JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY ALFRED REHDER EDITOR JOSEPH H. FAULL anp CLARENCE E. KOBUSKI ASSOCIATE EDITORS VOLUME XIV JAMAICA PLAIN, MASS. 1933. Reprinted with the permission of the Arnold Arboretum of Harvard University KRAUS REPRINT CORPORATION New York 1968 DATES OF ISSUE (pp. 1-99, pl. 55-60) issued January 25, 1933 (pp. 101-197) issued April 20, 1933 (pp. 199-334, pl. 61-71) issued July 24, 1933 (pp. 335-441, pl. 72-81) issued October 25, 1933 Printed in U.S.A. TABLE OF CONTENTS An ENUMERATION OF THE HERBACEOUS PLANTS COLLECTED BY J. F. ee FOR THE ARNOLD ARBORETUM. By Alfred Rehder and Clar- e DP DUMERIT i o8a Sawind.ms 36.09 ed 4 Maw Eo Pe eee rer Gaareeuios TO THE FLora oF THE NEw HeEpripEs; PLAnts CoL- y S. F. Kayewsk1 1n 1928 anv 1929; Supplement. By SupPLEMENT To C. T. WuiTE, LIGNEoUus PLANTS COLLECTED IN THE TERRITORY OF PAPUA (BritisH New GuINEA) IN 1925-26 BY BRASS etry ee CRO 25.6 shea Geese ek een eee ue INHERITANCE IN AN Oak Species Hyprip. With eight text figures. By TE VGN oe eek ok yk Shas OO Se Rew OME eae TAXONOMIC AND CyTOLOGICAL RELATIONSHIPS YuccA AND ele With plate 55. By Susan Delano Monee: and Karl Casi Numsers IN UtMus AND RELATED GENERA. With Ce De OE IE slic tiv dae ec a ewe aeeher eeimenm) se ae RRHIZAL AND OTHER FEATURES OF THE Root SYSTEMS OF PINUS. With plates 57-60. By A. B. Hatch and K. D. Doak ............ AppDITIONAL Notes ON THE ORCHIDS OF THE NEW HEBRIDES AND SeuNTA CHUL ISLANDS: By (OGRES Ames ids ieeyewscnwens ens ess Ture CHROMOSOME COMPLEMENT OF CYPHOMANDRA BETACEA. With six text figures. By Thomas W. Whitaker ..........+-+++0005: STUDIES ON THE PRECIPITIN REACTION IN PLantTs. III. A Bio- CHEMICAL ANALYSIS OF THE “NoRMAL PRECIPITIN REACTION.” With four text figures. By Kenneth S. Chester and Thomas W. LEE E hes sae eee eee aio. oies artes avai eyes 1 New SPECIES, VARIETIES AND COMBINATIONS FROM THE HERBARIUM AND THE COLLECTIONS OF THE ARNOLD ARBORETUM. With nine 1 text figures. By Alfred Rehder ........ce cece ecee cece ee eeenes Notes ON THE LIGNEOUS PLANTS DESCRIBED BY LEVEILLE FROM Eastern Asia. By Alfred Rehder ............ be ceecseeceecs VARIATION IN FLOWER COLOR IN HAMAMELIS VERNALIS. With two 25 text figures. By Edgar Anderson .......ceeeeerec essen cerecres Notes ON THE GENUS per THE BLACK CONE OF PINUS PONDER- Ww Tue CAMBIUM AND ITs DERIVATIVE TissuEs. No. es STRUC- TURE, DISTRIBUTION AND DIAGNOSTIC SIGNIFICANCE OF VESTURED MY 1 ah aaa ci ep Sian Peet re SPECIES Hysrips IN PLATANUS AND Campsis. With two text fig- 2 pres. By Karl Sat .occs ccastveee spawns ey ase bei so 8G one ie a oe es CHroMosoME BEHAVIOR IN CALYCANTHUS. With four text figures. By Karl S@¢ :cxisceshcyeue ees Pad ees 6k see ee NESS weees OrIGIN AND BEHAVIOR OF THE NUCLEOLUS IN PLANTs. With plates 2 64-67. By Haig Dermen. 22. ..ccccssnec cds cedessnsseccsevess Some HyMENOMYCETES FoRMING MycorRHIZAE WITH PINUS STRO- pus. With plates 68-71. By A. B. Hatch and C. Talbot Hatch . iv TABLE OF CONTENTS Notes ON THE DISTRIBUTION OF WHITE SPRUCE AND BANKSIAN PINE IN yah Canapa. With plates 72 and 73. By PURO Ti RP 4p deans EOE OR 4 eR a Oe Sees NEw esha ee AND COMBINATIONS FROM THE HERBARIUM AND THE COLLECTIONS OF THE ARNOLD ARBORETUM. With plate 74. BY SCE ROMEO 565545 i025d0ns cy eeneenee eee eeeeereeas es > THE COMPARATIVE ANATOMY OF THE STEMS OF BETULA PUMILA, BETULA LENTA AND THE Hypsrip BetruLa Jacki. By Sarah M. OUMINS. Goda ban ehc 44 oe Ce RSE Eee ba eae ous nice NUMBER AND MorRPHOLOGY IN THE CONIFER With plates 75-79 and 3 text figures. By Karl Sax and Hally he Sax 3 CHROMOSOME NUMBER AND RELATIONSHIP IN THE MAGNOLIALES. With plate 80 and 4 text figures. By Thomas W. Whitaker ...... CHROMOSOME NUMBER IN ACER AND STAPHYLEA. With plate 81. Ue ODENE Le PO as Seed 55 6s oo a oS OR RR eae STUDIES ON THE “PRECIPITIN REACTION” IN PLANTs. V. APPLICA- TION TO PLANT RELATIONSHIPS. By K. S. Chester, E. C. Abbe and eke FE OUIOU ex sae ey wae L VET AERG TETAS UERER LORE RD REE CEES 3 THE ARNOLD ARBORETUM DURING THE FISCAL YEAR ENDED JUNE 30, 1933: The Arboretum; The Pathological Laboratory; The Cyto- logical Laboratory; The Herbarium; The Library; Bibliography of ‘ ‘ 19 THE STAFF OF THE ata p ARBORETUM 1933-1934 ...........005. DA PROGR 6c aes A ee as a JOURNAL OF THE ARNOLD ARBORETUM VoLuME XIV JANUARY, 1933 NuMBER 1 AN ENUMERATION OF THE HERBACEOUS PLANTS COLLECTED BY J. F. ROCK FOR THE ARNOLD ARBORETUM! ALFRED REHDER and CLARENCE E. KOBUSKI FUNGI (LICHENES) Determined at the Farlow Herbarium Cladonia gracilis (L.) Willd. var. elongata (Jacq.) Floerke Southwestern Kansu: Tao River basin, nos. 13187, 14937. Lobaria pulmonaria (L.) Hoffm. var. hypomela Crombie Southwestern Kansu: Lower Tebbu country, nos. 14848, 14867. Peltigera aphthosa (L.) Willd. Southwestern Kansu: Tao River basin, no. 13188. Sticta Henryana Muell.-Arg. Southwestern Kansu: Lower Tebbu country, no. 14839. POLY PODIACEAE Determined by E. B. CopELAND Woodsia lanosa Hook. Southwestern Kansu: Lower Tebbu country, no. 14879. Woodsia macrospora C. Christensen & Maxon in Jour. Wash. Acad. Sci. xvit. 499 (1927). Southwestern Kansu: Tao River basin, no. 13712; Lower Tebbu country, no. 14780 Cystopteris montana (Lam.) Bernh. Southwestern Kansu: Lower Tebbu country, no. 14871. Pteretis Struthiopteris (L.) Niewl. Southwestern Kansu: Lower Tebbu country, no. 14731. 1For the enumeration . Bay poke ere collected for the Arnold Arboretum by J. F. Rock see Jour. A d Arb. 1 27, 37-125 (1928); xr. 385-409 (1932). Except when stated eee ie pant in the following enumeration have been sapprioea ed at the Botanic Museu Berlin-Dahlem, New species and varieties based on cael collected by J. F. Rock are marked with an eelerik (*). JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Dryopteris paleacea (Swartz) C. Chr. var. khasiana (Clarke) C. Christensen in Contrib. U. S. Nat. Herb. xxvr. 280 (1931). Dryopterts filix mas (L.) Schott var. khasiana Clarke Southwestern Kansu: Lower Tebbu country, no. 14841. Dryopteris Robertiana (Hoffm.) C. Chr. Southwestern Kansu: Lower Tebbu country, no. 14778. Polystichum Braunii (Spenn.) Fée Southwestern Kansu: Lower Tebbu country, no. 14840. Polystichum molliculum Christ Southwestern Kansu: Tao River basin, no. 13135; Lower Tebbu country, no. 14822. Eastern Tibet: Radja and Yellow River gorges, no. 14027 (in part). Asplenium varians Wall. Eastern Tibet: Radja and Yellow River gorges, no. 14027 (in part). Athyrium acrostichoides (Sw.) Diels Southwestern Kansu: Lower Tebbu country, nos. 14842, 14869. Athyrium filix femina (L.) Roth Southwestern Kansu: Upper Tebbu country, no. 14634; Lower Tebbu country, no. 14697. Athyrium filix femina (L.) Roth var. eyclosorum Rupr. Southwestern Kansu: Lower Tebbu country, no. 14705. Athyrium spinulosum (Maxim.) Milde Southwestern Kansu: Lower Tebbu country, no. 14779, Notholaena Delavayi (Baker) C. Christensen in Contrib. U. S. Nat. Herb. xxvi. 307 (1931). Gymnogramme Delavayi Baker Gymnopteris Delavayi Underw. Southwestern Kansu: Lower Tebbu country, no. 14881. Cheilanthes argentea (Gmel.) Kze. Eastern Tibet: grasslands between Labrang and Yellow River, no. 13910. Adiantum latedeltoideum (Christ) C. Christensen in Act. Hort. Gothob. 1. 94 (1924). Adiantum monochlamys var. latedeltoideum Christ Southwestern Kansu: Lower Tebbu country, no. 14714, Adiantum pedatum L. Southwestern Kansu: Tao River basin, no. 14647; Lower Tebbu country, no. 14870. Polypodium clathratum C. B. Clarke Eastern Tibet: Radja and Yellow River gorges, no. 14028. South- western Kansu: Lower Tebbu country, no. 14880. Polypodium eilophyllum Diels in Act. Hort, Gothob. 1. 100 (1924). Southwestern Kansu: Lower Tebbu country, no. 14719. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK Polypodium lineare Thunb. Southwestern Kansu: Lower Tebbu country, no. 14854. Polypodium lineare Thunb. var. Southwestern Kansu: Tao River basin, no. 12140. Cyclophorus taeniodes C. Chr. Southwestern Kansu: Tao River basin, no. 12085. Cyclophorus sticticus (Kze.) C. Chr. Southwestern Kansu: Lower Tebbu country, no. 14715. Cyclophorus pekinensis C. Chr. Southwestern Kansu: Tao River basin, no. 12139. LYCOPODIACEAE Selaginella spec. Southwestern Kansu: Tao River basin, no. 12085 A. EQUISETACEAE Equisetum arvense L. Western Szechuan: no. 12040. Equisetum spec. Eastern Tibet: Radja and Yellow River gorges, no. 13984. JUNCAGINACEAE Triglochin maritimum L. Southwestern Kansu: Tao River basin, no. 12273. Eastern Tibet: Radja and Yellow River gorges, no. 14009; Ba Valley, no. 14246. GRAMINEAE Determined by A. S. Hitcucock Torresia odorata (L.) Hitchc. Eastern Tibet: Radja and Yellow River gorges, no. 14107. Stipa conferta Poir. Eastern Tibet: Kokonor region, no. 13383. Stipa splendens Trin. Eastern Tibet: Kokonor region, no. 13393. Stipa mongholica Turcz. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14530. Aira caespitosa L. Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14454, 14456, 14528. Trisetum spicatum Richt. Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14479, 14488. JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Trisetum spec. Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14457, 14473. Beckmannia erucaeformis (L.) Host Central Kansu: no. 12680. Phragmites communis Trin. Eastern Tibet: Kokonor region, no. 13387. Koeleria cristata (L.) Pers. Eastern Tibet: grasslands between Labrang and Yellow River, no. Poa attenuata Trin. Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14446, 14477, 14480. Poa bulbosa L. Southwestern Kansu: Tao River basin, no, 13741. Poa flexuosa Muhl. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14452. Poa sphondylodes Trin. Southwestern Kansu: Tao River basin, nos. 12562, 13177. Poa aff. P. arctica R. Br. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14453, Poa spec. Eastern Tibet: grasslands between Labrang and Yellow River, no. Festuca ovina L. Eastern Tibet: grasslands between Labrang and Yellow River, no. Agropyron longearistatum Boiss. Southwestern Kansu: Tao River basin, no. 13711. Elymus sibiricus L. Eastern Tibet: grasslands between Labrang and Yellow River, no. 44 CY PERACEAE Determined by G. KUKENTHAL Eriophorum comosum Nees Western Szechuan: no. 12023. *Kobresia kansuensis Kukenthal in Notizbl. Bot. Gard. Mus. Ber- lin, x. 882 (1930). Southwestern Kansu: Tao River basin, no. 13714 (holotype). Kobresia Prattii C. B. Clarke 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 5 Eastern Tibet: grasslands between Labrang and Yellow River, no. Carex atrata L. var. aterrima (Hoppe) Hartm. Southwestern Kansu: Tao River basin, no. 13726. Carex atrata L. ssp. pullata (Boott) Kukenth. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 62 Carex atrofusca Schkuhr Southwestern Kansu: Tao River basin, no. 12311. ARACEAE Arisaema consanguineum Schott f. latisectum Engl. Southwestern Kansu: Tao River basin, no. 12431. JUNCACEAE Juncus leucomelas Royle Central Kansu: no. 12715. Juncus Thomsoni Buchen. Eastern Tibet: Ba Valley, no. 14248. LILIACEAE Hemerocallis Dumortieri Morr. Central Kansu: no. 12770. Gagea pauciflora Turcz. Southwestern Kansu: Tao River basin, no. 12093. Allium chrysanthum Regel Southwestern Kansu: Tao River basin, nos. 13704, 14622. Allium cyaneum Regel Western Kansu: nos. 13252, 13253. Southwestern Kansu: Tao River basin, no. 13722. Allium Henryi C. H. Wright Southwestern Kansu: Upper Tebbu country, no. 13048. Allium kansuense Regel Southwestern Kansu: Tao River basin, no. 13728; Lower Tebbu country, nos. 14885, 14887. Allium monadelphum Turcz. var. thibeticum Regel Southwestern Kansu: Tao River basin, no. 12401. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14389. Allium polyrhizum Turcz. Eastern Tibet: Ba Valley, no. 14364. Allium Przewalskianum Regel Southwestern Kansu: Tao River basin, no. 12588. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Allium tanguticum Regel Southwestern Kansu: Tao River basin, no. 12903. Allium victorialis L. Southwestern Kansu: Tao River basin, no. 12847. Allium spec. Southwestern Kansu: Tao River basin, nos. 12616, 12852; Upper Tebbu country: no. 13037. Central Kansu: no, 12789. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14449; alpine region between Radja and Jupar ranges, nos. 14095, 14389. Lilium Davidi Duch. Southwestern Kansu: Tao River basin, no. 12592 A. Lilium Duchartrei Franch. Central Kansu: no. 13618. Lilium Duchartrei Franch. var. Farreri Krause Southwestern Kansu: Tao River basin, no. 13145; Upper Tebbu country, no. 14608. Central Kansu: nos. 12666, 12668. Lilium tenuifolium Fisch. Southwestern Kansu: Tao River basin, nos. 12589, 12592, 13606. Fritillaria cirrhosa Don var. ecirrhosa Franch. Southwestern Kansu: Tao River basin, 12197, 12481; Upper Tebbu country, no. 13085. Fritillaria Roylei Hook. Eastern Tibet: Radja and Yellow River gorges, no. 14102; alpine region between Radja and Jupar ranges, nos. 14160, 14225. Lloydia tibetica Bak. var. lutescens Franch. Southwestern Kansu: Tao River basin, no. 12600. Lloydia tibetica Bak. var. purpurascens Franch. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14143. Asparagus brachyphyllus Turcz. Southwestern Kansu: Tao River basin, nos. 12172, 12915. Asparagus trichophyllus Bee. Southwestern Kansu: Tao River basin, no. 12176. Clintonia udensis Trautv. & Mey. Southwestern Kansu: Lower Tebbu country, no. 14812. Polygonatum bulbosum Lévl. Eastern Tibet: Radja and Yellow River gorges, nos. 13987, 14082. Polygonatum sibiricum Ledeb. Southwestern Kansu: Tao River basin, nos. 12561, 12639. Polygonatum spec. Southwestern Kansu: Tao River basin, nos. 12241, 12315. Eastern Tibet: Radja and Yellow River gorges, no. 14106. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 7 Paris polyphylla Sm. Southwestern Kansu: Tao River basin, no. 12224. DIVUSCOREACEAE Dioscorea quinqueloba Thunb. Southwestern Kansu: Lower Tebbu country, no. 14752. IRIDACEAE Iris dichotoma Pall. Southwestern Kansu: Lower Tebbu country,-no. 14746. Iris ensata Thunb. Southwestern Kansu: Tao River basin, nos. 12111, 12323; Upper Tebbu country, no. 13637. Iris gracilis Maxim. Southwestern Kansu: Tao River basin, no. 12340. Eastern Tibet: Radja and Yellow River gorges, nos. 13983, 14078. Iris Potanini Maxim. Eastern Tibet: Radja and Yellow River gorges, nos. 13945, 13964, 13975; Jupar Range, no. 14338. Iris tenuifolia Pall. Southwestern Kansu: Tao River basin, nos. 12100, 12101. Eastern Tibet: Radja and Yellow River gorges, no. 13950. Iris spec. Southwestern Kansu: Tao River basin, nos. 12138, 12851. ORCHIDACEAE Determined by C. SCHWEINFURTH Cypripedium luteum Franch. Southwestern Kansu: Upper Tebbu country, no. 12536. Cypripedium nutans Schlecht. Southwestern Kansu: Upper Tebbu country, no. 12499. Cypripedium tibeticum King. Southwestern Kansu: Upper Tebbu country, no. 12521. Eastern Tibet: Radja and Yellow River gorges, nos. 13989, 14165. Aceratorchis tschiliensis Schlechter Eastern Tibet: Radja and Yellow River gorges, no. 14190. Orchis chusua D. Don. Central Kansu: no. 12735. Southwestern Kansu: Tao River basin, no. 12833; Upper Tebbu country, no. 13115. Orchis salina Turcz. Eastern Tibet: Radja and Yellow River gorges, no. 13978; Ba Valley, nos. 14251, 14356. Orchis spathulata Reichb. JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Central Kansu: no. 12713. Amitostigma monanthum (Finet) Schlechter Southwestern Kansu: Tao River basin, no. 12622. Herminium tanguticum Rolfe Central Kansu: no. 12736. Habenaria conopsea Benth. Southwestern Kansu: Tao River basin, no. 12944; Upper Tebbu country, no. 13120. Habenaria cucullata (L.) Hoefft Southwestern Kansu: Lower Tebbu country, no. 14700. Habenaria spiranthiformis Ames & Schlechter Eastern Tibet: Radja and Yellow River gorges, no. 14210. “Amnesia longibracteata Schweinfurth in Jour. Arnold Arb. x. 172 (1929). Central Kansu: no. 13216 (holotype). *Oreorchis Rockii Schweinfurth in Jour. Arnold Arb. x. 173 (1929). Central Kansu: no. 12744 (holotype). MORACEAE Humulus lupulus L. Southwestern Kansu: Tao River basin, nos. 13207, 13716; Lower Tebbu country, no. 14753. URTICACEAE Urtica dioica L. Eastern Tibet: Radja and Yellow River gorges, no. 14024. POLYGONACEAE Rheum acuminatum Hook. f. & Thoms. Central Kansu: no. 12673. Rheum palmatum L. Eastern Tibet: Radja and Yellow River gorges, nos. 14071, 14167; Jupar Range, no. 14313. Rheum palmatum L. floribus rubris Southwestern Kansu: Tao River basin, no. 12844. Eastern Tibet: Radja and Yellow River gorges, no. 13982. Rheum pumilum Maxim. Southwestern Kansu: Tao River basin, no. 12409b. Eastern Tibet: Radja and Yellow River gorges, nos. 14038, 14173. Rheum spiciforme Royle Eastern Tibet: Jupar Range, no. 14330. Polygonum Hookeri Meisn. Southwestern Kansu: Tao River basin, nos. 12409a, 12604. East- 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 9 ern Tibet: alpine region between Radja and Jupar ranges, nos. 14221, 14376. Polygonum sphaerostachyum Meisn. var. Eastern Tibet: Jupar Range, no. 14335. CHENOPODIACEAE Eurotia ceratoides C. A. Mey. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14545, CARYOPHYLLACEAE Arenaria kansuensis Maxim. Southwestern Kansu: Upper Tebbu country, no. 13043. Eastern Tibet: alpine region between Radja and Jupar ranges, nos. 14149, 14222 Arenaria melanandra (Maxim.) Mattfeld apud Handel-Mazzetti, Symb. Sin. vir. 202 (1929). Eastern Tibet: alpine region between Radja and Jupar ranges, nos. 14157, 14379. Arenaria Przewalskii Maxim. Southwestern Kansu: Upper Tebbu country, no. 13082. Eastern Tibet: Amnyi Machen range, no. 14431; grasslands between La- brang and Yellow River, no. 14508. Arenaria spec. Southwestern Kansu: Upper Tebbu country, no. 13083. Silene Fortunei Vis. Southwestern Kansu: Lower Tebbu country, no. 14745. Silene repens Patrin Southwestern Kansu: Tao River basin, no. 12865. Silene tenuis Willd. Eastern Tibet: Radja and Yellow River gorges, no. 14209. Melandryum apetalum (L.) Fenzl Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14476, 14500. Melandryum apetalum (L.) Fenzl forma Central Kansu: no. 12737. Melandryum glandulosum F. N. Williams Eastern Tibet: Radja and Yellow River gorges, no. 14121. Gypsophila Gmelini Bge. Central Kansu: no. 13236. Dianthus chinensis L. Southwestern Kansu: Tao River basin, nos. 13209, 14650. Central Kansu: no. 12752, 10 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Dianthus superbus L. Central Kansu: no. 12686. Southwestern Kansu: Tao River basin, nos. 12892, 13155. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14540. RANUNCULACEAE Paeonia anomala L. Southwestern Kansu: Tao River basin, nos. 12260, 12276, 12456, 13593. Paeonia Veitchii Lynch Southwestern Kansu: Tao River basin, nos. 12235, 13127. Caltha palustris L. Southwestern Kansu: Tao River basin, no. 12229. *Caltha scaposa Hook. f. & Thoms. var. Smithii Ulbrich in Notizbl. Bot. Gard. Mus. Berlin, x. 864 (1929). Southwestern Kansu: Tao River basin, no. 12317 (syntype). Trollius pumilus D. Don Southwestern Kansu: Tao River basin, nos. 12354, 12606, 14623; Upper Tebbu country, no. 13002. Central Kansu: no. 12697. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14220. *Trollius pumilus D. Don var. alpinus Ulbrich in Notizbl. Bot. Gard. Mus. Berlin, x. 865 (1929). Eastern Tibet: Radja and Yellow River gorges, nos. 13976, 13988, 14050 (syntypes). Helleborus thibetanus Franch. Southern Kansu: no. 12059. Paraquilegia anemonoides (Willd.) Ulbr. Southwestern Kansu: Tao River basin, nos. 12393, 12629, 12812; Upper Tebbu country, nos. 12494, 12514. Eastern Tibet: Radja and Yellow River gorges, nos. 13973, 14067, 14068, 14069, *Urophysa Rockii Ullbrich, n. gen. & sp., in Notizbl. Bot. Gard. ‘Mus. Berlin, x. 869 (1929). Western Szechuan: no. 12015 (holotype). Souliea vaginata (Maxim.) Franch. Southwestern Kansu: Tao River basin, nos. 12304, 12412. Actaea spicata L. var. erythrocarpa Ledeb. Southwestern Kansu: Lower Tebbu country, no. 14849. Aquilegia ecalcarata Maxim. Southwestern Kansu: Tao River basin, nos. 12206, 12439, 12945. Aquilegia oxysepala Trautv. & Mey. Southwestern Kansu: Tao River basin, no. 12887. Delphinium albocoeruleum Maxim. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 11 Southwestern Kansu: Tao River basin, no. 13194. Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14459, 14467, 14501, 14542, 14543. Delphinium coelestinum Franch. Southwestern Kansu: Tao River basin, no. 14625. Delphinium densiflorum Duthie Eastern Tibet: grasslands between Labrang and Yellow River, no. Delphinium Forrestii Diels Southwestern Kansu: Tao River basin, no. 13740. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14535.. Delphinium grandiflorum L. Southwestern Kansu: Tao River basin, no. 13178; Upper Tebbu country, no. 14594. Delphinium Henryi Franch. Southwestern Kansu: Tao River basin, nos. 13154, 13193. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14547. *Delphinium labrangense Ulbrich, spec. nov. ined. Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14482, 14483, 14504, 14513. Delphinium Maximowiczii Franch. Southwestern Kansu: Upper Tebbu country, nos. 13086, 13093, 14644. Delphinium Pylzowi Maxim. Southwestern Kansu: Upper Tebbu country, no. 13024; Tao River basin, no. 13185. Delphinium Souliei Franch. Southwestern Kansu: Upper Tebbu country, no. 13092. Eastern Tibet: Kokonor region, no. 13398; grasslands between Labrang and Yellow River, nos. 14460, 14502, 14518, 14523. Delphinium sparsiflorum Maxim. Southwestern Kansu: Tao River basin, no. 13191. Delphinium tanguticum Huth Southwestern Kansu: Upper Tebbu country, nos. 13023 (in part), 13041, 13088; Tao River basin, no. 14641. Delphinium tatsienense Franch. Eastern Tibet: Kokonor region, no. 13361; grasslands between La- brang and Yellow River, no. 14538. Southwestern Kansu: Tao River basin, nos. 13136, 13730. Delphinium tongolense Franch. Southwestern Kansu: Tao River basin, no. 13161; Upper Tebbu country, nos. 13109, 13111. JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XIV Aconitum Anthora L. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14494. Southwestern Kansu: Lower Tebbu country, no. 14677. Aconitum Anthora L. var. anthoroideum (Rehb.) Regel Southwestern Kansu: Tao River basin, nos. 13182, 13710, 13722, 13735, 14620; Upper Tebbu country, no. 13000. Eastern Tibet: Kokonor region, no. 13363. Aconitum gymnandrum Maxim. Southwestern Kansu: Tao River basin, nos. 12183, 12242, 12285, 12351, 12877. Aconitum laeve Royle Central Kansu: no. 12739. Southwestern Kansu: Upper Tebbu country, no. 13079; Tao River basin, 13162. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14541. Aconitum Napellus L. var. semigaleatum Pall. Southwestern Kansu: Tao River basin, no. 13721. Aconitum rotundifolium K. & K. var. tanguticum Maxim. Southwestern Kansu: Tao River basin, nos. 14621, 14643; Upper Tebbu country, nos. 13023 (in part), 13039, 13050. Central Kan- su: no. 12694. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14515; Amnyi Machen range, no. 14440. Aconitum volubile Pall. Southwestern Kansu: Tao River basin, nos. 13205, 13743; Lower Tebbu country, no. 14692. Aconitum volubile Pall. var. flexuosum (Rchb.) Rap. Southwestern Kansu: Tao River basin, nos. 14615, 14860. Aconitum volubile Pall. var. Southwestern Kansu: Tao River basin, no. 14649. Anemone demissa Hook. f. & Thoms. Southwestern Kansu: Tao River basin, nos. 12596, 12838. Anemone imbricata Maxim. Eastern Tibet: Radja and Yellow River gorges, nos. 14049, 14146; alpine region between Radja and Jupar ranges, no. 14231; Jupar Range, no. 14371; Amnyi Machen range, no. 14421. Anemone japonica (Thunb.) S. & Z. var. tomentosa Maxim. Central Kansu: no. 12766. Anemone narcissiflora L. Southwestern Kansu: Upper Tebbu country, no. 13625. Anemone rivularis Buch.-Ham. Southwestern Kansu: Tao River basin, no. 12357. Eastern Tibet: Radja and Yellow River gorges, no. 14128. “Anemone Rockii Ulbrich in Notizbl. Bot. Gard. Mus. Berlin, x. 877 (1929). 19331 REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 13 Southwestern Kansu: Tao River basin, nos. 12408a, 12487 (syn- types); Upper Tebbu country, nos. 12520, 13061, 13626 (syn- ypes). Anemone rupestris Wall. Southwestern Kansu: Tao River basin, no. 12254. Eastern Tibet: Radja and Yellow River gorges, nos. 13992, 14017, 14103, 14105, 14196. Anemone vitifolia Buch.-Ham. var. tomentosa Maxim. Southwestern Kansu: Tao River basin, no. 13213; Upper Tebbu country, no. 14609. Pulsatilla ambigua Turcz. Southwestern Kansu: Tao River basin, no. 12182. Pulsatilla spec. Eastern Tibet: Radja and Yellow River gorges, no. 14048. Oxygraphis glacialis Bge. Southwestern Kansu: Tao River basin, no. 12390. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14099. Ranunculus acris L. Southwestern Kansu: Tao River basin, no, 12566. Ranunculus affinis R. Br. Southwestern Kansu: Tao River basin, no. 12280. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14405. Ranunculus affinis R. Br. var. flabellatus Franch. Southwestern Kansu: Tao River basin, no. 12460. Ranunculus affinis R. Br. var. tangutica Maxim. Southwestern Kansu: Tao River basin, no. 12446. Ranunculus Flammula L. Southwestern Kansu: Tao River basin, no. 12282. Ranunculus japonicus Thunb. Southwestern Kansu: Tao River basin, no. 12908. Ranunculus pulchellus C. A. Mey. Southwestern Kansu: Tao River basin, no. 12281. Eastern Tibet: Radja and Yellow River gorges, no. 13991. Ranunculus pulchellus C. A. Mey. var. sericeus Hook. f. & Thoms. Eastern Tibet: Radja and Yellow River gorges, no. 14197. Ranunculus yunnanensis Franch. Southwestern Kansu: Tao River basin, no. 12342. Thalictrum alpinum L. Southwestern Kansu: Tao River basin, no. 12248. Thalictrum baicalense Turcz. Southwestern Kansu: Tao River basin, no. 12862. Thalictrum javanicum Bl. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Southwestern Kansu: Tao River basin, no. 12835. Eastern Tibet: Ba Valley, no. 14271. Thalictrum Przewalskii Maxim. Southwestern Kansu: Upper Tebbu country, no. 13054. BERBERIDACEAE Podophyllum emodi Wall. Southwestern Kansu: Tao River basin, nos. 12259, 12350, 12424, 14901. Epimedium brevicornu Maxim. Western Szechuan: no. 12016. Epimedium pubescens Maxim. Southwestern Kansu: Tao River basin, no. 12302. Leontice robusta B. Fedtsch. Southwestern Kansu: Lower Tebbu country, no. 14716. PAPAVERACEAE Hypecoum erectum L. var. lactiflorum (Kar. & Kir.) Maxim. Southwestern Kansu: Tao River basin, no. 12272. Dicranostigma spec. Southwestern Kansu: Tao River basin, no. 13144. Macleaya microcarpa (Maxim.) Fedde Southwestern Kansu: Lower Tebbu country, nos. 14655, 14690. Bocconia spec. Southwestern Kansu: Lower Tebbu country, no. 15095. Meconopsis integrifolia Franch. Southwestern Kansu: Tao River basin, nos. 12416, 13651; Upper Tebbu country, no. 13073. Eastern Tibet: Radja and Yellow River gorges, nos. 13996, 14086. Meconopsis psilonomma Farrer Southwestern Kansu: Tao River basin, no. 12613; Upper Tebbu country, no. 13632. Meconopsis punicea Maxim. Southwestern Kansu: Tao River basin, nos. 12195, 12423, 12840, 12942; Upper Tebbu country, nos. 13053, 13642. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14506. Meconopsis quintuplinervia Regel Southwestern Kansu: Tao River basin, nos. 12360, 12608, 12611, 13620; Upper Tebbu country, no. 13046. Central Kansu: 12700, 12742. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14462; alpine region between Radja and Jupar ranges, nos. 14144, 14224; Jupar Range, no. 14309. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 15 Meconopsis racemosa Maxim. Southwestern Kansu: Tao River basin, nos. 12621, 12625, 13160; Upper Tebbu country, nos. 12537, 13009, 13627. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14377; Jupar Range, no. 14339; Amnyi Machen range, no. 14434. Meconopsis spec. Southwestern Kansu: Upper Tebbu country, no. 13077. Eastern Tibet: Amnyi Machen range, no. 14430 Corydalis adunca Maxim. Eastern Tibet: Radja and Yellow River gorges, no. 14046. Corydalis curviflora Maxim. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 92 Corydalis curviflora Maxim. var. cytisiflora Fedde Southwestern Kansu: Upper Tebbu country, no. 12503. Corydalis curviflora Maxim. var. pseudo-Smithii Fedde Southwestern Kansu: Tao River basin, no. 12576 Corydalis curviflora Maxim. var. Smithii Fedde Southwestern Kansu: Tao River basin, no. 12414. Corydalis dasyptera Maxim. Southwestern Kansu: Tao River basin, nos. 12387, 12602; Upper Tebbu country, no. 13007. Eastern Tibet: Radja and Yellow River gorges, no. 14189; alpine region between Radja and Jupar ranges, no. 14237; Jupar Range, nos. 14311, 14312, 14336. Corydalis glycyphyllos Fedde Eastern Tibet: alpine region between Radja and Jupar ranges, nos. 14374, 14375 Corydalis linarioides Maxim. Southwestern Kansu: Upper Tebbu country, no. 13018. Corydalis melanochlora Maxim Southwestern Kansu: Upper Tebbu country, no. 13021. Eastern Tibet: Jupar Range, no. 14329. Corydalis Rheinbabeniana Fedde Southwestern Kansu: Upper Tebbu country, no. 13003. Eastern Tibet: Amnyi Machen range, no. 14419 Corydalis scaberula Maxim. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14226. Corydalis straminea Maxim. Southwestern Kansu: Tao River basin, nos. 12178, 12462, 13138. Corydalis stricta Steph. Eastern Tibet: Radja and Yellow River gorges, nos. 14019, 14114, 14184 16 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Corydalis trachycarpa Maxim. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14155. Corydalis spec. Eastern Tibet: Radja and Yellow River gorges, no. 14047. CRUCIFERAE Megacarpaea Delavayi Franch. Southwestern Kansu: Upper Tebbu country, no. 13087. **Megacarpaea Delavayi Franchet var. grandiflora O. E. Schulz in Notizbl. Bot. Gard. Mus. Berlin, x. 557 (1929). Southwestern Kansu: Tao River basin, no. 12597 (holotype). Dilophia fontana Maxim. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14398. Dilophia macrosperma O. E. Schulz Eastern Tibet: Amnyi Machen range, no. 14414. Cochlearia scapiflora Hook. f. & Thoms. Eastern Tibet: alpine region between Radja and Jupar ranges, nos. 14236, 14396 Eutrema compactum O. E. Schulz Eastern Tibet: Jupar Range, no. 14340. Eruca sativa Lam. var. lativalvis ( Boiss.) Coss. Southwestern Kansu: Tao River basin, no. 12179. Brassica juncea (L.) Coss. Eastern Tibet: Radja and Yellow River gorges, no. 14180. Cardamine macrophylla Willd. Eastern Tibet: Radja and Yellow River gorges, nos. 14025, 14066. Cardamine macrophylla Willd. ssp. polyphylla (Don) O. E. Schulz in Notizbl. Bot. Gard. Mus. Berlin, rx. 1071 (1927). Southwestern Kansu: Upper Tebbu country, no. 12508. Cardamine tangutorum O. E. Schulz Southwestern Kansu: Tao River basin, no. 12318. *Draba lanceolata Royle var. latifolia O. E. Schulz in Notizbl. Bot. Gard. Mus. Berlin, x. 555 (1929). Eastern Tibet: Radja and Yellow River gorges, no. 14134 (holo- type). Draba lanceolata Royle var. leiocarpa O. E. Schulz in Notizbl Bot. Gard. Mus. Berlin, 1x. 1077 (1929). Southwestern Kansu: Upper Tebbu country, no. 12491. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14223; Radja and Yellow River gorges, no. 14130. Draba lichiangensis W. W. Sm. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 17 Eastern Tibet: Jupar Range, no. 14369. Draba oreades Schrenk proles chinensis O. E. Schulz in Engler, Pflanzenr. tv.-105 (Heft 89) 109 (1927). Southwestern Kansu: Tao River basin, no. 12388. Draba oreades Schrenk var. commutata (Reg.) O. E. Schulz in Engler, Pflanzenr. tv.-105 (Heft 89) 109 (1927). Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14228 Draba oreades Schrenk var. racemosa O. E. Schulz in Engler, Pflanzenr. 1v.-105 (Heft 89) 109 (1927). Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14091, Draba oreades Schrenk var. Tafelii O. E. Schulz in Engler, Pflanz- enr. Iv.-105 (Heft 89) 108 (1927). Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14142. *Draba Rockii O. E. Schulz in Notizbl. Bot. Gard. Mus. Berlin, x. 555 (1929). Southwestern Kansu: Tao River basin, no. 12405 (holotype). Arabis alaschanica Maxim. Southwestern Kansu: Upper Tebbu country, nos. 12490, 12515. Cheiranthus roseus Maxim. Eastern Tibet: Jupar Range, no. 14409, *Cheiranthus roseus Maxim. f. caespitosus O. E. Schulz, msc. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14096 *Cheiranthus roseus Maxim. f. elatior O. E. Schulz, msc. Eastern Tibet: Jupar Range, no. 14345. Malcolmia africana (L.) R. Br. Eastern Tibet: Radja and Yellow River gorges, no. 14133. Torularia humilis (C. A. Mey.) O. E. Schulz Southwestern Kansu: Tao River basin, no. 12243. Eastern Tibet: Radja and Yellow River gorges, no. 14131. Torularia humilis (C. A. Mey.) O. E. Schulz prol. Piasezkii (Maxim.) O. E. Schulz Southwestern Kansu: Tao River basin, no. 12269. Torularia humilis f. grandiflora O. E. Schulz Southwestern Kansu: Tao River basin, no. 12171. Dontostemon glandulosus (Kar. & Kir.) O. E. Schulz in Notizbl. Bot. Gard. Mus. Berlin, x. 554 (1929). Arabis glandulosa Kar. & Kir. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14399. 18 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Parrya Villosa Maxim. Eastern Tibet: Radja and Yellow River gorges, nos. 14051, 14088. *Parrya Villosa Maxim. var. albiflora O. E. Schulz in Notizbl. Bot. Gard. Mus. Berlin, x. 557 (1929). Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14392 (holotype). CRASSULACEAE Sedum Aizoon L. Southwestern Kansu: Tao River basin, no. 12914. Sedum Aizoon L. var. scabrum Maxim. Southwestern Kansu: Tao River basin, no. 12355. Sedum algidum Ledeb. Southwestern Kansu: Tao River basin, no. 14153. Sedum algidum Ledeb. var. tanguticum Maxim. Eastern Tibet: Jupar Range, nos. 14346, 14365. Sedum Henryi Diels forma gracilis ( ¢ ) Southwestern Kansu: Tao River basin, no. 12839. Sedum Kirilowi Regel Eastern Tibet: Radja and Yellow River gorges, no. 14174. *Sedum progressum Diels in Notizbl. Bot. Gard. Mus. Berlin, x. 887 (1930). Southwestern Kansu: Tao River basin, no. 12849 (holotype). Sedum Purdomii W. W. Sm. Central Kansu: no. 12705. Sedum quadrifidum Pall. Eastern Tibet: Radja and Yellow River gorges, no. 13981. Sedum venustum Praeger Southwestern Kansu: Upper Tebbu country, no. 12516. Eastern Tibet: Radja and Yellow River gorges, no. 14042. Sedum spec. Western Kansu: no. 13259. SAXIFRAGACEAE Saxifraga confertifolia Engl. & Irmsch. Southwestern Kansu: Upper Tebbu country, no. 13027. Saxifraga diversifolia Wall. var. Soulieana Engl. & Irmsch. Eastern Tibet: Kokonor region, no. 13356. Southwestern Kansu: Upper Tebbu country, no. 14585. Saxifraga diversifolia Wall. var. Southwestern Kansu: Tao River basin, no. 14642. Saxifraga flagrans H. Smith var. platyphylla H. Smith. Southwestern Kansu: Tao River basin, nos. 12196, 12817. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 19 Saxifraga Giraldiana Engl. Southwestern Kansu: Upper Tebbu country, no. 13051. *Saxifraga kansuensis Mattfeld in Notizbl. Bot. Gard. Mus. Berlin, x. 887 (1930). Southwestern Kansu: Tao River basin, no. 12624 (syntype); Upper Tebbu country, no. 12525 (syntype). Saxifraga lumpuensis Engl. Central Kansu: no. 12743. Southwestern Kansu: Tao River basin, no. 12832. Saxifraga melanocentra Franch. f. Franchetiana Engl. & Irmsch. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14158. Saxifraga melanocentra Franch. var. pluriflora Engl. & Irmsch. Eastern Tibet: Amnyi Machen range, no. 14441. Saxifraga melanocentra Franch. var. Southwestern Kansu: Upper Tebbu country, no. 13042. Saxifraga montana H. Smith var. splendens H. Smith Southwestern Kansu: Tao River basin, no. 14613. Saxifraga Przewalskii Engl. Southwestern Kansu: Upper Tebbu country, no. 13017. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14232; Jupar Range, no. 14327. Saxifraga pseudo-hirculus Engl. Southwestern Kansu: Upper Tebbu country, -no. 13089. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14527. Saxifraga tangutica Engl. Southwestern Kansu: Upper Tebbu country, no. 12510. Eastern Tibet: Radja and Yellow River gorges, no. 14169. Saxifraga unguiculata Engl. Eastern Tibet: grasslands between Labrang and Yellow River, no. Tiarella polyphylla D. Don Central Kansu: no. 12679. Chrysosplenium nudicaule Bge. Eastern Tibet: Jupar Range, no. 14334. Chrysosplenium sphaerospermum Maxim. Southwestern Kansu: Tao River basin, no. 12097. Parnassia Delavayi Franch. Central Kansu: no. 12790. Parnassia setchuenensis Franch. Southwestern Kansu: Tao River basin, no. 13157; Upper Tebbu country, no. 14589. JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV ROSACEAE Aruncus sylvester Kostel. Southwestern Kansu: Tao River basin, no. 12815. Fragaria elatior Ehrb. Eastern Tibet: Radja and Yellow River gorges, no. 14036. Potentilla Anserina L. Southwestern Kansu: Tao River basin, no. 12180. Eastern Tibet: Radja and Yellow River gorges, no. 14022; grasslands between Labrang and Yellow River, no. 14532. Potentilla biflora Willd. Southwestern Kansu: Upper Tebbu country, no. 13006. Potentilla bifurca L. Southwestern Kansu: Tao River basin, no. 12223. Eastern Tibet: Kokonor region, no. 13351. Potentilla multicaulis Bge. Eastern Tibet: Radja and Yellow River gorges, no. 14176. Potentilla Potaninii Th. Wolf. Southwestern Kansu: Tao River basin, no. 12859. Potentilla Salesoviana Steph. Northwestern Kansu: no. 13321. Eastern Tibet: Ba Valley, no. 14280. Potentilla Saundersiana Royle Eastern Tibet: alpine region between Radja and Jupar ranges, nos. 14150, 14378, 14391. Potentilla sericea L. Southwestern Kansu: Tao River basin, no. 12857. Potentilla spec. Southwestern Kansu: Tao River basin, no. 12911. Eastern Tibet: Radja and Yellow River gorges, no. 14175; grasslands between Labrang and Yellow River, no. 14486. Coluria longifolia Maxim. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14141. Sanguisorba canadensis L. Central Kansu: no. 12777. Southwestern Kansu: Tao River basin, no. 13204. Sanguisorba officinalis L. Central Kansu: no. 12778. LEGUMINOSAE Sophora alopecuroides L. Northwestern Kansu: no. 13317. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 21 Thermopsis alpina Ledeb. Southwestern Kansu: Tao River basin, no. 12155. Eastern Tibet: Radja and Yellow River gorges, no. 14055. Thermopsis lanceolata R. Br. Southwestern Kansu: Tao River basin, no. 12338. Eastern Tibet: Radia and Yellow River gorges, no. 13919. Medicago ruthenica Trautv. Southwestern Kansu: Tao River basin, no. 12924. Gueldenstaedtia diversifolia Maxim. Eastern Tibet: Radja and Yellow River gorges, no. 14123. Astragalus melilotoides Pall. Southwestern Kansu: Tao River basin, nos. 12896, 12901. Astragalus monadelphus Bee. Southwestern Kansu: Tao River basin, no. 12958. Astragalus skythropus Bge. astern Tibet: alpine region between Radja and Jupar ranges, nos. 14151, 14394; Jupar Range, no. 14315. Astragalus tanguticus Batal. Eastern Tibet: Radja and Yellow River gorges, no. 14182; Ba Valley, no. 14269. Astragalus spec. aff. A. subumbellatus Klotzsch Eastern Tibet: Radja and Yellow River gorges, no. 14004; Ba Valley, no. 14253. Astragalus spec. Southwestern Kansu: Tao River basin, nos. 12084, 12092, 12098, 12220, 12221, 12284, 12327, 12447, 12628, 12802, 12925, 12926, 12953, 13146; Upper Tebbu country, no. 13045. Central Kansu: nos. 12704, 12718. Western Kansu: no. 13254. Eastern Tibet: Kokonor region, nos. 13375, 13382; grasslands between Labrang and Yellow River, no. 14490; alpine region between Radja and Jupar ranges, nos. 14148, 14163, 14233, 14243, 14381; Ba Valley, nos. 14249, 14250, 14254, 14256, 14257, 14258, 14286, 14361; Jupar Range, nos. 14343, 14370. Oxytropis spec. Southwestern Kansu: Upper Tebbu country, no. 1306S. Hedysarum esculentum Ledeb. Southwestern Kansu: Tao River basin, no. 13142; Upper Tebbu country, no. 14607. Hedysarum obscurum L. Eastern Tibet: Radja and Yellow River gorges, no. 14171; alpine region between Radja and Jupar ranges, no. 14382. Hedysarum pseudastragalus Ulbrich JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Eastern Tibet: Amnyi Machen range, no. 14423. Hedysarum spec. Southwestern Kansu: Tao River basin, no. 12623; Upper Tebbu country, nos. 13010, 13063. Central Kansu: no. 12675. Eastern Tibet: Jupar Range, no. 14347. Vicia amoena Ledeb. var. elliptica Freyn Southwestern Kansu: Tao River basin, no. 12922. Vicia cracca L. Southwestern Kansu: Tao River basin, no. 12959. Vicia unijuga A. Br. Southwestern Kansu: Tao River basin, no. 12583. Vicia unijuga A. Br. var. Southwestern Kansu: Upper Tebbu country, no. 13064. Lathyrus pratensis L. Southwestern Kansu: Tao River basin, no. 12585. GERANIACEAE Geranium eriostemon Fisch. Southwestern Kansu: Tao River basin, nos. 12225, 12575, 12920. Geranium Pylzowianum Maxim. Southwestern Kansu: Tao River basin, nos. 12450, 12582, 12825, 12918. Eastern Tibet: Radja and Yellow River gorges, no. 14212. Geranium spec. aff. G. pratense L. Southwestern Kansu: Tao River basin, no. 12855. Eastern Tibet: Radja and Yellow River gorges, no. 14216; Ba Valley, no. 14360. Erodium Stephanianum Willd. Eastern Tibet: Radja and Yellow River gorges, no. 14205. Biebersteinia heterostemon Maxim. Southwestern Kansu: Tao River basin, no. 12904. LINACEAE Linum nutans Maxim. Southwestern Kansu: Tao River basin, no. 12170. ZYGOPHYLLACEAE Peganum Harmala L. Central Kansu: no. 13229. Eastern Tibet: Jupar Range, no. 14317. Zygophyllum mucronatum Maxim. Central Kansu: no. 13241. Tribulus terrestris L. Northwestern Kansu: no. 13397. POLYGALACEAEF Polygala sibirica L. Eastern Tibet: Radja and Yellow River gorges, no. 14204. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 23 EUPHORBIACEAE Euphorbia spec. Eastern Tibet: Radja and Yellow River gorges, no. 13941. GUTTIFERAE Hypericum Przewalskii Maxim. Southwestern Kansu: Tao River basin, no. 12875, 12909. Eastern Tibet: Radja and Yellow River gorges, no. 14213. Hypericum spec. Central Kansu: no. 12772. TAMARICACEAE Hololachne songarica Ehrenb. Central Kansu: no. 13235. VIOLACEAE Viola biflora L. Southwestern Kansu: Upper Tebbu country, no. 12493. Eastern Tibet: Radja and Yellow River gorges, nos. 14015, 14084. Viola bulbosa Maxim. Eastern Tibet: Radja and Yellow River gorges, no. 14029. Viola mongolica Franch. var. flor. carneis Eastern Tibet: Radja and Yellow River gorges, no. 13966. THYMELAEACEAE Stellera Chamaejasme L. Southwestern Kansu: Tao River basin, nos. 12238, 12286. Eastern Tibet: Radja and Yellow River gorges, no. 14020. Stellera Chamaejasme L. fl. albo Eastern Tibet: Radja and Yellow River gorges, no. 14199. OENOTHERACEAE Epilobium angustifolium L. Central Kansu: no. 12755. Southwestern Kansu: Tao River basin, no. 12954. UMBELLIFERAE Determined by C. NorMANn Bupleurum longeradiatum Turcz. Central Kansu: no. 12769. Bupleurum spec.* Eastern Tibet: Radja and Yellow River gorges, no. 14211. Perhaps this specimen represents B. microcephalum Diels. However, the speci- n at hand is without true leaves, hence it cannot be determined with certainty. 24 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Pleurospermum Candollii C. B. Clarke’ Central Kansu: no. 12703. Pleurospermum cnidiifolium Wolff in Act. Hort. Gothob. 1. 292 (1926) Eastern Tibet: Kokonor region, no. 13359. *Pleurospermum Dielsianum Fedde & Wolff in Fedde, Rep. Spec. Nov. xxvu. 121 (1929). Eastern Tibet: grasslands between Labrang and Yellow River, no. 14466 (syntype). Pleurospermum Franchetianum Hemsley in Hooker, Icon. xxiII. 2244 (1894). *Pleurospermum Rockit Fedde . oo in Fedde, Rep. Spec. Nov. xxvur. 120 (1929).—Synon. n *Pleurospermum Pilgerianum Fedde & Wolff in Fedde, Rep. Spec. Nov. 121 —Synon. no Seiten Kansu: Tao River ck no. 12227 (holotype of P. Rockii Fedde & Wolff), no. 12614 (holotype of P. Pilgerianum Fedde & Wolff). Pleurospermum linearilobum W. W. Smith in Notes Bot. Gard. Edinb. virt. 342 (1915). *Pleurospermum Dielsianum Fedde & Wolff in Fedde, Rep. Spec. Nov. xxvil. 121 (1929), as to no. 14178. Eastern Tibet: Radja and Yellow River gorges, no. 14178 (syntype of P. Dielsianum Fedde & Wolff). *Pleurospermum pseudo-involucratum Wolff in Fedde, Rep. Spec. Nov. xxvit. 119 (1929).° Southwestern Kansu: Upper Tebbu country, no. 12496 (holotype). Pleurospermum thalictrifolium Wolff in Act. Hort. Gothob. 1 297 (1926). Eastern Tibet: Jupar Range, no. 14372; Amnyi Machen range, no. 14432 Pleurospermum spec. aff. P. Candollii C. B. Clarke Eastern Tibet: Jupar Range, no. 14307 (incomplete material). Carum carvi L. *This slg onan is given with some hesitation as the specimen for study is very young. However, it agrees in all essentials with the other material canines, The involucels are ee than those of the Indian specimens, which may be due to the immature pene of the specimens. It is a widely dispersed species hav ing been recorded from Kashmir, Kumaon, Tian ‘han and Nepal. It is doubital if it has been ae before as far east as Kan *The specimen no. 14178 is cited in the oiginal description of P. Dielsianum Fedde & Wolff as identical ee no Haprieie . 66 is tt fragmentary an is referred here to P. Dielsian ee seem to be the same plant as nc. 14178, which is undoubtedly Pp ein pi W. W. Sm °"Wolff was in doubt whether this specimen should i described under Pleuro- spermum or Ligusticum, but the doubt lies rather between kale aaiaaid and Trachydium. The final decision will depend on more ample mater 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 25 Southwestern Kansu: Tao River basin, no. 12449. Eastern Tibet: Radja and Yellow River gorges, no. 14117; Ba Valley, no. 14255. Tongoloa elata Wolff in Act. Hort. Gothob. 1m. 291 (1925). Southwestern Kansu: Upper Tebbu country, no. 14629. Ligusticum Pilgerianum Wolff in Fedde, Rep. Spec. Nov. xxvu. 307 (1930) Central Kansu: no. 12724. Ligusticum sinense Oliver in Hooker, Icon. xx. 1958 (1891). *Ligusticum Pilgerianum Fedde in Fedde, Rep. Spec. Nov. xxvii. 322 (1930), non Wolff (p. 307) Southwestern Kansu: Upper Tebbu country, no. 14590 (syntype of L. Pilgerianum Fedde). *Ligusticum Weberbauerianum Fedde & Wolff in Fedde, Rep. Spec. Nov. xxvir. 312 (1930). Central Kansu: no. 12727 (holotype). Heracleum millefolium Diels Southwestern Kansu: Tao River basin, no. 12222. Eastern Tibet: Radja and Yellow River gorges, no. 14207; grasslands between Labrang and Yellow River, no. 14491 *Heracleum millefolium Diels var. longilobum Norman, var. nov. A typo differt foliorum segmentis ultimis longioribus acutioribusque, et ovarii indumento densiore. Central Kansu. Lien hoa shan: among rocks between Tao Chow and Titao, alt. 3450 m., no. 12734 (type), July 1925 (plant white, woolly; flowers, carmine); among rocks between Tao-chow and Titao, alt. 3050 m., no. 12674, July 1925 (plant rosette-like; flowers pinkish lavender). PYROLACEAE Pyrola rotundifolia L. Central Kansu: no. 12745. PRIMULACEAE Primula L. Determined by W. W. SmitH Primula aerinantha Balf. f. & Purdom Central Kansu: nos. 12676, 12711. Primula alsophila Balf. f. & Farrer Southwestern Kansu: Tao River basin, no. 12819. Primula chionantha Balf. f. & Forrest Southwestern Kansu: Tao River basin, no. 12392; Upper Tebbu country, no. 13631. "Dr. Diels, to whom part of this specimen was Submitted, intimated that he considered it ‘at least a variety or an allied species. JOURNAL OF. .THE ARNOLD ARBORETUM [VOL. XIV Primula chionantha Balf. f. & Forrest (farinose form) Southwestern Kansu: Tao River basin, nos. 12345, 12375. Primula chionantha Balf. f. & Forrest var. Southwestern Kansu: Tao River basin, no. 12470. Primula chionantha Balf. f. & Forrest forma Southwestern Kansu: Tao River basin, nos. 13656, 14920. Primula conspersa Balf. f. & Purdom Southwestern Kansu: Tao River basin, nos. 12451 (in part), 12563, 12381, 12874. Central Kansu: no. 12683. Primula flava Maxim. Southwestern Kansu: Tao River basin, no. 12363; Upper Tebbu country, nos. 12492, 13639. Eastern Tibet: Radja and Yellow River gorges, nos. 13965, 14043, 14120; grasslands between La- brang and Yellow River, no. 13914. Primula gemmifera Batal. Southwestern Kansu: Tao River basin, nos. 12451 (in part), 12818, 13159, 14624; Upper Tebbu country, nos. 13004, 13029. East- ern Tibet: Radja and Yellow River gorges, no. 13993. Primula graminifolia Pax & Hoffm. Southwestern Kansu: Tao River basin, nos. 12384, 12404; Upper Tebbu country, no. 13014. Primula limbata Balf. f. & Forrest Southwestern Kansu: Tao River basin, nos. 12381, 12408, 13654; Upper Tebbu country, no. 13066. Eastern Tibet: Radja and Yellow River gorges, no. 14064. Primula longipetiolata Pax & Hoffm. vel aff. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14227. Primula moupinensis Franch. Southern Kansu: nos. 12043, 12069. Primula optata Farrer apud I. B. Balf. Southwestern Kansu: Upper Tebbu country, no. 13633. Eastern Tibet: alpine region between Radja and Jupar ranges, nos. 14097, 14145, 14147, 14385. Primula optata Farrer apud I. B. Balf. forma Eastern Tibet: Amnyi Machen range, no. 14424. Primula polyneura Franch. Southwestern Kansu: Tao River basin, nos. 12251, 12316, 12438, 13652 Primula pumilio Maxim. Southwestern Kansu: Tao River basin, no. 12406. Primula Purdomii Craib 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 27 Southwestern Kansu: Tao River basin, nos. 12346, 12383, 12403. Eastern Tibet: Radja and Yellow River gorges, no. 14089. Primula reginella Balf. f. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14238. Primula reginella Balf. f. (dwarf form) Eastern Tibet: grasslands between Labrang and Yellow River, no. 13907. Primula sibirica Jacq. Southwestern Kansu: Tao River basin, nos. 12277, 12313. Eastern Tibet: Radja and Yellow River gorges, no. 13994; Ba Valley, no. 14276. Primula sikkimensis Hook. f. Eastern Tibet: Radja and Yellow River gorges, no. 14166; Amnyi Machen range, no. 14437. Primula stenocalyx Maxim. Southwestern Kansu: Tao River basin, nos. 12091, 12.109, 12307, 12421, 12488; Upper Tebbu country, no. 12523. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14470. Primula stenocalyx Maxim. (farinose form) Southwestern Kansu: Tao River basin, nos. 12361, 12402. Central Kansu: no. 12733. Primula tangutica Duthie Southwestern Kansu: Tao River basin, nos. 12253, 12261, 12612, 13619, 13655; Upper Tebbu country, no. 12512. Eastern Tibet: Radja and Yellow River gorges, nos. 14008, 14065; Jupar Range, no. 14310. Primula Woodwardii Balf. f. Southwestern Kansu: Tao River basin, nos. 12653, 12656, 12837, 13658. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14234. Primula spec. Southwestern Kansu: Tao River basin, no. 12407. Androsace L. Determined by H. HANpEL-MAzzetTTI Androsace Mariae Kan. var. tibetica (Maxim.) Handel-Mazzetti in Act. Hort. Gothob. rr. 114 (1926). Southwestern Kansu: Tao River basin, nos. 12094, 12133, 12308, 12619. Eastern Tibet: Radja and Yellow River gorges, nos. 13902, 13940, 14083; Ba Valley, no. 14261. Androsace tapete Maxim. JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Southwestern Kansu: Tao River basin, no. 12650. Central Kansu: no. 12741 Androsace yargongensis Petitmengin Southwestern Kansu: Upper Tebbu country, no. 13040. Cortusa Matthioli L Central Kansu: no. 12720. Glaux maritima L. Southwestern Kansu: Tao River basin, no. 12245. Eastern Tibet: Radja and Yellow River gorges, no. 14135. PLUMBAGINACEAE Statice bicolor Bge. Central Kansu: no. 12758. GENTIANACEAE Gentiana L. Determined by C. V. B. MARQUAND Gentiana algida Pall. forma Eastern Tibet: grasslands between Labrang and Yellow River, no. 14510. Gentiana dahurica Fisch. Eastern Tibet: Kokonor region, no. 13368; grasslands between La- brang and Yellow River, no. 14531. Gentiana Farreri Balf. f. Southwestern Kansu: Tao River basin, no. 14886; Lower Tebbu country, no. 14862. Gentiana Futtereri Diels & Gilg Eastern Tibet: Kokonor region, nos. 13360, 13399. Gentiana gracilipes Turrill Southwestern Kansu: Tao River basin, nos. 13175, 13719. Gentiana hexaphylla Maxim. Southwestern Kansu: Tao River basin, no. 13713. Gentiana hexaphylla Maxim. var. caudata Marquand in Kew Bull. Misc. Inform. 1931, p. 81 Southwestern Kansu: Tao River basin, no. 13736. Gentiana leucomelaena Maxim. Southwestern Kansu: Tao River basin, no. 12309. Eastern Tibet: Ba Valley, no. 14277. Gentiana Przewalskii Maxim. Southwestern Kansu: Tao River basin, nos. 13720, 13734, 14645; Upper Tebbu country, no. 13011. Gentiana quinquenervia Turrill Southwestern Kansu: Tao River basin, no. 13179. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 29 Gentiana riparia Kar. & Kir. Eastern Tibet: Radja and Yellow River gorges, no. 14021. Gentiana siphonantha Maxim. Eastern Tibet: Kokonor region, no. 13364; grasslands between La- brang and Yellow River, no. 14499 (partly browsed). Gentiana spathulifolia Kusnez. Southwestern Kansu: Upper Tebbu country, no. 14606. Gentiana straminea Maxim. Eastern Tibet: Kokonor region, no. 13380. Gentiana straminea Maxim. forma (leaves wider than the type). Southwestern Kansu: Tao River basin, no. 13715. Gentiana striata Maxim. Southwestern Kansu: Lower Tebbu country, no. 14863. Gentiana Szechenyii Kanitz (forma plicis longioribus). Southwestern Kansu: Lower Tebbu country, no. 14861. Gentiana tetraphylla Kusnez. Southwestern Kansu: Tao River basin, no. 14638. Gentianella ( § Crossopetalum) spec. Eastern Tibet: Radja and Yellow River gorges, no. 14218; Ba Val- ley, no. 14247. Southwestern Kansu: Tao River basin, no. 12353; Upper Tebbu country, nos. 13031, 14597; Tao River basin, no. 13709. Central Kansu: no. 12691. Gentianella spec. Eastern Tibet: Kokonor region, nos. 13347, 13388; grasslands be- tween Labrang and Yellow River, no. 14497. Swertia spec. Eastern Tibet: Kokonor region, no. 13350, 13358. Southwestern Kansu: Tao River basin, nos. 13707, 13708, 14619. ASCLEPIADACEAE Cynanchum inamoenum (Maxim.) Loes. Southwestern Kansu: Tao River basin, no. 12226. Central Kansu: no. 13230. CONVOLVULACEAE Convolvulus Ammannii Desr. Central Kansu: no. 13233. Eastern Tibet: Ba Valley, no. 14358. Convolvulus arvensis L. var. sagittifolius Fisch. Eastern Tibet: Radja and Yellow River gorges, no. 14200; Ba Val- ley, no. 14359. POLEMONIACEAE Polemonium coeruleum L. ssp. vulgare (Ledeb.) Brand Southwestern Kansu: Tao River basin, nos. 12574, 12848, 12919. 30 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV BORAGINACEAE Determined by |. M. JOHNSTON Lappula Redowskii (Horn.) Greene Southwestern Kansu: Tao River basin, no. 12358. Eritrichium strictum Dcne. Eastern Tibet: Radja and Yellow River gorges, no. 14203. Lithospermum officinale L. Southwestern Kansu: Tao River basin, no. 12356. *Microula Rockii Johnston in Contrib. Gray Herb. Lxxxr. 82 (1928). Anoplocaryum Rockit (Johnst.) Brand in Fedde, Rep. Spec. Nov. (1929). Southwestern Kansu: Tao River basin, no. 12605 (paratype). East- ern Tibet: grasslands between Labrang and Yellow River, no. 14511 (holotype); alpine region between Radja and Jupar ranges, no. 14384 (paratype). Microula sikkimensis (Clarke) Hemsley Southwestern Kansu: Tao River basin, nos. 12250, 12347. Eastern Tibet: Radja and Yellow River gorges, no. 14026; Ba Valley, no. 14275. Microula tangutica Maxim. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14397; grasslands between Labrang and Yellow River, nos. 14469, 14509. Myosotis alpestris F. W. Schmidt var. Southwestern Kansu: Tao River basin, nos. 12386, 13005. LABIATAE Ajuga calantha Diels Southwestern Kansu: Tao River basin, nos. 12214, 12455, 12610, 12636, 12806. “Ajuga calantha Diels f. albiflora Diels, msc. Central Kansu: no. 12787. Ajuga lupulina Maxim. Southwestern Kansu: Tao River basin, no. 12321. Eastern Tibet: Radja and Yellow River gorges, no. 14129; Jupar Range, no. 14373. Scutellaria amoena C. H. Wright Southwestern Kansu: Tao River basin, nos. 12564, 13198. *“Scutellaria scordiifolia Fisch. f. pubescens Diels, msc. Southwestern Kansu: Tao River basin, no. 12900. *Scutellaria Rehderiana Diels in Notizbl. Bot. Gard. Mus. Berlin, x. 889 (1930). 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 31 Southwestern Kansu: Tao River basin, no. 12827 (holotype). Marrubium incisum Benth. Eastern Tibet: Radja and Yellow River gorges, no. 14194. Nepeta coerulescens Maxim. f. major Diels, msc. Eastern Tibet: Radja and Yellow River gorges, no. 14119. Nepeta macrantha Fisch. Southwestern Kansu: Tao River basin, nos. 12891, 12905, 12963; Upper Tebbu country, no. 14628. Dracocephalum heterophyllum Benth. Southwestern Kansu: Tao River basin, no. 12894. Central Kansu: nos. 12768, 13247. Western Kansu: no. 13251. Eastern Tibet: Kokonor region, no. 13353; Radja and Yellow River gorges, no. 14198. Dracocephalum imberbe Bge. Southwestern Kansu: Tao River basin, nos. 12593, 12871. Central Kansu: no. 12767. Dracocephalum tanguticum Maxim. Southwestern Kansu: Tao River basin, nos. 12860, 12866. Western Kansu: nos. 13249, 13260. Phlomis rotata Benth. Eastern Tibet: Radja and Yellow River gorges, no. 14172; alpine region between Radja and Jupar ranges, no. 14406. Phlomis umbrosa Turcz. Central Kansu: no. 12791. Southwestern Kansu: Upper Tebbu country, no. 14576. Galeopsis Tetrahit L. Southwestern Kansu: Tao River basin, no. 13153. Stachys baicalensis Fisch. Central Kansu: no. 12665. Eastern Tibet: Radja and Yellow River gorges, no. 14181. Salvia Prattri Hemsl. Eastern Tibet: Amnyi Machen range, no. 14426. Salvia Przewalskii Maxim. Southwestern Kansu: Tao River basin, no. 12579. Salvia Roborowskii Maxim. Southwestern Kansu: Tao River basin, no. 12184. *Thymus Serpyllum L. ssp. mongolicus Ronniger in Notizbl. Bot. Gard. Mus. Berlin, x. 890 (1930). Southwestern Kansu: Tao River basin, no. 12927 (syntype). Cen- tral Kansu: no. 12664. Eastern Tibet: grasslands between La- brang and Yellow River, no. 14539 (syntype). Elsholtzia densa Benth. Southwestern Kansu: Tao River basin, no. 13150. JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV SOLANACEAE Anisodus spec. Southwestern Kansu: Tao River basin, no. 12337. Solanum septemlobum Bge. Eastern Tibet: Ba Valley, no. 14273. SCROPHULARIACEAE Scrophularia incisa Weinm. Eastern Tibet: Radja and Yellow River gorges, no. 14137; Ba Val- ley, no. 14245. Scrofella ieee Maxim. Southwestern Kansu: Tao River basin, nos. 13152, 13733. Central Kansu: no. 12690. Mimulus nepalensis Benth. Southwestern Kansu: Tao River basin, no. 12947. Veronica ciliata Fisch. Southwestern Kansu: Upper Tebbu country, no. 14600. Lagotis brachystachya Maxim. Southwestern Kansu: Tao River basin, no. 12130. Eastern Tibet: Radja and Yellow River gorges, no. 13951. Lagotis brevituba Maxim. Eastern Tibet: Jupar Range, no. 14332. Lagotis glauca Gaertn. Southwestern Kansu: Tao River basin, no. 12391. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14098. Euphrasia hirtella Jord. Southwestern Kansu: Tao River basin, no. 12246. Euphrasia tatarica Fisch. Central Kansu: no. 12788. Pedicularis alaschanica Maxim. Southwestern Kansu: Tao River basin, nos. 12459, 12568. Eastern Tibet: Kokonor region, no. 13374 Pedicularis anas Maxim. var. Southwestern Kansu: Upper Tebbu country, no. 13035. Pedicularis armata Maxim. Southwestern Kansu: Tao River basin, nos. 12880, 13148; Upper Tebbu country, no. 13091. Pedicularis cheilanthifolia Schrenk Southwestern Kansu: Tao River basin, no. 12312. Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14465, 14526; Ba Valley, no. 14272. Pedicularis cheilanthifolia Schrenk var. isochila Maxim. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 33 Southwestern Kansu: Tao River basin, nos. 12247, 12278. Eastern Tibet: Radja and Yellow River gorges, no. 14202. *Pedicularis chenocephala Diels in Notizbl. Bot. Gard. Mus. Ber- lin, x. 892 (1930). Southwestern Kansu: Tao River basin, no. 12607 (holotype). Pedicularis chenocephala Diels forma Eastern Tibet: Amnyi Machen range, no. 14435. Pedicularis chinensis Maxim. Southwestern Kansu: Upper Tebbu country, no. 13101. Eastern Tibet: Kokonor region, no. 13355 Pedicularis cranolopha Maxim. Eastern Tibet: Radja and Yellow River gorges, no. 14186. Pedicularis cranolopha Maxim. var. longicornuta Prain Southwestern Kansu: Tao River basin, no. 12928. Pedicularis cristata Maxim. Southwestern Kansu: Tao River basin, nos. 12828, 12879, 12910. Pedicularis Davidi Franch. Southwestern Kansu: Tao River basin, nos. 12448, 12906; Upper Tebbu country, no. 13032. *Pedicularis decorissima Diels in Notizbl. Bot. Gard. Mus. Berlin, x. 891 (1930), Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14546 (holotype), 14537 (paratype). Pedicularis ingens Maxim. Eastern Tibet: Radja and Yellow River gorges, no. 14217; grass- lands between Labrang and Yellow River, no. 14481 Pedicularis kansuensis Maxim. Southwestern Kansu: Tao River basin, nos. 12287, 12794, Eastern Tibet: Radja and Yellow River gorges, nos. 14118, 14208; Ba Valley, no. 14270. Pedicularis labellata Jacq. Southwestern Kansu: Upper Tebbu country, no. 13098; Tao River basin, no. 13164. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14403. Pedicularis lasiophrys Maxim. Eastern Tibet: Amnyi Machen range, no. 14439; grasslands between Labrang and Yellow River, no. 14525. Pedicularis lasiophrys Maxim. var. sinica Maxim. Southwestern Kansu: Tao River basin, no. 12850. Pedicularis longiflora Rudolphi Eastern Tibet: Ba Valey, no. 14354. Pedicularis macrosiphon Franch. Southwestern Kansu: Tao River basin, no. 12293. 34 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Pedicularis muscicola Maxim. Eastern Tibet: Ba Valley, no. 14279. Pedicularis pilostachya Maxim. Eastern Tibet: Jupar Range, no. 14331. Pedicularis Przewalskii Maxim. Eastern Tibet: Amnyi Machen range, no. 14438. Pedicularis recurva Maxim. Southwestern Kansu: Upper Tebbu country, no. 13097. Pedicularis rudis Maxim. Southwestern Kansu: Tao River basin, no. 12868; Upper Tebbu country, no. 13104. Pedicularis scolopax Maxim. flor. majoribus Eastern Tibet: Radja and Yellow River gorges, no. 14132. Pedicularis semitorta Maxim Southwestern Kansu: Tao oe basin, nos. 12268, 12795. Central Kansu: no. 12696. Eastern Tibet: Radja and Yellow River gorges, nos. 14127, 14215; grasslands between Labrang and Yel- low River, no. 14548. *Pedicularis striata Pall. var. poliocalyx Diels in Notizbl. Bot. Gard. Mus. Berlin, x. 892 (1930) Southwestern Kansu: Tao River basin, nos. 12798 (holotype), 42916 (parat Pedicularis szetschuanica Maxim. Southwestern Kansu: Tao River basin, nos. 12244, 12620. Eastern Tibet: Radja and Yellow River gorges, no. 14191; alpine region between Radja and Jupar ranges, no. 14161; Ba Valley, nos. 14252, 14274; Amnyi Machen range, no. 14416; grasslands be- tween Labrang and Yellow River, no. 14503. Pedicularis szetschuanica Maxim. var. longispica Bon. Southwestern Kansu: Tao River basin, no. 12643. Eastern Tibet: Radja and Yellow River gorges, no. 14192; Amnyi Machen range, no. 14415. Pedicularis torta Maxim. Southwestern Kansu: Tao River basin, nos. 12615, 12923. Pedicularis tristis L. var. macrantha Maxim. Southwestern Kansu: Tao River basin, no. 12637. Central Kansu: no. 12702. Eastern Tibet: Amnyi Machen range, no. 14422. Pedicularis versicolor Wahlenb. Southwestern Kansu: Tao River basin, nos. 12413, 12601, 12642; Upper Tebbu country, no. 13124. Eastern Tibet: Radja and Yellow River gorges, no. 13990; alpine region between Radja and Jupar ranges, nos. 14094, 14235, 14395. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 35 Pedicularis spec. aff. P. plicata Maxim. Central Kansu: no. 12784. Pedicularis spec. Southwestern Kansu: Tao River basin, no. 12255. Pedicularis (Bidentatae) spec. Southwestern Kansu: Upper Tebbu country, no. 12529. Pedicularis (Longirostres) spec. Central Kansu: no. 12689. BIGNONIACEAE Incarvillea compacta Maxim. Southwestern Kansu: Tao River basin, no. 13653. Northwestern Kansu: no. 13301. Eastern Tibet: Radja and Yellow River gorges, nos. 13974, 14087. Incarvillea principis Bur. & Franch. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 401 Incarvillea sinensis Lam. Central Kansu: no. 13245. OROBANCHACEAE Xylanche himalaica Beck Southwestern Kansu: Upper Tebbu country, no. 13123. LENTIBULARIACEAE Pinguicula alpina L. Southwestern Kansu: Upper Tebbu country, no. 12535; Tao River basin, no. 12644. RUBIACEAE Asperula odorata L. Southwestern Kansu: Tao River basin, no. 12570. Galium boreale L. Southwestern Kansu: Tao River basin, no. 12559. Galium verum L. Southwestern Kansu: Tao River basin, no. 12591. Galium spec. Southwestern Kansu: Tao River basin, no. 12580. Eastern Tibet: Radja and Yellow River gorges, no. 14195. CAPRIFOLIACEAE Triosteum pinnatifidum Maxim. Southwestern Kansu: Tao River basin, no. 12216; Upper Tebbu country, no. 12509. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV VALERIANACEAE Valeriana pseudodioica Pax & K. Hoffm. Southwestern Kansu: Tao River basin, no. 12218. Valeriana tangutica Batal. Central Kansu: nos. 12712, 12716. Eastern Tibet: Kokonor region, no. 13348; Radja and Yellow River gorges, nos. 13980, 14122. Valeriana spec. Southwestern Kansu: Upper Tebbu country, no. 14595. Nardostachys Jatamansi DC. Eastern Tibet: Radja and Yellow River gorges, no. 14168; grass- ‘lands between Labrang and Yellow River, no. 14484. DIPSACACEAE Morina betonicoides Benth. Southwestern Kansu: Tao River basin, nos. 12617, 14611; Upper Tebbu country, no. 13119. Eastern Tibet: Radja and Yellow River gorges, no. 14170. Morina chinensis (Batal.) Diels apud Limpricht in Fedde, Rep. Spec. Nov. Beih. x1. 497 (Bot. Reise Hochgeb. Chinas & Ost- Tibets) (1922). Southwestern Kansu: Tao River basin, nos. 12556, 12952. Eastern Tibet: Radja and Yellow River gorges, no. 14187. CAMPANULACEAE Adenophora liliifolioides Pax & K. Hoffm. Southwestern Kansu: Tao River basin, no. 13151. Adenophora Potanini Korsh. Southwestern Kansu: Tao River basin, no. 13140. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14544. Adenophora Smithii Nannfeldt in Act. Hort. Gothob. v. 21 (1929). Southwestern Kansu: Tao River basin, no. 12929. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14551. Adenophora spec. aff. A. gracilis Nannfeldt in Act. Hort. Gothob. v. 17 (1929). Southwestern Kansu: Tao River basin, no. 13149. Adenophora spec. aff. A. marsupiiflora Fisch. Central Kansu: no. 12688. Adenophora spec. aff. A. Potanini Korsh. Western Kansu: no. 13250. Codonopsis ovata Benth. Eastern Tibet: grasslands between Labrang and Yellow River, uo. 14522. Codonopsis viridiflora Maxim. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 37 Central Kansu: no. 12738. Southwestern Kansu: Upper Tebbu country, no. 13125. Cyananthus Hookeri C. B. Clarke var. hispida Franch. Southwestern Kansu: Tao River basin, no. 13725. COMPOSITAE Aster Bowerii Hemsl. Eastern Tibet: Kokonor region, no. 13352; grasslands between La- brang and Yellow River, no. 14487. Aster flaccidus Bge. Southwestern Kansu: Upper Tebbu country, no. 13008, 13019. Eastern Tibet: alpine region between Radja and Jupar' ranges, no. 14230. Aster Fordii Hemsl. Southwestern Kansu: Upper Tebbu country, no. 14627 Aster Heterochaeta Benth. Southwestern Kansu: Tao River basin, nos. 12202, 12252, 12489, 12830. Eastern Tibet: Radja and Yellow River gorges, nos. 13985, 14109. Aster hispidus Thunb. Southwestern Kansu: Tao River basin, no. 13706. Western Kansu: no. 13255 Aster tongolensis Franch. Southwestern Kansu: Tao River basin, no. 12842. Eastern Tibet: Ba Valley, no. 14285. Aster trinervius Roxb. Southwestern Kansu: Tao River basin, no. 13668; Upper Tebbu country, no. 14631. Aster Vilmorini Franch. Southwestern Kansu: Tao River basin, nos. 12801, 12950, 14612; Upper Tebbu country, no. 13099. Central Kansu: no. 12684. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14533. Aster spec. Southwestern Kansu: Tao River basin, no. 12569. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14447. Erigeron acris L. Southwestern Kansu: Upper Tebbu country, no. 14601. Central Kansu: no. 12779. Leontopodium R. Br. Determined by H. HANpEL-MazzetTTI Leontopodium calocephalum (Franch.) Beauvd. Southwestern Kansu: Upper Tebbu country, no. 13033. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Leontopodium Dedekensii (Bur. & Franch.) Beauvd. Southwestern Kansu: Tao River basin, no. 13147. Eastern Tibet: Radja and Yellow River gorges, no. 14206. Leontopodium haplophylloides Hand.-Mazz. Western Kansu: nos. 13390, 13394. Eastern Tibet: grasslands be- tween Labrang and Yellow River, no. 14550. Leontopodium Jacotianum Beauvd. Central Kansu: no. 12714. Leontopodium linearifolium Hand.-Mazz. Southwestern Kansu: Upper Tebbu country, no. 13016; Lower Tebbu country, no. 14726. Eastern Tibet: Kokonor region, nos. 13373, 13392; grasslands between Labrang and Yellow River, nos. 14445, 14448, 14461, 14475, 14489, 14492, 14534; alpine region between Radja and Jupar ranges, nos. 14152, 14239, 14242; Jupar Range, no. 14333; Amnyi Machen range, no. 14417. Leontopodium linearifolium Hand.-Mazz. ad L. Souliei Beauvd. Eastern Tibet: Amnyi Machen range, no. 14418. Leontopodium nanum (Hook. f. & Thoms.) Handel-Mazzetti in Beih. Bot. Centralbl. xtiv. Abt. 2, 111 (1927). Southwestern Kansu: Tao River basin, no. 12134. Leontopodium Smithianum Hand.-Mazz. (L. conglobatum X L. leontopodioides ) Southwestern Kansu: Tao River basin, nos. 12239, 12565, 12902. Anaphalis Hancockii Maxim. Southwestern Kansu: Tao River basin, no. 12240. Anaphalis lactea Maxim. Southwestern Kansu: Tao River basin, nos. 12882, 12899. Anaphalis margaritacea (L.) Benth. & Hook. Central Kansu: no. 12780. Southwestern Kansu: Tao River basin, no. 13158. Anaphalis spec. Southwestern Kansu: Tao River basin, no. 13724. Inula ammophila Bge. Central Kansu: no. 13234. Carpesium Lipskyi C. Winkl. Southwestern Kansu: Upper Tebbu country, no. 14605. Achillea Ptarmica L. Southwestern Kansu: Tao River basin, no. 13667. Central Kansu: no. 12774. Chrysanthemum fruticulosum Ledeb. Western Kansu: no. 13258 Chrysanthemum indicum L. Southwestern Kansu: Tao River basin, no. 13705. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 39 Chrysanthemum tatsienense Bur. & Franch. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14512. Tanacetum tenuifolium Jacq. Southwestern Kansu: Tao River basin, no. 13727; Upper Tebbu country, no. 14598. Eastern Tibet: Kokonor region, no. 13362. Artemisia biennis Willd. Eastern Tibet: Kokonor region, no. 13384. Artemisia laciniata Willd. Western Kansu: no. 13248. Artemisia salsoloides Willd. Eastern Tibet: Kokonor region, no. 13385. Artemisia Sieversiana Will Southwestern Kansu: Tao River basin, no. 12913. Artemisia spec. Southwestern Kansu: Upper Tebbu country, no. 13013. Stereosanthus Souliei Franch. Southwestern Kansu: Tao River basin, no. 12854. Serratula centauroides L. Southwestern Kansu: Tao River basin, no. 12864; Upper Tebbu country, no. 14610. Central Kansu: no. 12773 Petasites tricholobus Franch. Southwestern Kansu: Tao River basin, no. 12082. Doronicum stenoglossum Maxim. Southwestern Kansu: Tao River basin, no. 12941; Upper Tebbu country, no. 14599. Doronicum thibetanum Cavill. Southwestern Kansu: Tao River basin, nos. 12192, 12389; Upper Tebbu country, no. 13020. Cacalia deltophylla (Maxim.) Mattfeld, comb. nov. Senecio deltophylla pine Southwestern Kansu: Upper Tebbu country, no. 14586. Cacalia Potanini (C. Winkl.) Mattfeld, comb. nov. Senecio Potanini C. Winkl. Southwestern Kansu: Upper Tebbu country, nos. 12998, 12998a. Cacalia spec Southwestern Kansu: Tao River basin, no. 13717; Upper Tebbu country, nos. 13028, 13076, 14603. Central Kansu: no. 12722. Senecio acerifolius C. Wink. Southwestern Kansu: Tao River basin, no. 12946; Upper Tebbu country, no. 14602. Central Kansu: no. 12710. Senecio argunensis Turcz. Southwestern Kansu: Tao River basin, nos.’ 13195, 13669. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Senecio campestris DC. Southwestern Kansu: Tao River basin, nos. 12320, 12352. Senecio nemorensis L. Southwestern Kansu: Lower Tebbu country, no. 14856. Senecio thianshanicus Reg. & Schmalh. Eastern Tibet: Kokonor region, no. 13400; Radja and Yellow River gorges, no. 14214; grasslands between Labrang and Yellow River, nos. 14468, 14498. Senecio spec. Southwestern Kansu: Tao River basin, nos. 12836, 13139; Upper Tebbu country, nos. 12997, 14573, 14584. Ligularia altaica DC. Southwestern Kansu: Tao River basin, nos. 12557, 12633. Ligularia plantaginifolia (Franch.) Mattfeld, comb. nov. Senecio plantaginifolia Franch. Eastern Tibet: Radja and Yellow River gorges, no. 14179; alpine region between Radja and Jupar ranges, no. 14402. Ligularia Przewalskii Maxim. Southwestern Kansu: Tao River basin, nos. 12846, 12907. Ligularia sagitta (Maxim.) Mattfeld, comb. nov. Senecio sagitta Maxim. Southwestern Kansu: Tao River basin, nos. 13165, 14616; Upper Tebbu country, nos. 13078, 13117. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14536. Ligularia sibirica Cass. var. speciosa DC. Southwestern Kansu: Tao River basin, no. 12943; Upper Tebbu country, nos. 13126, 14592. Ligularia tangutica (Maxim.) Mattfeld, comb. nov. Senecio tangutica Maxim. Southwestern Kansu: Upper Tebbu country, no. 13074 Ligularia virgaurea (Maxim.) Mattfeld, comb. nov. Senecio virgaurea Maxim. Southwestern Kansu: Upper Tebbu country, nos. 12999, 13038; Tao River basin, nos. 13163, 13181, 14614. Eastern Tibet: grasslands between Labrang and Yellow River, nos. 14521, 14524, 14549. Ligularia yesoensis Franch. var. sutchuensis Franch. Southwestern Kansu: Lower Tebbu country, nos. 14651, 15038. Ligularia spec. Southwestern Kansu: Upper Tebbu country, no. 13110. Cremanthodium bupleurifolium W. W. Sm. Southwestern Kansu: Tao River basin, no. 12630. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14507. Cremanthodium Decaisnei C. B. Clarke 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 4] Eastern Tibet: grasslands between Labrang and Yellow River, no. 14464; alpine region between Radja and Jupar ranges, nos. 14159, 14393; Amnyi Machen range, no. 14428. Cremanthodium discoideum Maxim Eastern Tibet: alpine region Geeveed Radja and Jupar ranges, no. 404 Cremanthodium humile Maxim. Southwestern Kansu: Upper Tebbu country, no. 13022; Tao River basin, no. 13183 Cremanthodium Limprichtii Diels Southwestern Kansu: Tao River basin, no. 12618. Cremanthodium lineare Maxim. Eastern Tibet: Kokonor region, no. 13378. Cremanthodium plantagineum Maxim. Southwestern Kansu: Upper Tebbu country, no. 13030. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14471; alpine region between Radja and Jupar ranges, nos. 14383, 14387; Ba Valley, no. 14357; Jupar Range, no. 14308 Cremanthodium spec. Southwestern Kansu: Upper Tebbu country, nos. 13012, 13049, Echinops Turczaninowii Ledeb. Northwestern Kansu: no. 13396. Saussurea amara Southwestern Kansu: Tao River basin, no. 13201. Saussurea apus Maxim. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14474 Saussurea arenaria Maxim. Eastern Tibet: grasslands between Labrang and Yellow River, no. 1 ; Saussurea Giraldii Diels Central Kansu: no. 12717. Saussurea hypsipeta Diels Eastern Tibet: Jupar Range, no. 14367; alpine region between Radja and Jupar ranges, no. 14388; Amnyi Machen range, no. 14411 Saussurea medusa Maxim. Southwestern Kansu: Tao River basin, no. 12385; Upper Tebbu country, no. 13044. Eastern Tibet: alpine region between Radja and Jupar ranges, no. 14390; Jupar Range, no. 14366; Amnyi Machen range, no. 14429 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Saussurea nigrescens Maxim. Southwestern Kansu: Upper Tebbu country, no. 13067. Saussurea phaeantha Maxim. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14516. Saussurea polystichoides Hook. f. Southwestern Kansu: Tao River basin, no. 13737. Saussurea poophylla Diels Southwestern Kansu: Upper Tebbu country, no. 13001. Saussurea pygmaea Spreng. var. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14458 Saussurea stella Maxim. Eastern Tibet: Kokonor region, nos. 13349, 13379. Southwestern Kansu: Tao River basin, no. 13666 Saussurea Thoroldi Hemsl. Eastern Tibet: Kokonor region, nos. 13376, 13377. Saussurea spec. aff. S. poophylla Diels Central Kansu: no. 12721. Saussurea spec. Southwestern Kansu: Tao River basin, nos. 13156, 13180, 13199, 13200, 13739, 14617, 14618; Upper Tebbu country, nos. 13034, 13036, 13121, 14604, 14626; Lower Tebbu country, no. 14817. Central Kansu: nos. 12740, 12771. Northwestern Kansu: 13386, Eastern Tibet: Kokonor region, nos. 13700, 13702; grasslands between Labrang and Yellow River, no. 14485; Jupar Range, nos. 14337, 14368: Amnyi Machen range, nos. 14413, 14425, 14433. Carduus euosmos Forrest apud W. W. Eastern Tibet: Kokonor region, no. 13371. Southwestern Kansu: Upper Tebbu country, no. 14582. Cirsium arvense Scop. Eastern Tibet: Amnyi Machen range, no. 14443. Cirsium setosum Bieb. Central Kansu: no. 12764. Southwestern Kansu: Tao River basin, no. 13718. Cirsium Souliei (Franch.) Mattfeld, comb. nov. Cnicus Souliet Franch. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14496 Cirsium spec. Southwestern Kansu: Upper Tebbu country, no. 13122. Scorzonera austriaca Willd. 1933] REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 43 Southwestern Kansu: Tao River basin, no. 12169. Eastern Tibet: Radja and Yellow River gorges, no. 14104 Taraxacum mongolicum Hand.-Mazz. Eastern Tibet: Radja and Yellow River gorges, no. 13979. Lactuca spec. aff. L. Souliei Franch. Southwestern Kansu: Tao River basin, no. 13729. Crepis Hookeriana C. B. Clarke Southwestern Kansu: Tao River basin, no. 12645. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14517; alpine region between Radja and Jupar ranges, no. 14386; Amnyi Machen range, no. 14442. Crepis paleacea Diels aff. Southwestern Kansu: Upper Tebbu country, no. 13062. Crepis rosularis Diels aff. Southwestern Kansu: Upper Tebbu country, no. 13025. Crepis spec. aff. C. trichocarpa Franch. Eastern Tibet: grasslands between Labrang and Yellow River, no. 14529. Prenanthes spec. Southwestern Kansu: Upper Tebbu country, no. 14591. LIST OF SEEDS OF LIGNEOUS AND HERBACEOUS PLANTS COLLECTED IN 1925 AND 1926 BY J. F. ROCK AND DISTRIBUTED BY THE ARNOLD ARBORETUM 12131. Pyrus pashia Ham. 13244. Ampelopsis aconitifolia Bge. 12141. Viburnum fragrans Bge. var. palmiloba (Carr.) 12193. Lonicera hispida Pall. Re hid. 12270. Lonicera syringantha 13256. Berberis Vernae Schneid. Maxim. 13257. Lycium chinense Mill. 12297. Lonicera syringantha 13261. Berberis Boschanii a Maxim. 13262. Berberis dasystachy 12432. Prunus setulosa Batal. axim. 12475. Lonicera hispida Pall. 13263. Berberis diaphana Maxim. 12527. Prunus stipulacea Maxim. 13264. Rosa Sweginzowii Koehne 12532. Lonicera tangutica Maxim. 13265. Rosa bella Rehd. & Wils. 12797. Berberis parvifolia Sprague 13266. Rosa Sweginzowii Koehne 12893. Lonicera coerulea L. var. 13267. Sorbus thianschanica Rupr. edulis Reg. 13268. Sorbus Koehneana Schneid. 13197. Ribes ona oe 13269. Cotoneaster tenuipes Rehd. 13202. Berberis Mouillac & Wils. Schneid. 13270. Hippophoé rhamnoides L. 13225. eee uniflora Batal. 13271. Rosa omeiensis Rolfe 13240. Nitraria Schoberi L. 13272. Berberis diaphana Maxim. 13243. Ailanthus reels ( Mill.) 13273. Berberis diaphana Maxim. Swingle 13274. Berberis Boschanii Schneid. 44 13275. 13276. 13277. 13278. 13279. 13280. 13281. 13282. 13283. 13285. 13286. 13288. 13289. 13290. i291; 13292. 13293. 13294. 13295. 132906. 13297. 13298. 13299, 13300. 13301. 13502. 13303. 13304. 13305. 13307. 13308. 13309. 13310. 13311. JOURNAL OF THE ARNOLD ARBORETUM Hippophoé rhamnoides L. Lonicera syringantha Maxim. tangutica Korsh. Rehd. Clematis var. obtusiuscula ils iadadendcon Przewalskii axim. Rhododendron di noides Maxin Caragana brevifolia Komar. Picea asperata Mast. Picea asperata Mast. Betula japonica Sieb. var. Rockii Rehd. Rubus idaeus L. var. strigosus Maxim. Sorbus thianschanica Rupr. Lonicera nervosa Maxim. Berberis Vernae Schneid. Berberis kansuensis Schneid. eres a alia Max Berberis sca Schneid. Rosa bella Rehd. & Wils. Cotoneaster acutifolia Turcz. Berberis diaphana Maxim. Cotoneaster multiflora Bge. Rhamnus leptophylla Schneid. var. scabrella Rehd. Lonicera syringantha Maxim. Rosa Willmottiae Hemsl. Berberis Boschanii Schneid. Incarvillea compacta Maxim. Rhododendron Przewalskii Maxim. Rhododendron thymifolium Maxim. Picea asperata Mast. Juniperus zaidamensis Komar Picea asperata a Mas Clematis rae Ta DC. Picea asperata Mast. Picea asperata Mast. Evonymus Przewalskii Maxim. 13313. 13314. 13315. 13316. 13517, 13318. 13319. 13320. 13320 a. 13321. EK yA 13341. 13422. 13423. 13424. 13425. 13426. 13427. 13428. 13429. 13430. 13431. 13432. 13433. 13434. 13435. 13430. 13437. 13438. 13439. 13440. 13441. 13442. 13443, 13444. 13445. [VOL. XIV Picea asperata Mast. Berberis Caroli Schneid. Berberis Vernae Schneid. Berberis Boschanii Schneid. Sophora alopecuroides L. Sorbus thianschanica Rupr. Elaeagnus angustifolia La Nitraria Schoberi L. Zygophyllum xanthoxylum Potentilla Salesoviana Steph. ep Clematis fruticosa a var. canescens Turc Picea asperata Mast. Abies Faxoniana Rehd. & Wil ils. Abies Faxoniana Rehd. & Wils. Picea purpurea Mast. Abies Faxoniana Rehd. & Picea Wilsonii Mast. Picea purpurea Mast. Picea asperata Mast. Abies Faxoniana Rehd. & Wils. Picea purpurea Mast. Picea asperata Mas Juniperus foriapaand Hay. Picea asperata Mast. Picea purpurea ie Abies — sutchuenensis (Franch.) Rehd. & Wils. Abies Faxoniana Rehd. & Wils Abies Paxndien Rehd. & Wils. Picea asperata reat Picea purpurea Abies Faxoniana Rehd, & Wils. Picea purpurea Mast. Picea purpurea Mast. Picea Wilsonii Mast. Abies Faxoniana Rehd. & Wil ils. Abies Faxoniana Rehd. & Wils. 1933] 13446. 13447. 13448. 13449. 13450. 13451. 13452. 13453. 13454. 13455. 13456. 13457. 13458. 13468. 13469. 13470. 13471. 13472. 13473. 13474. 13475. 13476. REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 45 Picea asperata Mas Abies Faxoniana =) & Wils. sae oeees Rehd. & Pinus ones Carr. Abies sutchuenensis (Franch.) Rehd. & Wils. Abies sutchuenensis (Franch.) Rehd. & Wils. Picea Wilsonii Mast. Picea Wilsonii Mast. Abies — sutchuenensis (Franch.) Rehd. & Wils. Picea purpurea Mast. Abies sutchuenensis Cap ) Rehd. & me nOrTaRe Mast. . Picea purpurea Ma . Pinus tabulaeformis Ga . Picea purpurea Mast. . Pinus Armandi Franch. . Picea asperata Mast. . Larix Potaninii Batal. . Abies Faxoniana Rehd. & Wils. . Abies sutchuenensis (Franch.) Rehd. ils. Juniperus saltuaria Rehd. & Wils. Juniperus distans Florin Juniperus squamata Buch.- Ham. var. Fargesii Rehd. ils. Juniperus saltuaria Rehd. & Wils. Picea purpurea Mast. Viburnum betulifolium atal Berberis diaphana Maxim. Cotoneaster adpressa Bois Viburnum betulifolium Batal. 13477. 13478. 13479. 13480. 13481. 13482. 13483. 13484. 13485. 13486, 13487. 13488. 13489. 13490. 13491. 13492. 13493. 13494, 13495. Sorbus Koehneana Schneid. Cotoneaster horizontalis ecne Sorbus hupehensis Schneid. var. aperta (Koehne) neid. Cotoneaster multiflora Bge. ar. calocarpa Rehd. & ils. Aralia chinensis L. var. nuda akai ene liaotungensis oidz. eae macrophylla Wall. Tilia chinensis Maxim Virburnum Sargenti Keshne var. calvescens Rehd. Tilia chinensis Maxim. Quercus liaotungensis Koidz. Acer Maximowiczii Pax Acer tetramerum Pax var. betulifolium Rehd. Clematoclethra lasioclada axim. Acanthopanax Giraldii Harms ar Viburnum Sargenti Koehne var. calvescens Rehd. Cotoneaster multiflora Bge. var. calocarpa Rehd. & il ils. Sorbus Prattii Koehne Caragana frutex K. Koch 13495a. Rosa Sweginzowii Koehne 13496. 13497, 13498. 13499. 13500. 13501. 13502. 13503. 13504. 13505. 13506. 13507. 13508. 13509. Pyrus pashia Ham Malus kansuensis Schneid. Malus baccata Borkh. Pyrus ussuriensis Maxim. var. ovoidea Rehd. Crataegus kansuensis Wils. Crataegus kansuensis Wils. Malus baccata Borkh. Celtis Bungeana BI. Prinsepia uniflora Batal. Ribes Giraldii Jancz. Syringa pekinensis Rupr. Syringa pekinensis Rupr. Evonymus phellomana Loes. Lonicera heteroloba Batal. 46 13510. T3511 13512. 135135: 13514. 13515. 13516. 13517. 13518. 13519. 13520. 13921, 13522. 13523. 13524. 13525. 13526. 1352/, 13528. . Berberis diaphana . Berberis parvifolia eeu . Daphne tangutica Maxim. . Sorbus tapashana Schneid. . Caragana jubata Poiret 13534. tone: 13536. 13537. 13538. 13539. 13540. 13541. JOURNAL OF THE ARNOLD Daphne Giraldii Nitsche Corylus Sieboldiana BI. var. mandshurica Schneid. Lonicera chrysantha Turcz. ar. longipes Maxim. oe Davidiana ee dasystachya Max im. Viburnum mongolicum (Pall.) Rehd. Lonicera chrysantha Turcz. var. longipes Maxim. Cotoneaster multiflora Bge. axim. Viburnum glomeratum Lonicera Ferdinandi Franch. Lonicera chrysantha Turcz. var. longipes Maxim. Evonymus phellomana Loes. Berberis Silva-Taroucana Schneid. Cotoneaster racemiflora K. Koch var. soonga- rica Schneid. Sorbus Koehneana Schneid. Berberis diaphana Maxim. Berberis kansuensis Schneid. Prunus salicina Lindl. Berberis kansuensis oS Acanthopanax Giraldii arms Cotoneaster adpressa Bois a Maximovicziana Komar. Malus transitoria (Batal.) id. Sorbus Prattii Koehne Caragana jubata Poiret 13542. 13543. 13544. 13545. 13546. 13547. 13548. 13549. 13550. 13551. 13a. 13554. Ioooe: 13556. 13557. 13558. 13509, 13560. 13561. 13562. 13563. 13564. 13565. 13566. 13567. 13568. 13569. 13570. 13571. 13572. 13573. ARBORETUM [VOL. XIV Acer tetramerum Pax var. betulifolium Rehd. Malus lia aa (Rehd. ) Hug Malus A een (Rehd. ) Hughes g Berberis brachypoda Maxim. Rubus pileatus Focke Rosa omeiensis Rolfe Cotoneaster acutifolia Turez Pyrus pashia Ham. Lonicera syringantha Maxim. Lonicera coerulea L. var. edulis Reg Rubus idaeus L. var. strigosus Max Evonymus congas Loes. & Prunus tangutia ( Batal.) Koehn tania heteroloba Batal. Berber ae Sch Pyrus gee shia Har Malus toringodes ‘(Rehd. ) Hug ae aneess Mill. Berberis dasystachya Maxim. Caragana brevifolia Komar. Sorbus Prattii Koehne Berberis parvifolia Sprague Malus toringoides (Rehd.) Hugh 1es Sorbus tapashana Schneid. Caragana jubata Poiret Malus toringoides (Rehd.) Hughes Evonymus Giraldii Loes. var. ciliata Loes. Rosa Sweginzowii Koehne? Hippophoé rhamnoides L. Daphne tangutica Maxim. Rosa Sweginzowii Koehne 1Under this number two sets of seeds were distributed. One set corresponded to the species cited above while the second set proved to be seed of Prun var. commutata nus Padus 1933] 13574. 15379: 13576. 13577. 13578. 13579. 13580. 13581. 13582. 13583. 13584. 13585. 13586. 13587. 13588. 13589. 13590. 13591. 13592. 13593. 13594. 13595. 13596. 13597. 13598. 13599. 13600. 13601. 13602. 13603. 13604. 13605. 13606. 13607. REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 47 Berberis diaphana Maxim. Rosa omeiensis Rolfe Lonicera deflexicalyx Batal. ene dus L. var commutata Dipp. hae chrysantha Turcz. var. longipes Maxim. Ribes Meyeri Maxim. Lonicera tangutica Maxim. var. glabra Batal. Rubus amabilis Focke Lonicera nervosa Maxim. Lonicera thibetica Bur. & h ranch. Lonicera hispida Pall Lonicera heteroloba Batal. Ribes Meyeri Maxim Evonymus Przewalskii axim. Lonicera deflexicalyx Batal. Sorbus Koehneana Schneid. Rosa omeiensis Rolfe Rosa Willmottiae Hemsl. Lonicera szechuanica Batal. Paeonia anomala Spiraea gemmata Zab. Sibiraea pect Maxim. var. angustata Rehd. Rhododendron pares Rhododendron anthropogo- i Maxim. aod Saale thymifolium Re asien iecai rufum Batal. Rhododendron capitatum axim. Rhododendron rufum Batal. Hydrangea Bretschneideri ipp. ae ie pekinensis . kansuensis Rehd. ee Bretschneideri face aenilco, capitatum Maxim. Lilium tenuifolium Fisch. & Mey. Acer caudatum Wall. var. multiserratum Rehd. 13608. 13609. 13610. 13611. 13612. 13613. 13614. 13615. 13616. 13617. 13618. 13619. 13620. 13621. 13622. 13624. 13625. 13626. 13627. 13628. 13629, 13630. 13631. 13632. 13633. 13634. 13635. 13636. Rhamnus leptophylla Schneid. var. scabrella Arctous ruber Nakai Rhododendron fe a noides Max Rhododendron Capitatarh Rive eines Przewalskit axim. Rhododendron rufum Batal. Clematis brevicaudata DC. Clematis tangutica Korsh. var. obtusiuscula Rehd. & Wils. Betula albo-sinensis Burk. Buddleia alternifolia Maxim. Lilium Duchartrei Franch. Primula tangutica Duthie. Meconopsis quintuplinervia eg. Betula albo-sinensis Burk. Rhododendron capitatum axim. Clematis Fargesii Franch. Anemone narcissiflora L. Anemone Rockii Ulbrich Meconopsis racemosa axim. Rhododendron rufum Batal. Rhododendron Przewalskii Maxi axim. Rhododendron rufum Batal. Primula? chionantha Balf. f. Forrest. Meconopsis psilonomma Farrer penal optata Balf. f. & Farre Rhododendron capitatum Maxim. oo capitatum ee ak re noides Max . Iris ensata Thunb. . Caragana densa Komar. . Primula flava Maxim . Rhododendron rufum Batal. 13641. Betula albo-sinensis Burk. 48 13642. 13643. 13644. 13645. 13646. 13647. 13648. 13649. 13650. 13651. 13652. 13653. 13654. 13655. 13656. 13657. 13658. 13659. 13670. 13671. 13672. 13673. 13674. 13675. 13676. 13677. 13678. 13679. 13680. 13681. JOURNAL OF THE ARNOLD Meconopsis punicea Maxim. Rhododendron rufum Batal. Betula japonica Sieb. var. szechuanica Schneid. Rhododendron rufum Batal. Betula albo-sinensis Burk. Rhododendron does Batal. albo-sinen trionalis Schneid Rhododendron rufum Batal. Rhododendron rufum Batal. Meconopsis integrifolia Franch. Primula polyneura Franch. Incarvillea compacta xi axim. Primula limbata Balf. f. & Forrest Primula tangutica Duthie Primula chionantha Balf. f. Forrest Betula japonica Sieb. var. szechuanica Schneid. Primula ? Woodwardii Balf. f Philadelphus _ pekinensis Rupr. var. kansuensis Rehd. Sorbaria arborea Schneid. var. glabrata Rehd. a Bretschneideri pee aa fruticosa L. var. parvifolia Wo Potentilla fruticosa L. var. dahurica Ser. Rhododendron capitatum axim. Rhododendron rufum Batal. a Przewalskii Max Rhoden Przewalskii xim. Rhododendron rufum Batal. Rhododendron Przewalskii axim. Rhododendron rufum Batal. Rhododendron Przewalskii Maxim. 13682. 13683. 13684. 13685. 13686. 13687. 13688. 13689. 13690. 13691. 13692. 13693. 13694. 13695. 13696. 13697. 13698. . Picea asperata Mast. . Berberis parvifolia Sprague . Juniperus glaucescens — . Juniperus tibetica Kom . Picea asperata Mast. . Prinsepia uniflora Batal. . Prunus salicina Lindl. . Cotoneaster multiflora Bege. . Rhamnus leptophylla hneid 14801. 14802. ARBORETUM [VOL. XIV Rhododendron Rhododendron Rhododendron Rhododendron rufum Batal. rufum Batal. rufum Batal. Przewalskii Maxim. Rhododendron Przewalskii axim. Philadelphus — pekinensis Rupr. var. kansuensis ehd. Rhododendron capitatum axim. Spiraea alpina Pall. Spiraea longigemmis Maxim. (pubescent form Rhododendron Rhododendron Rhododendron aa rity rufum Batal. rufum Batal. rufum Batal. Przewalskii Max Rhododendron Przewalskii Max Ri cdedeninon rufum Batal. Rhododendron rufum Batal. Syringa microphylla Diels chneid. . Cornus macrophylla Wall. . Meliosma cuneifolia — . Aralia chinensis L. va nuda Nakai . Betula japonica Sieb. var. szechuanica Schneid. . Lonicera deflexicalyx Batal. . Lonicera chrysantha Turcz. var. longipes Maxim. . Lonicera Ferdinandi I Franch. Cotoneaster multiflora Bge. Crataegus kansuensis Wils. f. aurantiaca Wils. 1933] 14804. 14805. 14806. 14807. 14810. 14816. 14823. 14825. 14829. 14834. 14857. 14860. 14864. 14866. 14875. 14892. 14893. 14894, 14895. 14896. 14897, 14898. 14899. 14900. 14901. 14902. 14903. 14905. REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 49 Berberis kansuensis Schneid. Cotoneaster nitens Rehd. & Wils. Pyrus pashia Ham Lonicera Geechiatics Batal. Pyrus pashia Ham. albo-sinensis Burk. var. septen- trionalis Schneid. Aralia chinensis L. var. Cotoneaster Zabeli Schneid. Cotoneaster obscura Rehd. Wils. var. cornifolia Rehd. & Wils. Prunus pubigera Koehne var. Prattii Koehne Aconitum volubile Pall. var. flexuosus Rchb. Juniperus distans Florin Acanthopanax Giraldii arms Juniperus distans Florin . Berberis Vernae Schneid. Prunus tangutica Koehne . Lonicera heteroloba Batal. . Lonicera deflexicalyx iF Bata Rosa Willmottiae Hemsl. Prunus Padus L. var. commutata Dipp. Daphne tangutica Maxim. Rubus amabilis Focke Rosa Sweginzowii Koehne Cotoneaster multiflora Bge. Berberis Mouillacana Schneid. Berberis kansuensis Schneid. Rosa omeiensis Rolfe Podophyllum emodi Wall. Sorbus Prattii Koehne Rosa Sweginzowii Koehne Cotoneaster multiflora Bge. ’ 14906. 14907. 14908. 14909. 14910. 14911. 14912. 14913. 14914. 14915. 14916. 14917. 14918. 14919. 14920. 14922. 14923. 14924. 14925. 14926. 14927. 14928. 14929. 14930. 14931. 14933. 14938. 14939, 14940. 14941, 14942. . Berberis Vernae Schneid. . Rosa Willmottiae Hemsl. 14945. 14946. Lonicera Ferdinandi Rosa Sweginzowii Koehne Crataegus kansuensis Wils. Prunus salicina Lindl Malus transitoria Schneid. Berberis parvifolia Sprague Rosa Willmottiae Hemsl. Berberis Silva-Taroucana neid. Cotoneaster multiflora Bge. Berberis Vernae Schneid. Betula japonica Sieb. var. szechuanica Schneid. Crataegus kansuensis Wils. Malus kansuensis Schneid. Sorbus Prattii Koehne Primula ? chionantha Balf. f. & Forrest forma Malus transitoria Schneid. Malus toringoides (Rehd.) Hughes ug Pyrus pashia Ham. Malus kansuensis X toringoides Rehd. Philadelphus _pekinensis Rupr. var. kansuensis Rehd. Rosa Sweginzowii Koehne Rhododendron rufum Batal. Abies Faxoniana Rehd. & ils Wils. Abies sutchuenensis (Franch.) Rehd. & Wils. Abies sutchuenensis (Franch.) Rehd. & Wils. Caragana jubata Poiret Malus kansuensis Schneid. Crataegus kansuensis Wils. Quercus Baronii Skan Quercus liaotungensis Koidz. Quercus Baronii Skan Clematoclethra lasioclada Acer Maximowiczii Pax 50 14947. 14948. 14949. 14950. 14951. 14952. 14953. 14954. 14955. 14956. 14957. 14958, 14959, 14960, 14964. 14965. 14966. 14967. 14968. 14969. 14970. 14971. 14972. 14973. 14974, 14975. 14976. 14977. 14978. 14979. JOURNAL OF THE ARNOLD ARBORETUM seta ee Franch. . tripar cco. Rupr. var. kansuensis Rehd. Acer tetramerum Pax var. betulifolium Rehd. Quercus Baronii Skan. Juniperus chinensis L. var. pendula Franch. Juniperus chinensis L. Ribes Giraldii Jancz. Lonicera Ferdinandi ranc Evonymis nanoides Loes. & Rehd. Crataegus kansuensis Wils. Prinsepia uniflora Batal. Sageretia theezans Brongn. Quercus liaotungensis Koi idz. Lonicera Ferdinandi Picea Wilsonii Mast. Caragana tangutica Maxim. Betula albo-sinensis Burk. var. septen- trionalis Schneid. Quercus liaotungensis oidz. Cornus macrophylla Wall. Betula albo-sinensis urk. var. septen- trionalis Schneid Clematoclethra lasioclada Maxim. Viburnum betulifolium atal. f. aurantiacum hd Cotoneaster adpressa Bois Caragana densa Komar. Quercus liaotungensis Koidz. Picea Wilsonii Mast. Hydrangea Bretschneideri ipp. Lonicera nervosa Maxim. Evonymus alata Reg. var. aperta Loes. Rubus pileatus Focke 14980. 14981. 14982. 14983. 14984. 14985. 14986. 14991. 14992. 14993. 14994. 14995. 14996. 14997. 14998. 14999, 15000. 15001. 15002. 15003. 15004. 15005. 15006. 15007. 15008. [VOL. XIV Cotoneaster lucida Schlecht. Cotoneaster nitens Rehd. & Wils. Spiraea Wilsonii Duthie Acanthopanax Giraldii arms Acer tetramerum Pax var. betulifolium Rehd Clematoclethra integrifolia Maxim. Viburnum glomeratum m. . Sorbus Koehneana Schneid. . Rosa Sweginzowii Koehne . Abies Faxoniana Rehd. & Wils. . Sorbus hupehensis Schneid. (Koehne ) Hippophoé rhamnoides L. Cotoneaster multiflora Bge. var. calocarpa Rehd. & S. Lonicera chrysantha Turcz. var. longipes Maxim. Koelreuteria paniculata Laxm Quercus Baronii Skan Exochorda Giraldii Hesse Acer tetramerum Pax var. betulifolium Re Acer Maximowiczii Pax Tilia chinensis Maxim. Acer pictum Thbg. var parviflorum Schneid. Tilia chinensis Maxim. Quercus liaotungensis oidz. Viburnum Maxim. glomeratum var. Rockii Bi caleanan micranthum urcz. Hydrangea Bretschneideri Dipp. var. glabrescens Wikstroemia chamaedaphne Clematis fruticosa Turcz. Clematis aethusifolia Turcz. 1933] 15009. 15010. 15011. 15012. an03. 15014. 15015. 15016. 15017. 15018. 15019. 15020. 15021. 15022. 15023. 15024. 15025. 15026. 15028. 15029, 15030. 15031. 15032. 15033. 15034. 15035. 15036. 15037. 15038. 15039. REHDER AND KOBUSKI, PLANTS COLLECTED BY ROCK 51 Abies Faxoniana Rehd. & Wils. Acer caudatum Wall. var. multiserratum Rehd. Viburnum betulifolium Batal. Buddleia albiflora Hemsl. Viburnum betulifolium atal. Rhododendron rufum Batal Clematoclethra lasioclada Maxim. Hydrangea longipes Franch. Betula japonica Sieb. var szechuanica Schneid. Betula japonica Sieb. var. szechuanica Schneid. Pinus Armandi Franch. Abies sutchuenensis (Franch.) Rehd. & Wils. Picea Wilsonii Mast. Lonicera chrysantha Turcz. var. longipes Maxim Acanthopanax jevicorthizes arms Sorbus hupehensis Schneid. var. aperta (Koehne) Schneid Sorbus Koehneana Schneid. Actinidia tetramera Maxim. Viburnum betulifolium Batal. Berberis brachypoda Maxim. Quercus liaotungensis Koidz. Acer Maximowiczii Pax Acer tetramerum Pax. var. Tilia chinensis Maxim. Jasminum humile L Syringa oblata Lindl. var. Giraldii Rehd. Quercus Baronii Skan Ligularia yesoensis Franch. ar. sutchuensis Franch. Hydrangea Bretschneideri Dipp. var. glabrescens Rehd. 15040. 15041. 15042. 15043. 15044. 15045. 15046. 15047. 15048. 15049. 15050. 15051. 15052. 15053. 15054. 15055. 15056. 15058. 15059. 15060. 15061. 15062. 15063. 15064. 15065. 15066. 15067. 15068. 15069. 15070. 15071. 15072. 15073. 15074. Acer caudatum Wall. var. multiserratum Rehd. Acer Maximowiczii Pax Tilia chinensis Maxim. Pyrus pashia Ham Abies Faxoniana Rehd: & Wils. Picea asperata Mast. Philadelphus pekinensis Rupr. var. kansuensis Rehd. Acer Maximowiczii Pax Picea Wilsonii Mast. Spiraea uratensis Franch. Rhamnus leptophylla Schnei Acer tetramerum Pax var. betulifolium (Maxim.) Rehd. Tilia chinensis Maxim. Acer Grosseri Pax Pyrus pashia Ham. Aster Limprichtii Diels Indigofera Bungeana Walp. Syringa oblata Lindl. var. Giraldii (Sprenger) Rehd. Exochorda Giraldii Hesse Ailanthus altissima ( Mill.) ingle Caragana densa Komar. Caragana tangutica Maxim. Syringa oblata Lindl. var. Sea (Sprenger ) a ae Mast. Picea asperata Mast. Juniperus squamata Buch.- Ham. var. Fargesii Rehd. il & Wils. Smilax rubriflora Rehd. Evonymus phellomana Loes. Quercus liaotungensis Koidz. Malus toringoides (Rehd.) ughes 52 15075. 15076. 15077. 15078. 15079. 15080. 15081. 15082. 15083. HERBARIUM, ARNOLD ARBORETUM Harvarp UNIVERSITY JOURNAL OF THE ARNOLD ARBORETUM Quercus liaotungensis oidz. Philadelphus pekinensis Rupr. var. kansuensis Rehd Pertya sinensis Oliver Lonicera Ferdinandi Franch. Abies sutchuenensis (Franch.) Rehd. & ils. Picea asperata Mast. Abies Faxoniana Rehd. & Wils. Abies Faxoniana Rehd. & ils. Betula albo-sinensis Burk. 15084. 15985. 15086. 15094. 15095. [VOL. XIV Abies Faxoniana Rehd. & Wils. Aralia chinensis L. var. nuda Nakai Picea Wilsonii Mast. . Picea Wilsonii Mast. . Picea asperata Mast. 15093. ee Beene Buch.- i. r. Fargesii Rehd. Wi Is Lycium chinense Mill. Bocconia spec. 1933] GUILLAUMIN, FLORA OF THE NEW HEBRIDES 53 CONTRIBUTION TO THE FLORA OF THE NEW HEBRIDES PLANTS COLLECTED BY S. F. KAJEWSKI IN 1928 AND 1929 SUPPLEMENT} A. GUILLAUMIN MENISPERMACEAE Pycnarrhena sp. Banks Group: Vanua Lava Island, rain-forest, sea level, no. 410, July 5, 1928 (vine growing on rain-forest trees; fruit orange when ripe). The eastern limit of this genus had been until now New Guinea and Queensland. CAPPARIDACEAE Capparis sp. aff. C. artensis Montr. (Art Island) vel C. subacuta Miq. (Borneo). E fate: Undine Bay, common at sea level, no. 228, April 28, 1928 (vine climbing on top of trees; fruit purple when ripe). FLACOURTIACEAE Flacourtia sp. (teste E. D. Merrill). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 120 m., no. 812, Feb. 23, 1929 (large tree up to 18 m. high; fruit yellow when ripe). Xylosma sp. aff. X. lifuana Guillaumin (Loyalty Islands). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 60 m., no. 745, Feb. 12, 1929 (small tree up to 6 m. high; flowers greenish yellow). PITTOSPORACEAE Pittosporum naruaiao Guillaumin, sp. nov. Arbor, trunco 1 m. diam., ramis gracilibus fulvis, foliis ad ramulorum apicem confertis lanceolatis (usque ad 14 cm. X 3 cm.) apice acutis acuminatisve basi cuneatis membranaceis, nervis parum conspicuis, petiolo gracili 1-2 cm. longo. Inflorescentiae umbellatae, 2-2.5 cm. longae, pedunculo rufo-puberulo, bracteis parvis triangularibus extra rufo-puberulis, pedicellis 5 mm. longis rufo-puberulis, floribus albis (tantum in alabastris 3 mm. longis), calyce glabro campanulato 1 mm. longo, lobis ovatis tubo 2-plo brevioribus, petalis ovatis 1.7 mm. longis, staminibus petalis fere aequilongis, filamentis antheris 2-plo brevioribus See Vol. xu. 221-264, x11. 1-29, 81-126 for the preceding parts. 54 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV complanatis ovatis, antheris linearibus apice apiculatis, ovario turbinato staminibus fere aquilongo. Fructus rubri (4 cm. x 2.5 cm. Aneityum: west side, Nowhow, common in rain-forest, alt. 600 m., no. 1004 (coll. J. P. Wilson), Sept. 1929 (large tree, up to 1 m. in diameter; flowers white; fruit red). Vernacular name “Naruaiao.” GUTTIFERAE Garcinia sp. E fate: Undine Bay, common in rain-forest, alt. 100 m., no. 213, April 26, 1928 (tree 15 m. high; fruit orange). TERNSTROEMIACEAE Eurya japonica Thunberg, Nov. Gen. 67 (1781). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 60 m., no. 747, Feb. 12, 1929 (small tree up to 6 m. high; petals white; fruit black when ripe). Tanna: Mt. Tokosh Meru, common in rain- forest, alt. 300 m., no. 158, March 15, 1928 (tree up to 15 m. high).— Already found on Aneityum, (MacGillivray 58)! also in the Fiji, Caro- line and Palau Islands and in Malaysia. STERCULIACEAE Kleinhovia hospita Linnaeus, Sp. Pl. ed. 2, 1365 (1763). Efate: Undine Bay, common in rain-forest at sea-level, no. 199, April 24, 1928 (shade tree with dense leaves 15-20 m. high; flowers pink ).—Already found on Aneityum and Efate; also Fiji, Society, Solo- mon and Bismarck Islands, New Guinea and Malaysia. ELAEOCARPACEAE Elaeocarpus sp. Aneityum: Anelgauhat Bay, common in rain-forest, alt. 300 m., no. 938. March 17, 1929 (medium sized tree up to 15 m. high). BURSERACEAE Canarium aneityense Guillaumin, sp. nov. Arbor magna, foliis usque ad 28 cm. longis, foliolis 3-5 ovatis apice longe caudato-acuminatis basi obtusis rotundatis truncatisve rigidis, nervis 7-jugis, petiolulis terminali 4 cm., lateralibus 2 cm. longis apice tumidis, petiolo 3.5-6 cm. longo, stipulis ?. Inflorescentiae ad apicem axillares, paniculatae, 13 cm. longae, ramis gracilibus inferioribus 8 cm. longis, floribus albo-viridibus, ¢ minutis, pedicello 2-3 mm. longo medio articulato basi minute bracteolato, sepalis ovato-triangularibus 2 mm. longis, petalis ovatis aequilongis, staminibus filamentis brevissi- 1Not mentioned in my “Liste des plantes connues [des Nouvelles Hebrides] in Bull. Soc. Bot. France, LXxtv (1927). 1933] GUILLAUMIN, FLORA OF THE NEW HEBRIDES 55 mis, antheris petalis 2-plo brevioribus, disco annulari, pistillodio mi- nimo, stylo terete brevi. Aneityum: Anelgauhat Bay, common in rain-forest, alt. 150 m., no. 943; March 19, 1929 (large tree; Eonar green-white ) Valuable cabinet wood; vernacular name “Inyat MELIACEAE Melia Azedarach Linnaeus, Sp. Pl. 384 (1753). Aneityum: Anelgauhat Bay, common in rain-forest at sea level, no. 897, March 11, 1929 (large tree up to 18 m. high; fruit red when ripe).—Vernacular name “White Cedar.” DICHAPETALACEAE Dichapetalum sp. aff. D. missionum Krause (Bismarck Islands) vel D. validum Krause (New Guinea). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 300 m., no. 763, Feb. 14, 1929 (vine climbing up rain-forest trees; fruit brown when ripe). OLACACEAE Ximenia americana Linnaeus, Sp. Pl. 1193 (1753). Aneityum: Anelgauhat Bay, common along sea shore at sea level; no. 797, Feb. 21, 1929 (rambling tree growing to the height of 4.5 m.; fruit yellow when ripe, with almond taste).—Also in New Cale- donia, Loyalty, Fiji, Society and Mariana Islands, New Guinea, Aus- tralia (Queensland) and Malaysia.. AQUIFOLIACEAE Tlex sp. aff. 7. peduncularis F. Muell. (Queensland). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 60 m., no. 746, Feb. 12, 1929 (large tree up to 18 m. high; fruit red when ripe). HIPPOCRATEACEAE Salacia aneityensis Guillaumin, sp. nov. Scandens, glaberrimus etiam novellis, foliis atro-viridibus subdis- coideis (usque ad 11 cm. x 8 cm.) apice rotundatis vel obtusissimis basi brevissime subitoque cuneatis integerrimis membranaceis, nervis 5-7- jugis tenuibus subtus prominulis, petiolo 1 cm. longo. Flores fascicu- lati, pedunculo 6 mm. longo, sepalis ovato-triangularibus vix 0.5 mm. longis, petalis ovatis fere 1.5 mm. longis brevissime unguiculatis, disco crasso conico sepalis altiore, staminibus filamentis complanatis brevi- bus, antheribus transverse ovatis transversaliter dehiscentibus filamen- tis vix longioribus, ovario conico, disco fere continuo et in stylum sensim attenuato, loculis 3, 2-ovulatis. Fructus obovati (7.5 cm. 1 cm.), 2-loculares. 56 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Aneityum: Anelgauhat Bay, common in rain-forest, at sea level, no. 853, March 4, 1929 (vine growing on rain-forest trees; leaves dark green; petals light green). Related to S. prinoides DC. of Malaysia, New Guinea, Bismarck and Solomon Islands, but the leaves are different. SAPINDACEAE Pometia pinnata Forster, Char. Gen. 110 (1776). Tanna: Lenakel, common in heavy rain-forest soil at sea level, no. 19, Feb. 21, 1928 (tree up to 15 m. high; fruit brown when ripe) .— Already found on Tanna; also Fiji, Samoa, Marshall and Bismarck Islands, New Guinea and Malaysia. Cupaniopsis aneityensis Guillaumin, sp. nov. Arbor, ramis lepidotis, foliis 5-foliatis, jugis disjunctis, petiolo 3-4 cm. longo, foliolis lanceolatis (usque ad 8 cm. x 3 cm.) apice attenuatis basi asymmetrice cuneatis coriaceis vel chartaceis, nervis lateralibus 7-11-jugis, tenuibus, venis subtus densissime reticulatis, petiolulo 6-10 mm. longo; thyrsi ad apicem axillares, usque ad 20 cm. longi, dense lepidoti, densiflori, alabastris globosis dense lepidotis, floribus 2 mm. longis pedicello aequilongo suffultis, bracteis triangularibus minimis, sepalis 5 ovato-triangularibus extra lepidotis intus glabris, petalis mini- mis sepalis brevioribus cuneatis apice leviter emarginatis ciliolatisque, disco 5-glanduloso glabro, staminibus 2 mm. longis, filamentis dimidio inferiore pilosis superiore glaberrimis, antheris ovatis, germinis rudi- mento tereti apice rufo-penicillato. Aneityum: Anelgauhat Bay, common in rain-forest, alt. 275 m., no. 827 pro parte (Typus) Feb. 28, 1929 (medium-sized tree 12 m. high); same locality, common in rain-forest, alt. 100 m., no. 742 Feb. 12, 1929 (large tree, 20 m. high). Harpullia arborea Radlkofer in Sitzb. Math.-Phys. Cl. Akad. Muench, xx. 278 (1890). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 150 m., no. 939, March 17, 1929 (large tree up to 15 m. high; petals white; calyx brown, hairy).—Already recorded from Tanna; also Samoa, Fiji and Solomon Islands and Malaysia. Elattostachys vitiensis Radlkofer in Sitzb. Math.-Phys. Cl. Akad. Muench. rx. 602 (1879). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 240 m., no. 827 in part, Feb. 28, 1929 (medium sized tree up to 12 m. high). Eromanga: Dillon Bay, common in rain-forest at sea level, no. 356, June 3, 1928 (large tree up to 25 m. high; petals red).—Ver- nacular name “Ne-el-lar-ru.” 1933] GUILLAUMIN, FLORA OF THE NEW HEBRIDES 57 ANACARDIACEAE Spondias dulcis Forster, Fl. Ins. Austr. Prodr. 34 (1786). Eromanga: Dillon Bay, common in rain-forest at sea level, no. 311, May 26, 1928 (large tree about 25 m. high; fruit used for jam- making).—Also Fiji, Samoa, Society and Bismarck Islands, New Guinea and Malaysia.—Vernacular name “Nevie.”’ Dracontomelum vitiense ? Engler in De Candolle, Monog. Phan. Iv. 253 (1883). T anna: Lenakel, common in rain-forest at sea level, no. 22, Feb. 21, 1928 (tree up to 12 m. high; leaves dark green). Eromanga: Dillon Bay, common in rain-forest at sea level, no. 244, May 14, 1928 (tall tree up to 20 m.; fruit brownish yellow when ripe).—Vernacular name “Nar-ah.” Dracontomelum vitiense Engl. has been already found on Efate. LEGUMINOSAE Ormocarpum cochinchinense (Lour.) Merrill in Philipp. Jour. Sci. v. 76 (1910). Ormocarpum sennoides De Candolle, Prodr. 11. 315 (1825). Aneityum: Anelgauhat Bay, common along sea shore at sea level, no. 929, March 17, 1929 (small tree up to 6 m. high; flowers white with purple veins).—Also in Australia (Queensland), New Cale- donia, Bismarck Islands, New Guinea and Malaysia. GYROCARPACEAE Gyrocarpus americanus Jacquin, Select. Stirp. Am. Hist. 362 (1788). Eromanga: Dillon Bay, common in rain-forest at sea level, no. 371, June 5, 1928 (large tree 25 m. high; wood soft, used for canoe- making).—Already found on the Torres Islands; also in New Cale- donia, Fiji and Society Islands, Australia (Queensland, North Aus- tralia) and Malaysia—Vernacular name ‘“‘Nep-bleb-le.” MYRTACEAE Eugenia sp. Aneityum: Anelgauhat Bay, common along sea shore at sea level, no. 787, Feb. 20, 1929 (small tree up to 6 m. high, stunted, gnarled; leaves affected by salt water; fruit dark deep red when ripe, 2-50 in a bunch). CORNACEAE Alangium vitiense (A. Gray) Baillon ex Harms, Engler & Prantl, Nat. Pflanzenfam. 111.-8. 262 (1898). 58 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Aneityum: Anelgauhat Bay, common in rain-forest, alt. 300 m., no. 770, Feb. 14, 1929 (large tree up to 12 m. high; leaves dark glossy green; flowers yellow, sweetly scented). The plant from Aneityum differs from that of the Fiji Islands and of the west coast of Australia in the much larger leaves and particularly in the filaments being nearly three times shorter than the anthers. RUBIACEAE Dolicholobium aneityense Guillaumin in Jour. Arnold Arb. xt. 2 (1932). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 240 m., no. 767, Feb. 14, 1929 (plant up to 1.8 m. high; flowers white).— Already reported from Aneityum. Psychotria sp. Banks Group: Vanua Lava Island, common in rain-forest, alt. 30 m., no, 418, June 5, 1928 (small tree up to 10 m., growing in shade of larger trees; fruit red when ripe). MYRSINACEAE Maesa banksiana Guillaumin, sp. nov. Arbor parva, 8 m. alta, ramis gracilibus rubris glaberrimis, foliis ovatis vel ovato-lanceolatis (circa 14 cm. x 7 cm.) membranaceis apice basique acutis integris discoloribus in sicco supra brunneis subtus viridibus, nervis 6-jugis brunneo-rubris haud vel vix prominentibus, petiolo 1.5-3 cm. longo. Inflorescentiae axillares, 2-2.5 cm. longae, race- mosae, erecto-patulae, glabrae, floribus subsessilibus, bracteis lanceo- latis minimis margine leviter erosis, bracteolis ovato-triangularibus margine leviter erosis 0.5 mm. longis receptaculum haud formantibus, sepalis 5 liberis triangularibus acutis margine leviter erosis 0.75 mm. longis inconspicue punctatis, petalis 5 fere usque ad medium connatis ovatis apice rotundatis fere 1 mm. longis haud punctatis nec lineolatis, staminibus petalorum tertiam supremam partem attingentibus, an- theris late ovatis filamentis vix longioribus, ovario ovoideo, stylo tereti, stigmate rotundato. Banks Group: Vanua Lava Island, common in rain-forest, alt. 300 m., no. 456, July 12, 1928 (small tree up to 8 m. high). In the key given by Mez (in Engler, Pflanzenr. 1v.-236, p. 23) this new species should be placed beside M. denticulata Mez of the Phil- ippine Islands. OLEACEAE Ligustrum neo-ebudicum Guillaumin, sp. nov. Arbor magna, 20 m. alta, ramis pallide fulvis rugose lenticellatis, 1933] GUILLAUMIN, FLORA OF THE NEW HEBRIDES 59 foliis ovatis (usque ad 10 cm. x 4 cm.) basi cuneatis apice longe saepeque falcatim acuminatis subcoriaceis, nervis 8-10-jugis tenuibus, venis immersis, petiolo circa 1.5 cm. longo. Paniculae ad apicem axillares, ad 5 cm. longae, minute puberulae, floribus pedicello brevi (0.5 mm.) basi bractea 1 apice 2 ad apicem minute puberulis munito, calyce dentibus 4 triangularibus acutis margine minutissime puberulis, petalis albis 1.5 mm. longis ovatis quarta inferiore parte connatis, staminibus filamentis brevissimis corollae tubo vix aequilongis, an- theris ovato-linearibus apice truncato rotundatis nigroque punctatis, ovario ovoideo minimo stylo aequilongo, stigmate capitato. Aneityum: Anelgauhat Bay, alt. 300 m., rain-forest, common, no. 762, Feb. 14, 1929 (large tree up to 18 m. high; petals white). Near L. novo-guineense Lingelsh. of New Guinea, but differs in the calyx and particularly in its stamens with very short filaments and not apiculate anthers. Linociera ? ramiflora Wallich, Cat. 2824 (1831), nomen.—De Can- dolle, Prodr. vitr. 297 (1844). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 300 m., no. 765, Feb. 14, 1929 (large tree up to 18 m. high; leaves dark glossy green; petals white); same locality, alt. 210 m., no. 823, Feb. 28, 1929 (leaves light yellow-green; petals pink-white). The plant of Aneityum differs from that of New Guinea, Queens- land and Malaysia in the petals which elongate after the opening of the flower and attain finally a length of 5 mm. before dropping. MYRISTICACEAE Myristica inutilis Richard apud A. Gray, Bot. U. S. Expl. Exped. 1. 34 (1854). Aneityum: Anelgauhat Bay, common in rain-forest, alt. 60 m., no. 757, Feb. 13, 1929 (large tree up to 18 m. high; leaves golden ae below; fruit eolden brown). Tanna: Lenakel, common in rich rain-forest soil, alt. 60 m., no. 44, Feb. 24, 1928 (tree 15 m. high, 25 cm. in diam.; leaves brown below). E fate: Undine Bay, com- mon in rain-forest, alt. 60 m., no. 220, April 27, 1928 (tree 12-15 m. high; leaves brown below; fruit brown).—Also in the Samoa and Tonga Islands. LAURACEAE Cryptocarya tannaensis Guillaumin, sp. nov. Arbor ultra 15 m. alta, cortice brunneo griseo-maculato, ramis pri- mum rufo-puberulis deinde glabris in sicco nigris, foliis ovatis (usque ad 10 cm. x 5 cm.) apice acutis basi cuneatis leviter rigidis, glaberri- mis, nervis lateralibus 5, venis densissime reticulatis, petiolo circa 1 60 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV cm. longo. Paniculae 2 cm. longae, axillares terminalesve, ramis pedi- cellisque rufo-puberulis; flores pedicello brevissimo (1 mm.) apice bracteolato, perianthii tubo campanulato lobis fere aequilongo extra puberulo intus glabro, lobis 6 ovatis extra puberulis intus densius pilosis, staminibus generis, filamentis breviter ut connectivo hirsutis, antheris ovato-triangularibus obtusis filamentis vix longioribus, stami- nodiis brevius stipitatis acutis, ovario elongato glabro, stylo tereti antherarum apicem attingente. T anna: Lenakel, common in rich rain-forest at sea-level, no. 18, Feb. 21, 1928 (tree up to 15 m. high; bark brown with round grey spots). Litsea ? sp. Tanna: Lenakel, rain-forest, not common, alt. 30 m., no. 140, March 8, 1928 (small tree about 7 m. high; flowers white). EUPHORBIACEAE, Glochidion sp. aff. G. manono Baill. of the Fiji and Society Islands. Aneityum: Anelgauhat Bay, common in the country, in poor red soil, alt. 15 m., no. 700, Feb. 4, 1929 (small tree up to 8 m. high; leaves light yellow-green; petals yellow). Baccaurea sp. Aneityum: Anelgauhat Bay, common in rain-forest, alt. 60 m., no. 752, Feb. 12, 1929 (large tree up to 12 m. high; fruit red when ripe). Baccaurea sp. Aneityum: Anelgauhat Bay, common in rain-forest, alt. 240 m., no. 772, Feb. 14, 1929 (large tree up to 15 m. high; leaves dark green). Baccaurea ? Tanna: Mt. Tokosh Meru, common in rain-forest at 600 m., no. 168, March 15, 1928 (tree about 15 m. high). Aleurites moluccana Willdenow, Sp. Pl. 1v. 590 (1805). Aneityum: Aname, common at sea side, alt. 15 m., no. 989 (coll. J. P. Wilson), Sept. 1929 (large tree 1.3 m. in diam.; fruit brown).—Already found on Efate; also New Caledonia, Loyalty, Fiji, Tonga, Samoa, Cook, Society, Marquesas, Gambier and Mariana Islands, New Zealand, Australia (Queensland), New Guinea, Malaysia and Hawaii.—Vernacular name “Inhatch.” 1933] GUILLAUMIN, FLORA OF THE NEW HEBRIDES 61 MORACEAE Paratrophis tahitensis (Bur.) Drake, Ill. Fl. Ins. Mar. Pacific. 296 (1892); Fl. Polyn. Frang. 193 (1892) .2 Aneityum: Anelgauhat Bay, common in rain-forest, alt. 210 m., no. 818, Feb. 28, 1929 (large tree up to 12 m. high; flowers white; ies boiled and eaten by natives). URTICACEAE Leucosyke corymbulosa Weddell in De Candolle, Prodr. xvr'. 235*° (1869) Aneityum: Anelgauhat Bay, common in rain-forest at sea level, no. 791, Feb. 20, 1929 (small tree up to 9 m. high; fruit white when ripe, with green seeds). UNDETERMINED NUMBERS The following numbers have not been identified owing to insufficient material: nos. 159, 166, 297, 379, 750, 807, 808, 821 and 841 (= 166). MUSEUM NATIONAL D’HISTOIRE NATURELLE, ARIS 1This name is usually attributed to Bentham & Hooker (Gen. Pl. m1. 364), but these authors have only identified the genus Uromorus Bur. with Paratrophis and did not make the ‘binomial combination. 62 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV A SUPPLEMENT TO C. T. WHITE, “LIGNEOUS PLANTS COLLECTED IN THE TERRITORY OF PAPUA (BRITISH NEW GUINEA) IN 1925-26 BY L. J. BRASS”? ALFRED REHDER PALMAE Determined by M. BuRRET Calamus nannostachys Burret in a Arnold Arb. x1. 264 (1931). Kuraudi, Eastern Division, no, 137 Areca paliuscainea ue Burret, |. c. 265. Ihu, Vailalo River, rain-forest, no. 921. Actinophloeus microcarpus Burret, 1. c. 267. Loloki River, in dumps on river bank, no. 1659. Actinophloeus linearis Burret, 1. c. 268. Lower Mori River, Eastern Division, rain-forest, no. 1566. MORACEAE Ficus Determined by V. S. SUMMERHAYES In my account? of the Figs collected in Papua by L. J. Brass on behalf of the Arnold Arboretum I inadvertently used for two new species of Ficus names which had previously been used for quite other species. It therefore becomes necessary to rename the Papuan species; these are as follows: Ficus aechmophylla Summerhayes, nom. no etc Brassii Summerhayes in Jour. Arnold fe x. 146, 207 (1929) ; n R.Br. Loloki "River, no. 1660. Ficus skytinodermis Summerhayes, nom. nov. Ficus clusiaefolia Summerhayes 1. c. a 209; non Schott. Sogeri, foothill forest, alt. 500 m., no. 641 Ficus Decaisneana Miquel, F'1. ia Bat. I. pt. 1. 312 (1859). Loloki River, no. 549, and Bongwina River, Eastern Division, no. 1620; cited (1. c. 206) under Ficus subulata Bl Conocephalus sp. Lowa, Vailala River, alt. 60 m., on rain-forest trees, no. 1146, March 15, 1926 (climbing; attaching itself to trunks of trees by adventive roots; flowers axillary, pink). 1See Jour. Arnold Arb, x. 197-274 (1929) for the original paper. The deter- minations : ous supplement were mostly received from the Botanical Museum Berlin-Dah *Jour. nae Arb. x. 142-154 (1929). 1933] REHDER, PLANTS COLLECTED BY BRASS 63 MYRISTICACEAE Myristica sp. Bisiatabu, alt. 300 m., no. 610, Nov. 11, 1925 (compact tree, 10 m. high; bark channelled. LEGUMINOSAE Pithecolobium sp. aff. P. umbellatum Benth. Keuru, Gulf Division, sea beach, no. 1197, March 22, 1926 (small tree, 5 m. high; leaves dark bluish green; stamens pink.) Mucuna pruriens De Candolle, Prodr. 11. 405 (1825). Domara River, Eastern Division, no. 1594, June 31, 1926, (large climber; flowers purple; fruit silky-hairy). Pueraria novoguineensis Warburg in Bot. Jahrb. xm. 325 (1891). Thu, Vailala River, no. 957, Feb. 12, 1926 (climbing over bushes on river bank; flowers large, blue). CELASTRACEAE Siphonodon sp. Thu, Vailala River, rain-forest, no. 1074, Feb. 26, 1926 (small slen- der tree; fruit clusters on trunk). SAPINDACEAE! Determined by H. MEtcuior Allophylus micrococcus Radlkofer in Schumann & Lauterbach, FI. Deutsch. Schutzgeb. Siidsee Nachtr. 307 (1905). Thu, Vailala River, rain-forests, no. 950, Feb. 12, 1926 (slender tree, 12-15 m., with thin gray bark; fruit yellow); Hewa, Vailala River, rain-forests, no. 1132, March 13, 1926 (small tree with close gray bark; fruit bright red); Kuraudi, Eastern Division, on river bank, no. 1385, May 12, 1926 (small tree)—Common name “Odia” (no. 1132) and “Pidi-Pidia” (no. 1385).—See also Jour. Arnold Arb. x. 235. Alectryon ferrugineus Radlkofer in Act. Congr. Bot. Amsterd. 1877, p. 84 (1879). U-uma River, on river bank, Eastern Division, alt. 300 m., no. 1522, May 20, 1926 (spreading tree 10 m. high; seeds glossy black, partly enclosed in a pink arillus). Alectryon reticulatus Radlkofer in Sitzb. Math.-Phys. Cl. Akad. Muench. xx. 255 (1890). Bomgwina River, Eastern Division, no. 1618, June 2, 1926 (small river bank tree; flowers pink, in axillary panicles). Guioa rigidiuscula Radlkofer in Sitzb. Math.-Phys. Cl. Akad. Muench. rx. 614 (1879). 1See also Jour. Arnold Arb. x. 235 (1929). 64 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Bisiatabu, alt. 450 m., no. 594, Nov. 8, 1925 (slender tree, 10 m. high, with thin pale gray bark; fruit rufous) —Common name “Ete.” Guioa spec., aff. G. aryterifolia Radlkofer. Bisiatabu, edge of rain-forest, alt. 450 m., no. 614, Nov. 12, 1925 (tree 10 m. high, with thin gray bark; fruits brown in clustered panicles on main branches). Cupaniopsis spec. Iawarere, rain-forest, alt. 300 m., no. 706, Nov. 29, 1925 (slender unbranched tree, 3 m. high, with a crown of long spreading pinnate leaves; fruits orange-tinted in axillary racemes or panicles). —Common name “Sisimana.” Elattostachys tetraporandra Radlkofer in Sitzb. Math. Phys. Cl. Akad. Muench. xx. 267 (1890). Kapa Kapa, tidal flats, no. 795, Dec. 7, 1925 (spreading tree to 7.5 m. high, with brown close bark); Rigo, coastal brush, no. 815, Dec. 9, 1925 (spreading tree 7.5 m, high, with rough gray bark). Elattostachys spec. Domara River, rain-forests, Eastern Division, no. 1601, May 31, 1926 (tree 9-12 m. high; fruit red, seeds black). Lepidopetalum hebecladum Radlkofer in Sitzb. Math.-Phys. Ci. Akad. Muench. xx. 269 (1890). Mowabula, Eastern Division, no. 1871, May 10, 1926 (small river- bank tree; fruit red). Dodonaea viscosa Jacquin, Enum. Pl. Carib. 19 (1760). Bomgwina, Eastern Division, no. 1612, June 1, 1926 (small beach tree).—See also Jour. Arnold Arb. x. 236. Harpullia camptoneura Radlkofer in Sitzb. Math.-Phys. Cl. Akad. Muench. xx. 360 (1890). Kapa Kapa, tidal flats, no. 806, Dec. 8, 1925 (erect tree, 3 m. high; fruit yellow). arpullia cupanioides Roxburgh, Hort. Beng. 86 (1814), nomen, Fl. Ind. L. 645 (1832). Kerema, Gulf Division, rain-forests, no. 1215, March 24, 1927 (slen- der tree 3-4.5 m. high; flowers urceolate, white, in axillary panicles) ; Kuraudi, Eastern Division, common in riverine rain-forests over a large area of the south coast, no. 1894, May 12, 1926 (large bush or sparsely branched slender tree; flowers greenish in axillary or supra-axillary panicles); U-uma River headwaters, on river bottom, alt. 300-450 m., Eastern Division, no. 1450, May 18, 1926 (large sparsely branched bush or small tree, 2.5 m. high; flowers white, sweet-scented in axillary panicles); lower Mori River, Eastern Division, riverine rain-forests, no. 1558, May 28, 1926 (slender small tree or bush 3-4.5 m. high; 1933] REHDER, PLANTS COLLECTED BY BRASS 65 flowers greenish white; fruit scarlet, seeds yellow).—Common name “Gari” (no. 1558). Alphitonia moluccana Teysmann & Binnendijk, Cat. Hort. Bogor, 221 (1866), nomen.—Braid in Kew Bull. Misc. Inform. 1925, p. 184. Alphitonia eed Gare C. T. White in Jour. Arnold Arb. x. 236 (1929) ; not y. Nos. 1497 and 1687 (see White, I. c.). According to K. W. Braid in his Revision of the genus Alphitonia (1. c.), A. moluccana is not identical with A. zizyphoides which is re- stricted to Polynesia. O 5 ey RUBIACEAE Psychotria polyneura Val. vel aff. Thu, Vailala River, rain-forest, no. 962, Feb. 13, 1926 (small soft- wooded bush, leaves somewhat fleshy). COMPOSITAE Microglossa volubilis De Candolle, Prodr. v. 320 (1836). Uniori, alt. 250-300 m., light rain-forest, no. 739, Dec. 2, 1925 (soft rambling under shrub). APPENDIX LIST OF HERBACEOUS: PLANTS COLLECTED IN NEW GUINEA BY L. J. BRASS CYPERACEAE Remirea maritima Aubl. Mapania macrocephala (Gaudich.) No. 1614 K. Schum. No. 1158 ARACEAE Rhaphidophera novo-guineensis Schismatoglottis calyptrata Engl. oxb.) Zoll. & Moritzi No. 1034 (juvenile stage). No. 916 Cyrtosperma Merkusii (Hassk.) Alocasia lancifolia Engl. Schott No. 913 No. 931 COMMELINACEAE Pollia macrophylla Benth. Forrestia hispida Less. & Rich. No. 652 No. 1124 Aneilema sp. No. 999 LILIACEAE Dianella ensifolia Red. No. 1430 66 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV AMARYLLIDACEAE Crinum macranthum Engl. No. 1157 TACCACEAE Tacca pinnatifida Forst. No. 1169 DIOSCOREACEAE Dioscorea bulbifera L. Dioscorea tiliifolia Kunth o. 1030 No. 867 MUSACEAE Musa sp. No. 1364 ZINGJBERACEAE Curcuma longa L. Costus speciosus (Koenig) Smith No. 548 No. 922 Tapeinochilus sp. Alpinia novae-pommeraniae K. No. 608 Schum. No. 1373 MARANTACEAE Donax canniformis (lorst.) Rolfe Phrynium pedunculatum Warb. No. 926 No. 564, 960 in part Phrynium macrocephalum kK. Schum. No. 960 in part ORCHIDACEAE Agnostophyllum superpositum Oberonia diura Schlechter Schlechter No. 1523 (not no. 1216 as cited in No. 1493 Jour. Arnold Arb. x. 273). Appendicula cleistogama Schlechter AMARANTACEAE Deeringia amaranthoides (Lam. ) No. 869 CERATOPHYLLACEAE Ceratophyllum sp. an C. demersum No. 1621 BALSAMINACEAE Impatiens Hawkeri W. Bull. No. 554 BEGONIACEAE Begonia sp. No. 919 1933] REHDER, PLANTS COLLECTED BY BRASS 67 CONVOLVULACEAE Calonyction bona-nox (L.) Boj. Ipomoea sp. No. 857 No. 10 eae ics ( Desv. ) Jaquemontia paniculata ( Burm.) Hallier f Nos. 1029, 1639 No. 1404 LABIATAE Dysophylla auricularia BI. Orthosiphon stamineus BI. 1122 No. 833. Coleus scutellarioides Benth. No. 767 SOLANACEAE Solanum sp. No. 972 SCROPHULARIACEAE Limnophila rugosa (Rostk.) Ilysanthes veronicifolia (Retz. ) Schlechter Urb. No. 1651 No. 1014 GESNERIACEAE Cyrtandra sp. aff. C. bracteata arb. No. 1453 ACANTHACEAE Strobilanthes sp. aff. S. novo- megapolitanus Lindau No. 1082 RUBIACEAE Nertera depressa Banks & Sol. var. papuana No. 1 COMPOSITAE Vernonia cinerea (L.) Less. Blumea pubigera (L.) Merr. No. 75 No. 1363 Vernonia lanceolata (Warb.) Moq. KEclipta alba (L.) Hassk. No. 740 No. 893 sa eae viscosum Forst. Wedelia biflora (L.) DC. . 1036 Nos. 719, 958, 1382 Blumea Lessingii Merr. Wedelia spilanthoides F. v. Muell. No. 756 No. 775. HERBARIUM, ARNOLD ARBORETUM HARVARD UNIVERSITY 68 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV INHERITANCE IN AN OAK SPECIES HYBRID S. H. YARNELL With eight text figures A report of the first generation of the cross Quercus virginiana (Live Oak) & Q. lyrata (Overcup Oak)* was made by the late Helge Ness (1918) who, a few years later, was successful in obtaining three second generation families totaling twenty-three plants. While Professor Ness raised this second generation to fruiting maturity, he never made a second report. This paper presents a study of segregation of charac- ters in the above F,. The original cross was made in 1909 using the Live Oak as female; this was duplicated the following year. This species cross is easily made in spite of the fact that the parents belong to extremes of the section Lepidobalanus. Two of the four F, trees planted in front of the Station buildings are still growing vigorously and correspond very closely to the early description. They show definite evidence of hybrid vigor. They are now 13.4 and 14.8 inches in diameter at four and one-half feet above the ground. Eleven Live Oaks planted three years later average only 5.8 inches in diameter at the same height. Twenty Overcup Oaks planted with the Live Oaks average 5.97 inches. While the F, hybrids are perhaps a little more advantageously situated than the others, their diameter, which is more than twice that of plants of either parental species, can not be accounted for either by their situa- tion or by their three-year start. All F, plants of this cross were especially vigorous growers as seedlings. The 23 second generation hybrids belong to three families as follows: five trees from F, parent No. 1, ten from parent No. 2 and nine plants from parent No. 5. F, parent No. 1 resulted from the 1909 cross; par- ents No. 2 and No. 5 from the later cross. Since the F, plants are quite uniform and since the numbers in the second generation are small, all F, plants have been considered together as a single family. Seed to produce the second generation was collected in 1919, the trees being set out in 1923, spaced ten feet apart each way. Live Oak (Quercus virginiana) and the Overcup Oak have been found in several e€ es along the Gulf coast. Specimens from a number of trees found in the vi- cinity of Natchez, Mississippi, by Miss C. C. Compton were sent to the Arnold Arboretum in 1915 and 1916, and Professor Sargent published a description of them under the name Quercus Comptonae in the Botanical Gazette, Lxv. 456-458 (1918). The earliest discovery of the tree, however, seems to have been by Dr. Chas. H. 1933] YARNELL, INHERITANCE IN AN OAK HYBRID 69 LEAF CHARACTERS The leaves of the Overcup Oak are approximately twice as long as those of the Live Oak. While the leaves of the F, plants are large, they are smaller than those of the male parent. A random sample of 50 mature leaves were secured from each second generation tree and from representatives of the parental species. Measurements were taken, in centimeters, of the length and greatest width. The means of these two dimensions were added to serve as an arbitrary index of leaf-size. In relation to leaf-area this method exaggerates the size of the /yrata-type leaf. The results are presented in Figure 1. The F, plants fall into Ls | —— ~~. vy bf NUMBER OF INDIVIDUALS 5 10 15 LEAF INDEX IN CENTIMETERS *IGURE 1. SEGREGATION OF LEAF SIZE IN THE F,. The indices of the parental eae are marked by arrows. two distinct groups, one ranging around the virginiana parent, the other in a distribution below the /yrata parent. The curve of the first group shows a decided tendency to flatten out, the other group forms approxi- mately a straight line. The two groups are about equal in size, there being 12 individuals in the first and 11 in the second. This occurrence oe along pier oe Prange ke Co., Texas. A specimen of ca cna ay , Dec. 1880, n the he rbarium of a Arnold Arbor Nee nes this cimen also. pies the notation: “A large low wide- prank ra Me Rear folia age.” The foliage is ce so “small- Teal” group of Dr. Yarnell. he tree seems to have disappeared from this locality, or at least a diligent search a it by the writer in 1916 failed to find any trace of it. It is not ater on Dr. Mohr’s label whether the tree found by him was native or plant cam t in the absence of such notation, it may be assumed that it was yild. Wit no ragnideita of this and of two seedlings found in the woods near Natchez, ait of the trees known seem to have been planted either in parks, as street trees, or about dwellings vole plantations. But no pa s were obtained as to where they came from, ex in the case of some trees in New Orleans, which were reporte : hag have a brought from across Lake Pontchartrain 30 or years previou Miss s Co ABE a as stated by eae pert some 20 or 30 trees of this hybrid pan g in the vicinity of Na ee he sp ripe in the herbarium here show it o have bene ollected near Selm Alab cana ng the streets snd in Audubon Park New Orleans. These je Sa er a bern variation in leaf and fruit m i wo similar to aes peri d by Dr. Yarnell for the Ness hybrid and its descendants. The er wide natural distribution would seem to indicate that the two species, in pan of their obvious morphological differences, hybridize readily when growing together ARNOLD heoeetes ERNEST J. PALMER. 70 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV of two distinct groups of approximately equal size has been interpreted as evidence that two principal factors are involved in size of leaf and that the distribution obtained represents a 9:7 ratio, the lack of agree- ment being due to small numbers. Since segregation is taking place in the second generation of a rather wide interspecific cross, it is reasonable to assume that a number of factors from one species influence the expression of characters normally found in the other. Such an influence is evident in the expression of each of the characters under study, but does not, it is believed, invali- 2 3 4 5 6 7 8 Figures 2-8. PARENTAL, F, AND F,, LEAF SHAPE Types, X 3. 2. QUERCUS VIRGINIANA. 3. HE F,. 4. Q. LYRATA. 5-8. F, PLants. date an estimate of the number of principal factors involved. This, together with a lack of complete dominance, accounts for the amount of variation found in the second generation. Perhaps the most difficult classification to make, on this account, is that of leaf shape. The male parental species derives its name from the lyrate character of its foliage. The leaf of the female parent, while for the most part entire, does show a tendency toward the production of slight irregularities on some of the leaves, producing a type of margin which can not be distinguished, except in degree, from certain of the F, segregates. The virginiana type may be seen in Figure 2, the /yrata type in Figure 4. The F, is outlined in Figure 3. The figures have been traced from representative leaves and have been reduced % in reproduction. Figures 5 and 8 illustrate extreme types in the F,. The leaf in Figure 5 is the closest approach to the /yrata type. The leaf in Figure 8 is virginiana-like except for the apex which is pointed. There is one such plant. Three others have leaves very close to those of the 1933] YARNELL, INHERITANCE IN AN OAK HYBRID 71 female parent, with the addition of small lobes. Seven others are con- sidered nearer this type. All of these are in the “small-leaf” group dis- cussed above. The leaf in Figure 6 illustrates the only plant in the “small-leaf”’ group which is distinctly lyrate in shape. All trees in the “large-leaf” group have lyrate leaves. The exact form in some cases varies rather widely from the /yrata type as may be seen in Figure 7. This is presumably due to this influence of incidental factors from the Live Oak parent. The fact that one recessive was recovered in the F, and that all such plants can be placed in two general categories is good evidence that not more than two principal factors are involved in deter- mining leaf shape. Since the grouping with respect to size and shape of leaves coincide with one exception, the relationship of the two groups of factors can not be determined. A number of things might be responsible, such as identity or close linkage of factors, gametic selection, or inviability of certain gametic or zygotic combinations. The epidermis of the Live Oak is somewhat thickened, giving the leaves a shiny appearance, in contrast to the duller and thinner leaves of the Overcup Oak. The coriaceous character is not completely dominant in the F,, heterozygous plants being classified as modified coriaceous. The leaves of all plants in the “small-leaf” group were coriaceous. In addition, the leaves of three trees in the “large-leaf” group were given the same classification. This gives a total of 15 cori- aceous and 8 non-coriaceous plants. This is a little nearer the 9:7 than to the 3:1 expectancy (13:10 and 17.25:5.75 respectively). A related character is the habit of dropping or retaining the leaves at the end of the summer season. The name “Live Oak” is due to the evergreen character of Q. virginiana. The pollen parent, on the other hand, sheds its leaves relatively early. The first generation hybrid holds only a part of its leaves during the winter. All of the trees with large, non-coriaceous leaves are completely deciduous. In addition, two of the three trees with large modified coriaceous leaves were fully deciduous, while the third was classified as deciduous with a tendency to hold a part of its leaves later in the season. Five of the twelve trees with “small” coriaceous leaves were completely evergreen, the re- mainder being partially evergreen. It is evident that not more than two principal factors are involved here and quite possibly only one pair. FRUIT CHARACTERS Fruits of the two parental species differ widely in appearance, espe- cially in regard to size, shape and character of the cupule. Twelve of the twenty-three second generation trees fruited during the 1931 season. Fruits of Q. virginiana are approximately twice their diameter in 72 JOURNAL OF THE ARNOLD ARBORETUM [VoL, XIV length while those of Q. /yrata are somewhat wider than long. In the F, the fruits are wider in proportion to length than those of the female parent. Acorns of eight trees are distinctly long, of two are interme- diate, and of two decidedly round, giving a total of 8 long to 4 round. Inheritance of fruit shape seems to be similar to that of leaf shape with the former giving more nearly complete dominance. The fact that no fruits precisely like those of the /yrata parent have been found lends support to a two-factor hypothesis. There is a striking difference between the two species in size of fruit, the larger diameter of /yrata fruit giving the greater volume. Acorns of the first generation usually exceed those of either parent in size since they are somewhat longer than those of virginiana and nearly as wide as in lyrata. In every case in the F, the fruits have been classified as small and in three individuals the fruit is considerably smaller than in the virginiana parent. How much of this may be due to irregularities of development resulting from the interspecific nature of the cross can not be said. It is clear that the results can not be explained on a simple factorial basis. The Overcup Oak is so named because the cupule covers from three- fourths to nearly the entire nut. In the Live Oak fully three-fourths of the acorn is exposed. The cupule of the first generation is inter- mediate in size, covering from one-third to one-half of the large acorn. In nine F,, plants the cupule is about as found in the female parent. In two trees the cupule is intermediate in size, and in one the cupule approaches but does not equal that of the pollen parent in size. While two factors are indicated in this case it is evident that the results are complicated by other factors such as size of fruit. Thickness of cupule scales is a related character. While the scales are nearly equal in number in the two species they are much enlarged in the Overcup Oak and have a protuberance at the tip. The cupule scales in the F, may be said to be intermediate in size, although they are much closer to those of the Live Oak parent. In the F, eight individuals have thin scales and five have scales intermediate in thick- ness. None has scales approaching those of the Overcup Oak in prom- inence. Since this character shows such a close relationship with the preceding it is likely that the two have one factor in common. BARK CHARACTER It has been possible to obtain data in regard to one other character, type of bark. In Q. lyrata the bark flakes off, while in Q. virginiana the bark of the trunk is rough. The first generation trees are inter- mediate in this respect. In the second generation 7 trees have bark exhibiting different degrees of flakiness while 16 trees have a rough 1933] YARNELL, INHERITANCE IN AN OAK HYBRID 73 trunk. This ratio is probably the result of segregation of a single factor pair. This character, in common with the others under consideration, exhibits considerable variability in the F,,. DISCUSSION Augustine Henry (1910) studied large numbers of seedlings (971) of what was very likely the second generation of a natural cross be- tween Ulmus foliacea and U. glabra. The F, had received the name U. vegeta, known popularly as the Huntingdon Elm. Seedlings of both parental species were quite uniform. Three characters, opposite vs. alternate leaves, leaf size, and length of petiole, were found to segregate on a 3:1 basis in the second generation. Different F, trees were found to vary greatly in degree of fertility. Plate 24 of the paper cited above illustrates the Lucombe Oak, par- ents and seedlings. The leaf size and shape combinations in the F, are very similar to those of the Live & Overcup Oak cross. The Lucombe is evidently a seedling of a Q. cerris (Turkey Oak) & Q. suber (Cork Oak) cross. It is subevergreen and was propagated by grafting at Exeter, England in 1763. In 1792 acorns of this Oak were sown which gave a variable progeny (Henry 1910). The author recognizes hybrid vigor in the first generation of these and other tree crosses and the resulting increase in usefulness, and later made numerous controlled crosses with this end in view. H. A. Allard (1932) has made a study of a number of natural Oak hybrids growing in or near the District of Columbia. Open pollinated seed of Q. Saulii, considered to be a hybrid between Q. alba and Q. montana, were planted about 1922. Forty seedlings grew, most of which survived. Characters of Q. montana are more evident in the “F,” than those of Q. alba. The author points out that the generation involved is necessarily unknown. No factorial analysis was attempted. He reports that D. T. McDougal (1907) raised 55 seedlings of Bar- tram’s Oak (Q. heterophylia), a natural hybrid between Q. phellos and Q. borealis maxima. Plants similar to each parent and various types of intermediates were obtained. The first generation of the Live and Overcup Oak cross shares with other interspecific tree crosses the characteristic of rapid growth. In habit of growth it is intermediate between its two parents, being more erect than the Live Oak and having a greater spread than the Overcup parent. As a result, it is much more attractive as an ornamental than the latter species. Growth of the second generation trees varies widely, from somewhat better than the parental species to individuals that are decidedly stunted. 74 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV The tendency for characters of one parent to be inherited together in the F, is striking. It is a problem of importance to the plant breeder who makes use of interspecific crosses. Normal chromosomal linkage would account in some measure for this result, especially since it is likely that there is less crossing over in the F, than in either pure spe- cies. The partial sterility found in the second generation is good evi- dence that the chromosomal complements of the two species are sufficiently differentiated to prevent their normal functioning in certain combinations. It is but a step to assume that other combinations are unable to produce a viable individual. The chance for securing the de- sired grouping of characters under such conditions will depend upon the size of F, or back cross populations, which present an opportunity for viable chromosome combinations carrying the desired characters. The results suggest, for example, that the evergreen character might be combined with hardiness to extend the range of the Live Oak type considerably north of Virginia. This would involve a number of genes from each species such as those for thickened epidermis, leaf-size, and also for physiological characters. Such a successful combination, while not common in the F, would be expected to occur frequently enough to justify a systematic search for such individuals. There is ample evidence that the normal expression of a character is frequently modified by new combinations of genes from the two species. The idea that a gene may influence the expression of characters which are determined primarily by other genes has been frequently advanced. This idea usually carries with it the assumption that either allelomorph has the same incidental effect regardless of differences in the morpho- logical expression of the allelomorphic pair. The fertile species cross with its high variability of character expression in the second genera- tion bears out the above assumptions, in that what are presumably allelomorphic genes in the two species seem to have different inci- dental effects in contrast to the situation for an allelomorphic pair within a single species. The low number of principal factors involved for many characters increases the chance of securing a satisfactory combination in the second generation. Inheritance in forest trees thus appears to be no more complex than inheritance in most material of interest to the plant breeder. Division OF HORTICULTURE, Texas AGRICULTURAL EXPERIMENT STATION. 1933] YARNELL, INHERITANCE IN AN OAK HYBRID 75 LITERATURE CITED ALLARD, rar A. (1932). A progeny study of the so-called oak species Quercus Saulii, with notes on other probable hybrids found in or near the District of Columbia. (Bull. Torrey Bot. Club. 59:267-277. Henry, AuGusTINE (1910). On Elm-seedlings showing Mendelian re- sults. (Jour. Linn. Soc. 39:290-300. Pl. 2 McDoveat, D. T. (1907). Hybridization of wild plants. (Bot. Gaz. 43 :45-58. Ness, Hetce (1918). Hybrids of the Live Oak and Overcup Oak. 8. 3 pl.) (Jour. Hered. 9:263-26 76 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XIV TAXONOMIC AND CYTOLOGICAL RELATIONSHIPS OF YUCCA AND AGAVE Susan DELANO McKELVEY AND KARL SAX With plate 55 The senior author, during the past two years, has made an extensive study of Yucca and related genera. These genera have been collected from western Texas, New Mexico, Arizona, and the southern areas in California, Nevada, Utah and Colorado. The distance covered in this area has approximated 25,000 miles, and no day of travel passed with- out encountering numerous species of these plants. Identification of Yucca species in the field has been difficult because of the extreme variability of all species, and the overlapping range of two or more similar species. For example, in the group of filiferous- leaved Yuccas with dehiscent fruit, the leaf characters have at times been used as a basis for species differentiation, but as a practical guide in the field these characters are often unreliable and confusing. On a single inflorescence the flowers vary little in size and form, but two inflorescences from adjoining plants, apparently of the same species, may differ in habit; the flowers may differ in shape and size of corolla-segments, ovary and style; and vary in length, form and pubescence of filaments. Among the fleshy-fruited Yuccas, one species seems to have two distinct forms; one with a vigorous fleshy in- florescence and large flowers, and the other with a compact, more ligneous inflorescence and more numerous small flowers. In regions where closely allied species occur, their separation in the field is extremely difficult. Within the range attributed to Y. angus- tissima are plants of vigorous growth quite different than the expected type. In some respects, they resemble Y. glauca whose range is sup- posed to begin hundreds of miles further east. At the point of worst confusion has one encountered intermediate segregates from species hybrids, or are different types variants of a single species? In most respects the allied genera Clistoyucca, Hesperoyucca and Samuela, separated by Trelease (1902), are similar to Yucca. Yucca and Agave are also similar in various striking details although these genera are placed in different families by most taxonomists. The occurrence of transitional forms of Yucca species and the simi- larity of closely, and even distantly related genera, has led to a cyto- logical study of these plants in order to compare their chromosome num- er and morphology. Material was collected in the field by the senior 1933] McKELVEY AND SAX, YUCCA AND AGAVE 77 author. Flower buds at different stages of development were fixed in a mixture of acetic acid and alcohol, and later transferred to 70 percent alcohol. Chromosome counts were obtained from aceto-carmine smears by the junior author. Permanent smears were also made from pollen mother cells of Agave americana collected at the Bermuda Biological Station, from Agave virginica from the Missouri Botanical Garden,,and from Yucca flaccida and Y. filamentosa grown in the Arnold Arboretum The chromosomes of Yucca flaccida have been described in detail by O’Mara (1931). At the first meiotic division there are 5 pairs of long chromosomes and 25 pairs of very small chromosomes. The long chromosomes have an average chiasma frequency of 3 per bivalent at metaphase, while each small bivalent has only a single terminal chi- asma. The pairing is very regular and lagging univalents or other abnormalities were not observed. The extreme differences in chromo- some size and the large number of chromosomes is rather unusual in the plant kingdom. All of the species of Yucca examined have the same chromosome con- stitution. These include Y. flaccida, Y. filamentosa, Y. elata, Y. con- stricta, Y. rupicola, Y. macrocarpa, and Y. angustissima. The chro- mosomes of Y. flaccida and Y. filamentosa are shown at different stages of meiosis (figures 1, 2, and 3). The five pairs of large chromosomes are conspicuous at all stages. There is considerable variation in the size of the 25 small chromosomes, but all are relatively small. The closely related genera examined also have exactly the same chromosome constitution as Yucca. These include Hesperoyucca Whipplei, Hesperaloe parviflora, and Samuela Faxoniana. Unfor- tunately, favorable material was not obtained from the related gen- era Nolina and Dasylirion. Counts from somatic cells of Nolina show that there are about 38 chromosomes which differ consid- erably in size, and most, if not all, of those chromosomes have median or sub-median spindle fiber constrictions. The more distantly related genus Dracaena (D. arborea) also has about 38 somatic chro- mosomes. The species of Agave studied have 5 pairs of large chromosomes and 25 pairs of small chromosomes. The chromosomes of A. virginica and of A. americana are shown at different stages of the first meiotic divi- sion (figures 4, 5, and 6). Aceto-carmine smears of Agave consociata also show the same chromosome number and morphology. According to Mr. S. Horovitz of the University of Buenos Aires, Agave filifera also has 5 large and 25 small pairs of chromosomes. The Agavoideae include according to Engler and Prantl the genera 78 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. XIV Agave, Polianthes, Bravoa, Furcraea (Fourcroya) and Beschorneria, all natives of central America, and the Australian genus Doryanthes. According to Heitz (1926) Furcraea altissima has 10 large somatic chromosomes and about 40 small ones while F. Lindeni has 10 large chromosomes but only 30 small ones. We have found 10 large chro- mosomes and 50 small ones in. aceto-carmine preparations of root tips from Furcraea Bedinghousii, but in order to get satisfactory prepara- tions it was necessary to isolate single dividing cells and flatten the metaphase chromosomes by heat and slight pressure. According to Heitz, Miiller’s work on Beschorneria shows chromosome numbers simi- lar to those of Furcraea. As a rule pollen grain size is generally correlated with chromosome number in closely related species. The size and morphology of the pollen grains appear to be very similar in the following species ;— Yucca macrocarpa, Y. Treculeana, Y. baccata, Y. elata, Y. Reverchoni, Y. Thompsoniana, Y. mohavensis, Clistoyucca brevifolia, Hes peraloe parviflora, Hesperoyucca Whipplei, Agave Havardiana and several un- identified species or varieties of Yucca and Agave. Each species and variety examined had almost 100 percent of morphologically perfect pollen grains. DISCUSSION The confusing variation within species and the intermediate forms of Yucca might be interpreted as the result of extensive hybridization between species and varieties. All of the species examined show regu- lar chromosome behavior at meiosis and practically all of the pollen is morphologically perfect. If the questionable forms are hybrids the parental varieties or species must be closely related and possess very similar genoms. In respect to chromosome number and chromosome morphology all of the species studied seem to be very similar. If the chromosomes of different species are compatible with each other, a considerable amount of crossing might be expected and would be lim- ited only by geographic isolation and differences in time of flowering. The cytological analysis alone offers little help in solving the cause of variations within and between species of Yucca. The striking similarity of the chromosomes of Yucca and Agave and their allied genera indicates a close relationship between these two groups, even though they are referred to different families. The chro- mosome constitution, 5 pairs of large chromosomes and 25 pairs of small chromosomes, is so unusual that it cannot be due to chance. The two genera are also similar in many taxonomic characters. A brief comparison of the two genera, and of the families to which they 1933] McKELVEY AND SAX, YUCCA AND AGAVE 79 belong, has been prepared by Dr. Ivan M. Johnston of the Arnold Arboretum, and is quoted below. “Yucca belongs to the Liliaceae and Agave to the Amaryllidaceae. The Amaryllidaceae, having an inferior rather than a superior ovary, are evidently more specialized in basic floral structure. Taken as a whole the two families differ only in this character of the ovary. There is every evidence that the Amaryllidaceae have been derived from the Liliaceae and there are some very good reasons to suspect that the former are a polyphyletic group with the several points of origin in the Liliaceae. Yucca and Agave are similar in many striking details. Both are coarse fibrous perennials with usually firm, long-enduring, monocarpic leaf-rosettes that are not common in their families. The large panicles are similar in basic structure and pattern. Yucca has a superior ovary; the filaments are attached at the base of the corolla, bearing small firmly affixed anthers which do not surpass the corolla-lobes. Agave has an inferior ovary; the stamens are attached in the corolla throat, the linear filaments bear large linear versatile anthers usually protruding beyond the corolla-lobes. While Yucca is more simple than Agave in structure of ovary, and presumably belongs to a generally more primative family, its staminal structures and the complex symbiotic relation required in its pollina- tion are distinctly far in advance over conditions found in Agave. It is evident, therefore, that Agave could scarcely have been evolved from Yucca (including Samuela, Hesperoyucca, Cleistoyucca and Hespera- loe) as now constituted. Yucca is generally accepted as having relations in Nolina, Dasylirion and Dracaena of the Liliaceae. Furcraea, in the Amaryllidaceae, is a relative of Agave although it possesses many important details of habit and floral structure very similar to those found in the relatives of Yucca. If Yucca is a reasonably close relative of Agave, as I believe it is, then it is probable that the affinity is to be traced through the related genera mentioned.” In the Dracaenoideae only the Yucceae have 5 large and 25 small pairs of chromosomes. All of these chromosomes have terminal spindle fiber attachments. In the two species examined (Nolina sp. and Dracaena arborea) in the Nolineae and Dracaeneae the somatic chro- mosome number seems to be about 38. These chromosomes do show considerable size differences but not as extreme as in Yucca, and the spindle fiber attachment constrictions seem to be median. In the Agavoideae the two genera studied, Agave and Furcraea, have 80 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XIV chromosomes very similar to those of Yucca. The cytological com- parison shows a very close relationship between the Yucceae and cer- tain genera of the Agavoideae, but Nolina and Dracaena do not seem to be the connecting link between these groups. According to Engler and Prantl, all genera of the Yucceae, and with the exception of the Australian genus Doryanthes, all genera of the Agavoideae, are natives of Central America. These genera are closely related as indicated by both taxonomic and cytological characteristics. It is of interest to note that the African Aloinae, including the genera Aloe, Gasteria, Apicra and Haworthia, are in some respects similar to the Yucca-Agave group. All of these African genera have the same chromosome number, and the chromosome complex includes both large and small chromosomes. There are 4 pairs of large chromosomes and 3 small ones in each of these genera (Gaiser, 1930). The morpho- logical characters and the similarity in size differentiation of the chro- mosomes seems to indicate a remote affinity between the Aloinae of the Old World and the Yucca-Agave group of the New World. SUMMARY Yucca and Agave are similar in many taxonomic characters although one genus is placed in the Liliaceae and the other in the Amaryllida- ceae. Yucca and the closely allied genera Hesperoyucca, Hesperaloe, and Samuela, have 5 pairs of large chromosomes and 25 pairs of small chromosomes at the meiotic divisions. Exactly the same chromosomes constitution is found in Agave and in at least one species of the closely related genus Furcraea. The similarity in taxonomic characters and chromosome constitution indicates that these genera have had a com- mon origin and are closely related. The variability within and between species of Yucca is not correlated with any variation in chromosome number or irregularity in chromo- some behavior. HERBARIUM AND CyTOLOGICAL LABORATORY ARNOLD ARBORETUM, HARVARD UNIVERSITY LITERATURE CITED Gatser, L. O. (1930). Chromosome numbers in Angiosperms. (Bibl. Genetica, 6:171-466.) Heitz, E. (1926). Der Nachweis der Chromosomen. (Zeit. f. Bot. 7625-681. ) O'Mara, J. (1931.) Chromosome pairing in Yucca flaccida. (Cytologia, 3:66-— TRELEASE, W. (1902). The Yucceae. (Missouri Bot. Garden Ann. Rep. 13:27-133.) Jour. ARNOLD ArB., VoL. XIV. eS "> & e 4 y o ¢- rr ® a °* Whe 2 ? PLATE 55 t % XA: a & ot a S .° e° ‘ ‘f ’ we Sg : % @ 6 sams / TAXONOMIC AND CYTOLOGICAL RELATIONSHIPS OF YUCCA AND AGAVE. Tn 1933] McKELVEY AND SAX, YUCCA AND AGAVE 81 DESCRIPTION OF PLATE 55 e drawings were made from permanent smears of pollen mother cells. ec about 2500. 1. Yucca flaccida. Diakinesis. c AunbRwWh nN S Se s ° Q° Q ~ = ~~ 5 a o ry a. 3 o aS wn Agave virginica. Metaphase. Agave americana. Telophase. 82 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV CHROMOSOME NUMBERS IN ULMUS AND RELATED GENERA KARL SAX With plate 56 The genus Ulmus contains about 18 species and a considerable number of varieties and hybrids (Rehder 1927). The genus is repre- sented in the temperate regions of North America, Asia and Europe. Most species develop their pollen grains and flower early in the spring, but two species, U. parviflora, the Chinese Elm, and U. serotina, from south central United States, flower in early autumn. The other species have well developed flower buds in the fall but do not bloom until early spring. Chromosome counts at meiosis were obtained from aceto-carmine smear preparations. Fourteen pairs of chromosomes were found in the following species and varieties—the American species, U. racemosa and U. fulva; the European species, U. laevis, U. glabra, U. procera purpurea and U. foliacea; and the Asiatic Elms, U. laciniata and U. japonica. Ulmus hollandica varieties Pitteursii and superba also have 14 pairs of chromosomes, although there is occasionally some irregu- larity in the meiotic divisions of the variety superba. The American Elm, U. americana is a tetraploid with 28 pairs of chromosomes. There is, as might be expected, some indication of secondary chromo- some pairing in the species. The variety pendula is also a tetraploid. It is probable that U. americana is an autotetraploid and that only bivalent chromosomes are formed because of the limited chiasma frequency. The numerous varieties of U. Aollandica are presumably segregates from hybrids between U. glabra and U. foliacea. The two varieties studied show only slight irregularities in meiotic divisions and pollen fertility is almost complete. It seems probable that a number of species hybrids could be made in this genus. During the spring of 1932 crosses were made between U. americana and U. laevis. A considerable number of seeds were obtained, but owing to technical difficulties in pollination it is probable that many of the seeds were not of hybrid origin. Since U. americana is a tetra- ploid and U. laevis a diploid, the hybrid seedlings should be triploids. Chromosome counts from these seedlings have not been completed, but the seedlings examined seem to be tetraploids. In general, the species of Ulmus constitute a rather closely related 1933] SAX, CHROMOSOME NUMBERS IN ULMUS 83 group as indicated by chromosome number and morphology, the ocur- rence of fertile species hybrids, and the similarity of certain Old and New World species. The related genus Zelkova, also has 14 pairs of chromosomes. The reduction divisions in Z. serrata seem to be regular. The monotypic genus Hemiptelea is apparently a polyploid with a large number of chromosomes which are irregularly distributed at the meiotic divisions. About 70-80 per cent of the pollen of H. Davidii is morphologically imperfect. Celtis occidentalis also has very irregular meiotic divisions. The diploid chromosome number is 28, but at the first meiotic division there is little chromosome pairing and the univalent chromosomes are irregularly distributed to the daughter nuclei. In some cases 28 uni- valents were observed at the first meiotic metaphase stage and in most cases only a few bivalents were found (Fig. 11). The chromo- some behavior was the same in two different trees and in both cases at least 80 per cent of the pollen was sterile. A considerable number of fruits were produced on both trees. The irregular chromosome behavior and high pollen sterility in a good species is difficult to understand. If this species contains two different basic sets of chromosomes with only a few homologous chro- mosomes or parts of chromosomes it could breed true only by some form of apomixis. Such true breeding hybrids are found in nature. Malus theifera for example is a triploid which produces no functional pollen, but it reproduces the type and has a natural distribution. It is also possible that the prevalence of galls on Celtis may influ- ence the meiotic divisions. Kostoff and Kendall (1929, 1930) have found that gall mites and other parasites may cause chromosome irreg- ularity in certain solanaceous plants. Although no parasites were observed on the Celtis flowers collected it is perhaps possible that the residual effect of constant earlier infections may have had an influence on the meiotic divisions. If megasporogenesis is less easily disturbed a considerable number of fruits might be produced even though rela- tively few of the pollen grains are functional. Additional work on C. occidentalis and other species should be done to determine the cause of pollen sterility in this genus. The basic chromosome number is 14 for all of the genera of the Ulmaceae investigated. Among the allied families the basic chromo- some number is 16 for the Juglandaceae (Woodworth, 1930), 8 and 14 for Betulaceae, (Woodworth, 1931), 12 for the Fagaceae (Sax, 1930), and 14 for several genera in the Moraceae (Gaiser, 1930). From the standpoint of chromosome numbers the Ulmaceae might seem to be 84 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XIV allied with the Alnus,. Betula, Corylus group in the Betulaceae and Morus in the Moraceae, but the great diversity in morphological and anatomical characters would not seem to indicate any close phylo- genetic relationships among these families. CyToLoGIcAL LABORATORY, ARNOLD ARBORETUM HarvarpD UNIVERITY LITERATURE CITED Gatser, L. O. (1930). ba eaaaaas numbers in Angiosperms. II. (Bib bl. Genetica, 6:171-4 Sax, Hatry J. (1930). ome numbers in Quercus. (Jour. Arnold WoopwortH, R. H. seh . Meiosis of microsporogenesis in the Jug- landaceae. (Amer. Jour. Bot. 17:863-869.) —_—_——.. (1931). Polyploidy in the Betulaceae. (Jour. Arnold Arb. 12:206-217.) DESCRIPTION OF PLATE 56 Drawings from aceto-carmine preparations of pollen mother cells. X 2000. First meiotic metaphase shown in all figures except number 10 which is from a s ene —s plate. Zelkova serrata. Celtis tks hog Figure 1. Ulmus lae Figure 2. Ulmus ie farn pendula. Figure 3. Ulmus racemosa. Figure 4. Ulmus fulva. Figure 5. Ulmus laciniata Figure 6. Ulmus silo purpurea. Figure 7. Ulmus] Figure 8. Ulmus Ni hin Pitteursti. Figure 9. Ulmus Keine 0. ra = oe ory Lon 9 oO — Jour. ArNotp Arp., VoL. XIV. PLATE 56 ° o°2 r vd GgQ@ 6¢ eo © 2oses ae Pid [Se 2% ay \ é ®, & e$ y i‘ 3 3 @ e®@ “e | + ®e@ Oe. _@ dp SBS * qo , °é & ee @ 6 o® 2° ae 7 @e &¢ Phd . » @ \ @ a*s e ee a a eo. - "te 10 “se ce ee CHROMOSOME NUMBERS IN ULMUS AND RELATED GENERA. TT re 1933] HATCH AND DOAK, MYCORRHIZAL FEATURES 85 MYCORRHIZAL AND OTHER FEATURES OF THE ROOT SYSTEMS OF PINUS A. B. Hatcu anv K,. D. Doak With plates 57-60 The distinctive anatomical and morphological features of mycor- rhizal and non-mycorrhizal root systems, as illustrated by those of Pinus, are the subject matter of this paper. The discussion, although limited to Pinus, considers some features that are common to all ecto- trophic mycorrhizal organs of woody plants. The term ectotrophic mycorrhiza refers to a distinct morphological organ consisting of a “short root” and a fungus that are constantly arranged in an orderly manner with reference to each other. It is a well-defined organ and need not be confused even with its nearest rela- tive the pseudomycorrhiza. Frank (1885) coined the term ‘‘mycorhi- za”—a compound of Greek terms meaning “fungus root.” He described the structure as neither root nor fungus, but rather—analo- gous to the thallus of the lichen—an association of two differing components to form one morphological organ.* The use of the term mycorrhiza in reference to ectotrophic struc- tures (under which we also include the ectendotrophic forms for con- 1In his firs on mycorrhiza with reference to the ne — of ia Beech, etc., Fron pre that they consist of two kinds of elem are “a einem Kern, welcher die rapa Baumwurzel re repriiseritiert, und” a) us i mit jenem organisch verwachsenen Rinde, welche aus Pilzhyphen hc apres setzt ist.’ Der ganze Korper ist Ane weder Baumwurzel ae 4 Pilz allein, sondern ahnlich wie der Thallus der Flechten, eine Vereinigung zweier verschiedener Wesen zu einem einheitlichen morphologischen or bona eorameag passend als Pilzwurzel, Mycorhiza bezeichnet werden kann.” die k, 1885, The frequent subsequent use of the Bet mycorrhiza (or Pea orh eg to de- paren a condition rather than an organ is the res im “Of the absence of such a gan in Orchi Beg r that collectively have been converted into “ stile izae. oe In view of Fra original definition this is obviously incorrect. The word mycorrhiza is compou unded from the Greek wbxn, a rare variant of mixns “fungus,” and of plgfa “root,” with 66 (rrh) medially (as usual) an ai in the plural (since pita is a feminine ee of the first declension), which i botanical usage is latinized, becoming “ae. nek authors hire ea ee d the Blaval endings of their respective languages in their aot on ectotr mycorrhizae, but in English this is awkward phonetically. Kelley (931) atc pae adopted the latinized Greek plura ending, while retaining hs “r” of Frank. The spelling aged in this paper is the one considered eesti by authorities on the Greek languag 86 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV venience) is restricted by most European investigators to such an organ formed from short roots. Elsewhere, the term is frequently used to indicate indiscriminate fungal attacks on “long” as well as on “short roots.” McArdle (1932) for example has recently published photo- graphs (Figs. 1 to 6 and plate 2) of root systems of Pine and Spruce grown in open culture synthesis experiments to support his belief that certain fungi investigated form mycorrhizae with the roots of those trees. In figure 5 (ibid.) a dead radicle is illustrated from which three replacement radicles have emerged and grown geotropically. These are labeled “Mycorrhizae on Norway spruce, formed in synthesis with Lycoperdon gemmatum.” It is evident from this photograph and from the others mentioned above that long roots are considered to be mycor- rhizal organs. Neither these nor any of the visible short roots, however, exhibit the external characters possessed by a typical mycorrhiza. Since conversion of root tips into mycorrhizal structures is probably dependent on slow elongation rates it is improbable that long roots (rapidly growing roots) can be converted into “fungus roots” while the short roots (slow growing roots) of the system remain uninfected. Again, among 21 photographs and drawings of roots of culture plants reported to be mycorrhizal by Masui (1927, figs. 63-84) one drawing only (fig. 83) is included that illustrates a structure characteristically mycorrhizal (ectotrophic). Illustrations of non-mycorrhizal intracellu- lar infection, of cells evidently taken from long roots, are common (figs. 75 and 76). Further confusion in mycorrhizal studies has resulted from failure to qualify the use of the terms “large root” and “rootlet.” The latter evidently has been used to cover all elongating roots (McArdle, 1932). In connection with the former, Masui (1927, fig. 90) and others have used microchemical methods for studying the physiological functions of mycorrhizal roots and in so doing have apparently compared mycor- rhizal “short roots” with non-mycorrhizal “large” or “long roots.” The impropriety of comparing such roots microchemically may account for the diversity of opinion among these investigators concerning the physi- ology of mycorrhizal associations.* The profuse surface growth of non-mycorrhizal and probably para- sitic fungi on roots has frequently caused such roots to be mistaken for true mycorrhizal structures. Thus abundant and apparently parasitic IMcArdle (1932, p. 292) states that all a gs who have used this tech- nique agree that “mycorrhizal fungi are parasitic u n the associated higher pene i ioti However, Rexhause the relationship, and Weyland (1912) and Masui (1927). adie that parasitism exists. 1933] HATCH AND DOAK, MYCORRHIZAL FEATURES 87 attacks (with sporophore production) of certain -Gasteromycetes on mycorrhizal roots, roots already mycorrhizal through association with Hymenomycetes, have recently been observed in several instances by the authors (further communication by Doak). Recognition of the nature of such secondary attacks is often extremely difficult, a fact which undoubtedly accounts for much confusion in the literature. Attempts to learn of the abundance of the mycorrhizal roots of trees by reviewing the world’s literature of these structures are singularly futile. With few notable exceptions investigators have either neglected to define their use of the term “mycorrhiza,” failed to apply it ex- clusively (and quantitatively) to those roots known as short roots (absorbing roots, “Saugwirzelchen oder Ernahrungswurzeln’’) or omitted a tally and description of the roots they have not listed as mycorrhizal. In America, McDougal (1928) has reported repeatedly upon the existence of mycorrhizal structures on forest tree roots. One gathers from his papers that although they are widely distributed they are rela- tively nonabundant and in fact are somewhat infrequent if present at all on some species. For a number of years the authors have devoted their attention, almost exclusively, to the study of the mycorrhizal roots of forest trees both in America and Scandinavia. During this period they have rarely observed short roots (on forest trees over one year in age, growing in good forest soil with pH less than 5.5) that were not mycorrhizal. This is in line with the observations of numerous investi- gators, including Frank (1885) in Germany who recorded the invariable conversion of short roots into mycorrhizal roots, and also including Melin (1917, 1923, 1925 and 1927) in Sweden. Kelley (1932) re- ports similar phenomena for practically all woody plants he has exam- ined in the eastern and middle western United States during the past four years. The apparent lack of agreement among workers in this field arises in part from failure to distinguish the various root forms. The diag- nostic characteristics of the several types of roots occurring on Pine are discussed below. CLASSIFICATION OF PINE ROOT SYSTEMS Root systems of conifers exhibit several kinds of roots and are there- fore called “heterorhizic” by Noelle (1910). Two broad types are dis- tinguished: (1) long roots (‘“Bereicherungswurzeln”) and (2) short roots or absorbing roots (“Kurzwurzeln, Saugwirzelchen” or “Ernahrungswurzeln”). Anatomically these differ in Pinus as follows: 1. Long roots are furnished with root caps, while short roots lack 88 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV this structure, having a few layers only of cells beyond the plerome apex. (Plate 57, A and B.) 2. Long roots are diarch or polyarch, while short roots are monarch. 57, D.) The diameter quotient, that is, ratio of stele diameter to total ome of short roots is significantly lower than that of all other types of roots (Aldrich-Blake, 1930, p. 24). 4. Long roots exhibit secondary growth, while this condition has never been observed in short roots (Noelle, 1910, reports a questionable observation to the contrary). 5. Root hairs arise from the second or third layer of cortical cells in long roots (Plate 59, D; Noelle, 1910; Aldrich-Blake, 1930) and from the epidermal cells in short roots when the latter possess these struc- tures (Plate 57, B and D). 6. Long roots branch racemosely, while short roots branch dicho- tomously only. (Text fig. 1 and plate 58, C.) In addition to anatomical differences these roots have vastly differ- ent elongation rates. Finally, barring accidents, long roots, in nature, are permanent structures, while short roots are ephemeral. It should be mentioned that the monarch structure in uninfected short roots is demonstrated for the first time in this paper (Plate 57, D). Previous workers have apparently considered that the monarch structure resulted from infection by mycorrhizal fungi. Noelle (1910, p. 254), for example, writes—‘“Die unverpilzten Ernahrungswurzeln besitzen dagegen einen wohl ausgebildeten diarchen Protoxylemstrang, . .’ The materials Noelle studied were all from soils. It is prob- ails, therefore, that most of the short roots were converted into mycor- rhizal structures, and that Noelle’s “unverpilzten Ernahrungswurzeln” were in actuality arrested diarch lateral root initials. The latter are described in detail below in the section headed ‘‘Types of long roots.” The gross characteristics of long and short roots are illustrated in Text figure 1. The root system of a seedling grown in a mixture of sand and forest humus (the infektionsjord of Hesselman, 1927, p. 361; also Gast, unpublished) is shown. It has 24 long roots (totaling 65 cm. in length) in addition to a 25 cm. radicle or tap root. Including arrested laterals it had 435 short roots (quite a few of which were broken off in transit to America before the photograph was taken). Arrested laterals are to be noted along the radicle and on one upper- most side branch. A slow-growing lateral that was not attacked—as evidenced by absence of swelling and of dichotomy in the basal portion —by a mycorrhizal fungus until it had reached a length of about 1 cm. is marked by the digit “4.” The short roots are either mycorrhizal or 1933] HATCH AND DOAK, MYCORRHIZAL FEATURES 89 Oy Figure 1. Root System or Pinus. Entire root system of a three months old seedling of Pinus sylvestris L. gro in sand- Piel humus (infeksionsjord Joénaker, Sweden; unpub- ee cia of P. R. Gast’s, Harvard Forest); X 1.58. (1) Mother root, of which is mycorrhizal; (2) short roots attacked by M. R. ee ay ype of fungus; (3) dichotomously branched short roots ; (4) arrested lateral root initials 90 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV pseudomycorrhizal. Mycelium of the M ycelium radicis atrovirens type (Melin, 1923, p. 223) is visible on many of the short roots. The longest true short root on the system does not exceed 3 mm. in length. The comparative abundance of long and short roots on experi- mental seedlings is given in Table 1. These seedlings had an average of 19 long roots and 474 short roots per seedling. Thus 96 per cent of the growing root tips are short roots. This is of considerable sig- nificance when it is considered that ordinarily these are all mycor- rhizal, while the remaining 4 per cent—long root tips—on the other hand rarely acquire this structure. TYPES OF LONG ROOTS The term “long root” includes (a) the radicle or tap root (also a replacement radicle should the original radicle be injured), (b) diarch laterals called “mother roots” and (c) polyarch continuations of mother roots termed “pioneers.” Of these the radicle is the most strongly developed root in young seedlings. It is polyarch, has a large diameter quotient and a large and well-developed root cap. Lateral roots emerge from the radicle. These either abort (see below) or become mother roots. The latter are diarch and are characterized by a moderate elongation rate and by the production of numerous short roots and secondary laterals. At variable distances from their points of origin mother roots are converted into pioneer roots by augmenta- tion of the number of protoxylem strands. (In Scots Pine, however, all long roots remain diarch, Noelle, 1910). Pioneer roots are the most poorly-branched members of pine root systems, largely because of abortive secondary root initials (Aldrich-Blake, 1930). Pioneer roots elongate rapidly, have an intermediate diameter quotient, and a well-developed root cap. Collectively these roots markedly increase the spread of root systems. They likewise bear lateral and secondary roots in abundance (particularly true of mother roots). To what extent the growing tips of long roots, where the conduction systems are undifferentiated, function as absorbing organs is known by specula- tion only. The early development of secondary endodermis a short dis- tance back of the tips presumedly soon prevents absorption. Most workers have assumed that absorption is accomplished mainly by short roots. An experimental background for the assumption is lacking. Although a long root is capable of being converted into a mycorrhiza this condition is comparatively infrequent (Melin, 1927). A knowledge of the behavior of diarch lateral root initials is essen- tial to a proper understanding of the mycorrhizal habit. Three des- tinies are possible as follows: (1) they may abort before or during emergence or before they have reached an appreciable length; (2) TABLE 1. ROOT MEASUREMENTS OF 3 nS OLD PINUS SYLVESTRIS L. SEEDLINGS GROWN IN SAND XED HUMUS Seedling number (Average 6) 1 2 3 4 (Average 4) Average 10 r Six oe grown in puaiparranse earth Jae (x 0.31 to Number of long roots Total 18.9 Number of eset hone een cae agile cm. 16 26 12 27 12 25 18 25 10 23 18 18 14.3 24.0 12 17 9 7 13 20 8 21 10.5 18.8 12.8 21.9 0.30 w) and infra-red; low grown at the Royal Institute of ee aca Forestry, Stockholm, foeien r four in Total length of laterals cm. 46.0 58.8 Taw Length of onges lateral cm. 0 7.8 hall Total length of all lon Number of short roots 473.7 Number of ph roots per of long roots. eas iy ab Page kes 1927) in 50% solar radiation minus short ultra- 50% solar radiation minus ultra-violet; all were [€6l SAHNLVad TVZIHYYOOAW “MVOd GNV HOLVH 92 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV they may be attacked by microdrganisms with resultant conversion into mycorrhizal or pseudomycorrhizal structures; (3) they may, by con- tinued growth, become mother roots. The latter in turn give rise to initials having the same possible destinies and to monarch short roots. The distribution of lateral root initials along the radicle in Corsican Pine, their growth and their anatomical features have been studied by Aldrich-Blake (1930). He found that seedlings grown in a sandy soil produced initials opposite any of the four protoxylem points at dis- tances of 1.6 to 2.7 mm. apart. One-half of these became abortive before they grew one millimeter in length and only one-sixth grew beyond the very moderate length of 10 mm. The termination of growth in many initials at an early stage was associated with mycor- rhizal formation. The question of whether they were predestined to become mother roots or mycorrhizal roots was therefore examined. Internal as well as external factors were judged to be operative as evi- denced by the finding of an abortive initial that had ceased growth while still within the intact endodermis of the parent radicle (1. c., p. 22). It was further determined that the initial size of the roots which became mycorrhizal was distinctly smaller than that of those which continued growth and became mother roots. Aldrich-Blake writes, for example—‘The mean protoxylem diameter of the mother roots was 117 + 6 u, while that of the diarch basal portion of the dichotomized roots was 73 * 3 u; and the probability of significance of the difference between these two measurements was over 100 to 1” (1. c., p. 23). He concluded that the two types of roots were distinct from the very beginning in respect to size and that conversion to mycorrhizal organs was “the response of feeble roots to strong fungal infection” (1. c., p. 24). Melin (1925) also presents convincing evi- dence that many laterals are “feeble” from the very beginning. It is apparent from photographs (Melin, 1. c., pp. 60-62, figs. 23-27) of Pinus sylvestris that even in the complete absence of microorganisms and in a uniform substratum the common differentiation of laterals into slow-growing roots and mother roots still occurs. After three years many of the slow-growing laterals on pure culture seedlings had not reached the length of one centimeter (1. c., 1925, p. 61). During examinations of approximately 2500 seedlings of Pinus syl- vestris L. and P. Strobus L. grown in pot cultures (containing mixtures of forest humus or soil and sand) one of the authors has observed that only in the autumn, after elongation of the rapidly growing long roots has nearly ceased, do the tips of the latter frequently acquire the mycorrhizal structure (Plate 59, A). This leads us to believe, espe- cially in view of the slow elongation rate of short roots which normally are always converted into mycorrhizal organs, that rapid elongation is 1933] HATCH AND DOAK, MYCORRHIZAL FEATURES 93 the primary reason long roots ordinarily escape attack by mycorrhizal fungi. This hypothesis assumes (1) that mycorrhizal fungi can only successfully attack a root in the primary condition, and (2) that in rapidly growing long roots the secondary condition is acquired before mycorrhizal fungi are able to penetrate sufficiently to initiate the char- acteristic response. Rayner (1927, p. 99) has in part explained a somewhat similar phenomenon, namely, the absence of the endophyte from rapidly growing roots of Calluna vulgaris (L.) Salisb. in the early spring, by assuming differential growth rates for fungus and root at low spring temperatures. The same phenomenon is probably operative in tree seedlings, except that root growth rather than fungal growth is assumed to be retarded by low temperatures or other factors. In this connection a seedling of Pinus resinosa Ait. grown under very adverse conditions in pure culture has been observed to have all of its roots (long as well as short) converted into mycorrhizal roots (later com- munication by Hatch). We conclude, therefore, that every root tip in the primary condition is theoretically capable of conversion into a mycorrhiza. That this normally does not occur in nature we attribute to differences in rapidity and character of growth of long and short roots. Lateral root initials that become mycorrhizal (Plate 59, B) acquire the monarch structure in the dichotomized branches above their diarch basal portion, and they are ephemeral structures. They undoubtedly function in the same manner as true mycorrhizal short roots. For the purpose of our present discussion, therefore, we may include them in the category of short roots. TYPES OF SHORT ROOTS Three distinct kinds of short roots may be distinguished: (1) the uninfected short root, (2) the infected short root in which the mycor- rhizal structure is absent (pseudomycorrhiza), (3) the infected short root which has acquired the mycorrhizal structure (mycorrhiza). The first of these, or uninfected short roots, are distinguished by the following features (Plate 57, B): 1. Formation of root hairs from epidermal cells (when grown in a suitable nutrient solution and in a substratum that is adequately aérated). 2. Continuous although slow elongation. 3. Absence of hypertrophy of cortical cells. 4. Complete absence of fungus infection. 5. Frequent but not profuse dichotomy’ (Melin, 1925, figs. 23, 24). should be noted that dichotomous branching of the short roots is an oe sive ee so far as is known, of the genus Pinus among conifers. The ter 94 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XIV This type of short root is exceedingly rare. It appears consistently and probably exclusively in pure cultures and in open, mineral solution sand cultures that are free from organic materials. Root hair pro- duction is dependent on the condition of the substratum relative to aération and chemical composition (Melin, 1925, p. 61). It is notable that root hairs on short roots are elongations of epidermal cells (Plate 57, B and D). Its diagnostic value, considering that all root hairs on long roots of Pine originate from the second or third layer of cortical cells (Aldrich-Blake, 1930, p. 16; Noelle, 1910; and Plate 59, D of the present paper), has not been emphasized. In pure culture short roots exhibit renewed yearly growth in length and have been reported to reach a maximum length of 20 mm. in two years (Melin, 1925, p. 60). The absence of a root cap on mycorrhizal roots is considered to be a natural consequence of attack by mycorrhizal fungi by Laing (1932). McArdle (1932, p. 312) on the other hand attaches physi- ological significance to the presence of this structure in mycorrhizal roots. That it actually is not present on either uninfected or mycor- rhizal short roots in Pinus is demonstrated in Plate 57, A, B and C. The second type of short root exhibits infection, but of a distinctly non-mycorrhizal type. The name pseudomycorrhiza was given to this structure by Melin (1917). Melin applied it to short roots possessing the following features, which likewise are characteristic of American material: 1. Intracellular infection by soil inhabiting fungi that are usually of minute proportions. 2. Absence of root hairs. 3. Absence of hypertrophy in cortical cells of the roots. 4. Complete absence of intercellular net (Hartig’s net). 5. Early termination of growth in length (two to three millimeters is usually the maximum length). 6. Occasional dichotomous branching. The general habit of pseudomycorrhizal roots of Pinus sylvestris is illustrated in Plate 58, A. A longitudinal section of one of these is re- produced in Plate 57, A. A common type of infection of the cortical cells is shown in Plate 60, B, C and D, and by Melin (1917, fig. 39). Moller (1902) first called such a structure an “entotrof” mycorrhiza because it was characterized by endophytic infection only. Melin (1917, p. 360) pointed out that it exhibits no similarity to the true endotrophic infection occurring in the Ericaceae and Orchidaceae. “Gabelmykorrhiza” was coined by Melin (1923) in his studies on Pinus to cover mycorrhizal roots ceils ncaa by this type of branchin In this seen McArdle (1932, p. fe: has referred . the coralloid mycorrhiza of Spruce as the “Gabelmykorrhiza” of Melin; bu mycorrhizal short roots of Sy sor racemosely and should not be ates with the dichotomous type. 1933] HATCH AND DOAK, MYCORRHIZAL FEATURES 95 Neither may it be confused with an ectotrophic mycorrhiza since the fine endophytic hyphae wander from cell to cell without regard to structure, even occurring in the endodermis and parenchyma cells of the central cylinder (1. c., p. 359). Hyphae never occur in masses in the cells (ibid.). Several species of fungi may be present at once either in abundance or in extreme scarcity. In exterior appearance a pseudomycorrhiza differs from an uninfected short root chiefly in its shorter length and darker color, when the latter does not possess root airs. In contrast to the foregoing root types, attack by a mycorrhiza- producing fungus normally results in short roots that possess the fol- lowing well-known features: 1. Intercellular net between cortical cells (Hartig’s net). Plate 59, E; Plate 60, A. 2. Intracellular infection of the cortical cells (which may or may not be discernible; Melin, 1925). 3. A fungal mantle composed of a few to many layers of tightly packed and interlaced hyphae covering the region of cell hypertrophy (Plate 60, A). The mantle has been reported to be almost completely lacking on mycorrhizal short roots of certain experimental seedlings (Melin, 1925). 4. Hypertrophy of the cells of the cortical region. 5. Characteristic profuse dichotomous branching with age (Plate A ae © 6. Continued elongation of the dichotomous branches under favor- able conditions of soil and weather during one growing season. The external appearance and internal structure of a typical mycor- rhiza is illustrated in Plate 57, C; Plate 58, C; Plate 60, A. The cells of the cortical region of an uninfected short root of Pinus Strobus grown in pure culture (Plate 57, C) are on the average 29 y in radial and 61 y in longitudinal directions. Dimensions of similar cells of a mycorrhizal short root (Plate 57, B) [also pure culture P. Strobus and Lactarius deliciosus (L.) Fr.] are 43 y and 40 y respectively in radial and longitudinal section (average of 15 cells in both cases). Based on the uninfected short root, therefore, the increase in radial size (hypertrophy) of individual cells is 67 per cent. The diameter of the uninfected short root is 300 y, while the average diameter of the mycorrhiza in this case is 485 u, an increase of about 62 per cent. Comparative studies of pure culture seedlings indicate that mycorrhizal formation in short roots of Pine is not accompanied by either hyper- plasia or hypoplasia. It is evident, therefore, that Laing’s statements (1932, pp. 10, 11) that the root cap disappears during the develop- 96 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV ment of a pine mycorrhiza and that the number of cortical cells is less as a result of mycorrhizal formation arose from failure to distinguish between long and short roots. Attack by mycorrhizal fungi, if early in the season, almost invariably results in dichotomous branching of the short root. Judging from the exceedingly slow growth of uninfected short roots (see above), including slow growing laterals, the authors see little basis for the numerous references in the literature to the cessation of elongation of roots following attack by mycorrhizal fungi. Laing (1932), in concluding that mycorrhizal fungi exert a retarding influence on growth in length of “rootlets,” evidently failed to make comparisons with uninfected short roots or to include data supporting this view. In this connection McArdle (1932) reports that he observed the forma- tion of what he believed to be “‘mycorrhizae” through glass plates in sand cultures and that elongation did not proceed after attack by mycorrhizal fungi. McArdle’s photographs (ibid., figs. 1-6) as well as his measurements demonstrate conclusively that he was dealing entirely with long root tips. That mycorrhizal short roots do grow after attack by the causative fungus is adequately evidenced by profuse dichotomy which probably never occurs in the absence of true mycorrhizal infec- tion. Likewise the present authors have observed continued growth and branching of typical mycorrhizal short roots, individually, in their own pure cultures on several species of Pines in association with several species of Hymenomycetes. The profuse branching of a mycorrhizal short root of P. Strobus is illustrated in Plate 59, C. This short root has approximately 50 individual growing tips; another growing beside it had 84: in fact, all of the short roots on the long root from which these were taken were much branched. We can find no reason to assume that the total length (from the base of the mycorrhiza to the tip) of any of the numerous tips is less than the total length of the short root would have been had it remained non-mycorrhizal. Instead, mul- tiple tip growth represents a greatly accelerated development which increases the absorbing surface many times. In this connection, it is to be noted that by means of hypertrophy of the cortical cells and increase in surface area by means of absorbing fungal hyphae the surface area of a single-tipped mycorrhizal short root seemingly already exceeds that of an uninfected one. A comparative study of the actual absorbing power of mycorrhizal and non-mycorrhizal short roots should yield interesting results. We have shown that anatomically and morphologically a short root, the organ ordinarily involved in an ectotrophic mycorrhiza, differs markedly from a long root. It seems logical to assume that these root 1933] HATCH AND DOAK, MYCORRHIZAL FEATURES 97 types differ as fundamentally in their physiological functions. In in- vestigations designed to elucidate mycorrhizal nutrition, therefore, it is essential that these differences be recognized and that the types and subgroups of roots be adequately separated. Similarly in ecological studies, when the abundance of mycorrhizal roots is given in percentage or other figures, these data should apply to short roots only; and these should be separated into the three subgroups—mycorrhizal, pseudo- mycorrhizal and uninfected short roots respectively. The authors jointly wish to recognize their indebtedness to the Allegheny Forest Experiment Station maintained at Philadelphia in cooperation with the University of Pennsylvania. The senior author also gratefully acknowledges his indebtedness for advice and _ fa- cilities given by Professor Elias Melin, Mycological Laboratory, Royal Academy of Forestry, Stockholm, Sweden; by Professor H. Hesselman, Royal Institute of Experimental Forestry, Stockholm; Pro- fessors R. T. Fisher and P. R. Gast, Harvard Forest; and Professors Oakes Ames and J. H. Faull, Arnold Arboretum, Harvard University. The senior author is further indebted to the American Scandinavian Foundation for stipends to study in Sweden, as Fellow of the Founda- tion; and to the Arnold Arboretum for support that enabled him to continue the work at Harvard University, under the auspices of the Arboretum and the Department of Botany, as a graduate student. LABORATORY OF PLANT PATHOLOGY, eee eee AND THE DeE- PARTMENT OF BoTANy, HARVARD UNIVER ALLEGHENY Forest EXPERIMENT STATION, ae oF PLANT INDUSTRY, U. S. DEPARTMENT OF AGRICULTURE. LITERATURE CITED AAaut, Beder Chan T. (1923). Ueber der Mykorrhiza der Buche. (Allg. 0-50 ALDRICH-BLAKE, R. N. (1930). The plasticity of the root system of Corsican Pine in early life. (Oxford Forestry Memoir 12:1-60. Frank, B. (1885). Ueber die auf Wurzelsymbiose beruhende Ernahrung gewisser Baume durch unterirdische Pilze. (Ber. Deutsch. Bot. Ges. 3:128-1 Gast, P. R. (unpub.). Growth of Scots Pine in pot cultures of Swedish soils under varied light conditions. (To be published in the Meddel. fr. Stat. Skogsf6rs.-anst. Stockholm, Sweden. ) HEssELMAN, H. (1927). Studier 6ver barrtradsplantans utveckling 1 rahumus. I. Betydelsen av kvavemobiliseringen i rdhumustacket ar tall- och granplantans forsta utveckling. is German sum- y.) (Med ne fr. Skogsfors.-anst. 23:337-411 ere, A. P. (1931). The concept of mycorhiza. (Mycologia, 23:147- —————.. (1932). The literature of on (Manuscript, U. S. Dept. Agr. Library, Washington, D. C., pp. 1-948.) 98 JOURNAL OF THE ARNOLD ARBORETUM [voL. XIV Laine, E. V. (1932). Studies on tree roots. (Forestry Commission, Masur, K. (1927). A study of the ectotrophic mycorrhizas of wood lants. (Memoirs College of Science, Kyoto Imperial University, Series B, 3:149-279.) McArpie, R. E. (1932). The relation of mycorrhizae to conifer seed- lings. (Jour. Agr. Res. 44:287-316.) McDouaat, W. B. Cee Mycorrhizas from North Carolina and East- 5:141-148. ern Tennessee. (Amer. Jour. Bot. MELIN, E. Sse : Studier over de Norrlindske myrmarkernas vegeta- tion med sarskild hansyn till deras skogsvegetation efter torrlaggning. ( Akade misk Avhandling, Almqvist & Wiksells, Uppsala ————_.. (l Experimentelle Untersuchungen aber die Konstitu- tion und Oekologie der Mykorrhizen von Pinus silvestris L. und Picea ies (L.) Karst. (Mykologische Untersuchungen und Berichte, 2: ; ae ntersuchun liber die Bedeutung der Baum- mykorrhiza, eine dkologisch- a Studie. (Gustav eee ena. . (1927). Studier Over barrtradsplantans utveckling i rahu- mus. II. Mykorrhizans utbildning hos tallplantan i olika rahumus- former. (With German summary.) (Meddel. fr. Skogsfors.-anst. 23 :433-494. ) Mo ter, A. (1902). Ueber die Wurzelbildung der ein- und zweijahrigen Kiefer im miarkischen Sandboden. (Zeit. f. Forst.-u. Jagdwesen, 34:197-215 NoELLE, W. (1910). Studien zur vergleichenden Anatomie und Morph- ologie der Koniferenwurzeln mit Riicksicht auf die Systematik. (Bot. RAYNER, ML. C. (1927). Mycorrhiza, an account of non-pathogenic infec- tion by fungi in vascular plants and Bryophytes. (Reprint from New Phytologist, No. 15, pp. 1-246, London.) RexHausen, L. (1920). Ueber die Bedeutung der ektotrophen Mykor- rhiza fiir die hdheren Pflanzen. (Beitr. z. Biol. d. Pflanzen, 14:19-58.) WEYLAND, ( ; ou Ernahrungsphysiologie mykotropher Pflanzen. (Pringsh, Jahrb. f. Wiss. Bot. 51:1-80. Weevers, T. (1916). Das Vorkommen des Ammoniaks und der Am- moniumsalze in den Pflanzen. (Recueil des travaux botaniques néer- landais, 13:63-104.) DER Ren OF PLATES ea Professor E. Melin and the senior author.) See also Plate B. Longi tudinal (not medial) section of P. Strobus short root grown in pure culture, showing absence of hypertrophy and a root hair arising from an epidermal cell ; 69. Longitudinal _ medial) section of dichotomized mycor- rhizal short root P. Strobus grown in pure culture with Lactarius jae showing mantle, intercellular net and cell ae aes ry ; 7. (Ditto; Plate 58, C. sverse section of monarch short root of P. sylvestris n open sand culture showing root hairs arising from Sa cells; X 123. (Nutrient sand culture.) Jour. ARNOLD ArB., VoL. XIV. PLATE 57 MYCORRHIZAL AND OTHER FEATURES OF THE Root SysTEMS OF PINUS. Jour. Arnotp Ars., VoL. XIV. PLATE 58 MycorRHIZAL AND OTHER FEATURES OF THE Root SYSTEMS OF PINUS. PLATE 59 XIV. ARB., VOL. ARNOLD Jour. suwoen; op ge °°" MycorRRHIZAL AND OTHER FEATURES OF THE Root SysteMs oF PINUS. Jour. Arnotp Ars., VoL. XIV. PLATE 60 MycoRRHIZAL AND OTHER FEATURES OF THE Root SysTEMS OF PINUS. 1933] Plate 58. Plate 59. Plate 60. HATCH AND DOAK, MYCORRHIZAL FEATURES 99 A. Swag ag re short roots on mother root of P. syl- vestris ; 12. (Ditt ; Plate , A; material from unpub- or. Uninfected short roots possessing root hairs on mother root of P. Strobus in pure culture; 22. C. Mycorrhizal short roots on mother roots of P. Strobus grown in pure culture with Lactarius deliciosus; < 10. A. Mycorrhizal mother root tips of P. ‘Siabus: plants re- moved from soil after onset of cold weather, showing also sath mycorrhizal short roots; X rown in sand-mixed humus (Infektionsjord from forests in Sweden. See Hessel- man, 1927, and Gast, unpublishe B. Lateral oe initial from radicle of P. sylvestris converted into a mycorrhiza; basal portion pseudomycorrhizal ;. 5, a Profusely branched short root of P. Strobus; X oot hairs on long root of P. Strobus originating sub- epidermally ; be E. Int tercellular net (Hartig’s net) in a mycorrhiza of Tsuga canadensis (L.) Carr.; * 512. A. Mantle structure a roniawye net of cortical cells in a mycorrhiza of P. rigida Mill.; 30. B, C and D. Set taper type of infection in long roots of P. Strobus; X 512. Bn _ 7 7 : JOURNAL OF THE ARNOLD ARBORETUM VoLUME XIV. APRIL, 1933 NUMBER 2 ADDITIONAL NOTES ON THE ORCHIDS OF THE NEW HEBRIDES AND SANTA CRUZ ISLANDS Oakes AMES In an earlier paper’ on the orchids of the New Hebrides and Santa Cruz Islands, I included all of the species represented in S. F. Ka- jewski’s collections with the exception of several which were too frag- mentary for critical study. Since my paper was published additional material has been found and I am now able to complete the record of Kajewski’s 1928 and 1929 series. During my studies I have examined a rich collection of specimens containing species that have not yet been reported to be natives of the New Hebrides or Santa Cruz group. In the following paper I have included these, and to make the record more serviceable I have included also a number of species which although already reported from the New Hebridés, are now represented by specimens recently collected or by material referable to them. Un- fortunately many of the specimens found by Dr. Morrison on Aneityum lack flowers and are indeterminable. I have only included those species which are quite clear, and in a few cases, when a plant represented a genus new to the region, I have included it, although for want of flowers the specific name could not be arrived at. For the opportunity to examine the collections made on Malekula, Efate, Eromanga, Aneityum and the Banks Islands by Dr. Morrison and L. Cheeseman, I am indebted to Sir Arthur W. Hill, Director of the Royal Botanic Gardens, Kew. Habenaria physoplectra Reichenbach f. in Linnaea, xr. 17 (1876). Aneityum: Anelgauhat, Dr. Morrison, s. n., June 15, 1896; hills between Ithumu and Anelgauhat, Dr. Morrison, s. n., July 6, 1896. Efate: Undine Bay, hills, Dr. Morrison, s. n., August 27, 1896. 1Journal of the Arnold Arboretum, xr. 127-141 (1932). 102 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XIV The specimens referred to this species are too old to be serviceable for purposes of accurate identification, but vegetatively they bear a close resemblance to Reichenbach’s type specimen (MacGillivray no. 27) and the withered remains of the flowers on the plant collected be- tween Ithumu and Anelgauhat have the scrotiform spur that is one of the chief characters of H. physoplectra. MacGillivray’s specimens were collected on Aneityum. Fritz Kraenzlin refers Habenaria novo- ebudarum F. v. Mueller to the synonymy of this species. (Orch. Gen. & Sp. 905). Habenaria ponerostachys Reichenbach f. in Bonpl. mi. 213 (1855). Aneityum: Hill north of Anelgauhat, Dr. Morrison, s. n., June 1896.—Also Philippine Islands. In habit and floral structure the specimen from Aneityum resembles very closely the type and specimens from Leyte and Mindanao. The broad petals, aristate lateral sepals and the lobes of the labellum are similar to corresponding structures in the Philippine plants. The ex- tension of range is interesting, because H. ponerostachys is not com- mon in the Philippine Islands and has not been reported heretofore from any other part of the world. Habenaria stenodon Reichenbach f. in Linnaea, xi. 17 (1876). Banks Islands: Vanua Lava, L. Cheeseman, s. n., No- vember 1929.—Already found on Aneityum by MacGillivray. Corybas mirabilis Schlechter in Fedde, Rep. Spec. Nov. x1x. 22 (1923). Corysanthes mirabilis Schlechter in Bull. Herb. Boiss. ser. 2, v1. 296 (1906). Aneityum: Dr. Morrison, s. n., June 26, 1896 (dark glossy purple, variegated with whitish lines or meshes on lip and upper sepal). The type was collected by Dr. Morrison on the summit of the peak between Anumy and Ithug, c. 2300 feet altitude in June 1896. The specimens in the Kew Herbarium are accompanied by a note which reads: ‘“‘On peak crossed June 26, 1896.” Nervilia sp. Aneityum: Hill north of Anelgauhat, Dr. Morrison, s. n., June 15, 1896. A single leaf constitutes the specimen examined. This leaf resem- bles very closely the leaves of Nervilia MacKinnoniu (Duthie) Schlechter. Epipogon roseum (D. Don) Lindley in Jour. Linn. Soc. 1. 177 (1857). 1933] AMES, ORCHIDS OF THE NEW HEBRIDES 103 Malekula: South West Bay, in bush at sea level, L. Cheese- man, S. n., January 1930.—Also Java, Ceylon, New Guinea, tropical India and Australia. Goodyera triandra Schlechter in Bull. Herb. Boiss. ser. 2, v1. 298 (1906). Efate: Undine Bay, Dr. Morrison, s. n., August 20, 1896; Mount Macdonald, Dr. Morrison, s.n., August 18, 1896. The specimens in the Kew Herbarium are similar to the type in being triandrous. 4 Platylepis Morrisonii Schlechter in Fedde, Rep. Spec. Nov. rx. 161 (1911) Aneityum: Hills between Anelgauhat and Anumy Valley, Dr. Morrison, s. n., June 25, 1896; hills between Ithumu and Anelgauhat, Dr. Morrison, s. n., July 6, 1896. Cystopus aneytyumensis Schlechter in Fedde, Rep. Spec. Nov. rx. 282 (1911). Aneityum: Near summit of peak south of Ithumu, Dr. Mor- rison, S. n., June 30, 1896; ascent of peak south of Ithumu, 1000-1835 feet altitude, Dr. Morrison, s. n., June 30, 1896. Cystopus fimbriatus J. J. Smith in Bull. Dép. Agric. Indes Néerl. no, 10. p.3:.(1 907): Aneityum: Hills between Ithumu and Anelgauhat, Dr. Mor- rison, 8. n., July 6, 1896.—Also Dutch New Guinea. There are minor differences between the specimens from Aneityum and the type of C. fimbriatus, but they are hardly sufficient to justify specific separation. Zeuxine Erimae Schlechter in Schumann & Lauterbach, Nachtr. Fl. Deutsch. Schutzgeb. 90 (1905). Banks Islands: Vanua Lava, Avuas, in brush, at 250 feet altitude, L. Cheeseman, s. n., October 7, 1929 (roots scarcely lodged in soil).—Also Kaiser-Wilhelmsland. Anoectochilus sp. Aneityum: Hills between Ithumu and Anelgauhat, Dr. Mor- rison, s.n., July 6, 1896 (leaves dark green, velvety dull or with sheen according to incidence of light, veined with silvery whitish lines, under surface purplish red veined with greenish). The only specimens collected are sterile. The leaves resemble those of Anoectochilus Roxburghii Lindl. Vrydagzynea Cheesemanii Ames, sp. nov. In habitu V. albidae Bl. similitudinem gerit. Caules graciles, foliosi. 104 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Folia plus minusve congesta, petiolata, ovato-lanceolata, acuminata, acuta, membranacea. Racemus multiflorus. Bracteae racemi elongatae, lineari-lanceolatae, scariosae, margine glandulosae. Sepala lateralia irregulariter ovato-lanceolata, obtusa, uninervia, apice carinata. Se- palum dorsale anguste ellipticum, obtusum, apice incrassatum. Petala anguste elliptica, obtusa, uninervia, apice incrassata. Labellum in calcar conicum productum; lamina labelli suborbicularis. Columna generis. Terrestrial herb 17.5-21.5 cm. tall, in facies very similar to V. albida BI. Stems slender bearing several foliaceous bracts near the base, leafy above. Leaves crowded, including the petiole up to 10 cm. long, 1.5-3 cm. wide; lamina ovate-lanceolate, acuminate, acute, papyraceous. Pe- duncle slender, including the raceme up to 12 cm. long. Raceme 4-5.4 cm. long, about 1 cm. in diameter near the base, tapering gradually upward, densely many-flowered. Bracts of the raceme up to 1 cm. long, linear-lanceolate, glandular-pubescent on the margin, mid-nerve prominent. Ovary about 7 mm. long, smooth. Lateral sepals 3.5 mm. long, 2 mm. wide, asymmetrically ovate-lanceolate, obtuse, conspicu- ously thickened at the carinate apex, l-nerved. Dorsal sepal 3.5 mm. long, narrowly elliptical, obtuse, with the blunt apex considerably thickened, 1-nerved. Petals adherent to the dorsal sepal, 3.5 mm. long, 1.5 mm. wide, narrowly elliptical, obtuse, thickened at the tip, 1- nerved. Labellum calcarate, 6 mm. long from the tip of the spur to the tip of the expanded lamina; lamina 1.5 mm. long to the opening of the spur, 2 mm. wide, suborbicular, fleshy, distinctly thickened or bicallose at the base in front of the opening to the spur, somewhat thickened at the tip. Spur about 4.5 mm. long, about 1.5 mm. in diameter, tapering gradually to the obtuse or subacute tip; within on the posterior wall hang two pedicellate verruciform appendages, each appendage and its pedicel 1.5 mm. long. Column 2 mm. long. Malekula: South West Bay, at sea level, L. Cheeseman, s. n. (type in Herb. Kew; duplicate type in Herb. Ames, no. 37831), Jan- uary 1930. From Vrydagzynea elongata Bl., this species differs conspicuously in having a conical spur with an acute or nearly acute apex. The smaller flowers, dissimilar lip and larger leaves seem to differentiate it from V. neo-hibernica Schltr. Malaxis lunata (Schltr.) Ames in Jour. Arnold Arb. xur. 129 (1932). Microstylis lunata Schlechter in Fedde, Rep. Spec. Nov. 1x. 162 (1911). 1933] AMES, ORCHIDS OF THE NEW HEBRIDES 105 Aneityum: Hills between Ithumu and Anelgauhat, Dr. Mor- rison, Ss. n., July 6, 1896; near Anelgauhat, Dr. Morrison, s. n., June 1896; ascent of peak south of Ithumu, 1000-1835 feet altitude, Dr. Morrison, s. n., June 30, 1896. Oberonia sp. Aneityum: Gulley near Anelgauhat, Dr. Morrison, s. n., June This material collected by Dr. Morrison is in fruit and indetermin- able. In general appearance the plant resembles O. Betchei Schltr. It also suggests O. ensiformis Lindl. Oberonia sp. Aneityum: Anelgauhat to Anumy, Dr. Morrison, s. n., June 25, 1896. Sterile specimens that are vegetatively similar to O. aporophyilla Reichb. f. Oberonia glandulosa Lindley, Fol. Orch. Oberonia, 6 (1859). Efate: Undine Bay, Dr. Morrison, s. n., August 18, 1896. Aneityum: Hills between Ithumu and Anelgauhat, Dr. Morri- son, 8. n., July 6, 1896 (fruiting specimen ).—Also Tahiti. Liparis elegans Lindley in Wall. Cat. no. 1943 (1828), nomen; Lindley, Gen. & Sp. Orch. 30 (1830). Aneityum: Peak between Anumy and Ithug, Dr. Morrison, s.n., June 26, 1896—Also Malay Peninsula and Borneo. Chrysoglossum aneityumense Ames, sp. nov. Rhizoma gracile, verisimiliter statu juvenili vaginis laxis vestitum. Folium chartaceum, anguste ellipticum, subacutum, basi rotundatum in petiolum sulcatum attenuatum. Scapus erectus, pauciflorus. Sepala lateralia oblongo-lanceolata, acuminata, cum labello saccum scroti- formem formantia. Sepalum dorsale simile. Petala lanceolata, acuta. Labellum conspicue trilobatum; lobi laterales anguste semielliptici, valde obtusi; lobus medius semiorbicularis; discus inaequaliter tri- carinatus. Columna crassa, apice alata, basi in pedem producta. Rhizome slender, partly concealed by the fibrous remains of sheath- ing bracts, 2-4 mm. thick in dried specimens. Leaf including the petiole 7.5-9.5 cm. long, about 3.5 cm. wide, chartaceous when dry, narrowly elliptical, subacute, rounded at the base; petiole about 1.5 cm. long, terminal on a slender abbreviated pseudobulbous stem. Scape, the continuation of a leafless stem arising from near the base of the leaf, including the few-flowered raceme about 1 dm. long. Raceme 4-6 cm. long, loosely five- to eight-flowered, flowers about 1 cm. apart. Bracts 106 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV of the inflorescence about 1 cm. long, lanceolate, acute, membrana- ceous. Pedicels slender, including the ovary 1.5 cm. long, sharply curved. Lateral sepals 1 cm. long, 4 mm. wide, oblong-lanceolate, acuminate, acute, forming with the labellum a conspicuous scrotiform sac, spreading. Sac 5 mm. long, about 3 mm. in diameter. Dorsal sepal about 1 cm. long, 3.5 mm. wide, similar to the laterals. Petals 9 mm. long, about 3 mm. wide, lanceolate, acute, spreading. Labellum connected with the sac by a narrow claw, including the sac 1.4 cm. long, conspicuously 3-lobed; lateral lobes 4 mm. long, 2 mm. wide, narrowly semi-elliptic, rounded at the apex, separated from the middle lobe by an abbreviated isthmus; middle lobe semi-orbicular, 6 mm. long, 8 mm. wide; disc 3-carinate with the outer carinae auriculate near the base of the lateral lobes and extending nearly to the center of the middle lobe where they become conspicuously elevated into semi- elliptical plates, the central carina is shorter than the lateral ones and hardly expanded at the tip. Column fleshy, free portion 4.5 mm. long, produced at base into a conspicuous foot, at the summit becoming con- spicuously winged round the androclinium. Aneityum: Anelgauhat, below 800 feet altitude, Dr. R. Mor- rison, s. n., June 1896 (type in Herb. Kew; duplicate type in Herb. Ames, no. 39055). The closest ally of this species appears to be Chrysoglossum papu- anum (Schitr.) J. J. Smith which differs from the New Hebridean plant in having the lateral lobes of the labellum acute and in having slightly different carinae on the disc. Vegetatively these species are very similar with flowers that are about equal in size. Dendrobium calcaratum A. Richard, Sert. Astrol. 18. t. 7 (1834). Aneityum: Hills between Ithumu and Anelgauhat, Dr. Mor- rison, s. n., July 6, 1896; between Anelgauhat and Anumy, Dr. Morrt- son, s. n., June 25, 1896; ascent of peak south of Ithumu, at 1000-1835 feet altitude, Dr. Morrison, s. n., June 30, 1896.—Already found on Vanikoro in the Santa Cruz Islands. Dendrobium gnomus Ames, sp. nov. Caules dense caespitosi, graciles, perbreves. Folia disticha, lineari- oblonga, apice inaequaliter bidentata. Racemi pauciflori, laterales. Sepala lateralia mentum obtusum formantia, triangularia, acuta, valde membranacea. Sepalum dorsale lanceolatum, acutum. Petala lanceo- lata. Labellum elongatum, simplex, infra medium constrictum, bical- losum, supra medium in laminam lanceolatam acutam expansum. Columna generis. Roots finely fibrous, white, smooth. Stems up to 3.5 cm. long, about 1933] AMES, ORCHIDS OF THE NEW HEBRIDES 107 2 mm. in diameter when dry, yellow, deeply furrowed longitudinally, four- to eight-jointed, with the internodes about 5 mm. long and nearly equal in diameter. Leaves alternate, 1.7-2.1 cm. long, 2.5 mm. wide, linear-oblong, tapering gradually toward each end, unequally bi- dentate at the blunt apex. Leaf-sheaths somewhat complanate, heav- ily nerved, remaining as loose fibres at the nodes of the stem after the fall of the leaves. Inflorescence breaking forth at the nodes of the naked stems; rachis about 6 mm. long. Bracts of the inflorescence scarious, those subtending the flowers lanceolate, acute. Flowers sev- eral, bright purple, about 11 mm. long from the tip of the dorsal sepal to the tip of the mentum formed by the lateral sepals, membranaceous. Lateral sepals slightly spreading, triangular, acute, including the men- tum 1 cm. long, about 3 mm. wide near the middle; mentum 4 mm. long, about 2.5 mm. in diameter, closed in front, protuberant anteriorly at the tip. Dorsal sepal lanceolate, acute, 5.5 mm. long, 2 mm. wide. Petals lanceolate, acute, erect, diverging slightly from the dorsal sepal, 5 mm. long, about 1 mm. wide. Labellum 8 mm. long, expanded above the middle into an elliptic-lanceolate lamina which is 2 mm. wide, basal portion narrowly oblong, slightly constricted near the base of the ex- panded upper portion with a rounded thickening or callus on each side of the constriction. Column produced into an elongated foot. Santa Cruz Group: Vanikoro, in rain forest at 800 meters altitude, growing in moss, S. F. Kajewski, no. 605 (type in° Herb. Ames, no. 38083), November 11, 1928 (flowers bright purple, very pretty; (only one specimen seen). One of the smallest representatives of the Pedilonum section of Dendrobium, about equal in size to D. cyanocentrum Schltr. Dendrobium Mohlianum Reichenbach f. in Bot. Zeit. xx. 214 (1862). Dendrobium neo-ebudanum Schlechter in Bull. Herb. Boiss. ser. 2, vi. 456 (1906). Aneityum: Between Anumy and Ithug, at 2300 feet altitude, Dr. Morrison, s. n., June 26, 1896; hills near Anumy, Dr. Morrison, Ss. n., June 6, 1896. Eromanga: Peak south of Dillon’s Bay, Dr. Morrison, s, n., August 5, 1896; Traitor’s Head, summit of old crater growing on rotten wood, 2400 feet altitude, L. Cheeseman, no. 72, August 21, 1930 (flower pale vermillion, not deep).—Also Fiji Islands and Samoa. Dendrobium Mooreanum Lindley in Jour. Hort. Soc. vr. 272 in footnote (1851). Dendrobium Fairfaxii Rolfe in Gard. Chron. ser. 3, v. 798 (1889). Aneityum: Peak south of Ithumu, Dr. Morrison, s. n., June 108 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV 30, 1896 (flower generally pure white, labellum greenish, veined with purple); Anelgauhat to Anumy, Dr. Morrison, s. n., June 25, 1896; peak between Anumy and Ithug, Dr. Morrison, s. n., June 26, 1896. Efate: Undine Bay, Dr. Morrison, s. n., August 27, 1896. Ero- manga: Hill south of Dillon’s Bay, Dr. Morrison, s. n., August 5, 1896. Dendrobium occultum Ames, sp. nov. Caules lageniformes vel cylindracei, prope apicem diphylli. Folia oblongo-lanceolata, apiculata, in sicco chartacea. Flores singuli. Se- pala lateralia mentum elongatum formantia, oblonga, prope apicem attenuata, apiculata. Sepalum dorsale elliptico-oblongum, apiculatum. Petala oblanceolata, apice rotundata, breviter apiculata. Labellum oblanceolatum, apice apiculatum, prope basim lamella transversa ornatum; discus leviter tricarinatus. Columna generis, clinandrio den- ticulato. Stems 1.5-2.5 cm. long, lageniform or cylindrical, bearing at the summit two obliquely ascending leaves, when young concealed by tubu- lar sheaths which persist as stiff erect fibres at the nodes on the mature stems. Leaves 8-13 cm. long, up to 1.5 cm. wide, oblong-lanceolate, apiculate, papyraceous when dry. Flowers 2.8 cm. long, borne singly at the nodes of the leafless stems. Pedicel with the ovary 8 mm. long, ascending. Lateral sepals including the mentum 2.8 cm. long, about 5 mm. wide, oblong, gradually tapering from above the middle to an acute apiculate tip, the nerves somewhat raised on the outer surface in dried specimens, the central one subcarinate near the apex; mentum 8.5 mm. long, slender, 4 mm. in diameter, tapering gradually to the tip where it is about 1 mm. in diameter, closed in front for about one- half of its length. Dorsal sepal 2 cm. long, 5 mm. wide, elliptic-oblong, tapering to an apiculate tip, narrowed toward the base, mid-nerve lightly carinate toward the distal end. Petals about 1.9 cm. long, 7 mm. wide across the upper third, about 2 mm. wide near the base, oblanceo- late, rounded at the apex, shortly apiculate, 5-nerved. Labellum about 2 cm. long, about one-third shorter than the lateral sepals, oblanceo- late, apiculate, about 6 mm. wide above the middle, the three central nerves more or less prominent, the middle one subcarinate toward the base of the disc; 6 mm. from the base there is a fleshy transverse callus. Column 9 mm. long; clinandrium denticulate. Santa Cruz Group: Vanikoro, in the moss on rain- forest trees, at 800 meters altitude, S. F. Kajewski, no. 604 (type in Herb. Ames, no. 38084), November 11, 1928 (flowers, some cream with yellow-edged labellum, others are purple-pink, but all have the yellow-edged labellum). 1933] AMES, ORCHIDS OF THE NEW HEBRIDES 109 Apparently a close ally of D. asperifolium J. J. Smith, but the leaves and flowers are not verrucose. In habit similar to D. pentapterum Schltr. and D. Cuthbertsonii F. v. Muell. The specific name alludes to the tendency of the flowers to be hidden by the moss in which the plants were growing. Dendrobium purpureum Roxburgh, Fl. Ind. 11. 484 (1832). Dendrobium Morrtsonii Schlechter in Bull. Herb. Boiss. ser. 2, vt. 456 (1906). Aneityum: Hill northeast of Anelgauhat, Dr. Morrison, s. n., June 18, 1896; near Anumy, Dr. Morrison, s. n., June 1896; hills be- tween Ithumu and Anelgauhat, Dr. Morrison, s. n., July 6, 1896 (flow- ers white) —Also Moluccas. Dendrobium Quaifei Rolfe, ined. Dendrobium Qnaifei Guillaumin in Bull. Soc. Bot. France, -x1v. 707 (1927), sphalm. New Hebrides: Santo Peak, 4500 feet altitude, W. T. Quaife, s.n., May 1903. The narrowly oblanceolate petals and the dissimilar stelidia of the column separate this species from D. pseudo-Tokai Kraenzl. Appar- ently it is closely allied to D. montis-yulei Kraenzl., but it differs suffi- ciently in the flowers to be regarded as distinct. In habit it is very similar to D, Mooreanum Lindl., but from that species it differs con- spicuously in having much smaller flowers and a different labellum. The type consists of a single plant: Caules plus minusve 28 cm. longi, verisimiliter caespitosi, usque ad apicem dilatati, 6 mm. in crassi- tudine, valde sulcati, flavidi, ad apicem triphylli. Folia plus minusve 9 cm. longa, usque ad 3.3 cm. lata, valde coriacea, apice bidentata. Inflorescentia 20 cm. longa; pedunculus infra racemum plus minusve 11 cm. longus, paucibracteatus. Racemus plus minusve sex-florus, 9 cm. longus. Bracteae racemi vix 5 mm. longae, lanceolatae, concavi- usculae, acutae. Pedicelli ascendentes, graciles, plus minusve 3 cm. longi. Sepala lateralia 2 cm. longa, 5 mm. lata, triangulari-lanceolata, acuminata, acuta, nervo medio prominenti. Mentum 8 mm. longum, curvatum, obtusum. Sepalum dorsale simile. Petala 2.8 cm. longa, 7 mm. lata, membranacea, anguste oblanceolata, quinquenervia, obtusa. Labellum plus minusve 1.9 cm. longum, prope medium 1 cm. latum, trilobatum, lobis lateralibus rotundatis, lobo medio 9 mm. longo, 5 mm. lato, triangulari-lanceolato, acuto. Discus supra medium glaber, infra medium callo bisulcato ornatus. Columna generis, stelidiis recurvatis, acutis, terminalibus instructa. Dendrobium ruginosum Ames, sp. nov. Caules caespitosi, inferne graciles, prope apicem in pseudobulbum 110 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV longitudinaliter ruginosum flavidum producti. Folia ad apicem pseudo- bulbi congesta, coriacea, elliptico-oblonga vel lanceolata, acuta. In- florescentia plus minusve quinqueflora. Sepala lateralia mentum obtusum formantia, oblonga, usque ad apicem attenuata, acuta, extus carinata. Sepalum dorsale lanceolatum, apiculatum. Petala prope basim valde angustata, membranacea, supra basim suborbicularia, acuta. Labellum trilobatum. Discus glaber, callo elongato truncato ornatus. Columna generis. Roots glabrous, whitish, coarsely fibrous. Stems caespitose, 25 cm. tall, seven-jointed; the uppermost internode conspicuously swollen and pseudobulbous, yellow, 7 cm. long, up to 1 cm. in diameter when dry, longitudinally sulcate, tapering to a slender base; the remaining inter- nodes slender, 3.7-4.4 cm. long, 2-3 mm. in diameter, smooth, brownish or yellowish, with the fibrous remains of tubular sheaths persisting at the nodes. Leaves terminal on the pseudobulb, coriaceous, three in number, approximate, elliptic-oblong to oblong-lanceolate, tapering gradually toward the tip, 7.5-8.5 cm. long, 1.7-2 cm. wide, obliquely ascending. Inflorescence longer than the leaves, up to 11 cm. long, rigid, slender, about five-flowered, arising from the axil of a leaf or breaking out from the summit of the pseudobulb below the lowermost leaf. Bracts of the inflorescence about 4 mm. long, lanceolate, much shorter than the pedicels, concave or cymbiform, acute. Pedicels slender, with the verruculose ovary 1.7 cm. long, curving sharply toward the upper end. Flowers rather showy with the sepals and petals white and the labellum pale green with purple stripes. Lateral sepals membranaceous, forming a rounded mentum, including the mentum 2 cm,-long, about 5 mm. wide across the middle, oblong, tapering from about the middle toward the acute thickened apex, heavily carinate on the exterior surface along the median nerve, the carina extending beyond the apex of the sepal in a sharp apicule. Mentum 5 mm. long. Dorsal sepal 1.7 cm. long, 7 mm. wide across the middle, lanceolate, apiculate, 5-nerved, lightly carinate on the outer surface. Petals mem- branaceous, 2.4 cm. long, 1.3-1.5 cm. wide, cuneate at the base, dilated upward into a suborbicular or subrhombic lamina, lightly retuse at the apex with an apicule in the sinus or simply acute, 5-nerved at the base. Labellum about 2 cm. long, 3-lobed; lateral lobes rounded, 1.2 cm. long from the rounded tips to the point of insertion at the base of the labellum, about 5 mm. wide; middle lobe about 7 mm. long, 7 mm. wide, apiculate. Disc smooth, with a heavy truncate median keel on the inner surface which extends to the middle of the labellum. Column about 9 mm. long. Santa Cruz Group: Vanikoro, in rain-forest at 800 1933] AMES, ORCHIDS OF THE NEW HEBRIDES 111 meters altitude, S. F. Kajewski, no. 606 (type in Herb. Ames, no. 39056), November 11, 1928 (orchid growing on rain-forest trees; petals white, labellum pale yellow with purple stripes; only two specimens seen). This species is allied to D. atroviolaceum Rolfe, but varies from it in having smaller, differently colored flowers and differently formed petals. The three-lobed labellum serves to distinguish it from D. Mooreanum Lindl. and the very dissimilar petals differentiate it from D. Quaifei Rolfe, its closest ally in the New Hebrides and Santa Cruz Islands. Glomera Macdonaldii (Schitr.) Ames, comb. nov, Glossorrhyncha Macdonaldii Schlechter in Fedde, Rep. Spec. Nov. 111. 19 (1906). Eromanga: Traitor’s Head, summit of old crater on mossy trunk of Metrosideros, 2400 feet altitude, L. Cheeseman; no. 70, August 21, 1930 (straggling semi-erect branches, flower white) —Already re- ported from Aneityum. The type was collected by Macdonald on Aneityum. In the Kew Herbarium there is a sterile specimen collected by Dr. Morrison on Aneityum, between Anelgauhat and Anumy on June 25, 1896. Vegeta- tively this specimen is very similar to the one from Eromanga collected by Cheeseman. Phajus amboinensis Blume in Mus. Bot. Lugd.-Bat. 1m. 180 (1856). Aneityum: Dr. Morrison, s. n., June 22, 1896.—Also Java and Amboina. This is a small flowered form of the species. Eulophia macrostachya Lindley, Gen. & Sp. Orch. 183 (1833). Aneityum: Anelgauhat, on ground under trees, Dr. Morrison, s.n., June 9, 1896. Eromanga: Terrestrial, Dr. Morrison, s. n., July 18, 1896—Also Ceylon, Java, Sumatra, Dutch New Guinea and Borneo. The specimens on which my determination is based have somewhat smaller flowers than usual. Vegetatively the specimens cited above approach EF. novo-ebudae Kraenzlin in Guillaumin in Bull. Soc. Bot. France, sér. 5, v. 301 (1929), but the labellum is very different if Kraenzlin’s description of admittedly poor material is dependable. Phreatia calcarata J. J. Smith in Bull. Dép. Agric. Indes. Néerl. no. 19, p. 31 (1908). Banks Islands: Vanua Lava, rain-forest, 3120 feet alti- tude, L. Cheeseman, s. n., October 29, 1929 (flowers white, growing on rotten bole).—Also Papua. 112 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Thrixspermum sp. Eromanga: Dillon’s Bay, Dr. Morrison, s. n., August 5, 1896. There are three specimens, all without flowers, although one is said to have had pale yellow flowers. Vegetatively this species resembles Thrixspermum Vanoverberghii Ames, a native of the Philippine Islands. ARNOLD ARBORETUM, HarvARD UNIVERSITY. 1933] WHITAKER, CHROMOSOMES OF CYPHOMANDRA 113 THE CHROMOSOME COMPLEMENT OF CYPHOMANDRA BETACEA THoMAs W. WHITAKER With six text figures The genus Cyphomandra contains thirty or more species of herbs, shrubs, and small trees, all of which are of South American origin (Bailey, 1925). It is technically distinguished from Solanum by the fact that the two cells of the anther are separated by a thickened con- nective, which appears as a ridge on the back of the anther. Cyphomandra betacea (Cav.) Sendt., the so-called “Tree Tomato,” is cultivated, to a certain extent, by the natives of the American tropics for its edible fruit. It was first described in 1801 under the name Solanum betaceum Cavanilles. Later it was placed in the genus Cy- phomandra by Sendtner (1845). In the course of an investigation of the chromosome number and behavior of various members of the Solanaceae, several plants of C. betacea were made available to the writer through the courtesy of Pro- fessor Karl Sax. A preliminary examination indicated that the meiotic divisions were irregular, and, for this reason, it was thought that a further cytological examination would contribute some information on the cause and effect of the irregularities noted at meiosis. Attention was focused on three points: ( 1) A description of the nature of the irregularities taking place during meiosis; (2) the effect of these irregularities on the amount of pollen sterility; (3) the probable taxonomic relations of C. betacea from a study of its cytology. Observations on the meiotic divisions were made from aceto-carmine smear preparations. Preparations showing microspore divisions are, as a rule, unsatisfactory when prepared by the usual methods, because the pollen grains are thick and opaque. By fixing with aceto-carmine and exerting considerable pressure on the cover slip, combined with gentle heating of the slide, excellent preparations were secured. C. betacea has twelve pairs of chromosomes. The haploid comple- ment is shown particularly well at metaphase of the microspore divi- sion (Fig. 1). It is apparent that the attachment constrictions are approximately median in all cases, and there are no conspicuous mor- phological variations among the individuals of the haploid set. In root tip preparations, twenty-four chromosomes are plainly discernible. The somatic chromosomes are characteristically long and slender. Each 114 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Figure 1. Tue 12 GAMetTic CHROMOSOMES IN THE FIRST DIVISION oF THE MICROSPORE. — FiGuRE 2. THE 24 SOMATIC CHROMOSOMES AT METAPHASE. X 2100. Ficure 3. THe 12 BIvALENT CHROMOSOMES AT METAPHASE IN THE First DIVISION OF THE POLLEN MoTHER CELL. — Ficure 4. First META- PHASE SHOWING ONE UNIVALENT AT EACH POLE, THE REMAINDER AT THE Equator. — Figure 5. SHOWS THE SAME SITUATION AS IN FIGURE 4 BUT WITH Two BIVALENTS SHOWING Precocious Divis1on. — FIGURE 6. SHows A LAGGING PAIR OF CHROMOSOMES. X . 1933] WHITAKER, CHROMOSOMES OF CYPHOMANDRA 115 chromosome has a median attachment constriction, and several have secondary constrictions (Fig. 2). Good figures of diakinesis were difficult to obtain, owing perhaps to the unusual amount of chromatin contained in the relatively small nucleus, but where satisfactory figures were examined, it is clear that two of the bivalents have only one terminal chiasma. The remaining bivalents have at least two chiasmata and, in certain cases, aS many as three chiasmata. The structure of the individual chromosomes is very clear at meta- phase of the first meiotic division, and the spiral chromonemata are easily visible. At this point several abnormalities were observed to interrupt the ordinary sequence of the meiotic process. Briefly these abnormalities were: (1) One and sometimes two of the univalents were located at each pole, while the remainder of the set were still at the equator (Fig. 4); (2) two pairs of chromosomes frequently showed precocious division (Fig. 5); (3) lagging of one pair of chromosomes was repeatedly observed (Fig. 6). It is estimated that these abnor- malities occurred in about 50% of the cases. The irregularities observed at meiosis are reflected in the pollen, since fully 25% of the pollen grains are aborted. The percentage of aborted pollen in some flowers is much greater, reaching an extreme limit of 50% pollen sterility. DISCUSSION On comparing the size of the metaphase chromosomes of C. betacea with those of Solanum Capsicastrum (arbitrarily selecting S. Capsi- castrum as a representative of the other Solanaceae), it was found that the chromosomes of C. betacea averaged slightly over twice as long as those of S. Capsicastrum in similar preparations at the same stage. This same relationship existed regarding the width of the chromosomes. The diameters of the pollen mother cells of C. betacea and S. Capsi- castrum stand in the ratio of 1:1.33. Thus the volume of the pollen mother cells of C. betacea is approximately three times that of S. Capsicastrum. Vilmorin and Simonet (1928) have reported the chromosome number of a considerable range of solanaceous plants. In their studies they Aside from the striking difference in the size of the chromosomes and pollen mother cells, C. betacea is very similar, cytologically, to other 116 JOURNAL OF THE ARNOLD ARBORETUM [voL. XIV members of the Solanaceae. The closely related genera Salpichroa, Solanum, and Lycopersicum all have a basic number of twelve chromo- somes and ring bivalents. In this connection it is interesting to note that although the individual chromosomes of C. betacea are consider- ably larger, both in width and in length than those of other species of Solanaceae, the type of association is identical with that found in species with smaller chromosomes. The amount of pollen sterility (25%) found in C. betacea is quite unusual in a pure species, although by no means exceptional. Professor Jack, of the Arnold Arboretum, informs me that C. betacea is con- sidered a good species and that it is quite unlikely that it could have hybridized with any other member of the genus, as it is very distinct from other species of Cyphomandra. Therefore the irregularities at meiosis and high pollen sterility cannot be accounted for by assuming that the plants under consideration were a direct result of hybridization with a closely related species. It has been suggested that the irregularities under consideration might be due to the fact that the plants were grown under a lower temperature than that to which they are accustomed in their native habitat. Acting on this suggestion, the writer has examined pollen of plants from Cuba. (This pollen was supplied through the kindly cooperation of Mr. R. M. Grey of the Atkins Institution of the Arnold Arboretum.) Pollen sterility of the Cuban plants averages approxi- mately 5%, compared with an average sterility of 25% of plants flow- ering at the Bussey Institution. It appears as if this considerable difference in pollen sterility is significant, particularly since the pollen sterility was as high as 50% in a few cases of the plants grown under different environmental conditions. There is good evidence that meiosis in plants is, in some measure, disturbed by lowering of the temperature. Belling (1925) first sug- gested that pairing of the chromosomes might be more easily disturbed by cold in tropical or subtropical plants than would be the case with plants of temperate regions. He has also suggested that cold treatment might be an important means of inducing tetraploidy in tropical plants. Sax (1931) kept plants of Rhoeo discolor at about 45 degrees Fahren- heit for several days. The pollen mother cells of these plants showed a complete lack of pairing of the chromosomes at the first meiotic divi- sion. When the cold treated plants were again placed under normal conditions, the ordinary meiotic behavior was restored. Since the plants of C. betacea were not critically tested in a similar manner to find out if the temperature was the factor responsible for the 1933] WHITAKER, CHROMOSOMES OF CYPHOMANDRA 117 pollen sterility observed, it is impossible to state with certainty that temperature was the environmental variable. However, what little evidence available is highly suggestive and indi- cates that the irregularities during meiosis are directly traceable to the low temperatures prevailing when the plants were examined. The manner in which temperature seems to operate, in this case, is by reducing the number of chiasmata in certain chromosomes, which re- sults in very loose pairing at diakinesis. This naturally leads to speculation in regard to the way in which plants cope with conditions of decreasing temperature. From the evi- dence at hand, it seems as though the answer to this question may be that decreasing temperature causes greater variability, which, in turn, increases the chances of a mutant’s being produced that would be better able to survive under the new set of conditions, SUMMARY 1. Cyphomandra betacea has twelve pairs of chromosomes. The chromosomes of this species are considerably larger than those of any solanaceous plant yet reported. The irregularities which occur during meiosis are discussed, and evidence is brought forth to indicate that these irregularities were due to the temperature conditions prevailing at the time the plants were in flower. 3. Pollen sterility averaging about 25% has been found in the plants examined. This value, when compared with 5% pollen sterility found in plants grown in Cuba, is considered significant and may be attributed to the difference in temperature under which the plants were grown. 4. The indications are that temperature operates on the meiotic division by reducing the number of chiasmata per bivalent in certain chromosomes. CyToLocicaL LaBoratory, ARNOLD ARBORETUM, HARVARD UNIVERSITY. LITERATURE CITED BarLey, L. H. (1925). Manual of cultivated plants. pp. 851. (Mac- + millan Co. New York. BELLING, JOHN (1925). Production of triploid and tetraploid plants. (Jour. Hered. 16 :463-464. ) Sax, Kart (1931). Chromosome ring formation in Rhoeo discolor. oo: SENDTNER, QO. (1845). Monographia Cyphomandrae, novi Solana- cearum generis. (Flora 28:161-176.) VILMORIN, ROGER pE & Marc eects: (1928). Le nombre des chromosomes chez les Solanées. (Verh. V. Internat. Kongr. Vererb. Wiss. Berlin, 1520-1536.) 118 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XIV STUDIES ON THE PRECIPITIN REACTION IN PLANTS III. VII. Ill. A BIOCHEMICAL ANALYSIS OF THE “NORMAL PRECIPITIN REACTION”? Kennetu S. CHESTER AND THomMas W. WHITAKER With four text figures CONTENTS INTRODUCTION ..ccc ccc cece eect e cence eer eeenesrenseeeneees 119 MATERIALS AND METHODS .......ee eee eeeee terre nreentees 120 NATURE OF THE REACTIONS OCCURRING BETWEEN PRUNUS, PLataNus, ROBINIA, AND RIBES ......eee eee eset neers 5 A. Effect of salt concentration on the reactions .........+-- 125 B. Effect of hydrogen ion concentration on the reactions .... 127 C. Effect of heating on the reactions ..........-.eee eee 130 D. Effect of dilution on the reactions .......-.. 00ers ees 130 E. Effect of dialysis on the reactions ......-....eeee+e eee 135 F. Effect of various solvents on the reactions ........-++-+: 144 G. Effect of carbohydrate removal on the reactions ......... 151 H. Effect of enzyme digestion, contamination, and ageing on the TFOACHONS “6si%s ocak she uese nae daa eR ee ONE 153 I. Effect of double-precipitation on the reactions .......-.. 155 PRooF OF THE CALCIUM OXALATE REACTION ....-+.--0 0005 158 PROOF OF THE EXISTENCE OF OTHER REACTIONS ...-++++-++++ 164 PROPERTIES OF THE AB, MN, Anp XY REACTIONS ......-5+: 170 A. Heat experimen re ee eS ee wee ee 170 B. Experiments in extraction and solution .......-+.+++++- 173 C. Experiments in dialysis ...........00 eee eee teens 175 D. Nature of the AB, MN, and XY reactions ...........-- 178 THEORETICAL AND PRACTICAL SIGNIFICANCE OF THIS STUDY .. 179 a of Kostoff’s results in the light of the present 1 ge ee tees oo a ee SURO a Oe Se ER OS Oe eee ees we eee en stu C. poaant er East’s results in the light of the present study .. 183 D. — of Chester’s results in the light of the present 18 L kulirk Eee te REP EA od aes epee Eee Ree EES 92 F. Direction of further studies in this field ...............4. 193 UCUAREP ADS: oa fen anes & REM ESSE SS HAG OS SEDER TERENAS ES 193 ACKNOWLEDGEMENT ....e eee e ee eee ee eee teen ee enees 196 LITERATURE. CITED .cccc 0000 cc cece ee see tee ence eee sabe eas 196 1For parts I and II see Vol. XIII. 52-80 and 285-296. 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 119 I. INTRODUCTION Thirty years ago Marshall Ward concluded his researches on the parasitism of fungi (22) with the words: “Immunity depends entirely upon the physiological reactions of the protoplasm of the fungus and the cells of the host. In other words, infection and resistance to infection depend on the power of the fungus protoplasm to overcome the resistance of the cells of the host by means of enzymes or toxins, and reciprocally on that of the protoplasm of the cells of the host to form anti-bodies which destroy such enzymes or toxins.” This state- ment, at the time little more than a shrewd conjecture, inaugurated an epoch of new conceptions in phytopathology and physiology. Soon there appeared the classic studies of Bernard in which it was shown that the symbiosis between the orchid and its mycorrhizal partner is not the static, passive equilibrium which it had heretofore been con- sidered, but that it is an active, dynamic parasitism in which the orchid host controls through the mechanism of specific internal prophy- laxis the erstwhile parasitic invader, that it is in the words of Giumann “a parasitism with the brakes on” (“ein abgebremster Parasitismus’’) (9). Long before this time, however, the conception of a dynamic internal prophylaxis had served to inspire fruitful researches in human medi- cine. Ward’s analogy was but an echo of the doctrine of acquired immunity in animals. It was thus necessarily resultant that the un- experienced phyto-immunologist should look for guidance to his more experienced colleagues in the field of human immunology. Phyto- immunology has assuredly profited from this new application of medical experience, but there is a grave danger in attempting to transplant in toto the techniques and conclusions of one branch of science to the very different soil of a kindred science. There is an unavoidable tendency to interpret analogies as homologies, failing, however, to consider the in- herent differences which characterize the materials of the two sciences. As early as 1902 (11, 16) phenomena in plants resembling immuno- logical behavior in animals had been observed. Plant saps were found to act toward animal proteins in a fashion resembling that of immune blood in the presence of its specific antigen; they agglutinated and hemolyzed blood corpuscles and precipitated foreign proteins. In suc- ceeding years the literature has become replete with similar observa- tions and in almost all cases the terminology and interpretations of human medicine have been bodily appropriated in the establishment of the science of plant immunity. As the most striking and pertinent example of such non-critical appropriation of zooimmunitary concep- 120 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV tions must be cited the researches dealing with the reactions of grafted plants. The work of Kostoff (12) has already been discussed in this series of studies. It is sufficient to state that this worker found in studying the Solanaceae that aqueous extracts of certain plants precipi- tated in vitro in the presence of other such extracts, that after grafting two such plants together the intensity of the precipitation sometimes increased, that in certain cases the overlayering of two extracts resulted in a clear zone at the plane of contact, that the plastids of grafted plants sometimes showed a clumping together in masses, and that various malformations and abnormalities accompanied these other phenomena. All of this, without reserve, was interpreted in the terms of antibody formation: the precipitating action was due to “precipi- tins” (which term has a very specific meaning in zodimmunity); the clear zones resulted from the activity of “lysins”; the clumping of the plastids was an “agglutination.” Kostoff’s work has been accepted as 4 most conclusive link in the chain of proof that plants may elaborate antibodies of the zodimmunitary type (2, 8). Indeed one worker, Silberschmidt (21), has published a long and painstaking critique of Kostoff’s work suggesting numerous modifications based entirely upon the conception that the reactions observed are of the nature of immuno- logical reactions in animals, although Silberschmidt recognizes in the display of reactions certain differences from the reactions of animal immunology. It is manifest that if the reactions observed in the Solanaceae are homologous with the zodimmunological reactions then a most important step has been taken in the demonstration of antibody formation in plants. If, on the other hand, the reactions observed in the Solanaceae are only analogous to those in serology, if they are not characterized by the same properties and are not of the same significance as are the re- actions in animal immunology, then it is of paramount importance to point out their true significance, be it immunological or merely bio- chemical in its more limited sense. For this reason the present study was undertaken, in order to determine the nature of the reactions here- tofore described as immunological. A short preliminary account of this analysis has already been published (5), and it is the purpose of the present study to substantiate the preliminary account with experi- mental data. Il. MATERIALS AND METHODS The plants employed in this study were of species from a variety of herbaceous and woody families; they are enumerated below: LecumMinosaE: Robinia fertilis Ashe. 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 121 OLEACEAE: Ligustrum ibota Sieb. & Zucc., L. obtusifolium Sieb. & Zucc., L. vulgare L.; Syringa vulgaris L. PLATANACEAE: Platanus acerifolia Willd. RosacEAE: Prunus Armeniaca L. var. ansu Maxim. and var. “Mikado” Hort. SAXIFRAGACEAE: Ribes Carrierei Schneid. (R. glutinosum albidum Janez. X R. nigrum L.); Hydrangea paniculata Sieb. var. grandiflora ieb. ie SOLANACEAE: Atropa Belladonna L.; Browallia viscosa HBK.; Capsi- cum frutescens L.; Cyphomandra betacea Sendt.; Datura ferox L., D. innoxia Mill., D. metel L., D. Wrightit Hort. = D. meteloides DC.) ; yeapeesien cerasiforme Dun.; Nicotiana alata Lk. & Otto, N. acumi- nata Grah., N. Cavanillesii Dun., N. glauca Grah., N. glutinosa L., N. Langsdorffii Weinm., N. nudicaulis Hort., N. Palmeri Gray, N. paniculata L., N. plumbaginifolia Viv., N. Rusbyi Britt., es rustica L., NV. Sanderae Hort. Sand. (CN. alata x Forgetiana), N. suaveolens Lehm., NV. sylvestris Speg. & Comes, N. Tabacum L., x tomentosa R. & P., N. trigonophylla Dun. ; Petunia zlolacea Lindl.; Physalis peru- wiana L.; Salpiglossis sinuata Ruiz & Pav.; Solanian Capsicastrum Link, S. Melongena L., S. nigrum L., S. tuberosum L. The Solanaceae enumerated were greenhouse plants for the use of which the writers are indebted to Professor E. M. East. The woody plants were mature specimens in the ornamental collection of the Arnold Arboretum. Leaves from the woody species were collected on Sept. 2, 1931, dried at room temperature for one month, and then pul- verized and stored in wide-mouthed glass bottles until used. The sola- naceous leaves were collected as needed. In nearly all cases reported below the solanaceous extracts were made from leaves which had been dried in a dessicating oven at 55°C. for from one to three days and then pulverized. The method of extraction varied according to the requirements of the various experiments. It is necessary to distinguish here between pre-extraction and solution. The procedure of pre-extraction and its significance in protein reactions has been thoroughly discussed by Silberschmidt (21). Briefly the purpose of a pre-extraction process is to remove from the tissues substances, chiefly lipoids, which give non- protein reactions and hence tend to render the readings inaccurate. Pre-extraction usually involves extraction with strong alcohol, ether, and chloroform. Such pre-extraction was employed in certain of these experiments to determine its effect upon the resulting reactions (see pages 000 and 000), but was not used in the remainder of the experi- ments because it was the purpose of these experiments to determine the reactive constituents of extracts of the type used by Kostoff, in which no pre-extraction was employed. 122 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV The leaf powders or pulps were then extracted in the solvent designed to dissolve the reactive principles. In most of the experiments reported below distilled water was used as a solvent. Physiological NaCl solu- tion was used in some cases in place of the water, particularly in certain of the dialysis experiments in which it was desirable to avoid the auto- precipitation of extracts due to the withdrawal of electrolytes. Here, too, in certain experiments to determine the relative value of various solvents, a selection of several processes was employed, but, unless otherwise stated in the experiments to follow, the reader may under- stand that distilled water was used as a solvent. The concentration customarily employed was one part by weight of fresh leaf pulp to two parts of solvent or one part dried leaf powder to ten parts of sol- vent. This concentration is here referred to as normal concentration (N) and further dilution is indicated by the appropriate fraction (N/2, etc.). Extraction of reactive substances occupied two hours at room temperature or 18 hrs. at 2°C. in all cases, it having been previously ascertained that there is no significant difference in the reactivity of extracts prepared under such conditions. The clarification of the extracts was accomplished by various means depending upon the difficulty in obtaining clear solutions. The dried leaves yielded clear extracts much more readily than fresh leaves, and on the whole little difficulty was experienced in obtaining crystal-clear extracts. The customary procedure was to extract in an open beaker or large test-tube, stirring from time to time, to pour the pulp and extract into a funnel lined with a fine filter paper (C. S. & S. #589) which in turn was lined with sterile gauze. The gauze was then drawn together to form a bag, the liquid squeezed out into the filter paper, and the clear filtrate used in testing. In a few cases it was necessary to repeat filtration through one or two other papers. When dealing with a very small quantity of tissue (we have been able successfully to extract .1 gm. of tissue in 1 cc. of water) it was customary to centri- fuge, rather than filter, drawing the clear supernatant liquid off into another tube with a fine pipette. In only a few cases was it impossible to obtain clear extracts. Salpiglossis, Browallia, and Hydrangea gave the most trouble in this connection, and when it was necessary to use extracts from such species the procedure employed was to catch the filtrate immediately it came through the filter and to test it at once. In any case these extracts were used in very few of the experiments reported below. Nicotiana Rusbyi proved to be rather refractory in aqueous solution but where physiological saline was used as a solvent, the extracts were satisfactory. On standing for several days, even at 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 123 2°C., there was frequently an autoprecipitation of the extracts, but in no case of the many tested did the subsequently cleared extracts exhibit an altered reactivity. Extracts which were to be preserved for any appreciable amount of time ‘were covered with 4”-'4” of toluol and kept in the cold. Bacterial or fungus contamination rarely occurred and then only in extracts which had stood for a long period in the cold. Although it was ascertained that such contamination did not affect the reactivity of the extracts (see page 000) such extracts were not used in the tests reported below unless specifically so stated. For testing, the procedure described in an earlier paper (4) was employed. With a fine pipette .2 cc. of the liquid of greater density was introduced into the usual small test tubes and over this was layered an equal quantity of the second liquid to be tested. The length of time of reading varied with the requirements of the experiments from 5 to 40 minutes. Where gross results were desired a duration of 5-10 min- utes was satisfactory. In fact the reactions observed were usually so marked that a maximum reading could be made in a very few minutes, while a continuation of the time merely permitted the precipitates to diffuse through the extracts making readings less accurate. In a few of the experiments where only gross differences were of importance, the tests were performed by mixing equal quantities of the two ex- tracts in 3” test tubes and observing the resulting cloudiness in com- parison with the unmixed liquids. Throughout this paper, in any case, each experiment is a unit, performed under homogeneous conditions and for the purpose of demonstrating one point. The whole series of experiments were not devised for numerical comparison with each other, and accordingly the demands of scientific caution have been met by performing each experiment under its own set of standard condi- tions. The numerical value assigned to any given reading was accord- ing to one of two scales. The first scale, and that employed in nearly all cases, is that described and figured in an earlier paper of this series (4) in which a trace of a reaction (t) is succeeded by reactions of strength 1, 2, 3, and 4 (maximum ever observed). A more accurate measure of reactivity which was employed in a number of cases was to calculate the numerical sum of the readings (according to the first scale) at 1, 5, 10, 20, 30, and 40 minutes respectively. To be sure neither scale is wholly free from error, and both are arbitrary measures: but it is the belief of the writers that the readings within the limits of these experiments are accurate within the range of 1 unit on the t-1-2-3-4 scale and that a greater accuracy of reading is not compatible within the inherent variability of the material and technique. All of 124 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV the readings reported in this paper were made by the same observer, a second observer being consulted in case of doubt. Numerous controls were employed in each experiment. The ex- tracts were tested against the pure solvents and against such other fluids as the individual experiments required. Protein determinations by the use of the Millon and xanthoproteic tests were resorted to, in general, only in the dialysis experiments and then only for the purpose of obtaining a measure of the efficiency of the dialyzing membranes. Silberschmidt (21) feels that such deter- minations should be made on all extracts. When it is possible to say with assurance that one is dealing with protein reactions of the zodimmunitary type, then the protein tests will be of value, but since the aim of the present study was to determine the nature, protein or otherwise, of the reactive substances, it was not felt necessary to employ protein tests in all cases. Similarly other tests (for chloride, oxalate, calcium, etc.) were employed only as the specific experiments required them. Since a brief summary of most of these results has already been published (5) it is possible to state at the outset the sequence of points which are to be considered in the present paper. It was pointed out in this earlier summary that a reaction responsible for the majority of positive tests which have been observed is the reaction between calcium and oxalate ions in two respective extracts. It then being possible to eliminate from discussion the calcium oxalate reaction, attention may be focussed on the remaining reactions. Having ascertained that there are at least three other such reactions, designated here respectively as the AB, MN, and XY reactions, and their respective properties having been determined, it is finally possible to analyze the experimental data of Kostoff, Silberschmidt, East and Chester with respect to the role of these various reactions in the published accounts of the “normal pre- cipitin reaction,’ and to point out the significance, immunological or otherwise, of the precipitin reaction in plants. Hence the remainder of this paper will concern itself with answering the following questions: 1. What is the nature of the reactions occurring between Platanus, Ro- binia, Ribes, and Prunus? 2. What is the evidence supporting the statement that a calcium oxalate reaction is responsible for the majority of the recorded positive precipitations? 3. What proof is there as to the presence and number of other reactions in the plants studied? 4. What are the properties of these additional reactions? 5. What light does this study cast upon the previously recorded studies of the pre- cipitin reaction in plants? 6. What is the theoretical and _ practical significance of the precipitin reaction in plants? 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 125 III. THE NATURE OF THE REACTIONS OCCURRING BETWEEN PRUNUS, PLATANUS, ROBINIA, AND RIBES Early in this study the question was considered whether or not the precipitations observed were due to inorganic constituents. Attention was diverted from the theory that they might be explainable in such simple terms because of the following facts: (a) the specificity exhib- ited by the reaction (4) appeared hardly in conformity with such a simple explanation; (b) accepting Kostoff’s work at face value it would be inconceivable from the complexity of his results and their apparent immunological significance that he was dealing with such a simple reaction; (c) when in one preliminary experiment extracts of Platanus and Prunus were ignited at red heat and then the ash redissolved in water, the reaction had disappeared. The influence of the immuno- logical school of thought was so strong that the writers as well as their various associates were strongly inclined to believe the reactions highly complex in nature. The series of experiments devised accordingly em- phasized the proof of the presence or absence of protein in the reactive fractions of the extracts. With the elimination of protein as a reactive principle, attention was directed to the other complex chemical con- stituents of the extracts, and not until most of the major chemical groups had been eliminated was attention forcibly drawn to the func- tion of the inorganic constituents. The evidence is accordingly of two sorts, negative evidence that the proteins and various other chemical groups could not be responsible for the reactions, and positive evidence that calcium-oxalate was responsible. The present section will treat with the first of these subjects, i. e. the negative evidence, while the following section will deal with the proof of the calcium oxalate reaction. The first experiments performed were with the species of woody plants enumerated above, but particularly with Prunus Armeniaca “Mikado,” Platanus acerifolia, Ribes Carrierei, and Robinia fertilis. Later the results were extended to include the other species listed above. Of the extracts employed the Prunus tests strongly against Platanus, Robinia and Ribes, while the latter three are negative to each other. A. EFFect oF SALT CONCENTRATION ON THE REACTIONS As a primary step in the investigation of the nature of these reactions it was felt desirable to study the effect of certain physical factors on the reaction, in order to determine the optimum conditions for reaction, to observe the specific effect of variation of such factors, and to obtain a standard technique for subsequent work. The following experiment 126 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV was therefore first set up for the purpose of studying the effect of the variation of the electrolytic concentration in the extracts. Five samples of dry leaf powder of Platanus, Ribes, and Robinia respectively were extracted in the usual manner in ten times the weight of solvent. For solvents were used a series of phosphate buffer solu- tions prepared according to the system worked out by Cohn (6). All the buffers had the same pH, namely 6.0 (which is relatively close to the normal pH of these extracts). On the other hand the five buffers of each of the three series had salt concentrations respectively of .06M, 12M, .30M, .60M, and 1.2M. This system of buffers has the advan- tage of introducing only one metallic ion [KH,PO, (acidic) : K,HPO, (basic) |]. The fifteen extracts were then cleared and tested against one TABLE I. EFFECT ON THE PLATANUS-ROBINIA-PRUNUS PRECIPITIN REACTION OF VARYING THE SALT CONCENTRATION. Hobinia...Phosphate content (Oohn) Hdatarus. | Phosphate content(Cohn) 206M s12M 30M -G6O0M1.2M Total 106M 12M »5OM »60M 1,2M Total 206M 15.0 5.0 5.5 300 2.5 3heO0 15.0 14.5 9.5 4.0 0.0 43.0 Prams 212 M 11.0 9.5 10.0 9.5 4.5 44.5 17.0 15.0 8.0 5.5 2.0 47.5 phosphate content ,30M 11.0 11.0 6.0 7.0 5.5 405 16.0 15.0 7.5 4.0 3.0 4545 260M 11.0 10.5 9.6 4.0 3.0 38,0 17.0 15.0 9.5 2.0 0.5 440 22200 15.0 9.5 4.5 25 25 34.0 15.0 12.5 4.5 0.0 0.0 32:0 Totals 65.0 458 35:5 260 18:0 BQ.0 72:0 89:0 15.65 Bed another. The results of these tests are given in Table I. The numeri- cal values assigned to the various reactions were computed by summat- ing the strengths of reaction at the various time intervals from 1 to 40 minutes. For control experiments the various extracts of Robinia were tested against one another in all possible combinations; the same was done with the extracts of Prunus and Platanus. All extracts were tested against the various strengths of pure buffer, and finally the Robinia series was tested against the Platanus series. All of these con- trol tests were completely negative. Several points are brought out by this table. In the first place one observes that although Robinia and Platanus behave in the same fashion with regard to salt concentration, both differ in marked fashion from Prunus. Variation of salt content through this wide range hardly affects Prunus while Robinia and Platanus show a steady decline in reactivity as the strength of salt increases. This fact would indicate that the reactive substances in the two types of extract are different, a 1933 | CHESTER AND WHITAKER, PRECIPITIN REACTION 127 situation not in conformity with that in animal immunology. Since the higher concentrations of salts used in this experiment are far greater than occur in ordinary water or saline extracts, this experiment also reveals that with regard to salt concentration the conditions under which the experiments have previously been performed have been optimal. And finally the relative insensitiveness, at least of Prunus, to salt concentration is not in accordance with the great dependence of serological reactions upon salt concentration, and affords a first sug- gestion as to the non-protein nature of the reaction. B. Errect oF HyprRoGEN 10N CONCENTRATION ON THE REACTIONS This first experiment having revealed that the reaction is appar- ently not affected by the weaker salt concentrations used, and having indicated the best salt concentrations for further study, the next step was to vary the pH while holding the salt concentration near the opti- mum. For this purpose Robinia and Prunus were chosen. Each was divided into 9 samples and each series of samples was extracted in the usual manner in a series of buffers (Cohn’s system) all having the same Salt concentration (.06M) but having a variety of hydrogen ion concentrations. The range of pH varied from 5.2 to 8.4, since this is the maximum range possible with Cohn’s system. However, it must be emphasized that the extreme values here used (5.2 and 8.4) with this buffer system are susceptible to a somewhat greater error than the remaining values. This pH range extends from a point somewhat more acid than the ordinary pH of the extracts (except Prunus which may fall as low as 4.0-4.5) to a point far more alkaline than has been observed in any normal extract. The extracts having been prepared, all the possible combinations were made in precipitin tubes and as in the preceding experiment the readings were made at intervals for 40 minutes and then summated. The results of this experiment are shown in Table II. The results of this experiment are of interest in confirming the earlier findings of Kostoff that the reaction bears no relation to pH of the extracts. It is seen from a study of the table that no significant difference is to be observed throughout the whole of the pH range studied. This does not directly afford evidence as to the nature of the reaction, since the precipitin reaction in animals is also independent of pH within reasonable limits (17) but it does indicate that there is no necessity for a careful checking of the pH of every experimental series. As controls for this experiment each member of both series was tested against the complete series of pure buffers, the reactions in all cases being negative. 128 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV A rather interesting phenomenon was observed in connection with this experiment, namely that variation of pH in such a graded series revealed that the plant extracts thus prepared contain coloring sub- stances with the behavior of chemical indicators. In both Prunus and Robinia the pH series showed a progressive darkening of color as one passed from the acid to the alkaline extremes. In Prunus the color ranged from pale lemon yellow at 5.2 to molasses brown at 8.4. The Robinia extracts were all lighter than those of Prunus and ranged from TABLE II. EFFECT ON THE PRUNUS-ROBINIA PRECIPITIN REACTION OF VARYING THE pH OF THE SOLVENT Robinia...pH BoB 56 $0 Gee 6:8 7-2 7:6 8.0 824 Total 5.2 14.0 14.0 14.0 14.0 14.0 14.0 14.0 14.0 14.0 126.0 17.0 17.0 17.0 17.0 17.0 17.0 15.0 12.5 12.5 142.0 17.0 17.0 17.0 17-0 17.0 17.0 17.0 15.0 15.0 149.0 17.0 17.0 17.0 17.0 17.0 14.0 14.0 14.0 14.0 137.0 5.6 6.0 6.4 6-8 17.0 17-0 17.0 16.0 16.0 16.0 15.0 15-0 15.0 140.0 7.2 17.0 17.0 17.0 17.0 17.0 15.0 14.0 14.0 14.0 7.6 8.0 —— 158.0 17.0 17.0 17.0 17.0 17.0 17.0 16.0 16.0 16.0 150.0 17.0 17.9 17.0 17.0 15.0 14.0 13.0 13.0 13.0 143.0 124.0 8.4 14.0 14.0 14.0 14.0 14.0 14.0 13.0 15.0 13.0 Total 147. 147. 147. 146. 144. 138. 13}. 127. 127. very pale lemon at 5.2 to amber at 8.4. Differences in the color of the extracts were also observed in the preceding experiment, where the extracts in more concentrated phosphate were slightly darker, but the effect was much less than in the present experiment. The variation of color intensity with pH has been repeatedly observed in other subse- quent experiments. It is of value since by comparing two extracts which have been placed under different experimental conditions (as in enzyme digestion) it is possible to make a rough estimate from the color as to the pH effect. This fact was made use of in later experi- ments. While speaking of the color of the extracts it might be well to mention, in passing, the effect of dialysis on the pigment in these extracts. In the many dialyses which have been employed in this study it has always been observed that the completely dialyzed dialyzates 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 129 always retained the same dark color as the undialyzed controls, i. e. the pigment is non-dialyzable. However a colorless color-base does pass readily through the membranes and if this colorless diffusate is boiled a rich brown color results, presumably from its oxidation. Beside the color change due to pH there is another pH effect on the extracts which should be mentioned, namely, the tendency for more alkaline extracts to become cloudy or opalescent. Silberschmidt recog- nized this fact (1. c. p. 146: “das Filtrat wurde leicht trub, vermutlich, da es etwas alkalisch war”). This effect was not noted below the point of neutrality, but it has been repeatedly observed in alkaline extracts. It is easily eliminated by making the extracts neutral or slightly acid, which results in a re-solution of the solid suspension. A second experiment on the effect of pH is incorporated into an experiment on heat in the following section (page 130). The result of this experiment was to demonstrate that the reaction does not take place at pH 1.4-1.5 or at pH 8.8. This failure in reactivity may be due to one of two possibilities: either (a) the reactive substances are present in extracts which have been so titrated but are unable to react because of the excessive alkalinity or acidity of the solutions, or (b) the process of titration has resulted in a precipitation of the active ingredients, the reactivity of which is irrevocably lost if the coagulum is removed. The answer as to which of these two possibilities is the correct explanation is seen in the following experiment. TABLE III. EFFECT OF REMOVAL OF ALKALINE COAGULUM ON oe REACTION IN PRUNUS-PLATANUS-ROBINI: Platanmus Robinia Normal: Alkaline Norma}: Alkaline supernatant: supermatant; Normal; 4 a 3 3 Prums &lkaline 4 & 3 3 supernatant; Normal extracts of Prunus, Platanus, and Robinia were prepared (in distilled water). Each was divided into two fractions, one fraction being titrated to pH 9.0 with KOH. The abundant precipitate was then centrifuged away, and the supernatant fluid titrated with HCl back to approximately the normal pH. The extracts were then tested as is shown in Table III A study of this table reveals that the alkaline coagulum does not contain the reactive substances but that the supernatant fluid from 130 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV such coagulation does contain the reactive principles in undiluted strength, and that, by comparison with Table V, the reaction in the latter table is inhibited not by absence of reactive substances but by an excessively high pH, the reaction being fully restored when the pH is again brought to normal. C. Errect oF HEATING ON THE REACTIONS Having now determined the effects of salt concentration and pH upon the reactions, the next important step was to determine the effect of heat upon the extracts, for a sharp decline in reactivity of extracts which have been heated to temperatures of 60°-80°C. would yield strong evidence as to the nature of the reaction. Several experiments in heating the extracts have been performed, all yielding comparable results. As a preliminary word it should be mentioned that serologi- cally active protein is usually coagulated at about 60°C., sera heated above this point and subsequently cleared having lost their specific reactive power. However, a somewhat different display of behavior might be anticipated with the plant extracts since according to Osborne (19, p. 65) “the seed proteins differ in a marked degree from the ani- mal proteins, for most of them are very incompletely coagulated by heat- ing their solutions even to boiling, and many of them are not coagulated at all under these conditions.’’ Neutral and alkaline protein solutions are heat-coagulable only with difficulty or not at all, so that in the experi- ments reported below the extracts were all prepared under such con- ditions as to give them a slightly acid reaction. In the first heat experiment performed, extracts of Robinia and Prunus were employed. A quantity of normal extract was prepared for each of the species, the solvent used being a Cohn’s phosphate buffer of pH 6.0 and salt concentration of .12M (6). The two extracts were then each divided into 10 samples. One sample of each was untreated, the other 9 samples were each heated in water baths for 2 hour at varying temperatures from 30°C. to 100°C., so as to afford a complete series for Robinia and for Prunus. All the extracts were then clarified by centrifuging where necessary, and tested in all possible combinations. Clarification was unnecessary throughout in Prunus as the very slight opalescence which developed at the highest temperatures was so slight as not to interfere with reading the reactions. On the other hand, the Robinia extracts threw out a very dense and flocculent precipitate which began at 60° and became intense at 70°-100°C. The removal of these precipitates, however, as will be seen below, had little effect on the reaction. As controls for this experiment the tests of the 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 131 treated extracts against the normal extracts and then against ane an- other suffice to eliminate possibilities of error due to non-experi- mental variables. The results of this first experiment are given in Table IV. TABLE IV. EFFECT OF HEAT UPON THE PRUNUS-ROBINIA PRECIPITIN REACTION. bint eeece ed 1/2 h t t e es None 30° 40° 50° 60° 70” 80° 90° 100° None Total None 17.0 17.0 17.0 17.0 17.0 12.0 11.0 10.0 7.0 125, 30° =: 117.0 17.0 17.0 17.0 17.0 12.0 11.0 10.0 7.0 125, Prom... 40° 17.0 17.0 17.0 17.0 17.0 12.011.0 10.0 7.0 125. Heated 50° =-:17.0 17.0 17.0 17.0 17.0 12.0 11.0 10.0 7.0 425, 1/2 hour at: 60° = 12.0 12.0 12.0 11.0 10.0 7.5 6.5 5.5 665 _83. 70° «=612.012.012.0 9.0 8.0 7.5 565 5.5 6.5 78, 80° =. 11.0 12.0 10.0 8.0 8.0 7.5 5.5 5.5 55 23, 90° 10.010.010.0 7.0 6.0 5.5 8.0 5.5 3.0 62, 100° 9.0 9.0 9.0 7.0 6.0 5.5 40 5.5 3.0 58, Total 22, 12d. 121. 110. 106. 81-5 70.5 67-5 52.5 As will be seen from a study of the data here presented there is a decline in the reactivity of both extracts on being heated above 60°C. At first glance this would seem to indicate the direct effect of a protein coagulation. Indeed, such was the conclusion of the writers on first viewing the results. But subsequent experiments have shown this conclusion to be erroneous. The decrease observed at 60° is, to be sure, probably an effect of the heat coagulation, but an indirect effect, the coagulum of protein having by physical action withdrawn from solution, in coagulation, some of the other, non-protein substances present in the extracts. As a further test of the effect of heating it was resolved to heat the extracts more thoroughly in an endeavor completely to eliminate the reaction. Accordingly normal extracts were prepared of Prunus and Robinia in distilled water. Each extract was divided into three samples, one sample being reserved as a control, the other two being titrated with HCl to pH 1.4 and with NH,OH to pH 8.8 respectively. The unaltered extract of Prunus had a pH of 4.3, that of Robinia 4.8. Each sample was again divided into two, one fraction being heated, the other 132 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV being.retained as a control. Heating was accomplished by autoclaving with flowing steam at a pressure of 2 lbs. for three hours. The extracts were then cleared by centrifuging and tested as in Table V. In all cases there were coagulations due to heating as well as coagu- lations due to acidification and alkalinization. ‘These were more con- TABLE V. EFFECT OF LONG CONTINUED HEATING ON THE PRUNUS-ROBINIA PRECIPITIN REACTION. P eee -Unhe e . Prunus. eevee -Heated: pH1.5 0 ; o Robinia 4.8 ° (Umheat ea) pH 4.8 17 ; 17 pH 8.8 0 ps 1.5 ° - pH I.E 0 ) (Heated) pH 4.8 17 : 7 pH 8.8 ce) a C spicuous in Robinia than in Prunus. In view of this fact, the results of Table V are most interesting. There was no perceptible influence on the reaction due to this prolonged heating. All the reactions were repeated in parallel several times and viewed by two observers to con- firm this point. It is thus manifest that the coagulum due to heating does not contain the reactive substances in these extracts and that accordingly the reaction is either due to a non-protein or to a protein highly resistant to heat. The effect of pH is here also worthy of com- ment. In the extracts which had been titrated there was a complete loss of reactivity even in the unheated controls. The statements made in the preceding section must accordingly be amended to include the conception that alteration of pH to very high or very low values com- pletely eliminates the reaction, which view is compatible with the other findings reported. It need hardly be emphasized, however, that such pH values as 1.4 and 8.8 never have been observed to occur in extracts prepared in the ordinary fashion. The effect of pH on the color of the extracts was again noted in this experiment, that of both extracts varying from light amber at 1.4 to dark mahogany at 8.8. As further controls for this experiment the Robinia extracts at vari- 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 133 ous pH values, both heated and unheated were tested inter se with com- pletely negative results. The same was true of the Prunus extracts. This completely eliminated any artefact due to the experimental pro- cedure. Furthermore, reactions between any two extracts of different pH were not included because of the artefact reaction introduced through pH coagulation. D,. Errect oF DILUTION ON THE REACTIONS One of the most striking effects of the precipitin reaction in animal serology is the zone effect, whereby an excess of reactive substance may inhibit the reactive display, where there is an optimum concentration of reactive principles, which optimum if passed in either direction, re- sults in a decrease of reactivity. The question arises whether such a zone phenomenon is characteristic of the reactions being studied in this paper. The answer to this question will at once afford a significant TABLE VI. EFFECT OF EXTRACT DILUTION ON THE PRUNUS-PLATANUS PRECIPITIN REACTION Platameg ext ity; eT a ES aes ae tee a 1/128 1/256 1/512 Total 10 29.0 27.0 24.0 20.0 15.0 14.0 12.0 10.0 7-0 40 1.5 260 0.0 164.5 5 27.0 25.0 20.0 17.0 16.0 12.011.0 9.0 5-5 1.0 0.0 0.0 0.0 143.5 Z %.0 21.016.016.0 9.0 8.0 7.0 5.5 4.0 0.0 0.0 0.0 0.0 109.5" 1 16.017.015.014.0 8.5 6.0 5.5 3.5 3.0 0.0 0.0 0.0 0.0 20.5 LR 15.012.012.011.0 6.0 6.0 4.5 3.0 85 0.0 0.0 0.0 0.0 75.0 2/4 12.012.011.0 9.0 7.0 5-5 3.5 3-0 2.5 0.0 0.0 0.0 0.0 §5s5 ee 12-0 12.011.0 7.0 5.0 4.5 25 0.0 0.0 0.0 0.0 0.0 0.0 54.0 Fomality 1/16 11.011.0 10.0 5.5 5.0 4.0 2.5 0.0 0.0 0.0 0.0 0.0 0.0 _42.0 LSB 7-0 6.0 4.0 3.0 1.8 165 0.0 0.0 060 0-0 0.0 0.0 0-0 25.0 1/64 5.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 B20 U/128 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 O20 1/256 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 QO L/BIZ 0-0 040 020 040 060 060 060 0.0 060 0-0 0-0 0-0 0.0 O00 Total: 156. 165. 125. LOJ- 75-0 61,5 48:5 2.0 MS 5:0 25 2:0 20 comparison between the plant precipitations and those of serology and at the same time instruct as to the desirability of controlling this pos- sible variable in subsequent work. Accordingly the following experi- ment was devised. Normal extracts of Platanus and Prunus were prepared as usual. 134 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Each extract was then boiled down to 1/10 its volume (reference to the work in heat effects assuring one that this would not seriously inter- fere with the reactivity of the extracts). They were then centrifuged. Using these 10 normal supernatants as bases, progressive series of dilu- tions were made. Then each grade of Prunus was tested against the various grades of Platanus according to the scheme of Table VI. The readings were made by the method of summation defined above, and the totals plotted against the normalities of the extracts as in Figure I. \ ———Prunus ----- Platanus Total Precipitin 90 fe) > a te ek eke eee 16 32 «64 128 256 512 Normality of extract Ficure I. VARIATION IN THE PRUNUS-PLATANUS PRECIPITIN. RE- ACTION WITH PROGRESSIVE DILUTION. It is at once apparent from a study of Table VI and Figure 1 that dilution of the extracts causes only a regular dilution of reactivity. There is not the slightest indication of a zone effect. In this respect, then, the plant reaction under consideration exhibits a marked differ- 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 135 ence from precipitin reactions in animals, and the evidence against an homologous reaction with that of mammalian blood is accordingly strengthened. The relative regularity of the curves also affords a measure of the relative accuracy of the method of reading, since there is no appreciable deviation at any point of the curves. E. Errect oF DIALYSIS ON THE REACTIONS The next, and probably the most important step in determining the nature of the reactions under consideration is the dialysis of the extracts with subsequent testing. If the reactive substances are completely or nearly completely held back by protein-impermeable dialyzing mem- branes, then there is strong but not absolute proof that the reaction is due to proteins or to radicals bound to proteins. If, on the other hand, the reactive substances easily pass through membranes which are inde- pendently proved to be impermeable to proteins, then the evidence is conclusive that the reactive substances are non-protein in nature. Accordingly great stress is here laid on the dialysis experiments reported. The technique of dialysis varied at the start until a technique was found which proved to be wholly satisfactory. For membranes, collo- dion and cellophane were first employed but they proved to have numerous disadvantages in comparison to the parchment which suc- ceeded them. The collodion membranes, in addition to the labor of making them, proved much more difficult to handle and there was the greater danger of their rupturing or being imperfectly constructed, while the cellophane (commercial: Dupont Corp. #300, not water- proof) proved to be so frequently imperfect that not more than one out of five sheets was satisfactory. Moreover the cellophane when wet is very easily broken and is handled only with difficulty. A high grade of parchment, first in the form of diffusion cells, later in sheets, proved entirely satisfactory. Although every sheet was tested in these experi- ments, very few were found to require rejection. As regards apparatus, the first designed was in the form of the cus- tomary shallow tray with a periodic water change, in which the liquid to be dialyzed was placed in diffusion shells. This was efficient but very slow in contrast to the other apparatus described below. Using the diffusion shells two weeks were necessary for a satisfactory dialysis (at 2°C.), and since the time was so extended there was danger of errors due to this long standing. A dialysis device employed by Dr. F. D. Hagar at the Massachusetts Antitoxin and Vaccine Laboratory was studied and then modified to meet the conditions of this study. The apparatus is constructed accord- ing to the plan of Figure IT. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV A. Complete teapots Guinn! Seeman uci appa ratus Ces ce, qemds ee comes — _Water__ _ | —— Sey — —- l \ pe. Siphon (D)-+-+- -----Dialyzing unit eee a es Glass plates Glass Water compartment ER fcr { Extract compartment LM embrane plates B. Transectional view of dialyzing unit Ficure II. APPARATUS FOR DIALYSIS OF EXTRACTS. 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 137 The dialyzing membrane, cut to a size of 8” x 4” (or preferably nar- rower) is fixed between two U-shaped rubber gaskets of the same size, one gasket being twice the thickness of the other. An automobile inner tube of heavy, smooth rubber may be used for the construction of the gaskets. These gaskets are in turn clamped between two heavy glass plates. The glass must be rather thick (34”, or we have used two plates on each side of 3¢”-'4” thickness each). On the side of the membrane separated from the glass by the thinner gasket is placed the extract to be dialyzed. The other side of the membrane is in contact with a moving stream of water, entering at the funnel C and leaving by the siphon D. Such an apparatus affords a high degree of efficiency and has the advantages of affording a continuous flow, yet of being small enough to be placed in an ordinary sized refrigerator. About 25 cc. of extract may be dialyzed at a time, although by constructing the apparatus on a smaller scale this could be reduced to 5 cc. if desirable. The water flow is at about the rate of two litres per day, and dialysis for 1-2 days has given perfectly satisfactory results. At first the dialyses were carried out at 2°C., but later tests showed that a dialysis for 24-48 hours at room temperature did not seriously interfere with the reactivity or clearness of the liquids, and accordingly most of the experi- ments reported below were performed at room temperature. Using such an apparatus the diffusate may be saved and tested in addition to the dialyzate if such proves desirable. The experiments on the Prunus-Platanus-Ribes-Robinia reactions soon showed that the re- active principles passed rapidly through the membranes, and accord- ingly the diffusates were studied. These were evaporated by boiling in small fractions and thus reduced to the original volume of the dialyzed extracts. Control tests were employed in which whole extracts were diluted to 3-4 litres and then boiled down. Since there was no reduc- tion in reactivity in such controls, it was assumed that the technique of boiling down the diffusate was justified. The efficiency of the membranes was tested in several ways. In the first place, a set of preliminary tests were made with the parchment to determine their permeability. These experiments are tabulated in Table VII. The tests for the presence or absence of substances in dialyzate and diffusate included the Millon and xanthoproteic tests for protein, the Ninhydrin test for protein cleavage products, the iodine test for starch, the Fehling test for sugar, and the silver nitrate test for chloride. These tests are all highly sensitive. The Millon test reveals the presence of one part of peptone in 3000 parts of water, the Nin- hydrin test one part of peptone in 50,000 parts of water (water-clear), 138 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV and the Fehling test is sensitive to at least one part of glucose in 2000 parts of water (these sensitivities being independently determined for this study). The sensitivity of the AgNO, test is also very high, that of iodine less so. TABLE VII. EXPERIMENTS IN DIALYSIS: TESTING OF PARCHMENT MEMBRANES. ___ Cag ysis tests fors Nature of ss Protein = Starch Sugar _Cl__ Exp. Substance dialyzed: extract: Millon Xanthoproteic Ninhydrin Jodine Fehling ARKO; a Peptone « glucose Normal 5 5 bad " Dialyzate 5 5 5 a bad Diffusate ° 1 3 B Ovalbumin » starch ¢ Normal 4 4 3 5 5 Slucose » Nacl. bd Dialyzate 4 4 3 4 4 sd ” Diffusate ie} 0 1 3 3 c " Normal 4 4 3 5 5 ° od Dialyzate 4 4 3 4 4 sd ” Diffusate ) 0 1 3 3 In Experiment A the solution to be dialyzed contained peptone (satu- ration) and glucose (.2%) in water. Dialysis lasted for 18 hours and there was no continuous flow of water in the apparatus, which contained 30 cc. of dialyzate and 40 cc. of diffusate. At the end of this experi- ment the tests showed that the peptone had passed through only in very small amount, while the glucose had passed through freely. In Experiment B the solution to be dialyzed consisted of powdered egg albumen (.5%), soluble starch (.2%), d-glucose (.25%%), and NaCl (about .2%). The clear filtered solution was dialyzed for 24 hours with no flow in the apparatus. The tests at the end of this time showed that no albumen was demonstrable in the diffusate, while it was unaltered in strength in the dialyzate, that the sugar and salt had passed through freely, and that starch had passed through but not so freely as the sugar and salt. Experiment C was a repetition in all respects of Experiment B, and it will be seen that the results were exactly the same as in the preceding experiment. Thus these three membranes, chosen at random, showed a porosity of the exact type required for this work, and offered confidence that the use of such membranes would be of value in the succeeding experiments. Beside these preliminary tests, each dialysis by itself was tested for degree of dialysis by employing on controls, dialyzates, and diffusates tests for protein, carbohydrates, and salts. The Millon, Fehling, AgNO,, 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 139 and oxalate tests were extensively used. The xanthoproteic and iodine tests were also used but less frequently, because the former is of little value in strongly colored extracts, while the latter is less sensitive than the other tests. It is to be emphasized that there is no intention that the presence or absence of Millon-testable protein directly indicates the presence or absence of (hypothetical) precipitin-active protein. Rather these chemical tests are to be considered tests of the degree of dialysis and nothing more. Thus if the diffusate of any given experiment shows no Millon protein, then it is a natural assumption that all protein is being reasonably well retained, while if the diffusate gives a strong AgNO, or Fehling test, this is indicative that all crystalloids of the same approximate molecular size are passing freely through the membrane. In each experiment a control tube of the same extract as was being dialyzed was placed under corresponding conditions, at room tempera- ture or at 2°C. respectively throughout the experiment, and then used as “Normal” in the dialysis tests. The protein, salt, and precipitin tests were all performed in the customary manner with the usual con- trols. All the dialyses lasted from 1-2 days, and all were performed using distilled water as the fluid outside the membrane. The results of the dialyses with Prunus, Platanus, and Robinia are here introduced in tabular form (Tables VIII-XI) since space does not permit the intro- duction of an intimate discussion of each experiment. Considering first the reactions of Platanus acerifolia (Table VIII) the control extracts in every case contained strong protein, chloride, and sugar as is shown by the tests. The controls were all highly and equally reactive with Prunus and completely negative with Robinia, in con- formity with the earlier experiments. In dialysis, except in one experi- ment (Exp. 9) the Millon protein was completely held back by the membranes, while the salt passed freely through. It is thus highly significant that the diffusates in every case offered strong reactions with the Prunus, while the dialyzates, although not negative in but one case, exhibited only very weak reactions. The membrane in Experiment 7 appeared to have been far more retentive than any of the others, as is shown by all the tests, and accordingly the results in this experiment are of less significance than the others. Thus it is evident that there is a definite and strong correlation between the diffusable substances through protein-impermeable membranes and the precipitin potency. The fractionation is not perfect, but in such experiments as these the significance depends not upon a single test but upon the trend exhibited by all. That the positive reactions of the diffusates were not artefacts 140 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV TABLE VIII. EXPERIMENTS IN DIALYSIS: DIALYSES OF PLATANUS. Nature of Exp. Dialysis tests for: Precipitin tests against normal: extract: #: ‘Protein Chloride Sugar Extract: Test: Extract: Test: Normal 5 4 4 4 Prunus 4 Robinia 0 « 6 4 4 4 " 4 " 10) " 7 oe 4 4 bid 4 " 0 " 8 oe 4 a * 4 m 0 " 9 & 4 4 " 4 ad e) " 10 r 4 4 = 4 " fe) Dialyzate 5 3 1 0 " 1 Me fe) " 6 3 ak 0 " 1 * fe) “i 7 3 3 oO ” 3 " 0 " 8 3 1 fe) " 1 " (0) " 9 4 1 0 sd 0 " fe) " 10 4 2 1 * 1 " 0 Diffusate 5 0 4 4 bd 4 S ie) " 6 0 4 4 " 4 e 0) " x 0 a id 4 ug 0 bel 8 i¢] 4 a bed 4 * te) ” 9 1 a a bad 4 * e) ” 10 0 os 4 sid 4 " ce) TABLE IX. EXPERIMENTS IN DIALYSIS: DIALYSES OF ROBINIA. Nature of Exp. Dialysis tests: Precipitin tests against normal; Extract: #: Protein Chloride Extract; Test: Extract; Test; Normal 19 ok 4 Prame 4 Platanus 0 21 4 ” 4 = (0) 2 22 4 4 ” 4 " 0 ” 23 oe a ” 4 - (@] sad 24 4 4 " 4 i“ ¢) Lag 25 4 4 Ls 4 ” 0 " 26 4 4 " 4 " 0 Dialyzate 19 3 1 m 1 " fe) 7 4 2 % 2 ” (a) ” 2 1 " te) (e) " 23 4 1 " 1 sg 10) " 2A 3 1 = 1 ce) - 25 4 2 J 2 id 0 " 26 4 2 " 2 ° ce) Diffusate 19 2 4 ” 4 ” v0) ” 21 1 4 " 3 * 0 " 22 2 4 ° 3 7 0 * 23 2 4 ” 4 vs (0) bad 24 2 4 - 4 7 .e) n 25 1 4 id 3 " 0) 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 141 TABLE X. EXPERIMENTS IN DIALYSIS: DIALYSES OF PRUNUS. Nature of Exp. _ Dialysis tests: _ _ FPrecipitin tests against normal: Protein Chloride Sugar Extract: Test: Extract: Test: Normal py 4 4 4 Platanus 4 Robinia 4 " 2 4 4 4 " 4 " 4 " 3 4 4 4 n 4 n 4 n 4 4 4 4 " 4 " 4 n 11 4 4 4 " 4 " ? " 12 4 4 4 " 4 " ? " 13 3 4 4 " 4 " 2 " 15 4 4 4 n 4 " 4 Dialyzate 1 3 1 Oo " ) " ) " z 3 b G 1 m r?) " 0 " 3 3 1 ft) n ) " fe) nm 4 3 2 0 nm 0 il 0 " 11 4 2 0) " ) " ? " 12 4 2 fy) " fe) " ? n 13 2 : ) " f°) ig ) " 15 3 2 0 " ) . ) Diffusate 1 fo) 4 ? " 4 * 4 Ly 2 0 4 4 ss a Md o bl 13 1 3 3 bs 2 * 2 " 15 2 4 4 " 3 “ 3 TABLE. Xi; reset Ag DIALYSIS: TESTS OF FRACTION AGAINST ACTION (PRUNUS-PLATANUS) Prums; Plat ams: Exp. Nature of Dialysis tests; Exp. Nature of __ Dialysis tests; _ Precipitin _# _extract; Protein Chloride Sugar _# extract; Pretein Chloride Sugar xeaction 11 Dialyzate 4 2 9 Dialyzate 4 1 ° te) 12 " 4 2 0) 9 " 4 1 0 f°) 15 " 3 2 fe) 9 " 4 1 ° fe) 1 Diffusate 0 4 4 5 " 3 1 ) ° 1 " ) 4 4 6 n 3 1 ) ° 1 " te) 4 4 7 " 3 3 0 te) - - 0 4 4 8 " 3 1 1°) 0 15 Ls 2 . 4 9 bd 4 1 ie) 0 1 Dialyzate 3 a 0 10 Diffusate 0 4 4 0 2 " 3 1 1 10 " 0 4 4 rf) 3 " 3 1 fr) 10 " ° 4 4 Ce) 4 " 3 1 ) 10 " te) 4 4 9) ripe " 4 2 ie) 9 id 1 a 4 Oo 12 n 4 2 ce) 9 " 2 | a a o 1 Diffusate 0 4 ? 10 " 0 4 4 S 2 " oO 4 4 10 " te) 4 4 4 1 ” re) 4 4 5 n 0 4 4 4 1 ” re) 4 4 6 * fe) 4 4 4 1 n fe) 4 4 7 " (0) 4 4 ca 1 ” fe) 4 4 8 " ) 4 4 4 142 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV is demonstrated by the fact that they were absent when tested against Robinia and also when tested against one another (additional control tests not shown in the table). In Table [X the results with Robinia fertilis are entirely comparable. Here the retention of protein is not as complete as in Platanus (though it must be remembered that the Millon test is positive not only to formed protein but also to certain of its cleavage products, e.g. pep- tone), but the dialysis results in a fractionation in which one fraction, the diffusate, contains the bulk of the crystalloids, and the other frac- tion, the dialyzate, the bulk of the protein. The fraction containing most of the crystalloids contains most of the precipitin-reactive sub- stances, and the fraction containing most of the protein shows only weak precipitin reactions. The results with Prunus Armeniaca (Table X) are most striking of all. Here the fractionation resulted in diffusates which were almost or entirely protein-free according to the Millon test but which contained the great bulk of the crystalloids, while the dialyzates showed little protein loss and little crystalloid retention. Subsequent tests have shown that the chloride which seems necessarily retained with the protein is prob- ably bound to the protein and is hence non-dialyzable, since the equally delicate oxalate test in later experiments showed such dialyzates to be perfectly calcium-free. The high-protein, low-crystalloid dialyzates were perfectly negative to active extracts of Platanus and Robinia, while the high-crystalloid, protein-negative diffusates were richly en- dowed with precipitin potency. The next table, Table XI, shows the tests of fraction against fraction. These tests were performed to obtain confirmatory data to Tables VIII-X and also to answer a question that arose from a consideration of the data in these preceding tables. It is known that in animal pre- cipitin reactions the presence of a small amount of salt is necessary in order that the reactions take place. In testing a dialyzate against a diffusate or a normal extract one would expect that if the loss of re- action were due to a loss of salt, the second extract would supply suffi- cient electrolyte to permit the reaction to take place. Such does not occur (see Tables VIII-X as well as XI). On the other hand, if the reaction is due to protein in either extract it would be highly improb- able that the testing together of two Millon-negative diffusates could produce a reaction equivalent in strength to the original reaction of the normal extracts. Granting that a trace of protein, non-testable by the Millon test, may have escaped through the membranes, could this give a positive reaction equal in strength to the controls? One could hardly 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 143 zx AgNOz test (Index of dialysis) < ree x 1 Robinia sa o ad >» fe) (@) 1 Zz 3 4 Precipitin reaction 440% é ‘ Ox. 310 Xe oO ox re) Millon test ( Index of dialysis) 2] 0. O. ° al re) ° x ox fe) fe) te) 1 2 3 4 Precipitin reaction Ficure III. Errect or DIALYSIS ON THE PRUNUS-ROBINIA-PLATANUS PRECIPITIN REACTION. EY: 0 — Prinus:..= Ropimias-2 =" Pipes, © = no. test: |1\ = very eel test; 2 = weak test; 3 = moderate 4 = strong test. 144 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV cling to such a remote possibility in the present case, for the dialyzates are in no case positive to one another nor to the diffusates, while the diffusates are in every case not only positive inter se, but positive to a strength equivalent to that of the normal control reaction. The results of all these experiments are summarized in the graphs of Figure III which illustrate strikingly that the reactive substances bear a direct and intimate relationship to the dialyzable crystalloids and none whatever to the non-dialyzable proteins. The experiments in dialysis thus prove definitely that whatever the nature of the Platanus-Prunus-Robinia reactions, the substances re- sponsible for them pass freely through membranes highly impermeable to protein. This applied equally to the reactive principles of Prunus, Platanus, and Robinia. It is thus inconceivable that the precipitin reactions among these species could be due to protein unless one clings to the very doubtful hypothesis that a very small fraction of protein may give as strong a precipitin test as a quantity of protein 10 or 100 times as great, provided that electrolytes be present in sufficient quantity. IF. Errect oF VARIOUS SOLVENTS ON THE REACTIONS Several types of experiments have been undertaken to demonstrate what solvents will dissolve the reactive principles of Platanus, Prunus, Robinia, and Ribes, as well as what solvents will denature the reactive principles. The solvents employed have been alcohol, ether, benzol, chloroform, and carbon tetrachloride. The purposes of the experiments were various, first, to investigate the possibility of improving the tech- nique by an extraction process which will eliminate many heterologous substances from the extracts, second, to shed light upon the chemical nature of the active principles by determination of their various solu- bilities, third, to observe the effect of denaturing, particularly with regard to protein, upon the reaction, and fourth, to determine whether the techniques used by Silberschmidt in pre-extraction and solution exert any significant effect on the reactions at present being considered. The first experiment performed upon this topic was one designed to determine whether the reactive substances of Platanus and Prunus are soluble in alcohol of various strengths. Samples of leaf powder of these species were extracted in the customary manner but using as solvents water, 20%°, 40%, 60%, 80°°, and absolute alcohol. These extracts were aliered. and (ies ested for their content in protein, sugar, and chloride, and for their precipitin potency. It was not possible to test all the possible combinations since the addition of a strong alcohol to a 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 145 water solution or one in weak alcohol resulted in an artefact coagulation effect. The results of this experiment are given in Table It is seen that there is a steady fall in precipitin potency as the strength of alcohol is increased until there is no reaction in alcohol of 60% or stronger. A comparison with the chemical tests shows that TABLE XII. DIRECT EFFECT OF ALCOHOL UPON THE PRUNUS-PLATANUS PRECIPITIN REACTION. Prums: Extracted in alc.of str hh: 100% Tei 60% rs oof af Protein: 2 2 3 4 1 3 an + | 3 3 3 4 4 1 2 4 Chloride; 0 + 3 tr. Prot. Sug. Cl. t ? O 00 8% 2 3 +t 0.0 6% 3 3 1 0.0 Plat ames 40% 4 3 2 11.0 20% 4 4 38 13.0 Ws 4 4 17.0 there was no correlation between precipitin test and sugar (Fehling) or between precipitin reaction and protein, while there was an excellent correlation between the precipitin reaction and salt. This is of sig- nificance only by analogy, but the analogy is striking. It so happened in this experiment that the 20% alcohol solution of Prunus showed much weaker protein than either of the adjacent grades, but this is by no means reflected in the precipitin tests. The same reasoning applies to the water extract of Platanus. Thus this experiment informs us that the precipitin reactive substances are progressively less soluble in the increasing strengths of alcohol, and that the same is true of the chloride present in the extracts, although it is equally true of neither the protein nor the sugar. An important question arises from this experiment, namely: Why were the extracts in strong alcohol non-reactive? Was it because the active substances are insoluble in alcohol or was it because they are 146 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV denatured by the stronger alcohol? This question was answered by the following experiment. Leaf powders of Platanus, Prunus, and Robinia were extracted for 24 hours at 2°C. each with water, 60% alcohol, and 95% alcohol in three lots. The water extracts were then filtered and put on ice. The alcohol extracts were likewise filtered and the filtrates put aside. The alcoholic residues were dried and then extracted in water for 12 hours at 2°C. These last water extracts of the alcoholic residues were then filtered and all the extracts were tested in the manner shown in Table XIII. TABLE XIII. EFFECT OF ALCOHOL UPON THE PRUNUS-PLATANUS-ROBINIA PRECIPITIN REACTION. 6 cohol. 9 cohol. Water Alcohol Water Alcohol Water GSontrol extract extract extract extract Water cont....4 4 4 Alc. 606 extr.... 1 alc. Water Platanus ext re. 0 A 3 4 Aloe 95% ext re eee 0 alc. Water extre.. 4 4 4 Water cont....3 3 Alc. 60h extre-.. 1 alc. Water Robinia extre---3 2 3 Alo. 95% oxtr... 0 alc, Water extr,...3 3 3 The results of this experiment are very striking. Prunus, Platanus, and Robinia all behaved similarly with respect to alcohol extraction. In no case was there evidence that 24 hours in alcohol had to any extent whatever weakened the reaction. Solution in 95% alcohol dis- solved none of the reactive principles. Solution in 60% alcohol dissolved a small fraction of the reactive principles in all cases. One may say with assurance, therefore, that the reactive principles involved in these extracts are insoluble but not denatured, under these experi- mental conditions, by strong alcohol. Proceeding next to a consideration of the same questions as were 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 147 raised regarding alcohol, but respecting lipoid solvents, the following experiment was performed. Dried tissues of Prunus, Robinia, and Platanus were thoroughly extracted with many changes of (a) ether, (b) chloroform, (c) benzol, and (d) carbon tetrachloride, in separate samples. The pulps were then dried and extracted in water in the customary way. Meanwhile other samples of each species were ex- tracted in water without the lipoid pre-extraction. The extracts were then all tested for their precipitin potency. The scheme followed in testing may be illustrated in a single example (Table XIV). TABLE XIV. EFFECT OF LIPOID PRE-EXTRACTION ON PRUNUS-PLATANUS PRECIPITIN REACTION Pre-extraction with ether: Strength of ipitin reaction after: Total imin, 5 min. 10 min. 20 min. 30 min. 40 min. reac.; Extracts tested together; Prums control ¢ Robinia control.. 2 3 3 3 3 3 17 = treated ¢ id control.. 2 3 3 3 3 3 17 Ld cont + cy treated., 2 3 3 3 3 3 17 " treated » n treated.. 2 3 3 3 3 3 17 It will be seen that the ether extraction removed no detectable quan- tity of the precipitin-reactive substances. Since space is limited the corresponding tables for chloroform, carbon tetrachloride, and benzol are omitted, but were they included they would simply be a duplication of Table XIV. In other words, long continued pre-extraction of these tissues with ether, chloroform, carbon tetrachloride, or benzol in no case resulted in the withdrawal of any appreciable amount of reactive substance. We may thus conclude that the reactive principles here are not of lipoid nature. Further tests were employed with regard to benzol, in order to con- firm the preceding experiment. Leaf powders of Prunus, Platanus, and Robinia were divided into two samples each. One sample was ex- tracted in water as a control, a second sample was extracted for {wo hours at room temperature in commercial benzol. The benzol was fil- tered off, part being saved (A), the remainder being evaporated to dryness and re-dissolved in water (B). The benzol residue was dried and extracted in water (C). The various possible precipitin tests were then performed. The results of the tests showed that in every case the extract prepared from the benzol filtrate was perfectly negative whether the test was performed in benzol (A) or in water (B), while the reac- tions of the extracts prepared from the benzol residues (C) were equal in strength to those of the control extracts which had experienced no pre-extraction. 148 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV A last experiment to test the solubility of the precipitating principles in ether and benzol took the following form. Leaf powders of Prunus, Platanus, and Robinia were divided each into two samples. One sample was extracted directly in water. The other was first extracted in cold C. P. anhydrous ether, the extraction being continued as long as color was present in the filtrate (about 15 changes in 3 hours). These last samples were then divided as follows: 1 gm. of each when dry was placed in a tube with 10 cc. of water and extracted for 2 hrs. at room temperature (Table XV: Ether controls). The latter were then fil- TABLE XV. EFFECT OF BENZOL PRE-EXTRACTION ON PRUNUS-PLATANUS- ROBINIA PRECIPITIN REACTION. Precipitin reac. of: Kormal Ether Benzol filtrate; Benz. res. control control In benzol In water In water Experimental: Experimental: Prums + Platanus 3 3 0 0 3 Prums Robinia 3 3 0 0 3 Platanus e Robinia te) 0 0 10) Experimental: Prams + Platames 3 0 3 + Robinia 3 0 3 Plat ams ¢ Prums 3 0 3 Platanus + Robinia fe) 0 (0) « Prumae 3 0 3 Robinia + Platams 0 0 (e) tered. Of the ether residue 2.5 gm. of each species was next extracted in commercial benzol. Many changes were employed, extending through two days. The last change was with boiling benzol, the others with cold benzol. After each extraction (lasting about 1 hour), the benzol was decanted and filtered. The filtrate was evaporated in mod- erate heat. Previous to final evaporation the clear benzol filtrates were tested against one another (Table XV: Benzol filtrate in benzol). After complete evaporation the evaporated residue was re-dissolved in water, which was brought to boiling momentarily and then allowed to cool and extract over night. The following morning the water extract was cleared and tested (Table XV: Benzol filtrate in water). Mean- while the benzol residue of powder was dried (with momentary boiling at the start), and extracted in water over night (Table XV: Benzol residue). In this, as in all the preceding and subsequent experiments the ratio of dried tissue to solvent was always 1:10, regardless of the solvent or amount of tissue. The results of these tests are seen in Table XV. These last experiments thus confirm the earlier ones in demonstrating 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 149 that thorough extraction with lipoid solvents neither dissolves nor de- natures the reactive substances. As has been mentioned above, Silberschmidt has devoted much of his critique of technique (21) to the methods of pre-extraction and solution. As it was thought that Silberschmidt’s results might have de- pended to considerable extent upon his methods of extraction, and as, further, it was desirable that the results of Kostoff and of Chester be reduceable to the same terms as Silberschmidt’s, it was of much inter- est to determine the effect of Silberschmidt’s techniques on our own material. Accordingly the following experiment was designed to include within one experiment all of the techniques described and used by Silberschmidt. Homogeneous samples of dried leaves of Prunus, Platanus, and Ribes were extracted according to the following scheme: Experiment- al tube number: Pre-extraction: Solution: 1 No Water 2 None Physiological saline 3. None Water + solid NaCl to form phys. sol. 4 None Phys. NaCl + MgO to neutrality 5 None N/20 NaOH 6 95% alc. + 1% tartaric Water acid (5 hrs.) 7 ~~ Ditto Physiological NaCl 8 Ditto Phys. NaCl + MgO to neutrality 9 95% alc. + tartaric acid Water 1% (5S hrs.); ether (5 hrs.); chloroform (5 (18 hrs.). 10 = Ditto Water + solid NaCl to form phys. sol. 11 Ditto Physiological NaCl iz Ditto Phys. NaCl + MgO to neutrality All of solutions 6, 7, and 8 for any given species were made from the same pre-extracted tissue, and the same applies to 9, 10, 11, and 12 for each species. In pre-extracting there were three changes of each pre-extractant, the first change being momentary, the second lasting Y hour, the third 1% hrs. The pre-extracted tissues were then dried and extracted for 2 hours, each in its proper solvent. In decanting, in all the experiments in which this process was employed, it was cus- 150 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XIV tomary to decant repeatedly through the same filter paper, adding the residue in the paper to the leaf tissue being extracted. The various experimental extracts were then tested against normal extracts accord- ing to the plan in Table XVI. TABLE XVI. APPLICATION OF SILBERSCHMIDT’S METHODS OF PRE-EXTRACTION AND SOLUTION TO THE PRUNUS-PLATANUS-RIBES REACTION. Binet Exper. Expere cians Saban. Aer thee Exper. Exper. Exper. umber Plababs Hiten Prams frume an Plat an- aon Prums regard + + + Jess of Nomal Yomal Nomel Komal Setree Qutiwe Grtree Cectree species Proms Prunus Flatan. Rides Platane Frums frame Rides 1 3 3 3 3 ft) 0 0 0 2 3 3 3 3 0 0 0 0 3 3 3 3 3 r) 0 0 0 4 3 3 3 3 ) 0) 0 0 5 3 3 3 3 ty) 0 0 0 6 3 3 3 3 0 0 0 0 7 3 3 3 3 Oo 0 0 0 8 3 3 3 3 ) ) 0 0 9 3 3 3 3 0 ry) 0 0 1o 3 3 3 3 ) 0 0 ty 11 3 3 3 3 ty) 0 0 0 12 3 3 3 3 ty) 0 0 0 In this table “Normal” Prunus, Platanus, and Ribes refer to extracts #1 of these species respectively (aqueous extracts in which there had been no pre-extraction). A word of explanation must be inserted with regard to the last four columns of this table. At the time that this experiment was performed, the authors had already determined that the reaction in these species is due to the interaction of calcium ion in Platanus and Ribes with oxalate ion in Prunus. Therefore it was pos- sible not only to determine the effect of the various extraction methods upon the calcium oxalate reaction, but it was also possible to eliminate the calcium oxalate reaction from consideration and observe whether these various techniques resulted in the demonstration of any additional reaction beyond that due to calcium oxalate. Accordingly, in the last four columns of Table XVI the ‘““Normal” extracts are extracts in which 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 151 in the case of Prunus the oxalate present had been precipitated by the addition of a slight excess of Ca(NO,)., and in the cases of Platanus and Ribes the calcium present had been precipitated by the addition of a slight excess of K,C,O,. Returning now to the experiment proper, one first observes that none of the various techniques of pre-extraction and solution made the slight- est difference in the strength of reaction resulting, in any of the three species used. This demonstrates that there is no step in Silberschmidt’s techniques which will eliminate the calcium oxalate reaction, and fur- ther that his techniques do not modify the calcium oxalate reaction, in these plants at least. The experiment also further confirms the earlier findings reported in this paper with regard to solubility of reactive sub- stances in the various solvents employed. Lest the objection be raised that there might have been different reactions from the extracts obtained by the different techniques which gave equal (but not homologous) precipitin reactions, one may observe that in the last four columns of this table the elimination of the calcium oxalate reaction completely eliminates all inter-reactivity of the extracts employed. Hence one may conclude that the calcium oxalate reaction is the only one concerned here. Summarizing, then, the experiments on extraction of the Prunus, Platanus, Robinia, and Ribes extracts, one may say that: . The reactive substances in these plants are insoluble in strong alcohol, ether, chloroform, carbon tetrachloride, benzol, and 95% alcohol + tartaric acid 1%. — 2. They are not denatured by treatment with any of these solvents. 3. They are equally soluble in distilled water and physiological NaCl solution. 4. The reactions take place with equal facility in the presence of NaCl (.85% ), MgO, and N/20 NaOH, and in acid and neutral solutions. G. EFFECT OF CARBOHYDRATE REMOVAL ON THE REACTIONS In order to test the possibility that the reactive principles might be of carbohydrate nature, an experiment was undertaken to eliminate carbohydrates from consideration. The technique of Rimington (20) was employed. Normal extracts of Platanus, Robinia, and Prunus were prepared in the customary manner. To each of these was added a slight excess of neutral lead acetate. There was a voluminous precipitate, containing mucilaginous substances, etc. This precipitate was redis- solved by washing and suspending in water and bubbling through H.S. To the filtrate (neutral Pb(C.H,O,), filtrate) was then added about 152 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV 1/10 the volume of the original extract of NH,OH solution. A second voluminous precipitate resulted, this containing the bulk of the carbo- hydrates (alkaline PbAc precipitate). NH,OH and Pb(C,H,O,), were then added in turn until the precipitate was maximum, and the filtrate (alkaline Pb(C.H,O,), filtrate) virtually negative to the Molisch test. The alkaline Pb(C,H,O,). precipitate was redissolved in acetic acid, and the excess lead was removed from it and from the alkaline Pb(C,H,O.). filtrate with H.S. These solutions were then titrated back to slight acidity and the precipitin tests performed. The H,S was removed by boiling, although tests showed that it did not interfere with the reaction. The result of these tests is given in Table XVII. TABLE XVII. EFFECT OF CARBOHYDRATE REMOVAL ON THE PRUNUS-PLATANUS- ROBINIA REACTION. [In this table PbAc = Pb(C,H,O,),] Control tests: Tests for effect of HAS gn re 3 Praunus (Normal) + Robinia (Normal) = 3 Proms (Normal ¢ H2S) + Robinia (Normal + 328) = a: " " « Platam id = 3 = bad “ » Platams " 23 Platams " + Robinia ad =O Platams " " » Robinia ' = 0 Prams Platamsa Robinia (Normal) (Normal) (Normal) Alx.Pbac filtrate,pH 5.2, Molisch t 3 ie) t¢} bin Alx. Phac precisitate, Be ge 8, se fare ao very strong..... 0 0 i?) Control (Normal ex 3 (0) (0) Alk.PbAc filtrate, pH 5.3, Molisoh test very weak........+- 3 0 ce] Platams Alk.Pbac pres “iP itate, a 4. 1, Moliech iid hie ae seen te) © f¢) Control (Norm 3 r¢) 0 Alx.PhAc filtrate, pH 5.8, 5.5, Molisoh test weak-......--- 0,0 0,0 0,0 Alk. PbAo abe eet » FH 5.0, 5. 8, Molisch eae very strong 0,0 0,0 0,0 Prums Neutral Phac 0 0 0) Neutral PhAc pir itate ° 3 3 Control (Normal extract) r0) 3 3 This experiment is of interest in distinguishing the reactive com- ponents of Prunus as opposed to those of Platanus and Robinia. The former are precipitated by neutral Pb(C,H,O,)., the latter by neither neutral nor alkaline Pb(C,H,O.),. The reaction appeared in full strength in a fraction showing only a trace of Molisch reaction and hence could not conceivably be due to a Molisch-carbohydrate in Plata- nus or Robinia. The fact that the active constituent of Prunus was pre- cipitated by neutral Pb(C,H,O.,), does not shed much light upon its nature, since this precipitate contains a highly heterogeneous mixture of substances, but it is of interest in view of the later findings regarding its oxalate nature, since many oxalates are precipitated but reclaimable by this technique. 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 153 H. ErFrect oF ENzYME DiGESTION, CONTAMINATION, AND AGEING ON THE REACTIONS Of considerable interest is the question of whether the reactive in- gredients of these extracts are affected by the action of enzymes. In- deed the failures of enzymes, suitably chosen and tested, to inactivate such reactive principles is very strong evidence against their protein nature. Accordingly the following experiments were planned to test this matter. In the first experiment on this head, Prunus, Platanus, and Robinia were again used. Normal extracts of each of these species were titrated with KOH to pH 9.0-9.5 and then divided each into two portions. One portion of each was treated with a few drops of Difco Bacto-trypsin. Both portions of each were then covered with toluol and incubated at 37°C. Each day a few drops of the trypsin were added to each experi- mental tube. The incubation continued for 9 days. During this time no additional alkali was needed, as the rise in pH due to tryptic diges- tion was not over pH 1 during the experiment. That slight rise in pH which did take place was during the first three days alone. On remov- ing from the oven the extracts all showed a greater or less precipitate, and this precipitate, for the most part, was not brought back into solution on acidification. The extracts were then acidified with HC] to their original pH, filtered or centrifuged, and used in testing. The results of these tests are shown in Table XVIII-A. Where two values are given for any given reaction (as “4, 4’) this indicates two separate performances of the experiment, with extracts separately handled throughout. In the table “normal” signifies fresh, unheated extract, “control” signifies extracts titrated and incubated but not trypsinized. As a second part to this experiment, 15 cc. each of normal extracts of the same three species were slightly acidified and treated with crys- talline pepsin. The treated and control tubes were incubated at 37°C. for 8 days and then tested (Table XVIII-B). For a third part to the same experiment, extracts of the same species were titrated to neutrality with KOH, placed in fermentation tubes at room temperature, and inoculated with baker’s yeast. Control tubes remained beside them at room temperature. After five days all were cleared and tested. There was no gas production in any case. The results, (Table XVIII-C) being a duplication of those in part B, are not given in tabular form. An examination of Table XVIII-A shows that in no case did the trypsin digestion result in a loss of reaction in any one of the three types of extract. The minor variations are not considered of signifi- 154 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV cance in view of the decided strength of all the positive reactions. The same is true of the trypsinized extracts (B) and of the extracts treated with yeast. One may conclude from this experiment that the reactive substances are highly resistant to the action of proteolytic enzymes and of the zymase complex, and accordingly the view of their non-protein nature is strengthened. TABLE XVIII. EFFECT OF ENZYME DIGESTION ON PRUNUS-PLATANUS-ROBINIA PRECIPITIN REACTION. A. Effect of trypsin digestion. Prums Platanus Robinia Normal Control Exper. Normal Control Exper. Normal Control Rxper. Normal 0 0 0) 4 A 4,4 3 3 3,5 Prumis Cont rol 0 co) 90 3 2 2 1 10) 0 Exper. 0 (a) to) 4,4 4 4,4 3,5 2 2,3 Normal 4 3 4,4 0 ie) fe) ° 0 () Platems Contro 4 2 4 ie) 0 ie) (9) ° ce) Exper. 4,4 2 4,4 0 t¢) ce) (0) c¢) ie) Normal 3 1 3,0 9) 0 0 fe) fe) fe) Robinia Control 3 fe) 2 ° 0 f¢) c¢) oO 0 Exper. 3,3 0 2,5 i?) 0 c¢) (0) ie) Be Effect of pepsin digestion. Prums....Normal 0 (e) 4 4 4 4 8....Treated ie) fe) 4 4 4 4 Robinia...Normal 4 4 fe) ie} 0 (e) Robinia. ..Treated 4 4 0 (0) 0 ie) Platanus..N 4 4 (e) (9) 0 i) Platams..Treated 4 4 0 0 fe) 0 C. Effect of yeast digestion.(Omitted. Results exactly as in "B" above) Beside this experiment in enzyme digestion certain other evidence may be cited relative to the same question. Frequently extracts kept for long periods of time, particularly at room temperature or without protection with toluol, become contaminated with various types of microorganisms. Incidentally such extracts have frequently been re- clarified and found still to retain strong precipitin potency. However, in order to have definite data to meet this question an experiment was arranged to illustrate the effect of contaminations on the extracts. In this experiment extracts of 6 species of Nicotiana were employed, all of which had become contaminated 3 days after they had been first prepared and tested. The contaminated extracts were cleared and tested again against the same solutions as had been used with the uncontaminated ones three days previously. The results of these tests are given in Table XIX. 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 155 It will be seen from an examination of this table that the contamina- tion exerted no inhibitory effect upon the reactive substances, as there is no significant difference in the results of the tests before and after contamination. It should be noted here that the Nicotiana species all contained excess calcium and that they were tested against potassium oxalate simply because this gives an accurate measure of the calcium oxalate reaction. The significance of the test in no wise differs from that if Prunus had been used in place of the K,C,O,. The tests against N. Rusbyi, on the other hand, are not due to the calcium oxalate re- action, as will be explained in the following sections, and for this reason the evidence on contamination of this experiment signifies the absence of its deleterious effect not only in the calcium oxalate reaction but in the MN reaction as well. In addition to the experiments on enzyme digestion and on con- tamination, one other type of evidence should be considered in this section, namely the influence of age of the extracts. The question has seriously concerned the writers as to whether it was safe to use an extract, however clear, which had been preserved by cold and toluol for long periods of time. Accordingly an experiment to study the effect of age on the extracts was carried out. A group of extracts (See Table XX) which had all been extracted and preserved at least four months were decanted and tested as is shown in the following table (Table XX). An examination of the data shows that in no case was there any significant change attributable to the long period of storage. This fact indicates that future workers in this field may with confidence store extracts, at least of this type, for considerable periods of time at 2°C. and covered with toluol, and also that the auto-precipitations occurring in such stored extracts do not weaken the reactions. I. Errect oF DousBLEe PRECIPITATION ON THE REACTIONS A final point to consider in studying the analogy between the plant reactions and those of animals is the phenomenon of double precipita- tion. In serology the specificity of the precipitin reaction is such that if an antiserum A is prepared immune to two antigenic sera B and C, then A may be fully precipitated by B, cleared, and subsequently strongly precipitated by C. Presence of such a phenomenon in plants would strengthen the view that the plant reactions are homologous to those of animals, and absence of this phenomenon would reduce the plant reactions to a relatively simple type of precipitation. In order to test this point, an experiment was planned involving: 156 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV TABLE XIX. EFFECT OF CONTAMINATION ON THE CALCIUM OXALATE AND MN REACTIONS. CON.: CONTAMINATED. UNCON.: UNCONTAMINATED. Precipitin reaction against the follow ing species of Nicotiana; ___Blata _ Langsdorffii A AaUnAaR Coren Leth ETS —glanca __ Cont. Uncon.Cone Uncon. Con. Uncon. Con. Yncon- Con. - Con. Yncon. Nicotiana Busbyi 2 2 1 1 t t 1 2 a 1 2 1 (Mm) Lotansian oxalate 3 3. 3 3 3 3 3 4 3 4 4 x (x/100) TABLE XX. EFFECT OF STORING EXTRACTS FOR LONG PERIODS OF TIME IN THE COLD. “a” — Test when extracted. “b” = Test after four months of storage. Prams Pramas Photinia Deutzia Spiraea Armeniaca Armeniaca villosa scabra virginiana Nikado plena serrulata Hydrangea panic.grfl,"a" 2 u iia 2 2 0) 0) Phot inig villosa "a" 2 2 0 "pe 2 2 ) Deutzia scabra pl, "a" t t 0 ™" 1 1 ) Pyracanthe coccinea "a" 2 2 0 ) cru 2 2 ) fy) Spiraea virg. serml. "Q" & 2 ) bs ed co 2 ) TABLE XXI. DOUBLE PRECIPITATION IN THE PRUNUS-PLATANUS-ROBINIA-RIBES- HYDRANGEA REACTION COMPLEX Tested against normal extracts of; Robinia Platems Ribes Hydrangea Prums control (Prum.: water = 1; 4).... 2 2 3 3 P 6 + Robinia. Cleared mixt 1: 4.. 0 0 fe) 0 Prumis ¢ Platanus.Cleared mixture, 1: 4... ce) 0 Oo ie) Prunus « Ribes. Cleared mixture, 1: 4... 0 0 (e) (0) P 6 + Hydrangea. Cleared mix. 1; 4.. oO 8) (¢) (e) Robinia (Normal)......erccccccsesssscvces e- O 0 (e) (s) Platemms (Normal) o.ccccsccscccccccssccsesese (9) (e) 0 re) ranges (Normal).......sccsc eee eeereeeree fe) (a) fe) (a) Ribes (Normal) ..ccscccssccsccccressevsece ee 0 0 (¢) (¢) 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 157 normal extracts of Robinia, Platanus, Ribes, Hydrangea, and Prunus. Prunus is positive to the other four, which latter four are negative inter se. As the Prunus control was prepared an extract consisting of one part normal Prunus extract + four parts water. The other extracts were prepared by mixing one part of normal Prunus extract with four parts of Robinia, Ribes, Platanus, or Hydrangea extracts respectively. A voluminous precipitate resulted in all cases. The latter was removed by centrifuging and the resulting cleared mixtures were tested as in Table XXI A consideration of these results shows that complete precipitation with an excess of any one of the four counter extracts completely elimi- nated any further reactivity of the Prunus with the other extracts. This is a situation very different from that of animal immunology, and points definitely to a qualitative difference between the plant and ani- mal reactions. Indeed one is assured by a study of Table XXI that the reactive principle in Prunus is a single substance fully precipitable by any one of the other four extracts, which latter might well be assumed to contain the same reactive ingredient. We have now reviewed the evidence regarding the analogy of the Prunus-Platanus-Ribes-Robinia-Hydrangea reactions to the zodimmuni- tary precipitin reaction. Before passing on to the next subject it would be well briefly to enumerate the points demonstrated in the present section. > . The reactive substances of the Prunus-Platanus-Ribes-Robinia com- plex differ from the specific protein reactions of animal sera in the following circumstances: . The reactive substance of Prunus is manifestly different from that of Platanus, Robinia, Ribes, and Hydrangea. The reactive substances in these last four species, however, appear to be identical. . The reactive substances in all extracts are highly resistant to heat, alcohol coagulation, variations in pH, enzyme digestion, coagulation by strong acids and alkalies, and bacterial and fungus contaminations. — bo 3. The reactive substances in all these extracts pass freely through dialyzing membranes impermeable to formed protein but per- meable to crystalloids. 4. There is no evidence of a zone phenomenon in these extracts comparable to that of animal serology. ee] . The reactive substances are non-lipoid in nature. QD . They are, at least in some cases, non-carbohydrate in nature. 158 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV s) . They are unaffected by any of Silberschmidt’s various techniques of pre-extraction and solution, including pre-extraction with alcohol, ether, and chloroform, and solution in water, physiological saline (acid or neutral), and N/20 NaOH. IV. PROOF OF THE CALCIUM OXALATE REACTION The preceding section having treated with a comparison of the plant precipitin reaction with that of animal immunology and having dealt with the evidence that the plant reaction is non-protein, non-lipoid, and non-carbohydrate, the present section will concern itself with a proof of the exact nature of the reaction in the Prunus-Platanus-Robinia-Ribes complex and an extension of this proof to a variety of other plants. The first successful line of attack of this problem was in experiments in ignition of extracts and precipitates. It has already been pointed out (page 125) that a preliminary experiment in the ignition of extracts failed to reveal the reactive substances in solutions of the ash. How- ever, a new type of experiment was resorted to in order to determine whether the precipitate itself was of organic or inorganic nature. Clear normal extracts were prepared of Prunus, Platanus, Robinia, and Ribes in the ordinary way. These were centrifuged for several minutes to throw down any further organic débris. Then the following mixtures were placed in centrifuge tubes: Prunus (5 cc.) + Platanus (5 cc.) . * + Robinia (5 cc.) i “ + Ribes (5 cc.) These were permitted to interact for 10 minutes, and were then centri- fuged. Each precipitate from centrifuging was washed in three changes of distilled water, suspending the precipitate in water and centrifuging after each change. The precipitates were then suspended in the last water and each was ignited in a small beaker at dull red heat (SO0°C.). In no case was there evidence of any appreciable amount of carbon in the process of ignition. The deposits remained perfectly white throughout, with only the slightest trace of pale brown momen- tarily appearing. This experiment was subsequently repeated several times but always with the same result. A second line of attack was the solubilities of the precipitates. It was found that these were highly insoluble in all ordinary solvents. In fact solution was successful only in strong acid, while alcohol, lipoid solvents, boiling water, benzol and toluol, weak acids, and alkalies all failed to dissolve them. 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 159 At this point qualitative analysis of the precipitates was resorted to. The analysis gave the following results: — . In physical appearance the precipitates resembled organic solids in being limey, granular, heavy, and easily centrifuged. . Qualitative analysis showed calcium and oxalate ions to be the only ones present in any appreciable quantity. . Microscopically the precipitates were in the form of regular gran- ules, not amorphous, identical in appearance with certain commer- cial samples of calcium oxalate. Recrystallization of the precipitates (by solution in 6N H,SO, fol- lowed by precipitation by neutralization with 4N KOH) gave an abundance of crystals of the characteristic size and shape of CaC,O, crystals and indistinguishable from crystals of a commercial sample ot CaC,O, similarly treated. . Treatment of the granules of the precipitates with 6N H.SO, under the microscope showed first a moderate solution followed by a very striking conversion of the remainder of the granules into the char- acteristic raphides of CaSO,. This is a fairly accurate test for CaC,O,, and was precisely the behavior of a sample of commercial CaC,O, similarly treated. . When alcohol was added to the H,SO, solution of the precipitates, there was a precipitation (test for calcium). . The acid solution of the precipitates strongly reduced potassium permanganate (test for oxalates). bo w f& wn On ~ Next, through the courtesy of A. D. Bliss of the Harvard chemistry laboratories, a quantitative analysis was made of specimens of the washed precipitate prepared as was that in the ignition experiments described above. Mr. Bliss’ analysis of the precipitate gave the following results. The preliminary identification of the precipitate as calcium oxalate on the basis of its qualitative reactions, etc., was well confirmed by the quantitative analytical data, for in view of the stoichiometric relation- ships involved nothing else could be present to any great extent. The small amount of substance available naturally rendered extensive analyses impossible, but the calcium content was determined and also the permanganate-reducing power of the organic radical involved. To determine the calcium, a specimen was weighed, ignited at red heat, then cooled and weighed. Calculating the weight of calcium oxalate equivalent to the calcium oxide obtained, allowing for the fact that when prepared from wet materials and in contact with moist air calcium oxalate exists as the monohydrate, CaC,O,.H.O, the result was 85.89% of calcium oxalate in the material. 160 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Since the calcium oxalate might not be pure, it was next dissolved in HCl and the calcium precipitated with ammonium oxalate in ammoni- acal solution. The calcium oxalate was filtered off, washed, and ignited to constant weight as the oxide. The product was pure white, as it should have been. With the weight of calcium oxalate thus obtained the calculated amount of calcium oxalate was 84.61%. This percent- age does not differ very much from the first, showing that the oxide from the first ignition was practically pure calcium oxide. The percentage of calcium in the original sample was then found to be 23.22% as compared with 27.43% in pure calcium oxalate. This discrepancy may have been due to one or more of several causes: (1) presence of metals other than calcium, (2) other organic aids than oxalic, of higher molecular weight, or (3) organic matter which would disappear on ignition. In order to check the results of the calcium determination, a titration of the reducing power of the material was made with standard perman- ganate solution. This titration gave for the percentage of calcium oxalate in the precipitate 87.94%. The difference between this value, 87.94%, and that found from the calcium analysis, 84.61% may have been due in part to the normal errors of analysis, to a possible differ- ence in composition of the material, since the two samples were inde- pendently obtained, and lastly to higher molecular weight acid radicals or organic débris in the sample, as mentioned above. The qualitative and quantitative analyses having agreed as to the calcium oxalate nature of the precipitates from the Prunus-Platanus- Robinia-Ribes reactions, the next step was to determine the behavior of extracts of these and other plants in the presence of solutions of calcium and oxalate. The first test of this type was with normal ex- tracts of Prunus, Platanus, and Robinia as tested against one another and against certain calcium compounds as well as certain oxalates. These tests were all performed in the customary manner, and the re- sults are given in the following table (Table XXII). A study of this table shows that Prunus behaves exactly as do the oxalates, while Platanus and Robinia behave in the same fashion as the calcium salts. This same type of analysis was next applied to a variety of all the woody plants which were at the time available. All were tested against potassium and ammonium oxalate and calcium chloride and nitrate as well as against one another (Table XXIII). All the plants were found to fall into two groups, one group including only Prunus which was negative to the oxalates and positive to the calcium salts, the other 1933] RELATION OF THE PRUNUS-ROBINIA-RIBES REACTION TO THE TABLE XXII. CHESTER AND WHITAKER, PRECIPITIN REACTION PRESENCE AND ABSENCE OF CALCIUM AND OXALATE. Robinia, Prums, Fotassium normal pormal oxalate: Potassium oxalate Ammonium oxalate Prams Ameniaca Platams acerifol. Robinia fertilis Ribes Carrieri Hydrangea pentcul. Syringa vulgaris Lig.obtusifolium L. vulgare Leibota Calcium chloride Calciun nitrate © Potassium oxalate fe) ¢ Parts soluble in 100 parts Of water; Calcium nitrate 93. Calcium chloride 70. Calcium acetate 42. Calcium sal te 2 Calcium tungstate of Calcium phosphate Trace Calcium fluoride T race Potassium oxalate 33. Ammonium oxalate 4.2 © Ammonium oxalate °o wr ecoooo0o0o0o90 oooocdoo wr TABLE XXIII. PRESENCE OR ABSENCE OF PRECIPITIN REACTIONS IN CERTAIN WOODY PLANTS IN RELATION TO PRESENCE OR ABSENCE OF CALCIUM AND OXALATE. © Prunus Armeniaca oO @ Platanus acertfolia ¢ Robinia fertilis Ribes Carrieri + + ¢ Hydrangea paniculata ¢t T t oo OoOoOOrNNN Syringa vulgaris 2 ¢ + Ligustrum obtusifolium ¢ ¢ ¢ oo ooOoOrnnNDN L. vulgare L.ibota ¢ L + Ammon COOrFNNN oo Calcium chloride + ¢ ¢ 161 {um ate: Calcium nitrate ¢ ¢ + 162 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XIV group containing the remainder of the woody species, which were per- fectly negative inter se and to the calcium salts but all positive to the oxalates and to Prunus. We thus see that all of the reactions observed in the woody plants are susceptible to explanation on the basis of a calcium oxalate precipi- tation occurring as a result of the interaction of oxalate ions in Prunus with calcium ions in the remainder of the extracts. No other hypothesis is necessary to explain any of the reactions in this group of plants, and that the reactions observed are due to calcium oxalate is manifest not only from the analytical studies above but also from the simple chem- ical fact that the tests of these species with calcium salts and with oxa- lates (Table XXIII) demonstrates that they must contain calcium or oxalate as indicated. Furthermore it follows from the chemistry of such ions that if one adds a solution containing free calcium to one containing free oxalate, a precipitate must result. We thus find that not only do these precipitations occur between the extracts under con- sideration, but that they must occur from the chemical laws governing the behavior of calcium in the presence of an oxalate. It was now felt desirable to extend these results to a greater variety of plants to observe the extent of distribution and the importance of the calcium oxalate reaction in plants, particularly in those in which much of the work on the precipitin reaction has been accomplished. Accord- ingly the same type of experiment as that of Table XXIII was applied to a variety of species of the Solanaceae. The results of these tests are gathered and arranged in Table XXIV. A glance at this table will assure one that the same factors are at work here as were observable in the woody plants. Of the 21 species chosen at random, 8 were negative to the oxalates and positive to the calcium salts, 10 others were negative to the calcium salts and positive to the oxalates. The 8 “oxalate” extracts were all in varying degree positive to the 10 “calcium” extracts. Three extracts, Cyphomandra, Browallia, and Datura Wrightii, were negative or practically so to both salts, and were correspondingly negative to both the “calcium” and “oxalate” extracts. Moreover the “oxalate” extracts were perfectly negative inter se, while the “calcium” extracts were nearly so. Thus the same remarks as were applied to the results within the woody plants under consideration apply equally to the Solanaceae. The calcium oxa- late reaction is of frequent occurrence in the Solanaceae as it was in the woody plants, almost all of the reactions in Table XXIV being explain- able in terms of calcium oxalate precipitations, and moreover the be- havior of these extracts with the pure solutions of the salts requires that 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 163 such “‘precipitin reactions” take place when the various pairs of extracts are combined. In observing the results in Table XXIV it will be seen that the “cal- However, the few reactions which do appear in this block of the table will be explained cium” extracts are not perfectly negative inter se. in the following section. Ammonium oxalate .02N ari at oxalate .O1N Atropa Belladonna ra ferox IR g ie oxalate. 202N oOo Ammonium oxalate Co FPotassium oxalate .O1N TABLE XXIV. PRECIPITIN maaaaains AND CALCI noes ne REACTIONS OF CERTAIN SOLANA oo Atropa Belladonna Datura ferox Solanum tuberosum Capsicum frutescens Datura metel D. innoxia oo Physalis beruviana Calcium chloride .OO5N Calcium nitrate .OO5N NNNNYNFEFNHY SF THIDODOOCIODOCCVCOCO re nw UNNNWNNDKHHNYFEIt rOIMOOOOOO0C oO UQAAAANMWAAANIO FOlIOOOOCO00O0 ga Salpiglossis simata N.paniculata Petunia violacea Solamum nigrum oocoo0ooo0oooor oo RPrPrnwnn cv aw oo 8. melongena PRR PN Ne RPrPnwernwaw ooo00000000 oo oo oo Calcium chloride ,o0o5N wre Calcium nétrate + OO5N oooo000o0o00o ow oo We may thus conclude this section with the following brief state- ment of the occurrence of the calcium oxalate reaction: between Prunus, Platanus, Robinia, and Ribes have been proved to be due to the presence in Prunus of free oxalate which combines with free calcium in the other extracts to produce a precipitation of calcium The reactions The existence of free calcium and free oxalate in various other 164 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV woody plants as well as in a variety of the Solanaceae has been demon- strated, and as a necessary corollary, calcium oxalate precipitations must occur whenever two extracts containing these two ions respectively are mixed. Indeed all the precipitations observed in the preceding tables in the woody plants and nearly all in the Solanaceae were pre- cipitates of calcium oxalate. Mentally reviewing the evidence on the nature of the reaction pre- sented in the earlier part of this paper, all that has been obtained is in perfect agreement with this conception, the relative insensitivity to pH, heat, enzymes, and alcohol, the effect of increasing salt concentration, the regular behavior in dilution, the readiness with which the reactive principles pass through dialyzing membranes, their insolubility in al- cohol and lipoid solvents, and their identity in the experiments in double precipitation, all are in agreement with the calcium oxalate interpretation. However, one very important question remains: Is the calcium oxa- late reaction sufficient to account for all of the reactions observed? The answer to this question will be considered in the following section. V. PROOF OF THE EXISTENCE OF OTHER REACTIONS At this stage in the development of the problem, it was felt that a very important confirmation of the calcium oxalate explanation would result from the behavior of the “calcium” woody extracts in the pres- ence of the “calcium” solanaceous extracts. The number of Solanaceae was increased to 36 species for this experiment, and the precipitin reactions of all the possible combinations of these 42 species were deter- mined. In addition, all the extracts were tested against two oxalates and against two calcium salts. The species were then arranged accord- ing to their reactivity with the salts and tabulated as in Table XXV. Again the 42 species fall into three classes, a “calcium”’ class, an “oxalate” class, and a class reactive to neither. The ‘oxalate’ class (Ca—, Ox-+) is negative inter se, as would have been predicted; so also is the (Ca—, Ox—) class. The “calcium” class (Ca++, Ox—) are all positive to the Ca—, Ox+-) class and reciprocally. Thus far the calcium oxalate reaction is perfectly sufficient for our explanation. But when one concentrates on the inter-reactions of the members of the (Ca+, Ox—) class there are many positive reactions unaccounted for. These, however, are not irregularly distributed through the table but appear to be the reactions with certain specific extracts, notably Robinia, Platanus, and Ribes. Moreover the strength of reaction of any given solanaceous extract is proportional toward all of the three 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 165 woody extracts mentioned. It is thus clear that there is another re- action present, occurring between the woody species and the majority of the Solanaceae. This reaction can be studied best by eliminating TABLE. XXV. A spe ce TABLE OF ALL THE PRECIPITIN Sri. IN NTS EMPLOYED IN THIS STUDY, eetCEL Mae THE CALCIUM OXALATE REACTI Q ¥ z r CaQx- Case Or - C] . Cy . a a = q A p rie on 3 asd gs “ ; Be FagedseseisSs an 22es9s B i otiit eds?! Pe eseen 4 47 eedag B B58 22 a Sad SAS 3 aoe Sa bck Bi beh ALEPEET UES ESEEE CEE CECE ELSES. Pulsesbeg ad eee eerie ee CEC LEPT Looe MAA AOARA ASRARA ALA ee ROA APA RRR At ropa Oootoo0oo000jo0t112332231113221233313221132323 Datura ferox jo 0 t 00 OcOld 10051 2.2)275 211-212) 2.1 8x8 se Sik 2s, 2 288 8.258 Prams Amen. ) Nit tiz.e2 12 DSiteh 2/2212 S282 ere els ira Ss Se Sol.tubervs. jo 0 0 Ol00' O|1 1.02125 2221.1 21,12 2251525125 2462 sicum ) t.1 252.2 2-2:3°1 €91 212d S221 822 6 181 5:2 Dat.metel ) ( 2¢-O.2 2:2.25.1°2:1 2:32.12 1:2 12252 .2'2.82 1.83.82 Physalis ) ) €T021 212.111 2:3:2-2 12 82 ee 2 8 2 2 2 2 7 Dat. innoxia ¢ 0 0/1 2 PRSLIDWVS11¢ 2:82 212°) 18 0121 1 Salpiglossis |o 0 21 Q0 0 D C 0 ofo 0 oo00ton00tod Ee Se ee Oootoltttoto ¢ t0010.1-€2:1t:0 6 8.10.12 .6 10:2) 7 ) ) ° CO0001tEt 0000020 9000000000000 2 t 1000 0000000 900000000 9000 22000 0/0t o|o00 ; 5 t 1/0 0 oo ) ) 900 00 0 1jt 1 38/1 0t 4] 23s 21 4 2 t lit 3% ¢ r- L 1 lolt t : : : 24 22 23 01/0t 0 0 2 l 00 t ) ) 100000 2 1 t t ) 0 ) l 0 0 t oo00¢ 200 2 cloto 0 C 0 0¢ ( bi 2 ojo 0 0 1 4 t 1 t 3 ( pave Ot 0 ) Z111l10i0oco c 0 ) ( ) 2 3 ojo 0 0 (0) C ) ) 4s ae » 31)t 00 gi ¢) C 22 Pl 2 ) 900 ) bet t ) 2 es 1 t c lil 1 oO ¢ t 0 ¢ 0 ¢ tt t QoQ ooo o¢ C L L 00 b 000 > ¢ 0 ¢ ¢ 2 Ot Ot > ¢ 00000000 o¢ 5 2 2/t t 0001 2000000 0000000 00C 000¢ 2 ie 2 230 ( 9010 o000000 nC C r 2 2/t 00 1 230¢ COOKDODD00200D00ND0000 2.2 t 10212000¢ DOF 0000000000000 0N 2 Llolooojoot 120¢ COKODDD00000000000 OC 12 110 00 00000 0c DOO0O00000000000000 2 1J0 0 0jo 00 000 0¢ CO00000000 000000 CC 0 ¢ yo oj/ot til 11 T1lz22Z2212212223333383324243 Ammon. oxal ate | 0 0:0 0'0:0 O00 0/t tt 1 22 2 fitstst 2 2 2 t: 22:21 21-1 151 1-2 tie 1 2 entirely all of the calcium oxalate reactions. Such elimination auto- matically occurs in the interreactions of the (Ca+, Ox—) group and can be made to extend to all of the extracts in Table XXV by adding a slight excess of calcium to the (Ca—, Ox-++) extracts and filtering. [ VOL. XIV JOURNAL OF THE ARNOLD ARBORETUM 166 Such was done, and the precipitin reactions of all these species were determined anew. They are tabulated in Table XXVI, the arrange- ment of species being in order of their decreasing reactivity toward TABLE XXVI. ALL THE PRECIPITIN REACTIONS IN THE PLANTS STUDIED AFTER ELIMINATION OF THE CALCIUM OXALATE REACTION. EMPHASIZING THE AB REACTION. prepareg Sete | Naa oe ee ee STITT wes VTUTQCE | aAANN A NHARAHA AH we vwIODOOTDIOIND *Fyse0e8 snws wld 4a AeA R ew ooooclo0sd a SOO vo ov¥rooovr 90H 0000 2000000 Nu o oooo°o My TOJTENAQO’*STT\u COOOC, ° SCCOLHO =F =o a COv, OOo TAqsTyOTN|H NOH Ade aA >A Ae Ov vy OH HOOCO|O0O » QIBZTnaqensyT|aU COO°O » COC OO ° o euBeTAnier streesyz ° POR A we|HOO BHA TT 30S TIM 3BC | O >OOKP OCH HAO KH AHOCH HOC HK OPO CHK vPOOKO Hv O}H O00 ww nH misya* tos |O ° oo ooooooococn$°o|s? uyzequmtd COCK OO HY vOD0 OHA ans 076qN3°T ° ° ° 2,00 colooOoOlwann STITT eMor ° Ovo eo cCo0CCO ° 90 900 SCO] Ow |JnHnw sy seo ts td TH co oo0°o ° ° ° 8 03UemI0 4 ° SCOOVCDOvOT|OOOO AAA TI JHOpesuey oo eocooOoOR OO a 8 yeuyumoe oo oo SCCOOVvYVOOA wWIOOO Han eue#uolTem’ [0 ° fo MoNoRonoMonom ao) lo) 20 8} [ueoTpun ° COOvKOOve HOODOO HNN SseTe eo°e e°o °o 2 Sue [Oesens °o ° SOC OOCOO nM D/OCOOOINAN BeoelOTA BTuNzo. SCO SAQDDDD OOlO00 2 By xouut: eooooocoocoo°0coed ePOvV A vwIHOO DANN Teyom"gq|;oCOCCOOCOCCCOCCOO YPOOCAHOHOCOvaaANN xoresz"4BT;/OC COC ODOC OOCOOO Syrysoatfa*y{O OOOO eoooo00o ooocCoo0eo SCH CAwBIOCDO vlan amoeqe | ooo oo CcOOOKMOOnDO 20 mofperedoofTICCOCOCOCDOOCODCOCOCOCOCOOCOCOOOoG COCO s fo) N N etzeprspg*H|O CODCOD CDDODCOOOCCOOCCCOCOCCACOSD eo0000 2 esouyanTs*N}OSCOCOCOCOOCOCOCOC OOOO OCCOCOCOCCOD C$OCHOOHAJOCOOvINAN Somtss|OC COC ODOC RCACOO OOOO COCOCOODOS SOSCHOOC AAO OvvINAN etrsqdouos7214 ° ° ° ° ° COCOVv OO ve v/000 2 *“gequlzy umoysdeg |O OO oCo°o a aN Buu0petted edaryy jo ° SCOKOCOKDDDCDVOCOHvHOOOvV OK Av HOKolaNN VOT ysnI oo ° ooo eCocOoOCOOOReY Sir) eared ° oe) eo00000 2 Bqernoyued ° ° eo eCCOLr0CO0”O N 2) PEPOLL) US ae) ey a ae o oo eo00oCRn OONAIOO AN mM 4m Oo oS S&S oo N oa eer a a 4 S = . - rt + we & 342 SER oh oe dg shetes 85.3 coe Ges oe as" es cows & 28 et at 2o 2 » 4 ERB Bp @oOf Sogese se Sk sAee Ss THs ssseh sg gi see Ese a4 aesensk ®@o4 SS eons eget sa wana Dae ow ° ao CE ee Ee ee ee cr ee Per eee eed sats eettga o3 >a 8 gaaeeg deere aen des oe B59 5 e cai iicdct dt ACE BBek ile teh BABE ede CPE Qmamnand a ROR AAA Se Robinia, Platanus, and Ribes, which latter are plainly similar with regard to this reaction. We now find that we are able to separate all the extracts into three groups, one group (A—, B-+-) which is positive toward the three woody species, a second group (A-++, B—) which is composed of these three 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 167 woody species, and a third group (A—, B—) which is negative to both the former groups. Assuming the existence of two reactive principles, A in Platanus, Ribes, and Robinia, B in the Solanaceae, and neither in the Oleaceae, A reacting with B to produce a precipitate, then nearly all of the reactions are accounted for. This AB reaction, the existence of which later experimentation has further confirmed, thus easily accounts for the majority of the reactions not reduceable to calcium oxalate. A study of Table XXVI but eliminating from consideration the reactions of Platanus, Robinia, and Ribes is of interest in the following connection. If the calcium oxalate and AB reactions were the only reactions present in this group of extracts, all the remaining combina- tions should give wholly negative results, since they all lack both oxa- late and principle A. Such is nearly but not entirely the case. There are a few reactions remaining, particularly those of Nicotiana Rusbyi, Datura Wrighti, and Cyphomandra. A reorganization of the data in Table XXVI so as to include only the oxalate-negative, A-negative extracts, and arranged according to their reactivity with Nicotiana Rusbyi gives the interesting situation of Table X XVII. The reactions of Datura Wrightii and Cyphomandra are plainly cor- related with those of N. Rusbyi, as possibly are those of Prunus and Ligustrum ibota. The table clearly points to a third pair of reactive substances, M in the species just mentioned which precipitates in the presence of N, present in most of the Solanaceae but absent in Solanum nigrum, Ligustrum vulgare, Syringa, Petunia, Browallia, and Salpi- glossis, The calcium oxalate, AB, and MN reactions together account for all but a very small percentage of the reactions observed in these species. There still remain, however, a few scattered reactions of which those of Ligustrum obtusifolium are most striking. Rearranging, then, the remaining oxalate-negative, A-negative, M-negative extracts as nearly as possible in order of their reactivity with the Oleaceae, we obtain Table XXVIII. Table XXVIII assumes the presence of a reactive principle Y present in the Oleaceae considered, lacking in all the other plants, and precipi- tating in the presence of X which is found in certain of the Solanaceae (Datura metel, D. innoxia, Physalis, Atropa, and Solanum Capsi- castrum). This scheme accounts for every one of the reactions observed in these 42 species of plants. It is, moreover, confirmed by another type of experiment which will be here briefly mentioned. [VOL. XIV JOURNAL OF THE ARNOLD ARBORETUM 168 It will be recalled (page 156) that when Prunus extract was thor- oughly precipitated with Platanus, Ribes, Robinia, or Hydrangea, the resulting clear mixture was perfectly negative to the remaining woody TABLE XXVII. REACTIONS IN THE PLANTS STUDIED AFTER PRECIPITIN REACTIONS. EMPHASIZING THE MN REACTION. Ne yéqera x NN HAHA HAAR AHAV HAHA KH A HARK HKOYvOwWOCOCO0CO ojooovd TIITIN WE Cer wOee wove HvODOOvseBOOvOwI0000 clowooe wrpuem@eqdsy | woe vpOveue ew QGO0000000000|\0000 orooo°0c°o MMT ETIT Ono woe Cow HOH 090NT00000H0000/0000_ COOv0 “ety SWNT. 6, ou ooooooGe6o ene soo00o eee ° Oo» 00 000 co So 9090 oo jo CO O|» o990 stesortatdres ° oo Oo °o © oo o jo “OTA BITNII | GQQnDDTOCCDDFCCODDDODODOOODOOOOD OO O|IC00D COlooooo "8tma BBUTTEE| GQ Dao DCDCDC ODOC DOC ONDOCOO CODD OD|0000 oloo000 eABBTMA'ETTo CoCo oO HODOOwOHDONDDOO0DO000900|0000 e00900 mateo loo000000 ojooo°oc°o OCoOoOoCO ooOowo0 wueFuoTem* TOS ©900000000000000000000000000 o]oo0ooOe mmo ;sredoofT} tae ats o) Foo Roem) ISPWE'N| Gogo g0o0000 ocooOn Memted’tloqgqga0000 CHOOw eye Toyued: oeocoo°o coloooLea WBWMEIDI"TIS|g 4500 cooo ss aa 80300M03"H]G gGG00 COOMs MIIEDiGqg00 coO oloo a TI3J AOpsRuey "5 ©000000000000000099000000000 o000K rores BINgzEC ooo°o ooooc;eo eV 4 "8T8dBD"TIS|qg DQ QD 0000 oloroon *uyzequmtd: oWoNeR- MoM k-lite N=) ooo Ro MoRo Me) ° COvee BTIMOTPUN | Gg Gg G00 0000 es eoecooulsoce Coase StTPLAT| 60 G000000000000000000000000n|0 w 0000/0 we wd SteTOOAEns ° CiAdataceeeoaccouseeoneeenoes ces C]looLno0d VT xLoUTy * 7H oo0°0 PVOTCDDDOO SOD O OCOD DO DONO vO CO OC|OMw en *qdouos 723° SoD O CC OOO0D oo coolooo0no Ceduna e TOT WNW GQQ00 CODDDR2RDRD0000 oo COn|DQCCCOO|OH4OweR PORATED DDD DD QDDOCODOOCDD0DRDD0000000\0000 cloo#on Syeupance* °° oho RoRoMoReRoRoRoMoRoNo) ooljo00°o COen “ST leg edor3 ooo° “9000000 20000000 °° 200 wee ven cioreest ¢ e200 ° ° C00 Oloowea . S719 804T 000000 00 0 00000 0 0 ccooc0odedoe peda WUT DDD DDD DCD OCDODDOOCOCOD DOOD OODD DO O|lD000 Ovrnna BOm@IIN OG oo eceoooeocKce o o eo wetitie THSOTITWBAB) GQ QD D000 0000 ° oon Ke) 270000 OlunAan BITE OTH GC aac conde dpa eae ngnipreipeipripalpoionpaiaes S. Wibuwe . “ : ~ a ee . « + ae Case “8s 7 4 = FI aoa 8 e= e he 32 + -_ a °o p48 a “~« e a- 33 £ @eona a? 8,28: = § ERs ad vad e 8348 = 2 228 20° © ond Sx = % =~” oO “~ “~aAoO 2 2 a he & fa @ 2 D> afagkssgeeseigsa BGS gRe Rae Rat age she spas gag thon age sees el gs Sasaksgsgas ges oc wWHPAR Bgeg8 §34 ape ge CH Gn Whe & of = eee oe eo te Be CR aS . *O- >o Ca med 3 = om in me mo an Poors | 1O mo AA AN AGA anoAaAn When, on the other hand, Prunus was allowed thoroughly to interact with Ribes, then cleared, the resulting clear mixture was pre- plants. cipitated strongly by Petunia, Lycopersicum, Nicotiana Tabacum, N. This accordingly offers additional evidence as to the presence of a second reaction, in this case the AB reaction. rustica, and N. paniculata. 169 These addi- AFTER TABLE XXVIII. THE PLANTS STUDIED REACTIONS IN REACTIONS. EMPHASIZING THE XY REACTION. CHESTER AND WHITAKER, PRECIPITIN REACTION ELIMINATION OF THE CALCIUM OXALATE, AB, AND MN There is, then, no doubt that beside the calcium oxalate reaction PRECIPITIN there exist certain other reactions in this group of plants. tional reactions have been tentatively referred to as the AB, MN, and 1933] Re ee ee ee eoo0oj0ce edoliyj A YX CHOC COCO oooooo°ocooeooooooooooor0cso euesuoTea"Ss}] COTIOD CDOT AODDODDO DOC OOOO OOO0G Aaa OOjoco O° msoreqny’TO9S|} PO TIO DCO CO DODD DOCOOCOOODOOOOCO OO Olo00 OO KOIeZ BANGSEG] OCOCKCIDDOODGD CDODCDODOOCOCOCDODAAOO OOO OO Olo00 OO stesso Tstdtesg| O° O ° ° ooo ° Bjunjyeg] PP PIO SCOSCOSCOCOOCOOCAOOODODOAOCG AGO O|jfo00 OO moysiradook&tT] POCOJOD DBOSCDDDDGDDDC0DCOD909G SGO0D0|000 9 O moysedeg] PCOSICCOODGD CODD ODDDACDODODODAOOO OO O|jO900 9 Oo eezepueg*n|} PSO] COCOOCOCOFPCOKOKGOOOCOOOOOoOO CAaoOojCo0 oO Oo JsouTeG’N]} PSS] PCOOCOOCDSOCONDGACGDOCOGCOCOG AGDaO|000 OO eyetnoyuad'g| OPS SO}IOC COCO CO OOO OKO OOOO COC OCORC OOO Oo|C00 OO e80j}uem04°N| COO loMeRoRowoxeoweokowomon hos onenonoRononeome) eoooa|jco0 0 9 TisJs1opesteT*M|SOSO] COOKCOCOKCOOCKORCOOOCOOoOOG ACOOaac000 0 O re ‘ujzeqamtd*gH]} PSO} COCO COCOCOODC ACO O OOOO eCOG CACC O000 oo sytimoppumsw]} SOOO] COCO CCOCOOCAOODOCOOKDOOO0G CAOCO|000 0 Oo Byete'HHO lO] COCO OCOOCKCOKO COO OOOKOOCOoOG CoOoOOolo0o0 Oo Oo ettfqdouoszaysw} COSC] DOCCODDOCCOD COOCOCGC0GCGO CGO0CO|000 0 O eoptwurn]H OCOSOISC CC OCOOC OOOO OCOAOC OOOO OOG aco a000 oO Oo QeujumoesM COCO] CODD DC ODD ADCO GC GOOG GG ODAGAGOlO00 oO mmosaeyaswHOOSOKO OOOO COCO OOCOOOKD OOOO OG AGO OjC00 oe sjaqseatTfesMH OOO] COCO DCDOCOODOCCOFOO0D Oo oo esouziais*w} SOOO] COCOOC OCC OOOCOOOCAOC OOO mo ANDO Ojo00 0 Oo somsts*m COO] COCO CDCOCDOCOOOCOOCOCOO CAA Olo00 o CO TISOL L3Ue RRS ‘NH SSCCOOCOKCOCOOCOOCOOD OC COCO O ao ooljco0 0 O SCOODCOCOOKCOOOCOOO OOOO o ooo OjGo00 0 Oo oo o o CO eCCoo ooo 9000 oO SCSCOCOCKOOCOOOCOOOCOOAaOO aoa Olooa ON . SPCOCOCDOCOCODOD OOOO OC OCC OOOO OlO4 yn wo N *Tosysny4Qo°T SSPKOKCSCAOOOCKOCOCODO OOOO OCOOCOOC OO On HH AN = ° . -“ . + 2 a @Qen ae on ei bd 2oa¢2 22.2.8 acess. 8 ddsaat. § Sereseesegsse BPssackag Fe ekes. 8 ee ee ee Cee Pee Pere Same eedaanghectsddaga qeazesaaead 3 a2 a SA >» he ra 3° a3 Ps as a OF Moco MUR ao BESO AEDK Sama meeasesakaan bo soe SD iccinc ie mieaesice weatcdeclestistialwe ire eR Ogee. RS AAMOR ee eRe Rhee ee ROTA AANABIAAL <4 It will be the purpose of the What is their nature? following section to answer this question as fully as the experimental data permit, although it must be freely confessed that the investigation in this direction has only begun. XY reactions. 170 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV VI. PROPERTIES OF THE AB, MN, AND XY REACTIONS A. Heat EXPERIMENTS The AB reaction, according to our analysis of the preceding section, results from the interaction of a substance A present in Robinia, Plata- nus, and Ribes with a substance B present generally in the Solanaceae but absent in the Oleaceae. The MN reaction results in a precipitation from the interaction of a substance M present particularly in Nicotiana Rusbyi and Datura Wrightii with a substance N present in many of the other Solanaceae. TABLE XXIX. EFFECT OF HEAT ON THE cna a RESPONSIBLE FOR THE AB REACT A B Precipitin reaction of A with B; Normal t B heated —Nommal Heated Unheated extract; extract: control(B) 40° 50° 60° 70° Platanus N.Ruabyi i-4 a S lox i) BSE NNONDFYPR Ne Platamus Dat.Wright. Ribes Dat Wright. 2 wNOoNYNYYeKY NH VNYNNEE WH non NM wMeE FY DY NYYNYEPRPNP cr MNYYNYEHYNP A wo NON NE RD is ~ ek The first test regarding the nature ofthe AB reaction was with regard to heat. Fresh leaves of Nicotiana Rusbyi (B), N. rustica (B), and Datura Wrightii (B) were extracted for 18 hours in physiological saline. They were then filtered and the normal extracts were used in the fol- lowing tests. Normal aqueous extracts of dried leaves of the following were also used: WN. Tabacum (B), Platanus (A), and Ribes (A). The solanaceous extracts were then divided each into 8 fractions and the various fractions were heated in water baths for /% hour intervals at temperatures ranging from 40°C. to boiling. The results of this experi- ment are given in Table XXIX Several considerations develop from a study of this table. — . The B factor (but not necessarily the A factor as far as we know at this point) is not rendered inactive by heating for 4 hour at tem- peratures up to 100°C. There was observed in some cases a very slight drop in potency in the extracts heated at temperatures above 58°C., but this was so slight as not to be scoreable. In any case 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 171 there was still a reaction approximately equal to the control reaction in all the various heated extracts. There was a heavy precipitation due to heating in all the extracts but NV. Tabacum, but removal of this precipitate did not influence the reaction. . The extracts of NV. Rusbyi and D. Wrightii were tested against NV. Tabacum and N. rustica at the beginning of this experiment, but the results (in contrast with those of Table X XVII) were negative or very weak. Thus the MN reaction is dependent either upon the variability of different plants or on the solvent used in extraction. This applies to both V. Rusbyi and D. Wrightti. Many subsequent extractions have confirmed this finding: There is considerable vari- ation with regard to the presence and strength of the M principle in experimental plants of the same species. This fact, however, does not apply to the other principles studied thus far. 3. The AB reaction is manifestly different from the MN reaction, in that extracts wholly negative for the one are strongly positive for the other. Both are distinct from the calcium oxalate reaction, since in this experiment the tests with Ca(NO,), remove the possibility that a calcium oxalate reaction could be involved here. bo As a further test of the effect of heat on these reactions, normal extracts of Ribes, Platanus, N. Rusbyi, and N. Langsdorffii were pre- pared. In this case the N. Rusbyi (M) was negative to the NV. Langs- dor ffii (N). Each extract was divided into two samples, one sample being retained as a control, the other heated in flewing’steam at 0-2 lbs. pressure for 3 hours. The Ribes and Platanus extracts remained clear throughout the heating, the other two species autoprecipitated and were cleared before testing. The results of the precipitin tests, which were then performed, are given in Table XXX. It will be seen from this experiment that the AB reaction is un- affected by heating the extracts for 3 hours at 100°C. This applies to both A and B principles. Further tests of the AB and MN reaction were performed under similar conditions. Active A, B, M, and N extracts were prepared in the customary way, heated and cleared as in the preceding experiment, and tested as in Table XXXI. Here the results confirm those of the preceding experiment in show- ing the AB reaction to be coctostable, while there is a weakening of both M and N principles due to the heating. In this respect, then, as well as in others which will be seen later, the AB reaction differentiates itself from the MN reaction. Tests with potassium oxalate demon- strated that the calcium oxalate reaction was not involved here. 172 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV TABLE XXX, EFFECT OF LONG-CONTINUED HEATING ON THE AB REACTION. =xeeensan reaction a N Nicotian dorfii Control TTT mentee Heated Ribes.....Control.. 2 2 2 2 Ribes.....Heated... 2 rs 2 2 Platanus..Control.. yi 1 1 nl Platanns..Beated... 1 1 1 i TABLE XXXI. EFFECT OF LONG-CONTINUED HEATING ON THE AB AND MN EACTIONS SESSA ATES reaction with: Ribe N. sb Control pace Control Heated N.alata...Heated 1 1 Zz 1 TABLE XXXII. APPLICATION OF SILBERSCHMIDT’S METHODS OF PRE-EXTRACTION AND SOLUTION TO THE AB AND MN REACTIONS Ko pre-extraction; Pre-ex.in alc. Pre-ex.in alc.eth.chi. Hormal Experim. H2Q+ 1 BoQe extract: ext > HeQ NaCl NaCl sMe0 NaQh HO NaCl sMg0 HO Nach NaCl shed reacti N.Rustica N.Rusbyi 2 2 (1) 0 oO N.eRusbyi N.rustica 2 2 2 2 (1) #1 1 t t t? ot t (AB react ion) N.Rusbyi Platams 2 2 2 fe) te) 1 2 () t t t (o) N.Rusbyi Ribes 3 3 3 1 (tt?) 2 2 () 1 1 2 ) N.rustica Platams j 1 1 0 (1) t t fe) t t t 0) K.rustica Ribe 2 2 2 2 (1) 1 1 t 1 t 1 t amas WN. 2 2 2 2 (1) t t 1 t t t 1 Platams N.rustica 2 2 2 2 (2) +t t 1 t t 1 1 Ribes NX. Rusbyi 3 3 3 (1) 1 2 2 t 2 2 2 Ribes Nerustioa 3 3 3 3 (3) 2 1 2 t 1 1 t 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 173 B. ExXxprERIMENTS IN EXTRACTION AND SOLUTION As an extension of the scope of the experiment reported on page 149 regarding the effect of Silberschmidt’s various techniques upon the resultant reactions, this experiment was repeated with the AB and MN reactions. The plan of the experiment was exactly as in the earlier one. Pre-extractions were of three types: (a) none, (b) with 95% alcohol + tartaric acid, and (c) with tartaric alcohol followed by ether, chloro- form, and chloroform vapor. Solution was in water, physiological saline, water + solid NaCl in concentration of .85%, physiological NaCl + MgO to neutrality, and N/20 NaOH. The results of this experiment are given in Table XXXII. The reactions involving NaOH are not considered significant because of the reaction of the controls with a pure solution of this solvent. These reactions thus represent alkalinity coagulations and not true precipitin reactions. The remainder of the reactions are all significant, however. With regard to principle M, this experiment shows that it is com- pletely soluble or denatured in 95% alcohol under these experimental conditions. N is somewhat affected by 95% alcohol. The use of physiological saline in place of water as the solvent exerts no per- ceptible effect on the reaction, regardless of whether the NaCl is present after or before extraction. Extracts in water may thus be freely tested against extracts in NaCl without artefact reactions appearing (this point has also been many times confirmed in subsequent tests). Principle A is moderately soluble or denatured in alcohol and the lipoid solvents. It is soluble preferably in acid solution, not in neutral solution, and it is equally soluble and reactive in water and physio- logical saline. Principle B, on the other hand, is equally soluble in water and in physiological saline and in acid and neutral solutions; it is soluble or denatured to a moderate extent in alcohol and the lipoid solvents. Thus this experiment establishes definitely the fact that the AB and MN reactions are not only distinct from the calcium oxalate reaction (which was eliminated in this experiment by the control tests) but that they are distinct from each other. That the A and B principles are definitely soluble or denatureable in alcohol points to their organic nature. This experiment likewise indicates that the techniques used by Silberschmidt may exert a very distinct effect on the strength of the reactions observed. Some of the inconsistencies which he was unable to account for undoubtedly are to be explained in terms of the solubilities of the reactive substances, and a subsequent part of this paper will 174 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV accordingly be devoted to a consideration of his results in the light of this and other experiments. The question of alcohol solubility thus appears to be a very im- portant one, and accordingly a second, more thorough experiment con- centrating on this factor was devised. Leaves of Platanus (A), Ribes (A), Nicotiana tomentosa (B, N), and N. Rusbyi (M) were dried and pulverized; 1 gm. of each was extracted in physiological saline for 2 hrs. at room temperature. One gm. of each was pre-extracted in many TABLE XXXIII. EFFECT OF THOROUGH PRE-EXTRACTION WITH ALCOHOL AND ETHER ON THE AB AND MN REACTIONS. Efrect on principle M: Effect on principle A: masbyi a» tomentosa ie 2 Ribes a+ tomentosa a...1 Platamus a s tomentosa a...1 be ed .0 " de bf a...1 e be ss Qeeol " Ce “ M4 52 "= Oe ” Bee 2 oe " 8...1 bd a+ sl a...0 mn ode ™ a...0 " de * a...0 ” ee ad Q...2 u @+ . a...t e+ " a ..t Effect on principle H; Effect on principle B: Busbyi a » Tomentosa a...2 Ribes a » tomefitosa a...1 Platams & « tomentosa a...1 “ ae be b.--0 "ae " bewot . ae ba beet " ae " 0... * @e J G.--t ad ae " O.--t ad ar * a...0 nn &e ud d...0 " ae " d..-0 ae * 0.2.2 "Be ” @.--1 al ae * e...1 Effect or M and B: _ Effect on principles 4 snd B 8 t Ensbyi a @ tomentosa 4...2 ase a+ tomenroes a...1 Platamus & ¢ tomentosa 4...1 " De * b.--0 be eee ie be # beet " Ge . Ooeed " oF " Geel Ks c+ " Co. ot " ae al a...0 node iad dee. ” de " da...0 ad e+ al @.--1 nn 64 - @.-61 - ee al e...t changes of absolute alcohol; 1 gm. of each was pre-extracted in many changes of anhydrous ether. There were about 9 changes of pre- extractant in each case, each change lasting 2 hours. The pre-extracted supernatants were evaporated down and redissolved in physiological saline for testing. The pre-extracted residues were dried and extracted in physiological saline. Thus for each species the following 5 extracts were availa a. No pre-extraction, whole extract (control) in physiological saline. b. Alcohol extractive in physiological saline. c. Alcohol residue in physiological saline. d. Ether extractive in physiological saline. e. Ether residue in physiological saline. These were tested against each other according to the scheme of Table XXXITI. 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 175 From a consideration of this table one may conclude that under these more thorough experimental conditions: 1. Principles A, B, M, and N are all insoluble in ether and unaltered y ether treatment. They are hence non-lipoid. 2. Principles A, B, M, and N are all non-coagulable by alcohol, even when treated for 24 hours with absolute alcohol at 50°C. 3. Principles M and N are wholly insoluble in absolute alcohol. 4. Principles A and B are both very appreciably soluble in absolute alcohol. 5. No difference was seen in the behavior of A from that of B nor in M from that of N. The results of these two experiments are summarized in the follow- ing table for the reader’s convenience (Table XXXIV). TABLE XXXIV. SOLUBILITY AND REACTIVITY OF PRINCIPLES A, B, M, AND N IN VARIOUS SOLVENTS. React ive Solubility in; Reactivity in; principle: Alcohol] Ether Water Vater » NaCl .85% NaCl NaCl + MeO Sach calcium None Yono Good Good Good Good Results not oxalate None Yone Good Good Good Good eignif- ioant mae Moderate None Good Good Good Poor becanse ° badd Moderate None Good Good Good Good artefact. -™” None None Good Good Good Good sd fed None None Good Good Good Good e C. ExXprerRIMENTS IN DIALYSIS Dialyses of the extracts to determine the diffusability of the AB and MN principles were performed in the same fashion as with Platanus, Prunus, and Robinia as described on page 135. No further comment is necessary beyond the statement that the dialyses of the Solanaceae were usually performed using physiological saline rather than distilled water as the fluid outside the membrane. This was desirable practically because it is difficult to obtain clear extracts of Nicotiana Rusbyi and certain other Solanaceae in water, and theoretically because the protein which might hypothetically be responsible for these reactions might be precipitated by removal of the electrolytes. The experiments of the previous section, however, have revealed that there is no choice in the use of these two solvents as far as demonstrating the precipitin tests is concerned, and furthermore in all experiments here and elsewhere where both aqueous and saline extracts were used in the same experiment, 176 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV control tests against the respective solvents confirmed the absence of artefact reactions due to choice of solvent. Since the AgNO, test for freedom from chloride could not be used in dialyses in NaCl, the oxalate-calcium test, of comparable sensitivity, was substituted where necessary. The results of the dialysis experiments are gathered in Table XXXV. Considering first the dialyses of Platanus (A) one observes that the TABLE XXXV., EFFECT OF a ON THE REACTIVE PRINCIPLES OF ACTIONS AB AND M A, Effect on principle A tract nat: Extract Heture of Exp. Dialysis — Precipitin te f experimental ex Gialyzed: _extract; 4 Prote Cl. Ca. Sucar Extract: Tes te Extract: Tegt: Extract: eat: Platenus Normal 5-10 4 4 3 N.tomentosa 3 Neglauca 3 ft Dialyz. § 2 t t e 2 ” p | ba ad 9 3 t t " 2 ad 2 Ld ad 7 2 2 t " 1 " 2 ” - 5 2 t t " 1 “ 2 * n 10 3 1 t n 2 " 1 ” Diffuse. 8 fe) 4 1 " 0 2 0 ? t * 3 " ie} wd 0 bed " 7 ie} 4 3 " 10] sg, 0 " ” 5 (0) 4 3 " 0 " 0 Robinia Normal 19-26 4 4 4 id 3 id 3 bad alyZ. 24 3 1 t " 3 bad 2 " ~ ' 20 2 1 t bal 2 ef 2 " " 25 4 2 t " 3 * 3 ” " 23 4 1 t e 2 ne 2 " 22 2 1 t 7 2 bad 2 " bad 26 4 2 1 ” 2 " 3 " bed 21 4 2 0 ” 4 al 3 " ” 19 3 si t " 3 as 2 " Diffuse. 19 2 4 2 " 2 - 1 " " 21 1 4 2 " 1 * 2 " " 22 2 4 2 Ls 1 " 2 " Lad 23 2 4 2 " 2 ° 2 " ” 20 4 3 1 ” 2 - 3 " " 2a 2 4 2 ad 2 ad 3 " " 25 1 a 2 " 1 ” 2 B. Effect on principle B ¥.tabacam — oH 3 4 3 Platams 1 .tosentos 4 3 4 Robinia 2 n 1 - 4 4 4 ” F bis 1 a ie sid 37 2 4 4 sd 2 Ribes 4 Nealata ” 38 2 4 3 * 2 7 1 . 2 - 39 2 3 3 * 2 ” 1 2 " = 40 4 3 2 & 2 bad z = 2 id " 41 4 4 2 " 2 = i ” 2 " " 42 2]2 4a 4 " i. " : 43 2 4 4 i 1 n " 44 2 3 4 2 m * 45 4 3 4 ” z N. tabacum dem itaial . 2 1 " t N.toamentosa 3 ie} * 1 ” 1 al 3 t 1¢) bs 1 ad A , a‘ ne rustica " 37 1 ie) 6) ° 1 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 177 TABLE XXXV. (Continued). Extract Nature of Exp. pe ay as Ereoipitin tests of experimentel extrent_savonet: Gialyzed: extract: # Cl, Ca. Sugar Extract: Test: Extract: Test: Extract: N.alata Dialyz. 38 Z t 0 Robinia Z Platames ° Ribes 1 ” « 39 1 t 0 L- 1 2 0 * L * * 40 2 te) fe) 1 " 0 “ a n " 41 2 re) 0 nw 1 " 0 " 1 " bed 42 2 0 t bed a LJ bod 43 2 0 t hed 1 Ls bal 44 . t t bod 2 wl “ 45 yf t t " 2 N.tabacum Diffus. 27 te) & het p a Nealata J 44 t . Oo \ 45 t 3 0 g. Effeot on principle M N.Rusbyi Normal 28 4 4 N.tabacum 1 H.rustioa 5 Ld vd 29 4 4 ied 1 by t-} = = 33 3 3 4 N.nudicomlis 1 bed 2 " 34 3 b § 4 " a " 2 Ms . 8 & ie} het pi bad 2 Ls i > * 0 Lal 1 ee 2 by " 33 3 ie) 0 * ak “ 2 n " 34 3 (0) 0 n n 2 D. Befect on principle H. Nealata Normal 38 2 4 3 N.Rusbyi 2 Lad 39 2 4 3 bed 2 LJ # 40 4 4 2 bed 3 " " 41 4 4 2 xd 3 " " 42 2 4 4 of 3 hed e 43 2 4 4 = $s ” ” 44 2 3 4 w 2 " " 46 2 3 4 ” 2 : Dialyz. 38 1 t 0 = 0 ® " 39 2 t i) bed f?) " " 40 2 0 0 Ld ° " " 41 2 0 0 LJ ° " " 42 2 0 t = oO " “ 43 2 0 t bad 0 " " 44 2 t t bed 0 " 45 2 t t Ls 0 “ Diffus. 44 t 3 hed 2 " " * p Notes: Prot.:Millon test oe ptotein; Cl.:AgNOz test for chloride; Ca.;potassium oralate test for calcium; Sugar: peta ik test. Extracts in right hand columns all normal. Dialyz.: Dialyzate; oie alee ate technique of dialysis was satisfactory as is evinced by the dialysis tests. The fractionation resulted in a retention of most of the Millon protein and very little of the crystalloid content. The reactive principle A was retained in amount comparable to the retained protein in the dialyzates. The diffusates, protein-free and containing almost all of the crystalloids, were absolutely negative to the precipitin tests. The results with Robinia are less clear cut, but still significant. The dialyzates contained the bulk of the protein and an appreciable but minor amount of crys- talloids. The bulk of the salts had plainly passed through to the diffusates, although some protein had also escaped. The results may be 178 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV considered safe by comparison, however, the precipitin reaction as in Platanus being strongest in the high-protein, low-crystalloid fraction. We may thus conclude from these dialyses that principle A diffuses very slowly through membranes relatively impermeable to protein, Turning now to principle B, the dialyses were technically highly efficient. Most of the protein was reclaimable in the dialyzates, but the latter were practically or entirely negative for calcium and sugar. How- ever, with this highly efficient fractionation, nevertheless a very appre- ciable amount of the precipitin reaction was lost in dialysis. This is true of the various species employed in these B dialyses all of which behaved similarly. It is thus apparent that, like principle A, principle B diffuses, although imperfectly and slowly, through membranes highly impermeable to protein but highly permeable to crystalloids. The results with the M and N principles are beautifully clear-cut and striking. With regard to principle M in Nicotiana Rusbyi the dialyzates retained all the protein and none of the calcium or sugar according to the dialysis tests. Moreover the precipitin reaction was retained in its full strength in the dialyzates. On the other hand, al- though the dialyses were almost equally sharp and clear-cut with regard to principle N, the latter was completely lost from the dialyzates in company with the crystalloid content, although very little protein passed through the membranes. We may hence conclude that principle M is retained by protein-impermeable membranes with a high degree of effi- ciency, while principle N readily passes through such membranes. D. Nature oF THE AB, MN, anp XY REACTIONS The experiments in dialysis conclude our investigation of the nature of the AB and MN reactions as far as it has been accomplished at the present time. However, although it is impossible at the present defi- nitely to ascribe these reactions to specific substances, yet from the data obtained one can eliminate many possibilities and arrive at an approximate idea of the nature of the reactive substances involved. We are now able to utilize the following facts which have been demon- strated above. Principles A and B are water soluble, alcohol soluble, coctostable, insoluble in lipoid solvents, not coagulated by alcohol, and slowly and imperfectly dialyzable through membranes relatively impermeable to protein. The precipitate resulting from their interaction chars markedly on ignition. It is thus plain that they are organic, non-lipoid, and non- protein (because of their alcohol solubility and positive, although im- perfect, dialysis). The only alcohol-soluble proteins, the prolamines, 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 179 are insoluble in water (19). Thus far no satisfactory distinction has been found between A and B. Principles M and N are likewise soluble in water and weak salt solu- tions, but are insoluble in alcohol, non-coagulable by alcohol, insoluble in lipoid solvents, and in contrast to A and B are thermolabile. The precipitate formed by their interaction chars on ignition. M is retained by dialyzing membranes with high efficiency, but N passes freely through such membranes. Thus we may conclude that M and N are both organic and non-lipoid. N is in all probability an organic sub- stance of relatively simple structure, and in any case it is highly unlikely that it is protein. M, on the other hand, gives strong indica- tions of being a protein or at least a complex organic substance of high molecular weight. With regard to the XY reaction, nothing can be said as to its nature at the present. This is the least frequent, and because of its weakness and rarity the least significant of the four reactions studied. It is plain from Table XXVIII that the species of Datura, Physalis, and Atropa are alike in possessing a component X which reacts with Ligustrum (Y). Whether or not the reactions of Solanum Capsicastrum against the Oleaceae are qualitatively the same or different is questionable. Because of the weakness and infrequency of this reaction or group o reactions, no analytical tests have been employed concerning it. VII. THEORETICAL AND PRACTICAL SIGNIFICANCE OF THIS STUDY In the light of the investigations reported in this paper, it is now of importance briefly to review the results obtained by the various investi- gators in the field of the plant precipitin reactions for the purposes of culling out the interpretations now known to be erroneous, of support- ing with experimental proof those which are correct, of evaluating the immunological significance of the plant precipitin reactions, and of thus placing the studies regarding such reactions upon a solid experimentally- determined foundation. Accordingly the present section will be devoted to these more important theoretical and practical considerations. A. ANALYSIS OF KosTOFF’S RESULTS IN THE LIGHT OF THE PRESENT TUDY, Kostoff’s studies on the precipitin reaction in plants (12) were re- stricted to 20 species and a hybrid of the Solanaceae. The materials on which the findings of the present study are based included 16 of the 20 species on which Kostoff worked. Of the 157 normal precipitin reactions of these 20 species tabulated by Kostoff we have repeated all 180 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV but 52, with the following results (compiled from Kostoff’s Table 10, l.c. p. 49, and from our Tables XXV and XXVI): Negative reactions reported by Kostoff and confirmed by us ....64 Positive reactions reported by Kostoff and confirmed by us ..... 33 Due to calcium oxalate .............. 0. cece cence eee aa Due to the MN reaction CORED TU 6 yc es dads ew as eines waser eens 53 Positive reactions reported by Kostoff and found by us to be RICO + 5h obese bade hehe wee be via dwae awed 4eeewes 157 Note: With regard to Capsicum, Kostoff used C. pyramidale while we used C. frutescens, but the reactions of these two species are clearly the same, judging from his and our data. The other species involved were identical and largely from the same stock. Of the 27 acquired precipitin reactions described by Kostoff in the Solanaceae we have repeated the normal reactions in 10 of the com- binations with the following findings: Marked acquirement of precipitins according to Kostoff ........ 5 Found by us to be due to calcium oxalate ............... 3 Found by us to be due to’the MN reaction ............. 1 Ot tested BY WG-24 ici Siew dee ce cave et ea ee dads nd 1 Slight acquirement of precipitins according to Kostoff .......... 7 Found by us to be due to calcium oxalate .............. 0 Found by us to be due to the MN reaction ............. 2 Found by us to be negative ................. 0000005. 1 Wot tested by Us ..54s sss sskedibesaueeshawctavesaas 4 No acquirement of precipitins according to Kostoff ............ 11 Found by us to be due to calcium oxalate .............. 1 Found by us to be negative ..............00 0. cece eens 9 Pee TOR OEMS 564 Pek eedeadwesinvaxeunek wens 1 Decrease of precipitins according to Kostoff .................. 4 Found by us to be due to the MN reaction ............. 3 DIOe TOSEC. BY UE ance ke es tee cc ese knwe be oes enes 1 We thus see that of the 33 normal precipitin reactions reported by Kostoff and repeated by us, 70% were due to calcium oxalate and 30% to the MN reaction. It is of interest to compare these results with those obtained by us in all our normal precipitin tests. The total number of combinations tested by us was 869. Of these, 444 were negative and 425 positive. Of the 425 positive tests 242, or 57%, 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 181 were due to the calcium oxalate reaction, 100, or 24%, were due to the AB combination, 73, or 17% were due to the MN reaction, and 10, or 2+ % were due to the XY pair. Thus it is evident that in so far as they have been repeated, Kostoff’s positive reactions were all due to the calcium oxalate and MN reactions, mainly to the former. Since the calcium oxalate reaction is inorganic and more or less incidental to the autonomy of the plant, it is manifest that no immunological conclusions can be drawn from an observation of such reactions. Moreover, with regard to the MN reaction, since one reactive group (N) at least is plainly non-protein, and since in the majority of cases where Kostoff dealt with this reaction he observed no change or even decrease, rather than increase, in precipitin reaction after grafting, no sound immunological conclusions may be drawn from this reaction. Kostoff’s experiments in nephelometric determination of precipitins (13, 14, 15, 15a) are subject to the same criticism as his gross precipitin tests. Kostoff’s experiments im vitro include a discussion of “lytic rings” at the line of junction of the two extracts. These, he believes, are comparable to lytic reactions in animals. In later papers (14, 15, 15a) Kostoff has reported a confirmation of the reactions believed by him to be lytic, by means of dialysis-Ninhydrin tests. Regarding these last tests the authors have nothing to say, since they have not yet been repeated. But with regard to Kostoff’s “lytic rings” these have been frequently observed and studied. In his first paper on this subject (3) the senior writer mentioned having observed them in his oleaceous tests but he found them to be “‘not consistent, weak, and apparently of no greater significance.” He was unable, however, to demonstrate lysins in the oleaceous extracts in preliminary experiments using the Nin- hydrin method. That the “lytic ring” phenomenon is not immunological in nature is evident from the following observation transcribed from our notes on one of the preceding experiments. “On May 4th as a control of Ex- periment 83 the Ribes Carrierei extract (N/2 in physiological saline) was layered with N/100 potassium oxalate. The ring resulting (cal- cium oxalate) was about +1 in strength, clear, and showed the clear- est, most striking double ring separated by a clear zone which I have ever seen” (Chester). This was confirmed by two other observers, and was undoubtedly Kostoff’s “lytic ring,” but occurring in the layering of an extract with an inorganic salt! The same was repeated several times the following day with similar results in all cases. It was not a function of any particular way of pipetting, but careful layering was necessary to bring it out. 182 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV It is therefore evident that Kostoff’s precipitin reactions and his “lytic rings” are both of the same types as the corresponding phe- nomena studied by the writers, which are assuredly non-immunological in nature, and that accordingly, on the basis of the data published thus far by Kostoff, one is not justified in drawing immunological conclusions. B. ANALYSIS OF SILBERSCHMIDT’S RESULTS IN THE LIGHT OF THE RESENT STupy! Silberschmidt’s experiments (21) dealt with the following species of Solanaceae: Nicotiana Tabacum, N. rustica, N. glutinosa, Solanum Dulcamara, Lycium barbarum, and Salpiglossis sinuata. According to our findings these species had the following complements of reactive principles. The three species of Nicotiana all contained an excess of calcium and of principle N; Salpiglossis contained excess oxalate and lacked both M and N. From the behavior of Silberschmidt’s other two species, Solanum Dulcamara probably contained a weak excess of cal- cium and principle N, while Lyctuwm probably contained a slight excess of oxalate and lacked M. The calcium ion was evidently strongest in Silberschmidt’s NV. Tabecum and N. rustica and rather weak in S. Dul- camara and N. glutinosa, while the oxalate was evidently stronger in Salpiglossis than in Lycium, but not very strong in either, Calcium oxalate reactions are plainly evident from Silberschmidt’s tables (e.g. Salpiglossis + N. rustica and N. Tabacum in his Table 2). The clearness of his readings was at times obscured by the cloudiness and dilution of his extracts. This applies particularly to his data in Tables 3, 5, and 6. In his Tables 9, 10, 11, and 12 the reactions are all negative or practically so as would be expected from such reactive complements as those outlined above. The only reactions requiring comment are two involving NV. Tabacum. In Silberschmidt’s Tables 3 and 4 N. Tabacum reacts positively with Solanum Dulcamara and with N. rustica, This N. Tabacum reaction behaves much as the MN, and it would be reasonably explained on the assumption that N. Tabacum (which is closely related to V. Rusbyi, the type of principle M), some- times contains a sufficient quantity of M to react with active N extracts. In accordance with this view is the fact that Kostoff’s N. Tabacum reacted, although weakly, with two of the other Nicotiana species used. As was mentioned earlier in this paper the presence of M in NV. Rusbyi varies considerably from plant to plant and a corresponding variability 14 second paper by this author has recently been perce This nde causes schmidt, re (1932). Studien zum Nachweis von Antik6rpern in Pflan (Planta: Arch. f. Wiss. Bot. 17:493-589)] is subject to the same ge pina as the paper ne discussion, 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 183 in N. Tabacum would perfectly account for the apparent discrepancies between Silberschmidt’s results and those of Kostoff and the present writers. Such an analysis accounts for such of Silberschmidt’s reactions as are significant, while his negative results in the few places where one would expect positive reactions are apparently due either to lack or to weakness of reactive principles in the various combinations. His one acquired reaction is hardly of strength to be considered significant. Hence one may conclude, with regard to Silberschmidt’s work, that it is confirmed in fact, but not in theory, by the present study. For the same criticisms regarding immunological interpretations as were made of Kostoff’s results are applicable to those of Silberschmidt. His positive reactions are all explainable in the terms of the relatively simple reactions described in this paper, and there is no evidence that any of them are due to protein interaction. C. ANAzysis oF East’s RESULTS IN THE LIGHT OF THE PRESENT STUDY East (7) has observed certain phenomena in vivo with regard to recovery of sugar cane from the mosaic disease and to temporary re- sistance to re-infection. In connection with these observations he carried out a series of precipitin tests in order to try to decide whether or not the proteins of various types of cane were different. His experi- mentation followed the same general type as that used by Kostoff and Chester, and accordingly his results may be considered comparable. In certain combinations of extracts of cane with those of cane and other plants East obtained positive results. He found some indication of a consistent difference in precipitin reaction between cane which had never had the mosaic disease and cane which had or may have had the disease. Since plants which exhibited the mosaic disease reacted in the same manner as plants which appeared free from disease but which might have been acting as carriers, and since the reactions of such plants consistently differed from those of cane known never to have had the disease, East believed that his results, including his observations on infection and recovery, pointed to the probability that sugar cane gains an apparent immunity by reducing the virulence of the mosaic virus. The results of the present study render it very questionable as to whether the precipitin experiments with sugar cane should be con- sidered immunological. Since the writers have not experimented with the species of plants used by East, it is impossible dogmatically to rele- gate his reactions to the same category as the non-protein reactions found to occur so extensively in the Solanaceae et al., but assuredly one 184 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV must be rather skeptical in accepting as immunological any precipitin experiments in which the possibilities of such non-protein reactions as the calcium oxalate and AB reactions are not eliminated. D. ANALYSIS OF CHESTER’S RESULTS IN THE LIGHT OF THE PRESENT StTuby The findings of Chester with regard to the precipitin reaction (3, 4, 5) have not been definitely interpreted as immunological, pending the results of the present study. Chester’s findings prior to this study may be stated in the following terms: 1. No normal precipitins were found in the Oleaceae studied. Graft-blighted Lilacs showing serious malnutritional symptoms gave strong positive tests with other Oleaceae regardless of systematic relationship to stock or scion. Lilacs suffering from similar symp- toms but ungrafted gave the same reactions. . In tests in a number of families of woody plants a clear-cut parallel was observed between the precipitin reaction and the systematic relationships of the plants tested. bo ios) It is the purpose of the present section to criticize and interpret these findings. The tests reported in this paper have shown that in every case tested, the Oleaceae were characterized by having an excess of calcium present in their extracts. In the morbid processes attending the slow dying of leaves of Lilac there is an accumulation of oxalate. This has been shown by the fact that reactive extracts of graft-blighted plants may be rendered inactive toward “calcium” extracts by precipitating them with a slight excess of calcium. Such oxalate accumulation is localized in the diseased and dying cells, since when the green portions of green- and-yellow mottled leaves are cut away from the yellow portions and extracted, the extracts of these green portions behave exactly like extracts of healthy Lilacs. This fact renders it still more improbable that the reactions of graft-blighted Lilacs are immunological, since hypothetical antibodies from grafting would be expected to be generally distributed through the leaf tissue. The fact that leaves of plants mechanically injured are also reactive in the yellow areas supports this simple interpretation, as does the fact that extracts of mottled leaves autoprecipitate in extraction, losing their reactive potency as auto- precipitation continues, due to the interaction of the excess oxalate in the yellow portions with the excess calcium in the green portions, Passing next to the highly absorbing topic of the specificity of the normal precipitin reactions thus far investigated, the results of this 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 185 study are of especial interest. A priori it was difficult to conceive that such specificity as was seen in the Prunoideae and the other groups of woody plants studied could be due merely to the interaction of rela- tively non-specific substances. However the distribution of many chemical substances in plants follows in general their relationships. Thus the presence and form of calcium oxalate crystals in plants have been frequently used as taxonomic criteria, and it is manifest from a study of the data on distribution of other substances, such as saponins, glucosides, alkaloids, etc., that homogeneous groups of plants are char- acterized by homogeneity in their content of such substances. Guided by such reasoning it was of interest to examine the distribu- tion of the various reactive principles differentiated in this study to observe whether they showed any species-, genus-, or family-specificity in their distribution among the various plants studied. From the re- sults of Tables XXV-XXVIII it was possible to ascribe to each species its “reactive formula.” This, for example, with Salpiglossis sinuata was: A—B+, M—N—, X—Y—, Ox-+Ca—. These were formulated for each of the species studied and then they were grouped in a tri- furcating dendritic system as in Figure IV. The order of arrangement was guided by the observation that certain of the reactions varied within the genus, others within the family, etc. A survey of Figure IV brings out some very interesting relationships. The five families used in this study fall into three orders, the Rosiflorae, including the closely related Platanaceae, Saxifragaceae, and Legumt- nosae (which three are adjacent in Rehder’s Manual), the Contortae, including the Oleaceae, and the Tubiflorae, represented by the Sola- naceae. With but one questionable exception the Rosiflorae studied all possess A and lack B, the Tubiflorae (Solanaceae) all contain B and lack A, while the Contortae (Oleaceae) all lack both A and B. Prunus lies very near the borderline, and might without difficulty be included in either group. Within the Rosiflorae subsequent division was not attempted because of the difficulty of eliminating the AB reaction in considering the MN and XY reactions. Within the Oleaceae all con- tain excess of Y and of calcium and are separated only by presence or absence of M and N, a highly homogeneous assortment of reactive principles as would be expected from their relative taxonomic homo- geneity. Within the Solanaceae none contain the principle Y (which was found in the Oleaceae) and the majority also lack X, although this is present in a few species. The Solanaceae are differentiated by both the MN and calcium oxalate reactions, the latter, however, being con- 186 JOURNAL OF THE ARNOLD ARBORETUM LVOL. XIV sidered, because of its simple nature, as least significant. Of the 18 species of Nicotiana all are identical in their reactive ingredients with the sole exception of NV. Rusbyi. Datura varies with regard to the XY reaction but is relatively uniform with regard to the presence of N and ( Prams) Salpiglossis D.ferox Capsicum co D.innoxia at Physalis ! pea oo ! Ce, Plateamus Ligustrum Syringa Ligustrum Solanum X. a Petunie Sol. tuberosum Ox a ibota Ligust. obtusifo. Capsic. Sol.nigrum 3.melongen vulgare N.Cavanillesii -Siauca N.glutinosa N. sylvestris N.Tabaocum i. pn rg = “Trigonophy1. Figure IV. CLaSssIFICATION OF ALL EXTRACTS STUDIED ACCORDING TO THE PRESENCE OR ABSENCE OF THE VARIOUS REACTIVE PRINCIPLES. of excess oxalate. The Solanum species are all uniform in their excess of calcium, although there was some individual variation of this in- S. tuberosum, while they are separated by the presence or absence of MN and X. 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 187 Taken on the whole, there seems to be a definite tendency for the various taxonomic units to show relatively homogeneous reactive formulae. There are exceptions, to be sure, but these exceptions are by no means as striking as the agreements. That such a dendritic system as that of Figure IV could result from chance is beyond belief. It is entirely possible that an extension of these results to include many more species might either favorably or unfavorably influence this rela- tionship, but the data presented appear to justify, at the present, the statement that there is a definite indication of a correlation between the distribution of reactive principles in the plants studied and their taxonomic position. Hence the results obtained by the senior writer on the Rosaceae, Saxifragaceae, et al. are interpreted by the results of the present study. If there is even a moderate and imperfect degree of correlation between the systematics of these plants and their complement of reactive prin- ciples, then it necessarily follows that this correlation will be reflected in a corresponding correlation of precipitin reaction with systematics. Before, however, drawing final conclusions regarding the specificity of the normal precipitin reaction in plants it is well to consider the variability of the reaction among individuals of a given species. Thus far none of the published accounts have been concerned with the re- actions of more than one or a few individuals of a given species. Kos- toff does not record the number of repetitions of each normal test. Silberschmidt’s data on the subject are so limited as not to afford sig- nificant comparative data. Chester’s experiments with the woody plants apart from the Oleaceae were based on one or very few extracts for each species, and his results with the Oleaceae were the only ones published in which a generous assortment of individuals of each species was used. In the latter case, however, although his results show the uniformity of the various individuals and varieties of a species, the fact that the basic reaction here is negative weakens the force of the argu- ment regarding the constancy of the precipitin reactions in a species. In order to test this matter an experiment was performed by the writers utilizing numerous individuals of a few selected species but under different experimental conditions. Three questions were investi- gated, namely: What is the effect of age of plant on the precipitin reactions? What is the effect of the addition of calcium and oxalate to the soil upon the precipitin reactions? What variability is there be- tween different horticultural varieties and individuals of a given species? For this experiment cultivated Tomato (Lycopersicum escu- lentum...4 varieties), Pepper (Capsicum frutescens...2 varieties), and 188 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Potato (Solanum tuberosum...2 varieties) were employed. In the testing of varietal differences all 8 varieties were used, likewise all the varieties were used in testing the effect of addition of salts to the soil, while for the tests on age variations the tomato variety Bonny Best was employed. Five or six individuals of each variety were used in each part of the experiment. The experiment on age differences consisted in weekly testing of 5 plants of the tomato variety from the time they were 6-8” in height to maturity. That on the effect of fertilization con- sisted in testing the 48 plants used after having treated one third of them with lime, one third with daily applications of 1:1000 potassium oxalate, and leaving one third untreated as controls. The treatments lasted for one month. The tests which were then performed in every case comprised tests with a N/100 potassium oxalate solution, to deter- mine the calcium oxalate reaction, with Ribes to determine the effect on the AB reaction, and with Nicotiana Rusbyi to observe variations in the MN reaction. The same tubes of Ribes and Nicotiana were used in all the tests. Unfortunately the N. Rusbyi plant selected at the beginning of the experiment, and therefore of necessity employed throughout the whole of the experiment, was negative for the MN re- action. However, the consistency of its negative reactions, under tlhe various experimental conditions is of significance, even though this significance is not as great as though it had been positive. The results of this experiment are gathered in Table XXXVI. Let us consider first the question of variation among the horticultural varieties of a given species. A comparison of the reactions, taken as a whole, of the four tomato varieties shows them to be entirely uniform with reference to each of the reactions. The average reactivity of the four tomato varieties to oxalate was 1.4, 1.4, 1.2, and 1.8 respectively; that of the four tomato varieties to Ribes was 2.0, 1.4, 2.0, and 2.2 re- spectively; that of the four varieties toward Nicotiana Rusbyi was negative in every case. The same reasoning may be applied to a com- parison of the two potato varieties with each other and of the two pepper varieties with each other. In addition to these data, the find- ings of Chester on the reactivity of 18 horticultural varieties of Syringa vulgaris may be cited, in which it was shown that when tested under comparable conditions there was no significant variability of reaction among the varieties employed. We are thus free to conclude that all the evidence thus far obtained indicates that the reactions of any se- lected horticultural variety of a species are approximately equivalent to those of any other selected variety of the same species. Passing next to the question of variation among individuals of a given 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 189 TABLE XXXVI. A. VARIABILITY OF THE ee OXALATE, AB AND MN REACTIONS SHOWING EFFECTS OF INDIVIDUAL AND VARIETAL VARIABILITY pe OF SOIL TREATMENTS. Precipitin reactions Before After Pptin. reaction Tomato Plan & Treat treatment: treatment: Potato Plant Treat- after treatment ment: Ox AB MH & OL EB variety: # Ox AB ME yariety: #£ ment: Pritchard 1 Lime 1-2 0 Zo 2 "0 Green Mt. 1 Lime Oavictis:. 0 LJ 2 Line ave o0: 12 0 Le iz Lime fa) t 3 0 " 3 Oxal. 220 2 <0 = 3 Oral. t 0 3 0 ” & Oxal 2 2 0 220 " 4 OQxale 1 O 2 O " 5 Kone 1-200 120 Ls 5 None Oo oO |2 0 " 6 Fone on 4 3.420 Break 0’ 1 = Lime 132.210 220 Var.? 1 Lin Oo t |2 0 dar 2 Lime 1. 0 ae 2,0 " 2 Lime t Oo = 0 " S$ oral. 1 1 0 12 0 « 3 Qralh 1 0 2 O " 4 Oral. 21 0 12 0 Ls 4 Qxrale 1 0 2 O oy 5 YWone 22 0 12 0 bd 5 Fone ©o 0 2 O Norton 1 Lime 120 1 2 0 P v " 2 Lime 12 0 220 Red Spot ofl Lime 1 0 2 0 e 3 orale 2 20 a 2 -0 sh 2 Lime 1 0 2 Oo " 4 Oxal 1 2 0 22 0 a 3 OQxale 1 0 2 O be 5 Fone 120 22 0 bal 4 Qral. ni 0 2 fe) " 5 None 2 0 1 0 Bonny Best 1 Lime 13 0 12 0 * 6 Hone 1 0 2 0 " 2 Lime 220 120 Chinese 1 Lime 2 0 2 0 be 3 orale 3 20 120 Giant 2 Lime io 32 .0 “J 4 Orxrale 1 2 0 12 0 “ 3 Oxalek 1 0 2 O bad 5 Xone 2e0 12 0 ad 4 Oxalek 1 © 2 O Le 5 None ZO) 2 30 ” 6 Hone 1 0 2 0 TABLE XXXVI. B. VARIABILITY OF THE CALCIUM OXALATE, AB, AND MN REACTIONS IN TOMATO VARIETY “BONNY BEST,” ORDING TO AGE OF PLAN Date of precipitin tests: —May 18 = _May 25 Plant May 5 May 11 el June 8 June 15 : S@2 ae & Gm 4B Ee & “BER & OB ER G@ BEN G@ AB LE 6 2 3 0 2 2 0 2 2 0 1s. 6 7 zy Sr SO 2 2. 0 : ee eee % +3 bo b. By quantitative analysis which showed calcium oxalate to be present in the washed precipitates to the extent of 85-90%. c. By complete elimination of the reactions by precipitating Prunus 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 195 a b mn ON with calcium salts or the other extracts with oxalates respec- tively. This calcium oxalate “precipitin reaction” has been found to occur extensively throughout the woody plants and Solanaceae studied. About half of all the reactions obtained in this study were positive, and of these positive reactions 57% were due to calcium oxalate. 70% of our positive tests of the plants used by Kostoff in this work were found to be due to calcium oxalate. . By now eliminating the calcium oxalate reaction from consideration it was possible to determine whether other reactions were present among these species. Three such other reactions have been dis- tinguished. One, the AB reaction, is characterized by the reactions of the Rosiflorae employed (A) with the Solanaceae (B). It ac- counts for 24% of all the positive tests observed in our material. A second reaction, MN, particularly results from the interaction of Nicotiana Rusbyi and Datura Wrightii (M) with other Solana- ceae (N). It is responsible for 17% of our precipitates. The third reaction, XY, occurs between the Oleaceae (Y) studied and certain of the Solanaceae (X). It is relatively insignificant, being responsible for only 2% of the total reactions observed. The iden- tity and distinctness of these three reactions have been established by various types of experiment. Principles A and B are water-soluble, alcohol-soluble, coctostable, insoluble in lipoid solvents, not coagulable by alcohol, and slowly and imperfectly dialyzable. The precipitate formed from their interaction chars upon ignition. Hence they are plainly non-lipoid and non-protein (because of their alcohol solubility and positive, though imperfect dialysis) but organic and of relatively high mole- cular weight. Principles M and N are soluble in water and weak salt solutions, but are insoluble in alcohol, are non-coagulable by alcohol, insoluble in lipoid solvents, and in contrast to A and B are thermolabile. The precipitate formed by their interaction chars on ignition. M is retained by dialyzing membranes with high effi- ciency, but N passes freely through such membranes. Hence they are both organic and non-lipoid. N is in all probability an organic substance of relatively simple structure and it is highly unlikely that it is protein. M, on the other hand, gives strong indications of being protein or at least a complex organic substance of high mole- cular weight. . The published immunological interpretations of precipitin reactions in plants are rendered untenable because of the lack of homology 196 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV between the animal and plant reactions, and because of the wide- spread occurrence in the plants heretofore used of simple, non- protein reactions. The works of Kostoff, Silberschmidt, and East are analyzed in this connection. . The findings of Chester with regard to the reactions of diseased Lilacs and to the specificity displayed by the normal reactions are interpreted and substantiated by these findings. . The experimental work has shown that the age of the plants does not have a significant effect upon the strength of the calcium oxalate reaction. Treatment of tomato and pepper varieties with lime, and with potassium oxalate did not appreciably alter the strength of the calcium oxalate reaction. The same treatment applied to potato varieties produced a demonstrable change in this reaction. In the later case the slight effect produced by experimental variation of the environment cannot be considered as seriously affecting the signifi- cance of the reactions reported. . The possibility, in the future, of demonstrating immunological re- actions from grafting is not eliminated by this study. There is proof that the reactions thus far reported were doubtless non- immunological in the zodimmunitary sense, but on the basis of the findings of the present study the desirable directions of future work in this field are pointed out. ~sI oo Ke) IX. ACKNOWLEDGEMENT The work reported in the present paper has been very appreciably furthered by the assistance and suggestions of a number of workers in related fields. The experiments in chemical analysis of the extracts were materially aided by the cooperation of Drs. Ronald Ferry and Benjamin White. For numerous suggestions and criticisms throughout the study the writers are indebted to Professors J. H. Faull and E. M. East. To them the writers wish, at the present time, to express their thanks. X. LITERATURE CITED 1. Besrepka, A. (1927). Local immunization. Plotz translation. (Williams and Wilkins, Baltimore, Md 2. CARBONE, D. C, ARNAUDI. (19 ae L’immunita nelle piante. (Monogr. Istituto Sieroterap. Milan Bs a gran K. S. (1931). Graft-blig it: MA disease of lilac related to e employment of certain unde es; samt in propagation. V. F. Spe- cific physico-chemical or immunological differences between stock and scion. Peas Arnold Arbor. 12:114-128.) (1932). Studies on the prec ipitin reaction in plants. I. The specificity: of the normal precipitin reaction. (Ibid. 13:52-74. ) 1933] CHESTER AND WHITAKER, PRECIPITIN REACTION 197 (1932). Studies on the precipitin reaction in plant ice Preliminary report on the nature of the “normal aes reaction.” (Ibid. 13:285-296. 6. Coun, E. J. (1927). The activity coefficients of the ions in certain phosphate solutions. A contribution to the theory of buffer action. (Jour. Amer. Chem. Soc. 49:173-193 7. East, E. M. (1931). Immunity to sugar cane mosaic acquired by the host. (Proc. Nat. Acad. Sci. 17:331-334. 8. Fiscuer, E. & E. GAumMann. (1929). Biologie der pflanzenbe- wohnenden parasitischen Pilze. (Fischer, Jena 9. Gann, . (1928). as Problem der Immunitat im Pflanzen ‘reich, (Vierteljahrsch. Naturforsch. Ges, Zurich. 73, Beitlatt: Festschrift Hans Schinz, 450-468. 10. HEIDELBERGER, M. & O. T. AVERY (1923). The soluble specific substance of pneumococcus. (Jour. Exper. Med. 38:73-79.) 1b & ee R. (1902). Lehrbuch der Intoxikationen. (Bd. 1. p. 161; , p. 695 ff. Stuttgart. Pa Kostorr, D. ( ). Acquired immunity in plants. (Genetics, 1 13, ——————_ 0). Tumors and other malformations on certain Nicotiana Senee (Zentralbl. f. Bakt. Abt. 2, 81:244-260. 14. —————— (1930). Induced immunity in plants by grafting. (Ex- trait du ler Congrés Intern. de Microbiol. Paris, 1930. 15. —————— (1931). Studies on the acquired immunity in plants. contribution to the methods for investigating the mutual activities between graft components. (Svedenia po Zemledelieto, 12:47-69. ) 15a. —————— (1931). ae on the acquired immunity in plants in- duced by grafting. (Zeit. f. tmmunitats®. 74: 16. Kraus, R. (1902). Zur Theorie der pelduncuon. (Zeit. f. Heilk. 17. Mason, V. P. (1922). Relation of pH to specific precipitation. ot ns Hopkins Hosp. Bull. 33:116. 18. Mu H. & E. FrAeEnKeEL (1911). (Cited in Much, H. Tuberku- eeerae sch. in Muenchener Med. Wochenschr. 78(1931) :137-139.) 19. OsxporneE, T. B. (1924). The vegetable proteins. Ed. 2 (Longmans, Green & Co. London. 20. Rimincton, C. (1929). The isolation of a aoe derivative from serum proteins. (Biochem. Jour. 23:430-443. ) 21. S1LBERSCHMIDT, K. (1931). Studien zum Na wea von Anti- korpern in Pflanzen. I. (Planta: Arch. f. Wiss. Bot. 13:114-168.) 22. Warp, M. 5). Recent researches on the parasitism of fungi. (Ann. Bot. 19:1-54). ARNOLD ARBORETUM, HARVARD UNIVERSITY. JOURNAL OF THE ARNOLD ARBORETUM VoLUME XIV. | 95 Sh ee AS: NuMBER 3 NEW SPECIES, VARIETIES AND COMBINATIONS FROM THE HERBARIUM AND THE COLLECTIONS OF THE ARNOLD ARBORETUM? ALFRED REHDER With nine text figures Celtis Rockii, sp. nov. Arbor 6-metralis, ramulis hornotinis dense flave crispulo-villosis angulatis vel subangulatis lenticellatis, annotinis tarde glabrescentibus; gemmae ovoideae, acutae, perulis adpresse flavo-pilosis. Folia papyra- cea, elliptico- vel rhombico-ovata, 4-8 cm. longa et 2.3-4.5 cm. lata, breviter acuminata, basi plus minusve oblique late cuneata vel fere rotundata, a medio vel infra medium ad apicem dentato-serrata, supra accumbenti-pilosa et scabrida, subtus tota facie satis dense molliter crispulo-villosa, ad nervos pilis longioribus patentibus flavidis instructa, triplinervia, nervis utrinsecus basalibus inclusis 2-3; petioli flave crispulo-villosi, 3-5 mm. longi. Flores non visi. Racemi fructiferi axillares in parte inferiore ramulorum plerique triflori, pedicello termi- nali incluso 1-1.5 cm. longi, dense flave crispulo-villosi, pedunculo 2-5, pedicellis 3-8 mm. longis; drupa subglobosa, parva, plus minusve villosula, vel maturitate glabra vel fere glabra, lutea (ex collectore) in sicco atro-fusca; putamen subglobosum, circ. 4 mm. diam., manifeste punctulato-foveolatum et leviter costatum. CHINA. Yunnan: region of Tungshan, Yangtze drainage basin, east of Likiang, J. F. Rock, no. 10522 (type), in 1923 (tree 20 ft.; fruits yellow). This new species seems closely related to C. Salvatiana Schneid., and also to C. cinnamomea Lindl., but from both it is easily distinguished by the rather dense pubescence of the leaves, the densely pubescent branchlets and inflorescence and the pubescent fruit; from C. cinna- momea it differs further in the serrate broader leaves, the few-flowered short-peduncled inflorescence and smaller fruit. 1Continued from vol. xm. 341. 200 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Clematis chinensis Osbeck, Dagb. Ostind. Resa, 205, 242 (1757); Reise Ostind. China, 267, 315 (1765); Voy. China East Ind. 1. 329; 11. 356 (1771).—Retzius, Observ. 11. 18, t. 2 (1781).—De Candolle, Syst. 1. 137 (1818); Prodr. 1. 3 (1824).—Forbes in Jour. Bot. xx11. 262 (1884).—Hemsley in Jour. Linn. Soc. xxi. 3 (1886).—Pritzel in Bot. Jahrb. xxrx. 332 (1900).—Finet & Gagnepain in Bull. Soc. Bot. France, L. 535 (1903); Contrib. Fl. As. Or. 1. 20 (1905), excl. synon. C. terniflora et C. Benthamiana.—Rehder & Wilson in Sargent, Pl. Wil- son. 1. 329 (1913); in Jour. Arnold Arb, vir. 106 (1927).—Léveillé, Fl. Kouy-Tchéou, 332 (1915) ——Merrill in Jour. Am. Bot. ur. 579 (1916).—Rehder in Jour. Arnold Arb. x. 187 (1929). Clematis sinensis Loureiro, Fl. Cochinch. 345 (1790). Clematis minor Loureiro, |. c. (1790).—De Candolle, Syst. 1. 136 (1818); Prodr. 1. 3 (1824).—Forbes in Jour. Bot. xx11. 263 (1884). eis recta § chinensis Kuntze in Verh. Bot. Ver. Brandenb. xxv1. 4 (Monog. Clem.) (1885). Pei pong! oe & Vaniot in Bull. Acad. Intern. Géog. Bot. x1. 168 (19 Clematis phi Pedi & Vaniot, l. c. xvi. no. 210-11, p. il (1907).—Léveillé, Fl. Kouy-Tchéou, 333 (1915 Clematis Cavaleriet Léveillé & Porter in Fedde Rep. Spec. Nov. 1x. —Léveillé, Fl. Kouy-Tchéou, 332 (1915). The fact that the generally accepted name Clematis chinensis Retzius would be invalidated by the older homonym C. chinensis Osbeck, if the two names referred to different species, has led me to investigate this question and I find that the two names are synonymous, as they have been already treated by a few authors. This will save the specific epithet chinensis for the species generally known under this name, and makes necessary the change of the author citation only, so that it will be C. chinensis Osbeck (1757) instead of Retzius (1781). Owing to the fact that Osbeck’s name was published with a rather incomplete description hidden away in the text of a work not primarily taxonomic, it has been neglected, while Retzius’ name published with an adequate description accompanied by a plate has been universally accepted. Also the misleading citation in Index kewensis of C. chinen- sis Osbeck as a synonym of C. recta L. may have caused the disregard of the name by later authors. Moreover E. D. Merrill (1. c.) has tried to identify C. chinensis Retz. with C. Meyeniana, but the description given by Osbeck (1. c. p. 205) “Plurima habet communia Clematide Vitalba, at folia lanceolata, angustissima, & flores minores” agrees much better with C. chinensis Retz. which has pinnate leaves like C. Vitalba with narrower leaflets and decidedly smaller flowers, while C. Meyeni- 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 201 ana has ternate leaves with large subcoriaceous leaflets, quite different from the pinnate leaves of C. Vitalba, and flowers scarcely smaller than those of the latter species. Additional data given by Osbeck (I. c. p. 242) “Pistilla 3 ad 6, stylis plumosis in orbem positis reflexis. Stam. O obseruavi. Frutex scandens, ramosissimus’” may apply as well to C. chinensis as to C. Meyeniana. In his more recent Commentary on Loureiro’s Flora cochinchinensis (msc.), however, Merrill identifies Clematis minor Lour. with C. chinensis Osbeck and states that accord- ing to his opinion C. chinensis Retz. and C. chinensis Osbeck are identi- cal; he also considers C. Benthamiana Hemsl. a synonym, following Finet & Gagnepain, which I keep separate (see Rehder & Wilson, 1. c.). The identity of C. chinensis of Osbeck and of Retzius is proven con- clusively by specimens before me, for the loan of which I am in- debted to those in charge of the herbarium of the State Museum in Stockholm. One of these specimens is marked on the back of the sheet “China: Osbeck,” a second “Herb. Swartzii—Osb.” and a third “ex Ind. Orient.,” the last note may have reference to the fact that it was collected during Osbeck’s East Indian voyages, but as the species does not occur in East India, the specimen probably came from China. l these specimens represent very early collections and were probably all collected by Osbeck; they are all named Clematis chinensis Retz., the determinations being apparently of a later date, for two of them bear the citation “DC.” The specimens vary somewhat in the shape and size of the leaves and only one has leaflets as narrow and small as in Retzius’ plate of C. chinensis. Clematis chinensis is widely distributed in southeastern and central China extending west to Szechuan and Kweichou. In this herbarium there are specimens from the following provinces: Kwangtung, Fukien, Hunan, Anhwei, Chekiang, Kiangsi, Hupeh, Szechuan and Kweichou; there is also a specimen from Annam, and a photograph of the type of C. minor Lour. from Cochinchina. Clematis grata Wall. var. likiangensis, var. nov. A typo recedit achaeniis glabris——Ramuli, petioli et inflorescentiae laxe villosula. Foliola ovata, trilobata lobis grosse paucidentatis, supra glabrescentia, subtus in costa venisque densius, in venulis sparsius et in facie sparsissime flavido-pilosis; flores ut in typo carpellis glabris exceptis. CuIna. Yunnan: Yangtze watershed, Prefectural district of Likiang, eastern slopes of Likiang Snow Range, J. F. Rock, nos. 3668 (type) and 3918. Clematis grata apparently varies like the related C. Gouriana Roxb., 202 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV and C. brevicaudata DC. with pubescent and glabrous akenes (C. Gouriana var. Finettt Rehd. & Wils. and C. brevicaudata var. lissocarpa and var. subsericea Rehd. & Wils.), which shows that the pubescence of the akenes is a character of secondary importance and cannot be used to define subdivisions of the genus. Deutzia Esquirolii (Lévl.), comb. nov. Styrax Esquirolii Léveillé in Fedde Rep. Spec. Nov. 1x. 446 (1911). Deutzia lancifolia Rehder in Sargent, Pl. Wilson. 1. 147 (1912); Jour. Arnold Arb. x11. 276 (1931).—Léveillé, Fl. Kouy-Tchéou, 387 (1915). Deutzia Chaffanjoni Léveillé, 1. c. (1915), pro synon. D. lancifoliae hd e : Deutzia Esquiroli (Lévl.) Léveillé, 1. c. (1915), pro synon. D. lancifoliae Rehd. When dealing with D. lancifolia in my Notes on the ligneous plants described by Léveillé (Jour. Arnold Arb. x11. 276) I overlooked that unfortunately Léveillé’s Styrax Esquirolii is one year older than my Deutzia lancifolia and that the new combination resulting from the transfer of the specific epithet of his name should be the valid name for the species. Léveillé himself had already published this combina- tion, but only as a synonym of D. lancifolia. Hydrangea umbellata Rehder in Sargent, Pl. Wilson. 1. 25 (1911). pee ioe meee ew in Engler & siren hs Pflanzenfam. ed. 2, xv , 203 (1930), pro parte —Syn Hydrangea Schindleri was only briefly agree in ‘hacia? s account of the species of Hydrangea (1. c.) without enumeration of specimens. Engler compared it with H. chinensis Maxim. and H. umbellata Rehd. In the Berlin Herbarium there are four specimens labeled H. Schindleri all collected by A. K. Schindler in August-September 1908 at Lu-shan, Kuling mountains, Kiangsi, the type locality of H. umbellata. Two of them, nos. 325 and 327, I cannot distinguish from H. umbellata while the other two numbers belong to the following species: Hydrangea paniculata Siebold in Nov. Act. Acad. Leop.-Carol. xIv, pt. 1. 690 (Syn. Hydr.) (1829). Hydrangea Schindleri Engler in Engler & dienes he Pflanzenfam. d. 2, xvi11-a, 203 (1930), pro parte — S$ Of the four numbers collected by Schindler at iu ae Kiangsi, and named by A. Engler H. Schindleri two belong to the preceding species, while the other two, nos. 322a and 324, are identical with H. paniculata Sieb. No. 324 bears the following note in A. Engler’s handwriting, “Hydrangea Schindleri Engl. n. sp., affinis Hydr. chinensi Maxim., differt foliis ab infima triente sursum angustatis, haud e medio utrinque angustatis, distinctius serratis, florum sterilium sepalis ovatis angusti- 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 203 oribus.” This seems to show that this specimen should be considered the type of H. Schindleri Engl., since the preceding characterization is apparently based on Schindler’s no. 324 rather than on any of the other numbers and is the same as given in German in the Pflanzen- familien. In 1911 in my Synopsis of the Chinese species of Hydrangea (in Sargent, Pl. Wilson, 1. 25, 1911) I stated that Wilson’s no. 1601, col- lected at Kuling, July 27, 1907, was to my knowledge the first speci- men of H. paniculata collected in China. Since then, however, many additional specimens have come to this herbarium and the species is now known from the following Chinese provinces: Kiangsu: Yii-du-hsien, H. H. Hu, no. 1179. Anhwei: Chu-hwa-shan, R. C. Ching, no. 2808; Wu-yen, VN. K./p,no. 7675. Chekiang: Tsing- Tien, Taishun-hsien and Chang-shan-hsien, Y. L. Keng, nos. 172, 310 and 841; Pang-yung, R. C. Ching, no. 2099; East Tien-mu, H. H. Hu, no. 1609. Kiangsi: Kuling, E. H. Wilson, no. 1601; Lu-shan (Kuling), A. K. Schindler, nos. 322a, 324; Lu-shan, A. N. Steward, no. 2613; Ningdu, Wang-Te-Hui in Handel-Mazzetti, Pl. Sin., no. 442. Fukien: Yenping, H. H. Chung, nos. 2844, 3301, 3556 and 3659. Kwangtung: Lokchong, Y. Tstang, no. 1219; between Bei-shen and Nan-shung, W. Y. Chun, no. 5683; road to Chang-kiang, W. Y. Chung, no. 5794; Siudsao, R. Mell, no. 1773. Hunan: Wukang, Handel-Mazzetti, no. 12527. Kweichou: Lou-tsong-koan, E. Bodinier, no. 1661; Kwei-yang, Handel-Mazzetti, no. 10478; Kweiting, Y. Tstang, no. 5627. Yunnan: Yunnanfu, O. Schoch, no. 423. The specimen collected in 1897 by E. Bodinier in Kweichou was de- scribed by Léveillé as H. Kamienskii (cf. Jour. Arnold Arb. x1. 277). Spiraea yunnanensis Franchet, Pl. Delavay. 200 (1890).—Schnei- der, Il. Handl. Laubholzk. 1. 463 (1905). Spiraea sinobrahuica W. W. Smith in Not. Bot. Gard. nates Mg (1917) ; xiv. 233, 260 (1924) ; xvir. 388 (1930).—Synon. n Spiraea stnobrahuica var. aridicola W. W. Smith in Not. Bot need Edinb. x. 68 (1917) ; xv11. 197, 363 (1930).—Synon. nov. Cutina. Szechuan: in valle fl. Ming, inter stationes Sim-pu- guanj et Pei-schuy-tchan, G. N. Potanin, Aug. 25, 1873 (frutex usque metralis); inter Tatsien-lu et Batang, ad stationem Natschuka sive Nachtschuka, V. Kashkarov, May 19, 1893 (frutex plus quam metra- lis); between Batang and Tschien-lu, John R. Muir, in 1911; Muli kingdom, Shou-chu valley, alt. 2435-2900 m., J. F. Rock, no. 16279, June 1928. (shrub 1-1.5m.) Yunnan: terrains calcaires, pierreux au dessus de Mo-so-yn, Lankong, alt. 2200 m. J. Delavay, no. 1082, (holotype of S. yunnanensis, photo. and fragments in A. A.), May 1, 204 JOURNAL OF THE ARNOLD ARBORETUM LVOL. XIV 1884 (arbrisseau d’un métre; fleurs blanches); eastern flank of the Lichiang range, Lat. 27° 10’ N., alt. 9000-10500 ft., amongst the scrub in side valleys, G. Forrest, no. 5580 (syntype of S. sinobrahuica; ex W. W. Smith, 1]. c.); descent of the Yangtze from the eastern boundary of the Lichiang valley, lat. 27° 15’ N., alt. 9000-10000 ft., G. Forrest, no. 10117 (syntype of S. sinobrahuica), June 1913 (shrub 4-5 ft.; flowers creamy white); descent of the Yangtze valley from the eastern range of the Lichiang valley, Lat. 27° 30’ N., alt. 9000-10000 ft., G. Forrest, no. 10084 (syntype of S. sinobrahuica var. aridicola), June 1913 (shrub 4-6 ft.; flowers creamy white); mountains of Chungtien plateau, lat. 27° 30’ N., alt. 11000 ft., G. Forrest, no. 12634 (syntype of S. sinobrachuica var. aridicola; photo. in A. A.); open stony slopes and on ledges of dry cliffs on the western flank of the Lichiang range, Lat. 27° 40’ N., Long. 100° 18’ E., alt. 10-11000 ft., G. Forrest, no. 21171, May 1922 (shrub 3-5 ft.; flowers creamy-white); Yangtze val- ley, northwest of Likiang, Lat. 27° 20-30’ N.; alt. 2000-2100 m., Handel-Mazzetti, no. 8792, June 2, 1916; Yangtze watershed, Prefec- tural district of Likiang, eastern slopes of Likiang snow range, J. F. Rock, no. 3639, May-Oct. 1922; western slope of Likiang snow range, Yangtze watershed, J. F. Rock, no. 8557, April 1923 (shrub forming globose bushes); Lotueshan, mountains of Labako, west of Yangtze bend at Shiku, J. F. Rock, no. 8471, April 1923 (shrub 3-4 ft.; flowers white); dry rocky slopes and on cliffs on the Chien-chuan-Mekong divide, Lat. 26° 36’ N., Long. 99° 40’ E., alt. 9-10000 ft., G. Forrest, no. 21465, July 1922 (shrub of 2-3 ft.; flowers white); open dry rocky slopes and ledges of cliffs on the Mekong-Salween divide, Lat. 28° 12’ N.; alt. 7-9000 ft., G. Forrest, no. 16410, May 1918 (shrub 1-3 ft., flowers creamy white). Southeastteern Tibet: Tsarong, open stony situations on the ledges of cliffs on the Salween-Kiu-chiang divide, Lat. 28° 40’ N., alt. 7000 ft., G. Forrest, no. 18881, July 1919; on ledges and in crevices of cliffs and dry bouldery slopes on the Salween-Kiu-chiang divide, Lat. 28° 40’ N., Long. 98° 15’ E., alt. 8000 ft., G. Forrest, no. 19147, Sept. 1919 (shrub 1-2 ft., widely branched). I have been unable to find any characters to separate Spiraea sino- brahuica W. W. Sm. from S. yunnanensis Franch. The chief differ- ence given by the author of the former name, the glabrous upper surface of the leaves of S. yunnanensis, does not hold, since the leaves of the type specimen of S. yunnanensis show the same pubescence of short accumbent hairs on their upper surface as the types of S. sinobrahuica, though Franchet describes them as glabrous, a statement which induced the author of S. sinobrahuica to consider Forrest’s specimens distinct. It also does not seem possible to separate var. aridicola from the type. 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS — 205 The species which is closely related to S. brahuica Boiss. is very vari- able in the size, in the serration and to some extent in the shape of the leaves, in the size of inflorescence and in habit; the two extremes merge imperceptibly into each other and are apparently only individual dif- ferences caused by difference of exposure and soil. The only form which seems worthy of a distinct name on account of its striking habit is the following. Spiraea yunnanensis f. tortuosa (Rehd.), forma nova. Spiraea tortuosa Rehder in Sargent, Pl. Wilson. 1. 445 (1913). Cutna. Szechuan: Mao-chou, arid regions of the Min val- ley, alt., 13-2000 m., E. H. Wilson, no. 2764 (holotype of S. tortuosa), May 25, 1908 (shrub 3-4 ft.; flowers white). This differs strikingly from the type in its distinctly zigzag branch- lets, the perfectly straight internodes forming sharp angles of about 130-150° with each other; the leaves are suborbicular to broadly oval, more or less 3-lobed and scarcely exceed 12 mm. in length; the in- florescence is 5-12-flowered. The only specimens enumerated under the type which show any approach to zigzag branchlets, are Potanin’s specimen from the same region and John R. Muir’s specimens from western Szechuan. The specimens described as S. sinobrahuica var. aridicola may be considered as being nearest to f. tortuosa. Spiraea siccanea (W. W. Sm.), spec. nov. Spiraea yunnanensis Fr. var. siccanea W. W. Sm. in Not. Bot. Gard. Edinb. x (1917). Frutex 1-1.5 m. altus ramis gracilibus teretibus, hornotinis adpresse villosulis, annotinis glabris purpureo-fuscis partim decorticantibus; gemmae ovoideae, pluriperulatae perulis ovatis glabris ciliolatis. Folia ramulorum floriferorum (turionum non vidi) obovata vel ovalia, rarius oblongo-ovalia, 0.8-2 cm. longa et 6-10 mm. lata, apice rotundata vel obtusa, mucronulata, basi late cuneata vel fere rotundata, supra medium inaequaliter dentata vel crenato-dentata dentibus mucronulatis, inter- dum indistincte trilobata, supra laete viridia et glabra, subtus glauca ad costam nervosque tantum laxe pilosa, basi 3- vel 5-nervia, ceterum utrinque nervis 1-2 instructa; petioli 1-2 mm. longi, villosuli. Flores albi, subumbellati, circiter 12-20 ramulos paucifoliatos pedunculo 5-10 mm. longo incluso 1-2 cm. longos terminantes; pedicelli 5-10 mm. longi ut pedunculus villosuli; calyx turbinatus, circ. 1 mm. longus, ut lobi triangulares acutiusculi subaequilongi laxe villosus; petala orbiculari- obovata, 3-4 mm. longa; stamina circ. 20, dimidia petala aequantia; discus conspicuus, annularis, 10-lobatus; carpidia villosula, stylo apicali fere 1 mm. longo coronata. Fructus non vidi. Cuina. Yunnan: Lang-kong-Hoching mountains, Lat. 26° 16’ 206 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV N., alt. 8000 ft., open dry situations, G. Forrest, no. 9912 (syntype of S. yunnanensis var. siccanea), May 1913 (shrub of 3-5 ft.; flowers white) ; Lang-kong-Hoching divide, Lat. 26° 10’ N., alt. 8000 ft., dry stony situations amongst scrub, G. Forrest, no. 9972 (syntype of S. yunnanensis var. siccanea) May 1913 (shrub of 3-5 ft.; flowers creamy white). This seems to be distinct enough to be specifically separated from S. yunnanensis Franch. from which it is readily distinguished by the leaves being quite glabrous above and only loosely pilose on the veins neath; no forms intermediate in pubescence were found among the numerous specimens seen of S. yunnanensis, except perhaps one speci- men collected between Batang and Tachienlu by John R. Muir, in which the pubescence though rather slight, is present nevertheless on both sides; at the same time the specimen has very small, 4-8-flowered in- florescences, very small leaves and slightly tortuous branchlets. Malus Rockii, sp. nov. Arbor 8-10 m. alta, ramis pendulis; ramuli hornotini villosi, annotini glabrescentes fusci vel fusco-rubri; gemmae ovoideae, perulis atrofuscis ovatis medio villosis ceterum fere glabris. Folia chartacea, elliptica vel ovato-elliptica, ovata ad ovato-oblonga, 6-12 cm. longa et 3.5-7 cm. lata, acuminata, basi rotundata, rarius late cuneata, argute et adpresse inaequaliter serrulata, dentibus mucronato-acuminatis, supra costa sparse villosa excepta glabra et in costa et nervis glandulosa, impresso- reticulata, subtus pallida, in costa, nervis et venulis manifeste elevatis crispo-villosa, interdum sparse in facie villosula, venulis trabecularibus conspicuis; petioli 2-4 cm. longi, villosi. Flores non visi. Fructus 2-4 vel solitarii, pedicellis 2-4 cm. longis villosis suffulti, ovoidei vel sub- globosi, basi in petiolum abrupte attenuata, apice juniores tandem plus minusve leviter attenuati 1-1.5 cm. longi, calyce tarde deciduo, juniores apice et basi villosuli lobis calycinis lanceolatis extus intusque villosis partim coronati, maturitate carminei, luciduli, 5-loculares. CuHinaA. Yunnan: west of Talifu, Mekong watershed, en route to Young-chang and Tengyueh beyond Lampba, along watercourses, alt. 7000 ft., J. F. Rock, no. 6842 (type), Sept.-Oct. 1922 (tree with long drooping branches; fruits carmine, cherry-like); Litiping range, Mekong-Yangtze divide, east of Weihsi, J. F. Rock, no. 11552, Oct. 1923 (tree 25 ft.); Yangtze watershed, western slopes of Likiang Snow Range, J. F. Rock, no. 5346, May 30-June 6, 1922 (tree 35 ft.). This new species is apparently nearest to M. baccata (L.) Borkh. but the fruits are larger, the calyx is tardily deciduous and the leaves are rather densely pubescent and reticulate beneath. From M. pumila 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 207 Mill. which it resembles somewhat in its leaves, it is farther removed by the deciduous calyx and the slender-stalked small fruit not impressed at the base and the apex; also the leaves are more strongly reticulate and rounded at base. One might compare M. Rockit with the hybrids be- tween M. baccata and M. pumila or M. prunifolia (Willd.) Borkh., but the calyx seems to be always deciduous and the leaves are pubescent and reticulate beneath; moreover, hybrids between these two northern species cannot be expected to occur in Yunnan even as escapes from cultivation, and the three specimens cited above are apparently from spontaneous trees. Malus hupehensis (Pamp.), comb. nov. Pirus communis Pavolini in Nuov. Giorn. Bot. Ital. xv. 415 (1908).— Non Linnaeus. Pirus hupehensis Pampanini in Nuov. Giorn. Bot. Ital. n. ser. xv1I 291 (1910).—Rehder in Sargent, Pl. Wilson. 11. 265, 300 (1915). Pyrus baccata Hemsley in Jour. Linn. Soc. xxu1r. 255 Seasep a plantam e Chekiang.—Diels in Bot. Jahrb. xx1x. 387 (1900).— on Linnaeus. Pyrus spectabilis Hemsley in Jour. Linn. Soc. xxii. 258 (1 ian quoad plantam e Kiangsi et Hupeh.—Diels in Bot. Jahrb. xx1x Malus baccata var. himalaica Schneider, Ill. Handb. Laubholzk. 1. 721, fig. 397s (1906), cree plantam chinens. et fig—Non Pyrus baccata var. himalaica Maxi Malus theifera Rehder in Seek Pl. Wilson. 11. 283 (1915); in Jour. Arnold Arb. v. 192 (1924); viz. 121 (1927); Man. Cult. Trees Shrubs, 395 (1927)—Chun, Chin. Econ. Trees, 173, fig. 65 (1922).—Hers in Jour. N. China Branch R. As. Soc. Li. 116 (1922) ; Liste Ess. Lign. Honan, 29 (1922) —Chung in Mem. Sci. Pyrus theifera (Rehd.) Bailey in Rhodora, xvii1. 155 (1916); Stand. Cycl. Hort. v. 2872 (1916).—Kew Handlist Trees Shrubs, ed. 3, oto Crono): It is rather unfortunate that the name of this species which as Malus or Pyrus theifera is already well known in horticultural literature as a highly ornamental Crabapple, has to be changed, but when examining last year in the Biondi herbarium at the Botanical Museum in Florence the type of Pampanini’s Pirus hupehensis, I saw at once that this species is identical with my Malus theifera. When describing the latter species I had not seen a specimen of Pampanini’s species, which accord- ing to the author’s remarks was most closely related to P. pashia Buch.- Ham. and also to P. communis L. Owing to the world war, I was un- able to obtain a specimen of the species from Florence and I, therefore, 208 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV mentioned (1. c. 265, 300) P. upehensis among the doubtful species of Pyrus, but stated that it could not belong to Pyrus in the restricted sense, since the author described it as having three connate styles villous below. Later Pampanini apparently revised his opinion regarding the affinity of this species, since on a note dated December 1921 and pinned to the sheet of each type specimen he referred it to Malus baccata var. himalaica Schneid., making a new binomial combination of that variety under Pirus which, however, was never published. This identification, of course, came much closer to the true relationship of his P. kupehensis. Malus hupehensis is widely distributed in mountainous regions of China at elevations of from 1000-2000 m. and extends south into Assam. It is represented in this herbarium by specimens from the following Chinese provinces: Shantung, Kiangsu, Honan, Chekiang, Kiangsi, Fukien, Hunan, Hupeh, Szechuan, Kweichou and Yunnan; also from Assam. The syntypes of P. kupehensis Pamp. were collected in north- ern Hupeh, Sian-men-kou (Silvestri, no. 939) and Ma-pau-scian (Sil- vestri, nos. 940, 9402); of nos.‘939 and 940 there are photographs in this herbarium. The holotype of M. theifera also came from Hupeh, near Ichang (Wilson, no. 451), and the paratypes from other localities in Hupeh, from Shensi, Chekiang, Szechuan and Assam. A form with rosy-pink flowers is the following: Malus hupehensis f. rosea (Rehd.), comb. nov. Malus theifera £. rosea Rehder in Sargent, Pl. Wilson. 11. 284 (1915) ; Man. Cult. Trees Shrubs, 395 (1927 Pyrus theifera var. rosea (Rehd.) Bailey in Rhodora, xviii. 155 (1916) ; Stand. Cycl. Hort. v. 2872 (1916). This form has been found in Hupeh; the type comes from Fang Hsien, (Wilson, no. 2980) and a paratype from Patung Hsien (Wilson, Veitch Exped. seed no. 766, Oct. 1900; specimen from Kew Bot. Gard., Wm. Bean, May 1914). Prunus Slavinii Palmer, hybr. nov. Prunus angustifolia var. varians Wight & Hedrick & P. gracilis nge Gray. Frutex, 4-12 dm. altus, raro arborescens et ad 2-2.5 m. altus, dense ramosus, ramulis spinescentibus, novellis brunneo-rubescentibus glabris vel pubescentibus, vetustioribus cinereo-brunneis. Folia lanceolata vel ovato-lanceolata, tenuia sed firma, supra fere glabra, infra glabra vel pubescentia, venulis reticulatis; petioli graciles, 1-1.5 cm. longi glabri vel pubescentes, eglandulosi vel raro glandulosi. Flores 2-6-umbellati, pedicellis 9-12 mm. longis glabris vel pubescentibus. Fructus ovoideus vel subglobosus, 1.5-2.2 cm. longus, 1-2 cm. latus, ruber, pallide punc- tatus, vel rubro-luteus, succosus, esculentus. 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS — 209 Slender or arborescent shrubs .5 to 1 m. or rarely 2 to 2.5 m. tall, with numerous spreading or ascending somewhat spinescent branches, those of the last year’s growth reddish-brown. Bark on old stems and branches dark gray or gray-brown, with pale lenticels. Leaves lanceo- late or ovate-lanceolate 3.5-7 cm. long, 1-2.5 cm. wide, rounded or slightly subcordate at base, rounded, acute or short-acuminate at apex, finely serrate with shallow gland-tipped teeth, bluish-green and glabrous or sparsely pubescent above, paler and usually more densely pu- bescent beneath, sometimes only along the prominently reticulate veins, thin but firm in texture, on slender eglandular or rarely glandular petioles. Flowers appearing in March or early April before the leaves in 2-6-flowered umbels, on slender pubescent or nearly glabrous pedi- cels; ovary usually somewhat pubescent, rarely glabrous; calyx-teeth lanceolate, usually with entire margins, glabrous or slightly pubescent without, pubescent within; petals ovate, clawed, 3-4 mm. long; stamens numerous; anthers yellow or rarely red. Fruit ovoid or nearly globose, 1.5-2 cm. long, 1-2 cm. broad, bright to dark crimson with pale dots, or orange-yellow with red cheek; flesh yellow, becoming soft and suc- culent. Stone compressed-ovoid, 10-12 mm. long, 9-10 mm. wide, rounded at base, pointed at apex, slightly keeled and grooved on ventral side. Growing in thickets, in sandy ground, within the range of the parent species and apparently always in close association with them. Range, from the Arkansas River valley in southeastern Kansas, through cen- tral Oklahoma, and probably to be expected also in eastern and central Texas. NortH AmerIcA. Kansas: Arkansas City, Cowley Co., B. H. Slavin, no. 164, April 10, July 4, 1914; E. J. Palmer, no. 21254, May 11, 1922. Oklahoma: Sapulpa, B. H. Slavin, no. 132, April 1, June 29, 1914; Muskogee, B. H. Slavin, no. 128, March 31, June 26, 1914; Oklahoma City, B. H. Slavin, no. 143, July 2, 1914, no. 144, April 4, July 2, 1914, no. 145, April 4, July 2, 1914, no. 146, April 4, July 2, 1914, no. 152, April 5, July 1, 1914 (type), no. 252, March 27, June 29, 1916; Norman, B. H. Slavin, no. 251, March 27, June 29, 1916; Chicka- saw, B. H. Slavin, no. 257, March 29, 1916, no. 259, March 29, June 30, 1916, no. 260, June 30, 1916, no. 262, March 29, June 30, 1916; Kingfisher, B. H. Slavin, no. 329, April 12, July 5, 1916, no. 330, April 12, July 5, 1916, no. 331, April 12, July 5, 1916, no. 332, April 12, July 5, 1916, no. 334, July 5, 1916; Anadarko, E. J. Palmer, no. 12601, July 20, 1917. Also cultivated in the Arnold Arboretum and in Durand- Eastman Park, Rochester, N. Y. The Chickasaw Plum (Prunus angustifolia Marsh.) is widely dis- 210 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV tributed in the southern states, from Maryland and Delaware to Florida, Oklahoma, and Texas. In the western part of its range, the var. varians Wight & Hedrick, is the commoner. form, and it is found abundantly in sandy soil in central Kansas, Oklahoma, and northwest- ern, central and eastern Texas, where it often forms large thickets of spiny shrubs, 4 to 6 or eight feet in height. The yellow or red fruit matures early and is often of excellent quality. The leaves are pre- vailingly lanceolate, thin, nearly or quite glabrous, and with only the mid-veins prominent. They are usually conduplicate, making them appear narrower than they really are. The Sand Plum (Prunus gracilis Engelm. & Gray) is found in the western part of the same range, from the valley of the Arkansas River, in southeastern Kansas, and along the western border of southern Arkansas, through most of Oklahoma and eastern Texas as far west as the Brazos River. It is a low slender shrub, usually from 1 to 4 feet in height. The leaves are oval or ovate, gray-green, of firm texture, slightly pubescent above and densely so beneath, and with prominent reticulate veins. The fruit is slightly smaller than that of the Chicka- saw Plum, and is edible. It is sometimes borne in such profusion as to weigh the slender branches to the ground. The two species bloom simultaneously and apparently hybridize freely, judging by the number of specimens found. Prunus Slaviniit is quite intermediate in habit and character between the two parent species, and different individuals differ considerably in size and in the pubescence and prominence of the reticulation of the leaves. The specific name is for Mr. B. H. Slavin, superintendent of the splendid park system of Rochester, N. Y., who first collected this interesting Plum and brought it into cultivation there and at the Arnold Arboretum. The type specimen and all of the other numbers cited here are in the herbarium of the Arnold Arboretum. ERNEST J. PALMER Calophaca sinica, sp. nov. Frutex erectus ramis robustis, hornotinis dense albido-pubescentibus, annotinis cortice purpureo-fusco laminis soluto ochraceo-albidis. Stip- ulae scariosae diu persistentes. Folia pinnata, cum petiolo 3-5 cm. longa, pleraque 7-foliolata; foliola chartacea, ovalia vel obovato-ovalia, 12-18 mm. longa et 7-12 mm. lata, apice roundel vel truncata, basi rotundata et saepe leviter subcordata, supra cinereo-viridia, maturitate fere glabra, subtus pallidiora, minute et laxe villosula, reticulo venu- larum satis manifesto, nervis utrinsecus 5-6, supra leviter subtus magis elevatis; petioluli villosi, 1 mm. breviores; petioli 5-12 mm. longi, ut rhachis albido-villosuli. Pedunculi circiter 4 cm. longi longe patentim 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 211 villosi et apicem versus stipitato-glandulosi, pauciflori; flores non visi; legumen oblongum, circ. 3 cm. longum, villosum, glandulis longe stip- itatis crebris munitum, calyce villoso et stipitato-glanduloso lobis lin- eari-lanceolatis tubum subaequantibus suffultum. Cuina. Shansi: Chiao-cheng hsien, alt. 3000 ft., W. Y. Hsia, no. 1032, May 13, 1929. The discovery of this new species in Shansi is of interest particularly for the reason that it extends the range of the genus farther to the East into the flora of China. The genus in its restricted sense (excl. of sect. Chesneya) ranges from southern Russia and the Caspian region through Furkestan (C. wol- garica Fisch.), Bokhara (C. grandiflora Reg.) Tian-shan (C. wolgarica var. tianschanica [B. Fedtch.| Popov) to Dsungaria (C. Hovenii Schrenk) and now extends into northeastern China. From C. wolgarica the new species differs chiefly in the fewer and larger leaflets, the absence of glands from the rhachis and the lower half of the peduncle; from C. grandiflora in the fewer leaflets, few- flowered raceme and the stalked glands of the legume and from C. Hov- enti chiefly in the larger leaflets the spreading pubescence and the presence of stipitate glands on the inflorescence and the legume. Acer sect. Macrantha Pax, emend. Rehd. The section MAcRANTHA is one of the most difficult of the sections of Acer on account of the rather uniform character of the inflorescence, flowers and fruit and of the variability of the foliage with apparently intermediate forms between the species. The section is here limited as proposed by me in 1911 (in Sargent, Pl]. Wilson. 1. 92) where I in- cluded into this section proposed by Pax in 1886 (in Bot. Jahrb. vit. 244) several species referred by Pax to his sect. INprvisa, namely A. stkkimense Miq., A. Hookeri Miq., A. Davidi Franch., A. laxiflorum Pax and A. crataegifolium Sieb. & Zucc., but I excluded A. parviflorum Franch. & Sav. and A. erosum Pax which is a synonym of A. caudatum Wall. var. multiserratum (Maxim.) Rehd. and belongs like A. parvi- florum to the sect. Sprcata. Another species erroneously referred to the Macrantua by Handel-Mazzetti and by Fang, is A. Wardii W. W. Sm. (A. mirabile Hand.-Mazz.) which belongs to the sect. SPICATA into the affinity of A. simense Pax; the inflorescence, though typically paniculate, is sometimes reduced to a simple raceme as in the type of A. mirabile, but in the bracted and opposite rather long and ascending pedicels and in the flowers it differs from the MACRANTHA section. Be- sides the seven Chinese species distinguished here, there are eight species in Eastern Asia outside of China (A. crataegifolium Maxim., 212 JOURNAL OF THE ARNOLD ARBORETUM : (VOL. XIV A. capillipes Maxim., A. tegmentosum Maxim., A. rufinerve S. & Z., A. micranthum S. & Z., A. Tschonoskii Maxim., A. morrisonense Hay. and A, rubescens Hay.), three in India (A. sikkimense Miq., A. Hookert Mia. and A. pectinatum Wall.) and one in Eastern North America (A. pennsylvanicum L.). f the Chinese species A. Davidi is the most widely distributed and is found in all provinces of China except in Hopei and Shantung. The other species are of more restricted distribution and each seems to occupy a fairly well defined area. Acer laxiflorum Pax is restricted to Szechuan and southeastern Tibet. Acer taronense ranges from north- ern Szechuan through western Yunnan to eastern Tibet and northern Burma. Acer Forrestii is found in southwestern Szechuan and western Yunnan. Acer Maximowiczii is a northwestern species and extends from southern Shensi and Kansu into northern Szechuan and _ into Hupeh; a rather distinct form is found in Kweichou. Acer Grosseri with var. Hersii is found in northern China from Kansu to Hopei and extends south to Hupeh and through Honan to Anhwei. Acer Metcalfit is known so far only from Kwangtung and from Hunan, if 4. Davidi f. trilobata is identical with it. As neither inflorescence, nor flowers or fruits in this section seems to show distinctive and reliable characters, the following key is based ex- clusively on the leaves which allows identification of both flowering and fruiting specimens. KEY TO THE CHINESE SPECIES A. Leaves not lobed, doubly crenate-serrate with obtusish or acutish COG Sy ides eulecauhie sens O04 ees aseseeewet igs ers 1. A. Davidi AA. Leaves more or less lobed. . Margin of leaves with 5 or more acute or acuminate teeth to cm. C. Leaves with rusty pubescence on the veins beneath ; D. Lateral lobes short, acute; flowers red us bles laxiflorum DD. cen lobes elongated, acuminate ; teeth acuminate ROS Seen ee heen ana 2 taronense el — ahron beneath except axillary feo in some spec D. ee doubly and sharply serrate with acuminate teeth ; middle lobe elongated, the lateral below the middle of the leaf, long acuminate, sometimes short on part of the leaves. E. Leaves 6-12 cm. long without or occasionally with small basal lobes, finely and closely doubly serrulate; lateral lobes sometimes short and acute, pointing forward. 4. A. Forrestii 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 213 EE. Leaves 4-8 cm. long, with distinct basal lobes rarely without, incisely doubly serrate or lobulate; lateral lobes always long- acuminate, spreading...5. A. Maximowiczii DD. Leaves unequally or doubly serrate with obtusish mucronulate teeth; middle lobe triangular-ovate ; lateral lobes short, acute or short acuminate, rarely long-acuminate. A. Grossert BB. Margins of leaves coarsely and ipreniarle dentate with 2-3, rarely 4, obtuse teeth to 1 cm., lobes acuminate with entire ae Fao | Se a ae Sere 7. A. Metcalf 1. Acer Davidi Franchet in Nouv. Arch. Mus. Paris, sér. 2 viii. 212 (Pl. David. m. 30) (1884).—Pax in Engler, Pflanzenr. tv.-163, p. 36 (1902).—Rehder in Sargent, Trees & Shrubs, 1. 167, t. 83 (1905); in Jour. Arnold Arb. vir. 221 (1926); vir. 163 (1927); Ix. 90 (1928).—Fang in Contr. Biol. Lab. Sci. Soc. China, vi. 174 (1932).— Fig. 1. Acer Davidi var. glabrescens Pax in Hooker, Icon. x1x. sub t. 1897 (1889) ; in Bot. Jahrb. xx1x. 449 (1900); in Engler, Pflanzenr. 1v.-163, p. 36 (1902).—Rehder in Sargent, Trees and Shrubs, (1905). 1. 167 ) Acer Davidi |. igi alle ee in Prolene ait, ox (1893). — Pax in Engler, Pflanze .-163, 6 (1902) “ Acer Davidi var. ale eke in oe nt, Trees rand Shrubs, 1. 168 (1905), pro parte, quoad specim. Wilson, no, 1882.—Non Pac ax Acer Cat aleriei Léveillé in Fedde, Rep. Spec. Nov. x. 432 (1911). Acer laxiflorum var. integrifolium Fang in Contrib. Biol. Lab. Sci. Soc. China, vit. 932). DisTRIBUTION: Kansu, Kiangsi, Chekiang, Anhwei, Kiangsi, Hunan, Kwangtung, Kwangsi, Kweichou, Szechuan, Yunnan and southeastern Tibet. I have seen numerous specimens from all the provinces of China named above. This shows that the species is widely distributed throughout China except the northern and northeastern provinces Shensi, Shansi, Hopei, Honan and Shantung. I do not think that var. glabrescens Pax is distinct enough to be sep- arated as a variety or form. Its holotype, Henry, no. 7085, shows even on the mature leaves remnants of the brown tomentum. Var. tomen- tellum Schwerin represents the typical form. Acer Cavaleriei Lévl. of which I have a duplicate of the holotype, Cavalerie, no. 3345, before me differs slightly in the rather narrow oblong leaves 2.8-3.2 cm. wide, rounded, not subcordate at base, quite glabrous beneath, with simple and rather slight crenate serration and an entire acumen; the wings of the fruit spreading horizontally. 214 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Acer laxiflorum var. integrifolium Fang from Mt. Omei, Szechuan, of which I have a duplicate of the holotype, Fang, no. 2692, does not seem to differ from A. Davidi except that the leaves are comparatively small. Fic. 1.. Acer Davipt Franch.: leaf (2/3 nat. size) from Wilson, no. 1005a Mupin (type locality).—Fic. 2. ACER LAXIFLORUM Pax: leaf (2/3 nat. size) from Faber, no. 433, Mt. Omei (syntype). 2. Acer laxiflorum Pax in Engler, Pflanzenr. 1v.-163, p. 36 (Acer.) (1902 ).—Rehder in Sargent, Trees & Shrubs, 1. 180 (1905) ; in Sargent, Pl. Wilson. 1. 93 (1911), excl. synon.—Fang in Contrib. Biol. Lab. Sci. Soc. China, vir. 178 (1932).—Fig. 2. 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 215 SZECHUAN: Mt. Omei, E. Faber, no. 453 (syntype), Z. H. Wilson, Veitch Exp. 3349a, W. P. Fang, no. 2874; Nanchuan Hsien, W. P. Fang, no. 1191; Pan-han-shan, FE. H. Wilson, nos. 1904, 4142; Kuan hsien, W. P. Fang, no. 2369; Wenchuan-Hsien, E. H. Wilson, nos. 1309, 4099, 4108; Wa-shan, E. H. Wilson, no. 1154; Mupin, E. H. Wilson, nos. 1007, 1007a, 1069, 1234; Tachienlu, E. H. Wilson, no. 1309, W. P. Fang, no. 3664. SOUTHEASTERN TIBET: ‘Tsarong, G. Forrest, no. 21671; Mt. Kenyichunpo and region of Champutong, Salween-Irrawadi watershed, J. F. Rock, no. 10242. Acer laxiflorum is closely related to A. Davidi Fr. from which it may be distinguished by the lobed leaves with sharper acute serration, longer and slenderer acumen and with pubescent veins beneath. It also is very close to A. Forrestii W. W. Sm. from which it differs in its broader and larger leaves pubescent on the veins beneath though the pubescence is sometimes rather slight. The flowers and young fruits of A. laxiflorum are more or less purple or red, but occasionally the flowers may be greenish as in Wilson’s no. 1309, though the fruits of the same number are reddish. The flowers of the two other species are always greenish. Acer laxiflorum var. longiphyllum Fang in Contr. Biol. Lab. Sci. Soc. China, vit. 179 (1932). This variety based on Fang’s no. 4513 from Ma-pien-hsien, which I have not seen, does not appear according to the description to differ much, if at all, from the type. Neither have I seen A. laxiflorum var. ning poense Pax in Engler, Pflanzenr. tv.-163, p. 36 (1902). 3. Acer taronense Handel-Mazzetti in Anz. Akad. Wiss. Wien, 1924, p. 84 (Pl. Nov. Sin. Forts. 25, p. 3).—Fig. 3. Aes laxiflorum Pax var. longilobum Rehder in Sargent, Pl. Wilson. 1. 94 (1911), excl. specim. Wilson, 4108. SZECHUAN: Chin-ting-shan, E. H. Wilson, no. 1927 (type of A. laxi- florum var. diiieiamiiee Tu-ti-liang Mts., Lungan-fu, E. H. Wilson, no. 4509. YUNNAN: “prope fines tibeto-birmanicas inter fluvios Ludjiang (Salween) et Djiou-djiang (Irrawaddi), Handel-Mazzetti, no. 9385 (type of A. taronense) ; Mt. Gitsa west of Mekong and north of Wei-hsi, J. F. Rock, no. 18425; without precise locality, G. Forrest, nos. 8990 and 9059. EasTERN TIBET: without precise locality, G. Forrest, nos. 26317 and 26581. Upper BurmMa: G. Forrest, 26501 and 27269. Acer taronense is closely related to A. laxiflorum Pax, but the leaves differ in the caudate-acuminate lateral lobes, broader and comparatively shorter middle lobe, finer and closer serration with aristate teeth and more densely pubescent veins beneath, though glabrescent at maturity. In the less fine and close serration the Szechuan specimens approach A. laxiflorum, but the lateral lobes are caudate-acuminate. On none 216 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV of the specimens is the pubescence quite as dense on the veins as in the type specimen. The racemes of the Burma and Tibetan specimens are quite long (about 7-8 cm.) and many-flowered, while in Wilson’s no. 1927 they are only 4-5 cm. long and about 10-flowered. It is also closely related to A. pectinatum Wall., which is distinguished by the usually 5-lobed leaves more deeply cordate at base and glabrous be- neath except the bearded axils; in its very close aristate serration it approaches the specimens of A. taronense from Burma. R TARONENSE Hand. oo leaf (2/3 nat. size) from Handel: Mazzetti, no. 9385 (holotype ).— 4. Acer Forresti1 Diels: leaf (2/3 nat. size ) from G. Forrest, no. ihe (holotype). 4. Acer Forrestii Diels in Not. Bot. Gard. Edinb. v. 165 (1912).— Fig. 4. Acer laxiflorum Rehder in Sargent, Pl. Wilson. 1. 93 (1911), in art.—Fang in Contr. Biol. Lab. Sci. Soe. China, vir. 178 (1932), in part.—Non Pax. SOUTHERN SZECHUAN: east of Ning-yuan-fu, C. Schneider, no. 959; between Ouentin and Kalapa, C. Schneider, no. 1462; between Hunke and Woloho, C. Schneider, no. 1499; Kingdom of Muli, J. F. Rock, nos. 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 217 17993, 18046 and 18230, G. Forrest, no. 21337. YUNNAN: Likiang range, G. Forrest, no. 2106 (holotype), C. Schneider, nos. 1909, 3281, 3338, 3338a, J. F. Rock, nos. 3490, 3761, 4231, 5105, 5404; north of Wei-hsi, J. F. Rock, no. 17062; near Pe-yen-tsin, S. Ten, no. 548; between Chien-chuan plain and Mekong drainage basin, J. F. Rock, no. 8630; Chien-chuan-Mekong divide, G. Forrest, no. 22380; Mekong- Salween divide, G. Forrest, no. 20009, J. F. Rock, no. 8893; without precise locality, G. Forrest, nos. 10063, 11226, 11279. Acer Forrestii is closely related to A. laxiflorum Pax, but differs in the glabrous and glaucescent under-side of its leaves; the forms with longer acuminate lobes approach A. Maximowiczti Pax, but can be dis- tinguished by the finer and closer serration, the absence of the basal pair of lobes and the glaucescent under-side. The flowers are mostly greenish, but Rock’s no. 8893 has red flowers; also the fruits of Rock’s no. 5404, of Schneider’s no. 1909 and Forrest’s no. 20009 are distinctly red or reddish. Acer Forrestii f. caudatilobum, forma nova.—Fig. A typo differt lobis longe caudato-acuminatis, lobis iaeeeslibus lobo medio plerumque plus quam dimidio longioribus. YUNNAN: Yangtze watershed, western slopes of Likiang Snow Range, J. F. Rock, no. 4149, May 30—June 6, 1922 (tree 10 m.; petioles and stems red). With its caudate-acuminate lobes, the lateral ones mostly more than half as long as the middle lobe, this form looks so strikingly different from the type that it seems desirable to distinguish it as a form. 5. Acer Maximowiczii Pax in Hooker, Icon. xrx. sub t. 1897 (1889); in Engler, Pflanzenr. 1v.-163, p. 70 (1902).—Rehder in Jour. Arnold Arb, vir. 223 (1926); rx. 90 (1928).—Fang in Contr. Biol. Lab. Sci. Soc. China, vir. 180 (1932).—Fig. 6. Acer urophyllum Maximowicz in Act. Hort. Petrop. x1. 105 (1890 ).—Rehder in Sargent, Trees and Shrubs, 1. 169, t. 84 (1905). HupeH: without precise locality, A. Henry, nos. 6857 (syntype) and 6783, E. H. Wilson, Veitch Exp. nos. 1891 and 2343; Mt. Ngo-san, Hugh Scallan; Chang-yang, E. H. Wilson, Veitch Exp. no. 724; South Wushan, E. H. Wilson, no. 229; Fang-hsien, E. H. Wilson, nos. 355 (in part), 1914 (in part) and 4427; Hsing-shan-hsien, EZ. H. Wilson, nos. 355 (in part) and 1914 (in part); Ichang, FE. H. Wilson, nos. 355 (in part) and 1914 (in part); Hsao-lung-T’an, W. Y. Chun, nos. 4220 and 4618; Shin-tien-tze, W. Y. Chun, 4030. SwHeEnsti: Tai-pei-shan, W. Purdom, nos. 947 and 948. Kansu: Lower Tebbu country, Want- sang forests, J. F. Rock, nos. 14682, 14703, 14706, 14730, 14814, 14855, 218 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XIV 15031, 15041 and 15047; Tsaushi-ku, J. F. Rock, no. 14735; Dayaya, J. F. Rock, no. 14784; Pezhu valley, J. F. Rock, no. 14946; Tsaoshiku, J. F. Rock, no. 14998; Lien-hoa-shan, Shanshen-miao, J. F. Rock, no. 13488; vicinity of Choni, R. C. Ching, no, 1009. SzECHUAN: Sing- pan-hsien, E. H. Wilson, nos. 4100 and 4513, W. P. Fang, no. 4171; Nanchuan-hsien, W. P. Fang, no. 931. KweicHou: Fan-ching-shan, Steward, Chiao & Cheo, no. 500. Fic. 5. Acer Forrestirt f. CAUDATILOBUM Rehd.: leaf (2/3 nat. size) ee F. Rock, no. 4149 (holotype). Acer Maximowiczii is very similar to A. Forrestii Diels, but may be distinguished by the smaller leaves with coarser, distinctly double and even lobulate serration, by the presence of two small basal lobes, though often much reduced or sometimes lacking, and by the more elongated and spreading lateral lobes. In the Kansu specimens collected by Rock the basal lobe is mostly lacking, but the serration and the general 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 219 shape and size of.the leaf is that of A. Maximowiczti, The specimen from Kweichou (Steward, no. 500) has much larger leaves up to 10.5 cm. long and 8.5 cm. wide with the teeth more acuminate, but they are distinctly 5-lobed and lobulate and the lobes abruptly contracted in the long acumen. Possibly it should be considered a distinct variety. ae c.6. Acer Maximowiczi1 Pax: leaf (2/3 nat. size) from Henry, no. 6783, Fic. 7. Acer GrossEeR1 Pax: leaf (2/3 nat. size) from Harry Smith, no. 7932, Shansi Acer Grosseri Pax in Engler, Pflanzenr. 1v.-163, p. 80 (1902).— Rehder in Sargent, Trees & Shrubs, 1. 181 (1905); in Jour. Arnold Arb. vit. 222 (1926); vit. 163 (1927); 1x. 90 (1928).—Fang in Contr. Biol. Lab. Sci. Soc. China, vir. 181 (1932).—Fig. 7. Acer Davidii var. Y horizontalis Pax in Engler, Pflanzenr. re ce p. 79 (1 902) ; in Bot. Jahrb. xxxvi. beibl. Lxxxir. 72 905) — Rehder in Sargent, Trees and Shrubs, 1. 168 (1905), excl. Bs eee no. 1882.—Hers in Jour. N. China Branch R. As. Soc. Lit. 106 (1922) ; Liste Ess. Lign. Honan Sept. 1 (1922). Acer Davidii var. glabrescens Pax in Bot. Jahrb. xxxvi. beibl. Lxxx1I. 73 (1905).—Non Pax (1889). Acer Pavolinit Pampanini i in Nouv. Giorn. Bot. Ital. xv1z. 422 (1910). Acer Hersii Rehder in Jour. Arnold Arb. 111. 217 (1922), pro parte. 220 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Hopei: without precise locality, Pére Chanet, no. 90. SOUTHERN SuHansi: Chich-hsin, Harry Smith, no. 7932, 5895; Shih-li-p’o-shan, Harry Smith, no. 6780; Mien-shan, Lin-shih-hsien, 7. Tang, no. 970; Chin-yuan, Lin-kon-shan, K. Ling, no. 9346. SHENSI: Kan-y-san, G. Giraldi, no. 2121 (holotype; photo. in A. A.); Lin-tou-san, G. Giral- di, July 14, 1897; Mt. Marg-hua-san, west of Singan-fu, G. Giraldi, Oct.-Nov. 1894; Tai-pei-shan, W. Purdom, no. 949; Thui-kio-tsuen and Mt. Ngo-san, Hugh Scallan, in 1899; Hua-shan, J. Hers, no. 3080; 60 km. south of Sian-fu, J. Hers, nos. 2950, 2999; Lung-chow, J. Hers, no. 2359. Kansu: Lower Tebbu country, Mayaku, J. F. Rock, no. 15053. Hupeu: Ku-tcen, C. Silvestri, no. 1377 (syntype of A. Pavo- liniti; photo. in A, A.); Ou-tan-scian, C. Silvestri, nos. 1370, 1371. NortH Honan: Sung-shien, J. Hers, no. 533; Lu-shi, J. Hers, no. 1169. ANHwetI: Chu-hwa-shan, R. C. Ching, no, 2613. Acer Grosseri is very close to A. Davidi from which it differs chiefly in its lobed leaves and the somewhat sharper serration, but the lobes sometimes are very short or nearly obsolete which makes it difficult to separate the two species. Such intermediate specimens are e. g. K. Ling, no. 9346, from Shensi, and Giraldi’s Mt. Mong-hua-san specimen and Purdom no. 949 from Shensi, but the slightly lobulate margin, the sharper serration and the glabrous under‘side refer them to A. Grossert. In the type specimen the middle lobe of the leaf is broadly triangular- ovate, while in other specimens it becomes elongated and oblong-ovate which gives the leaves a resemblance to those of A. laxiflorum Pax, but the latter has the leaves pubescent on the veins of the under side and a sharper and deeper serration. Acer Grosseri var. Hersii (Rehd.), comb. nov.—Fig. 8. Acer Hersii Rehder in Jour. Arnold Arb. 111. 217 (1922); vir. 222 (1926); vit1. 163 (1927).—Fang in Contrib. Biol. Lab. Soc. Sci. China, vir. 180 (1932). Acer sp. allied to A. Grosseri Bailey, Gent. Herb. 1. 35 (1920). Honan: ‘Teng-feng-hsien, J. Hers, nos. 219 (holotype) and 2780; Tsi-yuan-hsien, J. Hers, nos. 1739 and 2800; Kikung-shan, A. NV. Stew- ard, no. 9768 (in part) Aug. 3, 1925. Huprn: Kikungshan, 4. \. Steward, no. 9768 (in part) July 1925; L. H. Bailey, June 16, 1917. ANHWEI!I: Tsin-tai, Chu-kwa-shan, R. C. Ching, no. 2789. This variety differs from the type in the more elongated long- acuminate lateral lobes, most pronounced in Steward’s no. 9768 from the Kikung-shan, in which the lateral lobes are nearly as long as the middle lobe. 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 221 Acer Metcalfii, sp. nov.—Fig. 9. Acer Davidi forma trilobata Diels in Notizbl. Bot. Gard. Mus. Berlin, x1. 211 (1931).—Fang in Contr. Biol. Lab. Sci. Soc. China, vir. 177 (1932). Arbor 10-metralis, glaberrima, ramulis laevibus fuscis vel flavescenti- bus. Folia decidua, subcoriacea, trilobata, basi subcordata, lobis medio et lateralibus caudato-acuminatis acumine basi except integro, grosse inaequaliter serrato-dentatis dentibus obtusiusculis, nervis lateralibus Fic. 8. Acer Grossert var. Hersit Rehd.: leaf (2/3 nat. size) from A. N. Steward, no. 9768, Honan. lobi medii 8-9 in dentes exeuntibus et tantum dentibus duobus vel uno vel nullo inter nervos laterales, utrinque in sicco conspicue reticulata et brunneo-viridia subtus vix pallidiora; petioli circiter 2.5 cm. longi. Flores non visi. Racemi fructiferi 5-6 cm. longi, fructibus 4-6, nuculis compressis nervosis fere horizontalibus circ. 6 mm. longis, alis leviter ascendentibus angulum latum formantibus, nuculo excluso 1.5-2 cm. longis et 6-7 mm. latis. Kwanctunc: Lung-tan Mountain, near Iu, Herb. Canton Christ. 222 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Coll. no. 12135, May 22-July 5, 1924 (type). SouTHERN HUNAN: without precise locality, S. S. Sin, no. 298, May to Aug. 1926. This new species is closely related to A. Grosseri var. Hersti Rehd. but is easily distinguished by the subcoriaceous leaves reticulate on both sides remotely and coarsely dentate-serrate with obtusish teeth; be- tween the teeth terminating the lateral veins, there are only one or two or rarely three secondary teeth and none at all toward the apex of the lobes. Of Acer Davidi {. trilobata I have before me only a photograph 9. Acer Mercatrit Rehd.: leaf (2/3 nat. size) from Canton iiss Coll. no. 12135 (holotype). of the type which shows a serration somewhat less coarse and remote than that of the type of A. Metcalfii, but it certainly is referable rather to that species than to A. Grosseri var. Hersii. I take pleasure in naming this new species in honor of Dr. F. P. Metcalf, who recognized the specimen as new and marked it thus in this herbarium; his recent paper on the species of the section Integri- folia (Lingnan Sci. Jour. x1. 193-210. 1932) is a valuable contribution to the knowledge of the genus Acer. 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 223 NOTES ON THE LIGNEOUS PLANTS DESCRIBED BY LEVEILLE FROM EASTERN ASIA’ ALFRED REHDER RUTACEAE Zanthoxylum simulans Hance in Ann. Sci. Nat..Bot. sér. 5, v. 208 (1866), “Xanthoxylum.”—Rehder in Jour. Arnold Arb. vit. oc (1826). Zanthoxylum Bungei Planchon in Ann. Sci. Nat. Bot. s , XIX. (1853), nomen.—Hance in Jour. Bot. x11. 131 ne *Xanthory- lum”; non Ann. Sci. Nat. sér. v. 209 (1866 Ganthowsion Argyi Léveillé in Mem. Acad. oh Pigaa x11. 560 Cat. Pl. Kiang-Sou, 20) (1916).—Synon. n CHINA. Ki iangsu: montagnes, d’Argy, May (1846-66) (holo- type of Z. Argyi; merotype in A. A.). Zanthoxylum stenophyllum Hemsley in Ann. Bot. rx. 147 (1895).—Rehder & Wilson in Sargent, Pl. Wilson. 11. 127 (1914). Zanthoxylum Esquirolii Léveillé in Fedde, Rep. Spec. Nov. XIII. 266 (1914) ; Fl. Kouy-Tchéou, 377 (1915).—Synon. nov. Cuina. Kweichou: without locality, J. Esquirol, no. 425, June 1904 “arbrisseau” (holotype of Z. Esquirolii ; photo. and fragments in A.A Zanthoxylum Chaffanjoni Léveillé in Fedde, Rep. Spec. Nov. x11. 266 (1914). Zanthoxylum oxyphyllum Léveillé, Fl. Kouy-Tchéou, 377 (1915).— on geworth. Cutna. Kweichou: environs de Kouy-yang, mont du Col- lége, J. Chaffonjon in herb. Bodinier, no. 2171, April 12, 1898 “arbuste liane épineux” (holotype; photo. in A. A.). This species which belongs to the section Fagara D. Don differs from Z. oxyphyllum Edgew. to which it was referred by Léveillé in 1916 (1. c.) chiefly in the 1-3 pairs of minutely serrulate leaflets, in the small inflorescences and the small flowers with the perianth only about 3 mm. long. It is rather similar to Z. cuspidatum Champ. but easily dis- tinguished by the serrulate fewer leaflets, the smaller inflorescence and the distinct pedicels 1-3 mm. long; from Z. nitidum DC. it differs in the acuminate serrulate leaflets. 1Continued from vol. xm. 332; for preceding parts see vol. x. 108-132, 184-196 and vol. x. 275-281. 224 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Zanthoxylum dissitum Hemsley in Jour. Linn. Soc. xxmr. 106 (1886) —Rehder & Wilson in Sargent, Pl. Wilson. 1. 128 (1914).— Léveillé, Fl. Kouy-Tchéou, 377 (1915). Zanthoxylum Bodinieri Léveillé in Fedde Rep. Spec. Nov. x11. 266 (1914). Cuina. Kweichou: environs de Kouy-yang, mont du Col- lége, trou au pied de la montagne de Ste. Anne (item a Tsin-gay, Che-téou-tchay) E. Bodinier, no. 2058 (in part; flower buds), Feb. 10, 1898 “grande liane €pineuse,” (syntype of Z. Bodinieri); environs de Gan-pin, torrent des Ligularia, L. Martin & E. Bodinier, no. 2058 (in part; open flowers), Feb. 25, 1898 “grande liane épineuse”’ (syntype of Z. Bodinieri; photo in A. A.); Pin-fa, bois, J. Cavalerie, no. 748, Dec. 4, 1902 “fruit a odeur forte” (syntype of Z. Bodinieri; photo. in A. A.). Zanthoxylum Bodinieri was enumerated by Léveillé in 1915 as a syn- onym of Z. dissitum Hemsl. Zanthoxylum odoratum Léveillé in Fedde, Rep. Spec. Nov. xi. 266 (1914). Evodia odorata Léveillé in Fedde, Rep. Spec. Nov. 1x. 458 (1911); Fl. Kouy-Tchéou, 375 (key) (1915). Fagara ¢ eat Handel-Mazzetti in Akad. Anz. — 1921, p. 64 (Pl. Nov. Sin. Forts. 10, p. 2) (1921).—Synon LRRD giganteum (Hand.-Mazz.) sear in Tour. Arnold Arb. virr. 151 (1927). Cuina. Kweichou: Ma-jo, J. Cavalerie, no. 2978, Aug. 1908 “odeur forte et agreable” (holotype of Evodia odorata, in fruit; photo. in A. A.); Pin-fa, montagne J. Cavalerie, no. 1771, April 17, 1904, “trouvé un pied seulement, 3 ou 4 m. de haut; les nombreuses fleurs blan. avaient une odeur forte et agreable”’; cited under Zanthoxylum odoratum,; photo. in A. A.); Ta-hi-yen, Feng-hsiang-ping, on bushy slope, alt. 1700 m., Steward, Chiao & Cheo, no. 708, Oct. 18, 1931 (tree 8 m. high, 40 cm. ‘auc fruit reddish). Hunan: in monte Yun- schan prope urbem Warkang: in silva, alt. 1150-1250 m., Handel-Maz- zetti, no. 12327, Aug. 8, 1918 “arbor 15 m.; fl. virenti-flavi cum foliis citriodori” (holotype of Fagara gigantea; isotype in A. A.). Under Evodia odorata Léveillé enumerates only Cavalerie’s no. 2978 and under Z. odoratum he enumerates only Cavalerie’s no. 1771, though he cites E. odorata as synonym. Cavalerie no. 1771 bears inflorescences with flower-buds like Handel-Mazzetti’s specimen with which it agrees perfectly; Steward, Chiao and Cheo, no. 708, has ripe fruits. Orixa japonica Thunberg, Fl. Jap. 61 (1784)—Rehder & Wilson in Sargent, Pl. Wilson. rr. 135 (1914), Sabia Feddei Léveillé in Fedde, Rep. — Nov. 1x. 456 (1911); FI. Kouy-Tchéou, 379 (1915).—Synon. n 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 225 Sabia Cavaleriei Léveillé in Fedde, Rep. awe Nov. 1x. 456 (1911) ; Fl. Kouy-Tchéou, 378 (1915).—Synon. n Glochidion Vanioti Léveillé in Fl. Kouy- Tehéou, 164 (1914).— iS) Cuina. Kweichou: Pin-fa, J. Cavalerie, no. 22bis, April 4, 1902 ‘‘2-3m.; bonne odeur; fleurs vertes” (holotype of Sabia Feddei; merotype in A. A.); Pin-fa, J. Cavalerie, no. 23bis, April 4, 1902 “petal. 5, vert-jaunes” (holotype of S. Cavalerieit, photo. in A. A.); Pin-fa, bois de La-tong, J. Cavalerie, no. 575, Oct. 1, 1902 ‘“‘arbuste” (holotype of Glochidion Vanioti, photo. in A. A.). The specimen of Sabia Feddei bears staminate and that of S. Cava- leriei pistillate flowers, that of Glochidion Vaniott is in fruit. Boenninghausenia albiflora (Hook.) Reichenbach apud Heynhold, Nomencl. Bot. Hort. 1. 126 (1840).'—Léveillé, Fl. Kouy-Tchéou, 374 (1915 Bonners thalictrifolia Léveillé & Vaniot in Bull. Acad. Intern. Géog. Bot. x1. 48 (1902). Cutna. Kweichou: mont de Kao-po (Tsin-gay), haies, herbages de la haute montagne, J. Laborde in herb. Bodinier, no. 2702, Nov. 8, 1899 “petales d’un blanc pur” (syntype of Bodiniera thalictri- folia, photo. in A. A.); environs de Hoang-ko-chou, haies, buissons, J. Seguin in herb. Bodinier, no. 2499, Aug. 1898 “fleurs blanches” (syn- type of Bodiniera thalictrifolia, photo. in A. A.). Bodiniera thalictrifolia had been already identified with Boenning- hausenia albiflora as shown by a note on the sheet of Seguin’s specimen. If var. longipes Franchet, Pl. Delavay. 123 (1889) is to be considered a distinct variety, the specimens cited above should be referred to that variety, but the difference in the length of the stipe of the carpels seems to be too slight and too gradual to maintain var. longipes as distinct. In the typical form the stipe is about 3-4 mm. long and the carpels reach to about the middle of the petals, while in var. /ongipes the stipe is about 5-8 mm. and the carpels are nearer the apex of the petals. The typical form occurs in India and Japan, while the var. longipes seems to be the prevailing form in China, except in the southwest where the following variety occurs. Boenninghausenia albiflora var. q brevipes Franchet, Pl. Delavay. 123 (1889). 1The binomial is usually credited to Reichenbach, Consp. Reg. Veg. 197 (1828), but at the place aerate peg meres a description of the genus nor di he transfer the specific epit thet of R ora Hook. which is cited as a syno- nym. The aires eneric descri Sption * was published in 1836 by Meisner, Pl. Vasc Gen. 60, and the authority for the generic name therefore should be “Reichenbach apud Meisner.” 226 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Boenninghausenia sessilicarpa Léveillé in Fedde, Rep. Spec. Nov. xra. 252 (1913), Boenninghausenia brevipes (Franch.) Léveillé, Cat. Pl. Yun-Nan, 249 (1917). CuiIna. Yunnan: paturages des mont. derriére Tong-tchouan, alt. 2550-2700 m., E. E. Maire, July 1912 “plante vivace, sous-ligneuse, en touffes dressées” (syntype of B. sessilicarpa; photo. in A. A.); pied des mont. vallées de Tong-tchouan, alt. 2500 m., E. E. Maire, Aug. 1912 “Ancolie sous-ligneuse, rameuse, petites fleurs blanches” (syntype of B. sessilicarpa; photo. in A. A This variety is well distinguished from the type by the nearly sessile carpels, but as there are no concomitant characters except that the leaves are rarely thrice pinnate and the leaflets are smaller, it can hardly be considered specifically distinct. It seems, however, geographically well separated; ten specimens from Yunnan and one from southwestern Szechuan close to the Yunnan border have subsessile carpels except one which like the specimens from all other parts of China has long-stipi- tate carpels. Toddalia asiatica (L.) Lamarck, Tab. Encycl. Méth. mm. 116 (1793 ).—Rehder & Wilson in Sargent, Pl. Wilson. 11. 137 (1914). Aralia Labordei Léveillé in Bull. Acad. ger Bot. xxiv. 144 (1914) ; Fl. Kouy-Tchéou, 34 (1914).—Synon. n Cuina. Kweichou: environs de Tsin-gay, montagne boisée escarpée prés de la ville, J. Laborde, Nov. 6, 1898 “arbuste” (holotype of Aralia Labordei; merotype in A, A.). Glycosmis Esquirolii (Lévl.) Tanaka in Bull. Soc. Bot. France, LXxv. 709 (1928); in Tanaka Citrus Exp. Sta. 1. no. 2, p. 40 (1932). Clausena Esquiroliti Léveillé in se Rep. Spec. Nov. 1x. 324 (1911); Fl. Kouy-Tchéou, 374 (191 CHINA. Kweichou: without oe J. Esquirol, no. 744 (holotype of Clausena Esquirolii, photo. in A. A. Clausena Dunniana Léveillé in Fedde, ve Spec. Nov. x1. 67 (1912). Clausena id rec Léveillé, Fl. Kouy-Tchéou, 375 (1915); non Wight CHINA. Kweic a ou: Pin-fa, rochers, J. Cavalerie, no. 1072, June 18, 1903 “petit arbrisseau, fleurs blanches” (holotype; photo. in Leveillé reduced this species in his Flore de Kouy-Tchéou to C. Willdenowti, but T. Tanaka in revising Léveillé’s specimens maintained the name C. Dunniana. Also the determinations by Evans of Handel- Mazzetti’s no. 10386 from Kweichou as C. Dunniana was confirmed by 1933], REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 227 T. Tanaka according to his note on the specimen in the Arnold Arbore- tum herbarium. Citrus ichangensis Swingle in Jour. Agric. Research. 1. 1, fig. 1-7, t. 1 (1913); in Sargent, Pl. Wilson. 1. 144 (1914). ? Citrus Cavaleriei Léveillé apud Cavalerie in Bull. Géog. Bot. xx1. 211, 236 (1911), nomen.—Koidzumi, Fl. Symb. Or.-As. 55 (1930). Cuina. Kweichou: (no specimen in herb. Léveillé). There are no specimens of Citrus Cavaleriei in the Léveillé herbarium and the name was published without description by Cavalerie who states that it is a spiny Orange growing wild at an altitude of 1700 m. near Ma-jo and Kai-tchéou in the province of Kweichou. In the same article Cavalerie publishes a few other manuscript names of Citrus given by Léveillé. Léveillé states on p. 236 that W. T. Swingle is inclined to refer C. Cavaleriei tentatively to C. hystrix DC. Koidzumi, however, identifies it (1. c.) with C. ichangensis and makes the latter name a synonym of the former, in spite of the fact that C. Cavaleriez is a nomen nudum. SIMAROUBACEAE Ailanthus Esquirolii Léveillé, Fl. Kouy-Tchéou, 404 (Sept. 1915), nomen; in Monde PI. sér. 2, xvi. 23 (Nov. 1915). CuinaA. Kweichou: without locality, J. Cavalerie, no. 773 (not now in herb. Léveillé). Léveillé remarks in Monde des plantes (1. c.) that unfortunately the specimen which had been put aside for a future diagnosis has been mislaid and that only a short diagnosis could be given which runs as follows: “differt ab A. glutinosa [sic| foliis majoribus conspicue denta- tis et floribus coloratis.”” The plant may not be an Ailanthus at all. MELIACEAE Chickrassia tabularis A. Jussieu in Mém. Mus. Paris, x1x. 251 (1830). Dysoxylon Esquirolii Léveillé in Cat. Pl. Yun-Nan, 176 (1916).— Synon. nov. Cutna. Kweichou: Ycoca-may, J. Esquirol, no. 858 (? 898) June 1906 “grand arbre, fleurs jaunes” (holotype of Dysoxylon Esqut- rolii; merotype in A. A.). The specimen represents a form with the leaves soft-pubescent be- neath and sparingly pubescent above; also the rhachis of the leaf and the inflorescence is finely pubescent. Cipadessa baccifera Miq. var. sinensis Rehder & Wilson in Sar- gent, Pl. Wilson. 11. 159 (1914). Rhus Blinii Léveillé, Fl. Kouy-Tchéou, 411 (1915).—Synon. nov. 228 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV CutnaA. Kweichou: sud de Tin-fan, J. Cavalerie, no. 1911, Nov. 1904 (holotype of Rhus Blinii; photo. in A. A.). MALPIGHIACEAE Aspidopterys Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. rx. 458 (1911); Fl. Kouy-Tchéou, 271 (1914) in part——Hutchinson in Kew Bull. Misc. Inform. 1917, p. 97.—Niedenzu in Engler, Pflanzenr. Iv.-141 (Heft 91) p.32 (Malpigh.) (1928). Aspidopterys Dunniana Léveillé in Fedde, Rep. Spec. Nov. x1. 65 (1912); Fl. Kouy-Tchéou, 271 (1914).—Niedenzu, |. c. (1928). Cuina. Kweichou: Lo-fou, J. Cavalerie, no. 2993, April 1908 (holotype of A. Cavaleriei and syntype of A. Dunniana; photo. in A. A.); Lo-fou, J. Cavalerie, no. 3477, March 1909 “‘fleurs blanches”’ (syntype of A. Dunniana ; photo. and fragments in A. A.). Aspidopterys Dunniana was referred as a synonym to A. Cavaleriei by Hutchinson who drew attention to the fact that Léveillé cites Caval- erie no. 2993 under both species, under A. Dunniana together with Cav- alerie, no. 3477. In his Flore de Kouy-Tchéou, however, Léveillé omits Cavalerie’s no. 2993 under A. Cavaleriei and cites instead Cavalerie, no. 1882, and Esquirol, no. 712, which do not belong to Aspidopterys at all, but represent Combretaceae. Aspidopterys Esquirolii Léveillé in Fedde, Rep. Spec. Nov. x1. 65 (1912); Fl. Kouy-Tchéou, 271 (1914)W—Hutchinson in Kew Bull. Misc. Inform. 1917, p. 100.—Niedenzu in Engler, Pflanzenr. tv.-141 (Heft 91) p. 21 (Malpigh.) (1928). Cavalierella cordata Léveillé, Fl. Kouy-Tchéou, 61 (1914).— Synon. nov. Cuina. Kweichou: without locality, J. Esquirol, no. 593, Aug. 1905 “arbrisseau, fleur jaune” (ex Léveillé et Hutchinson; syn- type); Houa-kiang, J. Cavalerie, no. 2032, June 6, 1904 (syntype; photo. in A. A.); Gan-chouen, J. Cavalerie, no. 3961, May 1912 (holo- type of Cavalierella cordata; photo. and merotype in A. A.). This species is according to Hutchinson easily distinguished from the other species by the externally densely hairy sepals. Esquirol’s no, 593 I have not seen, but it is cited by Hutchinson and apparently does not differ from Cavalerie’s 3961. Cavalierella cordata Lévl. one of the two species on which Léveillé based the new genus Cavalierella placed into the Caprifoliaceae is undoubtedly an Aspidoptervs and seems referable to A. Esquirolii ex- cept that the leaves are subcordate to cordate at the base and generally broader and larger than in Cavalerie’s no. 3961; the specimen is in fruit and the fruit is suborbicular, about 3 cm. broad and slightly longer, 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 229 cristate between the wings which are furnished with setose hairs except near the margin. The other species of Cavalierella, C. Dunniana, be- longs to Dipelta. EUPHORBIACEAE Andrachne Bodinieri Léveillé in Fedde, Rep. Spec. Nov. x1. 187 (1913); Fl. Kouy-Tchéou, 158 (1914). Andrachne hypoglauca Léveillé, 1. c. (1913); 1. c. (1914): CuInA. Kweichou: montagnes de Lou-tsong-koan, rocailles, talus pierreux, E. Bodinier, no. 1662, July 12, 1897 “petit arbuste”’ (holotype of A. Bodinieri, merotype in A. A.); bord de la plaine de Tou-chan, J. Cavalerie, July 16, 1897 (holotype of A. hypoglauca ; photo. in A. A.). This species seems nearest to A. chinensis Bge., but differs in its narrower, oblong to lanceolate glabrous leaves, cuneate at the base, with revolute margin and of chartaceous texture with the midrib and lateral veins impressed above and prominent beneath. In the type of A. hypoglauca the leaves are narrower and more glaucous beneath than in A, Bodiniert. In the shape of the leaves it resembles A. lolonum Hand.-Mazz., but this has the leaves densely pubescent beneath. This and the other species of Andrachne described by Léveillé are only mentioned by name as species dubiae by Pax and Hoffmann, in Engler, Pflanzenr. 1v.-147, xxv. p. 178 (1922). Andrachne Esquirolii Léveillé in Fedde, Rep. Spec. Nov. 1x. 327 (1911); Fl. Kouy-Tchéou, 158 (1914). Andrachne persicartifolia Léveillé in Fedde, Rep. Spec. Nov. x1. 187 1913). Andrachne attenuata Handel-Mazzetti in Akad. Anz. Wiss. ee 1921, p. 178 (Pl. Nov. Sin. Forts. 13, p. 2) (1921). Synon. n Cuina. Kweichou: without locality, J. Esquirol, no. “110 (holotype of A. Esquirolii; photo. in A. A.); environs de Kouy-yang; bois de la pagoda de Kien-lin-chan, dans les rocailles, “item au Ke-ma- tong,’ E. Bodinier, no. 1695, July 20, 1897 (holotype of A. persicarii- folia, photo. in A. A.); Djitschangping prope oppidum Muyu ad austr.-occ. oppidi Dschenning, Handel-Mazzetti, no. 10402, June 22, 1917 (holotype of A. attenuata; isotype in A. A.). In the shape of its leaves and other characters A. Esquirolii resembles closely the type of A. attenuata, while A. persicariifolia has much narrower leaves varying from oblong-lanceolate to narrow-lanceolate, but in the acute sepals and the strigose-setose ovary it agrees with A. attenuata. Securinega suffruticosa (Pall.) Rehder in Jour. Arnold Arb. xi. 338 (1932) 230 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Securinega ramiflora (Ait.) Mueller Arg. in De Candolle, Prodr. xv. pt. 1. 449 (1866). Securinega fluggeoides Mueller Arg., : c. 550 (1866). Phyllanthus Argyi Léveillé in Mem. Acad. a xi. 550 Cat. Pl. Kiang-Sou, 10) reine ae Cutna. Kiangsu: without locality, P Arey, no. 78, (1846- 66) “arbrisseau” (holotype of PAyllanthus Argyi; merotype in A. A.) Phyllanthus emblica Linnaeus, Spec. Pl. 982 (1753). Phyllanthus Mairei Léveillé in Bull. Sie ial xxv. 23 (1915); Cat. Pl. Yun-Nan, 97 (1916).—Synon. n Curna. Yunnan: rives du fleuve Bleu, 3 a , Siao- ho, alt. 400 M., E. E. Maire, May 1912 “arbrisseau toujours vert; fleurs jaunes; fruits verts, acides, en forme de cerise” (holotype of P. Mairei, merotype in A. A.). Phyllanthus Mairei has been already identified by W. W. Smith with P. emblica according to a note on the sheet. Phyllanthus Franchetiana Léveillé in Bull. Géog. Bot. xxv. 23 (1915); Cat. Pl. Yun-Nan, 97 (1916). Cuina. Yunnan: river du fleuve Bleu, a Siao-ho, alt. 400 m., E. E. Maire, May 1912 “fleurs rougedtres, drupes vertes, acidulées, en forme de cerise” (holotype; photo. and fragments in A. A.). This species resembles P. pulcher Wall., but the leaves which are 6-8 mm. long are not mucronate and the sepals are crenate-dentate rather than laciniate. The fruits which are not present on the speci- men are cherry-like according to the collector’s note. Phyllanthus Dunnianus (Lévl.) Handel-Mazzetti in herb. BE clea rte Dunnianum Léveillé in Fedde, Rep. Spec. Nov. 11); Fl. Kouy-Tchéou, 166 (1914) —Handel- Mazzetti, Symb. nee vit. 224 (1931). Perens Cavalerici Léveillé in Fedde, Rep. ja Nov. Ix. 454 (1911); Fl. Kouy-Tchéou, 116 (1914).—Synon. Phyllanthodendron eal Lévl. var. te Léveillé, FI. Kouy-Tchéou, 166 (191 CuHInA. Kweichou: Lo-fou, J. Cavalerie, no. 2659, Nov. 1905 “fl. vertes” (holotype of Phyllanthodendron Dunnianum ; mero- type in A. A.); O-to prés Lo-fou, J. Cavalerie, no. 3284, April 1907 “arbrisseau” (holotype of Ph. Cavaleriei; merotype in A. A.); Lo-fou, J. Cavalerie, no. 3500, April, May 1909 (holotype of P#. Dunnianum var. hypoglaucum,; merotype in A. A.). The three specimens cited above agree in the distinctly winged branchlets and generally in the shape of the leaves, but differ in several other characters. In PAyllanthodendron Dunnianum var. hypoglaucum (Cavalerie, no. 3500) which is in fruit the leaves are glaucous beneath and the branchlets are pilose, while in the other two specimens they are 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 231 perfectly glabrous. Phyllanthodendron Dunnianum (Cavalerie, no. 2659) has thin membraneous leaves, while Ph. Cavaleriei (Cavalerie, no. 3284) has subcoriaceous leaves with prominent veins on both sides; both specimens are in flower. Cavalerie’s nos. 3284 and 3500 have dimorphic leaves, while no. 2659 which consists only of two small sterile branchlets (with a detached flower in a pocket) has only one kind of leaves. The specimen of no. 2659 bears in Handel-Mazzetti’s hand- writing the new combination “Phyllanthus Dunnianus (Lévl.)” but in his remarks on Phyllanthodendron Dunnianum under his Phyllanthus anthopotamicus (in Symb. Sin. vu. 224. 1931) he does not cite this combination; he only states that the two species are closely related. Phyllanthus Dunnianus, however, differs markedly from P. anthopo- tamicus in the strongly angular, slightly winged glabrous or only slightly pilose branchlets, the quite glabrous, more acuminate leaves dark green above and generally larger, and in the larger, longer-stalked glabrous flowers. Phyllanthus spec. Sterculia Bodinieri Léveillé, Fl. Kouy-Tchéou, 406 (1915).— Synon. nov. Cuina. Kweichou: environs de Hoang-ko-chou, grande cascade, au bord de l’eau, J. Seguin in herb. Bodinier, no. 2194, April 1898 “arbuste, fleurs rougeatres” (holotype of Sterculia Bodinieri, merotype in A. A.). This plant is undoubtedly a Euphorbiacea and seems referable to the sect. Eriococcus of PAyllanthus except that it has 4 anthers. Glochidion puberum (L.) Hutchinson in Sargent, Pl. Wilson. 1. 518 (1916). Glochidion Bodinieri Léveillé in Fedde, Rep. Spec. Nov. x1r. 183 (1913); Fl. Kouy-Tchéou, 163 (1914).—Synon. nov. CHINA. Kweichou: Kouy-yang, mont du Collége, E. Bo- dinier, no. 2307, July 21, 1897 (fruit), June 9, 1898 “arbuste de 1 m., trés branchu” (holotype of G. Bodinieri; merotype in A. A.). Glochidion villicaule Hooker f., Fl. Brit. Ind. v. 326 (1887). Glochidion Esquirolii Léveillé in Fedde, Rep. Spec. Nov. x11. 186 (1913); Fl. Kouy-Tchéou, 163 (1914).—Synon. nov. Cu1naA. Kweichou: Ouang-mou, J. Esquirol, no. 714, June 1904 “fleurs jaunes” (holotype of G. Esquirolii; photo. in A. A.). Baccaurea Cavaleriei Léveillé, Fl, Kouy-Tchéou, 159 (1914). Cu1na. Kweichou: Lo-fou, J. Cavalerie, no. 3299, April 1907 “petit arbre” (holotype, photo. in A. A.). In the general habit and in the leaf this species has some resem- blance to B. sapida Muell. Arg., but the material is insufficient for exact 232 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XIV determination; it consists of a leafy branch and broken fragments of immature inflorescences. Antidesma microphyllum Hemsley in Jour. Linn. Soc. Bot. xxv1. 432 (1894).—Handel-Mazzetti, Symb. Sin. vir. 218 (1931). Antidesma Seguini Léveillé in Fedde, Rep. Spec. Nov. 1x. 460 (1911); Fl. Kouy-Tchéou, 158 (1914)—Pax & Hoffmann in Engler, Pflanzenr. 1v.-147, xv. 166 (1922). Myrica Darrisii Léveillé in Fedde, Rep. tie Nov. x11. 537 (1913) ; Fl. Kouy-Tchéou, 281 (1914).—Synon. n Cutna. Kweichou: district de Tehen- Tin, bord du fleuve a la cascade de Hoang-ko-chan, J. Seguin, June 10, 1898 “fleur blanche” (ex Léveillé; syntype of A. Seguini) ; fleuve Hoa-kiang, J. Esquirol, no. 505, June 1905; (ex Léveillé; syntype of A. Seguini) ; Tchai-choui-ho, J. Esquirol, no. 1586, July 1909 (syntype of A. Seguini; merotype in A. A.); route de Mou-you-se 4 Houang-ko-chou, J. Cavalerie, no. 2058, June 10, 1904 (holotype of Myrica Darrisii; photo. and frag- ment in A. A.). Of the three syntypes of A. Seguini I have before me only Esquirol, no. 1568, which agrees exactly with Henry’s 9530a of A. microphvllum Hemsl., while Cavalerie, no. 2058, the holotype of Myrica Darrisit, differs slightly in, at least partly, somewhat broader leaves. Antidesma Seguini had been identified with A. microphyllum by Dr. Stapf accord- ing to a note in a letter from Dr. Handel-Mazzetti of Dec. 25, 1922. Croton Tiglium Linnaeus, Spec. Pl. 1004 (1753).—Handel-Maz- zetti, Symb. Sin. vir. 218 (1931). pe Vanioti Léveillé, Cat. Pl. Yun-Nan, 95 16). Cuina. Yunnan: Tong-tchouan, £. E. Maire, rete (holotype of Alchornea Vanioti; merotype in A. A.). Alchornea Vanioti was identified with Croton Tiglium by Handel- Mazzetti (1. c.). Speranskia cantonensis (Hance) Pax & Hoffmann in Engler, Pflan- zenr. Iv.-147, v1. 15 (Euphorb.) (1912). Mercurialis acanthocarpa Léveillé in Fedde, Rep. Spec. Nov. 11. 21 906). Speranskia tonkinensis Léveillé, Fl. Kouy-Tchéou 167 (1914). Cuina. Kweichou: Pin-fa, prés d’une riviére, trés rare, J. Cavalerie, no. 1585, Oct. 19, 1903 (holotype of Mercurialis acantho- carpa; photo. in A. A.). Mercurialis acanthocarpa was referred by Léveillé together with Speranskia Henryi Oliv., to Speranskia tonkinensis, a name cited with- out author and probably a mistake for S. cantonensis. Mallotus Leveillanus Fedde, Rep. Spec. Nov. x. 144 (1912).—Pax & Hoffmann in Engler, Pflanzenr. 1v.-147, vir. p. 165 (1914). 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 233 Mallotus Esquirolii Léveillé in Fedde, Rep. Spec. Nov. 1x. 461 (1911), non Léveillé, 1. c. 327, Mallotus Leveillei Fedde apud Léveillé, Fl. Kouy-Tchéou, 165 (1914). CuInA. Kweichou: Ouang-mou, J. Esquirol, no. 120, June 1904 (holotype of M. Esquirolii; photo. in A. A.); Lo-fou, J. Cavalerie, no. 3666, Aug. 1909 (duplicate in A. A.). Cavalerie’s no. 3666 is cited together with Esquirol’s no. 120 by Léveillé under M. Leveillei (1. c.) and by Pax & Hoffmann under M. Leveillanus (1. c.). Mallotus philippinensis (Lam.) Mueller Arg. in Linnaea, xxx1Vv. 196 (1865).—Pax & Hoffmann in Engler, Pflanzenr. 1v.-147, vu. p. 184 (1914). Evonymus eee Léveillé in Fedde, Rep. Spec. Nov. x111. 260 1914).—Syn Phyllanthodendron a ‘Léveillé, Fl. Kouy-Tchéou, 166 (1914). Cu1na. Kweichou: Lo-fou, J. Cavalerie, no. 2733, April 1906 (holotype of Evonymus hypoleucus and Phyllanthodendron sp.; photo. in A. A.). Evonymus hypoleucus is cited by Léveillé in his Flore de Kouy- tchéou as a synonym of Phyllanthodendron sp. Mallotus Esquirolii Léveillé in Fedde, Rep. Spec. Nov. rx. 327 (1911); Fl. Kouy-Tchéou, 165 (1914).—Pax & Hoffmann in Engler Pflanzenr. Iv.-147, vir. 196 (1914). CuInA. Kweichou: without precise locality, J. Esquirol, no. 898 (holotype; photo. and fragments in A. A.); Lo-fou, J. Cavalerie, no. 3114, March 1909 (cited in Fl. Kouy-Tchéou; duplicate in A. A.). Mallotus Milliettii Léveillé, Fl. Kouy-Tchéou, 165 (1914). Cuina. Kweichou: route de Gan-chouen a Hin-y-fou, Cava- lerie, no. 3967, July 1912 (holotype; photo. and fragments in A. A.). This species seems nearest to M. contubernalis Hance, but differs chiefly in the leaves being sparingly stellate-pubescent above and rather densely and softly so below, and in the subsessile larger cap- sules about 1.4 cm. in diam. with a heavier indumentum. In the pubescence of the leaves it seems to approach var. chrysocarpus (Pamp.) Hand.-Mazz. which I have not seen, the type of it being missing from the Herb. Biondi when I was in Florence in 1932. Léveillé cites Cavalerie, no. 3697, as the type of his species, but the specimen is numbered 3967 and bears in Léveillé’s handwriting the name Mallotus Cavaleriei in ink and below in pencil M. Milliettii, Léveillé apparently changed the name M. Cavaleriei which was never published for this plant, to M. Milliettii, because he had already given 234 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV the former name to a species which was later referred to Discocleidion rufescens (Fr.) Pax & Hoffm. Discocleidion rufescens (Franch.) Pax & Hoffmann in Engler, Pflanzenr. 1v.-147, vil. p. 45, fig. 6 (1914).—Leéveillé, Fl. Kouy-Tchéou, 161 (1914). Mallotus Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. x1. 296 (1912). Cutna. Kweichou: Gan-chouen, J. Cavalerie, no. 3825, June 1910 (holotype of Mallotus Cavaleriei; photo. in A. A.). Acalypha Mairei (Lévl.) Schneider in Sargent, Pl. Wilson. It. 301 (1916).—Léveillé, Cat. Pl. Yun-Nan, 94 (1916).—Pax & Hoff- mann in Engler, Pflanzenr. 1v.-147, xvi. 137 (1924). Morus Mairei Léveillé in Fedde, Rep. Spec. Nov. x11. 265 (1914). Cumna. Yun-nan:_ brousse derriére Mo-tsou fleuve Bleu, alt. 800 m., E. E. Maire, May (ex Léveillé; syntype of Morus Mairet) ; rochers de Ma-hong, alt. 3000 m., E. E. Maire, June 1912 (syntype of Morus Mairei, merotype in A. A.). Tragia involucrata Linnaeus, Spec. Pl. 980 (1753).—Pax & Hoff- mann in Engler, Pflanzenr. 1v.-147, rx-x1. 81 (1919).—Handel-Mazzetti, Symb. Sin. vi. 218 (1931). Alchornea Mairei Léveillé, Cat. Pl. Yun-Nan, 94 (1916). Curva. Yunnan: vallon de Yon-fong-keou, alt. 800 m., E. E. Maire, July 1912 (holotype of Alchornea Mairei; photo. in A. A.). Alchornea Mairei was first identified with Tragia involucrata by Handel-Mazzetti who states that it approaches var. intermedia Muell. Arg. Sapium rotundifolium Hemsley in Jour. Linn. Soc. Bot. xxv, 445 (1894).—Handel-Mazzetti, Symb. Sin. vit. 212 (1931). Baccaurea Esquirolii Léveillé, Fl. Kouy-Tchéou 159 (1914) .— Synon. nov Cuina. Kweichou: Mou-you-se, J. Cavalerie, no. 2137 (no. 13), June 1904 “arbre, fl. jaunes” (syntype of Baccaurea Esquirolii ; photo. in A. A.); Lo-fou, J. Cavalerie, no. 3458, Oct. 1908 (syntype of B. Esquirolii; photo. in A. A.); without precise locality, J. Esquirol, no. 517, June 1905 (syntype of B. Esquirolii; photo. in A. A. This species was collected in Kweichou also by Handel- Naweetii (nos. 10279 and 10371). DAPHNIPHYLLACEAE Daphniphyllum macropodum Miquel in Mus. Bot. Lugd.-Bat. m1. 129 (1867).—Rosenthal in Engler, Pflanzenr. 1v.-147a, p. 9 (1919). Webera Marchandii Léveillé in Fedde, Rep. Spec. Nov. xu. 178 (1911) ; Fl. Kouy-Tchéou, 372 (1915).—Synon. nov. 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 235 Cutna. Kweichou: moulins de Tong-tchéou, J. Esquirol & R. Marchand, no. 3252, June 22, 1912 (holotype of Webera Mar- chandii; photo. in A. A.). BUXACEAE Sarcococca Hookeriana Baill. var. humilis Rehder & Wilson in Sargent, Pl. Wilson. 11. 164 (1914). Maesa spec. Léveillé, Fl. Kouy-Tchéou, 287 (1914). Myrsine Chevaliert Léveillé, Fl. Kouy-Tchéou, 287 (1914).— Synon. nov. Pachysandra Mairei Léveillé, Cat. Pl. Yun-Nan, 97, fig. 23 (1916). Cutna. Kweichou: enfoncement de Ouan-ly prés Thou-ly, J. Esquirol, no. 2593, Feb. 1911, “fleurs jaunatres, fruits rouges” (holo- type. of Myrsine Chevalieri; merotype in A. A.). Yunnan: an pied de rochers, collines arides a l’est de Tong-tchouan, alt. 2600 m., E. E. Maire, March 1912 “arbrisseau toujours vert en touffes; fleurs blanches; fruits noirs” (holotype of Pachysandra Mairei; merotype in Esquirol’s no. 2593, Myrsine Chevalieri, agrees exactly with speci- mens of Sarcococca Hookeriana var. humilis and I have no doubt that it belongs here, though Esquirol states that it has red fruits; the speci- men before me bears only immature inflorescences. Under Myrsine Chevalieri Léveillé cites (1. c. p. 288) Maesa spec., Esquirol 2593, ap- pearing on the preceding page as being the same. Pachysandra Mairei was identified by Handel-Mazzetti (Symb. Sin. vi. 235) with S. Hook- eriana var. digyna, but I think it belongs with var. Aumilis which, how- ever, may be only a dwarf and smaller form of var. digyna. A topotype or perhaps an isotype of P. Mairei was distributed as Sarco- cocca spec. E. E. Maire, no. 355, by the Arnold Arboretum. Pachysandra stylosa Dunn in Jour. Bot. xtv1. 326 (1908).— Handel-Mazzetti, Symb. Sin. 236 (1931). Pachysandra Bodinieri Léveillé in Fedde, Rep. Spec. Nov. x11. 187 1913). Pachysandra axillaris Franch. var. Kouytchensis Léveillé, Fl. Kouy- Tchéou, 166 (1914). CuHInA. Kweichou: monts entre Ma-kay et Se-tchong-hien a Tien-sen-kiao, rochers a l’entrée du Tien-sen-kiao, E. Bodinier, no. 1525, Aug. 5, 1897 (holotype of P. Bodinieri; photo. in A. A.); Tsin- gay, E. Bodinier (holotype of P. axillaris var. Kouytchensis; photo. in A.A). Pachysandra Bodinieri is not enumerated in the Flore de Kouy- Tchéou, nor is Bodinier’s no. 1525 cited under Pachysandra. Accord- ing to Handel-Mazzetti both belong to P. stylosa. 236 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Buxus microphylla S. & Z. var. aemulans Rehder & Wilson in Sargent, Pl. Wilson. 11. 169 (1914). Buxus Bodinieri in Fedde, Rep. Spec. Nov. x1. 549 (1913) ; Fl. Kouy- Tchéou, 160 (1914).—Synon. nov. Cuina. Kweichou: environs de Kouy-yang, mont du Col- lege, E. Bodinier, no. 2079, Feb. 25, 1898 “arbuste plus ou moins grand” (holotype of B. Bodinieri; merotype in A. A.). The filaments in Bodinier’s plant are longer than in the type of var. aemulans. Buxus megistophylla Léveillé, Fl. Kouy-Tchéou, 160 (1914); Cat. Ill. Pl. Seu-Tchouen, pl. 26 (1918). Cutna. Kweichou: environs de Hoang-ko-chou, rocailles de la cascade, J. Seguin in herb. Bodinier, no. 2607, March 15, 1899 “arbuste de 0.60 cm., fleurs blanches” (syntype; merotype in A. A.); Kiao-men, prés Lo-fou, J. Esquirol, no. 2560, Dec. 1910 (syntype; photo. in A. A.). In its large leaves and subterete branchlets B. megistophylla re- sembles B. Henryi Mayr, but differs in the glabrous inflorescence, sub- sessile staminate flowers (in B. Henryi on pedicels 2-3 mm. long) and short stigmas; from B. microphylla S. & Z. it differs besides in the large leaves, chiefly in the short rudimentary ovary and the long filaments of the staminate flower. Buxus Myrica Léveillé in Fedde, Rep. Spec. Nov. x1. 549 (1913); Fl. Kouy-Tchéou, 160 (1914). Frutex ramulis tetragonis breviter pilosis gracilibus. Folia brevis- sime petiolata petiolis pilosulis, oblongo-lanceolata vel anguste lanceo- lata, 2-5 cm. longa et 0.5-1.4 cm. lata, acuta, mucronulata, basi cuneata, glabra, tenuiter coriacea et in sicco utrinque distincte reticulata, costa media utrinque elevata. Racemi axillares numerosi, rhachi dense pilosula elongata 5-7 mm. longa; bractearum paria 6-8, bracteae ovatae, acutae, 2 mm. longae, dorso dense pilosulae et intus ad marginem villosulis; flores ¢ breviter pedicellati pedicello 1-1.5 mm. longo dense pilosulo, sepalis ovalibus, 2 exterioribus carinatis carina ciliata, 2 interioribus longioribus glabris circ. 3 mm. longis, staminibus sepala superantibus 4 mm. longis, rudimento ovarii sepalis triplo breviore; flos ° terminalis, sepalis oblongo-ovatis 3-4 mm. longis, 3 exterioribus dorso dense 3 interioribus sparse pilosulis, ovario stylis complanatis apice recurvis multo breviore (ovarium 1.5 mm. longum, styli 3.5 cm. longi), stig- matibus ad medium stylum decurrentibus. Fructus non visus. Cumna. Kweichou: Pin-fa, J. Cavalerie, no. 3198, April 8, 1907 (syntype; merotype in A. A.); Lo-hou, J. Esquirol, no. 2566, Dec. 1910 (ex Léveillé; syntype); grande cascade de Hoang-ko-chou, dans 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 237 les rochers, J. Seguin in herb. Bodinier, no. 2266, April 3, 1898 (syn- type; photo. in A. A.). This species, of which I have given a description above, since that of Léveillé is inadequate, seems most closely related to B. Henryi Mayr, but differs in the quadrangular pilose branchlets, narrower and smaller thinly coriaceous and reticulate leaves, smaller and acute densely pilose bracts not scarious on the margin, in the shorter pedicels of the stami- nate flowers (in B. Henryi 2-3 mm. long) shorter filaments and shorter flattened styles. The two specimens which I have seen differ slightly in the size and shape of the leaves; Cavalerie’s no. 3198 has the leaves mostly 3-5 cm. long and up to 1.4 cm. broad, while Bodinier’s no. 2266 has comparatively narrower leaves 2-5 cm. long (a few even smaller) and 5-10 mm. broad; in the former the staminate flowers have mostly dropped while the latter is in full bloom. Buxus Harlandi Hance var. cephalantha (Lévl. & Vant.), comb. nov. Buxus cephalantha Léveillé & Vaniot in Fedde, Rep. Spec. Nov. 111. 21 Popes Aaa var. ge ie Léveillé, Fl. Kouy-Tchéou, 160 (1914) —Non Siebold & Zuccar Cuina. Kweichou: Pin-fa, ace dans ou prés ruisseaux, J. Cavalerie, no. 1797, Aug. 25, 1904 “tout petit buis, 1 pied de h.” (holotype of B. cephalantha; photo. in A. A.). This form differs from the type chiefly in its very small size; the plant is about 30 cm. high with short branchlets, the leaves measure 6-11 mm. in length and are mostly slightly emarginate and partly obovate (8:4 mm.), the inflorescences are mostly terminal. CORIARIACEAE Coriaria sinica Maximowicz in Mém. Acad. Sci. St. Pétersb. sér. 7, XXIX. no. I. 9, fig. (1881).—Léveillé Cat. Pl. Yun-Nan, 249 (1917): Morus calva Léveillé in Fedde, Rep. Spec. Nov. x1r. 265 (1914). CuHInA. Yunnan: coteaux arides 4 La-kou, alt. 2400 m., E. E. Maire, March 1912 “grand arbuste a 1’écorce fibreuse, fleurs jaunatres, d’abord rouges” (holotype of Morus calva; merotype in A. A.). Morus calva was first identified with Coriaria sinica by C. Schneider (in Sargent, Pl. Wilson. m1. 301. 1916). ANACARDIACEAE Spondias axillaris Roxburgh, Cat. Hort. Beng 34 (1814), nomen; FI. Ind. 1. 453 (1832).—Rehder & Wilson in Sargent, Pl. Wilson. 11. 172 (1914). 238 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Rhus hggor tity Léveillé in Fedde, Rep. Spec. Nov. x. 437 (1912).— Synon. CHINA. HH on - kong: bois de Happy Valley; bord de la routs a Pok-fu-lum, E. Bodinier, no. 1103, April 6 & 15, 1895 (holotype of Rhus Bodinieri; photo. in A. A.). Pistacia chinensis Bunge in Mém. Div. Sav. Acad. Sci. St. Pétersb. 11. 89 (Enum. PI. Chin. Bor. 15) (1833).—Rehder & Wilson in Sargent, Pl. Wilson. 11. 173 (1914). Rhus gummifera Léveillé in Fedde, Rep. pes Nov. x. 474 (1912) ; Fl. Kouy-Tchéou, 412 (1915).—Synon. n Rhus Argyi Léveillé in Mem. Acad. Ci. Barceona ser. 3, x11. 562 (Cat. Pl. Kiang-Sou, 22) (1916).—Syn CutnaA. Kiangsu:_ without meen "Tocality, Ch. d’Argy, (1846-66) (holotype of Rkus Argyi; merotype in A. A.). Kwei- chou: au sud de Pin-fa prés ruisseau, sur rochers, J. Cavalerie, no. 1078, June 18, 1903 “arbre produisant une espéce de gomme odorante” (holotype of Rhus gummifera; photo. and fragments in A. A.). The specimen from Kiangsu has the rachis and the midrib of the leaflets below densely and finely pubescent while that from Kweichou is nearly glabrous. This in general seems to constitute a difference be- tween the eastern and the western plants of this species; the specimens before me from the eastern provinces are mostly more or less pubescent on the rachis and the midrib of the leaflets, while the western speci- mens are glabrous or nearly so. Rhus punjabensis Stewart var. sinica (Diels) Rehder & Wilson in Sargent, Pl. Wilson. 11. 176 (1914). ? Rhus echinocarpa Léveillé in Fedde, Rep. Spec 475 (1912); Fl. Kouy-Tchéou, 411 (1915), pro parte, Balsa specim. Cavalerie, no. 2003. Rhus Esquirolit Léveillé in Fedde, Rep. — Nov. x11. 181 (1913) ; Fl. Kouy-Tchéou, 411 (1915).—Synon. n ? Rhus Mairei Léveillé, Sert. Yunnan. 2 (1916) ; Cat. Pl. Yun-Nan, 269 (1917).—Synon. nov Cuina. Kweichou: “Tsin-gai, bois, J. Cavalerie, no. 1157, July 13, 1903, “petit arbre, fruit rouge-velour” (holotype of Rd. Esquirolii; photo. in A. A.). Yunnan: pied de montagne, der- riére Tong-tchouan, alt. 2550 m., E. E. Maire, May 1912 “petit arbre cassant, rameaux rares, fleurs blanches” (syntype of Rk. Mairet; mero- type in A. A.) ; haies de la plaine a Tong-tchouan, alt. hae m., July 1912 “fruit d’un petit arbre cassant, fl. blanches en grappes,” E. E. Maire (syntype of RA. Mairei; photo. in A. A.); plaine de Tche-hai, haies, alt. 2500 m., E. E. Maire, Aug. 1912 (ex Léveillé; syntype of Rh. Mairet). Rhus Mairei 1 refer here with some doubt; it has partly only 5-7 leaflets and the rachis is not or very slightly winged, but otherwise it 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 239 seems to agree with RA. punjabensis var. sinica. I also refer here doubtfully Cavalerie’s no. 2003 numerated by Léveillé as a syntype of Rh. echinocarpa. Rhus trichocarpa Miquel in Ann. Mus. Bot. Lugd.-Bat. 1. 84 (1866); Prol. Fl. Jap. 16 (1866).—Rehder & Wilson in Sargent, PI. Wilson. 11. 180 (1914 Rhus piaeleatts Léveillé j in Fedde, Rep. Spec. Nov. x. 475 (1912) ; 1. ouy-Tchéou, 411 (1915); specim. no. 2003 excluso.— Synon. nov. Cuina. Kweichou: Mou-you-se, J. Cavalerie, no. 1016 (flowers), May 28, 1903 (syntype of Rh. echinocarpa; photo. and fragments in A. A.); without precise locality, fruit, J. Cavalerie, no. 108 (2) [?] (photo. in A. A.). Under Rhus echinocarpa Léveillé cites two specimens, nos. 1016 and 2003, of which only the first belongs here, while the second is identical with Rhus punjabensis var. sinica; both are flowering specimens. ‘The description of the fruit, from which the name of the species is de- rived, is apparently based on an unnamed specimen in the Léveillé herbarium marked only ‘J. Cavalerie, 108 (2)” on a slip and placed in the same cover with the other two specimens. As there is no other fruiting specimen of Rhus in the Léveillé herbarium which answers to the description of the very characteristic fruit of this species, this must be the specimen from which the description of the fruit was drawn, though it is not cite This record eins the range of Rh. trichocarpa west into Kweichou. AQUIFOLIACEAE Ilex suaveolens (Lévl.) Loesener in Ber. Deutsch. Bot. Ges. xxxm. 541 (1914) —Léveillé, Fl. Kouy-Tchéou, 201 (1914). Celastrus suaveolens Léveillé in Fedde, Rep. Spec. Nov. x11. 263 914). CHtInaA. Kweichou: environs de Kouy-yang, bois de Kien- lin-chan, E. Bodinier, no. 2663 [not 2683], June 19, 1899, “arbre de haute taille” (syntype of Celastrus suaveolens; photo. and fragments in A. A.); Pin-fa, J. Cavalerie, no. 17bis, June 3, 1902 ‘“‘arbrisseau 5-6 m., fleurs blanches parfumées” (syntype of C. suaveolens ; fragments in A full description of the species is given by Loesener (1. c.). Ilex purpurea Hasskarl, Cat. Pl. Bogor. 230 (1844).—Loesener in Nov. Act. Leop.-Carol. Akad. txxvit. 111 (Monog. Aquifol.) (1901). Callicarpa Cavaleriet Léveillé in Fedde, Rep. Spec. Es 1x.) 455 (1911); Fl. Kouy-Tchéou, 439 (1915).—Synon. no Embelia rubro-violacea Léveillé in Fedde, Rep. Spec. Nov. x. 375 (1912); Fl. Kouy-Tchéou, 285 (1914).—Synon. no 240 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Celastrus Bodinieri Léveillé in Fedde, Rep. Spec. Nov. x111. 263 ynon. nov. Cuina. Kweichou: environs de Tou-chan, J. Cavalerie, no. 2624, June 3, 1899 “grand arbuste, fleurs roses’? (holotype of Calli- carpa Cavaleriei; photo. in A. A.); Pin-fa, J. Cavalerie, no, 1334, June 23, 1902 (ex Léveillé; syntype of Embelia rubro-violacea); foréts, J. Esquirol, no. 429, June 1905, “grand arbrisseau” (syntype of E. rubro- violacea; merotype in A. A.); environs de Kouy-yang, mont du Col- lége, E. Bodinier, no. 2384, June 10, 1898 “arbuste’’ (holotype of Celastrus Bodinieri; photo. in A. A.). The specimens cited above belong to var. Oldhami (Miq.) Loes. which represents the type of J. purpurea Hassk.; all are staminate plants loom. ae pedunculosa Miq. var. continentalis Loesener in Nov. Act. Leop.-Carol. Akad. Lxxvur. 110 (Monog. Aquifol.) (1901); in Sar- gent, Pl. Wilson. 1. 76 (1911). Ilex purpurea var. § Leveilleana Loesener in Léveillé, Fl. Kouy- Tchéou, 201 (1914). Cuina. Kweichou: Pin-fa, bois, J. Cavalerie, no. 1066, June 1903 (holotype of J. purpurea var. Leveilleana; photo. in A. A.). Cavalerie’s no. 1066 was considered by Loesener according to a note by him on the sheet of the type specimen, a possible hybrid between I. purpurea Hassk. and J. pedunculosa Miq., but the specimen seems to agree well in all respects with 7. pedunculosa except that the petioles are not as deeply channelled as in /. pedunculosa and the margins of the leaves are somewhat more serrulate. In the solitary flowers (or in one case a 3-flowered cyme) and in the texture of the leaves it certainly agrees with 7. pedunculosa which often has slightly serrulate leaves; the midrib above is short-pilose as it is, though usually to a lesser degree, in J. pedunculosa; in I. purpurea it is perfectly glabrous. Ilex metabaptista Loes. var. myrsinoides (Lévl.), var. nov. Maesa myrsinoides Léveillé in Fedde, Rep. Spec. Nov. x. 375 (1912) ; Kouy-Tchéou, 286 (1914). Myrsine Feddei Léveillé in Fedde, Rep. a Nov. x. 376 (1912) ; Fl. Kouy-Tchéou, 288 (1914) —Synon. n Embelia Cavaleriei Léveillé, Fl. Kouy- Tehéou, 284 (1914).— Synon. nov. Ilex Fargesti Franch. var. Bodinieri, Loesener apud Léveillé, 1. c. 200 (1914).—Synon. no llex metabaptista Léveillé, L c. 200 (1914), vix Loesener CHINA. weichou: Pin-fa, ruisseau de La-tong, J J. Cavalerie, no. 579, Oct. 1, 1902 (holotype of Maesa myrsinoides; merotype in A. A.); Pin-fa, ruisseaux, J. Cavalerie, no. 842, May 23, 1902, “‘fleurs blanches” (holotype of Myrsine Feddei; merotype in A. A.); bords des 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 241 ruisseaux et torrents, J. Cavalerie in hb. Bodinier no. 2635, June 3, 1899, “fleurs blanches parfumées” (holotype of Embelia Cavaleriei, merotype in A. A.); environs de Kouy-yang, Mont du Collége, a la Cascade, J. Chaffanjon, May 28, 1898, et environs de Tou-chan, J. Cav- alerie, June 3, 1899, in herb. Bodinier, no. 2310 (2 sheets, one is the holotype of J. Fargesii var. Bodinieri, the other represents /. meta- baptista Lévl., vix Loes.; photos. in A. A.). This variety differs from the type chiefly in the smaller leaves 2.5-4, rarely up to 6.5 cm. long, shorter 2-4 mm. long petioles and plants inflorescences. Of the three specimens cited above, Cavalerie, no. 579, is in fruit, while his no. 842 bears pistillate flowers and no. 2635 stami- nate flowers; the latter differs in the slightly broader elliptic-oblong leaves perfectly glabrous below, while the other two numbers are like the type sparingly short-pilose on the midrib below chiefly toward the base and pubescent on the midrib above. Bodinier’s no. 2310 is repre- sented by two sheets, one containing two staminate specimens, and named by Loesener J. metabaptista, while the other with one pistillate specimen is named by Loesener “J. Fargesii var. vel forma 8 Bodinieri Loes. forma nova.” The labels on each sheet contain the same informa- tion and give two localities, though one sheet has only a single speci- men. All three are slightly different. The sheet named J. metabaptista by Loesener with two staminate specimens has slightly pilose in- florescences and thus approaching the type of the species, but the pubescence is less dense and the leaves are smaller than in the type. The one named /. Fargesii f. Bodinieri by Loesener is a pistillate speci- men with narrow leaves up to 6.5 cm. long and glabrous inflorescences; on account of the latter character I refer it to var. myricoides. From Ilex Fargesti to which Loesener referred it, it is at once distinguished by the short petioles which are only half as long as the inflorescence, while in J. Fargesti the petioles are 1-1.5 cm. long and exceed the inflorescence, also the leaves are larger and broader and distinctly serrulate to below the middle. The variety has also been collected in Kweichou by Y. Tsiang (no. 8525) near Tin-fan, Binshaw, and by Steward, Chiao and Cheo (no. 950) near Hung-shieh-lang, Kang-kou-hseh; in the latter specimen some of the leaves are up to 6 cm. long. Tsiang, no. 8525, was first identified by E. D. Merrill with Maesa myrsinoides Lévl. and dis- tributed as J. myrsinoides (Lévl.) Merr. Ilex corallina Franchet in Bull. Soc. Bot. France, xxxim. 452 (1886).—Loesener in Nov. Act. Leop.-Carol. Akad. txxvim. 327 (Monog. Aquifol.) (1901). Ilex Dunniana Léveillé in Fedde, Rep. aa. Nov. 1x. 458 (1911); Fl. Kouy-Tchéou, 200 (1914).—Synon. no 242 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Cuina. Kweichou: Long-ly, J. Cavalerie, no. 3000, May, 1908, “petit arbre, fleurs jaunes” (holotype of J. Dunniana; merotype in A. A.). Cavalerie’s specimen represents a form with broad leaves rather closely denticulate with upright-spreading not incurved teeth. Léveille had recognized the close relation of his species to J. corallina, but the differences he gives are either trivial or do not hold. Ilex corallina Franch. var. Loeseneri Léveillé, Fl. Kouy-Tchéou, 200, (1914), nomen. A typo recedit foliis angustioribus spinuloso-dentatis. Cuina. Kweichou: Kouy-yang, mont du Collége, J. Chaffan- jon in herb. Bodinier, no. 2242, Apr. 1898 (syntype; photo. in A. A.); Pin-fa, montagnes, J. Cavalerie, no. 580,” arbrisseau, fl. odor.’ (syn- type; photo. in A. A.). The two specimens cited above bear Loesener’s identification “Jlex corallina Franch. forma” on the sheets without any descriptive note, neither does Léveillé give a description. From the type this variety differs chiefly in the spinulose-serrate narrower leaves; in Cavalerie’s no. 580 the spinulose mucro of the teeth is about 1 mm. long, but in Chaffonjon’s specimens the mucro is much shorter and the leaves measure 1.5-2 cm. in width, while in Cavalerie’s specimen they are 8-10 2-2.5 cm. Ilex macrocarpa Oliver in Hooker, Icon. xvi. t. 1787 (1888).— Loesener in Nov. Act. Leop.-Carol. Akad. Lxxvi1. 489 (1901); Ber. Deutsch. Bot. Ges. xxxtr. 543 (1914).—Léveillé, Fl. Kouy-Tchéou, 200 (1914). Celastrus salicifolia Léveillé in Fedde Rep. Spec. Nov. x11. 263 (1914). Diospyros Bodiniert Léveillé, Fl. Kouy-Tchéou, 144 (1914).— FS) CuHina. Kweichou: without precise locality, J. Esquirol, May 10, 1906 “arbre” (holotype of Celastrus salictfolia; photo. in A. A.); Kouy-yang, mont du Collége, J. Chaffanjon in herb. Bodinier, April 1898 “arbre” (holotype of Diospyros Bodinieri; merotype in A. A Celastrus salicifolia was first identified with Jlex macrocarpa by Loesener (1. c.), and Léveillé enumerates in his Flore de Kouy-Tchéou the type of C. salicifolia under I. macrocarpa var. genuina Oliv., but does not cite C. salicifolia as a synonym. CELASTRACEAE Evonymus grandiflora Wallich in Roxburgh, Fl. Ind. ed. Carey, 11. 404 (1824).—Loesener & Rehder in Sargent, Pl. Wilson. 1. 484 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 243 (1913).—Loesener in Ber. Deutsch. Bot. Ges. xxxit. 540 (1914).— Léveillé, Cat. Pl. Yun-Nan, 34 (1915). Evonymus Maret peveile a eee Rep. Spec. Nov. xu. 260 (1914), eae specim. ‘‘Siao- Cu1na. Yunnan: ae ea de mont. a Tong-tchouan, alt. 2500 m., E. E. tae, May 1912 “petit arbre, feuilles caduques, fleurs blanc-jaunatre”; haies 4 pied de mont. a Tche-hai, alt. 2500 m., E. E. Maire, Aug. 1912 “petit arbre, toujours vert, fleurs jaunatres, fruits item’; brousse des rochers 4 Ma-hong, alt. 2700 m., E. E. Maire, May 1912 “petit arbre, fleurs jaunes, fruits rouges”; (all three syntypes of E. Mairei; photos. in A. A.). Evonymus Mairei was first identified with E. grandiflora by Loes- ener (I. c.). Evonymus yunnanensis Franchet in Bull. Soc. Bot. France, XxxtmI. 454 (1886). Evonymus Mairei Léveillé in Fedde, Rep. Spec. Nov. x11. 260 (1914), quoad specim. ‘‘Siao-ho. Cutna. Kweichou: mont. derriére Siao-ho, alt. 2800 m., E. E. Maire, June 1912 “grande arbuste, feuilles caduques, fleurs jaunatres” (syntype of E. Mairei; photo. in The specimen cited above was identified with E. yunnanensis by H. F. C[omber] according to a note on the sheet. Specimens recently collected in Kweichou by Y. Tsiang under nos. 9124 and 9133 belong here. Evonymus uniflora Léveillé & Vaniot in Bull. Soc. Agr. Sci. Arts Sarthe, trx. 320 (Bouquet Fl. Chine, 5) (1904); in Fedde, Rep. Spec. Nov. vr. 374 (1908); Fl. Kouy-Tchéou, 73 (1913).—Loesener in Ber. Deutsch. Bot. Ges. xxx1r. 539 (1914). Evonymus Blinii Léveillé in Fedde, Rep. Spec. Nov. x11. 259 (1914), quoad specim. Esquirol, no. 478. Cutna. Kweichou: Pin-fa, sud ouest, bord d’un ruisseau, J. Cavalerie, no. 256, Aug. 21, 1902 (holotype of E. uniflora; merotype in A. A.); without locality, J. Esquirol, no. 478, (fruit) (syntype of £. Blinii, and cited in Fl. Kouy-Tchéou; photo. in A. A.). There are two entirely different specimens marked Esquirol, no. 478, in the Léveillé herbarium. One is a fruiting specimen, named on the label E. uniflorus by Léveillé; this is apparently the syntype of E. Blinii, upon which the description of the fruit was based. The other is a flowering specimen which bears the name E£. Blinii in pencil in Léveillé’s handwriting, which may be the specimen referred by Loes- ener to E. theaefolia Wall., but I am unable to distinguish this speci- men which is in flower, from E. Forbesiana Loes. 244 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV This species has also been collected recently in Kweichou by Y. Tsiang (no. 4579). Evonymus theaefolia Wallich, Cat. 4293 (1828), nomen.—M. A. Lawson in Hooker f., Fl. Brit. Ind. 1. 612 (1875).—Léveillé, Fl. Kouy- Tchéou, 73 (1914).—Loesener in Ber. Deutsch. Bot. Ges. xxxt1. 541 (1914). ? Evonymus Blin Leéveillé in Fedde, Rep. Spec. Nov. x11. 259 (1914) quoad specim. “Esquirol, no. 478.” The syntype of FE. Blinti, J. Esquirol, no. 478 (flowers), was first identified by Loesener (1. c.) with E. theaefolia Wall. as “E. theifolia var. vel forma” and is cited by Léveillé as a synonym of the latter. The specimen, however, of E. Blinii, Esquirol, no. 478, which I have before me, I cannot distinguish from FE. Forbesiana Loes., but there may be or may have been another specimen of the same number which I have not seen and which may belong to E. theaefolia Wall. See also my remarks under the preceding species. Evonymus theaefolia Wall. occurs in Kweichou, as Y. Tsiang’s no, 7274 shows which was collected in 1930 near Gan-wu, Lohou in south- ern Kweichou. Evonymus Esquirolii Léveillé in Fedde, Rep. Spec. Nov. x1. 261 (1914); Fl. Kouy-Tchéou, 71 (1914).—Loesener in Ber. Deutsch. Bot. Ges. xxxiI. 539 (1914). Cuina. Kweichou: “Tang-tchang (Hoang-tsao-pa), J. Es- quirol, no. 1569, June 1909 (holotype; merotype in A. A.). Related to E. myriantha Hemsl. according to Loesener from which it differs chiefly in the much smaller leaves and denser less branched cymes. Evonymus Leclerei Léveillé in Fedde, Rep. Spec. Nov. x11. 260 (1914); Fl. Kouy-Tchéou, 72 (1914).—Loesener in Ber. Deutsch. Bot. Ges. xxx1r. 539 (1914). Cuina. Kweichou: Ma-jo, J. Cavalerie, no. 3058, Sept. 5, 1907 (holotype; merotype in A. A.). This species is like the preceding closely related to E. myriantha Hemsl. from which it differs in the longer leaves more densely and dis- tinctly reticulate above. Evonymus centidens Léveillé in Fedde, Rep. Spec. Nov. x1. 262 (1914); Cat. Pl. Yun-Nan, 34 (1915).—Loesener, in Ber. Deutsch. Bot. Ges. xxxi1, 262 (1914).—Rehder in Jour. Arnold Arb. x1. 164 (1930). Cuina. Yunnan: collines broussailleuses a Long-ky, alt. 700 m., E. E. Maire, June 1912 “grand arbuste a feuilles caduques, fleurs jaunatres, fruits rouges” (holotype; merotype in A. A.). 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 245 Of this species I gave a full description in 1930 (1. c.) based on the specimen cited above and Fang’s no. 5819 Evonymus Dielsiana Loes. var. Y latifolia Loesener in Bot. Jahrb. xxx. 455 (1902); in Ber. Deutsch. Bot. Ges. xxx. 540 (1914). Evonymus Cavaleriei Léveillé in Fedde, Rep. Spec. Nov. x11. 259 1914). Evonymus Dielsiana Léveillé, Fl. Kouy-Tchéou, 71 (1914), non Loesener sensu stricto. Cuina. Kweichou: Pin-fa, rochers de Ouen-pi, Li-tseou-gai, J. Cavalerie, no. 87, July 23, 1902 “haut 3-4 m.” (syntype of E. Cava- leriei; merotype in A. A.); Pin-fa, J. Cavalerie, no. 865, Feb. 17, 1903 (syntype of Z. Cavaleriei; photo. in A. A.). Evonymus Cavaleriei Léveillé in Fedde (not in his Fl. Kouy-Tchéou) was first identified with E. Dielsiana var. latifolia by Loesener, but Léveillé in his Flore de Kouy-Tchéou omits the variety. Evonymus Feddei Léveillé in Fedde, Rep. Spec. Nov. x11. 260 (1914); Fl. Kouy-Tchéou, 72 (1914).—Loesener in Ber. Deutsch. Bot. Ges. xxxi1. 539 (1914). CHINA. Kweichou: route de Pin-fa a Lo-fou, J. Cavalerie, no. 3353, April 1908 (holotype; merotype in A. A.). According to Loesener this species is related to E. attenuata Wall. and E. bullata Wall.; from the former it differs in the much larger leaves and more branched inflorescences and from the latter in the narrower leaves and less branched inflorescences. Evonymus Rehderiana Loesener in Sargent, Pl. Wilson. 1. 488 (1913); in Ber. Deutsch. Bot. Ges. xxxtr. 540 (1914).—Léveillé FI. Kouy-Tchéou, 73 (1914). Evonymus bicolor Léveillé in Fedde, Rep. Spec. Nov. x11. 260 (1914). Evonymus proteus Léveillé, 1. c. Cuina. Kweichou: Long-ly, Ma-jo, J. Cavalerie, no. 2238 (in part), Nov. 13, 1907 “arbre” (holotype of E. bicolor; photo. in A. A.); Paitchen, J. Cavalerie, no. 2238 (in part), March 30, 1905 ‘3 or 4 m. de haut, les fleurs et les jeunes feuilles vert-jaunes” (holotype of E. proteus; photo. in A. A.). Evonymus bicolor and E. proteus were first identified with E. Reh- deriana by Loesener (1. c. 1914). Both specimens cited bear the num- ber 2238. The published name E. proteus for the specimen from Paitchen does not appear on the original label, but instead another name referring to the color of the branches. Also collected in Kwei- chou by Y. Tsiang (nos. 5121 and 9029). 246 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Evonymus Forbesiana Loesener in Bot. Jahrb. xxx. 457 (1902); in Ber. Deutsch. Bot. Ges. xxx. 540 (1914). —Léveillé, Fl. Kouy- Tchéou, 72 (1914). Evonymus Crosnieri Léveillé in Bull. Soc. Bot. France, LI. p. exlvi Evonymus Blinit —. in Fedde, Rep. Spec. Nov. xu. 259 (1914), quoad specim. “Esquirol, no. 478.” Evonymus Vaniott Léveillé, be (1914 3 Cura. Kweichou: Pin-fa, bord des ruisseaux, J. Cavalerie, no. 1274, June 2, 1902 “fl. rousse” (holotype of E. Crosnieri; photo. in A. A.); Pin-fa, J. Cavalerie, no. 1272, March 10, 1902 “fl. rousse”’ (syntype of E. Vanioti; photo. in A. A.); moulins de Tong-tchéou, J. Esquirol, no. 3236, July 10, 1912 (syntype of E. Vanioti; photo. in A. A.); Pin-fa, J. Cavalerie, no. 1273, May 5, 1902 (syntype of E. Blinii; photo. in A. A.) ; without precise locality, ““coteaux,” J. Esquirol, no. 478 (flowers), June 1905 (? syntype of E. Blinii; photo. in A. A.). Evonymus Crosnieri, E. Vanioti and one of the syntypes of E. Blinu were first referred to E. Forbesiana by Loesener (1. c.), and the first two names placed as synonyms under E. Forbesiana by Léveillé in his Flore de Kouy-Tchéou. Evonymus Blinii is not cited as a synonym, but Esquirol’s no. 478, one of the syntypes of that species which, however, partly belongs to E. uniflora Lévl., while Cavalerie, no. 1273, which belongs here, is not cited. Evonymus aculeata Hemsley in Kew Bull. Misc. Inform. 1893, p. 209.—Sprague in Kew Bull. Misc. Inform. 1908, p. 33. ee eee hederaerhiza Léveillé in Fedde, Rep. Spec. Nov. x. 474 (1912).—Synon. nov. Evonymus acanthocarpa Léveillé, Fl. Kouy-Tchéou, 71 (1914); non Franchet. Cutna. Kweichou: Pin-fa, rampant sur les rochers, J. Cav- alerie, no. 2761, April 1906 (holotype of Echinocarpus hederaerhiza ; photo. in A. A.). Echinocarpus hederaerhiza was referred by Léveillé in his Flore de Kouy-Tchéou to Evonymus acanthocarpa Franch., but it differs from that species chiefly in the longer compressed spines of the fruit, in the black bud-scales persisting at the base of the branchlets for some time and in the non-verruculose branchlets. Evonymus acanthocarpa Franchet, Pl. Delavay. 129 (1889).— Sprague in Kew Bull. Misc. Inform. 1908, p. 32. Echinocarpus ii slip Léveillé in Fedde, Rep. Spec. Nov. x. 474 (1912).—Synon. nov. Evonymus panies hanes Léveillé, Fl. Kouy-Tchéou, 72 (1914). Cuina. Kweichou: environs de Gan-pin, dans la montagne, 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 247 L. Martin in hb. Bodinier, no. 2493, Oct. 9, 1898 ‘“‘arbuste lianeux” (holotype of Echinocarpus erythrocarpa; merotype in A. A.). Evonymus subtrinervis, nom. nov. Bend tol Esquirolit Léveillé in Fedde, Rep. a ae x. 474 re Goniems Esquirolu pe ae ys a "Cavalerici Léveillé, 1. ¢.— ; meee Blinii Léveillé, Fl. Kouy-Tchéou, 71 (1914) : non Léveillé n Fedde, Rep. Spec. Nov. x11. 259 (1914). Ends Cavaleriei Léveillé, Fl. Kouy-Tchéou, mye (1914) ; Léveillé in Fedde, Rep. Spec. Nov. x11. 259 (1914). Frutex glaber, ramis gracilibus tenuiter verruculosis quadrangulatis angulis subalatis; gemmae terminales parvae pauciperulatae, obtusae, pallide cinereo-virescentes. Folia persistentia chartacea, ovato-oblonga vel oblongo-lanceolata, 4.5-9 cm. longa et 1.8-3 cm. lata, acuminata, basi late cuneata vel rarius fere rotundata, fere ad basin denticulata denticulis mucrone fusco inflexo, basi trinervia, costa medio utrinque elevata, nervis lateralibus 4-6, pari inferiore superioribus aequilongo vel longiore, supra in sicco manifeste, subtus leviter elevatis, supra laete viridia, reti nervulorum in sicco visibili, subtus pallida, reti obsoleto; petioli 2-4 mm. longi, tenues. Flores desiderantur. Inflorescentiae fructiferae bis dichotomae; pedunculus circiter 1.5 cm. longus, gracilis, quadrangularis, ramulis circiter 1 cm., pedicellis 2-3 mm. longis: cap- sula subglobosa, aculeis inclusis circiter 1 cm. diam., pallide flavo- carnea, aculeis complanatis 1-1.5 mm. longis, dissitis non dense aggre- gatis; semina 6 mm. longa, nigra, arillo pallide aurantiaco apice lacini- ato aperto excepto inclusa. CHINA. Kweichou: without precise locality, J. Esquirol, no. 844 (holotype of Echinocarpus Esquirolii, E. Blinii and E. subtrinervis ; photo. and merotype in A. A.); without precise eres J. Esquirol (holotype of Echinocarpus Cavaleriei; photo. in A. The specimens cited above I have not been able to identity with any previously described species; they seem nearest to E. echinata Wall. from which they are easily distinguished by the leaves being 3-nerved at the base, with fewer veins diverging at a more acute angle. From E. Hemsleyana Loes. they differ in the shape of the leaves, shorter petioles, and the shorter spines of the fruit and smaller winter-buds. As Le- veillé’s description is insufficient I have drawn up a more complete description based on Esquirol no. 844. The other specimen, the type of Echinocarpus Cavaleriei, is rather poor with all the leaves dropped from the branches and very young fruit; it differs from the type in the broader leaves, ovate-oblong, partly rounded at base, about 5 cm. long and 2.2 cm. broad. Unfortunately none of the names proposed by Léveillé can be used 248 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XIV for this species. The two specific epithets under Echinocarpus cannot be transferred to Evonymus since the resulting combinations are pre- occupied and Evonymus Blinii, based on Echinocarpus Esquirolii is in- validated by an earlier homonym, which by Léveillé was considered non-valid, because it had been reduced to synonymy. It should be noted here that in 1914 three important publications relating to Evony- mus came out: Léveillé’s descriptions of many new species in Fedde’s Repertorium on May 5, 1914, Loesener’s corrections in Ber. Deutsch. Bot. Ges. on July 30, 1914, and Léveillé’s Flore de Kouy-Tchéou which contains these corrections, in September or October of that year. Evonymus Maackii Ruprecht in Bull. Phys. Math. Acad. Sci. St. Pétersb. xv. 358 (1857 ).—Loesener in Ber. Deutsch. Bot. Ges. xxx. 540 (1914).—Léveillé, Fl. Kouy-Tchéou, 71 (1914). Evonymus coreanus Léveillé in Fedde, Rep. Spec. Nov. vir. 284 Korea: in dumosis Chinnampo, U. Faurie, no, 520, Aug. 1906 (holo- type of E. coreanus, isotype in A. A.). Evonymus coreanus was first referred to E. Maackit by Loesener (1.6); Evonymus Hamiltoniana Wallich in Roxburgh, Fl. Ind. ed. Carey, 1. 403 (1824)—Loesener in Ber. Deutsch. Bot. Ges. xxx. 541 (1914). —Léveillé, Fl. Kouy-Tchéou, 72 (1914). Evonymus rugosa Léveillé in Fedde, Rep. Spec. Nov. xr. 261 (1914). Evonymus Darrisit Léveillé, 1. c—Synon. nov. Evonymus Maackii Léveillé, Fl. Kouy-Tchéou 73 (1914); non Ruprecht. Cutna. Kweichou: Hoang-tsao-po, colline de la pagode, J. Esquirol, no. 1532 (holotype of E. rugosa; photo. in A. A.); without precise locality, J. Esquirol, no. 711 (holotype of E. Darrisii; photo. in A. A.). Evonymus rugosa was first referred to E. Hamiltoniana by Loesener with some doubt, while he identified E. Darrisiti with E. Maackii. Though E£. Darrisii in its leaves resembles E. Maackii rather than E. Hamiltoniana, the fruit seems to agree better with that of the latter species. Also for geographical reasons it seems unlikely that the Kwei- chou plant belongs to E. Maackii which has not yet been recorded from central and western China. Evonymus lanceifolia Loesener in Bot. Jahrb. xxx. 462 (1902); in Ber. Deutsch. Bot. Ges. xxx. 541 (1914).—Léveillé, Fl. Kouy- Tchéou, 72 (1914). Evonymus Bodinieri Léveillé in Fedde, Rep. Spec. Nov. x11. 261 ). 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 249 Cutna. Kweichou: Gan-chouen, J. Cavalerie, no. 3824, June 1910 “petit arbre” (holotype of E. Bodinieri; photo. in A. A.). Evonymus Bodinieri was referred to E. lanceifolia by Loesener with the remark that it approaches E. Hamiltoniana. The type specimen consists of two branches, a flowering and a fruiting branch; according to the dark colored anthers the specimen belongs to E. lanceifolia, for E. Hamiltoniana differs from both closely related species E. Maacku and E. lanceifolia in the yellowish white anthers. The label of the type specimen bears in Léveillé’s handwriting an unpublished name referring to the color of the leaves, to which is added rather indistinctly written in pencil “vel Bodinieri.” Evonymus alata (Thbg.) Regel, Fl. Ussur. 40, t. 7 (1861).— Loesener & Rehder in Sargent, Pl. Wilson. 1. 493 (1913). Microrhamnus Taquetii Léveillé in Fedde, Rep. Spec. Nov. vir. 284 (1910).—Synon. nov. Korea. Quelpaert: secus torrentes, T. Taquet, no. 153, Sept. 1907 (holotype of Microrkhamnus Taquetii; fragments of type and isotype in A. A Besides the type specimen the cover of Microrhamnus Taquetti con- tains Taquet, no. 4095, consisting of flowering branches of E. alata; both specimens belong to the form aptera Regel with the branches not or very slightly winged. Evonymus disticha Léveillé in Fedde, Rep. Spec. Nov. x1. 261 (1914); Fl. Kouy-Tchéou, 71 (1914).—Loesener in Ber. Deutsch. Bot. Ges. xxxir. 540 (1914). CHINA. Kweichou: environs de Kouy-yang, bois de la pagode de Kien-lin-chan, E. Bodinier, no. 2455, June 27, 1898 “petit arbuste, fleurs jaunes” (holotype, merotype in A. A.). This species was collected also at the same locality in 1930 by Y. Tsiang, no. 8488. This specimen bears flowers and young fruits which differ from those of EF. alata in their shorter and broader lobes. Celastrus Vanioti (Lévl.), comb. nov. Saurauja Vanioti Léveillé, Fl. Kouy-Tchéou, 415 (1915, before October. ) Celastrus spiciformis Rehder & Wilson in Sargent, Pl. Wilson. 11. 348 (1915, Dec. 28).—Synon. nov. Cuina. Kweichou: environs de Kouy-yang, route du Col- lége, dans les haies, E. Bodinier, May 30, 1898 “‘liane ligneuse”’ (holotype of Saurauja Vanioti; photo. and merotype in A. A.). Saurauja Vanioti is without doubt identical with Celastrus spiciformis, the former differing only slightly in the somewhat more reticulate leaves resembling in this respect C. gemmata Loes., but paler on the under 250 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV surfaces. As Saurauja Vanioti was published several months earlier, the Flore de Kouy-Tchéou having been received at the Arnold Arbore- tum in October 1915, the transfer of the specific epithet to Celastrus becomes necessary. Celastrus gemmata Loesener in Bot. Jahrb. xxx. 468 (1902).— Rehder & Wilson in Sargent, Pl. Wilson. 1. 352, (1915). Embelia Esqutrolii Léveillé in Fedde, Rep. Spec. Nov. x. 374 n. nov. Cutna. Kweichou: College, J. Esquirol, no. 4, May 10, 1906 “arbre, fleurs blanc-verdatres” (holotype of Embelia Esquirolii, mero- type in A. A.). Celastrus gemmata has been collected in Kweichou by Y. Tsiang under nos. 4661, 5360 and 6450. Celastrus stylosa Wallich in Roxburgh, Fl. Ind. ed. Carey, 1. 401 (1924) —Léveillé, Fl. Kouy-Tchéou, 69 (1914). Celastrus Cavaleriei Léveillé in Monde Pl. sér. 2, xviir. 31 (1916) ; non Léveillé (1914). Cutina. Kweichou: Pin-fa, ruisseaux, J. Cavalerie, no. 496, Sept. 16, 1902 (holotype of C. Cavaleriei; photo, in A. A.). Celastrus Cavaleriei was by Loesener according to his note on the type sheet referred to C. stylosa or a related species. Celastrus flagellaris Ruprecht in Bull. Acad. Sci. St. Pétersb. sér. 3, xv. 357 (1857); Decas Pl. Amur. t. 4 (1859).—Loesener in Ber. Deutsch. Bot. Ges. xxx. 541 (1914).—Rehder & Wilson in Sargent, Pl, Wilson. 11. 357 (1915). Celastrus clemacanthus Léveillé in Fedde, Rep. Spec. Nov. vii. 284 10 Korea. Quelpaert: scandens in muris agrorum Haouen, 7. Taquet, no. 632, May 8, 1908 (holotype of C. clemacanthus ; photo. and isotype in A. A.). Celastrus clemacanthus Léveillé was identified with C. flagellaris by Loesener (1. c.). Celastrus Hindsii Benth. var. Henryi Loesener in Bot. Jahrb. XXIx. 444 (1900); xxx. 467 (1902). Erythrospermum Cavalerici Léveillé, Fl. Kouy-Tchéou, 51 (1914).— Synon. nov. CutInaA. Kweichou: Gan-chouen, rochers, J. Cavalerie, no. 3976, in 1912, “liane” (holotype of Erythrospermum Cavaleriei; mero- type in A Gymnosporia Esquirolii Léveillé in China Rev. Ann. 1916, p. 18 (1916).—Schneider in Oester. Bot. Zeitschr. Lxvr. 140, in nota (1918). Berberis Cavaleriet Léveillé, Fl. Kouy-Tchéou, 48 (1914); non Léveillé in Fedde, Rep. Spec. Nov. 1x. 454 (1911). 1933] REHDER, LIGNEOUS PLANTS DESCRIBED BY LEVEILLE 251 Berberis Esquirolii Léveillé, Fl. Kouy-Tchéou, 47, in clavi (1914). wa. Kweichou: Lo-fou, colline du fort, 800 m., terrain rocheux aride, J. Esquirol, no. 3645, June 20, 1912 (holotype of Ber- beris Cavaleriei; merotype in A. A.). Gymnosporia Esquirolii is closely related to G. diversifolia Maxim. but easily distinguished by the slender peduncles of the inflorescence 0.5-1.5 cm. long and by the closely and finely crenulate or serrulate leaves, often acute at the apex and of thinner texture. To this species I also refer Siméon Ten, nos. 169 and 378, from Pe-yen-tsin, Henry, no. 9391, from Mengtze, and Forrest no. 10738. Gymnosporia Esqutrolit seems to be restricted to Yunnan and Kweichou, while G. diversifolia Maxim. is widely distributed through southeastern China, Indochina, Formosa and the Liukiu Islands. Gymnosporia Esquirolii was published without reference to a speci- men, only with the citation of Berberis Cavaleriet as a synonym. This citation refers without doubt to B. Cavaleriei Léveillé, Fl. Kouy- Tchéou, 48, which appears in the key on the preceding page as B. Esquirolii. This species is based on Esquirol’s no. 3645 and fragments were sent to C. Schneider by Léveillé in November 1915 as B. Esqui- rolii, but without indication of collector and number. These fragments agree with the specimen cited above and were referred by Schneider to Gymnosporia (in Sargent, Pl. Wilson. 11. 359). Léveillé apparently had first given to Esquirol’s no. 3645 the name B. Esquiroli, but when Schneider, to whom Léveillé had sent in October 1913 specimens of his earlier B. Cavaleriei based on Cavalerie, no. 3209, and published in 1911 -(in Fedde Rep. Spec. Nov. 1x. 454), had referred this species to B. Griffithiana Wall., he changed the name of his not yet published B. Esquirolii to B. Cavaleriei, wishing to keep Cavalerie’s name attached to a valid species, but forgot to change the name in the key. Berberis Cavaleriei of 1911, however, which was in 1913 considered a synonym of B. Griffithiana by Schneider, and was accepted as such by Léveillé in his Flore de Kouy-Tchéou with the citation of “B. Cavaleriei Lévl. olim” as a synonym, was revived as a valid species by Schneider in 1918 (in Oesterr. Bot. Zeitschr. Lxvi. 140) and applied to the Chinese species formerly referred to B. Griffithiana which according to his re- vised opinion does not occur at all in China. Gymnosporia acuminata Hooker f., Fl. Brit. Ind. 1. 619 (1875). Evonymus yunnanensis Léveillé, Fl. Kouy-Tchéou, 73 (1914) ; non Cuina. Kweichou: Lo-fou, J. Cavalerie, no. 3530, March 1909 (cited under E. yunnanensis Franch. by Léveillé; photo. in A. A.). The specimen cited above differs from typical G. acuminata of which 252 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XIV I have isotypes of Hooker’s and Griffith’s specimens before me, in the smaller and narrower scarcely acuminate leaves 7-9 cm. long and 2-3 cm. broad and rather closely serrulate to near the base, but fruit and inflorescence does not seem to differ. It looks rather different and I was inclined to consider it at least a distinct variety, if not Hemsley’s no. 13437 from Szemao represented an intermediate form, agreeing in the size of the leaves with the type and in the serration, though less dense and fine, with Cavalerie’s specimen. Tripterygium hypoglaucum (Lévl.) Hutchinson in Kew Bull. Misc. Inform. 1917, p. 101.—Léveillé, China Rev. Ann. 1916, p. 23 (Mscr.). Aspidopterys hypoglauca Léveillé in Fedde, Rep. Spec. Nov. 1x. 458 (1911). Pentace Virginis Léveillé in Monde Pl. sér. 2, xvitr. 28 (1916).— Synon. nov Cuina. Kweichou: Ma-jo, J. Cavalerie, no. 3316, Aug. 1908 “petit arbre lianeux” (holotype of Aspidopterys hypoglauca; merotype in A. A.). Yunnan: brousse du plateau de Ie-ma-tchouan, 3300 m., EZ. E. Maire, June 1912 “arbuste grimpant, long rameaux, fleurs blanches, fruit rouge sombre” (holotype of Pentace Virginis; photo. in , ade Hutchinson first recognized Aspidopterys hypoglauca as a Tryptery- gium and communicated the resulting new combination to Léveillé who published it in his “China; revue annuelle 1916” distributed as manu- script to the larger botanical institutions before Hutchinson’s paper was printed. (To be continued ) HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1933] ANDERSON, FLOWER COLOR IN HAMAMELIS 253 VARIATION IN FLOWER COLOR IN HAMAMELIS VERNALIS EpGar ANDERSON With one text figure WuatTEVER the ultimate explanation, there seem to be certain species of plants in which the variation between one individual and another is unusually marked. Hamamelis vernalis is such a species, for the plants of a single locality vary markedly in flower color, color pattern, time of blooming, and the number and arrangement of the flowers on the flowering branches. In this respect the species stands in sharp con- trast to H. virginiana which, though it runs into geographical varieties, is comparatively uniform in any one locality. In Hamamelis virginiana, variation in flower color, while not unknown, is extremely rare. Exam- ination of many individuals in eastern Massachusetts has revealed only two with a faint flush of red at the base of the petals. At Sterling, Massachusetts, for instance, of 35 individuals examined, 34 were pure yellow and one was flushed with red on the calyx and at the base of the petals. There is at the Arnold Arboretum a bush presumably col- lected in eastern Massachusetts which has light red petals and a yellow- ish green calyx (Rehder, 1922). A similar bush was collected some years ago near Malden, Massachusetts, by Mr. Edward L. Rand (Sar- gent, 1893). Mr. O. A. Farwell has collected a similar form with red calyx lobes at Stony Creek, Michigan and it has been distributed as Farwell, no. 3943. Variation in flower color is therefore not unknown in H. virginiana, though it is comparatively infrequent. In H. vernalis nearly every bush has its own distinctive flower color and it is not at all uncommon to find pure yellow-flowered plants grow- ing side by side with red-flowered ones. These differences seem to be germinal, since there is little variation between the flowers on a single bush and since the peculiarities of a particular bush persist after trans- plantation. The specimens in the living collections of the Arnold Arboretum maintain their characteristic flower colors season after season and the colors are perpetuated in plants propagated vegetatively. One color form has already received taxonomic recognition, Hamamelis ver- nalis forma carnea Rehd. The following records of variation in natural populations of H. ver- nalis (Tables 1 & 2) were made in St. Francis County, Missouri, on February 28, 1931. In each case the bushes, as is usual in this species, were growing in gravelly creek beds. At each locality, for a distance of 254 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV approximately one-eighth of a mile along the watercourse, a single rep- resentative twig was taken from each bush then in flower. The speci- mens were then taken to the laboratory and a record was made of their color and color pattern. The underlying yellow pigment seems to be practically the same shade throughout all the plants and nearly all of the variation in color is due to variations in the intensity and distribu- tion of the red pigment, which was found to be a water soluble sub- stance occurring in the epidermis of the petals and sepals. In the petals it varies from none at all, to a faint flush at the base, through inter- mediate stages up to 85% of the length. In their general tone the flowers varied from Light Cadmium Yellow (Ridgway) to Dragon’s Blood Red (Ridgway). The coloring always extends up from the base of the petals and even the reddest flowers were yellow at the very tips of the petals. On the sepals the color seems to spread from the mid-vein. In the following records it has been summarized under four grades: “pure yellow.” no red pigment in the sepal. “line plus.” sepal red along the mid-vein with a faint flush of red at either side. “all but margin.” sepal red with a narrow yellow margin at either side. “entire.” red pigment distributed over the entire sepal. It will be seen that while the variation at the two localities was simi- lar, that the colors were on the average a little darker at Flat River than at Libertyville. While the intensity of the red coloring on the petals was associated with the coloring of the sepal, the correlation was not perfect. This is shown graphically in Table 2 which summarizes the data from both localities. The above tables present a graphic and objective summary of varia- tion in color pattern. The size and development of the flowering branches seem to be quite as variable, though it is difficult to record the variation objectively. My colleague, Mr. Ernest J. Palmer, tells me that there is a correspondingly great variation in the degree of pubescence of the leaves. This has been given varietal recognition in H. vernalis var. tomentella (Rehd.) E. J. Palmer. It is particularly interesting to find such variation in H. vernalis since it is a species of very limited distribution, being known only from the Ozark Mountains and adjacent lowlands to the south and west. It may well be an ancient species like many others in that area (Palmer and Steyermark, in press), since its closest affinities are with the Asiatic species of Hamamelis rather than with the American H. virginiana. Like them it is winter-flowering, like them it has an extensive develop- 1933] ANDERSON, FLOWER COLOR IN HAMAMELIS 255 ment of red pigment in the flower, like some of them it has a tendency to pubescent leaves. The genus is known from its fossil record (Berry, 1916) to be an ancient one which was formerly wide-spread in the northern hemisphere. It is commonly said that species of limited distribution are less vari- able than those more widely distributed. As far as Hamamelis vernalis and H. virginiana are concerned the reverse seems to be true. The approximate ranges of the two species are shown in Figure 1. Hama- melis virginiana covers an area many times as large as that of H. ver- nalis and yet in any one locality the variation in color, habit, and pubescence is much less. It is of course quite possible that variability Taste 1. Variation in petal color and sepal color of 63 plants of Hamamelis vernalis examined at Libertyville and Flat River, Missouri. Further explanation in the text. Libertyville, Missouri ~ a oO vo ks = x he = x s c= ba Ae, : a : a SS oH eS = 0 pure yellow, flush of red at 50 line plus petal base 50 entire, faint 25 line plus 25 entire, faint 75 entire 25 entire 25 entire, faint 25 entire 50 entire, faint 75 alt but margin 25 all but margin 25 entire, faint 0 line plus 25 all but margin 25 entire, faint 25 entire 50 entire, dark 25 pure yellow 25 entire, faint 25 entire 25 entire 25 entire 0 pure yellow, pure lemon yellow 25 entire 75 entire 25 all but margin, faint 75 line plus 25 all but margin, faint 0 faint flush, flush at petal base 50 line plus 75 entire 75 all but margin 25 line plus 25 all but margin 25 all but margin 256 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV TABLE I (Continued) Flat River, Missouri ~~ U4 vv 4 eT - 8 a @: Ss =. } io aoe 7 ae es 82 we 82 50 line plus 25 all but margin 25 entire 75 dark, all but margin 25 line plus 50 all but margin 25 all but margin 25 all but margin 75 entire 75 entire 25 all but margin 50 entire 25 line plus 50 entire 25 entire 75 entire 25 entire, dark 75 entire, dark 50 all but margin 75 entire 25 entire, faint 75 entire 25 entire, faint 75 entire 25 all but margin 50 line plus 25 all but margin (horizontal scale). Eac TaBLeE 2. Correlation in petal color (vertical scale) and sepal color t represents the combination found in a single individual. Solid dots, Libertyville; open circles, Flat River. , ENTIRE M ARGIN 4 LINE + ‘ YELLOW 0,8 € @ @ @ @.e @ ©.0 foe Me) ‘wae @ ~ eaua ee 8 e @.0ee ee wor 90 oO 1933] ANDERSON, FLOWER COLOR IN HAMAMELIS 257 between plants in any one locality and variation between one region and another are quite different processes. As far as inter-regional varia- tion is concerned H. virginiana is perhaps quite as variable as H. vernalis. SRL VAL pi +. Q - bs NG ieee tis a - Ficure 1. Approximate distribution of Hamamelis vernalis (solid black) and Hamamelis virginiana (diagonal lines). SUMMARY Hamamelis vernalis, a species of very limited distribution, is shown to exhibit a high degree of intra-regional variability. LITERATURE CITED M J., and J. A. logue of the plants of Missouri. (Ann. Mo. Bot. Gard.) EHDER, A. . Hamamelis virginiana L. f. rubescens. (Jour. Arnold Arb. 3:210.) SarGEnNT, C. S. (1893). Hamamelis virginiana. Silva of North America, 5:4.) ARNOLD ARBORETUM, HARVARD UNIVERSITY. 258 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV NOTES ON THE GENUS PINUS THE BLACK CONE OF PINUS PONDEROSA GEORGE RUSSELL SHAW IN SaRGENT’s SriLva and in his Manual, also in Sudworth’s Forest Trees of the Pacific Slope and in his Pine Trees of the Rocky Mountain Region, some of the cones of Pinus ponderosa Douglas are described as dark purple (nearly black) in color. The quotation below is copied from Sudworth in Bull. U.S. Dept. Agric. No. 460, p. 31 (1917). “The cones of some trees are a bright grass-green when mature, while those of other trees are a dark purple, there being no essential differ- ence between trees bearing cones so dissimilar in color.” In autumn when the cones are completely dried, the purple color dis- appears and is never found in’ herbarium collections. The ultimate color is brown, never purple or black. The explanation is simple. The purple color is a transient intermediate stage between the summer- green and the autumn-brown coloration of the maturing cones. Its life is brief and, as a consequence, it has escaped the notice of collectors. I have happened on it in five or six species, such as Pinus rigida Miller where the intermediate color is brown and the autumn-color is yellow, and also in Pinus Greggii Engelmann, where the intermediate color, strange to say, is a brilliant scarlet and the autumn-color is Naples yellow. The purpose of this article is to call attention to the inter- mediate color and to prevent further error from that source. 1933] BAILEY, VESTURED PITS IN DICOTYLEDONS 259 THE CAMBIUM AND ITS DERIVATIVE TISSUES NO. VII. STRUCTURE, DISTRIBUTION, AND DIAGNOSTIC SIGNIFICANCE OF VESTURED PITS IN DICOTYLEDONS I. W. BAILEY With four text figures and plates 61-63 INTRODUCTION IN CONNECTION with an extended investigation of plasmodesma, I have had occasion to examine the cell walls and the pit membranes of a wide range of Gymnosperms and Angiosperms. It is evident that many of the structures in the tissues of the higher plants which are hypothesized as evidence for the existence of protoplasmic connections can not be interpreted as such. The bordered pits in the vessels of the Leguminose and of certain other families of Dicotyledons are referred to as “‘sievelike” or “‘cribri- form,” a nomenclature based upon the assumption that the pit mem- branes are perforated by numerous small openings through which protoplasmic connections occurred in the immature vessel members. The sievelike appearance described by Dutailly (1), Jonsson (4), and others is not due to perforations of the pit membrane, but, as will be shown on the following pages, to minute outgrowths from the free surface of the secondary walls. Although the so-called cribriform pits of tracheary tissue lose much of their putative physiological significance, they appear to be of con- siderable value both in the systematic study of woods and in discussions concerning the relationships and classification of specific groups of Dicotyledons. STRUCTURE OF VESTURED (CRIBRIFORM) PITS In unstained longitudinal sections of the wood of the Leguminose, Myrtacee, Polygonacee, Lythracee, Combretacee, and of a number of other representatives of the Dicotyledons, the bordered pits have, in surface view, a punctate appearance due to the presence of refractive processes of varying forms. In sections treated differentially with Haidenhain’s hematoxylin and safranin, these processes are deeply stained and, in photomicrographs, appear as dark spots (Plate 61, figs. 1, 3, 5, and 7) or as reticulate structures (Figs. 2,4, and 6) on a lighter colored background. By carefully focusing at successive levels it is 260 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV possible to demonstrate that there are two entirely independent sets of deeply-staining processes in each bordered pit-pair and that the punc- tate appearance of the bordered pits in surface view is not due to an unevenly thickened or perforated pit membrane. Owing to the small size of the pits and to the thickness of the walls in the tracheary elements of most Dicotyledons, the internal structural details of bordered and half-bordered pit-pairs may be observed most accurately in sectional rather than in surface views. It should be emphasized in this connection, however, that for this purpose extremely thin sections, 5-7 microns, are essential. The pits illustrated in Plate 61, fig. 7 are shown in section in Plate 62, fig. 10. The thick, imperforate pit-membranes are in the median position, and the pit apertures and pit chambers are clearly visible. The dark-colored, toothlike processes obviously are attached to the overarching walls of the pit chambers and are not connected with the pit membranes. The more massive papillze B on TeXT FIGURE 1. CompBreTuM sPEcIFS. (A) Sectional view of bor- dered pit-pairs in the walls of adjacent vessels, showing coralloid out- growths from the overarching walls of the pit-chambers. (B) Sectional view of half-bordered pit-pair in the adjacent walls of a vessel (right) and of a parenchymatous element (left). The bordered pit is vestured, but the simple pit is not. are attached, close to the margins of the pit apertures and project diagonally towards the center of the pit chambers. Thus, in Plate 61, fig. 7, they are visible through the pit apertures, whereas the smaller peripheral papilla, on the contrary, are partly obscured by the inter- vening portion of the secondary wall. Plate 62, fig. 8 and Text fig. 1A are sections through the bordered pits illustrated in Plate 61, fig. 5. In these pits, as in the preceding ones, the deeply-stained processes are attached to the overarching walls and project into the pit chambers. They are characterized by having a distinctly branched or coralloid 1933] BAILEY, VESTURED PITS IN DICOTYLEDONS 261 structure, however. The pit membranes are shown in the median position in Text fig. 1, whereas in Plate 62, fig. 8 they are rumpled and deflected to the right or left. A somewhat different type of structure is shown in Plate 62, fig. 11, a section through the bordered pits illustrated in Plate 61, fig. 4. Here the dark-colored processes form loose mats of branching and anastomosing filaments which are attached to the over- arching walls of the pit chambers. Denser mats of finer texture which occlude the pit chambers are illustrated in Plate 61, fig. 2 and Plate 63, g. 17. TeExT FIGURE 2. EUGENIA DICHOTOMA DC. Sectional view of bor- dered pit-pair in the walls of adjacent fiber-tracheids, showing papillary projections from the margins of both the inner and the outer apertures. FER we we. 4 a: bry . Ae Lome > ey ei 4 “he erst PLA Sanus a * a's TEXT FIGURE 3. PARASHOREA PLICATA Brandis. (A) Sectional view of bordered pit-pair in the adjacent walls of a vessel (left) and of a short tracheid (right). Mats of fine texture fill the entire pit-cavities and project more or less into the lumens of the cells. (B) Sectional view of half-bordered pit-pair in the adjacent walls of a vessel (right) and of a parenchymatous element (left). The bordered pit is vestured, but the simple pit is not. The papillary, coralloid or filamentous processes are not confined to the pit chambers in all cases. As shown in Plate 62, fig. 12 and Text fig. 2, they may be attached to the margins of both the inner and the outer apertures. They may fill the entire pit cavities and project more 262 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV or less into the lumens of the cells (Plate 62, fig. 9 and Text fig. 3A). Not infrequently they occur on the inner surface of the secondary walls of the vessels (Text fig. 4 and Plate 63, fig. 16) as well as within the bordered pits. They appear to be confined to tracheary elements, how- ever. Thus, in half-bordered pit-pairs, they are present in the bordered pits of the tracheary elements but are absent in the simple pits of the adjoining parenchymatous cells (Plate 62, fig. 13 and Plate 63, fig. 18, Text figs. 1B and 3B). ay x Yt ea 5 {59 A \\ TEXT FIGURE 4. VOCHYSIA HONDURENSIS Sprague. Sectional view of bordered pit-pairs in the walls of adjacent vessels, showing branched and anastomosing projections from the overarching walls of the pit-chambers and from the inner surfaces of the vessels. In view of such facts as these, the terms sievelike or cribriform obvi- ously should not be used in discussing the structures originally described by Dutailly and Jonsson. The terms vestured pits and vestured walls have been substituted for them by the Nomenclature Committee of the International Association of Wood Anatomists. DEVELOPMENT OF VESTURED PITS A priori, the structure of vestured pits and of vestured walls might be interpreted as due to adhering extraneous material deposited in ma- ture tracheary cells during post mortem changes*in the drying of sap- wood or during the transformation of sapwood into heartwood. A study of living cells in sections of differentiating xylem reveals the fact that the curious processes in reality are formed by the cytoplasm during the later stages of the development of tracheary elements. It should be noted in this connection, however, that punctate appear- ances (Plate 63, fig. 15) may be produced at times by extraneous or coagulated material which accumulates in the bordered pits during post 1933] BAILEY, VESTURED PITS IN DICOTYLEDONS 263 mortem changes and particularly during the transformation of sapwood into heartwood. True vestured pits may be distinguished from such artifacts by constancy of form and distribution, as well as by differ- ential solubilities and other tests. In all the plants examined by me the vestured intervascular pits, when present in a species, occur in all vessels throughout a given specimen, in specimens from different parts of a tree, and in material from widely separated sources. The artifacts, on the contrary, are of extremely irregular and sporadic occurrence, not only in different specimens of a particular species, but even within the limits of a single section. DISTRIBUTION OF VESTURED PITS IN DICOTYLEDONS According to Record (6), pits with so-called cribriform membranes have been reported by Jénssen (4), Heiden (2), Ursprung (7), Moll and Janssonius (5), and Record as occurring in the secondary xylem of 20 families of Dicotyledons. In most cases, the number of species and genera listed is so restricted that it is not possible to formulate reliable conclusions concerning the occurrence and diagnostic value of vestured pits in specific groups of Dicotyledons. In 11 of the 20 families, the character is recorded in a single species or genus. It seems advisable, accordingly, to tabulate my observations upon 2660 species, 979 genera, 152 families, and 33 orders. Families with vestured pits are printed in italics in Table 1. The numbers of genera and species investigated in each family are recorded in the columns at the right. Families in which “cribriform pits” have been reported by other investigators are marked with an asterisk. TABLE I. mber of Nu r of Orders and Families eae studied ee VERTICILLATAE Casuarinacee 1 8 PIPERALES Piperacee 2 3 Chloranthacez 1 1 SALICALES Salicaceze 2 33 GARRYALES Garryacez 7 1 MYRICALES Myricacee z vA JOURNAL OF THE ARNOLD ARBORETUM TABLE I. (Continued) [VOL. XIV Orders and Families Number of genera studied Number of species studied LEITNERIALES Leitneriaceze JUGI ANDALES Juglandacez FAGALES Betulacez Fagaceze* URTICALES Ulmacez Moraceze Urticacez PROTEALES Proteacez SANTALALES Santalacez Olacacezx Octoknematacez Loranthacexe ARISTOLOCHIALES Aristolochiacez POLYGONALES Polygonacee* CENTROSPERMAE Chenopodiaceze Amarantaceze Nyctaginaceze Aizoacez Caryophyllacez RANALES Trochodendracez Himantandracez Cercidiphyllacez Ranunculacez Lardizabalacez Berberidacee Menispermacez Magnoliacez Annonacez Noe UH bd —= = De noes DAN —" —_- | HO eH CO 1933] BAILEY, VESTURED PITS IN DICOTYLEDONS 265 TABLE I. (Continued) sats Number of Number of Orders and Families genera studied species studied Myristicacez 4 6 Monimiacez 3 3 Lauraceze* 19 47 Hernandiacex* 4 6 RHOEADALES Papaveraceze a 4 Capparidacee 8 28 Crucifere a 4 Moringacez 1 4 ROSALES Saxifragacee 7 9 Pittosporacee S 17 Brunelliacez 1 if Cunoniacez 3 6 Hamamelidacez* 8 9 Eucommiacee 1 1 Platanaceze 1 4 Crossosomatacez 1 1 Rosaceze* 24 77 Connaracee 4 8 Leguminose* 91 198 Bauhiniez 2 19 GERANIALES Oxalidacez 1 1 Linacee 1 1 Humiriaceze Zz 2 Erythroxylaceze 1 + Zygophyllacez 3 Z Rutacee 22 46 Simarubaceze 9 16 Burseracee 7 1S Meliacez 18 56 Malpighiacee 13 os. Vochysiacee* 3 a2 Trigoniacez 1 5 Polygalacee 4 4 Dichapetalacee 1 1 Euphorbiacez* Bo ney Brideliee Po 6 JOURNAL OF THE ARNOLD ARBORETUM TABLE I. (Continued) [VOL. XIV Orders and Families Number of genera studied Number o species studied SAPINDALES Buxacez Coriariaceze Anacardiacez Salvadoracee Staphyleacee Icacinaceze Aceraceze Hippocastanacez* Sapindacee Sabiacez RHAMNALES Elaeocarpacee Tiliacez Malvaceze Bombacacez Sterculiacez Scytopetalacez PARIETALES Dilleniacez Eucryphiacez Ochnacea-Exalbuminose Albuminos Caryocaracee Marcgraviacee Theacez Guttiferz* Dipterocarpacee Fouquieracez Cistacez Winteranaceze Violacez — NRO + me Om O — Pr rR O&®ON DRY KEP NH Ne 1933] BAILEY, VESTURED PITS IN DICOTYLEDONS 267 TABLE I. (Continued) ake Number of Number of Orders and Families genera studied species ‘studied Flacourtiacez 15 31 Stachyuracez 1 1 Passifloracee 3 4 OPUNTIALES Cactacez 4 ’ 4 MYRTIFLORAE Oliniacee 1 2 Thymeleacee c | 3 Eleagnacee 3 6 Lythracee* 3 10 Sonneratiacee* 1 Z Blattiacee* 1 Z Crypteroniacee* 7 4 Punicacee i 7 Lecythidacez 18 45 Rhizophoracee 13 32 Nyssacez 1 4 Alangiacez 1 1 Combretacee* 7 24 Myrtacee* 15 89 Melastomatacee* 4 9 Oenotheracee* 1 3 UMBELLIFLORAE Araliacez* 25 66 Umbelliferze 2 : Cornacez* 7 15 ERICALES Clethracez 1 3 Ericacez 8 24 Epacridaceze 3 9 PRIMULALES Myrsinacez 8 19 EBENALES Sapotacez 11 49 Ebenacee 2 27 Symplocacee 2 6 Styracaceze 4 8 268 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV TABLE I. (Continued) . Number of Number of Orders and Families genera studied species studied CONTORTAE Oleacex* 16 59 Nathusia 1 1 Forestiera 1 5 Loganiacee 6 7 A pocynacee iS 29 Asclepiadacee* 1 2 TUBIFLORAE Convolvulacez 1 2 Polemoniacee 1 2 Hydrophyllacee 1 1 Boraginacee 5 19 Verbenacee 12 29 Labiate + 6 Solanaceze 6 11 Scrophulariacez* 6 9 Bignoniacee 11 25 Gesneriacee 1 1 Myoporacez a 3 RUBIALES Rubiacee* 41 78 Caprifoliacee* 5 18 Dipsacacez 2 2 CUCURBITALES Cucurbitacez 1 2 CAMPANULATAE Composite* 12 21 DIAGNOSTIC VALUE OF VESTURED PITS In the material examined by me, true vestured pits are either present throughout the secondary xylem of a species or genus or are entirely absent. A similar constancy in the presence or absence of these struc- tures appears to prevail in most families of Dicotyledons. In only four of the 152 families tabulated in Table 1, have I encountered vestured pits in certain representatives of a family and not in others. It should be noted in this connection, however, that in three of the four families the distribution of vestured pits correlates with major subdivisions. Thus, in the Leguminose, the vestured pits are present in all the species 1933] BAILEY, VESTURED PITS IN DICOTYLEDONS 269 and genera examined, with the exception of the Bauhiniex, whereas in the Euphorbiacee they are absent except in the Brideliee. They are present in the Exalbuminosz of the Ochnacee, but appear to be absent in the Albuminose. In other words, the Oleacez, are the only family in which the distribution of vestured pits fails to correlate closely with the systematic classification. It may be objected that cribriform structures have been reported in representatives of 13 families which are not italicized by me. families are listed in Table 2. TABLE 2 These Families Genera and species Reported by Fagacez Lauraceze Hernandiacez Hamamelidacez Rosacez Hippocastanacexe Quercus alba L. C-Cerris L. Q. pedunculata Ehrh. Q. obtusiloba Michx. Numerous genera and species Hernandia peltata Meisn. Altingia excelsa Nor. Cerasus serotinus hort. Prunus brigantiaca Vill. Aesculus Hippocastanum L. A. rubicunda hort. Jonsson (4) ““e cc Janssonius (3) Moll & Janssonius (5) Jonsson (4) Rhamnacez Phylica ericoides L. as Guttiferze Calophyllum Inophyllum L.’ Ursprung (7) C. Calaba Jacq. Record (6) Araliacez Hedera helix L. Jonsson (4) Cornacez Mastixia trichomata Blume Moll & Janssonius Scrophulariacee Veronica Andersoni hort. Jonsson (4) Caprifoliacez Composit Viburnum sundaicum Miq. Helichrysum moniliferum hort. Janssonius (3) Obviously it is essential to determine whether the ‘‘sievelike” appear- ances reported by Jonsson and others are due to the presence of ves- tured pits or to artifacts such as are produced during post mortem changes or during the transformation of sapwood into heartwood. As indicated in Table 1, true vestured pits do not occur in any of the 107 species of Fagacee that I have studied. Nor have I succeeded 270 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV in finding them in any of the numerous specimens of Quercus alba, Q. Cerris, Q. obtusiloba, or Q. robur L. (Q. pedunculata Ehrh.) that I have investigated. Not only are vestured pits entirely absent in 19 genera and 47 species of the Lauracee and in 77 species and 24 genera of the Rosacez, but also in numerous specimens of the Prunoidez, i. e., Prunus and its subgenera Prunophora, Amygdalus, Cerasus, and Padus. Similarly, I have been unable to find vestured pits in Altingia, the Bucklandiez, Altingiez, Parrotiee, and Hamamelidee of the Hama- melidacee; in Aesculus Hippocastanum, and five other species of the Hippocastanacez; in the Zizyphee and Rhamnez of the Rhamnacez; in Calophyllum Inophyllum and other Calophylloidee of the Guttifere; in Hedera helix and 65 other representatives of the subgroups, Scheff- leree and Araliee, of the Araliacee; or in the Mastixioidee and Cor- noidez of the Cornacee. Although I have failed to find true vestured pits in any of the families listed in Table 2, I have frequently encountered artifacts of various types. In such genera as Quercus, Altingia, Calophyllum, Hedera, Mastixia, etc., these artifacts may produce a punctate appearance which closely simulates that of vestured pits. Therefore, in view of the fact that punctate appearances were interpreted as evidence of a sievelike structure, it is not surprising that the species in Table 2 were recorded as having pits with cribriform membranes It should not be inferred from this that more detailed and extensive surveys may not reveal the presence of vestured pits in additional families of Dicotyledons. The vestured condition appears to have arisen independently a number of times. To assume that all plants with vestured pits are closely related or are derived from common ancestors which possessed such structures leads to a reductio ad ab- surdum. Vestured pits occur in the more highly specialized types of tracheary tissue and are absent in plants which have a primitive com- bination of structural characters in the xylem. If vestured pits have arisen independently a number of times, it is not unlikely that genera may ultimately be found in which these structures are present in certain species and are absent in others. It may be inferred, however, from a statistical analysis of Table 1, that sporadic distributions of vestured pits are likely to be of relatively infrequent occurrence in the case of subgenera, genera, and subfamilies That vestured pits are extremely useful diagnostic criteria in the systematic study of woods was clearly demonstrated during the course of the present investigation. Most collections of wood specimens, even when accompanied by herbarium specimens, contain a varying number of errors. In other words, the fact that an herbarium specimen and a 1933] BAILEY, VESTURED PITS IN DICOTYLEDONS 271 sample of wood bear the same number is unfortunately no guarantee that both specimens came from the same tree or that the herbarium specimen was correctly identified. Collections of woods commonly pass through a number of hands and may be subdivided, renumbered, or relabeled. Furthermore, transpositions are likely to occur during the preparation of microscopic slides unless a painstaking system of checking and rechecking is employed. In my reconnaissance of Dico- tyledons, I encountered true vestured pits in slides of putative repre- sentatives of a number of families which are not italicized in Table 1, and, conversely, ordinary bordered pits in families which are italicized. In all these families, with the exception of the Oleaceze, the aberrant specimens proved to be errors which could thus be eliminated from the collections. Similar combinations of anatomical characters occur not infrequently in the woods of families which are widely separated in the systematic classification of Dicotyledons. For example, the secondary xylem of the Osage Orange, Maclura pomifera (Raf.) Schneid., so closely re- sembles that of the Black Locust, Robinia pseudoacacta L., that it is extremely difficult to distinguish the woods by anatomical criteria. The woods of the two genera may be identified with certainty, however, by the fact that vestured pits occur in Robinia and are entirely absent in Maclura. As indicated in the accompanying text figures and photomicrographs, the number, size, and form of the refractive processes vary considerably in different plants. In certain species and genera the processes are confined to the intervascular pits, whereas in others they may occur as well in the half-bordered pit-pairs, in the fiber tracheids, or upon the inner surfaces of the tracheary walls. Such differences may obviously be utilized as diagnostic criteria in the systematic study and classifica- tion of woods. SUMMARY AND CONCLUSIONS 1. Many of the structures in the tissues of the higher plants which are hypothesized as evidence for the existence of protoplasmic connec- tions can not be interpreted as such. 2. The so-called sievelike appearance of the pits in the vessels of Leguminose and of other families of Dicotyledons is not due to perforations of the pit membranes, but to minute outgrowths from the free surfaces of the secondary walls. 3. These refractive processes which vary considerably in size, number, and form are not confined to the pit-chambers in all cases, but may occur on the inner surface of the secondary walls of the vessels. 272 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV 4. They appear to be restricted to tracheary elements; in half-bordered pit-pairs they are present in the bordered pits of the tracheary elements, but are absent in the simple pits of the adjoining paren- chymatous cells. 5. Pits which have these refractive processes may be referred to as vestured. 6. In the Dicotyledons (2660 species and 979 genera) examined by me, vestured pits are either present throughout the secondary xylem of a species or genus or are entirely absent. A similar con- stancy in the presence or absence of these structures appears to prevail in most, subfamilies and families. 7. Vestured pits, therefore, are of considerable value both in the sys- tematic study of woods and in discussions concerning the relation- ships and classification of specific groups of Dicotyledons. ACKNOWLEDGMENTS The text-figures used in this paper were drawn by my assistant, Mrs. Ernest C. Abbe. I am much indebted to Professor S. J. Record for his kindness in providing woods of various families of Dicotyledons. LITERATURE CITED 1. Dutattity, G. Sur l’existence de a al ease dans 7 bois de la racine nD Sa ae (Bull. Soc, Linn. Paris, 1:9-10. 1874.) 2. Herpen, H. natomische Char: akteristik der ear eR iS (Bot. Centralbl. 56:3. 1893.) 3. Janssonius, H. H. alee peng des Holzes der auf Java vorkom- menden Baumarten, I Leiden, 1926-193 4. Jonsson, B. Siebz Shaliche Poren in den trachealen a flag etre der P hanerogamen, hauptsachlich der Leguminosen. (Ber. Deutsch Bot. eo pase 1892.) 5. Mott, J. \ & Janssonius, H. H. oe re des Holzes der auf Java se ertck beni Baumarten, I-III. 1906-1918. 6. Recorp, S. J. Pits with cribriform Gea (Tropical Woods, peace 1925.) 7. Ursprune, A. Beitrage zur Anatomie und Jahresringbildung trop- ischer Sere Diss. Basel, 1900. DESCRIPTION OF PLATES PLATE 61 Fig. 1. Prosopis juliflora DC. Surface view of bordered pits, showing toothed appearance of pit apertures. Fig. 2. Duabanga moluccana Blume. Surface view ; of bordered pits, showing ly punctate appearance. ; Fig. 3. Fuchsia Colensot Hook. f. rface view of bordered pits, showing toothed appearance of elongated pit geri alee Fig. 4. Tibo uchina mutabilis Cogn. Surface view of bordered pits, showing mats of branching and a ee « 25( Fig. 5. Combretum spect of bordered pits, showing coarsely KC bi patois x 2500. Jour. ARNOLD ArB., VoL. XIV. PLATE 61 if hy hen of Oe Per ae her * ee st heh allt het ee ot oe et ee a i VESTURED PITS IN DICOTYLEDONS JOUR 8 . ARNOLD Arps., VoL. XIV. VESTURED Pits IN DICOTYLEDONS PLATE 62 Jour. Arnotp Ars., VoL. XIV. PLATE 63 VESTURED Pits IN DICOTYLEDONS 1933] NN oo Ke) — c—) —s — —" bo _— Ww —_— wn _— NX] BAILEY, VESTURED PITS IN DICOTYLEDONS 273 . Eugenia alternifolia Wight. Surface view of half-bordered 2500. pit-pair, showing irregularly punctate paar . Terminalia Chebula Retz. Surface view of bordered pits, 500. showing coarsely toothed appearance of pit ce x 2 PLATE 62 . Combretum species. Sectional view of pits illustrated in Fig. 5, showing c aga id outgrowths from the overarching walls of the pit chamber . Parashorea ‘plicata Brandis. Sectional view of three pairs of bordered pits in a adjacent walls of a vessel (left) and a short acnerd (right). ats of fine texture fill the entire pit cavities and project more or less into the lumens of the cells. 2700. . Terminalia chebula Retz. Sectional view of pits illustrated in Fig. 7, showing massive projections from the overarching walls of the pit chambers. The imperforate membranes are in the iti 00. Sa ae, . Tibouchina mutabilis Cogn. Sectional view of pits illustrated in F in Fig. showing attachment of mats to overarching walls of the pit chambers. 0. . Eugenia ce a DC. Sectional view of pit-pair in the adja- cent walls of fber- tracheids, showing toothed inner and outer erie x : . Parkinsonia Torre S. Wats. Sectional view of half- ya bordered pit-pairs in ‘the adjacent walls of a vessel (left) and parenchyma right). The bordered pits are vestured, but the simple pits are not. 2500. PLATE 63 . me Sectis retusa Spreng. Surface view of vestured inner surface of v ee) dein excelsum Griseb. Surface view of bordered pits, showing artifacts which simulate the deeply-staining X 25 Phat ae moluccana Bl. Sectional view of pits illustrated in Fig. 2, wing mats of fine texture within the pit chambers. Eugenia alternifolia Wight. Sectional view of half-hordered pit-pairs in the adjacent walls of a vessel (right) and paren- chyma (left). The bordered pits are vestured, but the simple pits are not. >< 2500. ARNOLD ARBORETUM, Ha ARVARD UNIVERSITY. 274 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV SPECIES HYBRIDS IN PLATANUS AND CAMPSIS KARL SAX With two text figures ACCORDING TO SEWARD (1931), Platanus is one of the oldest of broad-leaved trees, and fossil types have been found which are very similar to the fertile branches of modern species. This genus was one of the most widely spread of the earlier cretaceous dicotyledons. The cretaceous Platanus was more variable and had a much wider distri- bution than the existing species. The genus Platanus is now found in North America to Mexico and from southeastern Europe to India (Rehder, 1927). The American species, P. occidentalis, is hardier than the Old World species, P. ori- entalis ; but the two species must be rather closely related because they have produced the fertile hybrid P. acerifolia. According to Henry and Flood (1919), P. acerifolia originated as a natural hybrid between the Occidental and Oriental Plane. It probably originated in the Ox- ford Botanic Garden about 1670. The hybrid is intermediate in leaf and fruit characters and seems to possess unusual vigor. The hybrid is extensively used for planting in the streets of European towns where neither of the parental species will survive. Seedlings from the hybrid are variable, some of them resembling the parental species. The fact that the species hybrid is fertile, and segregates in the sec- ond generation, would indicate that the parental species have the same number of chromosomes and that their chromosomes are similar. The haploid chromosome number of Platanus has been reported to be 21 (Winge, 1917), 8 (Brouwer, 1924), and 20-22 (Bretzler, 1924). Per- manent smear preparations of pollen mother cells from P. occidentalis and P. acerifolia show clearly that the number of chromosomes is 21, as reported by Winge. The chromosomes are paired regularly at the first meiotic division, and there is no evidence of lagging chromosomes at any stage in the meiotic divisions. The chromosomes at the first meiotic metaphase, and at telophase, in the hybrid, are represented in Figure 1. One of the chromosomes is somewhat smaller than the others, especially when fixed in aceto-carmine solution. The chromo- somes are too small and numerous to permit an accurate determination of chiasma frequency, but both rod and ring chromosomes were ob- served. The average chiasma frequency is about 1.5 per bivalent. About 90 percent of the pollen is good in the hybrid. 1933] SAX, SPECIES HYBRIDS IN PLATANUS AND CAMPSIS 275 The apparent compatibility of the chromosomes from the two par- ental species in the first generation hybrid indicates that the Old and New World species have undergone no very fundamental changes since their segregation and differentiation. Although this genus has under- gone no very fundamental changes for a long period of time, and even though it is one of the oldest dicotyledons so far discovered, it does not possess characters which mark it as a primitive type or as an early member of an evolutionary series (Seward, 1931). It is of interest that Platanus is the only genus of the family Platancee. Another valuable hybrid between species from the Old and New World is Campsis Tagliabuana (C. hybrida) (Rehder, 1932). The parental species are C. radicans, from central and southern United States, and C. chinensis, from China. The American species seems to eng Ficure 1. Meiotic chromosomes of Platanus acerifolia at metaphase and telophase. The 21 bivalents divide regularly. X 2000. be hardier and more vigorous than the Chinese species, but the flowers of C. chinensis are more attractive. The hybrid has the good qualities of both parents. It is almost as attractive as the Chinese species and has the hardiness of the American species. The hybrid forms are grown in many gardens of southern Massachusetts (Anderson, 1933). The hybrid was first recorded by Visiani as having been raised by the brothers Tagliabue apparently some time before 1859, but doubtless it has originated independently elsewhere afterwards or even before 1859. Natural hybrids have also been obtained at the Botanical Gar- den in Washington, D. C. A more complete account of the origin and characteristics of Campsis Tagliabuana has been presented in the rnold Arboretum bulletin of popular information, by Edgar Anderson. The hybrids are partially fertile, and numerous segregates have been produced. The chromosome number is undoubtedly the same for both the parental species and the hybrid. There are 20 pairs of small chro- mosomes at the first meiotic division in C. radicans. No representa- 276 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV tives of C. chinensis were available for study, but a hybrid segregate much like the Chinese species, also has 20 pairs of chromosomes. The hybrid has 20 pairs of chromosomes which are perfectly regular in pairing and disjunction at meiosis. The chromosome number found in the hybrid is in accord with the count reported by de Vilmorin and Simonet (1927). The chromosomes are very small and usually form rod bivalents at the first meiotic division. The chiasma frequency is apparently little more than one per bivalent. e @ e.? ad oe? as e 8 a0e e e°e e e e . ° e ° ee «| fe »« e e e * «s ee ad ° ee © e e e e® e e @e a . & Ficure 2. Meiotic chromosomes of Campsis radicans, C. he and C. spec. (probably a hybrid segregate resembling C. chinensis). divisions are regular in the one parental species examined and in ie hybrids. > 2000. Although there is regular pairing and disjunction of the chromo- somes in Campsis Tagliabuana, about 50 per cent of the pollen is sterile. There is less than 5 per cent pollen sterility in C. radicans. Pollen sterility in species hybrids which have regular meiotic divisions is also found in other genera (Primula kewensis, diploid, Newton and Pellew, 1929; Deutzia and Philadelphus, Sax, 1931; Tradescantia, Sax and Anderson, 1933). Pollen sterility in such species hybrids might be caused by unequal interchange of chromosome segments in one of the parental species. The interchange of segments between non-homologous chromosomes has been found in a considerable number of different genera in slightly related families, and may be much more extensive than is indicated by the formation of rings or chains of chromosomes. In a species with a low chiasma frequency, segmental interchange would not result in chromosome rings, and if segmental interchange were unequal, chromosome pairing would be regular in both heterozygous and homozygous forms. The heterozygous types will be eliminated unless they possess a much greater survival value to compensate for their partial sterility or unless balanced lethals are involved. As a result, plants should originate which are homozygous for segmental inter- change chromosomes, as is known to be the case in Datura, Pisum, and Oenothera. If an individual homozygous for two segmental interchange chromo- somes (the minimum number possible) is crossed with a normal plant, the chromosomes should pair as bivalents and divide normally if the 1933] SAX, SPECIES HYBRIDS IN PLATANUS AND CAMPSIS 277 segmental interchange is unequal and if the chiasma frequency is low. But the random segregation of the chromosomes would result in 50 per cent non-disjunction of a chromosome segment. If both segments are essential for gametophyte development, the pollen sterility would be 50 per cent; if only the longer segment is essential, the pollen sterility would be 25 per cent. Pollen sterility would be almost complete in a plant heterozygous for four or five segmental interchange chromosomes. Segmental interchange might well be one of the factors involved in the differentiation of species. A form with relatively few segmental interchange chromosomes would tend to be isolated from the normal type because of the sterility of the F, hybrid between the two forms. Unless lethal factors are associated with the segmental interchange chromosomes, the homozygous forms should have a higher survival value. Variations originating in the different lines homozygous for chromosome structure would not be swamped by intercrossing and would tend to be developed more or less independently in different structural genoms. SUMMARY Platanus acerifolia, a natural hybrid between P. orientalis of south- eastern Europe and P. occidentalis of North America, has 21 pairs of chromosomes which pair regularly at meiosis. The hybrid is fertile, and genetic segregation occurs in the second generation. These facts indicate that the parental genoms are similar and are compatible with each other, even though the parental species have been isolated for a long period of time. Campsis Tagliabuana, a natural hybrid between C. chinensis from China and C. radicans from North America, has 20 pairs of chromo- somes which pair regularly at meiosis. Although the reduction divi- sions are regular, there is about 50 per cent pollen sterility in the F, hybrid. The association of regular chromosome pairing and partial or nearly complete pollen sterility in species hybrids may be the result of segmental interchange between non-homologous chromosomes in one or both parental species. The species hybrids in the above genera contain the desirable char- acters of the parental species and are especially valuable because of their hardiness. Such hybrids are good illustrations of what may be expected from many hybrids between Old and New World species. 278 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV LITERATURE CITED ANDERSON, EpGar (1933). Trumpet-creepers. (Arnold Arb. Bull. ser. 4, 1:1- 5 taints E. (1924). Beitrage zur Kenntnis der Gattung Platanus. ( Bot. Arch, 7 3388-417 Brouwer, J. (1924). Studies in Platanaceae. (Rec. Trav. Bot. Néer!. 136 2.) Henry, A., & M. G. FLoop Riders & The History of the London Plane. (Proc Roy. Irish Acad. NEwTon, W.C. F., & C. Pet w (1929). Primula kewensis and its derivatives. (Jour. Genetics, 20: 405467. Renper, A. (192 Eo big anua al ‘of cultivated trees and shrubs. Pp. 930. (Macmillan Co., a : ae Tagliabuana (Vis.), comb. nov. (Jour. Arnold Arb. 13:340. Sax, Kary (1931). erates numbers in the ligneous Saxifragaceae. (Jour. Arnold Arb. 12:198-206. Sax, Kart & EpGar Pecans (1933). Segmental interchange in chro- mosomes of beri (Genetics, 18:53-67. Sewarp, A. C, (1931). Plant life through the ages. Pp. 601. (Uni- versity Press, Cambridge, England. VILMORIN, R. pe & M. Simonet (1927). Nombre des chromosomes dans le genres Lobelia, oe et chez quelques autres espéces végétales (Compt. Rend. Soc. Biol. Paris, 96:166-8. WinceE, O. (1917). The chromosomes, their numbers and general im- portance. (Compt. Rend. Trav. Labor. Carlsberg, 13:131-275.) tn WwW CyToLocicaAL LABORATORY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1933] SAX, CHROMOSOME BEHAVIOR IN CALYCANTHUS 279 CHROMOSOME BEHAVIOR IN CALYCANTHUS KarL SAX With four text figures Two species of Calycanthus are found in eastern United States. Calycanthus floridus L. is found from Virginia to Florida while C. fertilis extends from Pennsylvania to Georgia and Alabama (Rehder, 1927). Herbarium material in the Arnold Arboretum includes C. fer- tilis from eleven localities in North and South Carolina and Georgia, and C. floridus from fifteen localities in South Carolina, Georgia and Alabama. The ranges of the two species overlap to some extent, but there is little evidence of hybridization although the two species are similar, differing chiefly in leaf characters. Varieties of the two species have been described, but in the case of C. floridus at least, the variety (ovatus) is rare and is apparently known only in cultivation, and appears to be of European garden origin. Chromosome counts were obtained from one plant of C. fertilis, two plants of C. floridus, and two varieties of these species. In both species there are eleven pairs of chromosomes at meiosis. The homologous chromosomes are united by terminal or subterminal chiasmata. The chiasma frequency is somewhat less than two per bivalent at early meta- phase. Some of the chromosomes are apparently heterobrachial and the separation of the short arms at late metaphase and early anaphase gives the impression of a prevalence of rod bivalents at the later stages although at early metaphase most of the chromosomes are ring bi- valents. The chromosomes of C. fertilis are shown at early metaphase (Fig. 1) and those of C. floridus are shown at late metaphase of the first meiotic division (Fig. 2). A variety of C. fertilis also had eleven pairs of chromosomes which pair and divide regularly at meiosis. The other variety in the Arbore- tum, C. floridus ovatus, is a triploid. At meiosis there are often eleven trivalents, although from one to four univalents are usually found. The trivalents are found in the form of chains, rings and rods, and Y’s (Fig. 3). There is some irregularity in the first meiotic division, in- cluding both trivalents and univalents, and the chromosomes are dis- tributed irregularly to the poles. The chromosomes at second metaphase following a comparatively regular division are shown in Figure 4. In one case a distribution of 20-13 was observed, but as a rule the numbers of chromosomes passing to each pole are approxi- 280 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV mately equal. Occasionally more than thirty-three chromosomes are found at the two poles, due presumably to a precocious division of one or more univalents. The pollen sterility of the triploid is about fifty per cent as compared with about five per cent in each of the pure species. SSS SN a & 2 4 Figure 1. CALYCANTHUS FERTILIS: sad peshesiipii at early metaphase of the first Fire division.—Figur THU : late metaphase.—Figure 3. Cae FLORIDUS OVATUS: a triploid show- ing trivalent chromosomes at meiosis.—Figure 4. The triploid variety gures were drawn from aceto-carmine preparations. Magnification x 1200 Since the two species of Calycanthus are similar in taxonomic char- acters and overlap in their distributions, the occurrence of natural hybrids might be expected. Some of the diploid varieties may be of hybrid origin, but there seems to be no extensive hybridization, and the two species are rather well differentiated. It is possible that tetra- 1933] SAX, CHROMOSOME BEHAVIOR IN CALYCANTHUS 281 ploid forms of these species exist and that the triploid variety is a hybrid between a tetraploid, C. floridus and a diploid C. fertilis, but it seems more probable that the ovatus variety is an autotriploid. The two plants of C. floridus in the Arboretum are typical for the species and both are diploids. The variety ovatus originated, or was first found, in a European garden and is not known to occur in nature. No species of Calycanthus is a native of Europe. Chromosome irregularity and pollen sterility have been considered as evidence of hybridity. In the case of Calycanthus floridus ovatus it is improbable that chromosome irregularity and pollen sterility can be attributed to species hybridization. An autotriploid originating within a species would be expected to show the chromosome irregulari- ties and pollen sterility. Chromosome irregularity at meiosis and pollen sterility can also be caused by segmental interchange with absolutely no change in the taxonomic characters of the plants involved. Tvra- descantia edwardsiana, for example, is a well marked species. The occasional segmental interchange plants show about fifty per cent pollen sterility, although they are taxonomically the same as the normal fertile plants (Sax & Anderson 1933). Chromosome irregularities may also be caused by variations in tem- perature and by genetic factors. On the other hand species hybrids often show regular chromosome pairing and division. In some of these hybrids there is much pollen sterility, but others are relatively fertile. Undoubtedly wide species crosses often result in hybrids which exhibit irregular chromosome behavior at meiosis, but chromosome irregularity is not necessarily evidence of species hybridization. LITERATURE CITED Reuper, A. (1927). Manual of cultivated trees and shrubs. Pp. 930. (O. McMillan Co., New York.) Sax, Kart & Epcar ANDERSON (1933). Segmental interchange in chro- mosomes of Tradescantia. (Genetics, 18:53-67.) ARNOLD ARBORETUM, HARVARD UNIVERSITY. 282 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV ORIGIN AND BEHAVIOR OF THE NUCLEOLUS IN PLANTS ‘- 2) Haic DERMEN With plates 64 to 67 TABLE OF CONTENTS INTRODUCTION 2... ccc ccc c ccc cece cece ccc eeecceeesessseess 285 MATERTAL. AND: METHODS © és 6.6:6-62245005:0.8oeeebbs Poa 08 ROR Ci ee 286 MorpPHOLOGY OF CHROMOSOMES IN RELATION TO THE NUCLEOLUS 287 Structural continuity of chromosomes in cycle ......... 287 II. Chromosome and nucleolar relationship .............. 290 III. Satellites and secondary constrictions ................. 292 ORIGIN AND BEHAVIOR OF THE NUCLEOLUS ........00-0eeeeee 294 Sne nucleolus in somatic cells 4...icicssscvencdencoves 295 Il. The nucleolus in microsporogenous cells .............. 297 IIT. Relationship of nucleolus to satellites and secondary OORIIORONE bod sori ne 0540 e nn eudseaneedeisatx 300 IV. Nucleolar symmetry in sister nuclei ................05. 300 V. The number and size of nucleoli in polyploid races ..... 302 THE BEHAVIOR OF THE NUCLEOLUS IN LIVING TISSUE ........ 305 I. The area around the nucleolus .................020005 305 II. Extranucleolar bodies ............. cece cece ecveeeeis 306 III. Chromosome and nucleolar relationship .............. 306 LV. ‘Ie wmucleolar vactoled ..ci0s bho ecaws eo 4 se see ea nee 307 GENERAL DISCUSSION AND CONCLUSIONS .........0.0c0eeeuee 310 I. Chromosome morphology and chromosome-nucleolar TEODOR vivorc tance ckancas Mouse eee enemas eises 310 II. Origin and development of the nucleolus .............. 310 III. Behavior and function of the nucleolus ............... 313 UMMA one te th codec aeunh cao sees REE RCE N ORES R SER eed eS 317 LETRA TUEE TRIN. ino 55 bs kk OFT KERR EA ed de becnee 4 319 19331 DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 283 A. INTRODUCTION THe Prosiem of the nucleolus is one which has attracted investi- gators for many years, but despite the large amount of work done on this nuclear body, the problem is far from solved. The origin, the development, the relation between nucleolus and chromosomes, and the nucleolar function are phases of the problem which have received atten- tion, but there is little concerted opinion regarding them. That the nucleolus is an essential component of cells is evidenced by the fact that it is present, very probably, in all types of cells of all organisms, a fact which may be responsible for Wilson’s question (1925) “whether the nucleolus may not play a more active and important part in cell metabo- lism than most writers have hitherto assumed.” The importance of a cell body is not necessarily dependent upon its continuity and permanence in a particular structural form, such as is characteristic of the chromosomes. Unlike the chromosomes _ the nucleolus possesses no constant structural form and is not even con- tinuous, in the strict sense of the word, appearing and disappearing during both mitosis and meiosis. This peculiarity alone has perhaps caused a confusion in evaluating its significance, but when its uni- versality is considered, it would appear that its function is one of vital import to the cell, and its appearance and disappearance in the cell cycle may be correlated with that function, which is probably of a metabolic nature. The present work is primarily intended.to determine the origin and to trace the development of the nucleolus through all the stages of mitotic and meiotic development ofthe nucleus in plants of diverse types and to see what relationship exists between the chromosomes and nucleolus. Many of the recent investigations are concerned with the origin of the nucleolus and its relation to the chromosomes. Its function has been, and still is, a matter of speculation. Van Camp (1924) believes that nucleoli originate from the chromosomes at telophase in the form of small globules which later fuse into larger forms, but still later, during the development of the chromosomes, through a direct contact established between a continuous spireme and the nucleolus, the nucle- olar substance flows back again over the chromosomes to be again disso- ciated during telophase. Zirkle (1928 and 1930), by developing specific fixatives, confirms the idea of direct transportation of nucleolar substance into the spireme and demonstrates the presence of this sub- stance in a skeleton form in the anaphase chromosomes. Heitz (1931la and 1931b) offers evidence to show that the nucleolus originates from the satellite chromosomes and only from them, and formulates a theory 284 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV that accounts for the symmetry of the nucleoli in number, size, position, and form in the telophase of the sister* (daughter) nuclei. Abele (1930) seems to believe that the nucleolar substance divides equally, as do the chromosomes, thus giving rise to nucleoli in sister nuclei that show symmetry in all respects and that “jeder Nukleolus hat seinen Partner in der anderen Tochterzelle.” As for the function of the nucleolus, Montgomery (1898) assumes that the nucleolar substance is connected with the nutritive processes of the cell and, as an assimilation organ, takes food from the cytoplasm. Meyer (Guillermond et al., 1933) believes that the nucleolus is com- posed of “ergastique” elements of the cell. It is commonly believed that the nucleolus is nothing but reserve food for the chromosomes which may build themselves from this supply. De Mol (1926), Zirkle (1928), and Fikry (1930) suggest that this body may break up into small particles and act as the carriers of hereditary stimuli from the genes through the cytoplasm. Further discussion of these theories, as well as others, will be found in the text. In addition to a study of the origin and behavior of the nucleolus an attempt has been made in the present study to clarify some of the points concerning the structural behavior of the chromosomes as a group throughout their developmental cycle and to throw light on the question of the supposedly continuous spireme and on the physical relationship often observed between chromosomes and nucleolus. Some of the conclusions are based on observations made from both living and fixed material. B. MATERIAL AND METHODS To make a study of this nature as comprehensive as possible, it was thought essential that the observations be made on various plants that are unlike taxonomically. The following plants were used: (a) Cal- lisia and Yucca (Monocotyledons); (b) Paeonia and Hepatica (Dico- tyledons); (c) Pinus ( Gymnosperm); (d) Polystichum (Pterido- phyte); and (e) a species of Spirogyra (Thallophyte). Some observa- tions were made from various other plants to see if there could be found any peculiarities not present in those mentioned above. The most in- tensive observations were made on Callisia and Paeonia. It was thought advisable to supplement the study of fixed material with that of living material as an aid in interpreting the phenomena observed. Living material was subjected to various toxic fluids, and *The use of the terms “sister cells’ and “sister nuclei” is suggested when cells or nuclei of the same “generation” are discussed, “daughter cells” and “daughter nuclei” being reserved for cases where derivation is implied. These terms are so used in this paper. 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 285 their effects on the protoplasm in general, and nucleolus in particular, were observed, and thus an interpretation of fixed objects could be arrived at. A fixative of 5% formalin and .5% chromic acid, in equal parts, was the principal killing and fixing solution used. This fixative seems to have been originally developed by Lewitzky (1931) and was recom- mended by Marshak (1931). The present author has made use of it with success. Its usefulness was proved by applying it to the delicate stigma hairs of Callisia, and it was found to cause less disturbance to the cell and its contents than any other reagents commonly used for cytological preparations. With this fixative the nucleolus takes a dark stain both with crystal violet-iodine and iron-alum haematoxylin. In the material fixed with this reagent and stained with crystal violet- iodine both chromosomes and nucleolus retain the stain remarkably well even after long washing in absolute alcohol. For the technical procedure with this fixative the reader is referred to my paper (Der- men, 1932). In addition to this method of preparation the Ehrlich- Biondi differential stain recommended by Van Camp (1924) was used. This is made up of Orange G, acid fuchsin, and methyl green. (For full details the reader is referred to Lee’s Vade Mecum, p. 177, 1928). Blotting paper tests of this stain described in this book were found essential and, accordingly, the proportion of reagents could be varied to give the proper reaction. C. MORPHOLOGY OF CHROMOSOMES IN RELATION TO THE NUCLEOLUS In order to make clear the nature of the relationship which may exist between chromosomes and nucleolus, it was considered essential that first a thorough analysis of the chromosome morphology should be made. For this reason this phase of the problem was undertaken. I. SrrucTURAL CONTINUITY OF CHROMOSOMES IN CYCLE. When the Lewitzky fixative is used, at no period of development are the chromosomes found aggregating into an unrecognizable mass. They always retain their linearity, and at prophase are found within the nuclear membrane as units grouped in the same relative positions which they occupied during anaphase. It seems that the individual chromo- somes retain their particular forms (Figs. 48, 49, and 2) as they move to their respective poles during division, bent at their polar constriction points. The prophase chromosomes (Fig. 2) differ from those at ana- phase (Fig. 49) only in that the former are more opened up to fill the space inside the nuclear membrane, while the latter are forced together somewhat into a more compact form in moving to the poles. This 286 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV compactness is extreme at telophase (Figs. 6, 10, 11, 16,and 17). From this point on, as soon as the new nuclear membrane is formed, there take place reverse morphological changes. A new nuclear membrane encloses the closely grouped telophase chromosomes; then follows the expansion of this membrane, due possibly to expansion of the tightly pressed chromosomes which start to develop into longer and _ finer threads, to the increase of nuclear sap, and to the development of nucleoli in number and size, resulting in an increase of nuclear volume as a whole. In comparing specifically Figs. 2 and 49 (in Fig. 49 only upper polar group should be considered), prophase and anaphase re- spectively, one can not fail to be impressed by the similarity of the orientation of the chromosomes. The difference is that in Fig. 2 the chromosomes are longer, spiral and zigzag in appearance, while in Fig. 49 they are short and straight, tending to crowd together as they are forced to the pole, and appear single in structure. It was found, how- ever, after critical study, that even at anaphase the chromosomes possess a dual structure. As in prophase (Fig. 2) and early metaphase (Fig. 29—in this figure the region of primary constriction is illustrated as faithfully as possible), anaphase chromosomes are found split length- wise and even show twists in the chromatid threads, as illustrated in Figs. 7 and 30, and these twists become more pronounced in telophase, as shown in Fig. 8 The chromosome contours at anaphase, as illustrated in Figs. 49-51, seem significant, and the interpretation may be that these chromosomes are split and their chromatids twisted together, as shown in Figs. 7, 8, and 30. Similarly, premeiotic chromosomes of microsporocytes show similar contours, indicating that they are already split before micro- sporocyte chromosomes go through characteristic meiotic development. he split nature of the chromosomes at this stage has been confirmed by a number of reputable cytologists, such as Sharp (1929) and Kauf- man (1931), working with plants, and McClung (1928) and Robertson (1931), working with animals. Darlington (1932), however, still main- tains that duality of the chromosomes at this stage is due to optical illusion, although previously (1926) he reported seeing chromosomes split at anaphase. This anaphasic split, as maintained by Sharp and others, seems to take place during the late prophase and metaphase. Such a split was seen in the Callisia satellite chromosomes, as illustrated in Fig. 5. The chromosomes at this stage, besides being fully split into distinct halves in readiness for mitotic division, also seem to show in each chromatid a secondary split, part of which, at least, could be dis- tinguished if observed carefully. At the upper end split there were also observed chromomere-like structures in pairs. It may be remarked 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 287 here that only rarely can these features be made out clearly. Delicate fixation and staining are primary requisites; and even with the best technique it requires special attention and considerable care on the part of the investigator that this feature may not be interpreted as an optical illusion. Here may be considered the arrangement of the chromosomes on the metaphase plate in preparation for mitotic division. The metaphase chromosomes (Callisia, Fig. 48) have a characteristic orientation which is considered normal; they are arranged in circular fashion, polar (primary) constrictions pointed toward the middle of the plate, and homologous chromosomes tending frequently to show a secondary pair- ing. If homologues are not always adjacent, at least there seems to be a tendency for them to orient themselves in such a fashion as to facilitate synapsis during meiosis. There may be a strong affinity between homologues at their polar constriction points that will play some part in bringing them together and result in their close pairing at meiosis. Structurally the chromosomes of Callisia vary considerably. The total number, as reported by Sax (1932), is 12, of which 4 have approxi- mately median constriction (Figs. 48 and 49), and 8 have subterminal constriction. Two of the latter possess satellites. No secondary con- strictions are observed. It is not uncommon to find in the literature references to a granular stage of nucleus during the resting stage. A careful study of Figs. 18 and 33 perhaps will throw some light on the exact nature of the nucleus at this stage. Fig. 33 represents the telophase of somatic chromosomes from Paeonia. The two sister nuclei are viewed from the side and illus- trated more or less diagrammatically. At this stage, as was pointed out above, the chromosomes are found already split and showing fine threads, each one representing a chromatid. The appearance of granu- lation often observed at these stages seems to be due both to curvatures and to points of strongly stained areas (Fig. 18). In mitosis these chromosomes will further elongate and open up to develop finally into a structural make-up similar to that illustrated in Fig. 2. If this stage were viewed from a different angle, it would no doubt present an entirely different aspect. Therefore, if it is true that at prophase the chromo- somes hold positions comparable to those at anaphase, then any possi- bility of their joining end to end into a continuous spireme is inconceiv- able. This point is further confirmed by the work of Sax and Ander- son (1933) and others who show that chromosomes may interlock during meiosis, a situation which would be impossible if the chromo- somes were continuous. The conception of the continuity of the chromosomes in a spireme 288 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV has probably been derived from impressions of oblique and polar views of nuclei of stages illustrated in Fig. 1, and especially of early stages similar to those in Fig. 18. (Fig. 1 is of the same stage as Fig. 2, except that in the former the nucleolus is included.) Even in a stage as late as Fig. 1 one can easily be misled concerning the continuity of the chromosomes, but with care they can clearly be made out as separate units. When root-tip sections were crushed under a cover-glass to flat- ten the cells, it was apparent that in all stages chromosomes invariably show linearity and are never broken up into scattered granules. Mar- tens (1928), having made a similar study from living and fixed material, finds the nucleus is not granular in make-up but primarily reticulate and filamentous, and that there is no continuous spireme at the begin- ning of prophase; that ‘Le réseau interphasique n’est donc que l’ensemble des chromosomes télophasiques—a peine plus évolués—et donc les mailles filamenteuses sont reliées par d’autres filaments d’union,” and confirms “la persistance—ou mieux, la continuité morpho- logique et génétique—des structures chromosomiques d’une cinése a l'autre, au cours de l’interphase et du repos.” II. CHROMOSOME AND NUCLEOLAR RELATIONSHIP. It can be seen from the preceding description of chromosomes that the idea of a con- tinuous spireme is incompatible with the facts, that the chromosomes retain a characteristic arrangement inside the nuclear membrane, and that the individuality of the chromosomes is maintained; therefore a direct flow of nucleolar material, postulated by Van Camp and many others, from the end of a chromosome through the rest of the chromo- somes, becomes an impossibility. Figure 1 shows a large central nucleolus around which the chromo- somes are distributed. In this nucleus all the chromosomes are found, with one exception, free from any physical connection with the nucle- olus. At the upper side of the nucleolus there was noted the only probable connection at the bend of a chromosome with the nucleolus. I was not able to discover, at this stage, the presence of a pair of satel- lites characteristic of this species of Callisia and determine their rela- tionship with the nucleolus, and only once, after a long search, I found two satellite-like bodies, one on each side of the nucleolus, at a con- siderable distance from the chromosome ends viewed from the side of the nucleus in a position similar to the one illustrated in Fig. 2; how- ever, these bodies may have been extra-nucleolar bodies which are described below. Van Camp (1924), using the Ehrlich-Biondi differential stain, found an intermediate coloration on the portion of the chromosomes attached to the nucleolus and therefore concluded that there was a direct flow 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 289 of nucleolar material over the chromosomes. Zirkle (1931), by using a selective fixation method, confirms this conclusion and finds that the nucleolar material which has flowed “into” the chromosome threads was fixed at anaphase in a skeleton form. The present author has been unable to confirm the findings of these investigators. In all the cells he has examined there was no sign of intermediate staining between any portion of the chromosomes and nucleolus. The nucleolus at all stages was always bright red and the chromosomes pale blue when material was stained with the Ehrlich-Biondi solution. Fig. 3 is a diagrammatic drawing of a prophase stage from differentially stained material where no such intermediate coloration could be observed. On the other hand, if Zirkle’s contention is correct, then the anaphase chromosomes were expected to show some red granular framework. Fig. 6 is such an ana- phase stage; the chromosomes were of the same pale blue color as in the prophase in Fig. 3. When Paeonia, Pinus, Polystichum, and other species were subjected to similar treatment, the results were the same as those found in Callisia; the differentiation was always perfect. It is quite generally assumed that the nucleolus decreases in volume during the development of the chromosomes, and that it supplies ma- terial to the “emerging” chromosomes. The facts on hand, however, seem to indicate that, at least in the plants studied, the situation is not such as described above, and that there is no diminution in size of the nucleolus from the resting stage, when it reaches its maximum volume, until the end of prophase before the nuclear membrane disappears. This fact can be demonstrated clearly by studying meiotic stages of species in which only one nucleolus is generally found, derived, no doubt, from fusion of a number of small nucleoli to form a large one. In Callisia there are usually two nucleoli present in the earliest meiotic stage, but as a rule there is only one at later stages. When there are two at early leptotene stage (Fig. 19), they are always smaller than the single nucleoli in adjacent nuclei (Fig. 20). Even at these early stages there are very few cells with double nucleoli, and these unquestionably later result in one nucleolus by fusion. Figs. 21-24 represent the zygo- tene stage, where the chromosomes are considerably thicker than in Fig. 20 and Fig. 25 the satellite chromosome pair at late diakinesis just before the nuclear membrane disappears, which, in this species, always holds the nucleolus at its satellite end, as reported by Sax (1932). In comparing the sizes of nucleoli in Figs. 20 and 25 and comparing both with the inner nucleolus of Fig. 18, which represents a cell from the earliest stage of meiotic development resulting from the division of the archespores, one is impressed by the constancy in size of the nucleoli all through these developmental stages. Similarly, Fig. 57 290 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV shows the diagrammatic representation of the volumes of the two nucleoli from Hepatica, a sketched from a very early leptotene, and b from the diplotene stage, showing the same constancy in size of the nucleolus as in Callisia. The assumption of transference of nucleolar material to the chromo- somes seems to be based on the fact that at later stages the chromo- somes take up stain more readily than at earlier stages, a fact which has led many investigators to believe that there is a correlation between the stainability of chromosomes at later stages and entire disappearance of the nucleolus by the time of chromosome division. In Yucca, illustrated in Figs. 55 and 56, it is shown that there is hardly any sign of decrease in volume of the nucleolus which is left out in the cytoplasm during meiosis. In general, it appears that the nucleolus reaches a maximum size at a very early stage in the nucleus soon after a nuclear membrane encloses the chromosomes; the size of the nucleolus remains constant all through further development of the nucleus to the very end of pro- phase, very soon after the nuclear membrane disappears; then the nucleolus generally disappears, the rate depending upon the species of plant, as will be indicated below. Belar (1928) seems to be justified in stating that there is not suffi- cient ground for assuming that the nucleolus has any role in the process of building up the chromosomes by increasing their mass. It was shown that when the differential stain was applied at all stages, the chromo- somes took a pale blue color, while the nucleolus stained bright red. On the other hand, when crystal violet-iodine, or even iron-alum haema- toxylin, is used, the chromosomes throughout do not show the same in- tensity of coloration as when Ehrlich-Biondi stain is used. Therefore this discrepancy of staining intensity of the chromosomes does not seem to be due to their being less chromatic at earlier stages, and more so at later stages, but rather to their “thinness” or “thickness,” depending upon their stage of development. Moreover, the stainability and the degree of retaining of the stain of the fine chromosome threads may be altered with different fixatives and stains. III. SATELLITES AND SECONDARY CoNsTRICTIONS. Belar (1928) states that there is no basis for assuming that there is any connection (physiological) between satellites and nucleoli, while, on the other hand, Heitz (193la and 1931b) has given great importance to these bodies. Heitz’s theory is that nucleolar bodies originate on and around the achromatic threads behind the satellites and at secondary constric- tion areas, and that the number of nucleoli at late telophase is the same as these satellite and constricted chromosomes. De Mol (1927) was perhaps the first to notice a correlation in the number of satellites and 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 291 nucleoli. In the present paper is reported a complete study of the satellite and constriction situation in Callisia, Paeonia, and Pinus, in order to have a basis for analyzing the relationship of the nucleolus to these features. The somatic chromosomes of Callisia, Paeonia, and Pinus were sub- jected to a detailed study to determine the presence and number of satellites and secondary constrictions. Pinus chromosomes were found to be devoid of typical globular satellites, but instead there were pres- ent five pairs of chromosomes with secondary constrictions which varied as to their position and length of achromatic threads, though appar- ently constant for each chromosome so characterized. Fig. 54 illus- trates Pinus Strobus chromosomes at anaphase (only part of the 12 pairs of chromosomes in each half is shown), showing some of these con- stricted areas on the chromosomes. Callisia (Fig. 48) has a pair of small satellites attached to relatively long threads which were hardly noticeable in this figure but are generally of good size, as shown in Fig. 5. Four plants of Paeonia were studied, two of which were of the species P. suffruticosa and two of different species,—namely, P. Dela- vayit alba and P. Woodwardii. Fig. 50 (a and 6), drawn from two adjacent sections, represents an anaphase from P. suffruticosa which was cut into two by the microtome knife. The exact number of the ‘satel- lites in this species was determined from this, as well as from other cells, and was found to be four, the position of which can be clearly seen. Two satellites are present on the short arms of a pair subtermi- nally constricted, and the other two on the short arms of a submedianly constricted pair. In Fig. 50b in the lower polar group one satellite was distinctly split, indicating a behavior similar to chromosomes at this stage discussed earlier. Here it should be emphasized that the satel- lites are normal components of some chromosomes and form their char- acteristic features with permanent attachment and are not free bodies picked up by chromosomes from the surface of the nucleolus, as was suggested by Navashin (1927). The Paeonia species studied had 5 pairs of chromosomes. They varied somewhat in size of chromosomes and satellites and in number of satellites. Fig. 50 represents a young seedling of P. suffruticosa from the Arnold Arboretum, while Fig. 51 represents a plant of the same species that has been growing at the side of the Bussey greenhouse. When slides made from these two plants were compared, a constant difference in number of nucleoli was found. The highest number reached in the greenhouse plant (Fig. 51) was always less than that of the Arboretum plant (Fig. 50). On the basis of Heitz’s theory it was suspected that this discrepancy in the number of nucleoli in two differ- 292 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV ent plants of the same species might be due to a difference in the number of satellites. To test this idea many cells at anaphase and metaphase were crushed by pressing on the cover glass and forcing the chromosomes to spread apart to make their satellited feature more obvious. Strikingly enough, the result was that actually the number of satellites of the plant near the greenhouse was less than in the plant of the Arboretum, three and four respectively. Fig. 51 represents an anaphase stage with three satellite chromosomes at each pole pressed flat and the chromosomes forced to spread. There was one pair of chromosomes submedianly constricted with very small satellites com- pared with a similar pair referred to in Fig. SO, and only one of the ‘chromosomes with a subterminal constriction possessed a satellite cor- responding in size to the ones illustrated in Fig. 50. Figure 52 is a metaphase stage of Paconia W oodwardii chromosomes. This plate also was crushed and flattened. The division at the polar constriction areas was quite distinct. There were six satellites, two more than in P. suffruticosa (Fig. 50), these two additional ones being on the shorter arms of a near-medianly constricted pair of chromosomes, while the other two pairs were on the same type of chromosomes as in P. suffruticosa. The satellites on these sub- and near-medianly con- stricted chromosomes were as large as the satellites of the submedianly constricted chromosomes of Fig. 50, while the ones on the subterminally constricted chromosomes were very minute and close to the end of the chromosomes and were determined only after considerable effort. Fig. 53 is from P. Delavayi alba, with one subterminally constricted chromo- some being shown which possessed a large satellite quite comparable to the ones in Figs. 50 and 51, in contrast to the small ones in Fig. 52, varying only in the length of the achromatic thread, which was some- what shorter. In other details the chromosomes of this species were comparable to those of P. Woodwardii (Fig. 52). On the basis of polar constriction the chromosomes of all Paeonia species reported here may be classified thus: one pair subterminal, one pair submedian, and three pairs (in varying degree) near-median. D. ORIGIN AND BEHAVIOR OF THE NUCLEOLUS Considerable space has been given to consideration of two important works by Van Camp (1924) and Heitz (193la and 1931b) concerning the origin of the nucleolus in plants. Van Camp, applying Ehrlich- Biondi stain, was able to show that the nucleoli originate from the chromosomes at telophase in the form of small globules which later, by fusion, collect into one large nucleolus, while Heitz limits the origin of the nucleoli to the satellite chromosomes, specifically on and around 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 293 the achromatic threads that connect either a satellite or a constricted arm with its chromosome. These two findings were subjected to an intensive study and their merits evaluated. I. THe NucLeotus 1n Somatic CELts. The origin of the nucle- olus is best studied in longitudinal root-tip sections of plants with large chromosomes. Fig. 9 represents two sister cells from Callisia root-tip sections. Here are shown, diagrammatically, small globules of nucleolus that were differentially stained red in contrast to the pale blue color of the chromosomes which are not shown. These globules collect, un- doubtedly by fusion, into two and finally one body (Figs. 10 and 11) and, as more globules are produced and added together, the nucleolus grows very rapidly and becomes constant in size during further develop- ment of the chromosomes. In general, the number of nucleoli in Callisia is one or two, and quite rarely three, as illustrated in Fig. 12. All these three nucleoli were homogeneously colored bright red by the Ehrlich-Biondi stain, indicating that they were of the same nature. Whenever it was possible to distinguish the outlines of the chromo- somes differentially stained, it was observed that these globules were on the surface region, around and between the split chromosomes. F ig. 31 is from Paeonia suffruticosa (greenhouse plant). The granules were found not only where the chromosomes are more compact, but also on the arms of the chromosomes that lay outside the region of the compact mass. Here it can be clearly shown that the origin of nucleolar sub- stance is not associated with the achromatic region of the satellite. Otherwise, assuming that these exposed arms may represent satellite arms, there should have been one granule to an arm, which is not the case. Fig. 32 is a diagrammatic representation of a similar stage, drawn on a larger scale to bring out the approximate number of these globules. Fig. 37 is from Pinus and represents the same very high number of small nucleolar globules. In Polystichum was found the same situation as in others; however, the higher number of chromo- somes in this species (2n = 60+) making it difficult to study the satellite or secondary constriction (if present), the representation of this situation was limited only to plants in which the number of satel- lites and secondary constrictions was determined so as to have a sound basis for discussing satellite and nucleolar relationship. Because of the same difficulty, a species of Arisaema (A. triphyllum, collected near Pepperell, Mass.) with 2n = 60+ chromosomes is not illustrated, but this also showed the same numerous small globular nucleoli at a stage similar to that illustrated for Paeonia and Pinus in F igs. 32 and 37 respectively. 294 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XIV In Fig. 29 is shown a part of a satellite chromosome from an early metaphase plate of Paeonia suffruticosa (Arboretum plant) to illustrate the supposedly true relationship between a satellite and the nucleolus. This small nucleolus is, perhaps, a remnant of a larger nucleolus that has not completely disappeared in the cytoplasm. It can be seen not only that this nucleolar body does not surround the achromatic thread— contrary to Heitz’s theory—but also that it lies away from the thread and is found very near the satellite (there may actually be an attach- ment between them). Both McClintock (1931) and Burnham (1932) have shown very clearly in Zea Mays that the nucleolus lies attached at the side of the achromatic thread of the satellite and not around it. In Callisia (Fig. 22, meiotic stage) and Paeonia (Fig. 29) this associa- tion seems to be between the satellite and nucleolus instead of between the achromatic thread and nucleolus. A behavior of extrusion of the nucleolus into the cytoplasm in micro- sporogenous cells will be more fully described below. In Pinus somatic cells not a trace of the nucleolus was found at metaphase, as was also reported by Zirkle (1931); in Paeonia very rarely one or two small pieces were found just before mitosis (Fig. 29); in Callisia often there were found one or two pieces just before (Fig. 4) and after (Figs. 16 and 17) mitosis. A similar phenomenon was observed in Polystichum root-tip sections. Fig. 45 is a diagrammatic representation of the meta- phase from side view. Outside the division sphere there is a nucleolar body that is a remnant from an earlier division stage. Similarly, in Fig. 46, is shown a large nucleolus outside the nucleus which is in pro- phase stage. Fig. 47 shows a nucleus with a typical nucleolus from an active region. The large nucleolus in the nucleus shows a_ bag-like feature as if containing a number of free bodies inside a bag, while the ones outside in the cytoplasm (Figs. 45 and 46) are uniformly round and homogeneous. Besides the large nucleolus shown in Fig. 46 there are a number of small nucleolar particles, staining red like the large ones, which may aggregate into various forms, as shown in Fig. 45, and finally be extruded into the cytoplasm, where they may further coalesce into a more compact form (Figs. 45 and 46). This peculiar form of nucleoli in such small granules may be comparable to the so-called amphi-nucleoli described by Wilson (1925). No speculations are jus- tified here as to either their analogy or homology, since the present author has not made a study of this situation in animals. However, in Polystichum these particles show the same behavior as is characteristic of larger nucleolar bodies, indicating that they are of the same nature as the large ones. Extrusion of nucleolar material, in part or in whole, is reported by many investigators, among whom may be mentioned 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 295 Davis (1903) and Wakayama (1930), working with fungi, Yamaha and Sinoto (1925) on some thirty species and forms of phanerogamic plants, while in some Protista (Belar, 1928) the nucleolus does not disappear at all during mitosis but divides into two which are included in the sister nuclei. In fixed material of Callisia, as well as in living, in the stigma hair nucleus there often were found around the region of the nucleolus some small extranucleolar bodies (Fig. 13). When subjected to differential stain, these bodies differed in their coloring when compared with the bright red of the nucleolus,.their color being of a darker shade of red. According to Heitz (1928), these bodies are considered to be chromo- some fragments, but recently this assumption was refuted by Scheuber (1932), who finds that these bodies vary in number, size, and shape and show differences in staining when compared with the nucleolus. In Callisia the situation was found to be identical with that reported by Scheuber. When fixed material is studied carefully, one will generally find a clear area around the nucleolus, no matter how small it may be, while around these bodies no such clear area is observable, and their staining is intermediate, though more like nucleolus than chromosomes. Similar bodies were observed in the microspore cells of Callisia. Fig. 27 represents a late prophase stage with six chromosomes, a large nucle- olus in the middle, and a body at the side. This body can not be con- fused with a satellite, since the four short chromosomes point away from it and also’ because of its characteristic stainability referred to before. Fig. 28 is the drawing of a metaphase with such a body at the distal end of the satellite chromosome. Again it should be stated here that this body stained characteristically and was no doubt an extranucleolar body. It seems to have a remarkable stability in size and does not disappear as the nucleolus would. Identical bodies were observed during anaphase and telophase where it could be seen that they are left outside the new nucleus in the cytoplasm, there perhaps to disintegrate eventually. Such bodies were also found in root-tip cells of Callisia. II. THe Nucreotus 1n MicrosporocENnous CELLS. Undoubtedly the process of nucleolar origin in the telophase of archespores is the Same as was described for the somatic cells of root-tip material. No favorable stage having been found, this point could not be illustrated. Figures 14-17 are from very young microsporogenous tissue of Callisia. At this stage quite commonly were found large extranuclear nucle- olar bodies. In Fig. 15 there are two sister cells that have developed at a parallel rate, both being at late prophase stage. At the out- side of each nucleus there is a large nucleolar body in an opposite 296 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XIV position to the corresponding one in the sisttr cell. Undoubtedly these two extranuclear nucleoli are remnants from a like cell shown in Fig. 14. Figs. 16 and 17 show the behavior of a nucleolus that has not been dissolved completely during the earlier stages of mitosis and is forced into the cytoplasm where it may remain for a considerable time during later stages of meiotic development, since these cells are destined to become microsporocytes. Fig. 18 represents the earliest stage of a microsporocyte after emerging from a mitotic division similar to the one shown in Figs. 16 and 17. In this, besides a large nucleolus being present among chromatic threads, one medium-sized nucleolus is found outside the nucleus in the cytoplasm probably resulting from a process illustrated in Figs. 15-17. In Fig. 23, which represents a much later stage, a nucleolus is found outside very near to the main nucleolus inside the nuclear membrane to which two chromosomes seem to be attached, while there is still another extra round body within the nucleus which is free from any such connection. Inside the cell represented in Fig. 21 there are present two such free round bodies, besides the larger nucle- olus with a chromosome attached to it, that also show no connection with any chromosomes. In Yucca microspore mother cells there are either two medium-sized or one large nucleolus present inside the nu- cleus, as shown in Fig. 55. This figure is drawn at the nucleolus level, and not all the chromosomes are included in it. It must be emphasized here that the nucleolus is free from any chromosome attachment. The nucleolus in Yucca retains its size till the time of disappearance of the nuclear membrane, then it is extruded away from the division plane, sometimes one nucleolus to one side and another to the other side of the division plane, or both to one side individually, or two closely to- gether, as shown in Fig. 56, there perhaps to disintegrate gradually. The nucleolar extrusion referred to in Callisia and Yucca may be analo- gous to a similar but seemingly more common behavior in Polystichum. As is described above, the nucleoli that have not completely dissolved after the disappearance of the nuclear membrane are not retained in the newly-forming cells but are forced outside; so this behavior may be considered the primary rule but perhaps not the only rule. As is the case in some Protista described by Belar (1928), Christoff and Gent- scheff (1932) have assumed that the nucleolus divides during mitosis, and the pieces of these nucleoli are retained in the newly-forming sister nuclei. So far most of the findings by many investigators and the present author indicate the contrary. However, in the present study there was encountered a difficulty which may be explained if it is assumed that rarely some of these pieces of nucleoli may be enclosed among the dividing chromosomes and retained in some of the nuclei. 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 297 As was pointed out, the nucleolar pieces shown outside the nucleus in Figs. 15 and 18 took a bright red color similar to the ones inside, while the free nucleoli inside the nucleus in Fig. 21, one outside and one inside the nucleus of Fig. 23, stained somewhat differently. If these free bodies are actually carried over from an earlier cell, the age of these nucleolus-like bodies may account for their reaction toward the stain, thus staining somewhat differently than the true nucleolus. They may remain there as inert particles. The above is presented as a mere suggestion, since the presence of these extra bodies is difficult to explain in the absence of experimental data. These bodies in the microsporo- cytes are of rare occurrence. Another feature in Callisia was the presence of a small bud-like pro- tuberance on the nucleolus, as shown in Fig. 22._ A phenomenon similar to this is reported by Gates and Latter (1927), Maeda (1930), and Selim (1930). Gates and Latter find at the points of chromosome attachment to the nucleolus dark staining bodies, varying in number, size, and distribution. Maeda similarly finds small nucleolar bodies on the ‘mother nucleolus” and sometimes free from it in the nuclear cavity. Selim, studying meiosis in rice, finds the nucleolus budding off a “secondary nucleolus” at late diakinesis. He supports the view that the secondary nucleolus contributes material to the chromosomes, while the primary may contribute to the spindle. Similarly, as in mitotic sister nuclei, in the sister nuclei of first meiotic division there were observed a number of small nucleolar glob- ules (Fig. 26) which were found clearly oriented along the chromosome threads and not in the clear space of the nucleoplasm. Owing to some technical difficulties arising from poor fixation, no critical drawings of meiotic stages of Paeonia and Pinus could be presented. Only in an outline fashion the number and depth of nucleoli in Pinus (P. Thun- bergi) from earliest to latest stages of meiosis are given (Figs. 39-44) to show the chromosome and nucleolus relationship. However, no radical differences from Callisia were observed in these plants. In Paeonia suffruticosa (greenhouse plant—other members of this genus were too young for the study of the meiotic phase of the problem) the nucleoli in the earliest stage numbered from six to one, while at later stages the number decreased to three to one, disappearing with the dis- appearance of the nuclear membrane. In Pinus at early leptotene (Fig. 39) and during early diplotene (Fig. 40), as many as nine nucleoli were found, the number decreasing as the development of the nucleus advances, but the volume of total nucleoli not showing any decrease. Fig. 41 is from late diplotene, while Fig. 42 is from late diakinesis stage when the nuclear membrane is disappearing or newly disappeared prior 298 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV to the chromosome arrangement in a metaphase plate. Fig. 43 is one of the sister cells after the first reduction division, while Fig. 44 is after homotypic division before the nuclei go into resting stage. In these two figures the number of nucleoli reached as high as nine to twelve. III. Retationsuip or NucCLEOLUS TO SATELLITES AND SECONDARY Constrictions. Heitz (1931a and 1931b) has made an extensive study of satellite and secondary constriction number in Vicia and other plants upon which he has based a definite theory concerning the origin of the nucleolus. His conclusion is that the number of nucleoli in any species in telophase sister nuclei should correspond to the satellites and secondary constrictions there present, for he assumes and gives evidence, as was mentioned above, that each nucleolus at telophase originates on and around the achromatic thread of each satellite or secondary con- striction in the form of a collar. This correlation could be more readily proved if there were never found more nucleoli than satellites and sec- ondary constrictions characteristic of a species. However, it is sig- nificant that the number of nucleoli in the greenhouse plant of Paeonia suffruticosa never reached the highest number found in the Arboretum plant of the same species. In the former the highest number was found to be five (Fig. 34), while in the latter seven, which have respectively three and four satellites. Both P. Woodwardii and P. Delavayi alba have six satellites; in the former the number of nucleoli reached as high as nine (Fig. 35), while in the latter quite rarely to as high as eleven (Fig. 36). Similarly, in Callisia quite rarely there were found three nucleoli (Fig. 12) instead of two to correspond with the number of satellites: in Pinus (P. Strobus root-tip) the number reached as high as fourteen (Fig. 38) where there should have been only ten if there are ten achromatic areas in the twelve pairs of chromosomes. There can be no doubt that a discrepancy to Heitz’s expectation exists, the cause of which should be looked for elsewhere. However, the above data clearly indicate that whenever there is an increase of satellites or sec- ondary constrictions, there is a similar increase in the number of nucleoli. The analysis of the situation is given below. IV. NucLEOLAR SYMMETRY IN SISTER NucLe!. Before I had occa- sion to review the literature on the nucleolus, the perfect symmetry in some sister nuclei in root-tip sections, as illustrated in Fig. 33, had impressed me, and I was led to assume that this symmetry was due to the spatial relationship existing between the chromosomes in the sister nuclei. Therefore the mirror image of some nuclei in having nucleoli similar in number, size, and position was thought to be primarily due to the proportionality of the spaces between the chromosomes of the 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 299 sister nuclei; hence the symmetry of nucleoli in these cells. However, further work on this phase made it necessary to alter somewhat this original assumption. Fig. 33 is a somewhat diagrammatic representation of the symmetry between four nucleoli in two sister nuclei. The symmetry is not only in number but also in size and position of the nucleoli. This drawing is from a root-tip section of Paeonia suffruticosa with the four satellites (as illustrated in Fig. 50). According to Heitz, the size of a nucleolus depends on the length of the satellite achromatic thread and its position on that of the satellite. In Fig. 33 the four nucleoli show a symmetry of number, size, and position of nucleoli but not a symmetry of the type expected according to Heitz, since in this plant two satellites are prox- imal and two are distal to the division poles (Fig. 50), while here one nucleolus is proximal and three are distal. There is also a size differ- ence; the polar nucleolus in each nucleus seems to be more than the total volume of the other three nucleoli. Heitz would explain this on the basis that the increase of a nucleolus is proportional to the space around it; however, there were found many cases, as illustrated in Fig. 34, where there are small nucleoli which seem to have ample space to draw material and grow in size if Heitz’s assumption is correct, but nevertheless they have remained small. De Semet (1913) has proposed a genetic relationship between certain chromosomes and the nucleoli on the theory that nucleoli originate from certain chromosomes, and owing to this relationship he assumed the existence of a symmetrical relationship between sister nuclei. Yeates (1925) and Sprumont (1928) have supported de Semet’s view, while Abele (1930) has explained this symmetry by assuming that “Die Nukleolar-substanz wird bei der Karyokinese in gleichen Mengen auf beide Tochterzellen verteilt, infolgedessen sind die Nukleolen oder Nukleolensatz beider Tochterkerne gleich gross.” It was pointed out earlier that the number of nucleoli during the early stages of nuclear development corresponded quite closely to the number of satellites present for the species in question, although quite often there were found more nucleoli than were expected according to Heitz’s theory. However, since there seems to be some correlation between the number of satellites and the number of nucleoli, and also since during meiosis the nucleolus is constantly attached to a satellite chromosome pair in Callisia (Figs. 21-25), (Sax, 1932), in Ranunculus (Sorokin, 1929), and in Zea Mays (McClintock, 1931 and Burnham, 1932), therefore the symmetry often found between iwo sister nuclei in the number, size, and position of nucleoli is suggested to be primarily due to a physical relationship between a satellite and a nucleolus, and 300 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV independently to corresponding spaces between the same set of chromo- somes in the sister nuclei, where nucleolar globules “exuded” from the adjacent chromosome surfaces fill in. The decrease in number and increase in size of the nucleoli may come about from the increase in these interchromosomal spaces during the thinning of the chromosome threads and increase in nuclear volume as a whole, allowing these spaces to fuse into fewer and larger spaces, followed closely by the fusion of smaller nucleoli into fewer and larger ones. V. Tur NuMBER AND SIZE OF NUCLEOLI IN POLYPLOID RACES. De Mol (1926 and 1928) has made some assumptions concerning the size and number of the nucleoli in the di-, tri- and tetraploid varieties of Hyacinths and has concluded that the size of “complex nucleoli” and the number of “simple nucleoli” are proportional to the number of the chromosomes. He believes that after nucleolar globules have been formed during telophase, they fuse into a “complex nucleolus” which later fragments into two, three, or four “simple nucleoli,” depending upon whether the variety is diploid, triploid, or tetraploid. The present author has observed the process of fusion between two nucleoli at pro- phase in living tissue but has never observed fragmentation. If frag- mentation ever occurs, certainly it must be in very rare circumstances instead of being the rule. The abundance of higher number of nucleoli in early stages of nuclear development and lower number in later stages must mean one thing only, that this decrease of number and increase in volume of nucleoli come about through fusion of higher number and small sizes into lower number and larger sizes. Here may be reported the results of my own observations from diploid, triploid, tetraploid, and pentaploid Petunia. The largest single nucleoli from longitudinal sections of root-tips of six plants (Table 1) were measured in order to establish the size relationship of nucleoli in the polyploid series. The first two of these plants were diploids (7Lx and 7Lnc,_,), the third a_ triploid (14S,_.. 7Lne,)-1, the fourth and fifth tetraploids (14S,_., and 14L,_.,) and the last a pentaploid (Trip.4 14S,_.,). For the descrip- tion of these plants the reader is referred to the author’s paper on “Poly- ploidy in Petunia,” 1931. On the slide of some of these plants there were either two or three root-tips in serial sections; in that case ten measurements were taken from each root, one from each serial section. It must be remembered that on one section there are hundreds of cells and as many or more nucleoli; therefore, even if only one measure- ment is taken from a section, it is from one of the largest among hun- dreds of nucleoli. During the measurements it was found that there was 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 301 some constancy in the size of the largest nucleolus for all the sections of the same root; therefore, taking measurements from more sections of the same root-tip was not considered essential. Further, in order to make these measurements more proportionate and uniform, only the largest and circular appearing nucleoli were chosen to avoid confusion and error in the measurements. Below, in Table 1, are given the diametric measurements of nucleoli in microns. It is apparent from Table 1 that no significant conclusion can be arrived at as to the increase in volume of nucleoli from the increase of chromosomes of diploid to polyploid series, perhaps for the following reasons: (1) the number of measurements appears to be small and (2) the races tabulated here are of mixed nature. At present, unfortunately, there are no available plants to make this study more extensive; there- fore, of necessity, the above measurements had to be limited to the numbers presented in this table. Even though these numbers are per- haps small for an exact conclusion, it seems evident that in order to have a basis of comparison on this phase of the problem, one is required to choose plants of polyploid series that are of the same origin. Table 1 shows that even though the second diploid plant (7Lnc) is from a mutated bud from the first diploid (7L), the difference in their measure- ments is quite striking. The 4n plants and the 5n are derivatives from an entirely different race of diploid, while the 3n strain is from a cross of this 4n and the second 2n race. Perhaps in the case of 3n the small measurement in general may be explained by assuming that the second 2n plant had a diminution factor affecting at least the size of the nucle- olus. The two tetraploids and the pentaploid measured practically the same and were somewhat smaller than the first diploid. In the writer’s paper on Petunia is discussed the origin of these poly- ploid strains. From an ordinary diploid race the writer obtained a tetraploid plant, and from these two strains were obtained some triploid Strains, and finally a pentaploid strain was obtained by crossing a triploid with a tetraploid. From the old slides of these strains, cross sections, similar measurements of nucleoli were taken. The origin of all polyploid strains, as can be seen, is from a pure-breeding diploid plant; therefore, the measurements from these plants are expected to give more of the true picture concerning the differences in nucleolar volume (Table 2). In this case measurements were taken from ten serial sections of one root-tip from each plant. Table 2 seems to indicate an increase in diameter of a half a micron from 2n to 3n, one micron from 3n to 4n, and a decrease of one micron from 4n to 5n. There seems to be no difference between the 3n and Sn strains. From these measurements, as from those in Table 1, it may be Table 1 20€ WOLIMOdTUV GIONUY AHL AO IVNYNOL Petunia series Nucleoli in microns of ten sections from each root—tip train Pedigree | Root- : ) tips [4.0] 4.24 4.4 4.4 4.4 5.d 5. 5.4. 5.@ 5.8] 6.0/6.2/6.4/6.6)] 6.8] 7.0 & 2n TL lst 2] 8 2d 3] 2] 5 = te 34a 427,237.25) 24 23 21 ¢ w 7Ln ™ “2-8 | Ist 1{ if af 4] ef 2 + 2d 5] 2] 1 2 w 4 Sd | 1] 6] 3 < 5n (14S 7 2-28 | 1st 2| 2] 4 1 1 7Lnce9)-1 - 2d 1} 21} 1] 21 4] 2 5 | rs 3a 1] 3S] 1} 2 4 ~ 4n 1485_99 lst 27-7) 2 0 = 2d 1] 1] 6] 2 6 4n 14L2_21 1st 2] 3] 3] 2 ~ ob 2d 1 1] 8 J Sn Trip.4 lst 2} 6] 1 uF xTet. 4 29] 94 4{ 5] 1 © a w AIX “10A] 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 303 difficult to conclude whether or not there is any increase or decrease in the volume of the nucleolus along with the increase of chromosome number from diploid to polyploid series, according to de Mol’s theory. However, there may be a tendency toward slight increase in nucleolar volume, due to an increase in number of chromosomes, and also some other factors may be involved in influencing the volume, as in the case of the difference between the two diploid strains in Table 1 and the equality of volume in the triploid and pentaploid in Table 2. Table I Nucleoli measurements in microns of ten sections of on fe train root—tip from each Petunia strain 5.2/5.4 |5.6|5.8 16.0/6.2 (6.4/6.6]6.8/7.0| 7.2|7.4 [7.6 17.8 | Range | Median 2n 1/2 );443 5.2-5.8 5.5 3n 201s | healed 5.8-6.4 6.1 4n 1] 3 a ae ee 2 | 2 [6.2-7.8 7.0 5n 4);4]2 5.8-6.2 6.0 E. THE BEHAVIOR OF THE NUCLEOLUS IN LIVING TISSUE An intensive observation was made on Callisia stigma hairs, kept alive either in tap water or 5% sucrose solution, in order to have a basis for determining how the fixed material differed from, or how closely it resembled, the living with regard to the behavior of the nucleolus and its relationship with the chromosomes. Entire pistils or styles with stigma hairs were placed in a drop of water on a slide and covered with a cover glass, and the material was studied under low or high power magnification with oil-immersion objective. These hairs, a few scores in number to a stigma, are fine and non-septate cells reaching a length of three millimeters, and were found to be very favorable material. I, THe AREA AROUND THE NucLEOLUS. The nuclei and nucleoli in the hairs are considerably larger than those in some other parts of the somatic tissue, such as the style cells, where they may be one-fifth as large. The nucleoli in these cells may measure 4 to 10 microns in diameter, small in the very young and large in the hairs just before the bud opens. With a slow change of form of the nucleus, the nucleolus also changes its position and shape, perhaps owing to pressure exerted by the chromosome threads of the ever moving nucleus in the streaming cytoplasm. There was never found a clear area (separation area) be- tween the nuclear reticulum and the nucleolus. This area appeared whenever these hairs were subjected to a disturbance by introducing a 304 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XIV toxic fluid under the cover glass, resembling the area so commonly observed around the nucleolus, inside the nucleus, of fixed material. Based on this observation the conclusion may be drawn that the clear area around the nucleolus in the fixed material is due to fixation and therefore is an artifact. Fig. 13 represents, in an outline, a nucleus in which is shown the clear area around a vacuolated nucleolus. As a result of any slight injury, physical or chemical, the nucleolus, as if sensitized, immediately pulls itself to one side, somewhat shrunken, close to a globular extranucleolar body (if any is present). Simultaneously, similar shrinkage takes place in the nucleus, enlarging this clear area. The creation of this area seems to be primarily due to sensitizing of the protoplasm (used in a broad term) and secondly, if at all, to plas- molysis. This assumption is substantiated by the fact that, in some cases, when the hairs were treated with Lewitzky’s fixative, a clear area was created immediately but very soon disappeared, and no appreciable area was left between the nucleolus and nuclear mass, while when a stronger reagent was used, such as Bouin’s or Flemming’s, the effect was permanent. Il. ExTRANUCLEOLAR Bopres. In the earlier part of this paper reference was made to the nature of the extranucleolar bodies in somatic cells in connection with their difference in stainability and their possible homology with extranucleolar bodies in meiotic cells. In the stigma hairs of Callisia, as illustrated in Fig. 13, these bodies are generally found close to the nucleolus. Whenever the hairs were subjected to toxic fluids, the nucleolus seemed always to pull itself away from the nuclear mass and to remain attached to one of these bodies like a bal- loon attached to a mast. No adequate explanation can be given at present as to the origin of these bodies, small in size and varying in number from one to five or more. Because of their being near the nucleolar region, they may be considered some sort of extruded material from the nucleolus, or perhaps pieces of nucleolar material that were carried and held among dividing chromosomes and were included inside the nucleus and retained there along with the nucleolus as inert pieces of nucleoli. The present author is inclined to consider the last assump- tion more likely than the first. In general there was only one body of this kind present in the nucleus, and often there were nuclei without it. It was said that they were always near the nucleolus, but never were any of these found fusing with the nucleolus, which may be considered indicative of their difference, although the nature of this difference is not known. III]. CuHromosoMe aND NuCLEOLAR RELATIONSHIP. One of the principal reasons for studying living tissues was to determine what rela- 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 305 tionship exists between the chromosomes and nucleolus. In the style where cells could be found in late prophase stages, the association between the chromosomes and nucleolus was identical with that illus- trated in Fig. 1. Obviously the chromosomes are separate units, and many at least are free from any connection with the nucleolus. In the stigma hairs the nucleus appeared very finely granular and compact but was undoubtedly made up of fine threads which never thicken, perhaps because of the lack of cell division in these hairs. The hairs themselves grow in length as the pistil matures and dry up and die after they have played their part (if they have any) during anthesis. With the growth of the hairs there is an increase both in nucleus and nucleolus and a slight decrease (at least in nucleolus) when the hairs become exposed at anthesis.’ These hairs were treated with fixative to bring about a clear area around the nucleolus to see if there was any close connection between these fine threads and the nucleolus. The only apparent con- nection which could be found was that between the extranucleolar body and the nucleolus; and sometimes, in addition to some achromatic threads, there were very rarely found threads crossing this clear area, apparently being dragged by the nucleolus, owing to the separation of the nuclear mass and the nucleolus. Therefore it seems that even at early stages there is no complete direct association between all the chromosomes and the nucleolus and that a single attachment point between a chromosome and a nucleolus can not be considered of any physiological importance. IV. THe Nucieorar Vacuotes. Contrary to Fikry’s belief ( 1930) that vacuoles in the nucleoli are artifacts, the present author found that vacuolation was one of the most important features observed in living as well as in fixed material. At times the nucleoli in all the stigma hairs were found to be vacuolated; at other times only part of them were vacuolated; and at still other times there were stigmas with hairs none of which contained vacuolated nucleoli. It is apparent that vacuoles are not permanent features (in the strict sense of the word) of the nucleoli, but vacuolation may be considered a normal phenomenon, and a record I should like to mention that in the case of ist ainia the size of the ee is not constant all through the plant system. a crosporogenous tissue it is constant in all cells, at all st measurin eae s; in the stigma hairs it varies from 4 to 10 pr depending upon Oe stage of growth of kag hairs; in the tissue of style and stamen filaments it measures about 2 microns; hile in root-tips, in regions ere Oe is active division, the si y m ; e epidermal and root-cap cells t 2 microns. In all these cases there is a correlation between size of nucleolus and nucleu erences (2 S e diff here recorded may be due to difference of nutrition n, so that whenever the supply of food is less, there is a corresponding diminution in the size of both nucleus and nucleolus and vice versa. 306 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV vacuoles may appear and disappear normally. To check up this assumption, stigma hairs in water cultures were kept as long as they could be kept alive to see if non-vacuolated nucleoli would become vacuolated. For this experiment young pistils or styles with stigma hairs cut off from pistils were put in a drop of water on a slide and covered with a cover glass. They were first studied under the microscope with oil- immersion objective, the desired regions spotted, the hairs and nucle- olus in its surroundings sketched, and the nucleolus measured if necessary; then the slides were put away in a petri dish. At the bottom of the petri dish was placed some wet filter paper to create a moist chamber. This method enabled me to keep some of this material alive in tap water for six days. After many trials not a single unvacuolated nucleolus had become vacuolated. Some appeared to contain vacuoles in the form of small droplets, but this proved to be a sign of degenera- tion which was followed by the death of the cell. The general effect of this treatment was that in the living cells the nucleoli always showed considerable decrease in volume. One nucleolus was measured soon after the material was put on the slide. It measured 9 microns in diameter; at the end of the second day it had decreased to 6 microns, to 4 microns at the end of the third day, and to 1.8 microns at the end of the sixth day; and finally the cell died. Results of this kind would lead one to suppose that the nucleolus was used up as reserve food by the “starved” cell and hence its extreme decrease in size. How- ever, a similar decrease was noticed in the nucleus itself but not in the same proportion (no measurements of nuclei were taken). Analogous behavior is reported by Meyer (Guillermond et al., pp. 183-184, 1933), who believes that the nucleolus is made up of “ergas- tique” elements of the cell and that the volume of the nucleoli is essen- tially variable during the physiological state of the cell. The volume of the nucleolus increases when the cell has an abundant supply of nutritive material and diminishes when the cell is in a starving condi- tion. In the mesophyll of Galtonia candicans leaves Meyer finds that there is a diminution in the volume of the nucleolus when these leaves are etiolated and states: “Si, dans la feuille vivant dans des conditions normales, on exprime ce volume par l’unite, on constate qu’il s’est abaissé a 0,38 dans une feuille maintenue a l’obscurité pendant 36 jours et A 0,18 dans une feuille maintenue a l’obscurité pendant deux mois.” During the change of the albumen of Galtonia this author finds some interesting changes in nucleolar volume which are described thus: “dans les cellules de l’albumen jeune, les nucléoles mesurent 52,3 environ; leur volume s’accroit A mesure que l’albumen se charge de mateériaux 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 307 nourriciers, jusqu’au volume de 101,3; puis il diminue au moment de la formation des parois cellulaires, et surtout au moment de la germina- tion et de lutilisation de réserves, jusqu’a devenir insignifiant.” To supply food matter for the cells to prevent them from “starving,” some material was. kept in 5% sucrose solution in vials and on slides. In this solution it was possible to keep these hairs alive for eighteen days. At the end of this time, owing to the growth of yeast cells and molds of various kinds, the tissue as a whole was destroyed. The striking difference between water and sugar cultures was that while there was invariably a decrease of nucleolus in the water, there was hardly any appreciable decrease in the sugar solution, except that whenever the water under the cover glass had evaporated in part, the nucleoli were found rounded up and the flow of the cytoplasm slowed down. As soon as more water was added again, the cytonlasm increased its speed of flow, and the nucleoli took on ellipsoid or other shapes due probably to the activation of the nucleus. In one case one cell was recorded dead because there could not be observed any movement in the cytoplasm, but when water was added, it soon revived and showed cytoplasmic streaming. It was observed that the vacuolated nucleoli did not all lose the vacuole when they were kept in sucrose solution, except in a few cells, indicating that the tissue can live normally in the sucrose solution, for a while at least, while if kept in tap water, some radical change (superfi- cially Eye) seems to take place. During these experiments, how- éver few in number, no vacuolation, de novo, was observed even in the sugar solution. This was perhaps due to the somewhat abnormal con- dition in the sugar solution (if it is assumed that vacuolation is a normal phenomenon), since this solution can not be considered an ideal me- dium for tissue culture. Some measurements were taken of nucleoli, some of which were vacu- olated and some not. In all cases there was a decrease in volume in water cultures irrespective of the vacuole. Exact measurements were taken of two nucleoli, one with a vacuole and the other without. For convenience, let the vacuolated be No. 1, and the nonvacuolated No. 2. The diameter of No. 1 was 6.2 microns and that of its vacuole 5 mi- cronms. The diameter of No. 2 was also 6.2 microns. Overnight, sixteen and a half hours, the vacuole of No. 1 had disappeared, and the nucle- olus had lecrenesd in size to a diameter of 5.4 microns. The diameter of No. 2 also had decreased to 5.6 microns. The measurements after 26 hours were, for No. 1, 4 microns, and for No. 2, 4.2 microns. After some hours both cells were dead. As can be seen hon these measure- ments, the disappearance of a vacuole does not seem to have any appre- 308 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV ciable effect on the volume of the nucleolus. Therefore, it is perhaps safe to assume that the vacuolated region has been transformed into true nucleolar substance by some physiological phenomenon, the exact nature of which it is difficult to ascertain. This region appears to be of the same concentration as the cell sap. This assumption is based on the fact that in one case a small proplastid-like body was found inside this region and showed the same rate of Brownian movement as took place outside the nucleus. F. GENERAL DISCUSSION AND CONCLUSIONS I. CHromosome MorpHoLoGy AND CHROMOSOME-NUCLEOLAR RE- LATIONSHIP. Conclusive evidence is presented in the text to show that the chromosomes in somatic nuclei keep a form of linearity which remains practically constant from anaphase to the end of prophase; that there exists, in general, a permanence in position and structural form during these same stages. The facts seem to indicate that at no time of development do the chromosomes pass a period during which time they are in a granular state and thrown about in a haphazard way, later to reestablish themselves in threads, joining end to end, again to break apart into separate units as chromosomes. There is no doubt, among those familiar with karyokinetic problems, of the individuality of chromosomes in structure. The truth of this matter seems to have been substantiated by Rable as early as 1885 and by Boveri (1909) and, more recently, by Kagawa (1926), Belar (1929), Koerperich (1930), and others. Rau (1930) states that in Cyanotis cristata pollen-mother-cells no continuous spireme appears to be formed at any stage. It is a fact that generally, morphologically, the chromosomes of a basic set in a species differ considerably from each other: therefore the possibility of diverse chromosomes, such as described for Callisia and Paeonia, forming a continual spireme is in- conceivable. The conclusion may be that, normally, the chromosomes never join end to end. This point may be further emphasized by refer- ring to the work of Sax and Anderson (1933) and others who show that during meiosis some chromosomes may interlock, a phenomenon which could not take place if these chromosomes were parts of the same con- tinuous spireme thread. These facts indicate that the direct flow of nucleolar material into and through all the chromosomes, as suggested by some investigators, is an impossibility. Il. Oricrn AND DEVELOPMENT OF THE NucLeEoLus. It is shown that the nucleolus originates in the form of small globules on the sur- face of the chromosome threads during the late telophase stage, that 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 309 very soon there is a close grouping of chromosomes, and that a de- marcation between the chromosomes as a group and the outside cell-sap and cytoplasm is established. It is difficult to show if there is a definite nuclear membrane, as such, around the chromosome group at this period, but for practical purposes it may be assumed that such a mem- brane exists from the beginning of the late telophase till the time of metaphase plate formation. The method of origin of the nucleolus here reported was first described, I believe, by Van Camp (1924). The process of nucleolar development appears to be a surface phe- nomenon which can be explained by assuming that a chemical reaction takes place between the substance on the surface region of the chromo- somes and the nuclear sap. The nucleolus, then, may be considered a by-product of the chromosome matrix and nuclear sap and, in this sense, may be identified with the matrix, as was suggested by Marshak (1931), but with this difference, that it is the matrix which produces the nucle- olar substance by going into a chemical combination with some sub- stance in the nuclear sap. The difference is indicated by their reaction to the differential stain. From this it may be deduced that there exists a striking difference between the nuclear sap inside a nuclear membrane and the cell-sap outside this membrane, because of the fact that the nucleolus will originate and develop inside a nuclear membrane and float in the nuclear sap, while outside this membrane the nucleolus dis- appears and apparently is dissolved in the cell-sap. It appears, there- fore, that the cell-sap has a dissolving effect on the nucleolus, the rapidity of which may differ in different species; hence the lagging of nucleolar particles in some species and the lack of it in others which may be assumed to depend on the varying strength of reaction of cell- sap in different species. The phenomenon of lagging may also be explained on the basis that there exists a difference in rate of chromosome development during the late prophase and metaphase, varying in different species. For in- stance, it was noticed that the chromosomes in Callisia (somatic stage) are well developed by the time they are forced onto the metaphase plate, already well split, and ready to divide and move toward the division poles. In Paeonia and, to a lesser extent in Pinus, this development is not so far advanced as in Callisia before the nuclear membrane dis- appears; the chromosomes of the former species stay at metaphase con- siderably longer, while in the latter the separation of sister chromo- somes is much more rapid. It is suggested that into the phenomenon of lagging of nucleolar particles the time element may enter. Thus, in species where the chromosomes divide more rapidly at metaphase, as in Callisia, more nucleolar particles are observed during the metaphase 310 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV and anaphase; but rarely in Paeonia and never in Pinus, where the division of chromosomes is delayed, are these particles found. Heitz recently has developed a theory to explain the origin of the nucleolus based on the fact that in sister nuclei there often exists a sym- metry in nucleoli number, position, form, and size. He believes that there is a correlation between number and position of satellites and secondary constrictions and number and position of nucleoli in somatic cells, and that the nucleolus originates in the form of a collar around the achromatic thread that holds the satellite or secondary constriction segment to the end of the main chromosome body. Brunn (1932) indi- cates that Primula seclusa does not possess satellite chromosomes but possesses nucleoli. On the other hand, Geitler (1932) presents evidence supporting Heitz’s theory by finding four nucleoli in telophase stages in a tetraploid form of Crepis capillaris which has four satellited chro- mosomes. It was mentioned in the text that the present author also found some such condition in the species where the satellite situation was thoroughly studied. For example, such correlation was particularly noticed in Callisia with two satellites with usually one or two nucleoli; in Paeonia suffruticosa in two individual plants with three and four satellites respectively with nucleoli constantly fewer in the former and more in the latter. However, it was found that this correlation was not complete, since it is decisively shown by Van Camp and by the present author that there is no localization of nucleolar development compar- able to Heitz’s assumption and that, even at resting stages, there are frequently found nuclei that contain nucleoli far above the number ex- pected on the basis of Heitz’s theory. It may therefore be pointed out that there are no definite nucleolus-producing chromosomes; that nucle- oli may be produced in the form of small globules on the surface of every chromosome, which latter collect into larger globules; that some of these larger globules may come in contact with the satellites and remain attached there; and that finally the number of nucleolar globules may decrease to one by fusion between the globules, so that generally one large nucleolus is found by the time the chromosomes have developed to the late stage of prophase. De Mol’s assumption of nucleolar frag- mentation is found contrary to all observations made in all species re- ported here, and by no known mechanism can such a process be adequately explained; on the contrary, the present author observed in living tissue a fusion between two moderately large nucleolar bodies but has never seen any indication of fragmentation. No adequate explanation can be offered for the origin of a protu- berance (bud-like structure) at the side of the large nucleolus in meiotic stages nor for the extranucleolar bodies found both in somatic and 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 311 meiotic nuclei. Based on the staining reaction, it is suggested that the extranucleolar bodies may be undissolved nucleolar particles that have been carried by the dividing chromosomes into the new nuclei, there to remain as inert particles. It is difficult to see if there may be a rela- tionship between the bud-like protuberance and extranucleolar bodies. Until the origin of both these bodies is carefully studied from living material, it will be futile to speculate as to their origin and function. Whenever there is a similarity of reaction of nucleolar bodies inside the same nuclear membrane, this may be explained by assuming that each represents a nucleolar globule which, owing perhaps to its being attached to a satellite, is not able to fuse with another such globule into a large single one. The similarity or dissimilarity between the true nucleolus and these bodies may be intelligible if a method of differential staining is applied. III. BEHAVIOR AND FUNCTION OF THE NUCLEOLUS. The present author is inclined to put considerable stress on the behavior of vacuo- lation of the nucleoli and considers this phenomenon of significance, since it was pointed out above that a large vacuole may disappear com- pletely and yet not cause any appreciable decrease in the volume of the nucleolus. Therefore it seems more accurate to assume that instead of the nucleolus being used up as reserve food, it is concerned in the general metabolic processes of the organism as a whole. This assump- tion may be more validated when we consider that the nucleolus is very probably universally found in all nuclei of all organisms. An exception to this rule is found in the male gamete, as cited by Wilson (1925), Sharp (1926), Tischler (1926), and Belar (1928). This exception is probably more apparent than real, however, for it seems likely that the nucleolus is not lacking but only delayed in its development. The chromosomes for a long time remain unmodified until the male gamete reaches the female gamete, and there they seem to go through a resting stage, during which time nucleoli in small globules come to appear. This conclusion is drawn from the figure presented by Wylie (1923) of Vallisneria spiralis and from the studies of Sax and Edmonds (1933) and O’Mara on Lilium (1933). Montgomery (1898) makes the statement that the nucleolar vacuoles in Spirogyra are normal structures. It was found by the present author that the nature of these vacuoles does not indicate any peculiarity char- acteristic only of that species but a similarity in behavior comparable to nucleolar vacuoles observed in Callisia stigma hairs and in all other parts of the plant. The number of nucleoli varied from a few small ones, differing among themselves in size, to a single large one. It was 312 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV also noticed that thinly opened fine threads of the chromosomes sur- round the nucleolus, giving an appearance similar to the nucleus illus- trated in Fig. 18. Conard (1931) seems to be justified in opposing some investigators who claim the origin of the chromosomes from the nucleolus. To this category may belong the findings of Kater (1928) and Faulkner (1929), who claim that nucleoli give rise directly to chromosomes. In this connection some Spirogyra were treated with Ehrlich-Biondi stain which was introduced under the cover glass, and there was seen a distinct demarcation between the chromosome threads and the nucle- olus. The chromosomes did not take any stain, while the nucleolus took a brick red color. It was also observed that the pyrenoids sur- rounded by the chloroplastic mass took a stain identical with that of the nucleolus. The similarity of staining reaction of the nucleolus and the pyrenoids may signify that they are composed of the same chemical material, or it may merely be a reaction due to their having the same electrical charge. The significance of their similarity in reaction with this stain will remain an open question until their chemical make-up is understood; therefore it is futile even to suggest that nucleoli and pyrenoids may have a homologous function, one in the nucleus and the other in the cytoplasm. As mentioned in the introduction, according to de Mol (1926), Zirkle (1928), and Fikry (1930), the nucleolus may play a part in transmitting hereditary stimuli from the chromosomes to the cytoplasm. This theory may be justified from the fact that the nucleolus is built up from small globules that may either be exudation products from the chromosomes or formed on the surface of the chromosomes from a chemical combination between some material produced from the meta- morphosing chromosomes during the late telophase or between matrix substance and material taken up from the nuclear-sap. It is shown that this “compound” product, the nucleolus, seems to stay unchanged in volume during the entire development of the nucleus, disappearing in the cytoplasm only after the nuclear membrane has vanished, or perhaps when it has gradually become freely permeable to cell-sap. The fact remains that the nucleolus disappears in the cell-sap either com- pletely or partially, and in some cases, as in Callisia (Figs. 15-17), Polystichum (Fig. 46), and Yucca (Fig. 56), it may be left, in its entirety, in the cytoplasm and stay there undissolved, at least for a long time; while in other cases, as was indicated above, the nucleolus may not be dissolved in the cytoplasm at all, or may be included in the daughter nuclei, as seems to be the case in some Protista. If the assumption is correct that the nucleolus is a compound product 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 313 originating, in part, from all the chromosomes and thus containing material from all the chromosomes, and hence from all the genes, and since, excepting perhaps in the case of some Protista, this substance is dissolved in the cell-sap during or after cell division, the suggestion of the above authors that the nucleolus may play a part in transmitting hereditary stimuli from the chromosomes to the cytoplasm deserves consideration. It may be assumed that the chromosomes fundamentally as self-perpetuating bodies divide up and are carried into, and become the basis of the daughter nuclei; that the chromosomes, as such, do not directly transmit gene stimuli, but that this transmission takes place, indirectly, through the matrix substance, which may be considered a by-product of the chromosomes. If this is correct, an explanation may be offered that would perhaps help in interpreting some data presented by Riley (1932) concerning the phenomenon of self-sterility in Cap- sella. The data of this author seem to indicate that a self-sterility factor, instead of exerting its inhibitory influence over only half the number of the pollen grains, as is the case in Nicotiana (East, 1929), affects all the pollen grains. In the case of Nicotiana the factors of self-sterility may not be so inhibitory as in Capsella; hence the differ- ence in these two genera. Therefore a hereditary factor of self-sterility may be transmitted through the nucleolus into the cytoplasm of the microsporocyte and through the four pollen grains resulting from the division of the microsporocyte. The presence of a bud-like protuberance on the nucleolus in rice meiotic cells has led Selim (1930) to assume that a single nucleolus buds off a secondary nucleolus which disappears during the develop- ment of the chromosomes and the primary one at late diakinesis. He supports the view that the secondary nucleolus contributes material to the chromosomes and the primary one to the spindle formation. Bud- ding is taken by him as a separation of the nucleolus into two different materials. Sethi (1930) has entertained principally the same ideas as Selim. In general, the transportation theory seems to have many followers, based on the fact that there has been found some sort of direct connection between some chromosomes and the nucleolus, and also because there has been observed a difference in stainability of chromosomes in early and late stages, a faint stain, or none, being taken in early stages and a deep stain at late stages. With this is associated the disappearance of the nucleolus by the time the chromosomes are arranged on a metaphase plate. Harper, as early as 1905, working on certain mildews, refuted the presence of direct contact between the chromosomes and nuleolus. Fikry (1930), working with Rumex, sug- gests that the connection observed between nucleolus and spireme is 314 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XIV due to chance, and that there is no constant and fixed connection be- tween them. ‘The present author found no connection between the nucleolus and chromosomes in Yucca (Fig. 55); in Callisia and others the connection is only between certain chromosomes and the nucleolus. It was asserted above that the diminution or disappearance of the nucleolus had no correlation either with the stainability of chromosomes or with building up of the chromosome mass. For this adequate ex- planations were offered. The theory of the formation of the spindle by one of these nucleoli must be refuted because many investigators have observed and reported that nucleoli may be in part or completely extruded from the cyto- plasm, but the chromosomes nevertheless divide, as illustrated in Fig. 56, without the assistance of the nucleolus and move toward opposite poles. In this connection should be discussed the question of the electro- magnetic quality of the nucleolus and chromosomes proposed in the following words by Zirkle (1928) as an addition to a number of theories proposed earlier by others: “The plastin, being electro-positive, changes the electro-negative spireme by flowing into it, to an electro-positive chromatin complex; thus the chroniatin, which had collected at the equatorial plate as far as possible from the poles of the spindle, reverses its motion with its electrical charge and migrates to the two poles.” Christoff and Gentscheff (1932) have subjected Zirkle’s proposed point to a test by putting an electrical charge through a maize sprout, and find that nucleoli do show a definite charge, thus giving support to the above idea. However, it was mentioned early that in living material the present author has observed two large bodies of nucleoli fused to- gether into one, that the nucleolus may be bodily extruded into the cytoplasm without an appreciable decrease in size, as is shown for Vucca in Fig. 56. Therefore, based on these two facts alone, the polarity phenomenon can not be responsible for dividing the chromo- somes, because two similarly charged bodies ordinarily could not fuse into one, but would repel each other, and because in Yucca at least, the nucleoli being extruded into the cytoplasm, they are eliminated from playing any part in the division phenomenon. It was stated above that the present writer is inclined to put con- siderable stress on the behavior of vacuolation of the nucleoli. It is very likely that the nucleolus may play a part in the organization of the cell as an organ that may primarily be concerned with assimilative processes by “imbibing” substances from the surrounding medium, con- verting them into material that may be utilized by the chromosomes especially for their growth and perpetuation. Schaede (1929) states 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 315 that “Directe Verbindung zwischen Kernfaden und Nucleolus in dem Sinne, dass ersterer in amdboide Fortsdtze des letzterer einmiinde, besteht nicht, auch kein unmittelbarer Ubergang von Nucleolarsubstanz in den Kernfaden,” and that “In Anbetracht ihrer gegenseitigen Beziehungen ist eine Abgabe von Substanz in abgednderter Form aus dem Nucleolus in den Kernfaden wahrscheinlich .. .’ As Schaede points out, the nucleolar substance itself can not be directly involved in this process, as Van Camp and many others have believed. It was amply shown that the nucleolar body remains undiminished during the chromosome development. Its assistance in this process may be of an indirect nature during the vacuolation and devacuolation described above. The latter phenomenon may not be abrupt but simultaneous, since no appreciable decrease or increase was observed during this process. G. SUMMARY 1. Because of the evidence offered that the chromosomes never lose their individuality in structure and position inside the nuclear mem- brane, it is argued that a direct flow of any nucleolar substance from one end of a chromosome through the entire spireme is not possible. In the case of Callisia some, but not all, chromosomes may be attached to a nucleolus. Applying a delicate differential stain (Ehrlich-Biondi), there was never found an intermediate staining at any point of any chromosome that may have come in contact with the nucleolus, and at all stages the chromosomes were stained pale blue and the nucleolus bright red, indicating that no direct interchange of material had taken place between chromosomes and nucleolus. No correlation was observed between a decrease of nucleolar volume and the process of “emerging” chromosomes. The volume of nucleolus in meiosis studied in some species, once the maximum is reached during late telophase-early leptotene, remains constant all through the chromosome development. 3. The satellite and secondary constriction phenomena in Callisia, Paeonia, and Pinus are analyzed to determine the nature of satellite- nucleolar relationship. The nucleolus is found to originate on the surface of the chromosomes during telophase in the form of small globules, as Van Camp has shown, and not on a specific region (achro- matic threads of satellite and secondary constriction) of satellite chro- mosomes, as Heitz believes; hence there were found numerous globules instead of a limited number corresponding to the number of satellites or segmented chromosomes. When there was a connection between a satellite and a nucleolus, it was found that the nucleolus is attached at 316 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV the side of the satellite instead of forming a collar around the satellite achromatic thread. 4. There was found some correlation between the number of satel- lites and nucleoli, but not exactly as was expected according to Heitz’s theory. The symmetry between two sister nuclei in the number, size, and position of nucleoli is suggested to be due primarily to a physical relationship between a satellite and a nucleolus, and to corresponding spaces between the same set of chromosomes in the sister nuclei where nucleolar globules “exuded” from the adjacent chromosome surfaces fill in. 5. A behavior of extrusion of nucleolar particles in the species studied is described. In Callisia there were found small extranucleolar bodies which showed a difference in stainability when compared with the nucleolus. Their origin could not be determined. It is suggested that they may be inert nucleolar particles representing pieces of nucle- oli that have been retained in the sister nuclei as lagging particles. On the side of the meiotic nucleolus there was often observed a protuber- ance, a bud-like growth. Its nature and relationship, if any, with some of the free nucleolar bodies observed in meiotic nuclei could not be determined. 6. A critical analysis is presented of de Mol’s finding that the size and number of nucleoli vary proportionately as the number of chromo- somes is increased in polyploid series. No significant correlation was found in a similar relationship studied in Petunia. 7. The results of some observations made from living tissue (stigma hairs of Callisia) are as follows: (a) A clear area around a nucleolus is an artifact which can be produced by applying toxic fluids to stigma hairs in culture. (b) Small extranucleolar bodies were often observed in the nuclei of the hairs. The nucleolus was found showing some affinity to these bodies when a clear area was artificially created. (c) No general chromosome and nucleolar connections were observed. The connections, when present, may be limited to certain chromosomes. Often specific connection was found between the nucleolus and the extranucleolar bodies. (d) The vacuole in the nucleolus is normal and not an artifact and probably appears and disappears normally. The phenomenon suggestive of vacuolation is considered by the present author to be of physiological significance. Some data are presented concerning this behavior. 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 317 ne a no wm N — et Oo OC is) oO N — LITERATURE CITED ABELE, K. (1930). Untersuchungen tiber die Nukleolen. (Proto- Bevar, K. (1928). Die cytologischen Grundlagen der Vererbung. andbuch der Vererbungswissenschaft. Lieferung 5 (1,B), and I. 1929). Beitrage zur Kausalanalyse der Mitose. III. Untersuchungen an den Staubfaden, Haarzellen und Blattmeris- temzellen von oo virginica. (Zeitschr. Zellforsch. Mikr. Anat. 10:73-1 3 odin (1909). “Die Blastomerenkerne von Ascaris megalo- ephala und die Theorie der Chromosomen- individualitit. (Arch. Zellf, 3:181-268. . Brunn, H. G. (1932). Cytological studies in Primula. (Symb. . BURNHAM, CR (1932). An interchange in maize giving low sterility and chain configurations. (Proc. Nat. Acad. Sci. 18 :434440. ) . CuristorF, M. & G. GENTSCHEFF (1932). Untersuchungen tiber das & Verhalten des Nukleolus wahrend der Kernteilung. (Zeitschr. Zellforsch. Mikr. Anat. 15:540-546. ) Conarp, A. (1931). Il n’y a pas d’ “amphinucléole” chez les Spiro- gyra. eae Be Aart oe Biol. 107:923. DARLINGTON, (1926). Chromosome studies in the Scilleae. (Jour. Gen. ere 3370081. 1 Recent advances in cytology. (P. Blakiston’s Son & Co., Inc. Philadelphia. ) . Davis, B. M. (1903). Odgenesis in Saprolegnia. (Bot. Gaz. 1233-2 . DERMEN, ue (193la). Polyploidy in Petunia. (Amer. Jour. Bot. 18:250-261. ai 931b). A study of chromosome number in two genera of Berberidaceae: Mahonia and Berberis. (Jour. Arnold Arb. ee . East, E. (1929). Self-sterility. (Bibl. Genet. 5:331-370.) : eee er (1929). The early prophases of the first odcyte division as seen in life, in Obelia geniculata. (Quart. Jour. Microsc. Sci. 73(2) 1225-242. Fixry, M. A. (1930). Phenomena of heterotypic division in the pollen- mother- aah of a tetraploid form of Rumex sculatus var. typicus. (Jour. Roy. Microsc. Soc. 50:387-419. . Gates, R. Latter (1927). Observations on the pollen de- Roa J, velopment of two species of Lathraea. (Jour. Roy. Microsc. Soc. 47:209-225. ; GEITLER, L. (1932). Das Verhalten der Nukleolen in einer tetra- loi en Wurzel von Crepis capillaris. (Planta, 17:801-804 , Goren , A., G. MANGENOT & PLANTEFOL (1933). Traité de cytologie végétale (pp. 183-184). (Librairie E. le Francois, aris. . Harper, R. A. (1905). Sexual reproduction and the organization of the nucleus i in certain Mildews. (Publ. Carnegie Inst. Washington, No. ; : Herz, E. 28). Das Heterochromatin der Moose. I. (Jahrb. 8 “(19 Wiss. Bot. 69:762-818.) Ww oO JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Die Ursache der gesetzmassigen Zahl, Lage, _ ——__— (193la). Form und Grosse pflanzlicher Nukleolen. (Planta, 12: 775-844, ) —_——_--— b ). ukleolen und Chromosomen in der Gattung Vicia. (Planta, 15:495-505.) _ Kacawa, F. (1926). Cytological studies on Triticum and Aegilops. 24. (Cellule, 37:229-3 . KaTER, J. McA. (1928). Structure of the nucleolus in the root-tip cells of Nicotiana longiflora. (Univ. Calif. Publ. Bot. 14(12): 319-322.) KaurMann, B. P. (1931). Chromonemata in somatic and meiotic mitosis. (Amer. Nat. 65:280-283.) _ KoerpeRIcH, JOSEPHINE (1930). Etude comparative du noyau, des chromosomes et de leurs relations avec le cytoplasme (Nothoscor- dum, Eucomis, Beschorneria). (Cellule, 39 :307-398. 1 . LEE, A. B. (1928). Microtomist’s Vade-Mecum (p. 177). (P. Blakiston’s Son & Co., Inc. Philadelphia. ) . Lewirzxy, G. A. (1931). eview of investigations performed at the Cytological Laboratory of the Institute of Plant Industry in L,. 5; (Bull. Appl. Bot. Gen. in Plant-breed. 27(1) :9- 17.) MAEDA, 7. (1930). The meiotic divisions in pollen-mother-cells of the Sweet-pea (Lathyrus odoratus L.) with special reference to the cytological basis of crossing-over. (Mem. Coll. Sci. Kyoto Imp. Univ. 5:89-123 . MarsHak, A. G. (1931). The morphology of the chromosomes of i 39. isum sativum. Sa ee 2:318-3 . Martens, P. 8). Recherches expérimentales sur la cineése. Mays. (Mo. Agr. Exp. Sta. Res. Bull ane . McCune, C. E. (1928). Differential ee of Mecostethus 1756-778. Ag ae Scag Zellforsch. Mikr. L, W. E (1926). The nucleolar pees regarded as bearers of erp or finishing materials of the genes. (Genetica, 8537-542. ) — . On chromosomal constrictions, satellites and nu- cleoli in Hyacinthus orientalis. (Beitr. Biol. Pflanzen, 15:93-116. ) eee 9 (7. ucleolar number and size in diploid, triploid and aneuploid Hyacinths. (Cellule, 38:7-64. 9 . Montcomery, T. H. . Comparative cytological studies, with special regard to the morphology of the nucleolus. (Jour. Morphol. 15 :266-582. ) . NavasHin, S. G. (1927). Zellkerndimorphismus bei Galtonia can- dt egy Des. und — verwandten Monokotylen. (Ber. Deutsch. ot. Ges. 45:415-4 : O'Maze J. (1933). San of the generative nucleus in the pollen 578. tube of Lilium. (Bot. Gaz. 94:567- . Ras, C. (1885). Ueber Zellteilung. (Morphol. Jahrb. 10: 214-330. ) .R (19 Au, N. S. (1930). On reduction division in the pollen-mother-cells of Cyanotis cristata. (Jour. Ind. Bot. Soc. 9:79-113. H. P..¢ 932 . RILEY, ). Self-sterility in shepherd’s purse. (Genetics, 231-295. . ROBERTSON, W. R. B. (1931). Chromosome studies. II. Synapsis in the Tettigidae with special reference to the presynapsis split. (Jour. Vy aeneli and Physiol. 51:119-146.) 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 319 . SAX, Kart (1932). The cytological mechanism of crossing over. (Jour. Arnold Arb. 13:180-212. Sax, Kart, & Epcar ANnperson (1933). Segmental interchange in chromosomes of Tradescantia. (Genetics, 18:53-67. Sax, Kart, & H. W. Epmonps (1933). The eke ania of the male gametophyte in Tradescantia. (Bot. Gaz.; in pre . SCHAEDE, R. (1929). ber das Verhalten des ae Son wahrend der Kernteilung. ee 5:41-54. . SCHEUBER, L. M. (1932). A cytological study of Timmia cucullata. (Cellule, 41:145-162. ) . SetiM, A. G. (1930). A cytological study of Oryza sativa L. . SEMET, E. DE (1913). Chromosomes, prochromosomes et nucléoles -) dans quelques Dicotylées. (Cellule, 1930 . SETHI, M. L. ). To ae in Cassia didymobotrya. 139. (Jour. Ind. Bot. Soc. 9(2/3) :126- ; Seri iD, W. (1926). Introduction to cytology. Ed. 2 (McGraw- rk. Hill Book Co., Inc. ew Yo Structure of large somatic chromosomes. (Bot. Gaz. 88:349-382. ) . Sorokin, H. (1929). Idiograms, nucleoli, and satellites of certain B —4 Ranunculaceae. (Amer. Jour. Bot. 16:4 . SPRUMONT, G. (1928). oe et satellites dans |’Ornitho- galum. (Cellule, 38:269-2 . TIscHLErR, G. (1926). ice Pflanzenkaryologie. Handbuch 2, Bd. der Pflanzenanatomie 2, Lief. II. (Verlag von Gebriider . VAN Camp, G. M. 1924), Le role du nucléole dans la caryocinése 0. somatique (Clivia miniata Reg.). (Cellule, 34:1- KAYAMA, K. (1930). Goats uations to the cytology of fungi. IT. Cytological studies in Morchella deliciosa Fr. (Cytologia, 2:27-36. ) Wison, E. B. (1925). The cell in development and heredity. (The oe Co., New York. Lig, R. B. (1923). Sperms of Vallisneria spiralis. (Bot. Gaz. 5: 191-202. ) 2.. YAMAHA, G. & Y. Stnotd (1925). On the behavior of the nucleolus in the somatic mitosis of a plants, with microchemical notes. (Bot. Mag. Tokyo, 39:205-2 i A EATES: Vy, oS; (192 ,, The teh of Tmesipteris tannensis, (Roy. Soc. Proc. 98:227-244. ernh, - Z1RKLE, C. (1928). Nucleolus in root-tip mitosis in Zea Mays. ( Bot. Gaz 84: nee 8.) ————— (1931). Nucleoli of the root-tip and cambium of Pinus Strobus. (Cytologia, 2:85-105.) Ly Ft a . 10. le; 3d, ig. 21. . 14, ae AD: i.e JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV EXPLANATION OF PLATES 64 TO 67 All sections are longitudinally cut. Figs. 1-28 Callisia repens Method: Lewitzky fixation; stained in iron-alum- haematoxylin. Stage: prophase, oblique view, nucleolus in 7 one probable attachment between nucleolus and chrom Method: same as in Fig. 1. eu Oe side view (nucle- olus not show Method: Pike + 0.25% saponin; stained in i.a.h. first; destained and restained in Ehrlich Biondi mixture. Stage: prophase drawn on nucleolar level, er stained homo- arco pale blue and nucleolus bright Method: 4. Me ame as in Fig. 3. Stage: nee ei with two nucle- olar bag 5. Method: same as in Fig. 3, before destaining. Stage: meta- phase, the satellite chromosome pair, fully split; en of the sister chromosomes on e : shows a furthe it. 6. Method: same as in Fig. 3. Stage: late anaphase, chromosomes homogeneously pale blue 7. Method: same as in Fig. 5. Stage: an anaphase chromosome with split and twist. 8. Method: Lewitzky fixation; stained in crystal-violet-iodine. Stage: split and twisted chromosomes from late an e. 9. Method: same as in Fig. 3. Stage: telophase; diagrammatic "Picker panaeh of nucleolar globules. Method: same as in Fig. 3. Stage: telophase; two small nucle- oli shov 11. Method: same as in Fig. 3. Stage: telophase, with one nucle- olus larger than in Fig 10. 12. Method: same as ig. 1. Stage: interphase, with three nucleoli (the a neh were tested with E.-B. stain; all took bright red color). Method: same as in Fig. 1. Stage: dormant stage nucleus in stigma hair, a vacuolated nucleolus pith clear area around and four extranucleolar bodies. pene 1-3 minutes Carnoy > followed by oe xa ative e; stained in c.v.i., destained and restained a : prophase from a cell of mnicrosporogenoss pee with 1ucleoli drawn from the level of the — Method: same as in Fig. 14. Stage: prophas of two micro- ocyte sister cells, each with an extranuclear see eit drawn renin the level of the nucleoli . 16 and 17. ethod: same as in Fig. 14. Stage: late anaphase o microsporogenous cell division, with median and polar lagging nucleoli, respectively. 18. Method: Fixative same as in Fig. 14; stained in i.a.h. sa leptotene, with a cea Ae Pt inside the nucleus a agging nucleolus outside. Method. same as in Fig. 14. Stage: leptotene, with two small normal —— one with a small vacu Method: as in Fig. 14. Stage: leptotene, with a large normal nucleo Method: 1 Fs re Carnoy, followed we Zirkle (1928) oh i stained in i.a.h.; later destained and re in ee one large normal nucleolus at chbaeouonie rere 1933] DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 321 ment and two pai smaller nucleoli tes stained somewhat ee ay, (Some of these nucleoli were in the form of a bubble with a very ein a mE sine a somewhat lighter color ; Poatieae inside. Method: sa me as in Fig. 21. Stage: pachytene, with a bud-like protuberance at the oe of the nucleolus; a satellite chromo- some attached to the nucleolus. a 9 tw NS Fig. 23 ethod: same as in Fig. 21. Stage: Pachytene, with one out- side and one inside additional nucleoli, which are of the same nature as ue oupaie in Fig. 21. Fig. 24. Method: 1-3 minutes Carnoy, followed by Lewitzky ; stained in lah. Stage: ation: nucleolus with four small va acuoles, olu dark stain in the nucleolar enlace. somewhat hyaline in the marginal portion . Method: Lewitzky fixation; stained in Stage: diakinesis ; at the side of the elaine a satellite es ay (Chromo- utline. ) el . nN on rig 5. Stage: interphase of first meiotic division, showing globular nucleoli. (Chromosomes could not be shown because of their very faint color; however, these ee were very closely associated with ‘the chromosome thre : Method: same as in eae 25. Stage: microspore prophase with a large nucleolus with one probable chromosome attachment, and an ge icléciar body below a long doubled-up chromo- me. 8. Method: same as in Fig. 25. Stage: microspore sare alte with one satellite chromosome and an extranucleolar body nea the distal end of the same 2 oe is) ON Ss o ot — o) Q. pe) a oO feb] n ey a9 bo N 2 A nh ore Figs. 29 and 30 Paeonia suffruticosa (Arboretum plant) Fig. 29. neeeeree Lewitzky fixation; stained in c.v.i. Stage: metaphase ; of a satellite chromosome to which a small piece of Bares is attache Fig. 30. Method: same as in F . Stage: ue: parts of chromo- somes which are split and show spiral twis Fig. 31 Paeonia suffruticosa (greenhouse plant) Fig. 31. Method: same as in Fig. 3. Stage: telophase, where small bright red globular nucleoli were seen along pale blue chromo- some threads. Figs. 32 and 33 Paeonia suffruticosa (Arboretum plant) Fig. 32. Method: same as in Fig. 29, but destained and restained in E.-B. Stage: telophase, sister nuclei. Small bri ar red glo i lar nucleoli are presented in a ee mmatic way how m symmetry in two sister nuclei in a number, posi , and size four nucleoli. fae number of nucleoli may be as high as seven and possibly mor 322 . 34. w mn as n - N - oe) fe Ke) JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV Fig. 34 Paeonia suffruticosa (greenhouse plant) Method: same as in Fig. 3. Stage: interphase of sister cells, each nucleus : which possesses five nucleoli; no symmetry of size and position. Three-satellited plant with five nucleoli. i 35 Paeonia W oodwardii Method: Lewitzky fixation; stained in E.-B. Stage: interphase. Six-satellited plant with nine nucleoli. Fig. 36 Paeonia Delar rayt alba Method and stage: same as Six-satellited plant with number of Asin pie as high as eleven, generally less than this number. Figs. 37 and 38 Pinus Strobus Method: Lewitzky fixation; stained in E.-B. Stage: telophase ; sister nuclei with numerous small nucleolar globules. Method: same as in Fig. Stage: interphase; only nucleoli wn. Ten-secondary- constricted plant with as high as four- teen nucleoli. Figs. 39-44 Pinus Thunbergit Method: Lewitzky fixation; stained in E.-B. Stage: early lep- totene with nine nucleoli. . Method: Lewitzky fixation ; stained in E.-B. Stage: early dip- lotene with nine nucleoli. . Method: same as in Fig. 40. Stage: late diplotene with six oli. nucle . Method: same as in Fig. 40. Stage: beginning of metaphase ne Pe with four nucleoli. . Method: peer on same as in Fig. 40, but = in ¢.v.i Stage: interphase ater meiotic division. nly one of the two sister nuclei shown possesses nine nucleoli. . Method: same as in Fig. 43. Stage: late telophase of tetrad formation, number of nucleoli varying from eight to twelve. Figs. 45-47 csiviiele stic he ea . Method: Lewitzky fixation; a “5; Stage : meta- phase, with small chrindan ‘cuales in ya division region, and one outside this region, remnant t of an earlier cell division ision., . Method: same as in Fig. 45. Stage: interphase from periblem ‘ ; ; aide. Gardtk : inds ) “bag.” One large round nucleolar body stained darker like the small ones outside the nucleus, this body being a remnant of n. . Method: same as in Fig. 45. Stage: a nucleus in early p phase with a single bag-like nucleolus containing darker ao ing smaller sors se This figure is from a more actively divid- ing region Pies 48 - 49 Call’sia repens Method: Lewitzky fixation + 0.25% saponin; stained in 1.a.h. Stage: metaphase an eo the ee Sram differences of the chromoso Method: Navashin fixation, half strength; stained in c.v.i. Stage: anaphase Jour. ARNOLD ARB. VOL. xIVv. PLATE 64 ORIGIN AND BEHAVIOR OF THE NUCLEOLUS IN PLANTS THE HELIOTYPE CORP. BOSTON Jour. ARNOLD ARB. VOL. XIV. PLATE 65 ORIGIN AND BEHAVIOR OF THE NUCLEOLUS IN PLANTS THE HELIOTYPE CORP. BOSTON Jour. ARNOLD ARB. VOL. XIV. PLATE 66 peer 9° 0 Resp) es 8 ete : pe S fy 9 ry oN ‘ F vag oO gs * 0) F "Oey 2 ° + Se o.:2 . es a Q 46 ———— ORIGIN AND BEHAVIOR OF THE NUCLEOLUS IN PLANTS THE HELIOTYPE CORP. BOSTON Jour. ARNOLD ARB. VOL. XIV. PLATE 67 ORIGIN AND BEHAVIOR OF THE NUCLEOLUS IN PLANTS THE HELIOTYPE CORP. BOSTON 1933] Fig. DERMEN, ORIGIN AND BEHAVIOR OF THE NUCLEOLUS 323 50. Lo al — wn bo Fig. 50 Paeonia suffruticosa aes plant) Method: Lewitzky fixation; stained in c.v.1. Stage: anaphase presented in two sections with four satellites in total. Fig. 51 Paeonia suffruticosa (greenhouse plant) . Method: same as in Fig. 31, but again restained in c.v.i liga anaphase, sed flat to show the satellited feature. There three wept Hin instead of four; one of the cabtanatnaity con- stricted chromosomes is without one. Fig. 52 Paeonia Woodwardit : sae same as in Fig. 35, but restained in c.v.i. Sta eta- ad side pressed flat to show the satellited pan This. site fe fe pairs of satellite Fig. 53 Paeonia Delavayi alba Method: same as in Fig. 52. Stage: metaphase, only one chromosome showing. This plant has three pairs of satellites like those shown in Fig. 52, but the satellite, as shown ae is larger than the one on the subter rminal chromosomes in Fig. 52 and is attached with a longer achromatic thread. Fig. 54 Pinus Strobus Method: same as in Fig. 37, but stained in i.a.h. oar e: ana- phase, gue part of the chromosom es showin t t five pairs of secondary constricted chromosomes are ee Vine actually were counted). Figs. 55 and 56 Yueca flaccida . Method: smeared and fixed in Fleming strong; stained in c.v Stage: diakinesis, drawn on nucleolar level. No eect was observed between oe of the chromosomes and nucleolus. Method: same as in Fig. 55. Stage: telophase; two attached nucleolar bodies at the upper pole. Fig. 57 Hepatica triloba Method: same as in Fig. 55. Stage: two nucleoli are shown es aaa aah a from very early leptotene, and b from diplotene. No difference in size is seen CyToLoGicAL LABORATORY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 324 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV SOME HYMENOMYCETES FORMING MYCORRHIZAE WITH PINUS STROBUS L. A. B. Hatcu ANbD C. TaLsott HATCH With plates 68 to 71 THE CLASSICAL EXPERIMENTS of Elias Melin (1922) yielded the first conclusive evidence that the Hymenomycetes are involved in the mycorrhizal associations of forest trees. The only fully satisfactory means available for exploring such relationships, namely, the artificial association of tree seedlings with fungi in pure culture, was employed. In earlier experiments (Melin, 1921) mycorrhizal fungi (Mycelium radicis subsp.) were isolated from tree roots and their relationships to tree seedlings demonstrated. By 1925 species among the genera Lactarius, Russula, Cortinarius, Tricholoma, Amanita, and Boletus had been shown to form mycorrhizaé with Pinus, Picea, Larix, Betula, or Populus (Melin, 1922, 1923a, 1923b, 1924, 1925a, 1925b, 1925c). Hammarlund (1923) and Masui (1927) have subsequently reported successful syntheses in pure culture. Precise knowledge on the etiology of the mycorrhizal habit of Ameri- can Pines is fundamental to studies of their nutrition." With this larger problem in view, a series of syntheses experiments were pursued with P. Strobus L. and P. resinosa Ait. in association with species of Bole- tinus, Boletus, Lactarius, Russula, Amanita, and Mycelium radicts subsp. cultured in both Sweden and America (See Table 1). The study was initiated in Professor Melin’s laboratory at the Mycological Laboratory, Royal Academy of Forestry, Stockholm, in the fall of 1929. Descriptions of technique used in syntheses experiments have been published by Melin (1921, 1923a, 1925a). But Melin has been so frequently misquoted and his technique so often incorrectly followed that recapitulation in connection with the present work is desirable. Cultures of known fungi were obtained by means of tissue isolations from the pileus and upper portion of the stipes of young sporophores. The following fungi were obtained in culture:° 1Absorbing roots (short roots, “Saugwurzeln”) of Pines in natural habitats are pci aogs mycorrhizal. us nutrients entering the tree by these channels must first pass through a fungal mantle that entirely separates the root cells of the tree from the soil. 2Boletinus eae and rina castaneus were identified by Mr. C. L. Krieger, Washington, C., Lactarius chrysorheus by the junior author and all other specimens by eae E Melin and T. Lagerberg. 1933] HATCH, HYMENOMYCETES FORMING MYCORRHIZAE 325 Lactarius chrysorheus Fr. deliciosus (L.) ee Plate 68, G. us subdulcis (Bull.) F Bets fragilis (Pers.) Sing. puellaris Fr. Amanita muscaria (L.) Fr., Plate 68, D. Brees chrysenteron (Bull. ) Fr. piperatus Bull. “ granulatus L., Plate 68, B. “luteus (L.) Fr. ‘““ castaneus Bull., Plate 68, F. “ bovinus (L.) Fr., Plate 68, E. “edulis Bull. Boletinus porosus (Berk.) Peck, Plate 68, A. The plan of the work included syntheses experiments with Mycelium radicis subsp. isolated from both American and Swedish conifers. The technique developed by Melin (1921, 1925a) for isolating mycorrhizal fungi was employed. Essentially this consists of: (1) selecting com- paratively young and clean long-roots bearing mycorrhizae; (2) wash- ing these thoroughly in a strong stream of tap-water; (3) cutting the roots into short lengths, each bearing one mycorrhiza; (4) surface sterilizing the latter in 0.1 per cent bichloride of mercury; and (5) rinsing in three or more changes of sterile water. The time required for surface sterilizing small mycorrhizae of the forked (Pinus) or race- mose (Picea) types was two to five seconds (Melin, 1923a, p. 125). A considerable number of contaminations were inevitable, but longer treatments are lethal to the true endophytes. For tuberous mycorrhizae of Pine one minute usually did not prove injurious. After rinsing, the agar in test tubes. Uncontaminated pieces were later transferred to suitable culture media. All of the culture media employed contained, or consisted of, malt extract (See Melin, 1925a, p. 10). For the more sensitive fungi (ob- tained from either known sporophores or from roots) 5 per cent malt extract, sterilized by passage through a Berkefeld filter, was a suitable medium. Rapidly growing forms (Boletus bovinus, etc.) were culti- vated on autoclaved 5 per cent malt extract with 2 per cent agar (See Table 1). In the latter case a minimum pressure and period of sterili- zation was conducive to rapid growth of the organisms. 1The American brands of malt extract experimented with were not suitable. They are apparently iat at high temperatures, which destroy some nutritive properties. .We have used Liebig’s malt extract obtained from Apoteksvarucentral Vitrum, Stockholm DATA ON ABLE T 1 FUNGAL CULTURES Fungus ee ee ia an Neer eee ES Lactarius chrysorheus Westtown, Pa. Lactarius deliciosus Lactarius subdulcis Russula fragilis Russula puellaris Amanita muscaria Boletus chrysenteron Boletus piperatus Boletus granulatus Boletus luteus Boletus castaneus Place collected Djursholm@ Djursholm Djursholm Djursholm Djursholm Experimental- faltet Djursholm Djursholm Djursholm Ansonia, Pa. Date Most abundant trees in stand pure P. Strobus 8/26/29 9/22/29 9/20/29 9/20/29 10/15/29 9/22/29 10/3/29 10/28/29 10/9/29 9/20/29 8/24/29 plantation Pinus, Betula Pinus, Picea Pinus, Picea Betula, Pinus Betula Quercus Pinus, Picea Pinus, Betula Pinus, Betula Tsuga, Betula Diam. growth of colonies in mm. 10-12 (ME)¢ 30-35 (ME) (ME) (ME) (ME) 5-7 (MA)f 2-4 (ME) 8-10 (ME) 50-55 (MA) 70-75 (MA) 57-62 (MA) Color change Color of hyphae of media Type of Colony Remarks yellowish white slight loose, submerged Cultivated on yellow to slight loose, submerged solid media 4 whitish yellow s. after iso- aérial, submerged | lation. Form (MA) (Pl. 68) strands in cul- ; ture with Pine. white slight compact, sub- merged white slight compact, sub- merged white slight compact, sub- merged white slight compact, sub- merged (ME) compact, aérial (MA) (PI. 68) light brownish slight compact, sub- yellow lemon yellow dark loose, aérial brownish white, brownish light loose, aérial Formed i brownish with strands strands when ha Phat ark c aérial (Pl. 68) es Wi Bee on in pure culture. white pinkish aérial even (PI. 68) “SoysIp MWjeaq “UW NOT Ul IeaB 4[eUI yusD aad g "[W Og UO SkUp Og AOZ YYMOIZ Yuasaidaa «(WW),, Pexyteu asoyy f “POBI}XS }[VUL po}yBoolsap JO puviq UBdIIaWY UB UO paInyNg ‘a ‘OAIN}[ND [BUIZIIO YBIM A[JowXe Yoayd ZEGI UI sn Aq opew SuOI}BlOsE JUaNbasqng “uol}eIg JUueuIIadxg| se1og Auaysatpy ‘YsnoWH ““W Aq peyejosy “p *(qoRayxg HEA S$.SIqelT) Zutavjooyne Aq pazi[iiajs rede yuso aad Z “yoRa}xa z[VU Jus. Jad ¢—VWW “19314 Plejoyteg Ysno1y} esessed Aq pazi[iia}s Jovi}xe [eu pINbI] juao sed cAN '9 ‘uapemsg ‘UlepurA “Fse1Oq [BjusUIIadxq Uspl[SYOBq(NY ‘usewm[oH epuvsg WoOIJ [1IOG “ysVy “Y “qd AQ Sjusausedxa yod pues-jlos ur uMOIz SSUl[pees Woy pszelosy “q ‘uspamMg ‘dD SUOIJO9UUOD 4 duis! (GI) pesaou yorlq (WIN) 21-01 Binqed snozawn Ni “qns ‘esooT = ystumoiq =—-ystyurd-oytyan = (IL) 8T-ST Boold ‘SNUId 62/62/6 B1oqeiny, T StyIqy (veg) “IT eadIg S261 Ul 209 euou uvei3 ystheid (VI) ¢9-09 ‘e[njeg ‘snulg 0&/2/P uspl[syouq ny Q@ SUPIIAOI}E “IY E261 UP 2° auou useis ystheis = (VI) S6-¢8 SNUId 62/L1/h ‘H'N ‘eusey aT SUAIIAOI}R “I'L (VW) [etzee ‘yoedwos (QI) poss Bo0Ig -qns “yoeduioa auoUu yoryq jel ‘e[njeg ‘snuIq 08/%/p pvUepl[syorqiny guinso3114s0131u IW o3e8 YIM elnyegq Telige ystumoiq = uMOIg ‘ayy = (WIL) OF-GE ‘eadtd ‘SNUIG 6Z/62/6 pdieqeiny, T staiseayas (gq) ‘IW 938 YYIM poompiey Telge ystumolq = uMorg ‘ayy (WIA) LE-s ‘IaFUOD 6%, “dy = ‘e_ ‘eluosuy oT Iqo13g (snulg) “IW SuoTJOaUUOD 4VlIes Uld duieja) -1ew ‘etpaut jo yoetq 0} MoOT[aA snunig ‘e3nsJ, SNOJoWNN) svjIns UO pajsnio UMOIG YIVp ystumoiq dsap s(wyW) 69-F9 ‘sn3eyq ‘ernjog ‘eq ‘UdlIeM a‘pSnsoiod snutyelog peso -qns ‘yoeduioa euo0u ayy AGW) F-€ B[njog ‘SNUIq 62/02/6 wyoysantqg SI[npa snyzelog YStUMOIg o3e YIM (89 ‘Id) [elage yrep ystumorg ‘ay1yM (WIN) OL-S9 Badld ‘SNUIY 62/02/6 wyoysantq SnulA0g snyeog SY1BULIz Auo[oy jo addy, Bipsewu jo aeydAy jo a0[0O9 ‘wu UL SaruoTOo pueys ul seer} ayeq pezoa[[Oo avoulLg snsuny asueyo I0[0D JO YyMoIs “WIBIG quepunqge 4so,y SdYNLTNO TWONNA NO VLVd penuryuoj—t WTAViL 328 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV In designating the imperfect stages of mycelia isolated from roots we have followed Burgeff (1932, p. 147) and included the generic names (in parenthesis) as well as the specific names of the vascular plants from which the mycelia were isolated. These names are of value to the individual investigator and likewise conveniently serve to inform the reader of the identities of the hosts from which they were isolated. In those cases where the mycelial characters are so marked that the fungus in question may easily be recognized by other investigators (particularly when the fungus is associated with more than one vascu- lar plant), the name of the plant from which it was isolated is of less value and may be substituted by a descriptive specific name. Examples of such fungi are Mycelium radicis atrovirens Melin, and M. r. nigro- strigosum Hatch. The following fungi were isolated from mycorrhizae of Pinus Strobus, P. sylvestris L., Picea Abies (L.) Karst.: Mycelium radicis (Pinus) Strobi 1, M.r. (Pinus) sylvestris 1, Plate 68, C, M. r. ntgrostrigosum, M. r. atrovirens 1, M. r. atrovirens 2, M. r. (Picea) Abietis 1, Plate 68, I. The tree seeds used in our experiments were obtained through com- mercial seed houses. They were soaked over-night in water, surface sterilized for two minutes in 0.1 per cent bichloride of mercury, and rinsed in several changes of sterile water. They were then sown on agar in petri dishes, and contaminations and infected seeds were re- moved with a sterile spatula as they became evident. Early experience demonstrated that the particular sample of P. Strobus seeds we used required an after-ripening treatment to obtain even nominal germina- tion. Barton (1928) had shown that cold storage treatment was effective in hastening germination of Southern Pine seeds. We, there- fore, surface sterilized seeds of P. Strobus and stored them in a frigidaire at four to ten degrees centigrade for a period of two months. Regard- less of the acidity of the media (moist filter paper, agar, and peat) which we varied from pH 3.5 to neutrality, germination was uniformly good (approximately 80 per cent). Germination was procured in a constant temperature room at 25 degrees centigrade. As they germi- nated, the seeds were transferred directly into the culture chambers. The culture technique developed by Melin (1925a, etc.) was em- ployed in the syntheses. Fluvio-gilacial sand was screened and that portion having particle sizes between 0.5 and 2.0 mm. was used (ade- quate aération of the substratum is not possible in an undrained flask if smaller particles are included). The sand was boiled in concentrated 1Barton (1930) has published a second paper in which success is reported with low temperature treatments of P. Strobus 1933] HATCH, HYMENOMYCETES FORMING MYCORRHIZAE 329 hydrochloric acid for two hours, washed several hours in running tap- water, and finally in five changes of distilled water. It was dried in an oven, and 150 gram samples were weighed into 300 ml. Erlenmeyer flasks. The nutrient solution added to this substratum contained the following: KH,PO, 0.5 grams CaCl, O03. ait NaC] 0.025 ‘“ MgsO, + 7 H.O O15. ™ (NE).PO, O75. ™ Iron citrate 0.025 <“ Dextrose 0.5 a Distilled water 1. litre Thirty-seven ml. of this solution were added to each flask. The pH of the solution was 6.57. After autoclaving with the sand, this changed to approximately 4.2. The nutrient conditions in our experiments differed from those in Melin’s (1923a, p. 159). We used a less concentrated solution and likewise added 37 rather than 50 ml. of solution to each flask. Further, the growth of P. Strobus is considerably greater in culture than that of P. sylvestris seedlings. The differences in the nutritional conditions of the two sets of Pines were, therefore, quite marked. Since ™ growth is greater when nutrients are present in comparatively small quantities, it was argued that more short roots would develop. The probability of obtaining mycorrhizae would therefore be enhanced. It is not necessary, however, to obtain large numbers of mycorrhizae in syntheses experiments, as has been emphasized by McArdle (1932, p. 314). The unquestionable demonstration of the presence of only one typical mycorrhiza is adequate proof that the organisms concerned enter into mycorrhizal association with each other. The assembled culture chamber consisted of the Erlenmeyer flask with an inverted beaker over the cotton plug (Plate 70, D). Between March 23, 1930 and April 9, 54 of these units, each with a germinated seed of P. Strobus, and 52 units with P. resinosa seedlings, were set up. They were placed in the diffused light of a west window and received direct solar radiation, passing through the foliage of a large oak tree, late in the evening only (Plate 70, D). On June 5th to 7th the seed- lings were inoculated with all of the fungi listed above. The sub- cultures from which inoculations were made were less than ten days old. A number of seedlings were reinoculated August 21. Half of the flasks were opened during November 1930. These were 330 JOURNAL OF THE ARNOLD ARBORETUM [voL. XIV chiefly P. Strobus syntheses. The remainder were kindly cared for by Professor Melin until the fall of 1931, when the seedlings were placed in fixing solution and shipped to the authors in America. Except for the latter, the substrata in all flasks were tested for contaminations at the close of the experiment by placing sand from the flasks on malt agar media in culture tubes. Of thirty-five cultures of P. Strobus opened in November 1930, three were contaminated. A number of contaminations from seed-coat infections with P. resinosa were observed during the course of the experiment. Similar contaminations with P. Strobus did not occur, since these seeds were on moist agar for nearly three months before they were transferred to the flasks and, conse- quently, seed-coat infections were eliminated. Fixation was with Karpchinko solution; it did not prove particularly good for mycorrhizal details. The roots were embedded in paraffin for sectioning; both the butyl and ethyl alcohol series were used. Gross photographs of mycorrhizal roots were made in distilled water between glass plates. Microplanar or Tessar lenses, and Ilford panchromatic soft gradation plates with a Wratten B (red) filter were used. The staining technique employed will be reported by Dr. K. D. Doak in a future communication. RESULTS Typical ectotrophic mycorrhizae were form with P. Strobus by twelve of the fungi investigated: serail chrysorheus Fr., Plate 70, E deliciosus (L.) FE, “ 69, A; Plate 71, A Amanita muscaria (L.) Fr., : 69.455 *- Jie Boletus castaneus Bull., “70, I “ bovinus (L.) Fr., “ 70, F “luteus (L.) Fr., “ 69, B - ranulatus L., “ 70, H Boletinus porosus (Berk.) Peck, * 90. ies Plate 71, © M peer radicis nigrostrigosum “ (Picea) Abietis 1 Plate 71, B 7 (Pinus) Strobt 1 “ “ (Pinus) sylvestris 1 The mycelia of the M. r. atrovirens type, as in Melin’s experi- ments, overgrew the aérial parts of the seedlings and failed to exhibit any indications of mycorrhiza-formation (Plate 70, A). The mycelia 1P. resinosa grew very poorly in the medium used in these experiments. Root development was adequate for mycorrhiza-formation in only two or three cases. These will be reported elsewhere. 1933] HATCH, HYMENOMYCETES FORMING MYCORRHIZAE 331 of all other fungi failed to develop in the substrata, and information on their ability to form mycorrhizae was not obtained. DISCUSSION Positive results with Boletinus porosus adds Boletinus to the genera of fungi that have been proved to contain mycorrhiza-forming species. Lactarius chrysorheus and Boletus castaneus are also added to the list of known mycorrhizal organisms. The remaining fungi (with the ex- ception of those isolated from tree roots) have previously been tested in pure culture with success as follows: Lactarius deliciosus, with Pinus mugo Turra (P. montana Mill.), P. sylvestris and Picea Abies (Melin, 1924, 1925a). Amanita muscaria, with Betula pendula Roth, B. alba Roth, Larix decidua Mill. (L. europaea DC.), Pinus sylvestris, and Picea Abies (Melin, 1923a, 1925a). Boletus granulatus, with P. sylvestris and P. mugo (Melin, 1923a, Boletus bovinus, with Pinus densiflora Sieb. & Zucc. (Masui, Boletus luteus, with P. sylvestris, P. mugo, Larix decidua Mill., L. occidentalis Nutt. and Picea Abies (Melin, 1923a, 1923b, 1925a).* Concerning Amanita muscaria, Melin (1925a, p. 100) reports: ‘Der Pilz scheint aber den Pflanzen gegeniiber eine ziemlich hohe Virulenz gehabt zu haben.” A similar tendency was exhibited by our culture and by Lactarius chrysorheus in a 6-months synthesis as exemplified by the development of a heavy intercellular net. A culture plant removed after 15 months’ association with A. muscaria, on the other hand, showed normal infection (Plate 69, C). Concerning Boletus bovinus, Masui (1927) inoculated seedlings of Pinus densiflora growing in large test tubes on nutrient agar and reports that the seedlings were killed by this fungus. In Erlenmeyer flasks the fungus grew up the stem of the seedlings (1. c., p. 203 and Plate x1., Fig. 7). We note that in the latter photograph the plugs of the flasks 1Masui (1927) oe syntheses experiments between Pinus x hei ei oe Parl., uercus eases ian ., Q. phillyraeoides Gray, Q. glauca Thunb., Q. mongolica Fisch Sis ana "Rehd. & Wils., Q. paucidentata Franch., ’and a “fun us which i Ueoiaied from the tuberous (compound) mycorrhizae of the last named Oak believed chee ae zae were formed by the mycelium of a Boletus, the sporophore of which was attached to the mycorrhiza from which isolations were made. He listed ed & js) ~~ Nn ~~ -, SS = On an Ss ae S aS. B. Boletus granu latu iD} E. Boletus na F, G. H. 0 Mycelium radicis (Picea) Abietis 1, Plate 69. sible aces on roots of Pinus Strobus seedlings grown in re culture. A. Whole root system of seedling inoculated with Lactarius deliciosus. All short roots mycorrhizal, x 4. (Photo, U. S. Dept. Agric. WE aa formed by Boletus luteus, * 9. . Mycorrhizae for by Amanita muscaria, X 9. Plate 70. A. Seedling of Pas (eran inoculated with Mycelium radi- cis atrovirens 2, showing mycelial growth on stem and lower leaves, X 2/3. (Top of flask removed for photo.) B. Uninfected short roots of P. Strobus. Seventeen months old seedling, x 4. On ; ge of. P. resinosa inoculated with Boletus bovinus, ce mycelial growth over sand substratum, « 2/ OQ ie) ® o = the Royal Academy of Saeoedes Stockholm. Mycorrhizae of P. Strobus formed Lactarius chryso- rheus, X 6. (Photo, U. S. Dept. eho short roots of P. Se al with Boletus bovin Mee of P. Strobus formed with Bolctinus porosus, x 8. ea! ee 2 en Plate 71. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV H. Mycorrhizal roots of P. Strobus formed with Boletus granulatus, I. Mycorrhiza of P. Strobus formed with Boletus castaneus, “0. Photomicrographs of mycorrhizae of P. Strobus from syn- theses experiments, showing scion of mycelia between the cortical cells, and mantle structu nocul m—Lactarius deliciosus, panel oi sec- s 1 and nigrostrigosum, somewhat oblique, longitudinal ies 400. M.r. nigrostrigosum forming a secondary mantle over that of M. r. (P.) Abietts 1. Phtomicror adh by K. D. Doak. C. Inoculum—Boletinus porosus, medial, longitudinal section, 370 D. Inoculum—Amanita muscaria, oblique, transverse section, x 460. From 17-months old seedling ARNOLD ARBORETUM, HARVARD Forest AND DEPARTMENT OF BOTANY, Harvarp UNIVERSITY. Jour. ARNoLpD ARB. VoL. NIV. PLATE 68 SOME HyMENOMYCETES ForMING MyCORRHIZAE WITH PINUS STROBUS L. SoME HyMENOMYCETES FoRMING MyYCORRHIZAE WITH PINUS STROBUS Jour. ARNotp Ars. Vor. XIV. PLATE 70 SOME HYMENOMYCETES FoRMING MycorRHIZAE WITH PINUS Srrosus L. re Jour. Arnotp Ars. VoL. XIV. PLATE 71 Some HyYMENOMYCETES FoRMING MycoRRHIZAE WITH PINUS STROBUS L. JOURNAL OF THE ARNOLD ARBORETUM VoLUME XIV. OCTOBER, 1933 NuMBER 4 NOTES ON THE DISTRIBUTION OF WHITE SPRUCE AND BANKSIAN PINE IN NORTHWESTERN CANADA Hucu M. Raup With plates 72 and 73 THE WHITE Spruce, Picea glauca Voss (P. canadensis of most American authors) is the predominating forest tree throughout the central part of the Mackenzie basin, where its dark green spire-like tops give primary character to nearly every landscape. A _ traveler descending the Athabaska or Peace River gets the impression that the country is mantled with a heavy spruce forest, but this proves to be an illusion if short inland excursions are made, even if not more than a hundred yards from the river. A considerable part of the timber is found to be on the immediate banks of the rivers, while large areas in the flood plains through which the streams flow are covered with marsh lands separated by low natural levees which hold other narrow bands of timber. This lowland spruce forest is a nearly pure stand of tall, straight-boled trees, with a relatively thick undergrowth of Gray Willow, Salix Bebbiana, Mooseberry, Viburnum pauciflorum, Red-osier Dogwood, Cornus stolonifera, and a wild Rose, Rosa acicularis. There is a mat of woodland mosses on the ground in the older of these forests, but it is everywhere thin and in places almost non-existent. A prominent herbaceous species is Equisetum pratense, which often makes a continuous greensward on the forest floor. The lowland Spruce is common in suitable situations throughout the central part of the Mackenzie basin and far northward along the main streams. As would be expected it is but scantily developed eastward in the pre-Cambrian country where the thinness of the soils and the absence of large streams has precluded the formation of extensive alluvial deposits. A spruce timber which is closely related floristically to that of the 336 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV lowlands is found on the gently rolling surfaces of the upland country south and west of the Athabaska-Great Slave Lake region where it is interspersed with wide areas of Banksian Pine, Pinus Banksiana Lamb, It also occurs, although in a modified form, on the lower slopes of the northern Rocky Mountains in the upper Peace and Liard River basins. Where not disturbed for long periods this forest is a nearly pure stand of the Spruce in the more northern sections, while it is accompanied by Abies farther south, A. lasiocarpa in .the mountains and A. balsamea in the southern and southeastern sections. Toward the mountains Banksian Pine is replaced by the Lodge-pole Pine, P. contorta var. latifolia. Various forms of White Birch, Betula papyrifera and var. occidentalis, and B. neoalaskana, are of frequent though not dominant occurrence. The ground cover is usually a mat of woodland mosses with a very sparse herbaceous growth; and the shrub cover is also thin, consisting chiefly of Viburnum pauciflorum, Alnus crispa, Salix Beb- biana, Shepherdia canadensis, and Rosa acicularis. The whole differs from that of the lowlands mainly in having a thicker moss cover, rela- tively lighter shrub growth, and but few herbs. In the mountains the shrubs are more numerous in species and individuals, with the addition of Acer glabrum var. Douglasti and Oplopanax horridum. The highest and oldest land surfaces on the uplands are on the tops of the erosion plateaus which make the heights of land between the larger streams. The Caribou and Birch Mountains, and the Buffalo Head Hills are typical of these. Their forests have not been exten- sively investigated, but they seem to be characterized by Cordilleran elements. In the Lesser Slave Lake region they contain the Lodge-pole Pine and the Subalpine Fir, Abies lasiocarpa, while at the eastern edge of the Caribou Mountains the former has been found. Still another type of spruce forest occurs in isolated localities in the central part of the Mackenzie basin as well as in the northern Saskatche- wan River drainage of central Alberta. About 20 miles west of the upper Slave River, near the southeastern end of Lane Lake, is a highly morainic country composed largely of sand. On the tops of the high- est hills there are open park-like woods in which the individual spruce trees grow with their lower branches on the ground, and in which there is very little shrub growth. The ground cover is principally of lichens, Cetraria nivalis, C. tslandica, Cladonia alpestris, etc., and some mat- forming heaths, Arctostaphylos Uva-ursi, and Vaccinium Vitis-Idaea var. minus. Farther westward, near Moose (Eight) Lake, other sandy hills have a similar vegetation. Dr. E. H. Moss has described an iso- 1933] RAUP, WHITE SPRUCE AND BANKSIAN PINE 337 lated stand of Spruce in central Alberta near Edmonton! which occu- pies a similar habitat and seems both floristically and structurally similar to those farther north. It should be noted that these open stands of Spruce occur as “islands” in the ordinary mossy spruce woods described above, or in the dry pine woods. On the high morainic hills, in fact, the Pines occupy the lower slopes while the park-like Spruce is confined to the tops. Characteristic features of the northwest shores of Lake Athabaska are sand plains and ancient beaches which are many feet above the present level of the lake and form gentle slopes away from the rocky hills. The hills have a scraggly and stunted growth chiefly of scattered Jack Pines, while the plains have an open park-like timber of the same species mixed with Black Spruce, Picea mariana, and White Birch. The ground cover in the latter woods is of lichens, Cladonia alpestris, Cetraria nivalis, etc., and a few mats of Arctostaphylos Uva-ursi, Vac- cinium Vitis-Idaea var. minus, and Crowberry, Empetrum nigrum. A few taller shrubs are present, the most common being the Blueberry, Vaccinium canadense. A further examination of the inland country reveals that the White Spruce may occur in some abundance, but always on local flood plains where small quantities of alluvial silts have been sorted and deposited much as they are along the main rivers. Otherwise it is rather occasional in rock crevices and small soil accumu- lations in exposed situations. The upland pine woods west of the Slave River grow on sandy soils, and in their younger stages, as will be noted below, closely resemble those just described. If the observations are continued farther northeastward in the pre- Cambrian country to the eastern part of Great Slave Lake it is noted at once that the Pine has nearly disappeared from the forests, its place being taken almost entirely by the White Spruce. Rocky hills have stunted Spruces with gnarled and twisted trunks, and the sand plains such as are found at the eastern end of the lake have an open spruce wood which has the general aspect of the pine woods farther south.® The park-like timber is nearly identical with that of the high morainic hills referred to above. iMoss, E. H. e Vegetation of Alberta, iv. The Dae oe and Related Ve getation of Gel Alberta. (Jour. Ecol. xx. 412-3. 2) 2Raup, H. A Survey of the Ab pers of Shelter Point, Athabaska Lake. (Univ. of renee Bull. xxv. Oct. 1928.) 3Raup, H. M. The Vaopenits of the Fort Reliance Sand Plain. (Annals of the Carnegie Mus. xx. 9-38. 1930.) 338 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV To these facts must be added some recent findings on the north shore of Lake Athabaska. About 25 miles northeast of Chipewyan, Sand Point juts out into the lake to a distance of about 2 miles. It is com- posed entirely of sand, at least at the surface, and has its greatest elevations about a mile from its base. The eastern side of the point has been greatly modified by wind and wave action so that back of a long sand beach there is a steep bank of sliding sand in places about 50 feet high. The tops and upper western slopes of the sandy ridges are more stable and have an open spruce timber similar to that on Great Slave Lake. Other sandy ridges and plains nearer the base of the point are given over to the Pine. To recapitulate, there are three types of white spruce forests in the central part of the Mackenzie basin. One is of recent development on the flood plains of streams where all the processes in its formation upon newly exposed alluvial deposits may be seen. It is characterized by a rather thin moss cover on the ground, and by a relatively dense under- growth. A second is widely distributed on the lighter and better drained soils of the uplands, especially where it has not been much disturbed by fire or other agency. It has a much thicker moss mat (4-5 inches) than that of the flood plains, and a rather sparse under- growth. In the southern sections and in the mountains the Firs are associated with both of these types. The third type has an open, park-like stand of trees in which there is very little undergrowth and no mat of mosses on the ground. The place of the latter is taken by a lichen and heath mat. This type has a scattered distribution in the southwestern portions of the region where it is confined to the tops of high sandy ridges or exposed sandy promontories like Sand Point on Lake Athabaska. On Great Slave Lake, however, it is much more common, covering large areas of sand plain and rocky upland. It is in the form of the last type, with minor modifications, that the Spruce reaches its extreme limit in the arctic tundra.’ Random notes made by travelers northeast and east of the Atha- baska Lake country suggest a similar arrangement of forest types. J. B. Tyrrell described spruce woods along the Dubaunt and Kazan Rivers which must be like those seen on Great Slave Lake.’ Also, judging from his descriptions the northern limits of the Jack Pine on 1TyRRELL, J. B. on the Doobaunt, Kazan and Fergusin Rivers and the North- west Const . Madson Bay. (Canad. Geol. Surv. Ann. Rept. rx. 163F and 214F. 8.) shies T. The Arctic Prairies. New York, Charles Scribner’s Sons (1911). 1933] RAUP, WHITE SPRUCE AND BANKSIAN PINE 339 these streams are at Selwyn and Theitaga Lakes, respectively, between 50 and 100 miles south of the northern limit of timber. In another report’ Tyrrell has this interesting observation, ‘One small isolated grove of White Spruce was found on a high sandy island in Hatchet Lake, standing out conspicuously in the midst of the surrounding forest of small Black Spruce.” Hatchet Lake is about 100 miles southeast of the eastern end of Lake Athabaska, and far into the range of the Pine, so that we may look upon this grove as another isolated southern stand of the third type similar to those described above. There have been striking changes in the land forms of the region in post-Pleistocene time. Topographic evidence indicates that. the ice front retreated from southwest to northeast across the country, holding impounded for considerable periods the waters of the Athabaska and Peace Rivers to form a series of post-Glacial lakes which were pro- gressively smaller, lower in elevation, and more recent, to the north- eastward.” Morainic deposits left by the melting ice included high ridges which must have stood out as islands in the lower lakes. Thus we may establish the soil surfaces of the sandy hill tops west of the Slave River as considerably older than those of their immediate sur- roundings, and they must have been available for plant cover long before these lower levels. The topographic history of the hill tops noted by Moss in central Alberta has not been studied. With these topographic changes there must have occurred a climate gradually shifting from one similar to that of the present arctic tundra to what we have now. So far as we know there has been no change other than one of amelioration.* The thick spruce woods, then, are on somewhat younger soils of the uplands, and on the still younger silts of the flood plains formed in the last of the lake bottoms. The open spruce woods west of the Slave River are on the oldest hill tops. The Great Slave Lake woods, how- ever, as well as those of Sand Point, though closely related to the ones Sota ae B., assisted by Dowling, D. B. Report on the country between Atha- bas and Churchill River with notes on two routes traveled between the enters ae Saskatchewan Rivers. (Canad. Geol. Surv. Ann. Rept. vim. 12. 1896.) 2CAMERO sey eta Lakes in the rages River Basin, Northwest Territories Cana (Jour. of Geol. xxx. 337-53. M. eo and affinities of i ee of the Athabaska— Gee Slave hake Reson (Rhodora, xxx. 187-208. 1930.) elas: i B. Ciges of climate in rahe western Canada since the glacial iod. e Veranderungen des Klimas sei m Maximum der Letzten Eiszeit. (ith pit Geol. Congress, Stockholm monk p. 389-91.) Raup, H. M. Botanical investigations in the Wood Buffalo Park. In prep. 340 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV on the old hills west of the Slave River, are on very recently formed lake shore sand plains and beach ridges. This indicates that the lengths of time since their probable origins will not of itself account for their nature or distribution since they occupy some of the oldest and youngest surfaces Elevation above the general land surface, with consequent local modification of climatic and edaphic factors may have something to do with the occurrence of this forest on the high ridges west of the Slave River and in central Alberta. But mere elevation appears ineffective since the type also grows very near the level of Lake Athabaska at Sand Point, and on Great Slave Lake. In the latter region the climate in general, as far as it is indicated by the vegetation, is more arctic in character than near the western end of Lake Athabaska, so that we might expect some sort of change in the forests. As noted above, this change is shown by the elimination of the Jack Pine to the north- eastward. The park-like spruce woods at Sand Point and west of the Slave River, although differing widely in the age and elevation of their soils, have this in common—that they are in places which are exposed to cold storm winds, mostly from the north and east, and that their soils are dry and sandy. It is probable that woods of this type in central Alberta may be described in the same way. Similar soils not so exposed have jack pine or mossy spruce woods, but the question of what any other type of soil would produce in such exposed places needs further study. The only evidence we have in this connection is on isolated granite hills in the Athabaska-Peace delta where small patches of clayey soils of early lacustrine origin have natural prairie on them. Undoubtedly a study of the various habitats in the so-called “Barren Lands” or “Arctic Prairies” would throw some light on the matter. H. E. Pulling, basing his conclusions upon field investigations of roots in north-central Manitoba, suggests that an important factor determining the northern limits of various trees is the relation between the root habits and the permanently frozen condition of the subarctic soils! He classes the Black Spruce, Tamarac, and White Birch as having a rigid shallow root habit, White Spruce as having a flexible shallow root habit, Black Poplar a deep flexible habit, and Jack Pine a deep rigid one. Northward, of course, soils are increasingly shallow due to the frozen subsoil. In general the distribution of trees in the 1Puttinc, H. E. Root habit and plant distribution in the Far North. (Plant World, xx1. 223-33. 1918.) 1933] RAUP, WHITE SPRUCE AND BANKSIAN PINE 341 central part of the Mackenzie basin may be correlated with these find- ings. White Spruce, with a shallow flexible root habit, is most success- ful in coping with subarctic conditions, and extends far out on the Barren Lands in favored places. Other shallow rooted species, but having a rather inflexible system—Birch, Tamarac, and Black Spruce— have a wide range in the Mackenzie basin but are nowhere so tolerant, with the possible exception of the Birch, as the White Spruce. Black Poplar and Jack Pine, with deep root systems, have somewhat similar ranges. The former, like the Pine, has its northeastern limit near the eastern end of Great Slave Lake. Another species, White Pine, also listed by Pulling as having a deep rigid root habit is entirely absent from the region so far as is known. On the other hand, if this relationship is everywhere of first im- portance, we should expect to find, for instance, that the sandy soils on Sand Point, on the high moraines west of the Slave River, and on some of the hills in central Alberta, differ from those very near them in being frozen nearer the surface. Although no investigations have been made in these places, it does not seem probable that they would show the sharp differences. Great variations in elevation such as occur between the Slave River and the Caribou Mountains are known to be accompanied by frost differences in the soils, but these are not to be expected, for example, on a low sandy lake shore like Sand Point. Soil frost, therefore, although it is probably of first importance in the gen- eral distribution of the species in question, may not account for all the facts relating to the occurrence of the isolated stands of open Spruce; and we must conclude that exposure to storm winds may be a prime factor. The southern occurrences of northern park-like Spruce may be looked upon as remnants of an earlier forest which occupied the sandy soils throughout the region when a slightly more arctic climate, similar to that now prevailing about the eastern part of Great Slave Lake, was general farther south. There are vast areas of sandy uplands, now covered with jack pine timber, whose soils show no evidence of the humus accumulations which would have arisen from an earlier, heavier timber with a moss mat. These areas must have had an open spruce timber with a thin lichen mat during a somewhat more arctic climate, since the Jack Pine, as indicated by its present range, would then have had its northern limit farther south. These facts suggest that the Jack Pine is the most recent arrival among forest trees in this part of the Mackenzie basin, and that it has replaced the Spruce except in situa- tions excessively exposed. 342 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV The mossy spruce woods on the uplands west of the Slave River appear to be developing from jack pine stands by stages which are readily seen. Young groves of the latter are of close-growing, rather spindling trees, but later they thin out and produce much larger crowns with arching branches. During this stage young and vigorous Spruces come up in the shade of the Pines, accompanied by the beginnings of the moss mat. Later stages show a predominating spruce stand in which there are very old, scattered, and dying Pines festooned with lichens, and with small bunches of green needles at the tips of their long branches. How the Pines invaded the country in the first place, crowding out a park-like spruce forest as suggested above, is uncertain. Study of the country between Athabaska and Great Slave Lakes would undoubt- edly reveal a series of intermediate stages which would indicate the transition. It is probable that Pines would invade the open spruce forest in small numbers at first, not enough to make much shade, but sufficient to crowd out the older Spruces and prevent the growth of younger ones. As the climate ameliorated we may suppose that the Pines became more abundant, with such changes in soil and general conditions that Spruces could again invade and form the mossy woods so common now on the more southern uplands. Among the areas thus far studied by the writer those of Lake Athabaska come nearest to illustrating these intermediate stages. As one passes along Sand Point toward the base, the change from open spruce to very open jack pine woods in which many of the trees are old is a gradual one. Still farther inland the Pines are in a closer stand. The Lake Athabaska pine woods, being farther to the northeastward, nearer the limit of range, and occurring on or near the windswept mar- gins of the lake, are growing under more difficult conditions than those west of the Slave River, and could therefore be considered, as a whole, nearer to an original condition. This is borne out by the above- mentioned gradual transition from old spruce to old pine woods on Sand Point, and by the fact that so far no place has been found where the sand plain pine woods are developing into mossy spruce woods as they are on the uplands. Whether this is due to youth or exposure, or both, is uncertain. They are surely growing upon much younger land surfaces. Furthermore, being near the northeastern limit of its range, the Pine is undoubtedly more selective in its choice of habitats than farther south, and more circumscribed in its ability to invade other associations. 1933] RAUP, WHITE SPRUCE AND BANKSIAN PINE 343 Burning probably aids the spread of pine woods by opening other timber associations, but the normal course. of development on the up- lands to spruce woods seems to be only retarded by this means. Whether the same agency furthered the replacement of park-like spruce woods by the Pine is uncertain. If it were a very important factor we should expect that the sandy hill top Spruce of the uplands would long ago have been removed. But since it has not we must look for other causes for the present relationships. SUMMARY In the distribution of the existing pine and spruce forests of the central part of the Mackenzie basin there is evidence that the earliest post-Glacial timber was one of White Spruce arranged in an open asso- ciation on sandy soils. The ground cover in this timber was probably a thin mat of lichens and heaths; and the whole persists to the present time as a well-developed type near the arctic timber line. With the amelioration of the climate after the recession of the glacier, other less tolerant species could migrate into the country, among them the Jack Pine. The Pine could successfully compete with the Spruce on the wide-spread sterile sandy soils, and soon formed an open association resembling in aspect that of the Spruce. It is probably typified now by the sand plain pine woods of the Athabaska Lake district. An increasing growing season, more available moisture, and more rapid decay then permitted the accumulation of organic materials in the soils at a faster rate than had been possible before. This inaugu- rated a forest succession on the higher and relatively older land sur- faces southwest of Great Slave Lake. The richest woods growing in the region now are of White Spruce with a moss mat and light under- growth; and they appear to develop directly from mature jack pine woods, although in many places burning and the invasion of Trembling Aspen and Black Poplar have prolonged the process. As noted else- where by the writer,’ there is floristic as we]l as topographic evidence of the relative immaturity of even these richer spruce forests. Many species of wide range across the continent in the coniferous forests are absent or very rare in the Athabaska-Great Slave Lake region, and suggest that the latter have not had sufficient time under the existing subarctic conditions to develop further. The most recently formed spruce woods are on local flood plain 1Raup, H. M. See p. 339, foot-note 2. 344 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV deposits along the present streams. They resemble the upland mossy Spruce but have a thinner ground cover and relatively heavier under- growth. In a few local areas far to the south and southwest of the present park-like spruce forest there are woods which appear to be relics of the former wide distribution of this type. They are on both old and very young land surfaces, at both high and low elevations, but are all in places unusually exposed to storm winds. It is supposed that the Jack Pine has not been able to invade the original Spruce in these exposed situations. EXPLANATION OF THE PLATES Plate 72. Lerr: Lowland spruce forest along Peace oi near Peace Point. Note ground cover of eh outgy prate Ricut: Upland spruce oa about 4 torsfi west of the upper Slave River, near Pine’ Lake.! Plate 73. Upper: Open spruce woods on sand plain at the eastern end of Great Slave Lake (Ft. Reliance). Lower: Open woods of Pinus Banksiana, Picea mariana, and Betula neoalaskana near Shelter Point, Lake Athabaska. HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1The writer is indebted to the National Museum of Canada for use of the two photographs in this plate. PLATE 72 CIV, Jour. ARNoLtp ARB. VOL. HITE SPRUCE AND BANKSIAN PINE r / DISTRIBUTION OF W 73 PLATE XIV. VOL. ARNOLD ARB. JOUR. + See eee epee ee pie ane | AND BANKSIAN PINE SPRUCE W HITE OF DISTRIBUTION 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 345 NEW SPECIES, VARIETIES AND COMBINATIONS FROM THE HERBARIUM AND THE COLLECTIONS OF THE ARNOLD ARBORETUM! ALFRED REHDER With plate 74 Ilex ficoidea Hemsl. in Jour. Linn. Soc. xxut. 116 (1886). — Loesener in Nov. Act. Leop.-Carol. Akad. Naturf. txxvimr, 328 (Mon- og. Aquifol.) (1901). Symplocos tetramera Rehder in Sargent, Pl. Wilson. 11. 598 (1916. — Chung in Mem. Sci. Soc. China, 1. 211 (Cat. Trees Shrubs, China) (1924). — Synon. nov. Cuina. Kwangtung: Hongkong, C. Ford (? isotype of I. ficoidea) ; same locality, without collector, Feb. 1892; Lantao, Herb. Hongk. no. 8251; Loh fan shan, C. O. Levine, nos. 606 and 1567, Oct. 27-30, 1916 and Aug. 15, 1917; Lung tou shan, S. P. Ko, no. 50283, Apr. 2, 1930. Fukien: in woods and open thickets, R. C. Ching, nos. 2227 and 2231, Aug. 3, 1924. Chekiang: 50 li north of Sia-chu, R. C. Ching, no. 1646, May 24, 1924; 130 li of Wen-chou, R. C. Ching, nos. 1848 and 1862, June 17, 1924; -80 li northeast of Tai- suan, R. C. Ching, no. 2206, July 22, 1924; Hang chow, T. Tang & W.Y. Hsia, no. 365, July 18, 1927; Tai-pai-shan, Y. L. Keng, no. 1145, Aug. 24, 1927. Anhwei: Wu Yuan, K. Ling, no. 7858, Aug. 24, 1924. Hunan: in monte Yun-schan prope urbem Wukang, Wang- Te-Hui, Handel-Mazzetti, no. 12810, Apr. 1919; prope urbem Tchang- scha, Handel-Mazzetti, no. 11609, Apr. 13, 1918. Yunnan: Puerh, A. Henry, no. 13273 (holotype of Symplocos tetramera): be- tween Muang hing and Szemao, J. F. Rock, no. 2740, March 2-12, 1922 Ilex ficoidea resembles in its general aspect and particularly in its leaves and in the axillary clusters of its flowers so closely certain species of Symplocos, as S. anomala Brand, S. congesta Benth. and S. crassi- folia Brand, that also other botanists besides myself had placed in the herbarium specimens of this /iex under the genus Symplocos. When 1Continued from p. 222. 346 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV describing Henry’s no. 13273 as Symplocos tetramera I explained the different shape of the pistil by the assumption that the ovary was rudi- mentary and the unusual reduction of the number of stamens by refer- ence to some South American species of Symplocos with only four stamens. Ilex ficoidea has not yet been recorded from western China and in central China only from Changsha, Hunan (Handel-Mazzetti, no. 11609) and from Wukang, Hunan (Handel-Mazzetti, no. 12810, as var. brachyphylla Hand.-Mazz.). Henry’s no. 13273 differs somewhat from typical J. ficoidea in the shorter and stouter petiole, about 5 mm. long, and in the shorter acumen of the leaves. Rock’s no. 2740 which also has been distributed as Symplocos tetramera, is similar but the leaves are comparatively narrower and generally oblong and of thinner texture. Acer sikkimense Mig. var. serrulatum Pax in Bot. Jahrb. vi. 215 (1886); in Engler, Pflanzenr. 1v-163, p. 34 (1902). — Rehder in Sar- gent, Trees & Shrubs, 1. 180 (1905). — Fang in Contrib. Biol. Lab. Sci. Soc. China, vir. 178 (1932). Cura. Yunnan: Feng-cheng-lin Mt., forests, south of Red River, alt. 7000 ft., A. Henry, no. 10640 (tree 10 ft.); west of Talifu, Mekong watershed, en route to Youngchang and Tengyueh, J. F. Rock, no. 6834, Sept.-Oct. 1922; Shweli river drainage basin, environs of Tengyueh, J. F. Rock, no. 8014, Feb. 1923; without precise locality, G. Forrest, nos. 9759 and 26233. Himataya: Bengal, Griffith, no. 936 (holotype; not seen). Acer sikkimense agrees with A. Davidi in having oblong-ovate leaves without lobes, but differs in their caudate-acuminate apex, subcordate base, glabrous under side, entire or nearly entire margin and in the shorter pedicels and longer racemes. In the variety the leaves are closely and sharply serrulate. In my account of the Maples of the sect. Macrantha (pp. 211-222) I unfortunately overlooked this Chinese variety of an Himalayan spe- cies. In the key it should precede A. Davidi from which it is easily distinguished by the closely and finely serrulate margin of the leaves and their glabrous under side which in A. Davidii is more or less rufous- tomentose along the veins at least when young. Acer Davidi Franchet in Nouv. Arch. Mus. Paris, ser. 2, vit. 212 (Pl. David. 1m. 30) (1884). — Fang in Contrib. Biol. Lab. Sci. Soc. 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS 347 China, vii. 177 (1932). — Rehder in Jour. Arnold Arb. xiv. 213 (1933). — Add the following synonym: Acer laxiflorum Pax var. B ningpoense Pax in Engler, Pflanzenr. 1v.-163, p. 36 (1902). — Rehder in Sargent, Trees & Shrubs, 1. 180 (1905). Cuina. Chekiang: Ningpo-Berge, E. Faber, in 1886 (holo- type in Herb. Berol.; photo. and fragments in A. A.). In my account of the Chinese Maples of the Macrantha section in the last number of this Journal I had omitted A. laxiflorum var. ning- poense, because I had not seen the type and doubted very much the possibility that a variety of the western A. laxiflorum should occur near Ningpo. I have now received through the kindness of Professor R. Pilger of the Botanical Museum at Berlin-Dahlem an excellent photo- graph and fragments of the type specimen which proves that the speci- men does not belong to A. laxiflorum but is a slight variation of A. Davidi with indistinctly lobulate leaves and horizontal wings of the fruit. It differs from A. laxiflorum in the ovate-oblong leaves not 3—5- nerved at the base and abruptly narrowed into a remotely serrulate acumen entire toward the tip, in the irregular crenate-dentate serra- tion, and the practically glabrous under side, while in typical A. laxi- florum the leaves are sharply serrulate with acuminulate teeth, dis- tinctly 3-lobed and 3—S-nerved at the very base, with an elongated triangular-ovate middle lobe gradually narrowed in a closely and sharply serrulate caudate acumen, and are even at maturity more or less rusty pubescent on the veins. From slightly lobed forms of A. Grosseri Pax Faber’s Ningpo specimen differs chiefly in the oblong or oblong-ovate shape of the leaf, its crenate-dentate serration and the not clearly 3-nerved base. Rhamnus crenata S. & Z. var. discolor, var. nov. A typo recedit foliis subtus albido-tomentosis, nervis utrinsecus 8-12. — Frutex; folia elliptica vel obovato-oblonga ad oblongo- oblanceolata, subito acuminata, basi late vel sensim cuneata, circiter 8 X 4to 10 X 3 Cuina. Chekiang: open thickets, alt. 4400 ft., R. C. Ching, no. 2536 (type), Aug. 31, 1924 (shrub 15 ft.); southern Chekiang, open thickets, alt. 4200 ft., R. C. Ching, no. 2461, Aug. 24, 1924 (shrub 8 ft.). On account of the whitish tomentose under side of the leaves this variety looks quite different from the type, but I can find no other characters to separate it from R. crenata which shows considerable 348 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV variation in the shape and also in the pubescence of its leaves, except that the leaves have about 8—12 pairs of lateral veins while in typical R. crenata they have only 5—9 pairs. The two specimens cited above differ considerably in the shape of the leaves: no. 2536 has generally elliptic leaves about 8 cm. long and 4 cm. broad, broad cuneate at base, while no. 2461 has narrow-oblong to oblong-oblanceolate leaves 8—10 cm. long and 2.5—3 cm. broad, cuneate at base, but one leaf is oblong-obovate 9 cm. long and 4.5 cm. broad. Rhamnus utilis Dene. var. hypochrysa (Schneid.), var. nov. ceri crenatus FE. Pritzel in Bot. Jahrb. xx1x. 460 (1900), pro rte. — Non Sieb. & Zucc. as ag get ee C. K. Schneider in Notizbl. Bot. Gart. Mus. Berlin, v. 76 (1908) ; Ill. Laubholzk. 11. ae fig. 198 p-q, 199 o-p (1909) ; in Gree Pl. Wilson. 11. 252 (1914). Cuina. Szechuan: Nanchuan, A. von Rosthorn, no. 1585 (syntype of R. Aypochrysus in Herb. Berlin); Nanchuan Hsien, W. P. Fang, no. 1415, June 6, 1926. Shensi: Lin-hua-zao near Kin-lin- san, G. Giraldi, nos. 931, 932; Fu-kio, G. Giraldi, no. 940 (syntypes of R. hypochrysus in Herb. Berlin). Northern Honan: Yung- ning, Tsi-li-ping, alt. 1000 m., J. Hers, no. 1368, Sept. 30, 1919; Yung- ning, Yo-tze-ping, alt. 900 m., 7, Hers, no. 820, Oct. 4,1919.Hopei: Tang-san, F. N. Meyer baud, only); plants raised from these seeds distributed as U. S. Dept. Agr. no. 17909 in Arnold Arb. coll. 1909 and 1914 (in part)k Fukien: Enghok Hsien, H. H. Chung, no. 1342, April 14, 1923; Diongloh, Ne-lan-san, Herb. Fukien Christ. Coll. no. 11729, Aug. 6, 1926. This variety differs from typical R. utilis chiefly in the pubescent branchlets and pubescent leaves. In the type specimens of R. hypo- chrysus and in Fang’s no. 1415 the pubescence is rather dense and dis- tinctly yellow, while in the specimens from Honan the pubescence is somewhat slighter and not or scarcely yellow. In the Fukien speci- mens which have rather small leaves, the yellowish pubescence is still slighter and the branchlets are only very sparingly pilose or nearly glabrous, so that these specimens closely connect typical R. utilis with the variety. From the seed collected by F. N. Meyer near Tang-san, which is apparently the same as Tang-shan between Tien-tsin and Yung-ping-fu, and distributed from the U. S. Department of Agricul- ture under no. 17909, both forms, the glabrous and the pubescent with distinctly villous branchlets, were raised. It, therefore, does not seem possible to maintain R. kypochrysa as a distinct species. 1933] REHDER, NEW SPECIES, VARIETIES AND COMBINATIONS — 349 Though Decaisne describes his species as with “foliis . . . subtus puberulis,” almost all the numerous specimens in this herbarium except those cited above, have the leaves nearly or quite glabrous. A speci- men, however, collected by C. Schneider in 1903 in the nursery of Simon-Louis at Plantiéres near Metz from a plant marked “R. utilis, original from Decaisne”’ has the leaves loosely and yellowish pubescent on the under side and when young also slightly so above, but the branchlets are glabrous; this agrees exactly with Decaisne’s de- scription. It, therefore, appears that the type of R. utilis is the form with only slightly pubescent leaves but glabrous branchlets, and that the form with the leaves quite glabrous or slightly pubescent only on the veins beneath, may be considered a distinct form and distinguished as R. utilis f. glabra, forma nova, of which E. H. Wilson, no. 622, from Fang Hsien, Hupeh, may be designated as the type. Ampelopsis Delavayana Planchon in De Candolle Monog. Phaner. v. 458 (1887). Amelopsis aconitifolia Bge. y (See Diels & Gilg in Bot. Jahrb. xxix. 465 (1900). — Syn Cuina. Szechuan: esate Bock & von Rosthorn, no. 1540 (holotype of A. aconitifolia var. tomentella; photo. in A. A.; flowering branch). Ampelopsis aconitifolia {. tomentella the type of which I have seen in the herbarium of the Botanic Museum at Oslo, though superficially quite similar to A. aconitifolia {. setulosa Diels & Gilg, agrees with A. Delavayana Planch. in the occasional presence of undivided only slightly lobed leaves, the shallow serration of the leaflets and the rather short peduncles about as long as the cyme itself. Ampelopsis aconitifolia Bge. never has undivided leaves, the leaflets are rather deeply serrate to pinnatifid and the loose and smaller cymes are borne on slender peduncles longer than the cyme. In fruit the two species differ considerably in the color of the berries; the pubescence varies in the two species and glabrous and pilose forms occur in both. The other forms described (1. c.) by Diels and Gilg: f. glabra, f. setulosa (pr. parte typ.) and f. cuneata belong without doubt to the variable A. aconitifolia. Wilson’s specimen no. 1070 (Veitch Exped.) placed in the Berlin Herbarium with Ampelopsis aconitifolia {. tomentella belongs to Vitis Piasezkii Maxim. To V. Piasezkii belongs also Henry, no. 6479, one of the syntypes of Ampelopsis aconitifolia {. setulosa. 350 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Agapetes stenantha, spec. nov. Plate 74 Frutex vel arbor parva, glabra, ramis cinereo-fuscis lenticellatis. Folia coriacea, lanceolata, 6.5—9 cm. longa et 2—2.4 cm. lata, longe acuminata, basi in petiolum brevissimum 1—2 mm. longum attenuata, integra, supra laete viridia, subtus pallidiora, costa media utrinque elevata, nervis utrinsecus 6-7 supra fere obsoletis subtus in sicco lev- issime elevatis. Racemi axillares e gemmis nudis in ramulis biennibus vel vetustioribus, 5—8 cm. longi, multiflori (floribus circiter 18—25) ; Flores rubescentes, graciliter pedicellati pedicellis 5—8 mm. longis, bractea lanceolata circ. 2 mm. longa rubescente membranacea decidua suffultis, basi bracteolis 2 bracteae similibus sed saepius paullo minori- bus munitis, apice articulatis; dentes calycinis triangulari-lanceolati, circ. 3 mm. longi, dorso costati costis in tubum ovarii fere aequilongum decurrentibus; corolla anguste cylindrica apicem versus leviter angus- tata, circ. 2 cm. longa et medio 3 mm. lata, subalato-costata, lobis linearibus 1 mm. latis apice obtusiusculo mucrone reflexo munitis, valde inaequalibus, sinu altissimo circiter mediam corollam attingente; stamina stylo paullo breviora supérne in tubum stylum arcte cingentem cohaerentia, antheris pilosulis 5 mm. longis apice in tubum gracilem circ. 1 cm. longum attenuatis, basi in calcar pilosulum, 1.5 mm. longum filamentum glabrum fere aequans productis; stylus glaber, corollam aequans, apice clavatus; ovarium 2-2.5 cm. longum, glabrum. Burma: between Sadon and the Yunnan Chinese border at Changti- fang and Kambaiti, alt. 7600 ft., J. F. Rock, no. 7514, Nov. 1922 (shrub or small tree, flowers bright reddish, very ornamental). This distinct new species is characterized by the very narrow tubu- lar flowers about 2 cm. long and only 3 mm. wide, borne in many- flowered elongated racemes and by the lanceolate long-acuminate leaves attenuate at base and 6.5—9 cm. long. It differs in the elongated many-flowered racemes from all Asiatic species except A. vaccinioides Dunn which is easily distinguished by the short campanulate corolla and the larger leaves rounded or subcordate at base, and except A. corallina Cowan which differs in the lanceolate corolla-lobes and appar- ently glabrous anthers without spur. (To be continued) HERBARIUM, ARNOLD ARBORETUM, HARVARD UNIVERSITY. Jour. ArNotp Ars. VoL. XIV. PLATE 74 Ye ~ F,; a’ \ Sh WY) ‘ ¢ PLANTS OF BURMA Qaa pees AVenm..Ol, Sakd ~ ow F ian arene weed ern “ a Panter 6 1 Ramagrenees: Pooks 28 chromosomes. FIGURE. 2. KADSURA JAPONICA ini 405 os su640e00800%: 28 chromosomes. Figure 3. TROCHODENDRON ARALIOIDES ..........0.. 38 chromosomes. FicurRE 4. TETRACENTRON SINENSE ......e0ceeeee005 38 chromosomes. FicureE 1. has been drawn from a root tip smear. The remainder are from permanent preparations of root tip material. > 2100. 380 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV alike and are considerably larger than those of Drimys and other mem- bers of the group having 19 as the basic number. Cytologically, Illicitum is clearly distinct and bears no resemblance to either the members of the Magnoliaceae or Drimys, supposedly its closest allies. Morinaga, et al. (1929) have listed the reduced chromosome number of Illictum anisatum as 14. The figures given by these workers, of the chromosomes during metaphase of the first division compare favorably in size and shape with those of Schisandra sphenanthera (P\. 80, fig. 10) Drimys Winteri has about 76 chromosomes in somatic cells. Stras- burger (1905) has stated there are about 36 pairs of chromosomes in this species. Because of the large number of chromosomes and their small size, it is difficult to make absolutely certain of the number. However, it is undoubtedly between 72 and 76, with greater likelihood of the latter figure’s being correct. In Schisandra sphenanthera the haploid chromosome number is 14 (Pl. 80, fig. 10). The chromosome complement of this species has no affinity with that of the Magnolias or Liriodendron. The chromosomes are larger than those found in the Magnoliaceae, and there is a tend- ency toward irregularity in the time of division,—i. e., precocious division of one or two bivalents. This does not seem to aifect the per- centage of good pollen (94%). Kadsura is a genus of climbing vines, closely allied to Schisandra. Kadsura japonica, the only species which has been examined, has 28 chromosomes (Text fig. 2). The counts were made from permanent root tip preparations. The chromosomes of this species resemble those of Schisandra and other representatives of the 14 chromosome group. Trochodendron aralioides has a somatic chromosome number of 38 (Text fig. 3). Cytologically, its relations are with Magnolia and Lirio- dendron rather than with Euptelea, with which it has been placed by Hutchinson (1921). Euptelea polyandra has 14 pairs of chromosomes (PI. 80, fig. 9). Some secondary pairing and irregularity in the time of division, simi- lar to that found in Schisandra sphenanthera, has been noted in this species. Cercidiphyllum japonicum, the representative of the monotypic fam- ily Cercidiphyllaceae has 19 chromosomes (PI. 80, fig. 3). The figure indicates that the chromosome size and shape are of the same general nature as those of Liriodendron and Magnolia. The chiasma fre- quency corresponds closely to that found in these two genera. 1933] WHITAKER, CHROMOSOME NUMBER IN MAGNOLIALES 381 Tetracentron sinense has 38 chromosomes in the root tip cells (Text fig. 4). Judging by its cytological characteristics, this genus is allied to the Magnoliales. Its chromosome number definitely places it with Magnolia, Liriodendron, Trochodendron, etc. DISCUSSION Engler and Prantl’s system (Die natiirlichen Pflanzenfamilien) has been criticized on the ground that the Amentiferae are regarded as a more primitive group than such polypetalous families as Magnoliaceae and Ranunculaceae. From morphological evidence, this criticism seems justified. Maneval (1914) has compiled a list of characters which are generally conceded to be more or less indicative of a primi- tive condition. These characters are all found in the Magnoliaceae. Sinnott (1914) has used nodal anatomy as a means of indicating phylo- genetic relationships. He states: ‘“The Magnoliaceae are perhaps more variable in nodal structure than any other family of dicotyledons.”’ The Magnoliaceae, as used here, include, in addition to the Magnolias and Liriodendron, the Schisandraceae and Winteraceae of Hutchinson (1926). This variability in nodal structure is precisely what one would expect and what is actually found among primitive types. If one accepts the customary thesis that the Magnoliales are a prim- itive group of dicotelydons, one would expect to find, among the fam- ilies of this order, several lines of cytological development. Thus far the evidence from chromosome numbers and other cytological features indicates that there are at least two such lines. The first comprises Magnolia, Liriodendron, Drimys, Trochodendron, Tetracentron, and Cercidiphyllum. In this group the chromosomes are characteristically small, short rods with a basic number of 19 pairs. In the second group occur Euptelea, Schisandra, Kadsura and Illicium, with 14 as the basic chromosome number. The chromosome size, shape, and configurations found in this group bear considerable resemblance to one another. In Table II an attempt has been made to compare chromosome number, type of nodal anatomy, and several other anatomical fea- tures in the genera under investigation. Perhaps the most important point brought out by this table is the correlation existing between chromosome number and type of nodal anatomy. In Group I are found genera with two types of nodal anatomy: (a) 3 traces from 3 gaps (Tetracentron, Drimys, Cercidiphyllum); (b) « traces from co gaps (Trochodendron, Magnolia, Liriodendron). The second cate- TABLE II A COMPARISON OF CHROMOSOME NUMBER, NODAL ANATOMY, ETC. JN GENERA OF THE MAGNOLIALES? vessels vessels genus leaf present or | scalariform | tracheids vena absent or porous Type of Tetracentron palmate absent scalariform nodal anatomy Drimys pinnate absent rarely 3 traces from 3 gaps scalariform GROUP I Cercidiphyllum palmate present scalariform basic chromosome : number—19 Trochodendron | pinnate absent scalariform oo ees Magnolia pinnate present scalariform co traces from oo gaps wc cadlatae Liriodendron pinnate present scalariform or porous Illicium pinnate present scalariform cial ew Schisandra pinnate present scalariform 1 trace from 1 gap ; ad eee GROUP II Kadsura pinnate present scalariform or porous basic chromosome number— Type of odal anat Euptelea pinnate present | scalariform 3-7 traces from 1 gap 1Morphological data supplied by I. W. Bailey. Z8E WOLANO0dUV GIONYV AHL AO TYNYNOL AIX “IOA] 1933] WHITAKER, CHROMOSOME NUMBER IN MAGNOLIALES 383 gory is essentially a multiplication of the fundamental condition of 3 traces from 3 gaps. In the 14-chromosome group two types of nodal anatomy are found; (a) 1 trace from one gap (Jllicium, Kadsura, Schisandra); (b) 3—7 traces from 1 gap (Euptelea). The latter condition simply represents a multiplication in the number of traces of the basic condition of 1 trace from 1 gap. The value of nodal anatomy as an aid in classification of angio- sperms and as a means of detecting phylogenetic relationships, has been fully discussed by Sinnott (1914). It is sufficient, for the present pur- pose, to note that it is extremely constant within most families and is of considerable value in determining broad lines of relationship. The fact that there is a close agreement between chromosome number and nodal anatomy is reasonably strong evidence that the genera placed together in Table II form a natural grouping which should be con- sidered in any future classification of this group. It may be well to point out, at this juncture, the significance attached to a basic number of 19 chromosomes. This is an unusual number in the plant kingdom and is very rarely met. It has not been reported in any of the families which may be considered as related to the Magno- liales. If the chromosome number had been 12 or some other number of common occurrence, it would not carry nearly the weight which it does in the present case. It-is difficult to account for the occurrence of this unusual chromosome number in Group I unless we assume that the genera composing this group are made up of closely related forms. McLaughlin (1933) has recently made a very thorough study of the systematic anatomy of the woods of the Magnoliales. In his suggested classification of the group he has eliminated Cercidiphyllum, Euptelea and IJllicium by placing them in the Hamamelidales rather than in the Hamamelidaceae. Cytologically, there is no evidence to justify this step. Professor Sax informs me that the Hamamelidaceae have 12 rather characteristically small, short chromosomes (unpublished data). Illicium and Euptelea, with 14 chromosomes, and Cercidiphyllum, with 19 chromosomes, would not fit into this proposed classification. The placing of Drimys and Illicitum in the same family, as proposed by Hutchinson (1921), can hardly be sanctioned in view of the present evidence. These two genera are very dissimilar cytologically. The same criticism can be made in regard to placing Trochodendron and Euptelea in the same family, as this author has suggested (Hutchinson, 1921 384 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV SUMMARY The evidence shows that in the Magnoliales thus far investigated there are two lines of cytological development. Group I has a basic chromosome number of 19. The chromosomes of this group are char- acteristically small, short rods. In this group the following genera occur: Magnolia, Liriodendron, Cercidiphyllum, Drimys, Trochoden- dron, and Tetracentron. Group II has a basic number of 14 chromo- somes. In this group the chromosomes are much larger than those of Group I. The following genera occur in Group II: Jllicitum, Schisan- dra, Kadsura, and Euptelea. There seems to be a correlation between basic chromosome number and type of nodal anatomy. Group I is characterized by the trilacunar condition and a multiplication of this condition. Group II is charac- terized by the unilacunar condition and a modification of this condition. Some of the more recent systems of classification of the Magnoliales have been criticised, and it has been pointed out where certain revi- sions would bring about a more natural system of classification. LITERATURE CITED Hutcuinson, J. (1921). The family Winteraceae. (Kew. Bull. 1921: 185-191.) —_— (1926). The families of flowering plants. I. Dicotyledons. (Macmillan Co., London. 328 pp.) IsH1KAWA, M. (1916). A list of the numbers of chromosomes. (Tokyo 8.) ManevalL, W. E. (1914). The development of Magnolia and Lirioden- dron, including a discussion of the primitiveness of the Magnoliaceae. (Bot. Gaz. 57:1-31.) McLaucGuHiin, R. P. (1933). Systematic anatomy of the woods of the Magnoliales. (Trop. Woods, No. 34 8 Morinaca, T., FukusHIMA, E., Kano, T., MARUyAMA, Y. and Y. Ya- MASAKI (1929). Chromosome numbers in cultivated ee (Tokyo 94.) Renper, A. (1927). Manual of cultivated trees and shrubs. (Mac- millan Co., New York. 930 pp. ) Sinnott, E. W. (1914). Investigations on a ia Nd of the Angio- sperms. I. The anatomy of the node as aid in the classification of Angiosperms. (Amer. Jour. Bot 1:303-322.) STRASBURGER, E. 905). Die Samenlagen von Drimys Wintert und die Endospermbildung bei Angiospermen. (Flora, 95:215-231.) Jour. Arnotp Ars. VoL. XIV. PLATE 80 CHROMOSOME NUMBER AND RELATIONSHIP IN THE MAGNOLIALES 1933] WHITAKER, CHROMOSOME NUMBER IN MAGNOLIALES 385 DESCRIPTION OF PLATE 80 Chromosome numbers in the Magnoliales. The drawings are from ceto-carmine and permanent smears of pollen mother cells. Magnifica- aoa about 2100. Fig. 1. Magnolia bripetala 2.0... cece cece cece cece eee eneee II M. ig ce OE AAP MO aes Sa Opa eae eae RR eee es I M. Fig, 3. Cércidipayliaon FapOnicume © nics vusuewsveeeesaesey I M. Fig. 4. Lirtodendron Tulipiferd .......cececceceee ee eneees I M. Pc 0s I GONOHE PP ORUES rei oss 5 ond pee eas Se wat I M. Fig. 6. Types of pee eae found in Liriodendron Tulipifera. Bip: 7. Magncud GOMMUNGIG < o4 be 6s cowed Sawa deca caw ee vas I M. Fig. 8. ante SeaTgeans var. Brossontt ..........-... 1M. PAS ee PrCLe Od DOPUINOE oo 6s. 5:5 Dorn ee Ee Oe MEP eats I M. Big. “10. (SCNSGNGIE SPRENGHINETE 0.6 o009scecueebin ees epee’ I M. IG Uk GUHA MEANT oe 4.46.640s +0 6 Hk s Hee Es Pe eke I M. CyTOLOGICAL LABORATORY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 386 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV CHROMOSOME NUMBER IN ACER AND STAPHYLEA Ropsert C. Foster With plate 81 INTRODUCTION ACCORDING TO HUTCHINSON (1926) Acer is a large genus dis- tributed widely throughout the northern temperate zone in North America, Europe, western Asia, China, Japan, and is found also in northern Africa. This large group of approximately 115 species is divided into at least 14 subgroups. Eight of these subgroups are rep- resented in this study by 13 species and varieties. Two more sub- groups, represented by one species in each, have been studied by Taylor (1920). In all, 10 of the 14 subgroups and 19 species and varieties have been studied by various workers, affording a rather broad basis for conclusions. Closely allied to the Aceraceae are the Staphyleaceae. In the genus Staphylea two species have been previously studied (Mottier, 1914, Winge, 1917). These two have been re-examined and two additional species studied in the present work. MATERIALS In the spring of 1933, branches bearing male flowers of Acer and Staphylea were brought into the laboratory and examined for reduction divisions in the pollen mother-cells and for microscope divisions. The reduction divisions in Acer apparently occur very early, when flower buds and anthers are extremely small. A similar condition is true of Staphylea. Out of a large amount of material, satisfactory stages were found in only 13 species and varieties of Acer and 4 species of Sta- phylea. All material examined for chromosome counts was studied from aceto-carmine smears prepared according to Belling’s formula (Belling 1926). In addition to chromosome counts, preparations of the fully devel- oped pollen were made and examined for pollen sterility counts. These counts were made on 53 species and varieties of Acer. All ma- terial was gathered in the Arnold Arboretum. RESULTS The chromosome counts and the pollen sterility counts on the species of Acer examined cytologically are summarized in Table I. For com- TABLE I. ae Distribution Group Species n 2n pollen Author (According to Rehder 1927) Ie Platanoidea platanoides L. 13 26 1% Meurman, 1933 pulls and Caucasus iyabei Maxim 13 5% Jap II. Campestria campestre L. 13 1-3% Eur Western Asia III. Saccharina saccharum Marsh. 13 5% Taylor, 1920 East ‘Canada, south to Ala., Ga., Mis IV. Spicata pseudoplatanus L. 26 52 1% Taylor, 1920 Europe, West Asia pseudo platens var. eryth- 26 1-2% rocarpum V. Palmata circinatum Bai 13 10-50% Brit. Columb. to Calif. palmatum Thunb. var. 13 20% intermediu pseudo-sieboldianum Komar 13 10% Manchuria to Korea VITI. Indivisa carpinifolium Sieb. & Zucc 52 Taylor, 1920 apan IX. Macrantha rufinerve Sieb. & Zucc 13 1-5% a el Tscho oe Maxim 13 2% Jap XII. Rubra rub 40 Darling, 1912 Newfoundlatid to an west to Minn. and Tex 36 {Taylor, 1920 oa 1914 a ot 50 88-94 Taylor 1920 ” 68-75 ” a 52 5% saccharinum L. 26 52& +91 Taylor, 1920 Que. to Fla., west to Minn. & Okla. Blooms very late in fall XIII. Trifoliata nikoense Maxim. 13 Japan, Central China griseum Pax 13 1-5% est Chi mandshuricum Maxim. 13 mak XIV. Negundo Negundo L. 13 20% aylor, 1920 New rene ree Dashing, 1909 Sinoto, 1929 Negundo var. interius Sarg. 13 25% Alberta & Saskatoon to Ariz, & New Mex, L8€ VATAHdVIS GNV YaAOV NI YAEWON AWOSOWOUHD “YALSOA [£€61 388 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV pleteness, this table also includes counts made by other workers and a brief indication of the geographic distribution of these species, accord- ing to Rehder (1927). Without exception, the haploid chromosome numbers found in this study were 13 or multiples of 13. Of Staphylea, four species were examined. ‘The results are sum- marized in Table IT. TABLE II. Species n Distribution Staphylea bumalda DC 13 Japan = pinnata 13 Central and southern Europe ws colchica Stev. 26 aucasus trifolia L. 39 Eastern Canada and U. S. Here, as in Acer, the haploid number is either 13 or a multiple of 13. DISCUSSION Of the 19 species and varieties of Acer which have been examined cytologically, 14 are diploids, with n = 13. Four are tetraploids, with n = 26; one of these, A. carpinifolium has had only its somatic number studied (Taylor, 1920). A. rubrum is an octoploid, with n = 52. The Maples, then, have 13 as a basic number and are, for the most part, diploids, but a polyploid series does exist. Differing counts have been made by other workers. Cardiff (1906) found n — 11 in A. platanoides, a count corroborated by Taylor (1920) who found, however, 2n = 26 in somatic counts on seedlings of this species. Meurman (1933) finds n = 13 in this species, as did Darling (1923). One seedling which the former examined proved to be a triploid, 2n = 39, which he regards as either an autotriploid, or a hybrid between a diploid and a tetraploid (loc. cit. p. 159). In A. rubrum, for which n = 52 was found, four other counts have been made by Darling (1912), Mottier (1914) and Taylor (1920). The last named worker, in one instance, made a meiotic count of n == +50, which approximates the count made in the present study. The varying counts made on A. rubrum make plausible Taylor’s (1920) suggestion of races within this species, possessing different chromosome numbers. Such a condition is known in other species, such as Musa sapientium L. (Tischler, 1910, 1928). The count of n = 13 in A, Negundo has been made by three differ- ent workers, Darling (1909), Taylor (1920) and Sinoto (1929), al- though Mottier (1914) found n = 12 or 14. The chromosomes in Acer are quite small in size. In shape they 1933] FOSTER, CHROMOSOME NUMBER IN ACER AND STAPHYLEA 389 have been described as “elongated” (Mottier, 1914), “ovoid” (Taylor, 1920), and “irregularly polygonal” by Sinoto (1929) who denies the accuracy of the first two descriptions. All three characterizations, as a matter of fact, are correct, as can be seen from the accompanying fig- ures. Elongated chromosomes are clearly shown in Fig. 6, and both ovoid and irregularly polygonal types in Figs. 2, 3, and 5. The appar- ent shape undoubtedly varies with the angle at which the mitotic figure is oriented with regard to the observer. All the forms studied showed complete formation of bivalents. Even in polyploids like A. rubrum, no univalent or multivalent formations were seen. Meurman (1933) found a similar completeness of pairing in A; platanoides, as did Sinoto (1929) in A. Negundo. This situation is probably due to the fact that with a low chiasma frequency, there is little chance for the formation of multivalents. At metaphase I in the octoploids A. rubrum most of the bivalents have one chiasma and are in the form of rods; there are few if any rings. Meurman (loc. cit. Figs. 13 and 17) shows this to be true of A. platanoides, and the accompanying Figs. 1 and 7 of 1st metaphases, show this is the case in A. Miyabei, and A. pseudoplatanus var. erythrocarpum. From this it is apparent that the chiasma frequency is probably 1, although the occasional presence of a ring would raise it slightly above 1. As a result of the regularly-formed bivalents, the chromosomes can be distributed regularly to the poles, even in polyploids like A. rubrum. The phenomenon of secondary pairing noted by Meurman (1933) in A. platanoides is present in the species included in the present study. It is especially well shown in Figs. 6 and 8. Meurman (loc. cit. pp. 160 and 162) also found that 2 large pairs of bivalents showed sec- ondary pairing at both the 1st and 2nd metaphases, but found no such pairing between other bivalents. It was not found practicable to include a study of somatic chromo- somes, but both Taylor (1920) and Meurman (1933) have published some details on this point. Taylor noted that the longest somatic chromosome studied was about 3 microns long, and the smallest, 1 micron long, with diameters of from 1/3 to 2/3 microns. Meurman found in the somatic chromosomes of A. platanoides lengths from 0.8—22 microns. Both writers, too, note the existence of regions of doubling of chromosome numbers in the root tips. Pollen sterility counts were made on 53 species and varieties. Most of them showed a high percentage of good pollen. The species and varieties showing more than 80% good pollen were as follows: A. 390 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV platanoides L. and its varieties cucullatum Nichols, Schwedleri K. Koch, Stollii Spaeth, dilaceratum Dieck, palmatifidum Tausch, columnare Carr., nanum Nichols., A. Miyabet Maxim., A. truncatum Bge., A. pic- tum Thunb., A. campestre L., and its varieties compactum Schwerin and postelense Lauche, A. saccharum Marsh., A. grandidentatum Nutt., A. pseudoplatanus L. and its variety erythrocarpum Carr., A. Heldreichit Orph. var. macropterum Vis., A. Trautvetteri Medwed., A. ginnala Maxim. and its var. atdzuense Franch., A. tataricum L., A. palmatum Thunb. and its varieties atropurpureum Nichols., sanguineum Lem., septemlobum K. Koch, intermedium Schwerin, elegans Koidz., Hessei Schwer., sinuatum Schwer. and laciniatum Schwer., A. Sieboldianum Miq., A. pseudo-Sieboldianum Komar. and its variety ambiguum Nakai, A, pennsylvanicum L., A. rufinerve Sieb. & Zucc., A. Tschonoskii Maxim., A. argutum Maxim., A. rubrum L. and its variety glaucum Marsh., A. griseum Pax and A. Negundo L. and its varieties pruinosum Schwer,. and texanum Pax. Certain species display a high percentage of poor pollen, as follows: A, Mayrii Schwerin 50% bad A. zoeschense Pax 50% * A, spicatum Lam. 55% A. circatum Pursh 50% * A. tegmentosum Maxim. 100% = <* A. barbinerve Maxim. 90% « A, Negundo L., var. interius Sarg. 25% “ A. Negundo, var. nanum Dieck. 50% * Such high percentages of pollen sterility, as contrasted with the lower figures for the others studied, indicate a possible hybrid origin of these species or structural hybridity. In the case of one species, A. zoeschense, Rehder (1926) indicates that it may be a hybrid between A. campestre L. and A. Lobelti Ten. Little data on hybridization appear available. Rehder lists about 15 species hybrids, but they are usually between species in the same subgroup or in closely related subgroups. The widest cross noted was that between A. opalus Mill., var. obtusatum Henry in the Campestria, and A. pennsylvanicum L. in the group Macrantha. Information on grafting supplied by William H. Judd of the Arnold Arboretum, indicates that A. griseum Pax and A. parviflorum Franch. & Sav. have not been used successfully in grafting with other species. 1933] FOSTER, CHROMOSOME NUMBER IN ACER AND STAPHYLEA 391 Ordinarily, too, a species can be used as stock or scion only with species closely related to it. Mr. Judd has found this to be particu- larly true of A. platanoides. Chromosome conditions in Staphylea present a close parallel with those in Acer. The basic number is 13, but a tetraploid, S. colchica, n = 26, and a hexaploid, S. trifolia, n = 39, are found. In S. pinnata Winge (1917) found n = 12, but noted n = 13 in one cell. Mottier (1914) found n = +36 in S. trifolia. Although the chromosomes are much larger than those of Acer it is difficult to determine their shape from polar views of meiotic metaphases. Like Acer, too, they appear to have a low chiasma frequency, and separation is quite regular. The secondary pairing is clearly shown in Figs. 10 & 11. As in Acer, this secondary pairing is between 2 bivalent chromosomes, and often makes an accurate count quite difficult. The data thus presented show that in two genera belonging to dif- ferent families there are identical basic chromosome numbers together with a similarity in polyploid series, secondary pairing, and low chiasma frequencies. In Acer there is found a great differentiation of species in a highly polymorphic genus. Yet this process of species differentiation has taken place with no change in the basic chromosome number of the genus. Although only 1/6 of the known species of Acer have been studied, their distribution throughout the subgroups of the genus is sufficiently wide to make this statement reasonable. What is true of a genus is apparently true also of families. Despite a clear relationship, the Staphyleaceae are admittedly different from the Aceraceae morphologi- cally, but the type genus, Staphylea, shows the same chromosome set- up and the same general behavior, even to secondary pairing, which is found in Acer. It is true that there are differences in chromosome size between the two genera, but this is probably of no great significance. Such differences in chromosome size often exist between species within a genus or between varieties of the same species. Considered with other similarities, the common chromosome number and behavior may well indicate a common origin for these two closely related genera. SUMMARY 1. Chromosome counts were made on the meiotic stages of thirteen species and varieties of Acer and four species of Staphylea. Thirteen was found to be the basic haploid number in each genus. 392 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV 2. These counts, with those given by other workers show that most of the species are diploids. A polyploid series, however, is found in each genus. 3. The chromosomes of Acer are small, have a low chiasma fre- quency, behave regularly since no univalents or multivalents are pres- ent, and exhibit secondary pairing. 4. The chromosomes of Staphylea are larger than those of Acer, but show the same low chiasma frequency, regularity of behavior, and second pairing. 5. Pollen sterility counts were made on 53 species and varieties of Acer. Forty-five showed more than 80% good pollen. The remaining eight showed from 25-100% sterility. 6. Evidence from hybridization and grafting is briefly considered. 7. It is concluded that the cytological details afford evidence of a common origin for the Aceraceae and Staphyleaceae. I wish to express my thanks to Dr. Karl Sax for his assistance in preparing material and drawings, especially those of Staphylea, and for his criticism of the manuscript, and to Dr. Haig Dermen who made all the pollen sterility counts. LITERATURE CITED —" . BELLING, JOHN (1926). The iron-aceto-carmine method of fixing and staining chromosomes. (Biol. Bull. 50:160-161. CarpirF, J. D. (1906). A study of synapsis and reduction. (Bull. r ¥ . Daring, C. A. (1909). Sex in ‘dioecious plants. (Bull. Torrey Bot. 36: ——_—_— 2). Mitosis in living cells. (Bull. Torrey Bot. Club, 39:407—409. ) ————. (1923). Chromosome behavior in Acer platanoides L. (Amer. Jour. Bot. 10:450-457. HutcHInson, J. ( . Families of flowering plants. I. Dico- tyledons. pp. 328. (Macmillan, New York. MeurMAN, O. (1933). Chromosome morphology, somatic doubling and secondary association in Acer platanoides L. (H 18 :145-173. . Morrtier, D. M. (1914). Mitosis in the pollen-mother-cells of Acer negundo L. and ps ee trifolia L. (Ann. Bot. 28:115-133. 9. REHDER, ALFRED (192 7). Manual of cultivated trees and shrubs. N DBD ww & WwW bb oo s in some dioecious plants with especial reference to the ine (Cytologia, 1:109-191.) . (192 i _— o —N Lal Z 3 oe} 2 < —~\ OQ: >= a ° 3 s) mn fe) 3 a) n ce 2 — — HF : o a wm e- 3 ie) — ae) = a 2) 2 Qa .— “ny s a a te + CS ° rR i ‘@) sere “wn ped co oh S ie) rh reproduction in the genus Acer. (Contrib. Bot. Lab. Univ. Penn. 4:271-300.) Jour. ArNotp Ars. VoL. XIV. PLATE 81 $e °e, ° a e Dien LPs \\ ite ote 8 ee 7 g ie XY .,” = ny & @ e » bee “n° 7s vow e *%e* ee ug es 4a @ 6 & & a°s VO & = “° ~ e e ®@ q 10 CHROMOSOME NUMBER IN ACER AND STAPHYLEA 1933] FOSTER, CHROMOSOME NUMBER IN ACER AND STAPHYLEA 393 12. TiscHier, Georc. (1910). Untersuchungen tiber die Entwicklung des Ban anenpollens. I. (Arch. f. Zellforsch. 5:622-670.) ———___ 28). Ueber i. Verwendung der Chromosomenzahl fiir phylogenetische Probleme bei den Angiospermen. (Biol. Zen- tralblatt, 48:321-345. 14. WINGE, O. (1917). The chromosomes, their numbers and general importance. (C. R. Trav. Labor. Carlsberg, 13:131-275.) EXPLANATION OF PLATE 81 The figures in the plate were drawn from aceto-carmine preparations of meiotic sta The drawings were made with the aid of a camera lucida. Mare (ee tlon 1800, except Figure 8 which is about 2450. Fig. 1. Acer Miyabei; 1st metaphase, side view. Fig. 2. Acer circinatum; 2nd telophas Fig. 3. Acer pseudo- sieboldianum; late diakinesis. Fig. 4. Acer Tschonoski; i metaphase, end view. Fig. 5. Acer griseum; diakinesis Fig. 6. Acer Negundo aie diakinesis. Fig. 7. Acer pseudoplatanus erythrocarpum, 1st metaphase, side view. Fig. 8. Acer rubrum; lst metaphase, end view. Fig. 9. Staphylea pinnata; Ist see He end view. Fig. 10. Staphylea colchica; 1st telophas Fig. 11. Staphylea trifolia; 1st Barri end view. CyTOLOGICAL LABORATORY, ARNOLD ARBORETUM, HARVARD UNIVERSITY. 394 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV STUDIES ON THE “PRECIPITIN REACTION” IN PLANTS V. APPLICATION-TO PLANT RELATIONSHIPS K. S. CuHester, E. C. ABBE, AND P. A. VESTAL IN AN EARLIER PAPER Of this series (2) the senior writer applied the “normal precipitin reaction,” better designated by Foster and Avery (5) as the “precipitation reaction,” to representatives of a number of families of plants for the purpose of determining whether or not the method is applicable in the study of plant relationships. In most of the material tested at that time there was an abundance of positive reactions, and these tended in general to harmonize with the findings of empirical taxonomy, although a uniformity of negative reactions in some groups limited the procedure to certain families in which well marked positive reactions were obtained. In 1933 Foster and Avery (5) applied a similar technique to the genus /ris with results which were satisfying and entirely consistent with those reported in the pres- ent series of studies. The purposes of the present paper are to provide additional data on the occurrence of the precipitation reaction to those already recorded, and then from a consideration of all data thus far obtained to analyze the significance of the precipitation reaction in a study of plant relationships with particular reference to the limitations and essential meaning of the reaction. The new groups tested in the present study are the “Amentiferae” and the Guttiferae. The follow- ing scheme gives the results of all precipitation tests thus far obtained including the results of the present study. Tests of Some posi- Entirely genera tive tests negative Family: Genera: inter se: obtained with: tests with: Reference: Tridaceae Tris 301 Numerous Solanaceae? Saxifragaceae Foster & Av- positive (CaC,0,) ery (5); reactions Oleaceae Table IV. correlated Caprifoliaceae with system- Rosaceae atic position. ‘“Amentiferae”’ uttiferae 1The number refers to the number of species employed in the tests. 2Signifies that the only reactions observed were determined as due to calcium oxalate. 3Refers to tables in the present paper. 1933] CHESTER, ABBE AND VESTAL, PRECIPITIN REACTION 395 Tests of Some pa Entirely nera tests negative Family: Genera: inter se: obtained with: tests with: Reference: “Amentiferae” Uniformly Solanaceae Rosaceae Tables I, Saas Salix 1 negative Oleaceae Saxifragaceae III, IV Myricaceae: Myric Iridaceae Guttiferae Leitneriaceae: Lcitneria 1 Caprifoliaceae Juglandaceae: Carya 1 (CaC,0,) Betulacaeae: Alnus 15 Betula 16 Carpinus 6 Corylus 8 Ostryopsis 1 trya Fagaceae: Quercus 1 Saxifragaceae: oa 1 As in Solanaceae Iridaceae Chester ad Fen Tridaceae Rosacea “Amentiferae” table I See Platanaceae Guttiferae ma 1 eguml Hydrangea 1 Oleaceae Jamesia 1 (CaC,0, Deutzia 1 Caprifoliaceae Itea 1 (CaC,0,) Ribes 1 Platanaceae: Platanus 1 Solanaceae Leguminosae Chester (2) Rosa Saxifragaceae Oleaceae (CaC,0,) Caprifoliaceae (CaC,O,) Rosaceae: Spiroideae: Spiraea 1 Rosoideae Pomoideae Chester (2) Prunoideae Platanaceae guminosae Saxifrag Pomoideae: Cotoneaster 1 Generally Iridaceae Platanaceae Chester (2); Stranvaesia 1 negative lanaceae Leguminosae table IV Chaenomeles 1 oe Amelanchier 1 Pyracantha 1 Mespilus 1 Crataegus 1 Caprifoliaceae oideae “Amentiferae arena Guttiferae Oleaceae (CaC,0,) ” 396 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Tests of Some posi- Entirely genera tive tests negative Family: Genera: inter se: obtained with: tests with: Reference: Pomoideae: Sorbus 1 (continued) Aronia 1 Photinia 1 Malus 3 Pyrus 1 Prunoideae: Prunus 15 As in Saxifragaceae Iridaceae Chester (2); Osmaronia 1 __Iridaceae Leguminosae “Amentiferae” table IV Prinsepia 1 Pomoideae Guttiferae Maddenia 1 Platanaceae Solanaceae cea (CaC,0,) Caprifoliaceae (CaC,0O,) Leguminosae: Robinia 1 Solanaceae Platanaceae Chester (2) Saxifragaceae Oleaceae (CaC,0,) Caprifoliaceae Guttiferae: Hypericum 10 Uniformly Saxifragaceae “Amentiferae” Tables II, negative Solanaceae , Oleaceae Iridaceae Caprifoliaceae (CaC,O,) Rosaceae Oleaceae: Syringa Uniformly Solanaceae Chester (1); Ligustrum 8 _ negative Iridaceae table IV Fraxinus 1 except for “Amentiferae” Chionanthus 1 calcium ox- Guttifer Forsythia 1 alate reac- Rosaceae tion CaC,O Saxifragaceae ,O Caprifoliaceae 2°74 Platanaceae (CaC,0,) Leguminosae (CaC,0,) 1933] CHESTER, ABBE AND VESTAL, PRECIPITIN REACTION 397 Tests of Some posi- Entirely genera tive tests negative Family: Genera: inter se: obtained with: tests with: Reference: Solanaceae: Lycopersicum1 As in Oleaceae Kostoff (6); Salpiglossis 1 Iridaceae _Iridaceae Chester & yphomandra 1 (CaC,0,) Whitaker Browalli Rosaceae (3); table Nicotiana 19 Saxifragaceae IV Capsicum 2 Caprifoliaceae Solanum 5 Platanaceae Physalis 1 eguminosa Petunia 1 “Amentiferae” Lycium 1 Guttiferae Datura 4 Atropa 1 Caprifoliaceae: Kolkwitzia 1 Uniformly Solanaceae Chester (2); Symphori- negati Iridaceae table IV carpus 1 osaceae Diervilla 1 Guttiferae Viburnum 1 e Linnaea 1 Ai 5s Sambucus 1 Saxifragaceae Dipelta 1 C,0,) Abelia 1 Platanaceae Lonicera 2 (CaC,0,) Leguminosae (CaC,0, “Amentiferae” (CaC,0,) Thus up to the present fifteen families and approximately two hun- dred species of plants have been tested more or less extensively with regard to the precipitation reaction. Of these fifteen families, four {Solanaceae, Iridaceae, Saxifragaceae and Rosaceae-Prunoideae) have yielded among themselves significant positive results from the taxo- nomic standpoint. In five of the other eleven groups (Oleaceae, Rosaceae-Pomoideae, Caprifoliaceae, Guttiferae, ‘“Amentiferae”) fairly extensive tests within the groups have yielded wholly negative results, while in the remaining seven families the results thus far obtained are madequate for sound conclusions because the number of species tested is too limited. The present paper reports the results obtained in tests of the pre- cipitation reactions of forty-five species of Betulaceae, one species each [ VOL. XIV The technique employed ..moderate reaction; “3”... Experiments performed with dried BETULACEAE no reaction. PRECIPITATION REACTIONS IN THE JOURNAL OF THE ARNOLD ARBORETUM TABLE I. “e” trace; “l”...weak reaction; ‘2’. tiferae), and two species each of the Iridaceae, Solanaceae, Oleaceae, strong reaction; “O”... of five other families of ‘‘Amentiferae,” ten species of Hypericum (Gut- Rosaceae, Caprifoliaceae, and Saxifragaceae. herbarium material. 398 *2¢ WiseTo SMTIY 5 OO ©e©o0°o o ©oS90 C000 Coa coo The tests lum specimens VusyseTeOeU"e GD CO cooo e©02 290 89000 oOC9O ea°o ° VsotmpueTs FINI o Go oo0o°o ° ea0 09 ByTosTNAeA"y oD CSO eooo oo e900 e000 800 coo ° eypoogM"'y ao CoO oeco°o0 2©0 800 COCO aod e090 ° ° ST TewepTto°y OD CO foo Rore) 20 C800 COCO OO”D eo5o BTTOSTQQOM! SMUTY 4 CO ecco o©S9 e902 e800ao COn e090 0 ° eizyu FINw~E 0 CO Scoo0 ©2 ©8890 C000 Ce00 coo eaqul S™MTY Qo oOo eoooo e202 800 CACC O COAG Coad Pueyuysiya 2e°o°o ecooo 202 8200 0800 C9200 oao0 woymodelQ Qo0 °o°o°00 22 S00 20k 9 co0e coo emetpraed syedo£i3y80 Qoo ©2000 0 oo eo°o suposeatds*) coo 2000 0 90o coo waumtoo ent Ato G00 eec0e 0 Coo o°°0 wueyuyToreo emmpdweg Qao eo0o0 80 8O eoo eotdsyqotsa"d gaa eoco ° eco 00°90 auepyooo wiezpasded*g Qoo ©020Q 0 090 coo TraeTpeaped’d GqQoo ecco Oo of ooo wd 'd goo ©0200 0 of O00 FRROCL wl nze% ° weodur: eoo coco °o ooo SCCoCOoOCoCeooo e000 8CC000000 woyuoteh*y coo 2000 0 of ooo vmou"y aa0 e000 0 of eco eqmsty emmy Goo eco000 0 09 eco ° wresyitded wind Goo e000 ° 220 9009000009 ooo ocooo0 Wusojieme entf£I0D Ooo ©8020 80 oo ooo Bttostawe3"¥ DO00O G9QA00 0 Oo STIAVMY GDQDCATCCADO coOCoOo trayeeds SMTY G~o0o GA00 co00 rs) woyrem"dp C00 CCo00o ° e0000000 ecoo oo0S9 ao000 VABPPICGETETD COCOCQODGDS2TO CaAGCSO Vue TT eaB* ec0eococoeofC oO oooo ° VI TAqdoreyeq ent sre *WLostImMdod*g Cao eo00 o of Vueyz0yaoN; WTI Coo e°ceoo oo e990 SCoOC00 COCO ooo ST pwers-etureco*g{ eimund¢ CGOgq eoo°o oo eoo of ° TITS’ COS ©eooo ° coo 8000 COO e090 9o9000 weeare winged OOo ege0900 oo eatnzeq") Oo STTeqWeTI0"g © ° ° ° B4WPI0O'D oO TyaouywesOINT*D 2 Oo ° woyucdef sumydwegn COG CoC O00 ° eo90000 wIrmmg§ CO0G 0000 ° ©2200 C900 wsya WInyed © ° waits*¥ © “Ig VisyzoemmtTy o = ba | or} 4 a tf 2 4 3 ges 2, 2328 9 3.4.52 0 9 t8s9s53 Edledeles a dint tiie @ he a3 a a | -2e° oo a CELILEELEL LEELA BRE BEES tre eee was essentially as described in the earlier papers of this series (1, 2). were used for material, and these tests were later confirmed by repeti- In testing the majority of the Betulaceae dried herbar tion of many of them using fresh leaves of the same plants. 1933] CHESTER, ABBE AND VESTAL, PRECIPITIN REACTION 399 of the other ‘““Amentiferae’ and other families mentioned above were performed with freshly collected leaves. The dried leaves were ex- tracted for a few hours in ten times their weight of distilled water, and TABLE II. PRECIPITATION REACTIONS IN HYPERICUM (GUTTIFERAE) “4 trace; “+”...weak reaction; “t+”... moderate reaction; “+++” strong reaction; “O”...no reaction. Experiments performed with fresh material. a 8 g De > ace g | a 8 oO 3 Fat bse ag d 3 rs g 9 f >» © 2 8 8 FI can be] aa oO 4 ky 4 3 ° a= Gn ee | com Ss &F 4 wt °o ao 8 s222¢8 § = 3 = S82 58888224 3 4 Poth mh Hm & mm mM m& OS Hypericum Oo O OQ o 0 0 0 0 0 0 Ascyron #1. H. galioides ©1260 (@) Oo Oo Oo oO 0 Oo O 0 t (@) H.amreum (0) 0 0 0 0 @) 0 0 e) 1) - 0 Hearmoldamum 0 0 0 0 000 0 0 0 0 4 #0 H.Kalmiamnmm 0 0 0 0000 0 0 0 0 ¢ #O H.mudiflorum 0 0 o0.6.6U6UO”# (@) o oO O 0 0 Oo ¢t 0 H.lobocarpum 0 0 0 0 0 0 QO Oo oO O O ee O H.Buckleyi 0 0 0 0 Oo O Oo 860 Oo O 0 oa) H.boreale c@) 0 (@] 0 0 0 0 0 (0) @) 0 t 0 H.perforatum 0 0 Oo O Qo oO 0 Oo oO oO O + QO H.Ascyron #2.0 0 0 0 0 00 0 0 0 0 # 0 the fresh leaves in three to four times their weight of distilled water. The results of these tests are given in Tables I, I, III, and IV below. As is indicated in Table I an experiment involving forty-five species of the Betulaceae showed no positive results. These species represent every section of every genus of the family as recognized by Winkler (8) 400 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV with the exception of the section Cremastogyne in Alnus. In every case the material used was authenticated by herbarium specimens from which dried leaves were obtained for the tests. Through the kindness of Mr. Rehder of the Arnold Arboretum the determinations of these TABLE III. PRECIPITATION REACTIONS IN THE “AMENTIFERAE” Notations as in Table II. Experiments performed with fresh material. 3 .. 3 4 ; “A a 2.24% 4 ae | ao. 8 4 @ A Bb o 3 es a a oe es fae g fi P a3 2 doa 5 @ oO f 9 u $ 9 & a Z ar as & 9° © g »& gE 3 a4 2 a a os 3° bal a at - 3 a F| ° Sg Ee E 5s a@n33 5 2? q o 7 so 3zue: 8 PRPRBERE REY e824 4 8B 8s ee PER Oa eA Gg ZG OG a 2 on > > © = ~~ & a 7 Oo 2 ef Ostryopsis © 0 0 0000 00 0 0 0 0 0 0 0 we Ostrya virginiana og 0 0 0 0 0 0 0 0 0 0 0 0 0 « ) Carpims cordate 9 0 0 0 © 0 0 0 0 0 0 f°) Oo + ) C. caroliniana 0 0 0 0 0 0 0 0 0 0 © 0 0 0 0 0 + 0 Corylus cormta 0 0 0 0 060 00 0 0 0 0 ce) 0 0 0 « ie) Betula nigra 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 + oO B. pumila °° 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 + oO B. occidentalis 0 0 0 000 0 0 0 0 0 0 0 0 0 0 + 19) B.Maximowiczilana 0 oO oO 0 0 0 0O 0 0 0 0 0 o 8.060 + 0 Alnus crispa 0 0 0 0© 0 60 0 © 0 0 6 0 0 © 0 O0 ew 0 a. incana o © © © 0 6 © © 0 © 0 6 0 60 0 0 60 ws 06 Quercus alba © 0 0 0 0 000 00 0 0 0 0 0 0 ww Oo Leitneria florid, 90 0 © 0 0 0 0 60 00 0 060 0 0 0 0 0 eo 0 Salix Lat sudana {e) ie) ¢) o nn «OO Carya alba © 0 0 0 0 0 0 0 0 0 0 8 0 0 © 9 + fr) kyrica © © 0 0 0 0 0 0 6 0 0 0 © 0 0 G0 Oe 0 Hypericum nudifl. 9 0 0 0 0 0 0 © 0 0 0 oo 0 00 0 tt (e) specimens were checked and the specific names are cited in accordance with his treatment of the family in his “Manual of Cultivated Trees and Shrubs” (7). In view of the discrepancy observed in Jris by Foster and Avery (5) between the results obtained with fresh as contrasted with dried leaves, it was thought advisable to check the observations presented in Table I using fresh leaves from certain of the same plants involved in the tests 1933] CHESTER, ABBE AND VESTAL, PRECIPITIN REACTION 401 previously made. Opportunity was also taken to introduce one species each of the following families: Fagaceae, Leitneriaceae, Salicaceae, Juglandaceae, and Myricaceae. These other families of the “Amenti- ferae” were introduced in the hope that, should differences of reactivity within the group occur, some light might be shed on the vexing ques- tion of relationships between these families. A member of the Gutti- ferae was employed as a check on possible wider relationships. Again the observations (Table III) show absence of reactivity between the leaf tissue extracts. To determine what reactive substances occur in the Betulaceae, in the Leitneriaceae, and in Hypericum of the Guttiferae, the concept of “test plant” of a “known” constitution was introduced. Fresh material of the species with the label-names indicated in Table IV was collected in the Arnold Arboretum and these were tested against each other. The results indicate a similar reactivity in Betula populifolia, Diervilla florida, and Malus Arnoldiana, since these show positive reactions only with Jris chrysophoenicia, Nicotiana alata, and Lycopersicum escu- lentum, and all contain an excess of calcium ion. This suggests the presence of but one principle of a reactive pair, probably “A” (3), since the Nicotiana alata and Lycopersicum esculentum have been shown in an earlier paper (3, p. 186, fig. Iv) to contain an excess of the com- plement “B.” The possibility of this being the ‘““MN” reaction (1. c.) is eliminated because of the negative reactions with Ligustrum obtusi- folium which is known from the earlier work to be “N--+.” An interesting situation exists in the case of Leitneria where but one odd reaction occurred, namely with Syringa velutina, suggestive of a fifth reactive pair. Further investigation is needed here. An analysis of the results presented in Tables I and III yields little of value in the light of these conclusions, beyond the fact that oppos- ing members of reactive pairs are absent. One of these is probably “B” and another the counterpart of the unknown existing in Levtnerta. This striking absence of reactivity within the Betulaceae might well be construed in confirmation of preexisting morphological evidence indi- cating close relationship between the members of the family. But in considering the ‘““Amentiferae” with their similar negative reactions, it should be kept in mind that widely varying treatments of the relation- ships between the families composing this group exist. Our evidence,. in the light of the hypothesis expressed in an earlier paper (2) and later accepted by Foster and Avery (5), i. e. that negative results indi- cate very close or very distant relationship, helps not at all in clarify- ing the inter-relationships of the “Amentiferae.” 402 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. xIV The genus Hypericum as represented here includes ten species of five different sections as recognized by R. Keller in Engler and Prantl (4). The species vary from the small herbaceous H. boreale (Britton) Bick- nell, the larger herbaceous H. perforatum L. and H. Ascyron L., through the shrubby forms to the dwarf mountain shrub H. Buckleyi M. A. Curtis. The list includes one known hybrid H. Arnoldianum Rehder with both parents (H. galioides and H. lobocarpum) (7). All the species used were collected in the Arnold Arboretum and authenti- cated by the junior author. The precipitation reactions in Hypericum show consistent negative results as may be seen in Table II. From this one may conclude that the genus is homogeneous regarding reactive substances or that they are absent. In Table III one species of Hypericum tested against the ‘“Amentiferae” shows no positive reactions. This affords no clue to the situation. However, when tested against other representative families (Table IV) it will be seen that the species is highly reactive. Unfor- tunately a wide enough variety of “known” plants was not used in the present study to make a complete analysis possible. It will be seen (Table IV) that good positive reactions were obtained with Jris chryso- phoenicia and with the solanaceous representatives, Nicotiana alata and Lycopersicum esculentum. Since these species did not react against each other but reacted consistently against other forms, one may con- clude the reaction to be of similar quality and doubtless due to the presence of the same reactive substances. The reactions with the olea- ceous members, Syringa velutina and Ligustrum obtusifolium, must be due to a second set of reactive principles, since the Oleaceae also re- acted with the Solanaceae. In the same manner the positive reaction with Lonicera Myrtillus probably represents a third set of characters. The trace shown with Philadelphus grandiflorus is of doubtful char- acter since this species did not react with any other plant, thus giving no indication of the reactive substances involved. From the evidence presented one is justified in concluding the pres- ence of three reactive principles in Hypericum ; whether these fall into the categories of the reactive substances designated in an earlier paper (2) as “AB,” “MN,” and “XY” is not known. Although we cannot be sure that the same principles occur in all species of Hypericum, at least the species tested do not show the presence of any of the opposing characters. The known reaction, calcium-oxalate, is of no significance in indicating degree of relationship in this genus as is shown by the harmony of negative results obtained when tested against calcium 1933] CHESTER, ABBE AND VESTAL, PRECIPITIN REACTION 403 chloride. Thus it would seem that in the genus Hypericum the precipi- tation technique would be of little value in determining relationships within the group. TABLE IV. PRECIPITATION REACTIONS IN REPRESENTA- TIVES OF ALL THE FAMILIES THUS FAR INTENSIVELY STUDIED WITH REGARD TO THE PRECIPITATION REACTION Notations as in Table II. Experiments performed with fresh material. Extracts containing an excess of oxalate neutralized with calcium chloride and re-filtered before testing. a » s rT .na4 So . 2 a eee FEE g24¢ 8 Regs 4309, F335 8 Sf Soc4gy = ae ee Sar ge ag8 tS Eg aoc RE Ta Bg a : 3 x ePagetaeiigads fifa g a «a S$ 2 a >| a « s e a zt § a oh eee eRe ea SEG EEE SE ie Vee d i & & 8 3 a3 awk Pea & Aa A wo & Irie ppeundacomea 60 0 0 0 0 0 0 0 09 0 0 © 0 0 0 6 I.chrysophosnicia 96 © 0 0 ee ¢ © 0 0 0 & ee O eee 0 we Micotima alata © 0 0 0 we «© &© 0 0 0 & e O et ore 0 vee lycopersicumes. © 0 0 0 oF © #¢ 0 0 0 © © 0 & te & 0 (ee) Gyringa velutinn 0 ees 2% 0 0 0 0 0 0 0 0 & O oF se 0 (eee) Ligust rum obtus. Oe © e€ © 0 0 0 0 0 0 0 0 O0 +6 O shee Malue Arnoldiana © ¢ e #» 0 0 0 0 000060 0 0 O O et Pranus sermulas. © 0 0 0 0 0000 0 0 0 0 0 0 O He Philedelphus gr. 0 0 0 066€6C00hmlhCOmlUC OCHO I aC Deutzia scabra ro © a © © SE © 0 EE © SE © UIE © I © EC © © © © © Diervilla florida 0 »« * #¢ 0 0 0 00 00 00 0 0 O ee Lonicera myrtil. © eo # # 0 0 0 0 0 0 0 0 0 G & see © (++) Leitneria florid. 0 0 0 0 e+ 00 00600 0 0 0 0 0 QO Betula populifol. 0 eo # # 2 0 0 O09 O 9 (oe Oe © ¢ Hypericum lobooarp.0 e++ eee see te ee 0 0 t O 0 + 0 0 0 DO Fre We are now in-a position to resolve more satisfactorily the question of the taxonomic significance of the precipitation reaction in plants. Within certain of the limited plant groups tested (families, subfamilies, and genera) there have been obtained numerous positive reactions. 404 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XIV These positive reactions show a marked correlation with the taxonomic positions of the plants involved (2,3, 5). Within others of the limited groups tested (Caprifoliaceae, Oleaceae, ‘“Amentiferae,”’ Guttiferae) entirely negative results have been obtained. In at least three of these (Oleaceae, Betulaceae, and Guttiferae) the genera and species tested are usually considered rather closely related to one another and the negative findings may well be of significance indicative of this close relationship. In such groups as the Caprifoliaceae, which while rather heterogeneous still display uniformly negative reactions, one may only conclude that chemical differences demonstrated by the precipitation technique are not necessarily associated with gross morphological differ- ences, since any category of characters used for classification may vary independently of any other, a conclusion consistent with the findings whenever any two techniques (e. g. morphology, cytology, anatomy, genetics, etc.) are used in a taxonomic study of a given group of plants. As regards systematics on a larger scale, that is, phylogeny of the angiosperms taken as a whole, the precipitation technique is appar- ently not applicable in a differentiation of widely separated groups. This is evident when one considers Table IV which gives the results of inter-family tests of nine widely separated families of plants. Here no useful correlation of precipitation reaction with systematic position is to be seen. The reason for this is quite apparent when one considers the nature of the precipitation reaction. The thousands of positive reactions which have been observed in fifteen families of plants do not represent thousands of specific reac- tions but rather represent a relatively small number of analyzable reactions, one of which has been chemically determined and three of which have to some extent been characterized. Two or three pairs of reactive substances suffice to account for all of the inter-family reactions of Table IV; four pairs of reactive substances accounted for all of the reactions of paper ITI of this series (Solanaceae et al.). The substances responsible for any given reaction seem to be rather widely distributed throughout the angiosperms. Thus (Table IV) the substance in the Solanaceae which reacts with oleaceous extracts is likewise present in Iris, and the oleaceous complement also in Malus and the Caprifolia- ceae. This fact does not detract from the value of the precipitation reaction when confined to limited groups of plants (families, subfam- ilies, genera). Within such limited groups it has been found that morphologically similar species tend to carry similar complements of precipitating substances (2, 3, 5). 1933] CHESTER, ABBE AND VESTAL, PRECIPITIN REACTION 405 The value of the technique within families of the higher plants when used to determine similarity or difference as far as the three or more unknown and one known pairs of variables are concerned, is to be evaluated as in the case of any other category of evidence with a simi- lar number of variables. If all four or more pairs of variables are present in a given family the technique is correspondingly significant. If the number of variables is less, the value of the technique is accord- ingly reduced, until we reach a condition such as exists in the Betula- ceae and apparently the “Amentiferae” as a whole where the precipita- tion technique indicates no difference between the species tested. It is obvious from an examination of the experimental data thus far obtained that the method in its present form with only a relatively few pairs of variables provides hardly enough combinations of characters to be of significant aid in determining the extent of relationship between widely separated families. Thus we must guard against a possible mis- interpretation of the statement made earlier in this series of papers (2) and later confirmed by Foster and Avery (5) that absence of reaction indicates very close or very distant relationship, while a positive re- action indicates an intermediate degree of relationship. Clearly this concept cannot be applied to groups too diverse morphologically—its use must be restricted to plants closely related as determined by other means. Within such groups it should be kept in mind that the applica- bility of the method is determined by the number of distinct reactive pairs of substances. The greater the number of these the greater the number of categories into which the plants being investigated can be placed. In pointing out this limitation of the applicability of the precipitation reaction in plant systematics let us compare it with the use of chromo- some number in taxonomy. Within limited groups of plants chromo- some number may aid in classifying species; the fact that the same chromosome number may be found in very distantly related plant groups does not detract from the use of chromosome number in plant classification. The same is true of the precipitation reaction; its appli- cability in closely related groups of plants is not to be belittled by the fact that the same reactive substance may occur in very distantly related plant families. The work done up to the present on the precipitation reaction in plants marks only a groping beginning toward a phase of plant sys- tematics which will doubtless develop more widely in years to come, namely, the use of chemical properties in the study of plant relation- 406 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV ships. In testing plants by the precipitation technique one is actually testing related groups of species against arbitrary and fortuitously selected “‘test extracts.” Since there is apparently only a very limited number of types of precipitation reaction present in plants, one would actually gain in time and efficiency if he were to select a few “test species” containing various known reactive complements, and use these as standards against which to test all species of selected limited groups. That this ‘‘test species” concept has its limitations is shown by a com- parison of the reactivity observed in Table IV with that reported by Foster and Avery and in previous papers of this series. The reaction indicated between Iris chrysophoenicia and I. pseudacorus is strictly negative according to our results but (-+ ++) according to Foster and Avery. Similar discrepancies exist in the behavior of Solanum lyco- persicum and Nicotiana alata which in work earlier reported (3) were negative to oleaceous forms but were found (Table IV) in this series of experiments to be positive. It would seem that in order to define the constitution of an unknown plant according to the suggestion made in paper III of his series (p. 185), it is necessary in every new series of experiments to redetermine critically the constitution of “test extracts.” To go one step further, would it not be more satisfactory to submit the few reactions found to detailed biochemical analysis in order eventually to substitute for the ‘‘test species” simple chemical solutions of known composition, containing only the active ingredients of the “test extracts”? The senior writer has already done this with respect to one, namely the calcium oxalate reaction. A relatively small number of such solutions would enormously simplify the precipitation tech- nique and would give results far more accurate and explainable than those thus far obtained. The inquiring botanist at this point must turn to the skilled biochemist for aid in resolving this problem. It is essential now that the precise nature of the precipitation reactions in plants be made the subject of investigations by someone adequately trained in analytical organic chemistry; the results of his research could not fail to be of value in advancing our knowledge of this phase of plant systematics. LITERATURE CITED 1. Cnrester, K. S. (1931). Graft-blight: A disease of Lilac related to the employment of certain understocks in propagation. Jour. Arnold Arb. 12: 6. pe ( . Studies on the precipitin reaction in plant I. The specificity of the normal precipitin reaction. Ibid. 13: 52-74 1933] CHESTER, ABBE AND VESTAL, PRECIPITIN REACTION 407 3. —————- and T Me WHITAKER (1933). Studies on the precipitin reaction in 1 plants. III. A biochemical analysis of the “normal pre- cipitin reaction.” Ibid. 14:118- 4. Encter, A. and K. Pranti (1895). Die naturlichen choad ta usw. III. Teil, ‘Abt. 6 und 6a. Lei eipsig, W. Engelma 5. Foster, R. C. and G. S. Avery (1933). Parallelism are precipitation reactions and breeding results in the genus /ris. I. Preliminary study and correlation with other evidence. Bot. Gaz. 94:71 728. 6. Kostorr, D. (1929). Acquired immunity in plants. -Genetics, a| . REHDER, A. (1927). Manual of cultivated trees and shrubs. New rk, Macmillan. 8. WinxKLER, H. (1904). Betulaceae in Das Pflanzenreich, IV, 61. N LABORATORY OF PLANT PATHOLOGY, ARNOLD ARBORETUM, AND HARVARD BIOLOGICAL INSTITUTE, Harvarp UNIVERSITY. 408 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV THE ARNOLD ARBORETUM DURING THE FISCAL YEAR ENDED JUNE 30, 1933 THE YEAR 1932-33 proved favorable to the growth of all collections in the arboretum and fruticetum. Late in February a heavy snow-fall caused considerable damage to Junipers, Spruces and small deciduous trees. Four hundred forty-two plants were added to the collections; these were taken from the Arborway nursery. Exchange of plants, cuttings, scions and seeds were continued and during the year 798 cuttings and scions, 2213 plants and 817 packets of seed were sent out. There were received from institutions in this country, Canada, Europe and Australia 1091 packets of seed, 1556 plants and 158 cuttings and scions. Eleven hundred and thirty visitors registered at the Administration Building. Thirty-seven states in the Union were represented and also British Columbia, Nova Scotia, France, Switzerland, Holland, Germany, England, Ireland, China, Japan, Philippine Islands and Australia.— Pathological Laboratory.—The Arboretum’s research laboratory in Plant Pathology submits its report for 1932-3, making reference to certain new diseases, extension services and investigative activities. One of the important functions of the pathologist’s office is to advise of diseases that loom on the horizon, that is, of potentially devastating diseases which once established would entail heavy losses and constant warfare. An example of what the Arboretum has done in this connec- tion is afforded by the warnings broadcast and the testing of effective control measures against the beech bark disease, a scourge that has within the last decade destroyed a high percentage of the beech trees in the adjacent Maritime Provinces and now is advancing into Maine. The predisposing factor of the disease, a bark-infesting insect similar to the woolly aphis, was discovered about four years ago by Mr. John Ehrlich of this laboratory to be sporadically distributed throughout metropolitan Boston. Subsequently, it was located elsewhere in east- ern Massachusetts by other observers. Through the united efforts of the Arboretum and the Park Department it has been largely eradicated within the Boston area. 1933] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 409 Two other diseases now demand notice, one an affection of Firs known as the “Gout disease,” the other of Elms, the so-called “Dutch elm disease.” The former was pointed out to me in Nova Scotia in 1929 by the Provincial Forester, Mr. Otto Schierbeck. Quoting from the Journal of the Arboretum, XI: 57. 1930 (Faull, J. H.: Notes on Forest Diseases in Nova Scotia)—‘Standing out as perhaps most interesting of all is a widespread condition of Balsam Fir, involving practically all of the trees of entire stands, what I would designate for want of a better term as ‘Gout.’ The trees are stunted, the trunks taper rapidly and never reach normal height, the joints are very much swollen and the twigs in general are thickened and tend to be deflexed.” To this might now be added that many trees are soon killed outright and others are grossly mis-shapen. At that time the etiology was not known, nor was it certain that a contagium was responsible. Through the researches of Mr. R. E. Balch of the Dominion Entomological Branch it has recently been determined that the cause is a bark insect of the woolly aphis type. In a recent eastern trip I observed that spread has been rapid and destruction extensive. It is found at present as far south as Brunswick, Maine. I am informed by Mr. H. B. Peir- son, Entomologist for the State of Maine, that control is entirely prac- tical in ornamentals by use of a contact insecticide. It is known to attack Abies balsamea, A. arizonica, A. Fraseri, A. nobilis, A. glauca, A, sibirica and certain other Firs. Dutch elm disease is an even more serious malady because among the highly susceptible hosts is our much-beloved and widely-grown American Elm. From tests made in Holland on plants supplied by the Arboretum, Dr. C. Buisman has demonstrated that the American Elm is among the most vulnerable of the Elms. This disease is wide-spread throughout Europe and it has worked havoc there. Three years ago a few cases were discovered in Cleveland and. one in Cincinnati; they were promptly extirpated. Report this year comes of an outbreak in New Jersey and southern New York which appears to be of much more serious proportions. The Arboretum is intensely interested in the prob- lems of this and other elm diseases and will participate so far as its resources permit in efforts towards their solution. The extension work of the laboratory, though more or less incidental to its main purpose, has continued to grow. A great many inquiries are received, and these are generally accompanied by specimens for diagnosis. To care for this phase of our work the necessity for ade- quate assistance in the laboratory, especially during the growing sea- 410 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV son, is a matter of importance. While much of the material that comes in to us can be readily disposed of, frequently more or less critical material is received that requires extended attention. Particularly has this been the case this year because of uneasiness among so many of our correspondents regarding the possibility of the presence of the Dutch elm disease, the certain diagnosis of which involves rather tedious laboratory technique. Consultation is welcomed, partly because of service that may be rendered, but also because unusual problems through this medium are from time to time disclosed. The research undertakings of the laboratory have materially ad- vanced during the year, yielding results of scientific interest and of economic application. A summary follows. Professor J. H. Faull, continuing his studies on the rust fungi that attack conifers, has added to the literature a monograph on the hitherto little understood genus Milesia (Contributions from the Arnold Arboretum, No. II). These curious rusts, so far as has been deter- mined, alternate between Firs and ferns. The monograph rectifies the tangled nomenclature, outlines the geographical distribution of species of Milesia, the world over, describes thirty-thrée species and two vari- eties, delineates thirteen new species and varieties, and records for the first time the life histories of three species of the nine whose life cycles are now more or less completely known. Dr. K. S. Chester, on leave of absence in Europe on a Sheldon Fel- lowship for nine months of the year, has continued to publish on his work on precipitation reactions—on a biochemical analysis of these reactions and on the question of acquired reactions due to grafting. Most of his time has been devoted to an investigation of bacteriophage in relation to crown gall. Dr>John Ehrlich has brought to a conclusion his immediate studies in America on the “beech bark disease’ to which reference was made above. He now goes to Europe on a National Research Fellowship to study the problem at its original sources. Miss L. M. Hunter has advanced her studies on the spermogonia of coniferous rusts and will continue her problem in Europe during 1933-4 supported by a traveling Fellowship. Mr. I. H. Crowell has completed two years of intensive investigation on the Gymnosporangium rust diseases of Juniperus and the various Pomaceae other than Hawthorns. This work has been jointly supported by Dr. and Mrs. Henry Lyman and the Arboretum. Mr. J. D. Mac- Lachlan is making parallel studies on Juniperus and the Hawthorns. 1933] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 41] Mr. A. B. Hatch has devoted himself to a study of the mycorrhizae of Pines. He has made a notable contribution to the technique em- ployed in the investigation of the biology of these remarkable structures and has published the first adequate account of the mycorrhizae of the White Pine. His investigations are conducted jointly under the aus- pices of the Harvard Forest and the Arboretum and have been made possible through the liberal generosity of an anonymous benefactor. — Cytological Laboratory.—During the past year the cytological in- vestigations have covered a wide field, including considerable work on the mechanism of chromosome pairing and division, as well as the cyto- logical analysis of various families and genera from the standpoint of relationships and origins. An analysis of the mechanism of crossing over was presented before the International Genetics Congress and will appear in the Proceedings. The behavior of segmental interchange chromosomes in Tradescantia has been studied and the results published in Genetics. A detailed study of chromosome pairing in Larix was completed by Mrs. Sax and published in Genetics. Mr. O’Mara completed his work in chromo- some division in the pollen tube of Lilium and it has been published in the Botanical Gazette. Dr. Dermen’s work on the origin and behavior of the nucleolus has been completed. A study of chromosome pairing in Paeonia has thrown some light on the mechanism of chromosome association. Cytological work in connection with species and generic relationships has been completed for Yucca and Agave, Ulmus, Cornus, Acer, Mag- nolia and a considerable number of conifers. Yucca and Agave seem to be more closely related than the taxonomic grouping would indicate, but Dr. Dermen’s work on Cornus, Mr. Foster’s work on Acer, the chromosome numbers found in Ulmus and Celtis, the results obtained in the conifers by Mrs. Sax, and Dr. Whitaker’s work on Magnolia and related genera all show a close correlation between chromosome char- acteristics and taxonomic grouping. Closely related families often have similar chromosomes. In the conifers most genera are different in chromosome morphology, but species within each genus have similar genoms. The cytology of hybrids between Old and New World species have been studied in species hybrids of Larix, Platanus and Campsis. The Larix and Platanus hybrids are completely fertile and the Campsis 412 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV hybrid partially fertile, even though the original parental species in each case are morphologically distinct, and have been isolated from each other for very long periods of time. Breeding work has been continued with Roses, Lilies and conifers. Crosses have also been tried between different genera of Pomoideae, some of which seem to be successful. In general the breeding work has been confined to crosses between closely related species. — K. S. The Herbarium.—During the past fiscal year 16,377 mounted specimens have been added to the herbarium, bringing the total num- ber up to 374,880 specimens. Of these accessions approximately 3800 came from the United States and Canada, 4000 from Central and South America inclusive of Mexico, 600 from Europe and western and central Asia, 1900 from eastern Asia, 1700 from southern Asia and Malaysia, 1200 from Africa, 2400 from Australasia and 400 were cul- tivated plants. Among the more important collections received during the year the following may be mentioned: 3400 specimens collected by J. and M. S. Clemens on Mount Kinabalu, Borneo; 1270 specimens from Lingnan University collected chiefly in Hainan and southeastern China; 1100 specimens from H. H. Chung, Wuhan University, collected mostly in Fukien and Hupeh; 3350 South American specimens from the U. S. National Herbarium including a large number of Argentine plants col- lected by Venturi; also from the U. S. Herbarium 270 specimens col- lected by J. F. Rock in Yunnan and Burma; about 1000 numbers of Australian specimens collected by C. T. White, S. F. Kajewski and L. J. Brass; nearly 1000 numbers with about four sets of duplicates of plants from the Solomon Islands collected by L. J. Brass; 950 speci- mens of Kweichou plants from the University of Nanking; 900 speci- mens comprising 18 fascicles of the Herbarium Florae Rossicae from the Botanical Institute in Leningrad; 600 Brazilian plants collected by B. A. Krukoff; 575 plants from Colombia collected by A. E. Law- rance; over 550 plants from Trinidad collected by W. E. Broadway, 650 South African plants collected by F. A. Rogers; 1500 specimens chiefly from Indochina and Madagascar received from the Museum of Natural History at Paris; 195 specimens from Indochina collected by R. W. Squires. To the fruit collection 370 specimens were added bringing the number up to 7815. Additions of 451 specimens were made to the wood collection bring- ing the total up to 2816. 1933] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 413 The collection of negatives of types and other important herbarium specimens consists now of 2524 negatives; 667 having been added, mostly types of Chinese species taken by the Curator in European herbaria. Besides constantly using the herbarium in the determination of plants sent in for identification and also in the determination of large collec- tions chiefly from eastern Asia and North America members of the staff have engaged in special work; Mr. E. J. Palmer is continuing his studies in the genus Crataegus; Dr. I. M. Johnston has published extensive notes on the Boraginaceae of the western United States; Mr. A. Rehder is continuing the revision of the ligneous plants described by H. Lé- veillé from Eastern Asia and Mrs. Susan D. McKelvey her studies in the genus Yucca. Among the visitors who have worked in the her- barium may be mentioned Dr. Shun Ching Lee of the National Nor- mal University, Peiping, China, who has spent five months here in the preparation of a work on the forest trees of China, Dr. R. E. Wood- son, Jr., of the Missouri Botanic Garden studied Apocynaceae and Mr. E. H. Walker, of the National Herbarium, Washington, Chinese Myrsi- naceae; also Dr. L. H. Bailey of Ithaca, Dr. L. Diels, director of the Berlin Botanic Garden, Dr. S. F. Blake, Mr. W. W. Eggleston and Paul Russell of the Department of Agriculture, Washington, and Dr. A. Gundersen of the Brooklyn Botanic Garden consulted the herbarium. For study outside the Arboretum herbarium 586 specimens were lent to 18 institutions and individuals in this country, Europe and China. There have been distributed 14,536 specimens to 38 institutions in the United States, Canada, Europe, Asia, Africa and Australia. Botanical exploration by members of the staff or by expeditions partly or wholly financed by the Arnold Arboretum has been carried on in both Americas, Eastern Asia and Australasia. Dr. H. M. Raup and Mr. E. C. Abbe started in June 1932, as men- tioned already in the last report, on a tour of botanical exploration of the Peace River region in the Canadian provinces of Alberta and British Columbia with financial aid from the National Research Council along with the support of the National Museum of Canada; they returned toward the end of September, having collected about 6400 specimens representing 1200 numbers. An account of the flora of this region has been prepared by Dr. Raup and will be published as No. VI of the Contributions from the Arnold Arboretum. Professor J. G. Jack spent the time from the end of January to the beginning of April at the Atkins Institution of the Arnold Arboretum 414 JOURNAL OF THE ARNOLD ARBORETUM [voL. XIV at Soledad, Cuba, and collected in the neighborhood of Soledad and Cienfuegos about 600 numbers with many duplicates and with fruit and wood specimens. Dr. I. M. Johnston left for England in October under a fellowship of the Guggenheim Foundation, chiefly to continue his work on the Boraginaceae and to study type specimens and other material of South American plants at Kew and the British Museum. In February he went to the Continent and worked first in Berlin and later in Paris. Mr. Alfred Rehder spent the time from the end of June to the end of September in Europe for the purpose of examining and photograph- ing type specimens of plants of China and the adjoining regions. He took about 600 photographs of types and critical specimens in the herbaria at Kew, the British Museum, Berlin, Florence, Geneva and Paris. In China botanical expeditions of three institutions have had the financial support of the Arnold Arboretum. The southern province of Kwangsi, botanically as yet very little known, is being explored by an expedition from the University of Nanking under the direction of Dr. A. N. Steward. The Fan Memorial Institute of Biology at Peiping has, since spring 1932, an expedition collecting in the less well known parts of Yunnan; it also maintains an expedition in cooperation with the Academy of Science of Western China in Szechuan, which collected during 1932 about 2500 specimens and is in the field again this year under the direction of Dr. H. H. Hu. Lingnan University had under the direction of Dr. F. P. Metcalf during the second half of last year and the first half of this year an expedition in the field collecting in Hainan and Kwangtung which had up to the end of last year collected about 2125 numbers; another expedition started in January for Kwangsi and will remain there until July. In the Solomon Islands Mr. L. J. Brass, who had already collected for the Arnold Arboretum in New Guinea from 1925 to 1926, con- tinued the work of S. F. Kajewski and visited during the time from July 1932 to the beginning of January 1933 San Cristobal and Ysabel Island and also several smaller islands; he collected nearly 1000 num- bers with approximately four sets of duplicates. In February he joined the Archbold Expedition to New Guinea. In the summer of 1932 Dr. C. Regel, Director of the Botanic Garden at Kaunas, Lithuania, undertook a collecting tour to Asia Minor, partly financed by this Arboretum. Owing to the attitude of the Turkish authorities this trip was not very successful. 1933] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 415 From April to the end of June 1932 Professor J. Bornmiiller of Weimar made a successful collecting tour to Tripoli, Sicily and south- ern Italy with financial help from the Arnold Arboretum. — A. R. The Library.—At the end of June 1933 the Library comprised 40,919 bound volumes, 10,085 pamphlets and 16,898 photographs, a gain of 271 volumes, 200 pamphlets and 112 photographs during the year. The increase, due to conditions, was smaller than in previous years but, nevertheless, presents an interesting breadth of inclusion, and more time could be given to analyzing important articles from periodicals and other works, thus enriching the material easily available. Though few photographs have been added, the use of the collection has been extensive. Nearly 100 prints have been made and sold for nursery catalogues, post cards, publications, and collections institu- tional and private. Eleven photographs of Lilies were loaned to the Royal Horticultural Society of London for the Lily Conference. Eighty-eight prints of photographs taken in China by the late Dr. E. H. Wilson were made for Dr. Shun Ching Lee of Peiping, China, besides prints made for our own publications, and many photographs were used in lectures given by members of the staff. Eight colored slides reproduced from photographs in the collection have been added to the lantern slide collection and seven slides were made for the New York State College of Forestry. Slides used during the year number over 300. Cards filed during the year include 737 in the card catalogue of books in the Library, 119 in the catalogue of photographs, 4,598 in the “Card-index of New Genera, Species and Varieties published by the Gray Herbarium,” and 4,900 in the manuscript “Index of Illustrations and of New Genera, Species and Varieties of Ligneous Plants published since 1915,” prepared at the Arboretum, bringing the total number of the latter to 97,639. A very large amount of time has been spent in reading proof for the “Catalogue of the Library of the Arnold Arboretum of Harvard Uni- versity, Volume iii, Serial Publications—Authors and Titles, Supple- ment, 1917-1933,” which has now reached completion and is ready for distribution. It comprises 346 pages and approximately 17,300 entries arranged alphabetically, with numerous references. Nearly 900 slips have been filed for entries received too late for inclusion in the printed Catalogue and 300 slips for a supplement to the “Subject Catalogue,” which is now in preparation. 416 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Volumes bound number 113, as against 570 of last year, and nearly 100 smaller books and pamphlets were put into pamphlet binders. Clipping files and scrap-books, with references in the card catalogue, bring together and preserve much interesting and valuable material which would otherwise be lost. A large number of visitors have registered in the Library during the year. Dr. Shun Ching Lee of the National Normal University, Peiping, China, spent several months studying the forest trees of China. Pro- fessor L. H. Bailey and Miss Ethel Zoe Bailey, of Ithaca, New York, Mr. E. H. Walker, of the Smithsonian Institution, Dr. S. F. Blake and Mr. Paul Russell of the United States Department of Agriculture, Dr. R. E. Woodson, Jr., of the Missouri Botanical Garden, Dr. Conway Zirkle of the University of Pennsylvania, and Miss Elsie Jack of Hatzic, British Columbia, spent some time consulting the Library. Among other visitors were Mr. Frederick A. Delano to whom the Library owes its fine collection, previously described, of Chinese paintings of fruits and flowers, Dr. David Fairchild of Cocoanut Grove, Florida, and Mr. and Mrs. Arthur B. Spingarn of Amenia, New York. The publications of the Arboretum, the “Journal of the Arnold Arboretum” and the “Arnold Arboretum Bulletin of Popular Informa- tion,” were issued regularly; of the “Contributions from the Arnold Arboretum of Harvard University” numbers II-IV were published during the year. Of the approximately 400 periodicals that come to the Library from all parts of the world, nearly 250 were received in exchange for these publications. More than SO new periodicals have been received during the year, a large number in exchange for our publications and for herbarium speci- mens, some by gift and a number by purchase. They include: AALSMEER—Vereeniging tot oprichting en instandhouding van den proeftuin. Jaarverslag. 1931. Aalsmeer. 1932. AMERICAN civic annual. Vol. i-iv. Washington. 1929-32. AMERICAN nurseryman. Vol. lvii, no. 2 — Rochester. 1933 —> AMERICAN orchid society bulletin. Vol. i, no. 1 — Washington. 1932 > AMERICAN rose magazine. Vol. i, no. 1 — Harrisburg, Pa. 1933 > BoLeTiIN de pro-cultura regional, S. C. L. Tom. i, n. 1-24, 28-30. Mazatlan. 1929-32. BULLETIN of applied botany, of genetics and plant-breeding. Ser. A. Socialistic plant-industry. No. 1 — Leningrad. 1932 —> 1933] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 417 BULLETIN of applied botany, of genetics and plant-breeding. Ser. ii. Genetics, plant-breeding and cytology. No. 1 — Leningrad. 1952 => FIELD museum news. Vol. ii, no. 10, 12; iii, 1-3, 5-11; iv, 1 >= Chicago. 1931 — FioratiA. Vol. liv, no. 13 — Haarlem. 1933 — Forest log. Vol. i, no. 10-12; ii, 1; ili, 2 Salem, Oregon. 1931 —> ForRTSCHRITTE der botanik. Bd. i, 1931. Berlin. 1932. GARDEN lover. Vol. ix, no. 1 — Melbourne; Sydney. 1931 — HARVARD UNIVERSITY—Farlow herbarium of cryptogamic botany. Contributions. Vol. [i] — Boston, etc. 1883 — (Incomplete) HEMLOCK ARBORETUM at “Far Country.” Bulletin. No. 1 — Ger- mantown, Pa. 1932 — Ipano forester. Vol. xiii — Moscow, Idaho. 1931 — KAKTEEN-KUNDE. Heft 1 — Berlin. 1933 — orms new series of “Monatsschrift der Deutschen kakteen-gesell- schaft.” Kaunas—Universitas Vytauti Magni—Hortus botanicus. Scripta. Tom. i — Kaunas. 1931 > Krvu—Laboratoire des produits végétaux et de Vherbier du Service forestier, Communications. No. 1 — Bruxelles. 1930 — LENINGRAD—A kademiya nauk—Byuro po evgenike. Izvestiya. No. 1-9. Petrograd. 1922-32. MakILING echo. Vol. xi, no. 2,4. Laguna. 1932. MINNESOTA studies in plant science. No. 1 — Minneapolis. 1932 > NATIONAL nut news. Vol. xv, no. 1 — Chicago. 1932 — Revista espafiola de biologia. Tomo i, cuaderno 1 — Madrid. iva2 == REVUE agrologique et botanique du Kivu. No, 1 — Bruxelles. SAGHALIEN CENTRAL EXPERIMENT STATION. Reports. Ser. 2. For- estry. No. 1 -> Konuma. 1932 — SOCIEDAD ESPANOLA DE HISTORIA NATURAL. Resenas cientificas. Tom. vi, num. 1 — Madrid. 1931 — TarHoku—IJmperial university—Faculty of science and agriculture. Annual report of the Taihoku botanic garden. No. 1 — Taihoku. a) => Toxyo bunrika daigaku. Science reports. Section B. Vol. i, no. 4-10. Tokyo. 1932. Warsaw—Université—Institut de botanique systématique et de phytogéographie. Publications. No. 2, 6, 7, 10, 12, 13, 15-17, 19, 20, 22-25, 27 — Cracovie, etc. 1927 — 418 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV On November 8, 1932, fifty-five periodical titles were sent for pub- lication in the “Union List of Serials.” Among other important accessions are: AEs, Mrs. Blanche Ames. Genera of the Gymnospermae, [ Mono- cotyledoneae, Archichlamydeae and Metachlamydeae or Sym- petalae|, with their more important economic species arranged pss Engler & Gilg. N. Pp. [1916-17.] 4 ang asi satieer sarin of charts made by ‘Mrs. Ames for the use of clas in economic bot any at Harvard Teivarsien: showing i ne ee “Of economic nite given in the form of family [Iwasak1, Tsounemassa. Phonzo zoufou. “Inder. 1, 2. Yedo. 1916?| LaNCKMAN, C. Catalogue d’une belle et riche collection de rosiers a fleurs doubles; ainsi que toutes d’arbres a fruit, plantes de bruyere, d’orangerie, et un grand nombre d’arbustes et plantes de pleine terre. Gand. Lowe, Edward Joseph. Beautiful leaved plants. By E. J. Lowe assisted by W. Howard. London. 1864. 60 colored plates. CONGRES INTERNATIONAL DE BOTANIQUE ET D’HORTICULTURE, Paris, 1878. Comptes rendus. Paris. 1880. SCHWIMMER, Johann Michael. Ex physica secretiori curiositates, non minus utiles, quam jucundae. Jenae. 1672 “A collection of very remarkable discourses on various subjects of natural history. ost remarkable is chapter vii, ‘“Conjugium et sexus duplex vegetabilium,” which proves Schwimmer to have been a mentioned by the ae is that b y Camerarius, 1694, i twenty-two years later t ook.” | Evxrotrt, Stephen. ae manuscript. 1810-14.] This appears to be an early draft of his “Sketch of the botany of South Carolina and Georgia,” which was much altered and enlarged before printing. ——————_ Letters to, and other manuscripts by, Stephen Elliott. 1790-1829. ] -These manuscripts cc sae with the preceding, given to Professor C. S. Sargent many year have but recently come to light and they prove to be of ee interest. The letters comprise thirty from Henry Muhlenberg, 1808-1815, written in a very fine hand on foolscap paper, one from Zaccheus Collins, 1815, informing Elliott of Muhlenberg’s death. tito are from piace Darlington, 1827, twenty-four from Dr. James MacBride, one from eae Arnott Walker Arnott, 1828, one ree William nat 1828. gcd -seven letters are from S. Boykin, F. Boie, J. Vau ughan, Wm. Sw n, 1933] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 419 Wm. Thouin, Thomas Say, Dr. Lewis Schweinitz, John Brace, and others. An interesting short letter is from John Abbot, 1817, Tre- stou silversmith in London, and carry Francillon’s name, book stamp and printed title-pages, dated 1792-1804. There are also volumes of them in the museums of Oxford, Paris, Zurich and elsewhere. “Lists of books for the Charleston Library peace by Stephen Elliott, Esq.,” invoices, and letters signed by ichaux, lists o Georgia, by Stephen Elliott,’ with signatures of subscribers, com- plete the collection of miscellaneous manuscripts In: addition to the Elliott letters the Library possesses several thou- sand letters from various writers addressed to different members of the staff of the Arnold Arboretum. These comprise letters from many eminent botanists, from collectors during their expeditions for the Arboretum, and other important letters on botanical matters. To arrange these satisfactorily will be part of the undertaking for the coming year The collection of nursery catalogues is becoming an important de- partment and an effort has been made to obtain the latest catalogues of as many firms as possible. A large number of books were sent out as inter-library loans to Gray Herbarium, Bussey Institution, Harvard Medical School, Harvard Bio- logical Laboratory, Harvard Forest, Harvard College, United Fruit Company, Michigan State College, University of Iowa, West Virginia University, Canada Department of Mines, Tufts College, Marine Bio- logical Laboratory, Yale Forestry School, New York State College of Agriculture, Dartmouth College, Smithsonian Institute, Massachusetts Institute of Technology, McGill University, University of Pennsyl- vania, American Museum of Natural History, Antioch College, Uni- versity of New Hampshire, Massachusetts State College, Texas Agri- cultural Experiment Station, Forest Products Laboratory, Madison, Wisconsin, University of Toronto, and Wellesley College. In addition to loans, photostats or typewritten copies of references have frequently been made when books could not be loaned. Fifteen books were borrowed for members of the staff from the libraries of the United States Department of Agriculture, Cornell Uni- versity, Ohio State University, Harvard College, Pennsylvania Academy of Natural Science, Massachusetts Horticultural Society, and Gray Herbarium. — E. M. T 420 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XIV Bibliography of the published writings of the staff and students July 1, 1932-June 30, 1933 ALLEN, = Kathryn. nograph of the American species of the genus Halenia. (In pee of the Missouri botanical garden, 1933, xx, 119-222.) AMES, Oakes The Arn old arboretum ; Mary 1931-32). (In Journal of the Ar- nold arboretum, 1932, 447-448. The ‘ ttleyeae. (In eae orchid society bulletin, 1932, i, An extension of range for Epidendrum rhynchophorum. (In Har- vard university, Botanical museum leaflets, 1932, no. 3, pp. Insect visitors to rd eee parviflorum. (In American orchid society bulletin, 1932, i, John Lindley (1790-1868). “(In American orchid society bulletin, 1932, i, 34-36. New or notewo rthy Philippine orchids. iii. (In Philippine journal :. sctence, 1932, xlix, 483-504. ee of Laelia Digbyana. (In American orchid society bulletin, 1932, i, 59. An addition to the genus Epidendrum. (In Harvard university, Botanical museum leaflets, 1933, no. 7, pp. 1 An addition to the genus Hexadesmia. (in Harvard university, Botanical museum leaficts, 1933, no. 6, pp. Additional notes on the orchids of the New as and Santa Cruz Islands. (In Journal of the Arnold arboretum, 1933, xiv, 101-1 The Hondur of Lepanthes. (In Harvard university, pom, ae ign 1933, no. 4, pp. 3-8. A new Bletia from Spanish Honduras. ow Harvard university, -) A new Epidendrum from Honduras. Cin tina university, Botan- ical museum leaflets, 1933, no. 8, pp. 2 Octomeria from Spanish an sueae: (ts Harvard university, Botanical museum leaflets, 1933, no. 4, pp. A new Oncidium from Honduras. (In Harv ane university, Botani- cal museum leaflets, 1933, no. 5, pp. The Reichenbachian gate um. (In pee orchid society bulle- } 8-99. ) Botanizing from an airplane. By Edgar Anderson and Oliver Ames. (In Arnold arboretum bulletin of Pie information, 1932, vi, Growin ng oranges in pe (In Arnold arboretum bulletin of popu- lar information, 1932, vi, 45-47. The propagation of woo a plants. By Edgar Anderson and W. H. Judd. (In Arnold arboretum bulletin of popular information, 1932, Basswood bark and its use by the Indians. (In Arnold arboretum bulletin of popular information, 1933, i, 33-37. Color variation in a Missouri colony of Hepatica acutiloba. (In Rhodora, 1933, xxxv, 66-67.) 1933] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 421 eee of Iris versicolor in relation . the post-glacial Great LB: (In Rhodora, 1933, xxxv, 154-160.) aeons [a discussion of bo tanical ere or Arnold arbore- tum bulletin of popular information, 1933, Leatherwood. (In Arnold arboretum bulletin of Hie informa- The River Birch (Betula nigra). By Edgar Anderson and E. Palmer. (In Arnold arboretum bulletin of popular information, 6-8. Segmental interchange in chromosomes of Tradescantia. By Karl 7 pe bulletin of popular information, 1933, i, 5 Trumpet-creepers. (In Arnold arboretum bulletin “of popular in- Vicinism in Aquilegia vulgaris. (In American naturalist, 1933, Ixvii, 190-192. ) BAILEY, Irving Widmer. The cambium and its derivative tissues. vii. Problems of identify- ing the wood of mesozoic Coniferae. (In Annals of botany, 1933, xlvii, 145-157. Investigation on cambium and its derivative tissues. (In Yearbook of the Carnegie institution of Washington, 1932, xxxi, 191-192. Preliminary notes on cribriform and vestured pits. (dl n Tropical d 46-48. ) The problem of acquired te ee immunity in plants. i. (In Ouarterly revue of biology, 1933, viii, 129-154. Studies on the precipitin penutis in ’ plants. iii. A biochemical analysis of the normal precipitin reaction. By Chester and W itaker. (In Journal of the Dosis arboretum, 1933, xiv, 118-197 tudies on the pr recipitin reaction in plants. iv. The question of acquired reactions due to grafting. By T. W. itaker an Chester. (In American journal of a ee 1933, xx, 297-308. ) CurrikE, Mary Parasitism of myxomycete plasmodia on fungous mycelia. Fr, 1. Howard and Mary E. Currie. (In Journal of the Arnold arbore- tum, he xiil, 438-447.) DERMEN, Hai Cytological studies of Cornus. (In Journal of the Arnold arbore- aca xiii, 410-416.) bade Joh The eh bark disease in Nova Scotia. (In Province of Nova Scotia, Report of the Department of lands and forests, 1932, pp. Nectria coccinea on beech. (In Phytopathology, 1933, xxiii, 10.) 422 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XIV FAuLL, Joseph Horac Arnold ccs au Pathological mecgeen ae Mt 1931-32). (In a of the Arnold arbore 1932, 448-450. onomy and geographical phe ore of the genus Miles (In C ontributions from the Arnold arboretum of Harvard sy HatTcH, Alden Bruce. ycorrhizal and other features of the root systems of Pin By A. B. Hatch and K. D. Doak. (In Journal of the Arnold pres tum, 1933, xiv, 85-99. Pure-culture technique for quantitative studies of plant growth in association with microorganisms. (In Phytopathology, 1933, xxiii, ) True mycorrhizal fungus in contrast to Mycelium radicis atrovirens. In ihe aie 1933, xxiii, 14.) Howarp, Fra . aren on fungous mycelia. By F. L. Howard and Mary a urrie. (In Journal of the Arnold arboretum, 1932, xiii, 8-44 Jounston, Ivan Murray. Studies in the Boraginaceae. ix. (In Contributions from the Ar- nold cpg Sek of Harvard university, 1932, ili.) Jupp, William enry. The Great Tree of Tule. = so chronicle, 1932, xcii, 422.) ardy coniferous esl or Engla - (In Ga rdeners’ chronicle of Am B1032, XXXVI, 307-308, 3 The propagation of wo a plants. By Edgar pean and W. H. = (In Arnold arboretum bulletin of popular aiemton. 1932, i, 53-56.) Raising the Dove Tree from seed. (In Horticulture, 1932, x, 386.) KELLEY, Katharine Eleanor Bibliography of the published writings of the staff and students July 1931—June 30, 1932. (In Journal of the Arnold arboretum, 1932, xiii, 458-461.) Kinc, Edward. oe behaviour in a triploid Tradescantia. (In Journal of heredity, 1933, xxiv, 253-256. ) Kosusk1, Clarence Emmeren. Enumeration of the oats plants collected by J. F. Rock on the Arnold arboretum expedition to northwestern China and north- eastern Tibet (additions and continuation). By Alfred Rehder and C. E. Kobuski. (In Journal of the Arnold arboretum, 1932, xiii, 385-409. ) An enumeration of the sme ger plants collected by J. F. Rock for the Arnold arboretum. By Alfred Rehder and C. E. Kobuski. (In Journal of the hea arboretum, 1933, xiv, 1-52.) McKe vey, Mrs. Susan Delan A white pine blister rust Guesteiaton. (In Horticulture, 1932, x, Taxonomic and cytological relationships of Yucca and Aga By Susan D. McKelvey and Karl Sax. (In Journal of the ap Pees: arboretum, 1933, xiv, 76-81.) O’Mara, Joseph. Division of the e generative nucleus in the pollen tube of Lilium. (In Botanical gazette, 1933, xciv, 567-579. ) 1933] THEARNOLD ARBORETUM DURING THE FISCAL YEAR 423 PaLMER, Ernest Jesse. rataegus arnoldiana, Arnold Thorn. (In Addisonia, 1932, xvii, 33- The Crataegus problem. (In Journal of the Arnold arboretum, 1932, xiii, 342-362. The ferns and fern allies of Missouri. By E. J. Palmer and as Steyermark. n American fern journal, 1932, xxii, 105-122 s note book. (In Journal of the Arnold arboretum, 1932, xiii, 41 The River Birch (Betula nigra a). By Edgar Anderson and E ie (In Arnold arboretum hulle a popular information, 1933, Some eae southern trees and shrubs. (In Arnold arboretum bulle- tin of popular 1% informatio n, 1933, i, 45-49. Trees used by the pioneers. (In Arnold arboretum bulletin of popu- lar information, 1933, i, 38-39.) Raup, Hugh Miller Lichens. (In “Arnold arboretum bulletin of popular information, 1933, i, [9-16].) Range ‘conditions in the Wood Buffalo Park 2 pobats Canada with notes on the history of the wood bison. (In American committee a international wild life protection. Special PiBRCAHOn, 1933, — "Alft ed. Arnold arboretum Herbarium; (report 1931-32). (In Journal of the Arnold arboretum, 1932, xiii, 451-453.) Enumeration of the ligneous plants collected by J. F. Rock on the Arnold arboretum expedition to northwestern China and _north- eastern Tibet (additions oer continuation). By Alfred Rehder and C. E. Kobuski. (In Journal of the Arnold arboretum, 1932, —409. ew species, varieties and combinations from the herbarium and the collections of the Arnold arboretum. (In Journal of the Arnold A second species of Gossypiospermum. (In as woods, 1932, 6-7. no. 32, pp Uniformity in plant names. (In Horticulture, 1932, x, 460.) An enumeration of the herbaceous plants coll y J. F. Rock for Jou SAX, Hally oes. Chromosome pairing in Larix species. (In Journal of the Arnold arboretum, 1932, xiii, 368-374 Chiasma formation in Larix and Tsuga. (In Genetics, 1933, xviii, 121-128.) Sax, Karl. ‘Arnold arboretum Cytological laboratory (report 1931-32). (In Journal of the Arnold arboretum, 1932, xiii, 450-451.) 424 JOURNAL OF THE ARNOLD ARBORETUM [VOL XIV Chromosome relationships in Pomoideae. (In Journal of the Arnold —367. Meiosis and chiasma formation in Paeonia i aaa (In Journal of the Arnold arboretum, 1932, xiii, 375-3 Recent advances in cytolo ogy. By C.D. Picci (Review. ) (In Collecting net, 1932, vii, 201-203.) Chromosome numbers in Ulmus and related genera. (In Journal of the Arnold arboretum, 1933, xiv, 82-84. sy aon interchange in c romosomes of Tradescantia. By Karl Sax and Edgar aca: n Genetics, 1933, XVili, ») yearn ey cytological relationships of Yucca and Agave. By Susan D. McKelve ey an arl Sax. (In Journal of the Arnold arboretum, ae xiv, 76-81.) sae Ethelyn Mar nold ee eae i alpha (report 1931-32). (In Journal of the pe arboretum, vie xiii, 453-458. ) WHITAKER, Thomas Wallac Fertile gourd-pumpkin pobade (In Journal of heredity, 1932, xxiii, —430.) The chromosome complement of Cyphomandra betacea. (In Journal of the Arnold pea ane 1933, xiv, 113-117. Cytological an phylogenetic studies in the Cucurbitaceae. (In 7 90. Studies on the precipitin reaction in plants. iii. A catia analysis of the normal precipitin reaction. By K. S. Chester and T. W. Whitaker. (In Journal of the Arnold so a ‘1933, xiv, 118-197.) Studies on the ‘sig age Agia in plants. iv. The question of acquired reactions due to grafting. By T. W. Whitaker and K. S. Chester. (In American ig es of ee. soma xx, 297-308. ) The above articles cover a total of about 882 pag K, E. K, Staff of the Arnold Arboretum, 1933-34 Oaxes Ames, A.M., Arnold Professor of Botany, Supervisor. Joun GEorGE Jack, Assistant Professor of Dendrology. ALFRED REHDER, A.M., Curator of the Herbarium. JosepH H. Fautt, Ph.D., Professor of Forest Pathology. IRVING WIDMER BaItey, Sc.D., Professor of Plant Anatomy. Karv Sax, Sc.D., Associate Professor of Cytology. Epcar ANDERSON, Sc.D., Arborist. IvAN Murray JOHNSTON, Ph.D., Research Associate. CLARENCE E. Kosusk1, Ph.D., Assistant Curator, Herbarium. ERNEST J. PALMER, Collector and Research Assistant. CaROLINE K. ALLEN, Ph.D., Assistant in the Herbarium. Hucu M. Raup, Ph.D., Research Assistant. Hatc DERMEN, Ph.D., Research Assistant. Ivan H. Crowe tt, A.M., Assistant in Phytopathology. ETHELYN Maria TUCKER, Librarian. Susan Detano McKeE vey, Research Assistant. ETHEL ANTOINETTE ANDERSON, Business Secretary. KATHARINE ELEANOR KELLEY, Assistant in the Library. Louis Victor SCHMITT, Superintendent. WIL1L1AM HENry Jupp, Propagator. Ww — 231, 236, ERRATA AND ADDENDA line 9 for brevicornu read pubescens line 11 for pubescens read brevicornu line 22 Bocconia spec., change to Macleaya microcarpa (Maxim.) Fedde line 7 Meconopsis spec., no. 13077 = M. racemosa Maxim? forma line 8 Meconopsis spec., no. 14430 — M. punicea Maxim. after line 6 from below insert: Gentiana Piasezkii Maxim. Southwestern Kansu: Tao River basin, no. 13186 line 14 from below refer no. 14181 to Stachys Sieboldii Miq. (fide Hand.-Mazz.) , line 4 Pedicularis spec., no. 12255 = P. semitorta Maxim. line 21 strike out (L. conglobatum X leontopodioides) lines 9 and 11 from below for 60 m. read 200 m. line 14 from below for Rhaphidophera read Raphidophora line 10 for 3/4 read 1/3 line 5 from below for 1/4 read 2/3 line 23 at end of line for chromosomes read chromosome line 14 for LxIv read LXxXIV , line 4 from below for 000 and 000 read 147 and 174 line 8 for 000 read 151 line 14 from below add Synon. nov. line 13 from below for Chickrassia read Chukrasia after line 1 from below insert: Cipadessa cinerascens (Pellegr.) Handel-Mazzetti in mb. Sin. vir. 632 (1933). lines 5, 6, 11 and 17 from below and page 229, line 2 for Cavalierella read Cavaleriella line 7 for 2659 read 3500 line 3 after Bodinieri insert Léveillé INDEX Synonyms are printed in italics; Asse, E. C., CHEsTER, K. S. ie P. , Studies on. the Seiian rea tion in plants. V. Application. 40 Bee relationships, 394 Acalypha Mairei, 234 Acer and Sta eee Chromosome num- ber in, 386, pl. Acer sect. Maca, 211 3 ——trilobata, 221 — Forrestii, 216, fig. audatilobum, 217, fig. ee 219, fi Hersii — sikkimense serrulatum, 346 — taronense, 215, — urophyllum, 217 —sp., 220 Aceratorchis tschiliensis, 7 Achillea Ptarmica, 38 Aconitum Anthora, 12 — -— anthoroideum, 12 —gymnandrum, 12 — laeve, 12 — Napellus semigaleatum, 12 —rotundifolium tanguticum, 12 —volubile, 12 new names in bold-face type Aconitum volubile flexuosum, 12 Actaea viaiee erythrocarpa, 10 Actinophloeus linearis, 62 — microcarpus, 62 —— pais on the orchids of the cee ides and Santa Cruz a ie Aden ae zat 36 — liliifolioides, 36 — ape ey 36 — Pota mers 36 —sp., 36 Adenostemma eae 67 oe latedelto —mo amys iain, 2 eins tum, 2 Auhoalee Geena: 350, pl. 74 Agave and Yucca, Taxonomic and cyto- logical iment of, 76, pl. 55 fe) ositum, 66 —lupulina, 30 Alangium vitiense, 57 sare plore 234 — Van Ale Seite terse 63 reticulatus, 63 Alesittiee ates 60 Allium chrysanthum, 5 han = monadelphum thibeticum, 5 — polyrhiz — Pr Bea 5 —tanguticum, 6 428 JOURNAL OF THE ARNOLD ARBORETUM Allium victorialis, 6 lo saving micrococcus, 63 Alocasia lancifolia, 6 Alphitonia moluccana, 65 — zizyphoides, 65 Alpinia novae-pommeraniae, 66 AMES, OAKES, Additional notes on the orchids of the aged Hebrides and Santa Cruz Islands, Anaphalis Hancockii, 38 — lactea, 38 gabe thle 38 —sp., 38 ANDERSON, Epcar, Variation in flower color in Hamamelis vernalis, 253, fig. Andrachne attenuata, 229 29 Androsace Mariae tibetica, 27 —tapete, 27 wzasonsenis 28 Aneilem ts 6S Anemone as 12 — rupestr Anisodus 30 ane microphyllum, 232 — Segu 232 Appen ndicula cleistogama, 66 Aquilegia ecalcarata, 10 — oxysepala Arabis alaschanica, 17 [VOL. XIV Arabis glandulosa, 17 Aralia Labordei, 22 Areca (Balanocarpus) nannospadix, 62 Arenaria kansuensis, 9 —melanandra, 9 — Przewalskii, 9 —Sp., 9 Arisaema c nar snagged a 5 Arnold Arboretuum, ration of the J esesean ats er ner by J. F. Rock the, Arnold erooniedy wn bo Geranium eriostemon, 22 — Pylzowianum, 22 —=sp., 22 Glaux maritima, 28 INDEX 433 Glochidion Bodinieri, 231 — Esquiroli, 231 — villicaule, 231 —sp., 60 Glomera Macdonajdii, 111 Glossorrhynch a Ma: cdonaldii, 111 es mi 26 S. F. Kajewski in 1928 nd 1929; Supplement, 53 Guioa aryterifolia, ¢4 — rigidiuscula, 63 & —sp., Gymnogramma Deluwvayi, 2 Gymnopteris Delava yi, 2 Gymn piney oe 251 — Esqu Fiancee sail 9 Gyrocarpus americajius, 57 Habenaria conopsea 8 — cucullata, — physoplectra, 101 — ponerostachys, 102 Piamaies 8 — sten , 102 Hamamelis eetialki: Variation in flower —camptoneura, 64 ee capri 64 Hatcu, A. and Doak, K. D., My- corrhizal a other atu of the root systems of Finus, 85, pl. 57-60, gs. Hatcu, A. B. and Elatcu, C. Tarsort, Some Hymenomy:etes forming my- corrhizae with Pinus Strobus L., 324, pl. 68- Hedysarum esculentt\m, 21 — obscurum, 21 — pseudastragalus, 21 2 —sp., 22 Helleborus thibetanus, 10 434 JOURNAL OF THE ARNOLD ARBORETUM Hemerocallis Dumortieri, 5 Heracleum millefolium, 25 ——longilobum, 2 er reir plants collected in New Guin y L. J. pra List of, 6 sari The nold Arboretum during the fiscal year ended June 30, 1933, 412 Herminium tanguticum, 8 Hololachne sey 23 Humulus lupulus Hydrangea pana, 202 — Schindleri, 2 — umbellata ee Hyme nsnineias forming mycorrhizae with Pinus Strobus L., Some, 324, 1. 68-71 Hypecoum erectum lactiflorum, 14 Hypericum Przewalskii, 23 —sp., 23 Ilex corallina, at 0 — metaba inte myrsinoides, 240 — peduncularis, 55 — peduncle «continental, 240 — purpure epee 240 — icone 239 5 =e hae ee 67 pat Ha , 66 Lee mente 35 — principis, 35 — sinensis, 35 Inheritance in an oak species hybrid, 68 figs. Inula ammophila, 38 Ipomoea denticulata, 67 , 6 Iris dichotoma, 7 —tenuifolia, 7 [VOL. XIV Jacquemontia paniculata, 67 Juncus leuce omelas, 5 — Thomsoni, 5 Eajewsk s iy Contribution to the flora of Ne ew Hebrides, plants collected a in 1928 and 1929, sup- plement, 53 Kleinhovia hospita, 54 Kobresia kansuensis, 4 — Prattii, 4 KosBuskKI, CLARENCE E. and REHDER, ALFRED, An enumeration of the her- baceous ro’ collected Rock for the old Ameria 4 — List of seeds ot en eous and herba- ceous plants collected in 1925 and 1926 b k and distributed b Arnold Arboretum, 43 Koeleria cristata, 4 Lactuca Souliei, 43 —sp., 43 Lagotis brachystachya, 32 — brevituba, 3 — glauca, 32 Lappula Redowskii, 30 Lathyrus pratensis, 22 Leontice robusta, 14 Leontopodium calocephalum, 37 iin ida X leontopodioides, 38, 426 — Dedekensii, 38 oe 38 — Jacot BB fiero < conglobatum, 38, 426 — linearifolium, 38 —na , 38 Serra 38 Lepidopetalum es mM 64 eucosyke corymbulos Léveillé, Notes on ae ae plants from eastern Asia, described by, 223 Library, The Arnold Arboretum during the fiscal year ended June 30, 1933, 41 mas yHes altaica, 40 — plantaginifolia, 40 — Przewalskii, 40 1933] Ligularia sagitta, 40 — sibirica speciosa, 40 pared — yesoensis Pe uahde. 40 = 8);, Ligusticum Pilgerianum, 25 — Pilgerianum, 25 —sinense, 25 — Weberbauerianum, 25 Ligustrum neo-ebudicum, 58 —tenuifolium, 6 Limnophila rugosa, 67 Linociera ramiflora, 59 Linum nutans, 22 Liparis elegans, 105 List of seeds of ligneous and _ herba- ceous plants collected in 1925 and 1926 by J. F. Rock and distributed by the Arnold Arboretum, 43 Lithospermum officinale, Litsea, sp., 60 Lloydia tibetica lutescens, 6 6 — — purpurascens, Lobaria pulmonaria hypomela, 1 Macleaya microcarpa, 14, 426 Maesa banksiana, 58 — myrsinoides, 240 sp., 235 Magnoliales, Chromosome number and relationship in the, 376, pl. 80, figs. Malaxis lunata, 104 colmi 3 — philippinensis, 233 Malus baccata himalaica, 207 — hupehensis, —w— rosea, 208 — Rockii, 206 — theifera, 207 —— rosea, 208 INDEX 435 Mapania macrocepliala, 65 o and Sax Kart, Taxonomic and cytological re- lationships of Yucca and Agave, 76, Meconopes integrifolia, 14 — psilono a, 14 — punicea, 14, 426 — quintuplinervia, 14 — racemosa, 15 —-—forma, 426 —sp., 15, 426 Medicago ruthenica, 21 re Delavayi, 16 iflora, 10 Meera fe apetalum, 9 —_ ee 30 — tangut Mimulus iid, 32 Morina betonicoides, 36 — Mairei, 234 Mucuna pruriens, €3 Musa sp., 66 Mycorrhizae with Pinus Strobus L., Some Hymenomycetes forming, 324, pl. 68-71 Mycorrhizal and other features of the root systems of Pinus, 85, pl. 57-60, AG Myce ng Var., 30 rica Darr 232 ie Ante !9 —sp., 63 Myrica Darrisii, 232 — Feddei, 240 Nardostachys Jataraansi, 36 Nepeta coerulescens major, 30 —macrantha, 31 436 Nertera — papuana, 67 Nervilia sp., 102 New Hebrides and Santa Cruz Islands, Additional notes on the Orchids of the, 101 New Hebrides, Contribution to the flora and 1929; Sissls New species, varieties and combinations the herbarium and the collec- tions “0 the Arnold Arboretum, 199, figs., 3 4 Notes on oo distribution of White and Banksia npegee in north- org and _be- -67 _ species hybrid, os in an, , figs gs. Paahe sis 66 — glandu , 105 —sp., oo Orchids of the New Hebrides and Santa Cruz Islands, Additional notes on the, 101 Orchis chusua, 7 — spathulata, 7 Oreorchis Rockii, 8 Origin and behavior of the nucleolus in rixa japonica, 224 Ormocarpum' cochinchinense, 57 — sennoides, 5 Orthosiphon stamineus, 67 Oxygraphis glacialis, 13 Oxytropis sp., 21 Pa cabuonads ‘ete Kouytchensis, 235 35 Paeonia gina 10 — Veitchii, 10 JOURNAL OF THE ARNOLD ARBORETUM [VoL, XIV Papua (British New Guinea) in 1925- 26 by L. J. Brass, A supplement to . White, Ligneous oo col- incindl in the territory of Paraquilegia anemonoides, _ Paratrophis ra 61 Paris polyp Parnassia Deny 19 tchuenensis, 19 Parra lo : albiflor Pathological as The Arnold fiscal year Pedicularis alaschanica, 32 —anas var., —armata, 32 -—— ann 32 — isochila, Reena ts 33 —— forma, 33 — chinensis, 33 —cranolopha, 33 — — longicornuta, 33 — decorissima, 33 —i s, 33 — kansuensis, 33 —labellata, 33 —lasiophrys, 33 ——sinica, 33 — longiflora, 33 —macrosiphon, 33 —muscicola, 34 —pilostachya, 34 m a, 34, 426 —striata poliocalyx, 34 — szetschuanica, 34 — — longispica, 34 —torta ees macrantha, 34 — versicolor, 34 —sp., 35, 426 1933] Peganum Harmala, 22 aphthosa, 1 Phajus patra 111 Phlomis rotata, 3 —unbros, 31 mites munis, 4 ile Sears 111 Phrynium macrocephalum, 66 — pedunculatum, 66 Phyllanthodendron Cavaleriei, 230 — Dunnianum, 230 —— hypoglaucum, 230 —sp.,2 Phyllanthus Argyz, 230 — Dunnianus, — Franchetiana, 230 — irei, 230 —sp., 231 Pinguicula ae a, inus, Myc fad afd other features of the nee systems of, 85, pl. 57-60, gs. Pinus pene Notes on the genus . The black cone of, 258 Pinus, cee is me Hymenomy- cetes ie mycorrhizae with, 324, pl. 68-71 Pirus communis, 207 — hupehensis seve Pistacia cinenst sis, Pithecolobium ee 63 —sp., 63 Pittosporum naruaiao, 5 Platanus and Campsis, Species hybrids, , figs , 274 pie Niovesanil 103 pac ciass ne 24 — cnidiifoliu ie rie — Dielsianum, 24 Franchetianum, 24 — linearilobu — Pilgerianum, 24 — pseu udo- invohuratum, 24 — Roc — eee: 24 INDEX 437 Pleurospermum sp., 24 Poa arctica, — attenuata, 4 — bulbosa, 4 — flexuosa, 4 — sphondylodes, 4 —sp., 4 oe emodi, 1 Polemonium eee ie vulgare, 29 Pollia macrophylla, \) Polygala sibirica, 22 aie eri bulbosim, 6 — sibiricum, 6 en. . Polygonum Hookeri, 8 _gphineroeencliyann’ ‘var., 9 Polypodium clathratim, 2 —eilophyllum, 2 — lineare, 3 a ba eerie Braunii, 2 ae ee 2 Pometia pinnata, 56 Potentilla yen 0 20 —bifurca, 20 — multicaulis, 20 — Potaninii, 20 — Salesoviana, 20 — Saundersiana, 20 — sericea, 20 —_ 20 Precipitin reaction in plants, seri on the, III. biochemical analysis of the normal precigitin eens ie figs. Precipitin reaction in plants, Studies on the, of Applicatior to plant relation- ships, 394 Prenanthes sp., Primula Sire 15 — alsophila, 25 —chionantha, 25 f a, 26 —gemmifera, 26 —graminifolia, 26 438 JOURNAL OF THE ARNOLD ARBORETUM Primula limbata, 26 — longipetiolata, 26 — moupinensis, 26 —optata, 26 —forma, 26 — polyneura, 26 — pumilio, 2 poderarall, 27 —sp runus angustifolia varians x P. gracilis, 20; —gracilis >< P. angustifolia varians, 208 — Slavinii, Psychotria polyneura, 65 —sp., 58 Pteretis Struthiopteris, 1 Pueraria novoguineensis, 63 Pulsatilla ambigua, 13 p. Pycnarrhena s Quercus Comptonae, 68 —lyrata « vieinlone. 68 — virginiana lyrata, 68 Ranunculus acris, 13 —_— flabellatus, 13 nicus, 13 — pulchellus, 13 — —sericeus, 13 8 13 Raup, Hucu , Notes on the distri- bution of ates Spruce and Banksian Pine in northwestern Canada, 335, pl. 72, 73 REHDER, ALFRED, A_ supplement to [voL, XIV C. T. White, icin plants col- lected in the territory of Papua (British New Prarie in 1925-26 by L. J. Brass, 62 —, New — Riga and combina- bie from herbarium and the collections 7 < Arnold Arboretum, , pl. 74 the ligneous plants de- éveillé from eastern —, Notes on sciited by ia, —,and KosuskI, CLARENCE E., An enu- meration of the herbaceous plants collected by J. F. Rock for the Ar- ——, List of seeds of ligneous and her- pore a cigucein in 1925 and Rock and distributed oe aia pie 347 — crenatus, 348 —- ae 348 —utilis glabra, 349 ——hypochrysa, 348 Rhaphidophora novo-guineensis, 65, 426 Rheum acuminatum, 8 —palmatum, 8 — pumilum, 8 — spiciforme, 8 — punjabensis . 238 == eo. k, J. F., for a Arnold Arboretum, sere enumeration a the herbaceous plants collected —, List of seeds of a ume and _ herba- ceous plants collected by, and dis- tributed by the Arnold Arboretum, 43 Sabia Cavaleriei, 225 eddei, 224 Salacia aneityensis, 55 1933] Salvia Prattii, 31 — Przewalskii, on — Roborowskii, Sanguisorba, oe 20 — officinalis, 2 Santa Cruz Islands and New Hebrides, Additional notes on the Orchids of the, 1 Sapium rotundifolium, 234 Sarcococca Hookeriana humilis, 235 aurauja Vanioti, 249 Saussurea amara, 41 — nigrescens, 42 — phaeantha, 42 mee 42 — aie aia Gia 42 —sp., 4 Sax, Bae Joriverte and Sax, Kart, Chromosome number and morphol- ogy in the Conifers, 356, pl. 75-79, $ , Kart, Chromosome behavior in Calyeanthus, 279, figs. — Chrom numbers in Ulmus and rate genera, 82, pl. in Platanus and gs. ELVEY, SUSAN Taxonomic and Slatin mre er and morphology in ae Boca. 356, pl. 75-79, figs. Saxifraga confertifolia, 18 — diversifolia Soulieana, 18 —— var = Hagrans platyphylla, 18 — Giraldiana, 19 — kansuensis, 19 — lumpuensis, 19 Riser aaner ies Franchetiana, 19 INDEX 439 Saxifraga melanocentra pluriflora, 19 ——va rbtaiie splendens, 19 _— Pees a culat Pt canie iors calypteat, 65 Securinega fluggeoi'es, 230 —ramiflora, 230 — suffruticosa, 229 — Henryi me 8 — Kirilow = Progress, 18 — Purdomii, 18 aaa 18 — venustum, 18 —sp., 18 Selaginella sp., 3 Senecio acerifolius, 39 — plantaginifolia, 4) — Potanini, 39 — itn 40 — tang » 40 —_— tial 40 — virgaurea, 40 —sp., 40 Serratula centauroides, 39 SHAW, RGE RusELL, Notes on the genus Pinus. The black cone of Pinus ponderosa, 258 Silene Fortunei, 9 —t uls, 9 Siphonodon sp., 63 440 JOURNAL OF THE ARNOLD ARBORETUM Solanum septemlobum, 32 —sp., 6 Some Hymenomycetes forming mycor- rhizae with Pinus Strobus L., 324, pl. 68-71 Sophora alopecuroides, 20 Souliea vaginata, 1 Species hybrids in Platanus and Camp- sis, 274, figs. Speranskia cantonensis, 232 — tonkinensis, 232 — yunnanensis, 203 — — siccanea, 20 —w—tortuosa, 205 aan axillaris, 237 mei ties 31 — Sieboldii, 4 426 Chromosome number in 1 Stelle Strobilanthes novomegapolitanus, 67 —sp., 6 Studies on the precipitin reaction in plants. III. A biochemical analysis of e normal precipitin reaction, 118, figs. Studies on the precipitin reaction in plants. V. Application to plant rela- to C. T. White, Ligneous plants collected in the territory of Papua (British New Guinea) in 1925-26 by L. J. Brass, A, 62 Swertia sp., os Symplocos tetramera, 345 Tacca pinnatifida, 66 [VOL. XIV Tanacetum tenuifolium, 39 Tapeinochilus sp., Taraxacum mongolicum, 43 Taxonomic and cytological relation- ships of Yucca and Agave, 76, pl. 55 Thalictrum alpinum, — baicalense, 13 — javanicum, 13 — Przewalskii, 14 Thri sansipohiat ae © s Serpyllum eo 31 Tiarella polyphylla, rios fid 35 Tripterygium hypoglaucum, 252 Trisetum spicatum, 3 sp., 4 Trollius nagar 10 pinus, 1 Ulmus ae related genera, Chromo- Valeriana pseudodioica, 36 — tangutica, 36 —sp., 36 Variation in flower color in Hamame- lis vernalis, 253, a Vernonia cinerea, — lanceolata, Gendee ciliata, 32 VestaL, P. A., CHESTER, K. S. and ABBE, E. C., Studies on the precipitin reaction in p . Application to plant sclationshina, 3 04 Vesture it dicotyledons, The cambium ivative tissues. ts der No. VIII. Sania distribution, diagnostic significance of, 259, pl. 1933] 63, figs. Vicia amoena elliptica, 22 — cracca, —unijuga, 22 —— Var., 22 Viola biflora, 23 INDEX 441 —,The chromosome Co Mbieniea of Cysliikandrs betacea, 113, figs —,and CuesTer, KENNETH S., Studies on the Ne gem reaction in II. A plants III. chemical analysis of the cc eae reaction,” 118, White C. T., Ligneous ae collected Brass, A supplement to, 62 (British y i. J sae gant and Barksian Pine in Canada, pire on the distribution of, 335, pl. 1 ospora, 1 Ximenia americana, 55 Xylanche himalaica, 35 Xylosma lifuana, 53 a NELL, S. H., per i in an oak species ‘hybrid, 68, Yucca and Agave, hens Ue and cyto- logical Se Al 76, pl. 55 Zanthoxy lum — giganteum, 22 , 103 Zygophyllum erential: 22