, Department of Nuevo Leon, and brought back about 8000 specimens, representing 1100 numbers. The collection is being identified chiefly at this institution and the Gray Herbarium, and already a number of new species has been described from this collection. About the middle of June, 1935, Mr. Mueller started on another collecting tour to the same locality. Mr. A. F. Skutch collected from July, 1934 to January, 1935, for the Arnold Arboretum in southwestern Guatemala in the De- partments of Huehuetenango, Quezaltenango and Quiché; he collected about 1350 numbers with many duplicates, also wood specimens. His collection is being determined chiefly at the Gray Herbarium and the Arnold Arboretum. In China, two institutions received financial support from the Arnold Arboretum in 1934 for carrying on botanical exploration. The Fan 460 JOURNAL OF THE ARNOLD ARBORETUM (VOL, XVI Memorial Institute of Biology under the direction of Professor H. H. Hu sent an expedition to Yunnan to collect chiefly near the border of Cochinchina, and another to Szechuan which penetrated into the in- terior Lololand not yet visited by any botanical collector. From Ling- nan University expeditions have been sent under the direction of Dr. Franklin P. Metcalf to the provinces of Kwangsi and southern Kiangsi. Mr. M. Dinklage who had made before important collections in tropical west Africa for the Berlin Botanical Museum, went to Liberia in 1934 to collect in the hinterland of Monrovia for the Arnold Arbore- tum. As the rainy season prevented his starting soon after his arrival for the Nimba mountains, his final goal, he collected first in the neigh- borhood of Monrovia and toward the end of the year he started for the imba mountains which, however, he was destined not to reach. He was taken sick on the way and died of dysentery on February 2, at the Ganta Mission Station. — A. R. The Library.—At the end of June 1935 the Library comprised 42,025 bound volumes, 10,917 pamphlets, and 17,573 photographs, a gain of 579 volumes, 432 pamphlets and 332 photographs. Many pamphlets and a few books have been received as gifts, including ‘‘Arbejder fra den Botaniske Have i K@benhavn,” Russian works, and miscellaneous articles. Important articles from periodicals and other works have been analyzed, and notice of all available obituaries of botanists have been filed in the author catalogue. There has been an increase over last year in the number of photo- graphs received, the largest accessions comprising those taken by Dr. Edgar Anderson during his trip to the Balkans, Mr. E. J. Palmer’s taken on collecting trips in the western and southwestern United States, Prof. J. G. Jack’s taken in Japan and Mrs. Susan D. McKelvey’s in Jamaica. The use of the collection has been extensive. Many photographs have been sold, including 700 taken by the late Dr. E. H. Wilson on his ex- pedition to eastern Asia, 1917-18, purchased by Jardin Botanique de Etat, Bruxelles. Sixty-three lantern slides have been added, the majority of them being colored. Cards filed during the year include 1,766 in the card catalogue of books in the Library, 336 in the catalogue of photographs, 4,617 in the “Card-index of New Genera, Species, and Varieties Published by the Gray Herbarium,” and 5,625 in the manuscript “Index of Illustrations and of New Genera, Species and Varieties of Ligneous Plants Publishe Since 1915,” prepared at the Arboretum, bringing the total of the latter to 106,675. For supplements to the author and subject “Catalogue of the Library” 2,589 slips have been filed. The subject slips now ready for publication number approximately 20,000. 1935] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 461 Volumes bound number 229 and more than 100 smaller books and pamphlets were put into pamphlet binders. Clipping files and scrap- books preserve much interesting material. Approximately 225 visitors registered in the Library during the year. These include Ynes Mexia, University of California, Berkeley; Rev. and Dr. M. H. Rioch, India; J. J. Wilder, Honolulu; E. Percy Phillips, National Herbarium, Pretoria; Jan T. Byhouwer, Velsen, Holland; Chi-yian Chiao, University of Nanking; E. Lowell Kammerer, Morton Arboretum, Lisle, Illinois; Alfred Gundersen, Brooklyn Botanic Gar- den; Clara W. Fritz, Ottawa; Z. H. Hellwig, Warsaw; H. B. Haddow, Toronto; G. Weidman Groff, Lingnan University, Canton, China; T. Yamamoto, Taihoku Imperial University, Formosa; David D. Keck, Carnegie Institution Laboratory, Stanford University, California; G. E. Gates, Judson College, Rangoon, Burma; A. S. Thurston, Uni- versity of Maryland, College Park; E. D. Merrill, New York Botanical Garden: Donald Wyman, Cornell University, Ithaca, N. Y.; Isabella Preston, Ottawa, and librarians attending the Convention of the Special Libraries Association held in Boston early in June. Dr. L. M. Ames of the U. S. Bureau of entomology and plant quarantine has spent more than a year studying Berberis and Mahonia. The publications of the Arboretum, the “Journal of the Arnold Arbo- retum” and the “Arnold Arboretum Bulletin of Popular Information were issued regularly. “Contributions from the Arnold Arboretum of Harvard University,” numbers vii and viii, and “Through the Arnold Arboretum” were published during the year. Of approximately 487 periodicals and reports currently received from all parts of the world, 220 are received in exchange for our “Journal” and “Bulletin” and 10 in exchange for our “Contributions.” The subscription to 20 periodi- cals was cancelled with a saving of about seventy dollars. Reprints from the “Journal” were also sent out as a medium of exchange. The following new periodicals have been received, a large number in exchange for our publications, some by gift and some by purchase. They are as follows: ACADEMY OF NATURAL SCIENCES OF PHILADELPHIA. Proceedings. Vol. lxxxii — 1930 — Philadelphia. 1931 — ACADEMY OF NATURAL SCIENCES OF PHILADELPHIA. Year book. 1930 — Philadelphia. 1931 — AustRALia — Council for scientific and industrial research. Journal. Vol. viii, no. 1 — Melbourne. 1935 — BLACK ROCK FOREST. Papers. Vol.i,no. 1 — April 1935 — Cornwall- on-the-Hudson. 1935 — 462 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XVI BiuMEA. Vol. i, no. 1 — August 25, 1934 — Leiden. 1934 — BOTANICAL review, interpreting botanical progress. Vol. i, no. 1 > January 1935 — Lancaster, Pa. 1935 — CHRONICA botanica. Vol. i— Leiden. 1935 > Institut J. B. Carnoy, Louvain. Travaux biologiques. No. 1 > January 11, 1929 — Lierre; Louvain. 1929 — INSTITUTUL DE CERCETARI SI EXPERIMENTATIE FORESTIERA. Analele. Ser. 1. Anul i, nr. 1 — Bucaresti. 1934 — Japan — Imperial forestry experiment station. Bulletin. No. 31 —> November 1931 — Tokyo. 1931 —> JouRNAL of South African botany. Vol. i, pt. 1 — January 1935 — Kirstenbosch. 1935 — LEXINGTON BOTANIC GARDEN. Lexington leaflets. Vol. i, no. 1 —> April 11, 1931 — Lexington, Mass. 1931 — Lisson — Instituto botdnico de faculdade de ciéncias. Trabalhos. i. 1925-32. Lisb6a. 1932? Nas vrt; revija hortikulturnog drustva. God. i, svezak 1-4, 7 > Zagreb. 1934 — New ZEALAND journal of science and technology. Vol. xi, no. 5 > Wellington. 1930 — PLANT science literature. Vol. i, no. 1 — Dec. 31, 1934 — Jan. 5, 1935 — Washington. 1935 — REGENSBURG, Germany — Koniglich-baierische botanische gesell- schaft. Schriften. Bd.i. Regensburg. 1792. Revista sudamericana de botanica. Vol. i, no. 1 — February 1934 > Montevideo. 1934 — RoczNnIKA nauk ogrodniczych. (Annales des sciences horticoles.) Tom. i — Warszawa. 1934 > ROSENJAHRBUCH. (Verein Deutscher’ rosenfreunde.) Berlin. SOVETSKIE subtropiki. No. 1-2. July-August 1929 — Sukhum. 1929 > THARANDT — Forstliche hochschule — Institut fiir auslindische und koloniale forstwirtschaft. Mitteilungen. Nr. 1 — Dresden; Tharandt. 1932 — ZEITSCHRIFT fur weltforstwirtschaft. Bd. i, heft. 1 — Okt. 1933 — Neudamm; Berlin. 1933 — Among other important accessions are: AMERICAN GEOGRAPHICAL SOCIETY. New England’s prospect. New York. 3 1935] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 463 AucusTINn, Samuel. Prolegomena in systema sexuale botanicorum. Viennae. 1777. BRITISH MUSEUM, London. A catalogue of the works of Linnzus, and publications more immediately relating thereto, preserved in the libraries of the British museum, Bloomsbury, and the British mu- seum, Natural history, South Kensington. 2d ed. London. 1933. BriTIsH MUSEUM, London — Natural history. Catalogue of the books, manuscripts, maps and drawings in the museum. Vol. vii. Supplement J—O. London. 1933. Bussey, Benjamin. Will of Benjamin Bussey (with four “codicils,” 1835-41). DEGENER, Otto. Flora hawaiiensis. [Book i. Honolulu. 1932.] DioscoripEs, Pedanios. The Greek herbal of Dioscorides. Illus- trated by a Byzantine, 512. Englished by John Goodyer, 1655. Edited and first printed, 1933, by R. T. Gunther. Oxford. 1934. Dorin, Karel. The beech forests of Czechoslovakia. Bern; Berlin. 1937. FESTSCHRIFT Cornelius Osten. Montevideo. 1933. GoEBEL, Karl. Organographie der pflanzen insbesondere der arche- goniaten und samenpflanzen. 3°, umgearbeitete aufl. Teil i-iii. Jena. 1928-33. — Erganzungsband. [i, ii.] Jena. 1924-31. INTERNATIONAL BOTANICAL CONGRESS, 5th, Cambridge, 1930. Inter- national rules of botanical nomenclature adopted by the Inter- national botanical congresses of Vienna, 1905, Brussels, 1910. Revised by the International botanical congress of Cambridge, 1930. Compiled by the Editorial committee for nomenclature from the Report of the subsection of nomenclature prepared by John Briquet. 3° ausgabe. Jena. Le Conte, John B. Observations on the genera Viola, Utricularia and Gratiola. A reprint from the Annals of the Lyceum of New York (1824-1826), to which are appended photographs of the hitherto unpublished plates of the author, destined to illustrate these works. Edited by J. A. Nieuwland. Notre Dame. 1917. MariE-VIcTORIN, frére. Flore laurentienne. Montréal. 1935. PARSONS & SONS, co. Descriptive catalogue of hardy ornamental, flowering shrubs and vines; including rhododendrons, roses, mag- nolias, Chinese and Ghent azaleas, camellias, Japanese maples and other rare and choice plants [in Kissena nurseries]. Flushing, L. I., [1887?] RaAFINESQUE, Constantine Samuel. Monographie des coquilles bi- valves et fluviatiles de la riviére Ohio contenant douze genres et 464 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI soixante-huit espéces. [Avec] Remarques sur les rapports naturels des genres Viscum, Samolus et Viburnum. Bruxelles. [1820. Sowerby, Arthur de Carle. The naturalist in Manchuria. 5 vol. Tientsin. 1922-30. Books were sent out as inter-library loans to Field Museum of Natural History, Library of the United States Department of Agriculture, Smith- sonian Institution, McGill University, United Fruit Company, Massa- chusetts State College, Wellesley College, University of Pennsylvania, Dartmouth College, New York Horticultural Society, Massachusetts Institute of Technology, Tufts College Library, Boston Society of Natural History, Massachusetts Horticultural Society and Michigan State College, as well as to many departments of Harvard University. In addition to loans, photostats or typewritten copies of references have frequently been made when books could not be loaned. ooks were borrowed for the use of the members of the staff and stu- dents from the Boston Public Library, Massachusetts Horticultural Society, Library of the United States Department of Agriculture and from the University libraries. A list of the forestry periodicals and reports prepared by the United States Forest Service was checked for our holdings, and additional material was prepared for inclusion in a supplement to “Index Lon- dinensis.” At the invitation of the Committee on Binding Advertisements in Periodicals the Library has agreed to cooperate with some thirty-four other libraries to bind 12 of its periodicals entire, all advertisements included. In the early spring Miss Margaret Hayes, under the direction of Mr. J. F. Ballard of the Boston Medical Library, spent more than two months in the Library indexing books and periodicals for a union list of the more important works to be found in special libraries in Boston. The Librarian attended the Convention of the American Library Asso- ciation in Denver from June 24-29, where she visited the public and university libraries. — E. M. T. Bibliography of the published writings of the staff and students July 1, 1934—June 30, 1935 AMEs, Oakes. An addition to the genus Vanilla. (In Harvard university, Botanical museum leaflets, 1934, ii, 101-103.) A contribution to our knowledge of the orchids of Spanish Honduras. t. ii. (In Harvard university, Botanical museum leaflets, 1934, iii, 17-36.) 1935] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 465 Critical notes on Costa Rican orchids. By Oakes Ames, F. T. Hub- arles Schweinfurth. (In Harvard university, Botani- cal museum leaflets, 1934, iii, 37-42. Epidendrum cystosum, a new species from the republic of Honduras, (In Harvard university, Botanical museum leaflets, 1934, ii, 105— 109. A nomenclatorial note [Epidendrum neoporpax ]. (In Harvard uni- versity, Botanical museum leaflets, 1934, ii, 112. otes on Mexican e “samearie ums eet largely on the Erik M. Ostlung collection. By mes, F. T. Hubbard and eens Schwein- furth. (In ae university, Botanical museum leaflets, 1934, ili, 1-16.) Outline of lecture on economic botany given in Biology 115, formerly Botany 15, Nov. 1934. Material prepared in 1930. Bro adside. Studies in Stelis. ili-iv. (In Harvard university, Botanical museum leaflets, 1934-35, iii, 45- 59, 134-135.) A four poly morphic alliance in Epidendrum. By Oakes Ames, F. T. Hubbard and Charles Schweinfurth. (In Harvard university, Botanical museum leaflets, 1935, iii, 93-110. Nomenclatorial studies in Malaxis and Spiranthes. By Oakes Ames an arles Schweinfurth. (In Har er university, Botanical museum leaflets, 1935, iii, 113-133.) Studies in Epidendrum. By Oakes Ames, F. T. Hubbard and Charles Schweinfurth. (In Harvard university, Botanical museum leaflets, 1935, iii, 61-76.) ANDERSON, Edgar. Rhododendrons. (In Arnold arboretum bulletin of popular informa- tion, 1934, 11, 21-24. ) Chromosome numbers in the Hamamelidaceae and their phylogenetic significance. By Edgar Anderson and ae Sax. (In Journal of the Arnold arboretum, 1935, xvi, 210-215.) An endemic Sophora from Rumania. (In Journal of the Arnold Gametic elimination in crosses between self-sterile species. (In American naturalist, 1935, ee 282-283. Nature’s bags of scent. (In Herbarist, 1935, pp. 5-6.) [Plants of current interest.] (In Arnold ean bulletin of popu- lar information, 1935, iii, 5-8, 13-16.) cience as a sto rehouse. (In Horticulture, 1935, xiii, 124.) aig perfoliata in Louisiana. By E. J. Palmer na Edgar Ander- S In Rhodora, 1935, xxxvii, 58-59.) A rk to the home of the lilac. (In Arnold arboretum bulletin of popular oe n, 1935, iii, ) BaILey, Irving he cambium oe its. derivative tissues. ix. Structural variability in aile Journal of the Arnold ‘arboretum, 1934, xv, 233-254. The cambium and its derivative tissues. x. Structure, optical prop- erties and chemical composition of the so-called middle lamella. By Thomas Kerr and I. W. Bailey. (In Journal of the Arnold arbore- tum, 1934, xv, 327-349.) 466 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI CROWELL, Ivan H. ompilation of reports on the relative roti tami of orchard vari- eties of apples to the cedar-apple rust disease. (In Proceedings of the American society for ae science, 1934, xxxii, 261- 272. ) The hosts, life cing A iiss control of the cedar-apple rust fungus Gymnospor rangium peri- cere Schw. (In Journal of the Arnold arboretum, 1934, xv, 163-232.) Eurtuicn, John A nectria disease of the beech following Cryptococcus fagi. pp. 104. (In eee oaetas from the Arnold arboretum of Harvard uni- versity, 19 FauL., Anna orwar e cambium and its derivative tissues. ix. Structural variability in the redwood, Sequoia sempervirens, and its significance in the identi- fication of fossil woods. By I. W. Bailey and Anna F. Faull Journal of the Arnold arboretum, 1934, xv, 233- Elaioplasts in Iris. (In Journal of the pant arboretum, 1935, xvi, —267. FAULL, Joseph Horace. Arthur Bliss Seymour (1859-1933). (In Proceedings of the Ameri- can academy of arts and sciences, 1934, 1xix 544. ) Winter foniiees a trees and shrubs growing in the Arnold arbore- tum By J. H. Faull, J. G. Jack, W. H. Judd and L. V. Schmitt. (In Arnold arboretum bulletin of popular information, 1934, ii, 29- 47, 53-6 Can we eradicate the Dutch elm disease? Boston. 1935. pp. 4. Hartcu, Alden culture aon for the study of 7 adam (In Journal of the Arnold arboretum, —— xv, 358-3 A jet-black mycelium for ng ecto- pcre mycorrhizae. (In Svensk botanisk tidskrift, 1934.3 XXxvili, 369-383. ) Hunter, Lilian M. A Pog gaa Me on life history studies of Eur ropean S of sia. (In oa of the Arnold arboretum, 1935, xvi, 143.) ers after a severe winter. (In Arnold arboretum bulletin of popular information, 1934, ii, 53-60. Plants of current interest. (In Arnold arboretum bulletin of popular information, 1934, ii Winter gee of trees and shrubs F tatac in the Arnold arbore- tum. By J. H. Faull, J. G. Jack, H. Judd and L. V. Se eet (In Arnold arboretum bulletin ass popular information, 1934, 29-47, 53-60. ) Winter injuries among trees and shrubs. (In Scientific monthly, 1935, xl, 332-338.) Boraginaceae. (Jn z, P. A. A manual of southern California botany, 1935, pp. piney Studies in Boraginaceae. The Boraginaceae of Hitigenlia South America. (In Journal of the Arnold arboretum, 1935, xvi, 1-64.) Studies in Boraginaceae. xi. e species of Tournefortia and oo in the Old World. Notes on Brand’s treatment of Cryptantha. or otherwise noteworthy species. (In Journal of the Arnold ee 1935, xvi, 145-205.) 1935] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 467 Jupp, William Henry. The association of Kew gardeners in America; the annual dinner, 4 0 Emil T. Mische. (In Journal of the Kew guild, 1934, v. 372-373.) [Notes on trip to the Pacific coast, 1933.] (In Journal of the Kew guild, 1934, v, 334-335.) Winter es a trees and shrubs oe in the Arnold arbore- tum. By J. H ull, J. G. Jack, W. H. Judd and L. V. Sc ee (In Arnold en bulletin of popular Sper dion. 1934, 29-47, 53-60. ) The association of ae gardeners in eee the annual dinner, 35; n Journal of the Kew guild, 1935, v Making horticultural records. (In Forheuiue *035, xiii, 166.) Kerr, Thom Action of ‘hydrofluoric acid in softening wood. (In Tropical woods, 1934, no. 40, pp. 37-42.) The cambium and its derivative tissues. x. Structure, optical prop- erties and chemical composition of the so-called middle lamella. B Thomas Kerr and I. W. Bailey. (In Journal of the Arnold arbore- 349. ) McKeE -vey, Mrs. Susan Delano Arctomecon californicum. “(In National horticultural magazine, 1934, xiii, 349-350. ) A verification of the occurrence of Yucca base in Arizona. (In Journal of the Arnold arboretum, 1934, xv, 350-352. Notes on Yucca. (In Journal of the Arnold arboretum, 1935, xvi, 268-271 MacLacuHian, John Douglas. The hosts of Gymnosporangium globosum Farl., and their relative susceptibility. (In Journal of the Arnold arboretum, 1935, xvi, 98-142. PALMER, Ernest Jes Adventures in ee collecting. (In American fern journal, 1934, xxiv, Game foods in the Arnold arboret By N. W. Hosley and E. J. Palmer. (In Arnold a Panay ere of popular information, 1934, ii, 49-52.) Indian relics in the Arnold arboretum. (In Arnold arboretum bulletin of popular information, 1934, ii, 61 Trees of the southeastern states. (In Journal of the Arnold arboretum, 1934, xv, 266 Supplement to the spontaneous flora of the Arnold arboretum. (In Journal of the Arnold arboretum, 1935, xvi, 81-97. Uvularia perfoliata in Louisiana. By E. iy) Palmer and Edgar Ander- son. (In Rhodora, 1935, xxxvii, 58-59.) REHDER, Alfred Amendments to the international rules of nomenclature, ed. 3. Pro- sed. maica Plain 4. pp. ie blutbuche. (In Mitteilungen der Deutschen dendrologischen peer 1934, xlv Notes on the ligneous plants described by hanes from eastern Asia. (In Journal ef the Arnold hia 1934, x Corrections and igeral i fe) s “M eeual of ely ated trees and shrubs.” Jamaica Plain. 1935. pp. 19. 468 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI Handeliodendron, a new genus of Sapindaceae. (In Journal of the Arnold arboretum, 1935, xvi, 65-67. ) Sax, Karl. sean ger e ey of cycadales. By Karl Sax and pe M. Beal. (In Jour- al of the Arnold arboretum, — xv, 255- bie for students. [Revie of “vito to cytology,” by L. W. Sharp.] (In Science, 1934, Ixxx, 407.) Structure of meiotic chromosomes in sic eee of Trades- cantia. By Karl Sax and L. M. Humphrey. (In Botanical gazette, 2.) Chromosome numbers in the Hamamelidaceae and their phylogenetic significance. By Edgar Anderson and Karl Sax. (In Journal of the Arnold arboretum, 1935, xvi, 210-215. ) Chromosome structure in the meiotic chromoso of Rhoeo discolor Hance. (In Journal of the Arnold aed, 1935, Xvi, 216-224. ) The aie en analysis of species-hybrids. (In Botanical review, 1935, i, 17. Variation in chias frequencies in Secale, Vicia and Tradescantia. (In Cytologia, 1035, vi, 289-293.) ss i Louis Victor. nter hardiness of trees and shrubs growing in the Arnold arbore- ein By J. H. Faull, J. G. Jack, W. H. Judd and L. V. cree (In Arnold arboretum bulletin of popular information, 1934, i 29-47, 53-60. WHITAKER, Thomas Wallac An improved technic for ‘the examination of aa in root tip smears. (In Stain technology, 1934, ix, 107-1 aryo-systematic study of Robinia. (In Journal of the Arnold arbo- retum, 1934, xv, 353-357. The shrubby ne (In Arnold arboretum bulletin of popular information, 1934, ii, 25-28.) The above articles cover a total of about 1037 pages. — E. M. T. 1935] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 469 Staff of the Arnold Arboretum, 1934-35 Oakes Ames, A.M., Arnold Professor of Botany, Supervisor. Joun GEorGE Jack, Assistant Professor of Dendrology. ALFRED REHDER, A.M., Associate Professor of Dendrology and Curator of the Herbarium. Josepu H. Fautt, Ph.D., Professor of Forest Pathology. IrnvING WipMe_eR BalItey, S.D., Professor of Plant Anatomy. Kart Sax, Ph.D., Associate Professor of Cytology. EpcarR ANDERSON, S.D., Arborist. Ivan Murray JOHNSTON, Ph.D., Research Associate. CLARENCE E. Kosusk1, Ph.D., Assistant Curator, Herbarium. Hucu M. Raup, Ph.D., Research Associate. ETHELYN Maria TUCKER, Librarian. ERNEST J. PALMER, Collector and Research Assistant. Susan Detano McKeE vey, Research Assistant. CaROoLINE K. ALLEN, Ph.D., Assistant in the Herbarium. EtHEL ANTOINETTE ANDERSON, Business Secretary. Louis VicToR SCHMITT, Superintendent. WILLIAM HENrRy Jupp, Propagator. 470 JOURNAL OF THE ARNOLD ARBORETUM [VOL XVI ERRATA Page 36, line 12 from below for (C. macrostachia) read (V. macro- stachia). “42, line 3 for 5- 10 cm. read 5-10 mm. ‘46, line 16 from below for east-northeast read east-northwest. ‘““ 38, line 10 for style sessile read stigma sessile. “94, line 3 for Ampelopsis humulifolia Bunge read A mpelopsis brevipedunculata (Dipp.) Koehne var. Maximowiczii Rehd. “186, line 15 from below for Heliotropium transandinum var. tiaridioides read Heliotropium transalpinum var. tiaridi- oides. INDEX Synonyms are printed in italics; new names in bold-face type Abelia Bodinieri, 334 — Cavaleriei, 335 — deutziaefolia, 334 334 — Mairei myrtilloides, 334 eas par , 4 racemosa, 319 ee eee ao 315 Agrostis can Akebia penta 358 — qui x trifoliata, 358 — cara pentaphylla, 358 Allium canadense, 90 Allocarya alternifolia, 193 — ones ot Bie Alomia sere Amelanchier § ee ee 449 morpha fruti Amit is ee Maximo- wiczil, 470 Amine ce 203 — Douglasiana, 197, a 203 a intermedia 197, 2 — Le ii, 202 —]1 yess 199 1 An endemic Sooo from Rumania, 76 pl. 123, 124, Anchusa Sellowiana, 194 ANDERSON, Epcar, An endemic Sophora rom ahylogeeti significance, 210, figs. mone qu patina 92 pee: canadensis, 9 — neglecta, 96 — plantaginifolia, . Antirhea bifida, 1 wo andhee ce Apios americana, 93 — tuberosa, 93 Aquilegia vulgaris, 92 Aralia hispida, 94 Arctostaphylos novoleontis, 450 Argusia, 161 Ar (ines 161 mme — nae ee 164 — oki 64 osmarinifolia, 164 avs create 88 ae Clematiin 91 — Kaempfer Arnold cannes Bibliography of the publis: Kajewski, on the, 206, pl. 1 . arborea, New hybrids A the heats asta 1934-35, Staff of the, 469 Aronia arbutifolia, 92 472 JOURNAL OF THE ARNOLD ARBORETUM Aster acuminatus, 96 —_ aria —" 96 — multi Atriplex ek 443 Avicennia marina, 70 — — resinifera, 70 Battey, I. W. and Kerr, THomas, The visible structure of the secondary wall d its significance in physical and chemical investigations of tracheary cells and fibers, 273, pl. 140-149 BAKHUIZEN VAN DEN Brink, R tes on some of the Ebenaceae aa collected on the Arnol pe Expedition, 1930-1932, 68, pl. 120-12 Benthamia lycopsoides, Boltonia asteroides, 96 Boraginaceae, X, Studies in; The oraginaceae of northeastern South America, 1 Werariicess, XI, Studies in the, 145 Borellia asper, Brandisia racemosa, 315 Brass and S WSKI, Lorantha- Expedition, 1930-1932, 206, pl. 129 Buchenavia capitata, 45 Buddleia oe 311 Callicarpa erie ulata, 70 — pentan 2: furfurace, 70 Calllixiche ier 93 — palustris, 93 Canthium Cavaleriei, 324 32 — Dunnianu 3 — Henryi, 328 Labordei, 323 Carex angustior, 89 — brevior, — Goodenowii, 89 [ VOL. XvI Carex laxiflora gracillima, 89 — — leptonervia, 89 — longirostris, 90 — Shc 90 — panicea, 89 — s eaaeeniia lucorum, 89 — stellulata angustata, 89 -—- alaede 89 — ten — varia a Carya mexicana polyneura, 448 — ovalis, 9 pabsarntek o_o 313 — fluminis, Cavaleriella Ge 333 Celastrus ioral eae ,9 Cephalan sce Cooatarak 319 — Nav i ee Cephalosporium wilt of elms, New facts me numbers in the Hamameli- daceae and their phylogenetic signifi- cance, 210, bhaineann erage and behavior in mitosis and meiosis, 423, pl. 161-164 Chromosome structure in the meiotic chromosom a oeo iscolor Hance, 2 a 130, 131 FS ey ita epee 96 lematis paniculata, 92 ndron amicorum, 71 Clerode — Buchanani, 70 — confusum — inerme, 7] Coldenia conspicua, 183 ecumbens, 185 Paes allii, 185 Iquhounia elegans pauciflora, 312 Cordia, 3 — aberrans, 173 — alliodora, — axillaris gymnocarpa, 179 1935] Cordia bahiensis, 43 — Braceliniae, 177 — calophylla, 1 — campestris, 178 — melanoneura, 30 — x, 10 — flavescens, 44 — formicarum, 13 — fulva 2 — galapagensis, 176 — hebecarpa, 21 — heterophylla, 21 — hirsuta, 13 — hirt — hispidissima, 13 ookeri 8 _ enkonbvctis 176 — linearis, 174 — Lockarti, 23 — lucida, 39 — mucronata, 30, 173 INDEX ‘Cordia multispicata, 43 — Neowidiana, 177 — nervosa, 18 — nodosa, 13 —— angustifolia, 14 — revoluta, 174 — ripicola, 180 — rufa, 32 — Sagotii, 2 — waa 44 — scabri — pene 22 — ucayaliensis, 181 Cotoneaster § M Eas 449 474 JOURNAL OF THE ARNOLD ARBORETUM Cotoneaster § Naegelia, 449 Crataegus or aaa Two new species of, 3 Crataegus Daniels, 355, fig. 2 — — glabra = tanetintensia, 353, fig. 1 w facts concerning Gachalsenscinsi wilt of elms, 45 Crepis capillaris, 97 Cristaria insularis, 44 CrowELL, Ivan H., The hosts, life his- tory and control of Gymnosporangium clavipes C. and P., 367, pl. 155-160 ah bros oe 171 albida, —— ne 169 — Candolleana, 170 —— saaslealitians:. compacta, 170 —— circumscissa, —- 195 -— eee a 171 1 —- ane ine 172 — diffusa, — diplotricha, 7 171, 172 — globulifera, 170, 171 — glomerata, 170, 172 — glomerata, 195 — glomerulifera, 169 — grandiflora anulata, 172 — granulosa, 172 — hispida, 172 [ VOL. XvI Bs Cacia — 172 — hisp — limensis, 172 m yunnanense, 325 Cyperus esculentus, 89 Cyphomattia lanata, 194 Cypripedium acaule, 90 Cystacanthus yangtsekiangensis, 315 Dactyliophora salomonia, 206, pl. 129 a- Danser, B. H., Loranthaceae collected in the ees Islands by n F. Kajewski, on th ‘Kbit ge 1930-1932, 206, pl. 12 Darker, Grant D., Hypodermella Hiratsukae, a new species of Hypo- dermataceae from Japan, 364, pl. 154 5 — Stramonium, 95 Dendrophthoé falcata, 209 Deutzia funebris, 315 1935] Dichapetalum scandens, 44 — vestitum scandens, 44 Diervilla Lonicera, 95 Diospyros ellipticifolia, 68 — ferrea, 68 — — salomonensis, 68 — insularis, se yo) Oey{opee Vat — maritima 92 Dispersal of cable basidiospores of the t Gymnosporangium rusts, The, 411, figs. Echinochloa crusgalli longiseta, 88 uricata, Echium vulgare, 95 Effect of temperature on nuclear differ- entiation in microspore development, The, 301, pl. 150, fig Ehretia Esquirolii, 320 Elaioplasts in Iris: a morphological study, 225, pl. 132-137 Elms, New facts concerning Cephalo- s ilt of, 453 Elsholtzia Cavaleriez, 311 — Dielsii, 313 — fruticosa, 312 mmenopterys Henryi, 318 Epifagus virginiana a, 95 Epilobium h senate 94 Eragrostis cilianensis, 88 — megastachya, 88 — pectinacea, 89 — peruviana, 443 ine Rengifoanum, 169 nium americanum, 90 —— a Smears 34 48 INDEX 475 Eurya sandwicensis Ae 352 Evonymus obovatus, 93 Faradaya amicorum, 71 — — salomonensis, 71 Fautt, ANNA F., Plaoolat in Iris: a morphological atnay; 225, pl. 132-137 Festuca ovina, 89 — rubra commutata, 89 Firensia, 44 — hirsuta, 13 — lutea, Flora of San Felix Island, The, 440, pl. 165 Galium asprellum, 95 Gardenia florida, 321 c , Gillenia trifoliata, 92 lyceri 2 eteniionai 89 Gmelina moluccana, 72 — salomonensis, en pl. 122 eae pian clavipes C. and P., The hosts, me ae and control of, 367, pl. 155-1 Gymnosporangium globosum Farl. and their relative ace uiney The hosts of, 98, 25-128, i pee ae ru | The dispersal f viable Se of the, 411, figs. Hackelia caerulescens, 194 een caerulescens, 194 , 194 Se 194 — patens, 194 Hamamelidaceae and their phylogenetic Cee Chromosome numbers in the, 210, figs. Bs ear, a new genus of Sapin- — Bodinieri, 66, pl. 119, fig. Havilandia opac — papuana, 191 — robusta, 191 Hedyotis Bodinieri, 316 476 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XVI Hedyotis oo. 316 Houstonia caerulea, 95 ; Hunt Luin M., A preliminary — ees a4 note on life rei ee of Euro- — yunnanensis, 332 pean species of Milesia, 143 Helenium nudiflorum, 96 Huodendron, a new see of Styraca- Heliophytum —— 59 ceae, 341, pl. 151, 152, — passerinoides, Huodendron, 34 Heliopsis pens 96 — biaristatum, 344 Heliotropium, 57 — — parviflorum, 346 — angiospermum, 186 — tibeticum, 342 — curassavicum, 57 Hypochaeris radicata — filiforme, 62 Hypodermella Hir ae a new species ~- Lean 61 of Hypodermataceae from Japan, 364, — Fum pl. 154 — cae 56, 165 Hypodermella Hiratsukae, 364 — helophilum, 63 Incarvillea Delavayi, 315 — humile, 186 ris: Elaioplasts a morphological — indicum, 59 study, 225, a - 137 — inundatum, 6 Ixora Henryl, 3 — japonicum, 164 JOHNSTON, aie - Studies in Boragi- — ania 58 naceae, X; The ee of — lineare, 167 northeastern South America, 1 — messerschmidioides, 166, 167, 168 — Studies in the each XI, 145 — micranthos, 167 — The see of San Felix Island, 440, — pannifolium, 168 pl. 165 — procumbens, 60 Juniperus communis depressa, 88 — salicoides, 62 Kajewski, Loranthaceae collected in the — scandens, 1 Solomon Islands by L. an — ternatum, 61 S. F. . on — Aik peers — — Fumana, 62 Repedition, 1930-1932, 206, pl. 129 — tiaridioides, 186 Kerr, THomas, and Battey, I. W., T — — genui 186 visible structure of i secondary apo 186 wall and its significance in physical — transalpinum, 186 and chemical investigations of trache- ——ti aridioides, 470 ary cells and fibers, 273, pl. 140-149 — trinitense, KosuskI, CLARENCE E., Studies in Thea- ae niin 167 ceae, I, Eurya subgen. Ternstroemi- — viridiflorum, 15 opsis, 347, pl. 153 — zeylanicum, ‘167, 168 Lactuca sagittifolia, 97 Sp., 61 — scariola, 97 Hibiscus Trionum, 94 —- ae integtifolia, 97 Hieracium florentinum, 97 Lagenaria leucantha, 96 — vulgatum, Lantana cor pighends 3 Hosts of Gymnosporangium globosum Lappula coerulescens, 194 Farl. and ue sina susceptibility, © — oe 194 The, 98, pl. 125-128, figs. — echinata, 194 Hosts, life history ei control of Gym- — grisea, 194 nosporangium clavipes C. and P., The, |. —subdecumbens coerulescens, 194 367, pl. 155-160 Lasianthus Biermanni, 3 1935] INDEX Lasianthus Dunniana, 323 Lithocardium mirandum, 13 — Esquirolii, — multispicatum, 43 — Hartii, 323 — muneco, 1 — Hookeri, 323 — Neowiedianum, 177 aes 323 — nervosum, 18 sp., — nodosum, 1 eae pinetorum, 187 — oxyphyllum, 36 Lathyrus pannonicus versicolor, 93 — paniculare, 26 Lechea intermedia, 94 — pubescens, 21 — tenuifolia, 94 — revoluta, 174 ee ioe : — scabridum, 30 — punctata, — Schomburgkii, 39 eae an 328 — Scouleri, 175 — Esquirolii, 326 — sericicalyx, 25 — Mai Z — Sprucei, 17 — peeneiadgle 327 — tetrandrum, 11 oblon — tetraphyllum, 45 — pilosa Peas 327 — tomentosum, 41 — Potanini, 326 — toqueve, — — glauca, 327 umbraculiferum, 11 _— baeouees 327 Puteaper in calycinum, 203 — ane SPH — clandestinum, 195 moi hirta, — lycopsoides, 202 Leucosceptrum eae 311 — minutum, 19 — plectranthoideum, 311 — Muelleri, 187 — sinense, 311 — viridiflorum, 153 Léveillé from eastern Asia, Notes on Litsea Muelleri, 449 the ligneous plants described by, 311 Lonicera acrophila, 337 Liatris scariosa, 96 alpigena Ligustrum Thea, 317 — — busifota, 336 vulgare, 95 — 95 Liriodendron Tulipifera, 92 - Suna 339 Lithocardium Anderssonii, 175 — Fauriei, 339 —A ii, 41 — fragilis, 337 — bicolor, 23 — Guilloni, 338 — calophyllum, 18 — gynopogon, 3 — cordifolium, 11 — Henryi setuligera, 338 — exaltatum, 29 — japonica, 339 — flavescens, 44 — Koeh a, 338 — galapagosenum, 176 — lanceolata, 337 — grandiflori 2 — ligustrina, 335 — guazumifolia, 179 — macrantha, 338 ec ye — Mairei, 339 — heterophyllum, 21 Menelii, 315 — hispidissimum, 13 — mission a) ianum, 1 — Pampaninii, 338 — leucophlyctis, 176 — pileata, 336 — Lockartii, 23 — Rocheri, 335 — Mezianum, 174 — tangutica, 335 478 Lonicera yunnanensis, 339 Loranthaceae collected in the Solomon pedition, 1930-1932, 206, pl. 129 Loranthus falcatus, 209 Luculia cimecsasinge Sly Luzula nemorosa, Lycapsus suite us, 44 Lycopersicon esculentum, 95 , J. D., The dispersal of viable ieee of the Gymno- sporangium rusts, 411, figs — The hosts of Gymnosporangium glo- bosum Farl. and their relative suscep- tibility, 98, pl. 125-128, figs Macromeria barbigera, 189 eontis, 188 Watatoue. 161 gnaphalodes, 16 Malus baccata tee 92 — verticillata crispa, 94 USAN — Notes on . 138, 93 en 2 Muelra, 451 — floribunda, 168 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XvI Mexico, Some new trees and shrubs m, Micromeria biflora, 312 Milesia, A preliminary note on life his- tory studies of European species of, 143 Mirabilis Jalapa, 91 M plied aera 36 — tome indie ae Sellowiana, 194 — Sellowiana, 194 Mussaenda Bodinieri, 320 Cavaleriei Myosti oe 192 ia, Pt ee Napaea dioica, 94 Neonauclea Navillei, 319 New facts concerning Cephalosporium wilt of elms, 453 New hybrids from the Arnold Arbore- tum, 358 Notes on some of the Ebenaceae and Verbenaceae of the Solomon Islands collected on the Arnold Arboretum xpedition, 1930-1932, 68, pl. 2 Notes on - ligneous plants described by ss é from Eastern Asia, 311 te) 268, ie 138, 139 Notts oct 209 Ocotea da. Oldentandia Bani 316 — Esquiro — Cavaleriei, 316 — violaceo-flammea, 31 Oxalis europaea villicaulis, 93 1935] Oxalis stricta, 93 Paederia Bodinieri, 325 — Cavaleriei, 326 — scandens, 3 — tomentosa, 324, 326 — Mairei, 324 — pu a -coerulea, 325 a __ Wallichii, PALMER, ie J., Supplement to the spontaneous flora of the Arnol Arbore ; — Two new ae of Crataegus from Missouri, 353, Panicum commu ine Ashei, 8 — depauperatum psilophyllum, a — dichotomiflorum, — lanuginosum Lindheimeri, 88 8 — oligosanthes Scribnerianum, 88 — Pariaenesee 88 — Scribne 8 Parietaria seb 443 —feliciana Pertya sre 340 Petraeovitex sae salomonensis, 73 teeny pubinervis, 315 Phlox paniculata, Physalis heterophylla en etinaeR 95 Plagio iobothrys calandrinioides, 193 ntago major, 95 Plectranthus con 314 — Mairei, 314 = eeeie 13 Pogostemon glaber, 313 INDEX 479 Pern sanguinea, 93 — acieun iitosachyum, 91 — Sieboldii, 9 Polystichum aaricies 87 Populus tremuloides, 91 Potamogeton epihydrus Nuttallii, 88 Potentilla anaes villosissima, 9 Prelimi ee note on life history emis ean species of Milesia, A, 143 Premna Trou 329 — ee - — Valbra Beast eo oe 328 w eee a 93 — Jud a susqueh , 93 — Sargentii x Fae 358 — yedoensis X Sargentii, 358 Z Pyrola americana, 94 Quercus ilicifolia * velutina, 91 — imbricaria X velutina, 91 —_ ar : ilicifolia, 91 — imbricaria, 91 Game pa 92 HDER, ALFRED, adeiodendion: a new genus of Sapindaceae, 65, pl. 119, fig. — Huodendron, a new genus of Styra- caceae, 341, pl. 151, 152, fig. — Notes on the ligneous plants described by Léveillé from Eastern Asia, 311 — Some new trees and shrubs from Mexico, 448 — and ANDERSON, Epcar, New hybrids from the Arnold Arboretum, 358 Rhoeo discolor Hance, Chromosome structure in the meiotic chromosomes of, 216, pl. 130, 131 480 JOURNAL OF THE ARNOLD ARBORETUM Ribes aan 92 — vulgare, 92 pie aa 194 — sp., 168 Rorippa sylvestris, 92 Rubus — 93 — flagellar — Idaeus us 92 — Jeckylanus, 93 — parvifolius, 93 = nee 93 Rumex ange a 91 sokeiiies gine 18 e€ 18 Sagina decumbens, 91 = sonereoni 90 — pyrifolia, mapraenied ae tee — javanica Argyi, 3 San niet stand sa a of, 440, pl. 165 Saponaria officinalis, 91 Sax, Harty JoLivetre and Sax, Kart, Chromosome structure and behavior in mitosis and meiosis, 423, pl. 161- 164 Sax, Karri, Chromosome structure in e meiotic sapere to) e discolor Hance, 216, pl. 1 131 — The effect of te inp on nuclear differentiation in microspore develop- ment, 301, pl. 150, fi , Epcar, Chromosome the Hamamelidaceae and their een significance, 210, figs. — and Sax, Hairy Jorivette, Chromo- some structure and behavior in mitosis and m , 423, e 161-164 Seeder py ee -— Hiei 58 — subracemosa, 62 [ VOL. XVI , 90 Zs ttlitia: 89 Sebestena aera 174 — revoluta, 174 Senecio aureus, 96 Scilla sibirica irpus Sida oe 94 Sideroxylon Bodinieri, 66 Silene antirrhina, 91 i 1 Si ansian angustifolium, 90 Sogerianthe versicolor, 207, pl. 129 c-e Solanum aculeatissimum, 314 — Bodinieri, 314 — Cav ae. 31 Solidago odora, 96 — ulmifolia, 96 oe Islands, Loranthaceae collected nt . Brass and S. F. eco. on the pea Arboretum Expedition, 1930-1932, 206, pl. 129 Solomon Islands, Notes on some of the Ebenaceae ‘il Verbenaceae of the. collected on the Arnold Asliabebiins Expedition, 1930-1932, 68, p 122 Some new trees and shrubs from Mex- ico, Sonchus arvensis glabrescens, 97 — oleracea, 97 Sophora from Rumania, An endemic, 76, pl. 123, 124, fig — oe oides osa, 78, ald , fig. eee 76, pl. 123, . fig. B Spergula arvensis, 91 ried flora of re Arnold Arbore- m,S Pree Seb be cleat 334 — hypericifolia, 334 — yangtsekiangensis, 315 Strobilanthopsis deutziaefolius, 334 — hypericifolius, 334 Studies in Boraginaceae, X; The Boragi- naceae of northeastern South Ameri- ca; 1 Studies in the Boraginaceae, XI, 145 1935] Studies in Theaceae I. Eurya subgen. Ternstroemiopsis, 347, pl. 153 Styrax biaristatus, 34 — parvi — .tibeticus, 342 Suaeda varia microphylla, 444 — nesophila, 4 ene to ae spontaneous flora of rnold Arboretum ee bona lia 362 — oblata Giraldii < spniatiioll, 362 — vulgaris, ae Tarenna incerta, 321 llissima — eri, Teucrium Se 313 Teysmanniodendron Ahernianum, 74 Thamnoseris crt 45 — lobata, 4 Thaumatearon dasyanthum Sellow- ia ? — Sellowi Thelypter penne intermedia, 87 Thymus Cavaleriei, 312 Toquera tomentosa, 21 Tourneorti se 145 — § Arguz — § ee 145 4 — acuminata, 159 — alba, 52 — angulosa, 166 — eae 166 — arborea, 164 -= Oe 159 — argentea, 164 INDEX Tournefortia Arguzia, 164 4 — Pee 52 — brasiliensis, 185 — Candollii, 154 — coriacea, 52 — cuspidata, 54 t=] a eS Bp 2 5 2 Pp On a a a] Q nn = Q = = _ wm rs — laevigata, 51 — macrophylla, 147 — melanochaeta, 55 — Messerschmidia, 167 — Meyeri, — Meyeri, 64 sm ee 164 482 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XVI Tournefortia micranthos, 167 Tournefortia Wightii, 153 — Miquelii, 51 — zeylanica, 168 — Mocquerysi, 158 Tradescantia virginiana, 90 — mollis, 167 Trientalis americana, 95 — mollis, 157 wo new pe of Crataegus from — montana, 153 Missouri, 353, figs. — Muelleri, 157 Typha latifolia, = — mutabilis, 167 Uncaria scandens, 319 — obscura, 54 Vaccinium corymbosum, 94 — orientalis, 147 Vangueria tomentosa, 167 — ovata, 153 Varronia canescens, 175 —— paniculata spigeliaeflora, 49 — corymbosa, 34 — peruviana, 48 — dichotoma, 34 — puberula, 158 — grandiflora, 32 — — Kirkii, 158 — — glabrata, 33 — reticosa, 152 — — Sprucei, 33 — rosmarinifolia, 164 — guasumaefolia, 179 — Roxburghii, 153 — guianensis, 36 — Royleana, 167 — lantanoides, 32 — Sampsoni, 153 — leucophlyctis, 176 — sarmentosa, 147 — linearis — — hirsuta, 1 — macrostachia, 470 — — magnifolia, 147 — macrostachya, 36 Schomburgkii, 52 — monosperma, 34 — setifera, 54 — polycephala, 33 — sibirica, 164 revoluta — angustior, 164 — scaberrima, 176 — spigeliaeflora, 49 — Scou — stenoraca, 167 — spicata, 43 — subulata, 167 — tomentosa, 41 — surinamensis, 48 — ulmifolia, 34 — syringaefolia, 48 Vernonia a 185 j Veronica peregrin — tetragona, 150 Viburnum sul 331 — tetrandra, 150 — barbigeru abiskitiaite: 151 — alee - ial 360 — es 150 — Bodinieri, 3 — — hirsuta, 147 — rile 328 — — longiflora, 150 — Cavaleriei, 329 — — Walkerae, 151 — Chaffanjoni, 330 oo ai 167 — congestum, 329 — Ule — corylifolium, 332 —U es 147 — crassifolium, 330 — velutina, 159 — tind. crassifolium, 330 viridifi 153 — Dunnia — — Griffithii, 154 —erosum punctatum, 333 — volubilis, 47 — — Taquetii, 333 Ww 1 — erubescens, 32 alkerae, 151 — Wallichii, 150 — foetidum, 331 1935] — sympodiale, 329 — Taquetii, 333 — ternatum, 330 — theiferum, 332 — Touchanense, 331 — trilobum, 95 Viola fimbriatula papilionacea, 94 — latiuscula, 94 — odorata, 94 — pedata lineariloba, 94 — sagittata, 94 — sororia, 94 Viburnum foetidum rectangulatum, 331 Juddii INDEX 483 Visible structure of the secondary wall cells and fibers. The, 273, pl. 140-149 Vitex Aherniana, 74 — cofassus, 7 W ebera oe 321 — salicifolia, 317 — uvariifolia Dunniana, 318 Yucca, Notes on, Yucca — brevifolia, — brevifolia Jaegeriana, 269, pl. 139 b 268, pl. 138, 139 arizonica, 270 270 erula, 270 Thornberi, 268, pl. 138 Treleasei, 270