THE BOTANICAL GAZETTE THE UNIVERSITY OF CHICAGO PRESS CHICAGO, ILLINOIS THE CAMBRIDGE UNIVERSITY PRESS LONDON AND EDINBURGH THE MARUZEN-KABUSHIKI-KAISHA TOKYO, » KYOTO, FUKUOKA, SENDAI THE MISSION BOOK COMPANY SHANGHAI ek THE SUES BOTANICAL GAZETTE *- iil JOHN MERLE COULTER VOLUME LXVI JULY-DECEMBER 10918 WITH TWENTY-FOUR PLATES AND ONE HUNDRED NINETEEN FIGURES THE UNIVERSITY OF CHICAGO PRESS : 0) . CHICAGO, ILLINOIS ae Published July, August, September, October, November, December, 1918 Composed and The University of Chicago Press Chicago, Illinois, U.S.A. TABLE OF CONTENTS Experimental investigations on the* = mses skya (with nineteen figures) - James R. Weir Chemical changes accompanying abscission in Coleus Blumei. cc from the Hull Botanical Laboratory 240 - - - - -Homer C. Sampson ae of the Taxineae (with plates I, IT) Mary C. Bliss — Significance of resinous tracheids (with five — Samuel J. Record Determination of wilting. Contributions from t Hull Botanical Laboratory 241 (with five fag Arthur L. Bakke Notes on American willows. I. The species related to Salix arctica Pall. - - - - Camillo Schneider Fecundation and formation of the primary endospe nucleus in certain Liliaceae (with plates III-V) Mildred N. othnagel Factors determining character and distribution of food reserve in woody plants (with two figures)Edmund W. Sinnott Suspensor and early embryo of Pinus. Contributions from the Hull Botanical ssecenpsaee See 242 fri plates VI-X and three figures) - John T. Buchholz Notes on North American trees. II. Carya - - C. S. Sargent Fertilization in Lilium. Contributions from the Hull Botanical ny 243 es a MI-XTM) = - - - Wanda Weniger Abnormal —— in preenre ve three ; fi gures J. G. Brown A contribution to the life iy of I sesaibens Sultani : (with plates XIV, XV) - - - - - - Alice M. Ottley Notes on American willows. II. The species related to Salix glauca L.- - - = - = = = Camillo Schneider The sporangia of Thismia americana (with plate XVI) Norma E. Pfeiffer Root variations induced by carbon dioxide 0s addi- _ tions to soil (with nine figures) - H.A. Noyes, J. F. Trost, and L. Yoder Vi CONTENTS Absorption of sodium and calcium by wheat urna For titles of book i ind Z thor’s name and reviews Papers noticed in “Notes for Students” are in- _ dexed under author’s name and subjects [VOLUME LXVI PAGE (with one figure) - - - Howard S. Reed 374 Morphology of Rumex crispus. Contributions from the Hull Botanical Laboratory 244 Sela Jeet XVII-XIX and twenty-one figures) Winfield Dudgeon 393 Notes on North American trees. III. Tilia. I. - C. S. Sargent 421 ‘Pine needles, their significance and — hes twenty-nine figures) - - - Jean Dufrenoy 439 Limiting factors in relation to specific ranges of toler- ance of forest trees (with seven figures) - - A.H. Hutchinson 465 Notes on North American trees. III. Tilia. II - C. S. Sargent 494 The purple hyacinth bean (with seven figures) - - George F. Freeman 512 A morphological study of Pallavicinia Lyellii. Con- tributions from the Hull Botanical Laboratory 245 (with plates XX-XXIV) - - - - - ArthurW. Haupt 524 BRIEFER ARTICLES— Meaptee en cones of Pinus montana Astaro one figure) - W.N. Steil 68 Healthy and sick specimens of Bryophyllum calycinum Jacques Loeb 69 Modified safety-razor blade holder for Se aane: control (with one figure) T. H. Goodspeed 176 Pollination of Asclepias cryptoceras- - - - ~- Charles Robertson 177 Cross-conjugation in Spirogyra Weberit (with one figure) - - - - - = - = = = Bert Cunningham 272 An endemic begonia of Hawaii =, = - i= = V. MacCaughey 273 Secondary parasitism in Phoradendron - - - J. Arthur Harris 275 Method of replacing paraffin solvent with paraffin - T.H. Goodspeed 381 Adaptation and natural selection - - - E.F. Andrews 382 Joseph Young Bergen (with portrait) - - - - RodneyH.True 455 Modification of hand microtome (with five figures) - 7. H.Goodspeed 534 CurRENT LITERATURE - - - - - = 70,178,277,383, 459,537 VOLUME LXvI] CONTENTS vii DATES OF PUBLICATION No. 1, July 15; No. 2, August 15; No. 3, September 16; No. 4, October 16; No. 5, November 15; No. 6, December 18. ERRATA VoL. LXVI P. 53, line 4 from top, for UNIVERsITY oF Onto read OnIO STATE UNIVERSITY P. 57, line 12 from top, for fig. 4 read fig. 7 P. 65, line 4 from top, omit (RPI) P. 109, line 12 from top, for average of 3 leaves read average indices of 3 leaves P. 132, line 9 from top, for 61 read 60 P. 217, line 16 from top, for YAzIN read YASUI P. 224, line 18 from top, for YAzIn read YaSsuI P. 241, line 24 from top, for Carya pallida read Carya pallida Ashe P. 246, line 9 from bottom, for southwestern read southeastern P. 338, line 2 from top, for MoLTE read MALTE P. 438, line 4 from bottom, for WoopHOUSE read WODEHOUSE VOLUME LXVI NUMBER 1 LEE BOTANICAL GAZETTE JULY 1978 EXPERIMENTAL INVESTIGATIONS ON THE GENUS RAZOUMOFSKYA JAMES R. WEIR (WITH NINETEEN FIGURES) Introduction This article is the first of a series of reports on culture experi- ments of mistletoes. The work was begun in September 1911, and will be continued indefinitely. The aim of these experiments is to determine the validity of the several species as now distinguished, their affinities to each other, hosts on which they may be of eco- nomic importance or on which they may occasionally occur, and influence of host and condition of host as governed by its environ- ment on the form, color, or other diagnostic characters commonly employed in the classification of these parasites. Since the system- atic position and host relationships of several of these plants are not definitely defined, and since they are of great economic importance in many forest regions, it is believed the work will be of consider- able value. The plan of these reports is to record as briefly as pos- sible the results of each series of cultures as completed. The present report includes considerable discussion, owing to the necessity of outlining the problems in hand. The detailed discussion of results and technical description of species will be reserved until the con- clusion of the experiments. 2 BOTANICAL GAZETTE [JULY Methods For that part of the work being conducted at Missoula, Mon- tana, the opportunities are very favorable. Practically all the species of Razoumofskya of any economic importance are of easy access from the laboratory. Members of the field force of the United States Forest Service are aiding in the work by sending in fresh mature specimens of R. pusilla on spruce and larch from the Lake states, and of the rare unclassified forms occurring on white and yellow pines in Oregon, Idaho, Utah, and Nevada. A great deal of material of the common forms from all parts of the North- west has also been contributed. The writer visits regularly the various forests of the Northwest and has made abundant collections of the mistletoes of these regions. The writer is under particular obligations to Professor W. C. WErR for service in connection with cultures at Bellingham, Washington; to L. H. Werr for collecting special material; to D. R. BREWsTER of the Forest Service Experi- ment Station, at Priest River, Idaho, and to J. Duncan, Super- intendent of Parks of the city of Spokane, for permitting cultures to be made on various exotic conifers; and to E. E. Huserr of this laboratory for assistance in making cultures. From 1911 to 1914 inclusive the inoculations were conducted in the open. Seeds were sown on trial hosts of species other than that on which they developed, either in the same vicinity or in widely separate regions. In the latter case trial hosts of the same species as that on which the mistletoe grew were also included. This served to check the viability of the seed, also to bring out differences due to change of environment between the plants resulting from inoculation on the same host species and the plants furnishing the seed. The same was true for the plants on trial hosts other than that on which the parent plant developed. This double procedure demanded copious notes on the conditions of growth and general morphology of the plants furnishing the seed used in inoculations in other regions and the saving of specimens of both sexes for comparison afterward. The same was done with plants resulting from inoculation. In the latter case, where neces- sary, the infected branch or stem was cut out to prevent the spread _of the parasite in new regions. A large number of specimens are 1918] WEIR—RAZOUMOFSKYA 3 accumulating, but this seemed desirable in case all necessary notes were not taken on both generations. In the case of continuing the inoculations of the same species of mistletoe through several generations on the same host but different individuals, either in the same or different localities, or on different host species, the saving of specimens fully recorded is doubly necessary. This should also furnish some information on the subject of the germinal transmission of characters. Cultures begun in 1914 are being conducted both in the field and in the greenhouse. This doubles the amount of work, insuring greater dependency on results; and in the case of the indoor work closer study is possible of the life history of a successful inoculation. Indoor work also permits the use of a larger number of trial host species. The seeds germinate more rapidly and results are sooner obtained. One of the chief reasons for maintaining outdoor cul- tures is to check, whenever possible, under natural conditions, any unusual result obtained in the greenhouse. Cultures in the open have so far proved more successful than those inside, where the same mistletoes and hosts were concerned. If, however, a few unusual hosts are obtained indoors, it must be remembered that it is a new association of host and parasite often not possible in nature; moreover, some of the mistletoes showing the greatest predilection for a particular host or host genus are Cece eee found on trees belonging to other genera. In making the inoculations great care is exercised to attach the seeds at the most vulnerable points, such as in the axils of the leaf sheaths, tender branches, base of terminal buds, and in the denser zone of needles at the nodes. Observations show that infection usually occurs at these places.‘ Before the seeds are transferred to the host they are allowed to stand for a few minutes in water. This causes the mucilaginous coat of the seed to expand. The seeds are then sucked against the point of a dropping pipette and placed firmly in the desired position. After a short time the mucilaginous layer dries, holding the point of the seed in place. The host material used in the inoculations ranges from seedlings 2 years old to the tender branches of mature forest trees. In case * WEIR, James R., Wallrothiella Arceuthobii. Jour. Agric. Research 4:377- tots. «® 4 BOTANICAL GAZETTE [JULY of the trial host possessing a suberized cortex the seeds are sown only on the first to the sixth year’s growth. It has already been demonstrated that infection will not normally take place on older tissues.? It has been experimentally proven, however, that by scraping away the dead surface tissues of the bark on parts of branches as old as 7 years, and which still contain chlorophyll, it may be possible to secure infection. The number of seeds sown on each trial host has been maintained at 20 for greenhouse cultures and, owing to possible accident to the seed, so for outdoor work. It seemed desirable to try to maintain the seed at a fixed number so that the relative susceptibility of all trial hosts to any one form of mistletoe may be compared. Because it is impossible, especially of cultures in the open, to know that all seeds sown remained on the trees, the relative susceptibility of all trial hosts was further tested in most cases where it was particularly desirable to do so and whenever it was possible, by using a fixed number of trees of any one genus. A record of the source of seed of all trial hosts and of the place where the trees were grown was kept. This seemed desirable in view of the question of influence on the morphology of the parasite. In the case of transplants the trees had not been trans- planted very long to the place where cultures were made. The seeds demand a period of rest before germination, and if stored under cool and moist conditions may be carried over and sown in the spring. Sowings as late as April have resulted in successful inoculations. The low temperatures of winter also seem beneficial to the seed, as it is observed that a higher percentage of seeds germinate which have undergone freezing temperatures. This probably accounts for the greater number of positive results obtained in outdoor cultures. If the seeds are stored in warm, dry air, they lose their vitality very rapidly, owing to the evapora- tion of moisture from the chlorophyllaceous endosperm. Germina- tion tests show that the seeds are capable of germination some 2 weeks before they are normally expelled from the capsule, so that it has been possible to sow the seeds of some species early in Sep- tember. Care must be taken in sowing seeds on Larix before 2 WEIR, J. R., Mistletoe injury to conifers in the Northwest. U.S. Dept. Agric. Bull. 360. p. 8. 1916. 1918] WEIR—RAZOUMOFSKYA 5 the leaves have fallen; other- wise the seeds placed on the foliar ‘spurs will be carried away with the falling leaves. The cultures of the false mistletoes may be considered difficult. There must first be considerable knowledge of the requirements for seed ger- mination, and of the plants afterward, in the case of the work done indoors. Much that is necessary has been learned, and the work is now going on more rapidly. The following is the first detailed report of the culture of mistletoes in this country. Some work of this kind Fic. nbs ponderosa: ith stems more or Wns Ctindeeal at base, stasitiates Oregon coast; reduced one-fourth. LE oa ga sainbelonede. on branching form but in another connection has already been reported by the Fic. 2.—R. campy de on Pinus ponderosa: short, thick form with angular stems, staminate and pistillate plants; n coast. writer (Joc. cit.). Cultures with yellow pine mistletoes Razoumofskya campylopoda (Engelm.) Piper and R. cryp- topoda (Engelm.) Coville, the largest and most conspicuous members of the genus in the United States, are supposedly 2 distinct species occurring on yellow pines. The former (figs. 1, 2, 6) is based on specimens from north Idaho or northeastern Washington, and is principally confined to the coast and northern Rocky 6 BOTANICAL GAZETTE [yuLy Mountain regions. The latter (figs. 3-5), based on specimens from New Mexico, is apparently limited to the southern Rocky Mountain regions. Both plants were originally described from specimens on Pinus ponderosa, which is their most common host. A large collection of these plants on P. ponderosa and a number of other hosts from their respective regions shows so few constant dis- tinguishing characters by Te which the plants from the Fic. 3.—R. cryptopoda on Pinus ponderosa: two geographical regions may pistillate; New Mexico; reduced one-fourth. readily be separated that it —Photograph by G. G. Hepccock. : seemed desirable to test them out by cultures. Color, branching, thickness of stems, parting of flowers, and position of anthers on the calyx lobes, characters (29% . usually employed to distinguish =, A one species from the other, are : not always constant in these plants from the several regions in which they are supposed to occur, but apparently merge into one form or the other with change of habitat, just as is the case in any other species having a wide distribution and range of hosts. In a series of experi- ments recently completed by the writer and not otherwise Fic. 4.—R. cryptopoda on Pinus ponderosa: staminate and pistillate; mentioned in this paper it has southern Utah. 3 Piper, CHARLES V., Contrib. U.S. Nat. Herb. 11:222. 1906. 1918] WEIR—RAZOUMOFSKYA 7 been demonstrated that size and color of flowers, stem, fruit, form and division of calyx lobes, slenderness and length of plant, com- pactness of individual colo- nies of the northern form depend upon age of the plants and of the infection, nourishment, condition, loca- tion, and species of host. In view of these results it seems desirable that the diagnostic characters as now employed in the separation of the large _ plants on yellow pines should Fic. 5.—R. cryptopoda on Pinus chihua- be substantiated by a large huana: sift reduced one-half.—Photo- : raph by G. G. Hepccock. number of cultures before they can be held specifically distinct. Experiments involving the transfer of seeds of the northern and coast plant from its various hosts to Rocky Mountain yellow pines, and vice versa, in their respective regions should be of some value in determining the validity of the two alleged species. R. occidentalis abietina Engelm. (figs. 7, 8) is a large form of mistletoe found on Abies throughout California, Washington, Oregon, and Idaho. It closely resembles the large mistletoes on yellow pines and is described as a Fic. 6.—R. campylopoda on Pinus varietv of the form R. cam py- ponderosa as it often appears growing from en a an advancing cortical stroma in branches of lopoda (figs. a 2, 6) (Arcew- witches’ brooms: plants pistillate, mature. ‘obium occidentale). The plant is not so large as the latter, but both have the same color variations and bloom and fruit in the same period. The facts that it is usually found in 8 BOTANICAL GAZETTE [JULY the same regions where the yellow pine mistletoe occurs, has the same diseases attacking it, and is not found in regions where the typical R. tsugensis is most abundant and which it also slightly resembles, indicate that it may be a biological form of the former. The results of a number of cultures involving the three plants mentioned are pre- sented in table I. It will be seen from table I that an effort has been made to sow the seed of the large mistletoes on Pinus ponderosa (figs. 1-6) from several localities on as many —R. occidentalis abietina on Abies different hosts as possible Canesten: staminate and pistillate plants; 0 Oregon. and on the same host in widely separate regions. The object of the latter was to try to determine the ae of the common mistletoes with thick, robust stems on yellow pine in the Rocky Mountain region to the more slender form on the same host in the Pacific Coast region. This problem has been sufficiently outlined previously. The cultures so far do not furnish any evi- dence that the two forms should be considered identi- cal. Plants in the Pacific Coast region resulting from seed collected in the northern Rocky Mountain region, and vice versa; exhibit various color varia- Fic. 8.—R. occidentalis abietina on Abies nobilis: staminate plants; Oregon. 1918] WEIR—RAZOUMOFSKYA 9 tions, depending upon the region where grown. The same varia- tion is noted in the robust form when grown in the North. This shows that these plants from the different localities cannot be held specifically distinct on a basis of color. Although color has been one of the chief distinctions between the two, the cultures show that there is no marked difference in the general morphology of each form when grown outside of its original place of collection. True, there are some differences to be noted with respect to size, but it is purely a matter of age of infection. Even after the first maturity these plants, which have a comparatively long life, grow larger by developing additional branches and increasing the thickness of the stem. The comparisons made. in the table are based on plants differing widely in age; consequently measurements must vary slightly. Excepting color changes, which were to be expected from varying habitats, the general morphology of the younger plants of the parent colonies were in no particular different from those of the cultures. The cultures have also demonstrated the fact that R. campylopoda will infect Abies, with considerable variation in color and size of the resultant plants, but closely resembling the form known as R. occidentalis abietina. It is inter- esting to note in this connection that CoviLtE‘ refers the plant found on Abies magnifica and A. concolor directly to R. campylopoda (R. occidentalis [Engelm.] Coville) with the statement that it is probably the plant that ENGELMANN’ had previously described under this name (Arceuthobium campylopodum). Itis further shown ' that R. campylopoda will infect Picea and Larix, but with difficulty. This mistletoe also will apparently readily infect Pinus contorta, a result repeatedly confirmed in the field. This tree, however, is not a common host. As will be shown in the case of R. americana, it is believed that this parasite may be expected to occur on any hard or yellow pine, but with predilection for certain species. The mere assumption that hosts are the determining factors of a species is here shown to be untenable. When a parasitic species will infect hosts from widely separate regions and even genera, and the result- ing plants have certain characters varying from those exhibited 4 Contrib. U.S. Nat. Herb. 4:192. 1893. 5 Gray, AsA, PI. Lindh. 2:214. 1850. BOTANICAL GAZETTE [JULY Io "peyed-y pue ¢ ‘SIMO {Suo] “wo $*1 ‘sayids [e1oy « fysty "Uy ‘yaoys ‘sayouviq {moras AATIO ‘oyeuTUTEys ‘uu Shr Aq o£ ‘paas ‘MOTTA aayo may fysry ‘woh “noysnosyy ysnqor ‘aseq 7B pm ‘mO][aA BATIO JYSI] SUII}S f‘ayeynsid ‘yuored f-wu ph Aq cok ena «J ‘t1-v-O1 ig161 ‘Sx Jaqo}Q | £1-o1-$ oyepy ‘IIATYy Ise [e00'T oy (z) vsoropuod snug *(g£9 ‘ou) JATIO YSLMOTAA ‘payouBsq Yonut fysry ‘wo g “uared {(gtg ou) UMOIG SNOIDRUTA sapou “9AqIO Ystmoppad ‘10[09 [eiaued ‘ysry ‘uo £ (q) {(L£9 ‘ou) Buoy ‘wd 1 sayids SuLaMoy ‘amnyeul f‘aaro dosp ‘ysry ‘wo £ (7) ayeuTUIe}s Vv" [Je {ys0y yea uo uoNsayut I :g161 ‘HI Jaqoy}DQ | £1-9z-9 oYyepy] ‘ATY ISI [e00'_T oe oat 2 Oe omen (z) esorapuod snurg uoneuruiiad $ mn (parassed Seen) ee gory peony aivd aNnv Ayeoory jo Sacnee pasn soquier pat aweN SHANLTAND AO SLTASAY aqvW axaM SAAALTAS GAIHM ALITVOO'T SLSOH ‘TVIN], GaYAVN HVLQ NI GaLOATTIOD aaas WONT saAALIND fg GawMAVW ‘OHVAT ‘SaTaVJT “LG LV GaLOaATIOD Gaas Wow SUAALTND fY CAXAVW ‘VINAOATIVD ‘LSTAOY TVNOILVN VAVAAVE VINVS AHL NI GALIATIOD GAAS WOU SAXALTAO ‘ppod oid 49 "yy aNv ppodojkdup? “y SV SNOIDAA AALLOAdSTA ALTHL NI NMONW ATNOWWOO Dsodapuod snuig NO SAOLATLSIN AQMVT AHL WOMT ‘HVLQ) NI GNV ‘NOLONIHSV\ ‘ANVNOS Lv ‘V161 ‘VI YAMWAAON OL F AAAWALdAG WOT STIVAMALNI LV ANV ‘e161 ‘of aaa#WaAlddS ‘VINYOATIV, ‘ISTAOY ‘IVNOLLVN Vavadvg VINVS ‘z161 ‘6c aaMWALaaS ‘OHVG] ‘SaTav “Ig ‘1161 ‘Sz UAANALAAS AALOATION AAAS I ATAVL WEIR—RAZOUMOFSKYA 1918] *pash d19M Spaes YOIYM wor sjuv[d azeI[19SId 07 ySorvaNy § ‘dIN}L[UIWIOU PUL SpIEpUeyS 1O[OI SABA] pry Wosy uexV SoueU JOjOD {Ty t *PpPy ur sainqng » *(P£g ou) poyied-y pue € siamoy ‘Buoy ‘woz sayids pesog ‘ysry ‘wo / ‘payoueig yonut ‘aArfo ystumorg 7Yy ST ‘oyvUTUTe}s “(€£9, ‘ou) ‘wu £$-1 Aq 1°€ poss ‘antq uvissns my ‘Jaqest suas fayeI[MWsId Juared {(1£9 ‘ou) ‘uu zh 1 Aq zo’ € paas ‘ainqzeul ‘aarjo YystMoT[ad ‘ysy ‘wis $°z 19y,R] {(z£9 ‘ou) poyed-£ Apsou “soy yora uo uoroayur 1 :g161 ‘St 1aqowQ “($9 ‘ou) dato YstMoTaA quared '(1%9 ‘ou) aatjo ystmoypad WYysT “ysTy ‘wo z ‘ayeysid fuonsayur 1 :g161 ‘Sr 1aqo}90Q "(£49 ou) paqied-b Asour ‘Suo] “woz sayids peioy ‘aato ystuaeis yrep ‘ysty “wo6 §ayeuruejs {(h49 ‘ou) “wu £1 Aq 0° paas 'Avi3 anjq-usei3 oped ymay J2AqO AumMe} suiays ‘fysry “wu 9 ‘aqeqasid “quale ‘uu Shr Aq co’f paas ‘Avid snosoeula £1-61-9 £1-61-9 €1-o7-$ zI-O1I-v V « zI-¥-01 (uon “81S qusuttadssf) OYeP] “IBANYT Seg ze) -P1S quawedx*y) ‘oyepy “aary yeu “yuo! #21-$z-O1 jUOyy “B[NOSsIPy [e90'T qd +‘lI-Ol-o1 “YseM "yseM “Weysurjeg | pur “[eD a (ue | [eo] T2O ee (z) vsorapuod snurgy ? (z) (umaojpndoss IBA) Bso1apuod snutg ke (z) vsorapuod snurg **(Z) Bsorspuod snurg [JULY BOTANICAL GAZETTE *(1S9 ‘ou) AAO YStuseIs Yep yNIy “aaqyo YStMoTaA FYSIT V # cI-t-o1 yuared £(oSg ‘ou) uMorq Aurre ‘ wale; Jspuyeuror Bon ‘poyor}}2 jurid suo ‘peap ysowye ‘pazunys 31M3 -B}g JuouIdxy) powozut ‘ayeutureys ‘UOT}IeFUT I :g161 ‘Si 1aqozIQ | £1-61-9 oyepy TOATY ysalig oyep] “UUTY, Bcc a (z) Bsoulsal snuig ‘(69 ‘oW) JATTO YSIMOTIAA qyusird ‘(gbhg ‘ou) MoTaA aatjo ‘Yysty “uD £ \vr1—zz-z1 ‘gyeurue}s sjurjd fsuoroeyur z :4161 ‘f Arenuef | S$1-g1-z "qUOP, ‘BPNOsstpy ‘Te Tea [°° °° °° (2) rAaagef snug "(929 ‘ou) : sjurjd ajeurumeys moped daap yi youriq aures uo ‘Mopjak Aauoy yuered ‘paonpoad you ymaiy ‘aato YSIMoyad ‘yBry “ud z (Of9 ‘ou) Jay}0 ay} {(6z9 ‘ou) psonpoid you ymgy ‘MoTpaA posed W938 JO Jopureutal ‘sapou 7¥ IeTIT ystdind ‘ysty ‘wo Se suo ‘azeNstd syurfd ‘saaq #2I-9-O1 yualayiIp z uo suornoajur z :g161 ‘PI IaqoyGQ | zI-gz-9 OYePpT “VAR Sag [P07 foo (z) ¥}10]U09 snutg _poyed-£ A]}SOUW SIaMOY BuO] ‘wo 1 soyids ‘paiamoy MmoyaA Aauoy ‘ysty “wo S$ ‘oy euTUIR}s (gP9 ‘ou) 9AT[O YSsT -MoyJaA yMIy ‘IayDO-aayo “Ysty “wd 4 ‘gy eyns1d quaied ‘(tg ‘ou) payied-t pue £ siamoy ‘Buo] "wd S$*1 sayids yeroy ‘aarfo ystuse3 yzep “Yysty V » Z1-9g-O1 “wo F ‘ay eUTUIR}S —— I :9161 ‘br JIqQOWQ | F1-9z-9 oyepy “IAT Wag [e00'T peoo’T |* "°° ** *(@) By40}U09 snutg 19q UO patp 99.1} »VI-OI-O1 ‘Suuvadde pee! ee I ne ‘1 aqUIBAON | S1-g7—-h “{uOW “B[NOssIyy “quo! "WN | °° °° (2) Bsorapuod snug “(Sq ‘ou) I0[oo Ae ‘ys1y ‘wo g yuored {(PPg ‘ou) aatjo Ayng ‘Yysty V x‘Z1-9-01 ‘wo £ ‘ayeysid ‘uonsayur 1 :g161 “br Jaqo}IQ | £1-gz-9 oYyepy] ‘ATY ISA [e00’] jwoo’T | °°" (z) esosapuod snutg 2I-b-o1 *"(£bg ou) anjq ueIssns yay ‘MoT[aA (uo) aayo jyuared ‘{(zhg ‘ou) aATfO-nIDa ‘Ys -B}$ JUoWLIadx]) ‘woz ‘ayeEysid ‘uonsoyur 1 :g161 ‘Sx JaqoyIQ | £1-61-9 oyepy ‘IAT Weg si @) "yeQ |° °° °° (z) wsorzspuod snutg asso soe umoud poas saanyyno ut (pasvisAv SyUsUIIINSvaW) S}Nsaxy feng oa aava QNV ao Aq20] Jo aamnog posh zaquinu put owey QaaHM ALITVOO'T SH4ALIND AO SLTASAY SLSOH ‘IVI8T, penusod-—T ATEVL WEIR—RAZOUMOFSKYA 13 1918] “IN}EPUSWON PUL SPIBpUR,S JOJOI $ ARMApPIY Wor] UdyR) SaureU JO}OI JTYT —“asnoyuaeg | —*pyay UT SaMyND » *(£¥q ‘ou) quared £(PS9 ou) aatjo ystueas = ysep ‘ys “woe ‘azyeyysid f(€$o ‘ou) 9ATIO YsTuseIs Yep ‘ainjeu ‘ysry “uD + @ ‘II-OI-11 syuvjd ayvurwmeys Ssuoroajut z :g161 ‘Er aqoqaQ |********* "yse mM ‘auvyodg ye9 ‘M0 "***"*(z) JO[OIUOD SaIqy "(£9 ou) quaied {(zSg ‘ou) Moyad Aauoy ‘yAry “uD £ «Z1-9-O1 syurd ayeysid ‘uonsayur 1 :9161 ‘hr Jaqoq}9Q | £1-9z~-9 oyrpy “IIATY Seg [B00'T [woo"T [°° °*(z) stpuesd saiqy ‘Ja}UIVIP Ul “WUT “urpnxyj0id Apereq ‘yews sjuvjd ‘youeiq jo Surjams uwosisny {v1—1z-o1 pessunouoid =fuonoayur 1 :g161 ‘1 sJaquiaAoN | £1-9z-z “quoyy ‘e[Nosstpy oyrpy adoingy |******** (1) vaspaoxe vaotg ynsar {vi—pi-o1 JOYyWNJ OU f1ayYxUIs Jo UoNvIaUag :g161 ‘x Avy | S1-g1-z “qUOP, “BNOsstpy oyepy adoing |********(b) vadoime xuey *yNser jvi—zz—z1 JIyWNy OU f1ayxUIs JO UOeIJOUAg :g161 ‘1 AvP | S1-g1-z *qUOP, ‘B[NosstTpy ‘uoWy |. «= “uOPY °°’ **(F) SITeJUapIO90 xe] ‘(Pg9 ou) MOTAA ured ‘ayeUTUTe]s yUIIed ‘paraMog prey ‘MorjeA ystuses ‘ysty “wo ¢ 21-9-O1 ‘ayeurureys syuejd ‘uonsejul t :g161 “PI 1aqo}9Q | F1-gz-9 oYyPpy ‘Ary seg oyepy oyepy [°° **(F) sqeyuapioso xueyT *(4z9 *OU) IATIO YsIMoT[ad sayids [voy ‘aArTO Yst -u90I3 yiep ‘youRIq des UO SUI9}S 9} eUTUIR}S ‘qeip ys ymay ‘MoyaA Aouoy ‘ysiy “wo 9g ‘ayeqnsid yuared {(gzg ‘ou) dAT[O YstmopjoA WS] ‘snosos1a “ysty ‘woS*f rayjO {(gzg ‘ou) MOTPPA aaTjo ‘“ysty “wu ‘poyunjs 1 ‘a}R] ylI-Ol-I1 : -isid yoq “soy yova uo I :g161 ‘hI AaquIadag |******” ; *yseM ‘aueyodg | sueyodg adoingq |*******(z) euvyuour snug "(zg ‘ou) ysry “wo + ‘aato ystmoz[aA ATurJO;TUN quaied ‘(¥zQg ‘ou) paonpoid you ymay ‘Avid ysyjdind sapousayur ‘aayo ystmorjeA Apu0y -lun jou ‘ysty ‘wo § ‘azerpystd “ueyd uonsayut pay *(€cg ou) aato Yysrmoyed qysT “ysiy wo Sz JajjR_ £(€z9 ‘ou) ainyeur ‘morpad xPM sayids [e1oy ‘qeip snoadeuta surays ‘ySty ‘uD + : Touoy fayeyposid 1 ‘ayeururejs 1 “qaede -wo z 41I-OI-I1 ulg}s 2UO UO z ‘suOT}DaJUT £ :Qg161 “PI JaquIadag |******** "ysem ‘auryods | oauvyods adoing | ******(z) suysaayAs snurg 14 BOTANICAL GAZETTE [JULY _ by the parent when growing on what we may term the mother host, the limitations of a species are naturally more difficult to define. Notwithstanding this change, however, a good species should be sufficiently characteristic on any host and under any of the ordinary conditions of growth as to be readily recognized by one having a wide knowledge of the plant in the field. We do not think of the low, scrubby Douglas fir of central Montana as anything different from the gigantic form of this tree occurring in the Puget Sound region. It is of considerable economic importance that R. campylopoda will infect Pinus resinosa, P. sylvestris, and P. montana, and may be expected to be a serious pest on these trees in localities where conditions are favorable. In drier sites of the Lake states and, in fact, throughout the Northeast, where it is proposed to plant P. resinosa, this mistletoe would undoubtedly grow luxuriantly, and care should be exercised against its introduction into these regions on nursery stock during the early period of infection. Seeds of R. campylopoda were sown on the following pines in most cases in the greenhouse, but either due to the poor quality of the seed, loss of seed, or low vigor of the trial hosts the results were mostly: negative. This does not mean, however, that all of the species mentioned here are immune. In a few cases infection did occur on species not mentioned in the table, but the results were of a nature that it is thought best not to report them at this time. These were Pinus Banksiana, P. mayriana, P. Strobus, P. Cembra, P. cembroides, P. edulis, P. Lambertiana, and P. monticola. Sowings made on Pseudotsuga taxifolia, Larix leptolepis, Tsuga heterophylla, Thuja plicata, T. occidentalis, Cupressus arizonica, Picea Engelmannt, P. canadensis, Populus tremuloides, P. trichocarpa, Betula occt- dentalis, Alnus tenuifolia, Acer glabrum, and Prunus demissa resulted negatively. SuMMARY.—Results of cultures so far indicate that the mistle- toes known under the names Razoumofskya campylopoda and R. cryptopoda are distinct. Each form, however, may exhibit con- siderable variation, due to geographic location and host. The relationship of the two forms will be further considered when a number of experiments now being conducted are completed. rg18] WEIR—RAZOUMOFSKYA 15 Cultures show that the plant known as R. occidentalis abietina on Abies is in all probability a biological form of R. campylopoda. The taxonomic position of the plant, however, cannot be established with any certainty until it is success- fully grown on yellow pine. Cultures with larch mistletoe From the fact that this parasite, R. laricis Piper (figs. 9, 10), exhibits con- siderable variation under different conditions of growth and will occasionally grow on other hosts than Larix, it seemed desirable to study the species in culture. The chief results of these experiments eceuee j i |! | meTRic.! system ©! Mi} beet gg Fic. 9.—R. laricis on Larix occidentalis: staminate and pistillate plants. are embodied in table II. These results indicate that Larix is the true host genus for R. Jaricis. Fic. 10.—R. laricis on Larix occidentalis: pistillate plants; uaa one-half, The fact that 6 trees of Larix occidentalis were in- fected out of 6 on which seed were sown demonstrated the close affinity of the host and parasite. The readi- ness with which R. laricis infects Larix europea and L. lepto- lepis, the common Japanese larch, shows that this parasite may be expected to cause serious injury to plan- tations of these species not only in America, but in many parts of Europe and Japan as well, wherever climatic conditions are [JULY BOTANICAL GAZETTE © - “($89 ‘ou) poyred-b Ajjsou ‘siamoy MOTIA Valo “yYouRIq sues uO suIa}s a}yeUTWIe}S *(6S9 ‘OU) BATTO YsTMOTIAA VysTT ynIy ‘UMOIq PU Yep surazs ozeyNsId ‘yuored “(489 pue 99 ‘sou) 9aT[o ystusers yep yn ‘uMoIg snosdeuIA yep szutof ‘Yysry "UID z 103}R] {(gS9 ‘ou) poyred-£ Apsour s1OMOT ‘use13 oul] SIMO “VIOL ut oInjeUT —— SNOIdeUIA Yep ‘YSry “wo € JaULIOF SO} eT[TSt z ‘97 RUT }S I ‘suorseyur € :9161 ‘Er 19qQ0100O "74 S1] Su0Iys ul Surmois ‘syutod sopuaary, ysmnyiq aed: yarm ynay “yng aAT]O_ yep quaied ‘fapeys asuap ut omnyno ‘u9013 SUIT] Inj ‘ayo 9AT]O swd}s uary ‘wd $b ‘az eTystd ‘suonsayur z : $161 ‘g raquiaydag ‘VY4SI] posnyip ut mais ‘ATO ystmorjad zYystT Apuojtun yuored {yy 31 qoalip ut pedojeaap MOTJIA DATTO soytds faydan PAI] yep surays ‘ysry ‘wo b ‘ayeT[Qstd sjyuyfd ‘ysoy 9UO UO sUOTPIaJUT Zz :Q161 ‘HI 19qQ0}90 ‘oerty ystjdind ymay jo wey joddn ‘u9013 oul] quored {Mora aArO suuaqs ‘posed -b pue © ‘morjad pear siamoy ‘ysty ‘wd Sz ‘ayeurureys sjurjd ‘fsuonsajut z :S$161 ‘Sz ysn3ny 4 6 ae Oe eee €1-9-9 £1-gI-g 11-11-11 ‘yse ‘aueyods z1-Of-o1 "JU, ‘BpNOsstpy Tret-o! oyepy ‘oyey ony ti-S-it oyepy ‘aus y,p 1n902 “YSeM [oO] eso] [e90"'] edoingy perk candi | [E00] *****(¢) vadoine xueyT "*(z) syeyuaproo0 xEwT "+ **(z) stfequaproo0 xUeT "+ >*(z) steyuapio00 Xue’ (pasvisAr sjusWaIMsvaUT) S][Nsay wor eure yo UoT}BAIOSgo ysiy Jo a2Rqy SHaOLTNO AO SLTASAY avd ANY aqvW AWM STAALIN aaaHM ALITVOO'T uMOI3 Ayye00'—T paas jo aoinog $ainj[nd UT posn Joqumu pur oweN SLSOH ‘TIVIM], ‘aTa1d AHL NI ACVW SAANLIND TIv “Sypjuapi990 X14vT NO S1914D] “Y WOUL [2161 ‘Of ANY O% AAMWALAAS ‘OHVA] ‘ATATY Isalag ‘1161 ‘bz wAAWALAAS ‘OHVA] ‘SAIAVJT “IS LV GALOATIOD GaAs I Wav WEIR—RAZOUMOFSKYA 1918] *(SgQ ‘ou) JINIJ paro[oo JayAIep YILM VATIO ystmorpad yySqy syuored ‘(99 ‘ou) ysydand poerveadde ynq Ajurez1909 YIM pourutajap jou JojOo [euIsu0 ‘ualyey Apsour pue prep azom sjuv|d 07 ep sty 72 nq ‘auIT} B 10J SNOIOSIA AT}UA -redde ‘ayeurueys fuorjejut I :g161 ‘FI 12qQ0}9Q9 ‘ayisoddo-ajeus193]2 Ajasop ‘pozred -b Ajjsour siamog ‘a[dind suasviyjUR 4soy owes UO SUd}s ‘ey BUTUE;s ‘(199 ‘OU) aAT[O YsI -MOy[oa qyst| yyy ‘umolq SNOIIRULA Yep Suld}s ‘ayeyusid fyuered {(099 “ou) poyzed- -€ A[jsou ‘ayisoddo-9}eu19}/2 Ajyue sip sIaMOV ‘qeip snoad “BuIA JEP syurof ‘aAo YstmoTjaA sutays “Ysty ‘wd © ‘ayBuTURs ‘UOT}IAJUT I :g161 ‘VI 19qG0}99 "(£99 “ou) MOTIA Aguoy yuered {(zg9g ‘ou) woH}IaTJOD JO aut} y@ Sno10sIA ‘ago ystuseis yaep ‘Yysry “WO a ‘gqeynsid syuyjd ‘uornoayut rt :$16r ‘Sz ysn3ny ($$g ‘ou) ajdind PLATT 2ep quered {(FS9 ‘ou) sdy ysiniq YM yy ‘BAO YSLMOT[PA JYSsIT sutays ‘YStY ‘uo z ‘ayesid ‘uorseyut © :gr61 ‘$1 139qo}9Q £1-g1-9 £1-gI-9g £1-61-9 (4 praee oyepl 4xv'T nig z1-9-O1 oyxpy ‘oye'y ani, 111-11 oyrpy ‘auayy,p ine0d zi-b-ol oyepl “TARY Sag [SOOT Teo] eoo'] ouepl [oT [eo] fOoT uvdef “s++***(g) stpuvid saiqy ****(Q) B]410]U0D snuTg eee (9) vsoiapuod snug “ts **(z) sidajoyda] xie'y 18 BOTANICAL GAZETTE [JULY favorable. The nature of the results on Pinus ponderosa and P. contorta, although demonstrating that this mistletoe under very favorable conditions will infect yellow pines, does not show any great affinity for the genus. When it is recalled that goo seeds were sown on 18 individuals of Pinus ponderosa, each receiving 50 seeds, result- ing in one infection, and one infection on P. contorta out of 12 trees tested with 600 seeds, the relationship between these 2 tree species and the larch mistletoe cannot be very close. The same is appar- ently true with regard to the infection of Abies grandis. Six trees were tested with the usual number of seeds, but only 1 infection resulted, which later died. These cultures also show that seeds germinating in the most vulnerable places only cause infection. Out of 500 seeds sown on Larix only 10 were able to cause infection, although apparently all the seeds which remained on the trees germinated. All were sown on parts of branches or shoots not over 6 years old, and care was taken to place the seeds favorably. It is to be expected that some of the seeds in outdoor cultures are removed by wind, rain, snow, insects, or birds. The observations relative to the favorableness of seed placement do not apply in the same way to the cultures on Pinus and Abies, since the larch mistletoe does not exhibit any marked affinity for these genera. That the same species of mistletoe growing on different hosts or under different conditions on the same host may exhibit different morphological characters is clearly demonstrated by these cultures. Since these experiments with the larch mistletoe were started, the following field observations have been made near Fernan Lake, Idaho. A large veteran western larch severely infected with R. laricis was left standing in a clearing which reseeded to Pinus ponderosa and P. contorta. From one each of these species growing directly under the larch typical, although small, specimens of the larch mistletoe bearing both pistillate and staminate plants were collected. The only true pine mistletoe in the immediate vicinity was R. americana. In a canyon near Missoula, Montana, where the larch is seriously infected with R. laricis and the pine mistletoes are not known to occur, specimens of the former have been collected from a single infection on Pinus contorta. These results are very 1918] WEIR—RAZOUMOFSKYA 19 much at variance with previous ideas of the host affinities of R. laricis, but they should not alter in the least the economic situation, since infections very rarely occur. The fact that both pines and larches are resinous may explain the occasional occur- rence of the parasite on the former hosts. Although great pains were taken to place the seed in favorable places on the trial hosts, the results on the following species were negative: Pseudotsuga taxtfolia, Pinus monticola, Picea Engelmanni, Thuja plicata, Tsuga heterophylla, Taxus brevifolia, Juniperus communis, Populus tremu- loides, P. trichocarpa, Betula occidentalis, Alnus tenuifolia, and _ Salix Bebbiana. Field observations on the intermingling of the branches of most of these species with severely infected branches of larch-bearing pistillate plants confirm the results of the cultures. Such observations, however, cannot be used as conclusive evidence for determining the host range for any one species of mistletoe. SuMMARY.—The hosts of Razoumofskya laricis are Larix occiden- talis, L. Lyalli, L. europea, L. leptolepis, Abies grandis, Pinus ponderosa, and P. contorta. The parasite is known to be of eco- nomic importance to the first named species only. The plant result- ing from an infection on any other host than that on which it normally grows exhibits considerable change in morphology and also in vigor. That different degrees of exposure with respect to light very greatly influence the color of the plants is very clearly demonstrated. Cultures with Razoumofskya species having purple flowers A group of small mistletoes found in the western United States has one character in common with R. pusilla of the East, namely, deep purple flowers.© They are R. Douglasii abietina (Engelm.) Piper? on Abies (figs. 11, 12), R. Douglasii (Engelm.) Kuntze on Pseudotsuga (fig. 15), and a small form on Picea (figs. 13, 14). A careful comparison of representative collections of these 3 plants from varied environments shows no constant characters by which they may be held as distinct species. All three have 2, 3, or rarely ° Wer, J. R., Wallrothiella Arceuthobii. Jour. Agric. Research 4:372. 1915. 7 Reported e ENGELMANN under the name Arceuthobium — var. abietinum in S. Watson, Bot. Cal. 2:106. 1880. { 20 BOTANICAL GAZETTE [JULY 2 / —2 22 y f — & oO % oe - a . — ee < = - poe = = Cae fie eo on F z > fe «= poo te & : : ge = ig E Ga f= 4 poe = fe wn : ee = NM Sn Se = = —< E = be = ‘a = es as = "ag = ae ees se Bey a ie be a= ALTA ; ag Ph 4 METRIC “212 || Fics. 11-15.—Fig. 11, R. Douglasii on Abies oo staminate and pistillate plants; fig. 12, R. Douglasii on Abies lastocarpa, staminate and pistillate plants; > 13s a purple- flowered form on Picea Engelmanni, staminate plant, natural small purple-flowered form on Picea ngelmanni, staminate flowers and Patillat oe: fig. 15, R. Douglasii on Pseudotsuga taxifolia, staminate and pistillate plan 1918] WEIR—RAZOUMOFSKYA 21 4-parted purple flowers, solitary or clustered, simple or branched, according to age of infection, bloom and fruit in the same season; and size of fruit, flower, plant, and color of stems show some varia- tions under different conditions of growth. Cross inoculations involving these forms should demonstrate whether or not all 3 are identical with R. Douglasii. The results of a series of cultures are given in tables III andIV. | At the time these cultures with R. Douglasit and R. Douglasii abietina were made seeds of the form on Picea were not available. The plant is not morphologically different from the other two, and cultures now under way indicate that it will infect Abies and Pseudotsuga. The evidence so far obtained is so pointedly in favor of the view that all 3 forms are identical that there can be little room for doubt. We find, for instance, that R. Douglasii will infect Abies grandis, A. lasiocarpa, and A. concolor, which are hosts for R. Douglasii abietina. No marked morphological differ- ences are found in the resultant plants and their parents, any more than is to be expected from a change of host or condition of growth. The same is true for the culture of this mistletoe on Picea Engel- manni. The evidence that all 3 forms are identical is further strengthened by the fact that R. Douglasit abietina from Abies lasiocarpa will infect Pseudotsuga taxifolia and Abies grandis, and that it is possible to fertilize the pistillate flowers of this form on the latter host with pollen from plants on Pseudotsuga. These results demonstrate the relationship of the 3 small purple-flowered forms here considered. The two forms on Abies and Picea should be considered identical with R. Douglasii in view of the foregoing results. It has already been pointed out that, in the writer’s experience, the plants on Abies and Picea are in most cases found in localities where R. Douglasii abounds. If the former were specifically distinct, with inherent tendencies to select their par- ticular hosts, they should in the light of our knowledge of the well defined species be more abundant. On the contrary, they are never found in any quantity. The conclusion that R. Douglasii does not abundantly infect other trees than Douglas fir is also shown by the following observations. The writer has looked several times in vain for infection of this species on Abies and Picea when the [yuLy BOTANICAL GAZETTE Pi “@SnoUUuseLty } “PPY BL SAangNy » onagn wey UTIOD yep sIaMOR ‘ajduits ‘(99 ‘ou) aayo ustmoie4 eae syurjd Star BIUOUIIA SIaMOT “perpaeiq ‘Areqyos §=‘ysry "wd Sz ‘ayeUTWeIS SUOTJIFUI I ‘apeys Ut MAID ‘poyed-£ ‘ajdind viuouiaA siaMoy ‘ynq aAT]O ‘saieds 104 yy? JoyIeI ‘payouvaq ‘pesaysnyo “ysty ‘wo S*% ‘ayeurlUEys = fy YS] Wap Uy MII3 ‘any ysttdind YA ‘aAqfo Yystmorpad * payouraq ‘paiaysnypo ‘ysty "wo ¢ ‘yuoied ‘(oL9 ‘ou) aprys ut ‘poyied-€ pure z ‘ajdind vruouisa saamoy :g16r ‘br pludy | €1-L1-9 471-S-O1 oyepy ‘IaAY ISI “*** (Z) TuURUNpesU vat ‘QayO Ystmopjed yYySy ‘zapuays ‘Arezyos “Ysry ,1I-O1-11 "wd z 9] BUIUIR}S “WorjdajUL fe 79161 “Pr judy ee ee “yseM ‘guvyods ew bee he eee we ee ee eS eee eee *(z) 40j09U09 soiqy "3INI} Po10[09 19743] Apysiys ya Avis aayo daap aiam “ysry “woe ‘ystumoiq quered ‘pansjopnas uO UMOS A]a}RIPIUTUT 319M pPsaIs ‘QuejysIp “Wz 919M YOTYM VSnsyopnasg ua u0y ~BNIOUL WO Surnnsar sjurjd wor pazt[n49j SIOMOP ‘9InzvUI ‘dAT[O ystuaers yaep yny ‘QAT[O ystmoypad yyst ‘Areqyos ‘ajduns ‘ysry ‘wo z «cI-S—OI sjuryd ‘ajeynstd Suonseyur tr :g161 ‘hi 1aqo}9Q | F1-LI-9 oyepy “IaATY Isa ‘PAPER et ee s**t) spe ay “u0r} |+1—bz-o1 -Ipuod 100d ut ysoy {yno parp ynq uoTdeJut UG | SI-9-z “quOyy ‘RpNOsstpAy “U0, ‘quoyy | °° °° (Z) SIpueis saiqy -paivadde Jaaau sjurjd ‘suonsayut | v1—-bz-o1 quaivdde ‘sjyurod jes9Aas ye Buljjams paounouorg | S1-g-< “quopy ‘B[NOssIPy "UOWT ‘quopw (°° °°‘ (z) vdavoorsey sarqy ‘OpRYS UT 197} P] qyslL WoIIp Ul IBULIO] “SIOMOZ pes 2 soy YM = aArfo ystmoyad yySt ‘yuU9 ‘din ysrppryq sromoy ‘Avi aatjo yrep ‘ArezyOs ‘Ysty 4c1-S—O1 ‘wo $'1 ‘ayeutueys fuoT}DIJUL I :g161 ‘FI judy €1-L1-9 oyep] ‘AY SAN | ee sg ee “*(@) BYyOyIxe] esnsjopnasg uornvurMsed (pasvisAe s}usUIsINsvaW) s}[Nsay jo uOTBAIasSqO ava Paced! yo ae pasn sequen pes aweny 1 JO VC | axv aaqvm AIM SISAL : S4aHM ALITVOO’T SISOH ‘TVIEL, SIANLINO 40 SLINsAy “‘pyofixn, DINSJOpNasST NO usp/Snog “y WOAd ‘VNVLNOJ ‘VINOSSIP Lv ‘F161 ‘I ATAOLOO ANV ‘e161 ‘oz ATAWALdag ‘1161 ‘ge AAOLNQ AALOATION dadag Ill ATAVL WEIR—RAZOUMOFSKYA 1918] ‘AO YsIMoTad 7yS] “Ysty “woe ‘poyouesq i I l t zI—g-O1 ‘Areyos ‘az eq ystd SUOTPIOJUL I 29161 ‘PT 19qoyIQ | £1-L1-9g oyepy “I9ATY ysaug [e00'T yeoo’T | °° **** *(P) sipuel3 saiqy ‘readde jou pip jurjd ‘yoursq z1—6—o1 jo Surjams ‘uonooyut yuoredde 1 :g161 ‘vr judy | €1-g1-9 oyepy ‘ayxe'T ong ye00'T [eoo'T | °*(¥) vryoyrxe) eZnsjopnasg *IO[OO 19;431y e@ jo puv radi], AYsys a19M syurtd yuored jo SIOMOY o}eUIUTE}s JO UOT}dadxa dy} YIIM JIN} vy qeynorjaed Aue ur Suteyip you ‘(649 ‘ou) quered ‘(g49 ‘ou) afdind ysryorq siomog ‘Qal[O YSIMoT]aA 34ST] ‘poyouvsq ‘Ares ‘YstYy ZI-9-O1 ‘woz ‘ayeulueys fuoljoejur rt :g161 “v1 Judy | £1-L1-9 ogepy ‘I9Ary yseIg [B00] [eoo’'T | -*(¥) vroyixe) eBnsjopnasg (padvisav sjuautainsvoul) s}[Nsay fog pie pare uMoI3 pose sainj[Nd ut ysayf Jo aye] AjI[290°T JO aoanog pasn Jaquinu pur ouieN AIVG GNV ALITVOO'T SaMALIND Ao SLTAsAY | SLSOH ‘IVIN, ‘Dg 4090180) Sa1qY NO PuIQIgv NISDPINOG “Y WOUA ‘161 ‘c AATOLNQ ATTA GNA ANV'] AVAN NIVINAOW ATV NO.aa19aT109 ata AI ATAVL 24 BOTANICAL GAZETTE [JULY latter grew in absolute contact with brooms on Douglas fir bearing pistillate plants. As previously stated, however, this is not con- clusive evidence of the host range of a species. Accident of infec- tion is too great; besides, trees growing in such juxtaposition are very often suppressed, thus reducing the amount of vulnerable tissue. These results were obtained only by the most careful placing of the seeds at the most susceptible points. In the course of years such conditions occur in nature. Afterward, as a matter of course, further infection from the parent tree to others of its kind may be easier. The type of broom produced by R. Douglasiz varies with age and host. On hosts with strongly excurrent growth, such as Abies lasiocarpa and A. grandis, the brooms are usually erect, but drooping or swaying forms occur. The erect type of broom is common on Pseudotsuga taxifolia during the first years of infection, but later may assume the weeping willow form. Seeds from plants on Abies lasiocarpa were sown on a single individual each of Tsuga heterophylla, Larix occidentalis, Pinus monticola, Thuja plicata, and Populus trichocarpa, but without results. Seeds from plants on Pseudotsuga taxifolia were without result on these hosts and also on Larix europea, Picea sitchensis, P. canadensis, P. excelsea, P. Parryana, Sequoia gigantea, Pinus ponderosa, P. contorta, P. Jeffreyi, P. sylvestris, Betula occidentalis, Alnus tenutfolia, and Pyrus. ‘The several species of Picea were not in a vigorous condition, having been transplanted only a short time before the seed were sown. SumMARY.—The foregoing cultures indicate that Razoumofskya Douglasii abietina is identical with R. Douglasii. The hosts. of R. Douglasit as known to the writer are Pseudotsuga taxifolia, Picea Engelmanni, Abies concolor, A. grandis, A. lasiocarpa, A. nobilis, and A. amabilis. The species is of economic importance only on Pseudotsuga taxifolia. Cultures with lodgepole pine mistletoe This species (R. americana |Nutt.| Kuntze) (figs. 16, 17) is one of the most characteristic of the genus. In order to determine its host range, the results of some recent cultures are presented in table V. It is shown that Pinus contorta is the true host of R. 1918] WEIR—RAZOUMOFSKYA 25 americana, but that occasionally other hard pines are attacked. The writer has previously reported the occurrence of this mistletoe on Pinus attenuata, P. Jeffreyi, and P. ponderosa, and it has long been known to be common on Pinus Banksiana in Canada. The fact that this mistletoe will infect Pinus montana, the common moun- tain pine of Europe, further supports the writer’s conten- tion that it may be expected to occur occasionally on any of the hard or yellow pines, and also is a warning that the parasite would probably find a favorable home in Europe. ae plant apparently attacks Fic. 16.—R. americana on Pinus bade. the yellow pines other than __,,.. oO Ger plants. Pinus contorta with difficulty. Such infections are by no means common, and frequently result in some morphological changes in the plant. These changes, how- ever, may not be any more marked than those the plant may exhibit when developing under various light intensities or varying conditions of nourishment on its regular host. If R. americana ex- hibits a certain antipathy to other yellow pines, it appar- ently has a much greater aversion to white pine. That Fic —R. americana on Pinus contorta: the species will infect white viseitate' binsita: : . a pines but with difficulty, and will never be of consequence in this respect, is shown by the discovery of two infections on Pinus albicaulis near Darby, isi Ayo a ay spay UW metric !isystem “f ‘asnoyuvedy | "pyay ul saangyna y <) SS 2 eR eet Bc hl bes ha *(699 *ou) |[ewWs poured nq uor}IaFur JO yurod yeurst10 aAoqe ysnf porvadde yuyd sayjoue ‘yno pap ‘S161 ut pareadde Ajareq syurjd yews zc ‘usar jo uoneapyut DAISSAIXA “YOuRIG JO ZuljjaMs poounouold ‘ysry ,1I-OI-I1 “wun $ ‘oyeUTUreys SuOTDayUT I 9161 ‘FI Jaquiadaqy |" ‘yse ‘ouryods | ouryodg wen tt (v) (Woy JAVMp) vuRyUOU snug ‘rvaddv jou prp syuryd ‘parp youraq {S$1-e1-1 ‘BuT[aMS WYySYs “UoNIeJur 1 :$16r ‘or 3saqoyQ | S1-g1-z "qUOPT “BPNOssT Py “quo wow *(1) esorapuod snutg “(899 ‘ou) mw) Why puv susazs yYoq ‘Avis aAl[o Yrep ATULIO] Tun ‘ysiy “wor ‘azensid fused adeviZ daap = u1a38 Jo syed Japfo yIIM aAr[o YsIMoTaA 731] N ysiy “wog ‘ayeurME}s {QUAI :(499 ‘ou) TZ «ANG Snooneys [Np yYSy yay ‘aatjo ystmoy[aA S sways ‘ysty "wo S 1937R, SmoTaA aato suraqs S ‘y8ry ‘wo F sauL0; ‘aqyeqnsid pure ozeurure}s eS sjurjd ‘payoayur 9013 1 ATUO ‘wajs a[suIs &B 4c1-Z-O1 > UO 9pou aUTRS dy} 7B SUOTJIaJUI z :gr61 ‘Ex Udy | £1-g1-9 - oYepy ‘ayey oni [v00"] [e007 |' °°’ **(¥) vsosapuod snurg S ‘sjurjd juored wo so1%ap payreu Aue ut S Burke you ‘(999 ‘ou) ayeI[NsId z ‘ayeuTUIE}s Z : ‘W938 UO YOURIG Jo SuITJeMs 7YySYs ‘9017 yore 4ZI-Z—-O1 Uo UOTPaJUT I :9161 “VI 1aq0}9Q pue £1 pudy | £1-gI-9 oyepy ‘ayxe'yT ong [v00'T [eoo'T | °° °° (#) By10}U09 snulg ht bab; meee umois paes sainyjno ut (padvioAR SyuaWIaINSvaW) s}[Nsay jo WOT} BAIASqO ss . 1g JO N° | save ptt >» of AyypRI0"7 JO 90n0g posn Joquinu pur owen aA4aHM ALIIVOO'T S€aNLIND AO SLTASAYy SLSOH ‘IVI4 J, "DJ40JUOI SNUIT NO DUDIAIMD “y WOAd ‘V161‘FI wAMWAAON ‘OHVG] “ANATY,d ANAOD IV GNV ‘2161 ‘I AAGOLIG ‘ONVAT ‘AAATY Isdaq ‘1161 ‘gt aad@OLOQ ‘VNVINOJ ‘AGAI] LY GXLIATION aaasg oO . " A ATEVL 1918} W EIR—RAZOUMOFSKYA 27 Montana. The parasite in this case caused unusually large and elongated swellings on the main stem of young trees, but the plants apparently were never able to come to maturity, remain- ing about 5-8 mm. high. One of the pines was transplanted into the greenhouse, and the context of the swelling shriveled up in a manner indicating that it was composed of very spongy tissues. The tree, however, remained living. The fact that 220 seeds of R. americana were sown on 6 different species of white pines with no result except the germination of the seeds further supports this observation. The trees tested were Pinus Lambertiana, P. monti- cola, P. Strobus, P. edulis, P. cembroides, and P. Cembra. R. ameri- cana is reported by CouLTER and NEtson® on Pinus flexilis. The results of sowings on Larix europea, L. occidentalis, Picea sitchensis, P. Engelmanni, P. excelsea, Abies nobilis, A. lasiocarpa, A. grandis, Tsuga heterophylla, Pseudotsuga taxifolia, Thuja plicata, Taxus brevifolia, Populus trichocarpa, Betula occidentalis, and Alnus tenuifolia were negative. SUMMARY.—The hosts of Razoumofskya americana are Pinus _ Contoria, P. Banksiana, P. attenuata, P. Jeffreyi, P. montana, P. ponderosa, P. flexilis, and with difficulty P. albicaulis. The plant is of economic importance so far as known only on the two | first named species. Morphological changes are induced by change of host or condition of growth, but not to an extent that this, the most characteristic of all members of the genus on pines, could be confused. Cultures with hemlock mistletoe In the St. Joe National Forest, Idaho, are several areas of almost pure stands of Tsuga heterophylla heavily infected with R. isugensis (figs. 18, 19). In the border zones of these areas a form of mistle- toe has been collected on Abies grandis and A. lasiocarpa which varies in a number of details from the form collected on the same hosts in regions where the large mistletoe on Pinus ponderosa occurs. In order to see whether this is a case of R. tsugensis infect- ing other hosts than the common western hemlock, and also to determine its host range in general, the cultures given in table VI were made. § New Manual of Botany of the Rocky Mountains. 146. 1909. 28 BOTANICAL GAZETTE [JULY R. tsugensis is not confined to species of Tsuga as heretofore believed, but will infect Adzes lasiocarpa. The mistletoe most closely resembling R. tsugensis in point of color and size is the form R. occidentalis abietina, but as the results from cultures stand at present there is apparently no relation between them. ‘The fore- going results indicate that the plant occasionally found on firs in the same vicinity with R. tsugensis is the common hemlock mistletoe, Fic. 18.—R. tsugensis on Tsuga heterophylla: staminate and pistillate plants; large form. and also that this species may be expected to occur occasionally on other hosts than hemlock. Cultures may be considered fully completed when the plants found on Abies in the vicinity of R. tsugensis, also the form on Abies which has been referred to the yellow pine mistletoe, are shown by culture to infect Tsuga and Pinus respectively. SUMMARY.—Seeds were also sown on Abies grandis, Pinus ponderosa, Picea orientalis, Larix occidentalis, and Pseudotsuga taxifolia, but the results were negative. The hosts of Razoumofskya 1918} WEIR—RAZOUMOFSKYA 29 isugensis are Tsuga heterophylla, Tsuga canadensis, and Abies lastocarpa. So far as the present cultures show, the hemlock mistletoe will not infect Pinus, Picea, Larix, and Pseudotsuga. The fact that this mistletoe will infect Tsuga canadensis indicates the possibility of it becoming a pest in the native regions of other species of hemlock and is a condition to be guarded against. Conclusion Cultures at present indicate that R. campylopoda and R. crypto- poda are not identical. Each form may exhibit considerable varia- tion, due to geographic loca- tion and host. It is shown that R. campylopoda will infect Pinus resinosa, and care must be taken to pre- vent it from getting a foot- hold in the eastern United States. It will also infect Pinus sylvestris and P. mon- tana, and should be _ pre- vented from entering Europe _F'- 19. tsugensis on Tsuga meriensi- . ana: staminate (center) and pistillate plants; or plantations of these trees mall form; reduced one-half. in America. It is also indi- cated that the plant known as R. occidentalis abietina is a biological form of R. campylopoda. R. laricis will infect Larix europea, L. leptolepis, Abies grandis, Pinus ponderosa, and P. contorta. All are new hosts for this species except the last. The parasite apparently readily infects the Japanese and European larch and would be expected to cause serious damage to these trees. Abies grandis, Pinus contorta, and P. ponderosa are infected with difficulty. This parasite so far as known at present is = economic importance only on Larix occi- dentalis. The mistletoe known under the name R. Douglasii abietina is shown to be identical with R. Douglasii and should be written under the latter name. R. Douglasii is only of importance on Pseudotsuga taxifolia. [JULY ‘asnoqudars) | “play Ur SaINyIND y *(£gQ ou) aAT[O ystuMOI FYSIT] JUdIVd £(zgq ‘oU) aAT]O Ystusais yep “Ysry ‘wor “Buypams Wyss 421-L-O1 Azaa ‘ayeutwmeys “uorajut I :9161 ‘zr raqo}IQ | S1-1-S “ySt Ay ‘gueyodg “STM WA To: (1) sisuspeurs Bans “QAT[O YSLMOTTOA ayaty ‘ysty “wo 9 “juared {(9g9 ‘ou) useI3 yeay ‘ystYy |S1-gz-1 "woz ‘ayeuTuEys ‘uoT}DayuI I :L161 ‘€ Arenuef | g1—-O6z-< “quo ‘B[NOssIPy *qUuOyT aww" (1) edavsorsey] sarqy WOT wergURIOS umoi3 paas sarny[no ut ee eee ee peice a@LvVd GNY TQVW aAaM Ayv90'] yO a2In0g pash zaquinu put awe oN SHMALIND TAAHM ALITVOO'] SAMALIND 40 SLIASAY SLSOH ‘TVIM], BOTANICAL GAZETTE bi61 ‘of aaaWALdaS ‘OHVA] ‘ISHXOJ IVNOILVN AOf “L§ NI LSOH ANVS AHL WOAT GNV ‘NOLONIHSV\\ ‘GuOdHSy Wow INAS DKYGo4ajay DINST NO SINVTd AOLATLISIN UNV SWOOUM OL ONIAAHGY ‘$161 ‘cz AUVANV[ ‘VNVINOJY ‘VINOSSIPY LY GALOATION aaag 30 TA ATAVL 1918] WEIR—RAZOUMOFSKYA : 31 R. americana will infect both hard and soft pines, the latter with difficulty, and is of importance only on Pinus contorta and P. Banksiana of the former group. This mistletoe will infect Pinus montana and may be of consequence if introduced into Europe. R. tsugensis will infect Abies lasiocarpa, thus possibly explain- ing the position of certain rare plants occasionally found on Abies in the vicinity of the hemlock mistletoe. This parasite will infect Tsuga canadensis and would probably cause serious damage to this tree in the East. Cultures show very clearly that many of the characters employed in the classification of the false mistletoes vary with change of host, geographical location, and with various other environmental factors. This indicates that only the broader and plainly evident lines of demarcation should be employed in their classification. OFFICE OF INVESTIGATION IN FoREST PATHOLOGY UREAU OF PLANT INDUSTRY Missouta, Mont. CHEMICAL CHANGES ACCOMPANYING ABSCISSION IN COLEUS BLUMEI CONTRIBUTIONS FROM THE HULL BOTANICAL LABORATORY 240 HomeER C. SAMPSON Introduction Von Monat (13) in 1860 was the first to announce that previous to the fall of the leaf there is formed near the base of the petiole a definite separation layer in which abscission always occurs by the separation of the cells from each other with their walls still intact. The xylem tubes not being included in this separation layer are finally ruptured mechanically, and the leaf falls. He also called attention to the fact that abscission and the formation of a pro- tective tissue are two very distinct processes, and that the latter process might either precede or follow the former. WHESNER (17) in 1871 confirmed the observation of voN MOH in the main, and formulated the theory that the dissolution of the intercellular sub- stance of the cells of the separation layer is caused by the action of organic acids developed in the leaves. In 1886 Motiscu (14) suggested that a gum ferment might be the cause of this dissolution process. Two years later MANGIN (11), upon his discovery of the pectic nature of the middle lamella of cell walls in plants, indirectly advanced the knowledge of abscission. Since this discovery the abscission process has generally been referred to as a dissolution of the pectose and calcium pectate of the middle lamella. LLoypD (8, 9), assuming the organic acid theory of WIESNER, speaks of the process as a hydrolysis of these pectic compounds, and later (11) of cellulose also. The anatomical workers disagree somewhat on the amount of the cell wall altered during abscission. Lrxr (7) in 1911 reported the disappearance of the middle lamella only, and two years later Hannic (6) reported the same condition in the abscission of flowers, with the exception of a species of Mirabilis and of Oxybaphus, in which the entire cell wall disappeared. On the other hand, T1son Botanical Gazette, vol. 66] [32 1918] SAM PSON—ABSCISSION 23 (15) as early as 1900, working with numerous species studied later by LEE, stated that in general the secondary membranes of the cell wall also undergo alteration and disappear, leaving only the thin tertiary membrane lining the cell lumen. Ltioyp (9, 10) has recently found the same condition in the abscission of cotton bolls and also in Mirabilis, instead of the disappearance of the entire cell wall as reported by HANNiIG. It seems improbable that the presence of weak organic acids would be sufficient to account for this extreme alteration of the cellulose walls of these cells. Certain authors - have suggested that the catalytic effect of enzymes may be an important factor in abscission, but experimental evidence has been wanting. Aside from his organic acid theory, WIESNER (21) in 1905 suggested increased turgor as a cause of abscission under conditions of forced leaf-fall. Frrrrnc (5) in 1911 accepts this view to account for the rapid abscission of petals when forced. The suggestion lacks experimental confirmation, and the work of Hannic (6) indicates less need for its assumption. The external factors capable of accelerating leaf-fall are extremely diversified. These have been summarized in the main by Lioyp (8). The more important are high and low light intensity, high and low water supply, high temperatures and frost, low con- centrations of anesthetics, toxic-concentrations of acids and salts, and wounding of the blade. On the other hand, low concentrations of oxygen and high concentrations of anesthetics retard leaf-fall, a state of rigor being produced by the latter. The internal changes affected by these various external factors have received very little critical study. The work discussed in the present paper was undertaken to determine some of the internal changes accompanying abscission of leaves in Coleus Blumei var. Golden Bedder. This plant was chosen for study partly on account of its ease of propagation, but mainly for its simplicity of analysis owing to the absence of protective tissue at the time of abscission. Anatomy In order to appreciate fully the chemical changes taking place in the abscission layer, it is necessary not only to compare the 34 BOTANICAL GAZETTE [JULY abscission layer with the adjacent regions of the petiole, but also to follow the changes in the abscission layer itself from the time of its formation to the fall of the leaf. This latter process is most easily accomplished in Coleus by beginning with the terminal bud and taking the leaves as they appear in order down the stem, from the youngest to the oldest. On this basis the following description is applicable to Coleus plants growing in 4-inch pots under green- _ house conditions, with each plant bearing 8 pairs of leaves, and with the eighth pair in the process of abscissing. In the first 2 pairs of leaves below the terminal bud there is no evidence of an abscission layer. The cells in the region where the abscission layer is later to occur are in the enlargement period of growth. The formation of the abscission layer in Coleus is usually initiated in the third pair of leaves below the terminal bud. Cell divisions in 2-4 layers of cells across the base of the petiole of these leaves begin in the epidermal and cortical region and gradually extend inward, reaching the phloem about the time the leaves appear as the fifth pair in order below the terminal bud. In the sixth pair of leaves the formation of the abscission layer is practically completed and involves all the tissue of the petiole except the xylem tubes. Growth of the leaf in general ceases long before this period is reached. In the majority of cases the fourth pair of leaves below the terminal bud are fully expanded. The formation of the abscission layer in Coleus therefore begins a short time before the maturity of the leaf and continues for a considerable period afterward. As a rule the layer is 8—12 cells in thickness. These cells always remain smaller than the neighboring cells of the adjacent regions of the petiole and their walls are somewhat thinner. At the time of abscission alteration of the walls of the cells of the abscission layer is quite general, but a continuous plane of separa- tion is finally formed somewhat nearer the distal side of this layer. This extreme alteration of cell walls is localized in the abscission layer and is not found throughout the entire leaf, as was reported by WIESNER (18). While this description is true in general for Coleus plants bearing 8 pairs of leaves, slight variations are not infrequent. The stages of development may be either retarded or accelerated. Under 1918] SAM PSON—ABSCISSION 35 conditions of forced leaf-fall all processes are greatly accelerated. The data in table I show that all processes, including the process of abscission, may be completed in the third pair of leaves in a period of 2 or 3 days as a result of the amputation of the blades. On the other hand, amputation of the blades of the first and second pairs of leaves before the beginning of the formation of the abscission _ layer inhibits its formation entirely. Method of abscission UNDER ORDINARY GROWING CONDITIONS.—The method of abscission has received much attention, but a critical survey of the exact changes in the cellulose and pectic compounds is wanting. An attempt to follow these changes in Coleus led to the discovery of certain facts which have a direct bearing upon the existing theories of the cause of abscission. Microchemical analyses show that there is a breaking down not only of the calcium pectate of the middle lamella of the cells of the separation layer, but also of the cellulose of the secondary membrane, leaving only a thin layer of cellulose surrounding the lumen of the cells. This cellulose is first changed to pectose, which, according to Cross (1), ToLLENS (16), and EuLER (3), contains more oxygen than cellulose and probably is an oxidized form. The pectose is then further changed to pectin and pectic acid; the excess pectic acid becomes gelatinous and is no longer able to hold the cells together, and the leaf falls. The changes taking place in the walls of the xylem tubes are still to be investigated. During this process there is not a disappearance of calcium from the cell walls, but the excess of pectic acid produced renders the amount of calcium present’ insufficient to maintain the solidity of this portion of the cell wall. The excess pectic acid appears to come from the transformation of the pectose rather than from the breaking down of the calcium pectate of the middle lamella. Since this description differs from all previous accounts of the method of abscission, further discussion is postponed until all the facts are brought together under the topic of microchemical analysis. FORCED LEAF-FALL.—Leaf-fall was accelerated by treatment with ethylene, amputation of the blade, and by allowing the soil 36 BOTANICAL GAZETTE } [yULY to become dry and then suddenly applying an excess of water. Under the first two treatments the leaves began to fall within 24 hours. The rate of petiole fall after the amputation of the blade is shown in table I. At the beginning of the experiment each plant had 8 pairs of leaves and 1 blade of each pair was removed. The numbers refer to the total number of abscissed petioles at the cor- responding dates. TABLE I RATE OF PETIOLE FALL FOLLOWING AMPUTATION OF BLADES January February March Plant 27 28 29 | 30 3r I 2 I 2 3 4 5 6 > Io Ber x I Be Ae Piel ee OT a Ee bee, ere Mag < Bo BoA ee ds ee OO ee ek ye eas 4 Gee cel cbvwug hire loge ly vas le cea be Re atees of oS obec vec n 6 |. 5 ee mee Wier alee heehee ce ie ea eae Ny 64, a gee Die obey ok Pecan vate. os pee ey ele a pat ae 6 61; Go * Blades amputated January 26; { blades amputated March 1. A microchemical analysis of the abscission layer in all these cases showed exactly the same changes in cellulose and pectic substances as noted under ordinary conditions of growth. Further- more, changes in oxidases, calcium in solution, and iron, to be . discussed later, were the same in all cases. These facts emphasize again the need of experimental investigation before accepting the turgor pressure theory of the cause of abscission. It is interesting to note that while the petioles usually absciss soon after amputation of the blade, there is one striking exception. If the blade is removed before the abscission layer is initiated, growth ceases throughout the entire petiole, the layer fails to develop, and the petiole is not dropped. The data in table I show that the 6 lowest petioles soon fall, but the upper 2 remain attached. This is an easy method of locating the period of forma- tion of the abscission layer. This extreme cessation of growth in the petiole induced by artificial means has some features in common with a retardation of growth and abscission formation in the petioles of the upper leaves under natural conditions of flowering and fruit- ing. Asa result of slowing up of growth and formation of abscission 1918] SAM PSON—ABSCISSION ‘ 37 layers in the petioles of the upper 3-5 pairs of leaves accompanying the development of the floral axis, these leaves remain on during the entire flowering and fruiting period. An investigation of the internal changes accompanying these 2 phenomena may throw some light upon abscission in general. Organic acids as a cause of leaf-fall WIESNER (17-21) cites three lines of experimental evidence as proof that the dissolution of the middle lamella is a result of the accumulation of organic acids in the aging leaves: (1) yellow leaves macerated and extracted with water when titrated were more acid than green leaves; (2) cuttings placed in 2.5 per cent oxalic acid dropped their leaves in a few days; (3) the exposed abscission surface of the petiole always gives an acid reaction to neutral red. Abscission in Coleus when examined from the point of view of this theory shows several facts in disagreement and not one in its favor. Cuttings placed in non-toxic concentrations of oxalic acid showed no acceleration of leaf-fall over that of cuttings in distilled water. Cuttings in 0.0002 N oxalic acid showed slight toxic effects. The immersed part of the stem and the tips of young leaves on axillary branches became brown in color. The concentration of acid used by WIESNER was 1500 times as great, but he fails to state what plants were used and whether toxic effects were produced. Cuttings of Coleus in 0.0016 N oxalic acid did not show an accelera- tion of leaf-fall, although the toxic effects were strongly pronounced. It was further found that the plants soon became adjusted to the oxalic acid. Plants started in 0.0002 N oxalic acid were transferred every third day to a concentration of acid double that of the previous concentration. This was continued until the plants were finally placed in 0.0512 N oxalic acid. There was no acceleration of leaf-fall during the entire period. At the end of the treatment the cells of the plant were found to be filled with starch. Similarly, potted plants infiltrated with non-toxic concentra- tions of oxalic acid showed no acceleration in leaf-fall. The plants were inverted under bell jars in vessels containing the various solutions, and the air was exhausted to 3cm. of mercury. The volume of solution entering the infiltrated plant was approximately 38 BOTANICAL GAZETTE [JULY equal to one-fourth the volume of the plant. No toxic effects were noted for concentrations of 0.0064 N oxalic acid and below. The results are given in table IT. TABLE II SHOWING EFFECT OF DIFFERENT CONCENTRATIONS OF OXALIC ACID RATE OF ABSCISSION February March Concentration 5} 16 8/19 3 6| 27| 28} | 2/3) 4] 5 @.0256 N oxalic acid... ...,.. no baat Bhs vb Bie Lg} SES. 1 Gis wee yf Ee Ta? Se Aor 6 6 On 6.0005" = * ee Pe I a oa eset gl SE Cie. 0.0032 “ . st ae I cols ciicpent Ste cP Sti. ele 9.0016 © % Spe ao an Cv BEC Eeee 2) 3| 4) S} 6. rete a el fd aT LC] ay of fe] ake Vintieated ik ei ech ky Be ae ay 8 ae i | age vei Mt ean an wa uo kel a sys i Jat ae Oa 8 The plants were infiltrated February 9, and again February 22. Leaf-fall was allowed to occur normally under greenhouse condi- tions. The numbers refer to the total number of leaves off at the corresponding date. Although concentrations of 0.0128 and 0.0256 N oxalic acid showed marked toxic effects, abscission was not accelerated; con- centrations between 0.04 and o.12N oxalic acid killed many blades without killing the petioles and stem. In such cases abscission of the petioles occurred within 2 or 3 days, just as in the case of petiole fall after amputation of the blade, or severe wounding of the blade. Likewise direct measurements of acidity do not agree with those of WiesNER. Table IV gives the acidity for 9 different regions of the plant. Falling leaves are not so acid as green leaves. Fresh yellow leaves in the act of abscissing when titrated with NaOH, using phenolphthalein as an indicator, had an acidity equivalent to 0.0069 cc. of normal acid per gram of wet weight. Fresh green leaves collected at the same time from the same plants had an acidity equivalent to 0.0089 cc. of normal acid per gram of wet weight. In both cases the leaves were weighed as rapidly as pos- 1918] : SAMPSON—ABSCISSION 39 sible after collecting, macerated in a mortar, made up to volume with distilled water, and after shaking for 30 minutes the solutions were filtered through a Buchner funnel and definite portions taken for titration. Fresh abscission layers treated in this way had an acidity of o.oroocc. per gram of wet weight, while that of the adjacent part of the petiole was o.0c0g5 cc. These two figures are not to be compared with those preceding, as the two sets of titra- tions were made at different times and on different plants. If the calcium pectate were being hydrolyzed by an organic acid, one would expect to find either an increase of calcium in solution in the cells of the abscission layer or an increase of crystals of calcium compounds in these cells. Such is not the case. Neither are there any calcium oxalate crystals in the middle lamella of these cells, such as one finds when the middle lamella is broken down by adding oxalic acid to sections under the microscope. Finally, the marked acidity of the abscission surface of a falling leaf was certainly not correctly interpreted by WiESNER. He ascribes the acidity of the abscission surface to the excretion of organic acids from the interior of the cells. This abscission surface is a continuous layer of pectic acid, formed during the abscission process, and the acidity of the abscission surface, therefore, is a result of the formation of pectic acid during abscission, and not of the escape of acids previously formed in the cells. This acidity of the middle lamella to neutral red may be seen in Coleus in any part of the plant, and it is increased in the walls of the abscission layer only after the formation of pectic acid during abscission. In conclusion, therefore, neither the turgor pressure theory nor the organic acid theory proposed by WIESNER to account for the cause of leaf-fall is in accordance with the facts observed in Coleus. Effect of salts on leaf-fall According to CZAPEK (2), the membranes of plant cells are colloidal in nature, and MANGIN has shown that the middle lamella is composed of pectic acid in combination with calcium. During the process of abscission the middle lamella undergoes a chemical alteration and the pectic acid present takes up water and swells. It was expected, therefore, that salts might show either a lyotropic 40 BOTANICAL GAZETTE ' [JULY effect on the intake of water by the pectic acid or a specific effect of salt formation with this acid and thus affect the course of abscis- sion. The following anions were used in the form of their potassium and calcium salts: PO,, SO,, Cl, NO;, and CNS; and the following cations in the form of their chlorides: K, Na, Ca, Ba. In all cases ©.o1 normal concentrations were used. The plants were treated by infiltration, by placing cuttings directly in the solutions, and by adding the salts to the soil. Abscission was slightly accelerated by treating the plants with 4 parts of ethylene per million of air. In all cases the results were the same. Neither lyotropic nor specific effects were noted. Similarly, concentrations of potassium and calcium chlorides between 0.04 and 0.00016 normal showed no marked effect. The experiment was repeated under conditions of rapid accelera- tion of abscission by treating the plants with 700 parts of ethylene per million of air. The results which agree with those above are summarized in table ITI. TABLE III SHOWING EFFECT OF SALTS ON ABSCISSION Jose FEBRUARY I5 FEBRUARY. 14 FEBRUARY 13 CONCENTRATION § P.M. Q A.M. 5 P.M. 9 A.M. 5 PM. 9 A.M, 5 P.M. ‘o-6) NBO oie a a eek 3 a ee irs 5 me ane Oe PUES Sb estiwe ce pe oge ees 2 3 Ae 4 Pe CASO ee ea ee 5 Or ee 6 Co UROL te oe ees 3 eer Spe es 5 oP RIN Ste oe 2 - GES is SIRI BERG 5 . CaC(NO De) So a ea mais a ree 4 8 RONG. se eee owe ee eee 4 4 WRANE eo. cic a i i 3 4 Se eh aes 2 5 ap ene ae ay com a nN 6 Pe ae kd eect bes eo waked Henn cere 3 3 WS icra ee tee Go ie. 6 6 Oey a ede as Ee ee Fe eee 6 6 6.08 MWB. cs ee is eee 3 4 LE 5 “ Ase Oo ie a Gs 3 4 ee eae 4 «© CeCe he. a Se 4 PENG aes 4 yg get, See Pe as ee 2 4 Bo bicwee, 4 _ Cuttings in solutions of the different salts gave similar results. The great diversity of the checks noted in this table is very unusual. 1918] SAMPSON—ABSCISSION 41 The failure of calcium to show a specific effect was unexpected. Later work, however, has thrown some light upon the matter, and it will be discussed under microchemical analysis. Oxygen pressure and leaf-fall Mo tiscu (14) grew plants half submersed in water and found that the aerial portions dropped their leaves sooner than the submersed portions. From this fact he concluded that low oxygen pressure retarded leaf-fall. This surmise proved to be correct. Coleus plants grown in a hydrogen atmosphere under bell jars with only sufficient oxygen to maintain a slow growth retain their leaves much longer than plants in normal air.. Under conditions of the experiment the plants in normal air usually retain 8 pairs of leaves. In the hydrogen atmosphere the plant retained 11 pairs of leaves. Inception of decay at the base of the stem destroyed the experiment at this point. Similarly, petiole fall, after amputation of the blade, is greatly retarded in very low concentrations of oxygen. Plants grown in o.1 normal oxygen pressure showed no retardation of leaf-fall. Whether the effect of oxygen in such cases is that of an essential factor influencing the general metabolism of the plant, or of a formative factor influencing directly the oxidase activity in the abscission layer, or of both acting simultaneously, is a problem still to be investigated. Likewise a critical investigation of the possi- bility of a double effect of carbon dioxide on leaf-fall might throw more light upon the causes underlying abscission. Macrochemical analysis In order to follow the chemical changes leading up to abscission, both macrochemical and microchemical methods of analysis were employed. About 2500 plants grown under greenhouse conditions were used for the macrochemical analysis. These plants were all grown at the same time under the same conditions, and collection of material was made at the same time each day. Series C was collected between February 17 and March 3, collections being taken from day to day as the lower leaves began to absciss. Series D and E were collected from these same plants on March 3 and 4. 42 BOTANICAL GAZETTE [JULY Material when collected was placed in 70-80 per cent alcohol and heated to 70°C. for one hour to destroy enzymatic activity. The material was then extracted with alcohol and ether. The residue was dried and analyzed for polysaccharides, calcium, and oxalates. The alcohol-ether extract was evaporated to dryness on a steam bath and then extracted with water at 70°C. The filtrate of this aqueous extract was analyzed for reducing and non-reducing sugars; ammonia, amino acid, and nitrate nitrogen; calcium in solution, and acidity. Table IV gives a summary of an analysis of 9 different regions of the plant. Series C represents leaves in the act of abscissing, series E represents leaves at, the time of the formation of the abscission layer, and series D represents leaves intermediate between these two points. Collection Ex represents approximately 5 mm. of the abscission end of the petiole, collection E2 an equal portion of the adjacent part of the petiole, and collection Ez a portion of the blades. In like manner, collections C1, C2, and C3 and collections D1, D2, and D3 represent these same three regions in their respective series. Attention should be called to the fact that while collections C1, Dr, and Ex represent the abscission end of the petiole, they do not represent the abscission layer only. In no case does the abscission layer represent more than about 5 or 6 per cent of the portion of the petiole taken. In collection Cr it represents still less, probably not more than 2 per cent, as in the abscissing leaf the petiole retains only about one-third of the abscission layer, the remaining two-thirds being attached to the stem. It is evident that chemical changes in the abscission layer might be overshadowed by the remaining 95 per cent of the collection, and even more so in collection Cr than in collections D1 and Et. This is especially true of the nitrates, which are frequently confined almost entirely to the abscission layer and are more abundant in this layer at the time of abscission than at any other time, although the figures in the table might lead one to think they were most abundant a short time before abscission. As a matter of fact, a large percentage of the nitrates in the abscission layer of collection Cr were left in the part of the abscission layer remaining attached SAMPSON—ABSCISSION 43 1918] Sgo0'o | zo-z | 6b'£ | z1°0 | 19°F |n€0'0 |zLo'o |bz0-o S€-b | gS-o lgg‘or izg'S1 | 6b'L lbE-Lz oa Or | 00'S [+t -- oe eas S$00'0 | oL-1 | $g-z | 1z'0 | go'€ |oS6'0 g40°0 |gf0°0 | og't | oo [zh or (zo'S1 | €S°¢ |16-z¢E Wr Oe | EE paces ee 6900°0 | Lo-z | oz‘z | z1'0 | z£-z lolo'1 Ogo'o jofo’o | gz*z | o°o |S6-or |1z‘°E1 | obo mo S¥ of Se | C€°¥ |: 2s: ey OOI0'O | L¥'I | Eb'z | G0'0 | zS-z |LE0'0 gSo'o |600'0 | Lz‘t | 0:0 |oL-61 \oo' bz Q¥'g joo-oz |PS-41 | bL'g |-°-::- ‘ett oe gsoo"o | 1€'1 | Gof | €€:0 | zh logr-o |LS0'0 |zlo-o | co-z OO /fo'c1 [fo'Fr |*****- PS Ee he ] Geek fete! eee fL00'0 | 1b | of-z | LE'0 | Lo'€ lozg-z |L60'0 glo°o | bl: | obo |gg11 |zo'br | $S-b l6e-S¢ ag GF | 19:8 +: ade L$00'0 | PE'1 | z1°z | go'0 | gr°z lobr-o gS0°o jozo'o | zg-1 | E10 |60°Sr |bo'L1 | 6:6 gL-ez |\€z°€1 | Svs eG 6L00°0 | Loz | c0°€ | oz‘0 | oz'€ Izto-o |€fo'0 grvo'o | o6'0 | ovo |€€-z1 fz Er gz '€ vo zl lol-gz | S6°€ |----:- 1 See ggoo'o | 46°1 | 6g-z | L1‘0 | go'€ |o06-0 gor'o |€90°0 | Sho | o'o |hb 11 |6g' 11 fg°z |SS-1€ |26-gz | b6-€ |--+- ++: Sees 3 fe) e) fe) 4 Zz > 2 bd Z a) 4 > > < a) agas|G82) Fr | 85] &€ | S| ek | B | ef | og | g SE | gge| 88 | & | 38 anpe | Ege | ge | Fe a B | 88 Bp | £6. a | & fo | ass] 28 5 Ee EGE = Cah | Fs 58 8 g Rw 5 BF 8 28 oe) 6s g | "eeg/B28/| 82] °F | B | € | SE | 2 | o* | Be} E | oe jeha| Pa] BY AS Zee | ae 5 5 5 g & ae By rte a a. = a NOLLOITIOD ghx & g o S Pant od pecechinmiasin ini - « na” . 8 LTHOIEM AY TVLOL 10 ADVINAOUAd SV dassaadx |] Aud Prices ean Oe d AI ATAV.L 44 BOTANICAL GAZETTE [JULY to the stem, and therefore are not included in the analyses. The data in the table, therefore, represent only the general chemical changes during the life of the leaves, while the detailed chemical changes occurring in the abscission layer itself will be given under microchemical analysis. Should one desire to make a macrochem- ical analysis of the abscission layers alone in Coleus no less than 40,000 plants would be needed. The data in table IV show an increase in dry weight in the abscission end of the petiole at the time of abscission; also an increase in alcohol-ether soluble material, but no increase in water soluble material. The significance of these changes is uncertain. In the older petioles there is a slight decrease in polysaccharides and an increase in reducing substances. There is no increase in ammonia and amino acids, as might be expected if the protoplasm were breaking down. Oxalates and total calcium remain fairly constant, but there is a slight decrease in the amount of calcium in solution and in the acidity. A more detailed discussion of the chemical changes in the abscission layer is given under micro- chemical analysis. In the older blades there is a decided decrease in the amount of accumulated starch at the time of abscission, but the amount of reducing substances remains fairly constant. Attention has already been called to the fact that the formation of the abscission layer is completed while the leaf is still in an active photosynthetic condition. Both photosynthesis and the translocation of foods continue for several days or weeks later. The data clearly show that the presence of the abscission layer does not prevent the movement of water and foods between leaf and stem. Microchemical analysis A microchemical investigation of Coleus showed a striking localization of physical and chemical changes in the abscission layer shortly before and at the time of abscission. The formation of the abscission layer usually in the third pair of leaves and the occurrence of abscission usually in the eighth pair of leaves (when the plants are grown in 4-inch pots in a greenhouse) make it possible to study the whole history of ‘the abscission layer by investigation of only 6 1918] SAMPSON—ABSCISSION 45 pairs of leaves in each plant. The fact that the leaves are opposite is also of advantage. Abscission of the pair may occur simul- taneously, or one of the pair may absciss long before the other begins, or it may occur at any stage in between. Since both abscission layers at each node may readily be obtained in a single free-hand section, it is possible to contrast all stages of abscission under exactly the same treatment. A study of the changes induced by forcing abscission in one of the leaves at each node is likewise facilitated. The investigations completed include a study of the distribution and amount of nitrates, carbohydrates, oxidases, iron, manganese, calcium in solution, and oxalates. NITRATES.—The data in table IV show a great increase of nitrates in the abscission end of the petiole as compared with the remainder of the leaf. Furthermore, the nitrates in this part of the petiole are least abundant at the time of formation of the abscission layer and most abundant a short time before leaf-fall. As already noted, these figures cannot be taken to represent the percentage of nitrates in the abscission layer. Microchemical tests show some interesting variations. In many plants the increase in nitrates is confined almost entirely to the abscission layer, while in others the petiole or the neighboring part of the stem may also show a like increase. In all cases studied there is an increase in nitrates in the abscission layer just before and at the time of abscission. In some plants this increase is gradual from the time of the formation of the abscission layer to the time of abscission. In other cases only traces of nitrates appear in the abscission layer until a short time before abscission, when they increase rather suddenly. CARBOHYDRATES.—The data in table IV show that the free reducing sugars, like the nitrates, increase in the abscission end of thé petiole with the increase in the age of the leaves, but, unlike the nitrates, they are less abundant in this part of the petiole than in the remainder of the leaf. This correlation of the amount of reducing sugars and the age of the tissue is still more striking when studied microchemically. From the terminal bud to the oldest leaves there is a gradual increase in reducing sugars in both stems and leaves. This increase is initiated last in the abscission layer. 46 BOTANICAL GAZETTE [JULY As a result the abscission layer has a lower percentage of reducing sugars throughout its entire history than the adjacent regions of the petiole. This difference is most marked in the fifth and sixth pairs of leaves near the close of the formation of the abscission layer, but it.is still quite evident at the time of abscission. In the cell walls of the abscission layer the change in form of the carbohydrates is still more pronounced and significant. During the process of abscission the first evident change in the cell walls is a conversion of cellulose of the secondary cell membranes to pectose. The second step is a conversion of some of this pectose to pectic acid and pectin. This is followed by the breaking down of the middle lamella of calcium pectate and the separation of the cells. The changes from cellulose to pectose can readily be followed by differential staining and crystallization methods, and by solubility tests. The evidence of the conversion of pectose to pectin and pectic acid is based upon solubility tests. Pectin is soluble in water, pectic acid is insoluble in water but soluble in dilute alkalies, while pectose is insoluble in both water and dilute alkalies. When an abscission layer at the time of abscission is treated with 3 per cent ammonium or potassium hydroxide or with 5 per cent sodium carbonate, the free pectic acid is dissolved. If the walls are then again examined a considerable portion of the secondary membrane, bordering the middle lamella which is still intact, is seen to have disappeared. A discussion of the changes in the calcium pectate is postponed until all the remaining facts have been stated. OxipAsEs.—In the stem and petioles oxidases are found in the epidermal and phloem tissues. In the abscission layer oxidases are found in all tissues except the xylem. Not only is this distri- bution peculiar to this region, but the increase in oxidases with the age of the abscission layer is also quite pronounced. Quanti- tative tests of the increase in oxidative activity are still to be made. Iron.—Slight traces of iron (Fe++*) are usually found through- out the plant, especially where chlorophyll is present. It is most abundant in the xylem tubes and in the epidermal region until a few hours before abscission, when it becomes extraordinarily abun- dant in the cells of the abscission layer. The path of diffusion of 1918] SAMPSON—ABSCISSION 47 the iron leading to its accumulation in the abscission layer has not been traced. No manganese was found. CALCIUM AND OXALATES.—The data in table IV show no marked difference in the distribution of oxalates in the leaves. Only occa- sional crystals of calcium oxalate are found in the cells, and none are found in the cell walls of the abscission layer at any time. Likewise the total calcium has a fairly constant distribution throughout the plant, but slight variations of the calcium in solu- tion are to be noted. The most striking and significant changes of the amount of calcium in solution, however, are shown by microchemical tests. Treatment of sections with 50 per cent - sulphuric acid or with 3 per cent oxalic acid or ammonium oxalate show an abundance of calcium in solution in all living cells of the petiole except those of the abscission layer at the time of abscission. The crystals of calcium sulphate or of calcium oxalate obtained by these treatments were very numerous in the cells of the abscission layer before the time of abscission, the latter averaging 30 crystals per cell, while during abscission only an occasional crystal was obtained. This decrease of calcium in solution is not always confined to the abscission layer, but breaks off rather abruptly in the first few layers of cells of the adjacent region of the petiole. In some cases cells not more than 5 cell layers distant from the line of cleavage showed no decrease in the number of crystals. These facts show that the calcium in solution in the abscission layer dis- appears during abscission, and it should be further stated that the disappearance takes place in the- first stages of the process. Summary of microchemical analysis 1. A pronounced increase in nitrates always occurs in the abscission layer at the time of abscission. This increase may be gradual, extending over the entire life-history of the abscission layer, or it may appear somewhat suddenly a short time before abscission. 2. A gradual increase in the amount of reducing sugars accom- panies the aging of leaves and stem. This increase is initiated last and is least pronounced in the abscission layer. 3. During the process of abscission the cellulose of the secondary membrane of the cell walls of the abscission layer is converted into 48 BOTANICAL GAZETTE [JULY pectose. This pectose is further transformed into pectic acid and pectin. The final stage is the breaking down of the calcium pectate of the middle lamella. 4. Oxidases are present in the epidermal and phloem tissues in both stems and petioles. In the abscission layer they are present in all tissues outside of the xylem, and increase in amount with the age of the abscission layer. 5. Slight traces of iron may be found in practically all parts of stem and petioles, but shortly before abscission there is a sudden CIO of iron in the abscission layer. . The amount of oxalates remains fairly constant throughout the ae life of the leaves. There is no evidence of an increase of calcium oxalate crystals in the cells of the abscission layer at the time of abscission, nor are there any crystals of calcium oxalate in the walls of these cells. 7. Calcium in solution is abundant in all living cells of the plant except those of the abscission layer at the time of abscission, where it practically disappears. Discussion According to ToLLENsS (16), pectose is an oxidized cellulose of the composition 9(CsH:0O;)—CcH:.Os. The first step in the breaking down of the cell walls in abscission in Coleus is evidently one of oxidation of cellulose. This process is possibly a result of the accumulation and subsequent activity of oxidases in the abscission layer, and also of the catalytic action of iron on these oxidases. This may be merely an acceleration of the conversion of cellulose into pectose which ordinarily goes on in cell walls of © plants with increasing age. Cellulases may play a part in this process, but the question is still to be investigated. Ever (4) succeeded in isolating a cellulase in a fungus, Merulius lacrimans, which was capable of altering cellulose, but cellulases in higher plants are still unknown. CzAPEK (2) and EuLER (3) have called attention to the fact that our knowledge of cellulases is very limited. The pectose formed from the cellulose is in turn readily trans- formed to pectic acid and pectin, and in this process the catalytic action of iron may again play an important réle. Whether acids 1918] SAMPSON—ABSCISSION 49 and pectic enzymes also play a réle in these changes is uncertain. Hydrolytic action may underlie some or all of these changes, but this must remain an open question until the molecular composition of these compounds is definitely known. At any rate, the trans- formation of the cellulose and pectose leads to the formation of an excess of pectic acid in the cell walls of the abscission layer. The most important question still open is the cause of the final breaking down of the calcium pectate of the middle lamella. There appear to be but two possibilities. Either the calcium ion of the pectate is captured by some anion, liberated in the cells of the abscission layer, and held in solution or precipitated, thus freeing the pectic acid, or the breaking down of the cellulose and pectose may lead to such an excess of pectic acid that the available calcium is no longer able to hold a sufficient proportion of the pectic acid as a salt, and thus maintain the solidity of the middle portion of the cell wall. The fact that calcium is not found in solution in the cells of the abscission layer, nor in crystalline forms either in the cells or in the cell walls, is decidedly against the first view, which is simply WIESNER’s organic acid theory stated in slightly diferent terms and which has already been discussed in detail. _ The second view is more easily understood when we recall the well known law of physical-chemical equilibrium. As soon as an excess of pectic acid is present in contact with the calcium pectate of the middle lamella there is undoubtedly a diffusion of calcium ions from the middle lamella and a diffusion of pectic acid into the middle lamella until an equilibrium of distribution of the two ions is established. A critical proportion of pectic acid to calcium would be reached in the middle lamella when the excess pectic acid breaks the continuity of the calcium pectate layer. This second view has the further advantage of being in accordance with all the experi- mental facts so far known, particularly the formation of excess pectic acid in the cell walls and the paucity of calcium in solution in the cells of the abscission layer. The fact already stated, in the discussion of calcium, that there is an abundance of calcium in solution in cells within 5 cell layers of the line of cleavage in abscission, indicates either that the process 5° BOTANICAL GAZETTE [JULY of abscission is a very rapid one or that the diffusion of calcium from cell to cell is a very slow one. This fact also explains why the addi- tion of calcium salts, already discussed, showed no specific effects on the rate of leaf-fall as the rate of diffusion of the calcium ions through the cells would again appear as a limiting factor. Extensive comparative investigations of abscission in the light of facts discovered in Coleus are still to be made. ‘Tison’s state- ment that in general the secondary membranes also are altered in abscission indicates that these processes may be rather general, particularly since the reports of HANNIG and Lioyp of a similar alteration in the abscission of floral organs. Investigation of the more fundamental factors underlying the ultimate chemical changes discussed in this paper must be made before general conclusions of the causes leading up to abscission can be drawn. The significance of the presence of an abundance of nitrates in the abscission layer at the time of abscission is uncertain. Their ability to affect the water holding capacity of colloids and similar effects of other ions which are changing in concentration in this region may influence the permeability of the cell membranes of these cells, a question that has not yet been touched upon experimentally. Conclusion Abscission of leaves in Coleus Blume? is a result of the conversion of cellulose into pectose, which is further transformed to pectin and pectic acid, leading to the formation of an excess amount of pectic acid over that of the available calcium sufficient to maintain the solidity of the middle lamella of the cell walls of the abscission layer. These processes are possibly initiated and probably accelerated by the presence of oxidases and ferric ions, both of which accumulate in the abscission layer. Microchemical methods employed In the microchemical study color reactions were used for orienta- tion. ‘These were followed by specific chemical reactions and solu- bility tests. A brief outline of the tests made for each substance follows. Details of these reactions may be found in recent micro- chemical texts. 1918] SAM PSON—ABSCISSION oe 51 CELLULOSE.—(1) Chlorzinc iodide: blue color; (2) hydro- cellulose reaction: blue color with iodine after treatment with 75 per cent sulphuric acid; (3) solubility: insoluble in dilute acids and alkalies, soluble in copper-oxide-ammonia; (4) crystallization: dissolve in copper-oxide-ammonia, wash with ammonia and water; colorless sphaero crystals or spiculate crystal clusters appear within the cells; (5) crystal reactions: insoluble in dilute acids and alkalies, soluble in copper-oxide-ammonia and sulphuric acid; blue color with chlorzinc iodide; (6) membranes of cellulose exhibit double refraction in polarized light. PECTIC COMPOUNDS IN GENERAL.—(r1) Ruthenium red: red color; (2) methylene blue: violet color; (3) membranes of pectic compounds do not exhibit double refraction in polarized light. PEcTOSE.—(1) Insoluble in copper-oxide-ammonia, dilute alkalies, ammonia, and alkali carbonates; (2) converted into pectic acid and pectin when gently heated with 2 per cent hydrochloric acid for 30 minutes. These latter substances are readily dissolved by 2 per cent potassium hydroxide or 5 per cent sodium carbonate, leaving the cellulose membrane intact. Prectic acim.—(1) Soluble in dilute alkalies, ammonia, and alkali carbonates; (2) insoluble in water. Prectin.—Soluble in water. CALCIUM PECTATE.—(1) Hydrolyzed by 2 per cent hydrochloric acid: calcium chloride is formed and pectic acid set free; (2) 3 per cent oxalic acid or ammonium oxalate: calcium oxalate crystals are formed, pectic acid set free; (3) 5 per cent sulphuric acid: calcium sulphate crystals formed, pectic acid set free. Catcrum.—(1) Two per cent oxalic acid: calcium oxalate crystals; (2) 5 per cent sulphuric acid: calcium sulphate crystals. Licnin.—Phloroglucin-HCl reaction: red violet color. SUBERIN.—(1) Sudan III or Scharlach R: red color; (2) insoluble in copper-oxide-ammonia; (3) phellonic acid reaction. Fructose.—(r) Fluckiger’s reaction: yellowish-red precipitate of cuprous oxide at once without heating; (2) phenylhydrazine reaction: yellow osazone crystals formed in 6-8 hours; (3) methyl- phenylhydrazine reaction: insoluble osazone; crystals formed in 15 minutes if preparation is heated, after 24 hours at room temperature. 52 ee BOTANICAL GAZETTE [juLY GiucosE.—(1) Fluckiger’s reaction: yellowish-red precipitate of cuprous oxide after heating 1-2 minutes; (2) phenylhydrazine reaction: yellow osazone crystals formed after about 24 hours. SucROSE.—Remove fructose and glucose. Invert with hydro- chloric acid. ‘Test for glucose and fructose as preceding. NitTRATE.—(1) Diphenylamine sulphuric acid reaction: blue color slowly changing to brown-yellow; (2) brucin-sulphuric acid _ reaction: red color. OxIDASES.—Benzedine reaction: blue or purple precipitate if tissue is acid, soon changing to brown; brown precipitate at once if tissue is neutral or alkaline. Iron.—(r1) Berlin blue reaction: sections in 2 per cent solution of potassium ferrocyanide 15 minutes, add a drop of 2 per cent hydrochloric acid. A dark blue precipitate indicates the presence of ferric ions. Similarly a red color with potassium ferricyanide indicates the presence of ferrous ions; (2) sodium thiosulphate: red color. MANGANESE.—Sections in 0.1 per cent hydrochloric acid, add 0.5 per cent sodium ammonium phosphate and ammonia vapor: ammonium manganese phosphate crystals, brown color in a 2 per cent solution of potassium permanganate. Matic Acip.—(1) Silver nitrate: sphaero crystals of silver nitrate, soluble in ammonia; (2) lead oxide: lead malate crystals; (3) sublimation: concentrated sulphuric acid, heat to 130°C.; slight charring. OXxALIc AciD.—(1) Uranium acetate: large yellow crystals of uranium oxalate; (2) strontium nitrate: strontium oxalate crystals; (3) ferrous phosphate: yellow precipitate of ferrous oxalate. AMINO Acips.—Crystallization: treat sections with absolute alcohol, crystals of amino acids; (1) compare with known crystal form; (2) specific reactions. TyROSINE.—Millon reaction: red color. ARGININE, HISTIDINE.—Picrolonic acid: yellow crystalline precipitate. Lreuctne.—Sublimation at 170°C. » ASPARAGINE, GLUTAMINE.—Quinone: red color. 1918] SAMPSON—ABSCISSION 53 The writer wishes to express his gratitude to Dr. WiLLIAM CROCKER and Dr. Sopuia H. Eckerson for numerous suggestions and criticisms during the progress of the work. 3 w UNIVERSITY OF OHIO CoLumsBus, OHIO LITERATURE CITED . Cross, C. F., Uber die constitution der Pectinstoffe. Ber. Deutsch. Bot. Gesells. 28:2609-2611. 1905. CzaPEK, F., Biochemie der Pflanzen (zweite Auflage), pp. 40, 371-375. . EuLer, H., Pflanzenchemie. 1908. Bar Kenntnis der Cellulase. Ztsch. Angewandt. Chem. 25:250- 251. 1912 . Frrrine, H. , Untersuchungen iiber die vorzeitge enblatterung von Bliiten. Jahrb. Wiss. Bot. 49:187-263. 1911. Hannie, E. Lpiagenypine iiber das abstossen von Bliiten u.s.w. Zeit- schr. Bot. 5:417-469. 1913. LEE, E., The OME of leaf-fall. Ann. Botany 25:51-106. 1911. Lioyp, F. E., Abscission. Ottawa Naturalist 28:41-52, 61-75. 1914. Abscission in Mirabilis Jalapa. Bot. Gaz. 61:213-230. 1916. ———, The abscission of flower buds and fruits in Gossypium and its relation to environmental changes. Trans. Roy. Soc. Canada 10:55-6r. Ig16. - MAncIN, M., Sur la constitution de la membrane des végétaux. Compt. 18 Rend. Acad. $c. 107:144-146. , Etude historique et Shihan s sur la presence des composés pectiques dans les tissues des végétaux. Jour. Botanique 6:12-19. 1892. - Mout, H. von, Uber des Ablésungsprozess saftien Pflanzenorgane. Bot. Zeit. 18:273-274. 1860. - MoriscH, Hans, Untersuchungen iiber Laubfall. Bot. Centralbl. 25: 393-394. 1886 . Tison, A., Recherches sur la chute des feuilles chez les Dicotyledonées. Mem. Sine: Linn. Normandie 20:121-327. 1900. Tot.ens, B., minal rapes iiber pisommeaess Liebig’s Annalen der Chemie 286: 278-292. - WIESNER, J., Seiden. iiber die herbstlische — der « Holzgewiichse. Litz. Akad. Wissensch. Wien 64:465-50 ———, Uber Sommerlaubfall. Bot. Gessells. 22:64-72. get , Trieblaubfall. Bot. Gesells. 22:316—323. 1904. - ———., Hitzelaubfall. Bot. Gesells. 22: 501-505. 1904. , Frostlaubfall. Bot. — 23:49-60. ees INTERRELATIONSHIPS OF THE TAXINEAE Mary C. Buiiss (WITH PLATES I, II) In considering the Taxineae it is interesting to note the taxo- nomic position to which this subtribe has been assigned at various periods in the history of the classification of the conifers. ENGLER and PRANTL (4) in 1889 placed it at the top of the group; PEN- HALLOW (6) in 1907 placed it at the bottom of the group; COULTER and CHAMBERLAIN (1) in rgor regarded the subtribe as the most . primitive of the conifers and placed it at the bottom, but in 1910 (2) shifted its position to the top of the group as the most modern. These facts show clearly that the family is a difficult one to inter- pret, and the difficulty is due in part to the fact that the Taxineae combine at the same time extreme simplification and specialization. The argument presented by PENHALLOW as evidence for his theory that the Taxineae are the most primitive of the conifers is based on the progressive development of the resin canals in Pinus and Picea from the isolated resin cells of Podocarpus “by various phases of aggregation.” In Taxus and Torreya of the Taxineae, which he investigated, PENHALLOW states that resin cells are entirely wanting. Isolated resin cells occur in abundance in Podocarpus of the Taxineae. In the true Coniferae isolated or aggregated resin cells are characteristic of all the genera except Picea and Pinus, where they are replaced by resin passages, of which the aggregations of resin cells form an essential part. From the genera Taxus and Torreya, characterized by the absence of resin cells, PENHALLOW traces a series through Podocarpus, where resin cells are scattered, to genera of the Coniferae, where first, as in Taxodium and Libocedrus, the resin cells are arranged in well defined zones as well as scattered, to resin sacs in Abies and Sequoia, to resin passages with constrictions in the canal in Larix, Pseudo- tsuga, and Picea, to the resin passages without constrictions, as in inus. Botanical Gazette, vol. 66] [54 1918] BLISS—TAXINEAE 55 Those who hold that the Taxineae represent a modern group in the evolution of the conifers interpret the facts already stated in PENHALLOW’S argument as evidence of an entirely different progression. Starting with the genera in which the resin canal is highly specialized and resin cells wholly lacking, as in Pinus, they trace a series in which there is a gradual reduction of the resin canal, to aggregations of resin cells as in Taxodium, to scattered resin cells as in Sequoia and Podocarpus, to entire absence of resin cells as in Taxus. As evidence of the fact that resin canals represent a primitive condition in the conifers, JEFFREY’s work on the genus Sequoia may be cited. If the presence of resin canals were evidence of modern development, we should expect to find them in the mature and more progressive parts of the plant, but in Sequoia gigantea JEFFREY (5) found the resin canals only in the first annual ring in the stem, in the ovulate strobilus, and in the leaf traces of very vigorous leaves of adult trees. In Sequoia sempervirens the resin canals were wholly lacking in these regions, but in injured stems and roots of both S. gigantea and S. sempervirens resin canals were present. In the case of Sequoia, then, the presence of resin canals represents a primitive condition in the conifers, retained only in the more con- servative regions of the plant in S. gigantea, and wholly absent in S. sempervirens. A reversion to the ancestral condition in both species may be induced by injury. According to this later view of the position of the Taxineae as contrasted with that held by PENHALLOW, we have a series of genera starting with Pinus as a representative of the most primitive group in which resin canals are normally present, proceeding through Seguoia as a type of a group in which resin canals are not normally present in the vegetative axis, until we come to Podocarpus, a representative of a group in which resin canals are never present. In this connection it is important to note that in those groups in which resin canals are normally absent the secretion of resin is carried on by resin parenchyma cells. These resin cells are char- acteristic of the Taxodineae, Cupressineae, and Podocarpineae. We should expect as the logical outcome of this gradual reduction and simplification of resin secreting structures the final passing out 56 BOTANICAL GAZETTE [yULY of the resin parenchyma cells in the more modern types. This question will be taken up later when we consider the genera within the group Taxineae. Much of the controversy in regard to the position of the Taxineae has been based on the character of the gametophytes and reproduc- tive structures, especially the ovulate cone and the method of development of the proembryo. The evidence which I have to offer is derived wholly from the study of the anatomical structure of the stem and root of various genera in the group. There are included in the family Taxaceae of ENGLER and PRANTL (4) the genera Phyllocladus, Ginkgo, Cephalotaxus, Torreya, and Taxus. In the most recent classification of the group by CouLTER and CHAMBERLAIN (2) Phyllocladus is included in the Podocarpineae, Ginkgo has been put in a family by itself, and the Taxineae include in addition to Taxus, Torreya, and Cephalo- taxus, the doubtful New Caledonian genera Acmopyle and Poly- podiopsis. Turning our attention to the 3 accepted genera of the group, Cephalotaxus, Torreya, and Taxus, I shall attempt to show that the Taxineae are the most modern group of the conifers, that Cephalotaxus is the most primitive genus of the subtribe and most nearly related to the Podocarpineae, that Torreya is intermediate, and that Taxus is the most modern genus of the family and repre- sents, so to speak, the last word in the evolution of the conifers. If we examine a transverse section of the stem of Podocarpus totara, we note the presence in great abundance of resin parenchyma cells (fig. 1). These parenchyma cells are even more evident in the longitudinal section of the stem (fig. 2) as cells which stain densely © with haematoxylin due to the presence of resin. These cells are narrower than the tracheids and are characterized by thin walls, by — the absence of pits, and end walls at right angles to the long axis of the cell. A section of the stem of Cephalotaxus drupacea (fig. 3) presents a very similar appearance to the stem of Podocarpus. That the resin parenchyma cells are widely distributed throughout the annual rings of the stem is evident from a consideration of the low power photograph (fig. 4). The location of these cells is shown by the deep staining of the resinous contents. 1918] BLISS—TAXINEAE 57 The root of Cephalotaxus drupacea (fig. 5) shows the presence of resin parenchyma in even greater abundance, and this is the con- dition we should expect to find, since the root is the more conserva- tive organ of the plant and would retain more fully the primitive or ancestral characteristics of the plant. . The stem of Torreya taxifolia presents a very different appear- ance from the stem of Podocarpus and Cephalotaxus already con- sidered. Resin parenchyma cells are present throughout the annual ring, but they are much less abundant than in the other stems. The distribution of the cells may be seen in the transverse section (fig. 6), and the character of the cells is shown very clearly in the longitudinal section (fig. 4). As previously stated, PENHALLOw did not find resin cells in any of the species of Taxus or Torreya which he investigated, and DEBary (3) also states that all investigated species of the Coniferae, with the single exception of Taxus, have resin passages or resin reservoirs. Asa result of my investigation, it is clearly evident that resin parenchyma is present in Torreya taxifolia, one of the species investigated by PENHALLOW. If we examine a transverse section of the stem of Taxus brevi- folia (fig. 8), we note the complete absence of resin parenchyma cells. A longitudinal section of the same stem (fig. 9) shows even more clearly that the vascular cylinder consists simply of thick-walled tracheids, with numerous bordered pits, and the characteristic spiral thickenings. So far then the condition in Taxus tallies with the investigations of PENHALLOW and DeBary; but if we examine a transverse section of the root of Taxus cuspidata (fig. 10) we note the presence of resin parenchyma diffused throughout the annual ring. A higher magnification of a portion of the root is shown in fig. 11. Here the resin parenchyma cells are very conspicuous. A longitudinal view of the same root also shows a view of these paren- chyma cells very clearly (fig. 12). The root of Taxus baccata also shows the presence of resin parenchyma diffused throughout the annual ring. Although in the normal stem of the species of Taxus investigated there was no resin parenchyma present, a wounded stem of T. bac- cata showed very clearly an extreme development of these cells. 58 BOTANICAL GAZETTE [JULY The location of the cells can be determined easily in the transverse section of the stem, due to the fact that the walls stain a deep blue with haematoxylin, and in the longitudinal section the characteris- tic shape of the parenchyma cells as contrasted with the tracheids make them easily recognizable. We find then in the normal stem of Taxus the condition which we should expect as the ultimate result of the gradual reduction of resin canals, namely, resiniferous parenchyma which finally com- pletely disappears except in the case of conservative organs. There are three important principles of evolution which have to be considered in working out the ancestry of any group of plants, namely, the principles of recapitulation in the development of the embryo and seedling stages of the plant; retention of ancestral characters in the more conservative regions of the plant, as the root, leaf, and reproductive axis; and reversion to ancestral conditions through injury. The first of these principles I have not been able to demonstrate, as I did not have access to the seedling stages of the genera investi- gated. The principle of the retention of ancestral characters in the most conservative organ of the plant is very clearly evidenced in the root of Taxus cuspidata by the presence of resin parenchyma cells which have entirely passed out of the stem; and finally the presence in abundance of resin parenchyma in the wounded stem of Taxus baccata seems to show clearly that we have in this instance a rever- sion to the ancestral condition. From a consideration of these facts the evidence seems to justify the conclusion that the Taxineae are the most modern group of conifers; that of the Taxineae, Cephalotaxus is the most primitive, in most nearly resembling Podocarpus in the abundance of resin parenchyma; that Torreya is the intermediate genus in the group, as shown by the reduction of resin parenchyma, especially in the stem; and that Taxus is the most modern genus in the group, since we find here entire absence of resin parenchyma in the stem, although it is retained in the root. Summary 1. Resin parenchyma is present in abundance in the stem and root of Cephalotaxus drupacea and shows clearly its close relationship 1918] BLISS—TAXINEAE 59 to the Podocarpineae, a family in which resin parenchyma is uni- versal. 2. Resin parenchyma is present in less abundance in the stem of Torreya taxifolia, showing in this respect an intermediate position between Cephalotaxus and Taxus. 3. Resin parenchyma is wholly absent in the normal stem of Taxus brevifolia, showing that this genus is the most modern one of the group. 4. Resin parenchyma in the root of Taxus cuspidata and T. bac- cata and in the wounded stem of T. baccata indicates the ancestral condition in this genus. 5. The Taxineae represent a modern group of conifers, as shown by the gradual reduction and final passing out of resin parenchyma in the more progressive organs. This investigation was carried on in the laboratories of Plant Morphology at Harvard University under the direction of Dr. E. C. JEFFREY, and I wish to express my thanks to him for his invaluable aid in the work and for the many courtesies extended to me during the year spent in his laboratory. Harvarp University CAMBRIDGE, Mass. LITERATURE CITED 1. COULTER, J. M., and CHAMBERLAIN, C. J., Morphology of the spermato- phytes. roor. eae oa a iOS MBER 3910: 3. DeBary, A., Pp g dferns. Oxford. 1887. 4. Enctrr, A., and Prantt, K., Natiirliche Pflanzenfamilien. 5- JEFFREY, E. C., The cuiaparative anatomy and phylogeny ‘i oe Conifer- ales. I. The genus Sequoia. Mem. Boston Soc. Nat. Hist. 1903. 6. PENHALLow, D. P., North American gymnosperms. 1907. EXPLANATION OF PLATES I, II PLATEI Fic. 1.—Transverse section of wood of stem of Podocarpus totara, X 250. Fic. 2.—Longitudinal radial section of wood of stem of same, X 250. Fic. 3.—Transverse section of wood of stem we Cephalotaxus drupacea, X 250. BOTANICAL GAZETTE [JULY 4.—Same as fig. 3, X40 5.—Transverse section a wood of root of Cephalotaxus drupacea, . 6.—Transverse section of wood of stem of Torreya taxifolia, X 250. PLATE II . 7-—Longitudinal radial section of wood of stem of Torreya taxifolia, . 8.—Transverse section of wood of stem of Taxus brevifolia, X 250. 9.—Longitudinal radial section of wood of stem of Taxus brevifolia, . 10.—Transverse section of wood of root of Taxus cuspidata, X30. 11.—Upper portion of same, more highly magnified, 12 25. . 12.—Longitudinal radial section of wood of root of Taxus cus pidata, PLATE I LXVI BOTANICAL GAZETTE, fea — a ¥ ‘ eX e fo i , ‘ ' — 1} ey pe@aatangee swan? OUR mee Oe RUM RD ciesiinipiis a jeeewanee’*)) ameasiiva Ni ~ bee bi hd o' 22 ems ® Aitys — ¥ an ge Se Sist ote @ , 08 | Po BLISS on TAXINEAE PLATE Il LXVI AL GAZETTE, BOTANIC D ef ve es, o8) g8i te et 8 8, (? & oe sx) 2 ve ae eo, & v¢ - “e ton = : Mi penetebenaitets se it ie BLISS on TAXINEAE pangs SSAA ANNI bl, 6 EPS: S71 C= ae ee ——— en | Ta he Pra ae: ef * hei ‘ 4 3 eats er fie pie . : if “ a + ieee srupy FS + alee SIGNIFICANCE OF RESINOUS TRACHEIDS SAMUEL J. RECORD (WITH FIVE FIGURES) The occurrence of resinous tracheids in gymnosperms has been noted by PENHALLOW’ in the woods of certain species of Cordaites, Araucaria, Dammara, and a few representatives of the higher Coniferales, namely, Pinus albicaulis, P. parviflora, Abies Fraseri, and A. grandis. Such tracheids do not differ structurally from other tracheids, but are distinguished by their resinous contents. The resin within them is localized and usually extends across the cavity to form an imperforate septum or plate, which, in unstained sections, may give the cell the appearance of being structurally septate. PENHALLOW figures the common form of these plates in Dammara australis. Figs. 1-3 illustrate the resinous tracheids in the wood of Pinus albicaulis, showing the characteristic form of the resin masses (RPI) and their association with the medullary rays. The resinous contents of the latter are omitted in order not to obscure the structure. It will be noted that there is considerable difference in the thickness of these plates, which are invariably thinnest in the middle, and not infrequently ruptured there. By comparison with PENHALLOw’s drawings it will be seen that the location, form, and distribution of the resin masses in the two species are identical. The close association of the rays with the resin plates in the tracheids clearly indicates the origin of the resin, which in some cases can be seen in the form of globules on the outside of the pit membrane of the parenchyma cell. When enough has exuded to form contact with the opposite wall of the tracheid, the surface tension of the liquid and the attraction of the cell wall cause it to assume a double concave form like a drop of water in a small glass * PENHALLOW, D. P., A Manual of the North American gymnosperms. Boston: Ginn & Co. 1907 (pp. 53-58)- 61] [Botanical Gazette, vol. 66 - andit is difficult to suggest an ad 62 | BOTANICAL GAZETTE [yULy tube. Sometimes there is enough resin to fill the tracheid for a considerable portion of its length, at others only enough to produce very delicate plates, while in the case of a cell with a wide lumen the resin may run down along one side or collect in masses without connection across the cavity. It is not unlikely that the presence of gas bubbles in the cells at this stage may play a part in forming the plates and in determining their location, since in some instances a thin plate may be found at a considerable distance from the ‘ray and without-visible connec- tion with it. What is the significance of these plates? PENHALLOW says as follows: The peculiar form in which the resin is deposited and the particular location of the plates point with much force to their connection with some functional activity, since if it were simply a ques- tion of the storage of secreted prod- ucts, the latter would hardly be disposed as found, but rather after the manner common to so many of the Cupressineae; and this suggestion gains strength from the fact that with respect to the peculiar form of the resin masses as well as their location in the tissue, the Cordaitales are peculiar among the ymnosperms. No exact comparison can be established with other plants, 4 juate expl i ne thing does seem clear, however, and that is that since these plates are of an impervious nature and developed in some cases, at least, in connection with a special constriction of the tracheid cavity, they offer and possibly are specially designed to afford a definite obstruction to circulation in a vertical direction. In this sense they may be designed to serve the same general purpose that is accomplished by the develop- ment of tyloses in the vessels of the angiosperms or in the resin passages of the higher Coniferales. It is possible, therefore, that they may be connected in some way, not at present clear, with a more complete restriction of the circula- tion to a horizontal direction, and particularly through the medium of the medullary rays as specialized channels for that purpose. Among existing RP Fic. 1.—Transverse section of Pinus albicaulis. 1918] RECORD—RESINOUS TRACHEIDS 63 gymnosperms resinous tracheids are almost exclusively confined to Dammara and Araucaria, though it is a noteworthy fact that similar structures occur rarely among the higher Coniferales..... The taxonomic value of the Fic. 2 API. Fics. 2, 3.—Fig. 2, radial section; fig. 3, tangential section of Pinus albicaulis resinous tracheids applies exclusively to the Cordaitales, where they are of ordinal value, though in Dammara and Araucaria they may also become ot specific value. 64 BOTANICAL GAZETTE [JULY The “peculiar form” and the ‘‘ particular location,” upon which PENHALLOW laid special stress, are readily understood when their origin is appreciated. The resin is produced in the parenchyma as a product, or more likely as a by-product, of the metabolic activity of those cells. The cavities of the adjacent tracheids become reservoirs for such portions of the resin as are excreted. Such excretions may retain the form of globules, or extend across the cavity and assume the form of thick or thin plates, or when the quantity is large nearly fill the cell. The writer does not consider the ‘“‘Cordaitales’’ (Araucarians) peculiar among the gymnosperms with respect to the form of the resin masses and their location. The same form and location obtain in Pinus albicaulis, not sporadically, but as a constant feature of the heartwood. Similar deposits have also been noted occasionally in the tracheids of P. resinosa, Picea sitchensis, and in abnormal sapwood of Pinus ponderosa, while globules of resin have been observed on the outside of the pit membranes of ray parenchyma cells in P. Strobus. These instances, taken in connec- tion with PENHALLOW’s note of the occurrence of resin plates in P. parviflora and in two species of Abies, lead the writer to believe that. they probably occur sporadically in many other representa- tives of the Coniferae. The writer has also noted tracheids in P. albicaulis with resin globules at several lateral pits connecting with the secondary epithelial cells of a vertical resin duct, showing that resinous tracheids in some instances may be independent of the rays. The association of the resin plates with special constriction of the tracheid cavity, as noted by PENHALLOw and figured by him for Dammara australis, appears to the writer to be without special significance. This constriction is due to increase in thickening of the tracheid walls where in contact with the rays, and has been observed by the writer as a common feature of the woods of various genera of the Coniferae, especially in the thick-walled cells formed late in the season. Such increase is presumably due to greater nutrition at that portion, and in most species is not in connection with resin plates, while the resin masses also occur where there are no such constrictions. rgr18] RECORD—RESINOUS TRACHEIDS 65 PENHALLOw states that no exact comparison can be established with other plants. The writer believes that exact parallels exist in many of the angiosperms. A good example is Nyssa sylvatica, and figs. 4 and 5 show resin-like plates (RP/) in this wood which exhibit virtually the same origin, form, and distribution as those Fic, 4.—Transverse section of Nyssa sylvatica in Pinus albicaulis. In Nyssa these masses occur in both the tra- cheae and wood prosenchyma and have their origin in the vertical strands of wood parenchyma (WP) and in the rays (R). Globules (a, 6) are shown emerging from the pits. The plates across the vessels are thin (c), but those in the libriform fibers may be very thick (d); in fact much of the fiber cavity may be completely 66 BOTANICAL GAZETTE [JULY filled except for occasional bubbles. Seen in transverse section a plate in a vessel appears as a thin imperforate membrane thickened in contact with the wall or with a rupture in the middle (fig. 4, c), presumably due to shrinkage. Exactly the same features charac- ‘terize the resin masses in the tracheids of the gymnosperms. The presence of secretions (or excretions) in the tracheae only or in both tracheae and prosenchyma in the dicotyledons is very 2oS sSS— TM TLET SE OES 28 ani ag Bet | Va & Se =< i . rk ‘seam : 2S a= Fic. 5.—Tangential section of Nyssa sylvatica common, and in a great many cases such material is in the form of a collar about the cell wall and with a diaphragm of greater or less thickness across the cavity. The writer believes that these various substances, although different in chemical composition, are alike in being excretions resulting from the metabolic activities of paren- chyma cells, and represent waste materials. Although produced in varying amount under normal conditions, the greatest production occurs when the cells are about to cease their vital functions and become heartwood, or when a similar condition is produced abnor- 1918] RECORD—RESINOUS TRACHEIDS 67 mally, as in severe wounds. The low vitality of the ray and wood parenchyma at that stage may be considered responsible for the excessive amount of waste matter produced, some of which finds its way or is excreted into the cavities of the adjacent cells. The parallel between the gymnosperms and angiosperms in the manner of their disposal of secreted or excreted products in the xylem extends further. Pinus, Picea, Larix, and Pseudotsuga, for example, have vertical resin canals in the wood. Similar canals are normal to the secondary woods of nearly all genera of Dipterocarpaceae and of Copaifera, Daniellia, Eperua, Kingioden- dron, Oxystigma, Sindora, and Prioria of the Caesalpineoidae. Ver- tical resin canals arise traumatically in Abies, Sequoia, Tsuga heterophylla, and others; similarly canals may be produced by injury in Liguidambar, Styrax, Terminalia, Drimycarpus racemosa, etc. Resin canals in the medullary rays occur normally in Pinus, Picea, Pseudotsuga, and Larix; similar canals have been observed by the writer in representatives of 11 genera of Anacardiaceae and 2 of Araliaceae,? and others have reported traumatic canals in both planes in Liguidambar and Styrax. The writer concludes that resinous tracheids in gymnosperms find numerous parallels in the angiosperms, that they represent one form of reservoir for excretions, and that the form of the resin masses is in response to well known physical laws. No direct functional activity is attributed to the resin plates, although they are in certain ways analogous to tyloses and reduce the permea- bility of the wood. As a diagnostic feature resinous tracheids appear of value in Pinus albicaulis and may prove to be so in other cases. YALE UNIVERSITY ORD, SAMUEL be Intercellular canals in dicotyledonous woods. Jour. Paani periann 19 BRIEPER ARTICLES BISPORANGIATE CONES OF PINUS MONTANA (WITH ONE FIGURE) In the latter part of June 1915 the writer found 3 clusters of bi- sporangiate cones of Pinus montana on a tree along the University Drive, Madison, Wisconsin. Nearly all the cones of each cluster bore both macrosporophylls and microsporophylls, the latter being in every case on the lower portion of the cone. The macrosporophylls were borne in most cases on only the upper portion of the cone. In a few instances the cones were almost wholly staminate or pistillate. The sporophylls and spore sacs appeared to be normal in every respect. No abnormalities were observed in the pollen grains which were stained for a micro- scopical examination. For the past two years the same tree has failed to produce cones of the type described. Bisporangiate cones have been reported in only one other species of pine, namely, in P. maritima by GOEBEL (1900). How- ever, in a number of other gymnosperms such cones have been described. More than 50 years ago DicKson (1860) reported them Fic. 1.—Cluster of bispo- in Picea excelsa, later SHAW (1896) in rangiate cones of Pinus é upbiskeas Carle oeick Sequoia, and more recently RENNER (1904) pistillate, lighter portion 12 Juniperus communis, and HILL and staminate; reduced one-half. DE FRAINE (1909) in Pseudotsuga Douglasit. In every instance, thus far reported, the microsporophylls and macrosporophylls occupied the same relative positions on the cone as in Pinus montana.—W. N. Steir, University of Wisconsin, Madison, Wis. Botanical Gazette, vol. 66) [68 1918] BRIEFER ARTICLES 69 HEALTHY AND SICK SPECIMENS OF BRYOPHYLLUM CALYCINUM Those who have worked with Bryophyllum calycinum, WALKER, Der Vries, GOEBEL, the writer, and probably many others, have all noticed that the leaves of Bryophyllum which form shoots when iso- lated will rarely or never do so when in connection with a normal and healthy plant. Miss E. L. Braun' makes the following statement: Pot-grown plants of B. calycinum in the writer’s possession have frequently grown both shoots and roots from leaf notches while the leaves were in con- nection with the plant. Early in the spring of 1917 a large plant of Bryo- phyllum began to produce shoots from the leaves more abundantly than the plants often do. The accompanying photographs were taken May 12, when shoot production had reached its maximum. It was not necessary to induce the notches to grow; they grew freely under ordinary room conditions, and with only the usual attention which a pot plant in a residence receives. A number of the leaves of the plant produced shoots from all the notches or from all except the basal notches, a phenomenon which, to accord with Loep’s theories, should take place only under very special conditions. The plant appears to be a “healthy plant,” as healthy and vigorous a plant as the writer has ever seen. Whether or not it is a ‘‘normal plant,” as a norma plant is conceived of by Logs, is difficult to say, for nowhere does he define a “normal plant.” He does state: ‘If, however, the flow of substances in a plant is abnormal, either because the roots or the apical parts or both have suffered, a growth of shoots may occur in moist air from the notches of leaves which are in contact with the plant.” There is no indication that either the roots or the apical parts have suffered; the plant appears healthy, and has had no accident. A glance at the photograph accompanying Miss BRAuN’s statement will show to those familiar with ‘normal’? Bryophyllum that the plant observed and photographed by Miss BRAUN was a sick specimen. The normal stem of Bryophyllum calycinum is perfectly straight and vertical (and unbranched). The specimen observed by Miss BRAUN has not a single straight stem. Stems so weak as not to be able to grow vertically upward are certainly abnormal in regard to nutrition. The bend in the stems acts like a partial block to normal circulation. Such sickly bent stems behave to all purposes like isolated pieces of stems whose leaves will in time give rise to shoots.—Jacques Logs, Rockefeller Institute for Medical Research, New York City. * Bor. Gaz. 65: Igt. 1918. CURRENT LITERATURE NOTES FOR STUDENTS Hybrid vigor.—The phenomenon of hybrid vigor has come to hold a very important place in practical plant breeding, and is of considerable theoretical interest to geneticists. The most generally accepted interpretation has been East’s' “‘heterozygosis,” according to which hybrids are vigorous because of their heterozygous sets. Heterozygosis has been very valuable in helping to organize our ideas on the general subject of hybrid vigor, but as a theoretical explanation of the phenomenon involved it has been unsatisfactory. When one says that hybrids are vigorous because of their heterozygous sets, he is making an accurate restatement of the fact of hybrid vigor in the language of genetics, but he is not providing any real explanation of the phenomenon. The only acceptable “real” explanation that has yet been presented is as ollows. In nature a “struggle for existence” occurs among species and indi- viduals. There occurs also a struggle for existence among unit characters. If a unit character is undesirable it is eliminated, since the species possessing it is eliminated. The unit characters, therefore, that have survived and appear in the plants of today are for the most part ‘‘desirable” ones, although some undesirable ones also may have survived, having been carried through in asso- ciation with the ‘desirable’ characters. The majority of unit characters today, however, may certainly be regarded as “desirable” ones, and a majority is sufficient for the present argument. The question then is raised as to what constitutes a so-called “desirable” character. It may, of course, be any one of a number of things, but is there not some feature common to all such ‘‘desirable’”’ characters? The character would seem to be vigor. Each ‘‘desirable” character must add somewhat to the vigor of the plant that contains it, and if vigor is increased, such things as size and productiveness be increased. Those plants, therefore, = be most vigorous which have in combination the greatest number of ‘desirable’ characters. The next question is, what plants, in general, have in combination the greatest number of desirable characters? The answer is hybrids, for they com- bine the ‘“‘desirable” characters of both parents. Thus, in general, hybrids have twice as many “desirable” characters as do pure races. At this point the objection is raised that though hybrids do actually contain this double quota, * East, E. M., and Haves, H. K., Heterozygosis in Ae and in plant breed- ing. U.S. Dept. Keke Bur. Pl. Ind. Bull. no. 243. pp. 58. 70 1918] CURRENT LITERATURE 71 each character is represented by only a single dose in the hybrid and by a double dose in the pure race, so that mathematically the two situations are equivalent. This objection is valid only if we assume complete lack of domi- nance. We are certainly within our rights in assuming some slight degree of dominance, and if we do this it follows that hybrids have more in the way of active “desirable” characters than have pure races, and, having more “desir- able” characters, hybrids are more vigorous. They are vigorous, not because they contain more heterozygous sets, but because they contain more positive dominant characters. This is a rather obvious explanation of hybrid vigor, and one that has probably occurred to a number of geneticists, being commonly referred to as “the hypothesis of dominance” (accounting for hybrid vigor). It involves 3 assumptions: (1) that there is such a thing as dominance; (2) that most domi- nant characters are “desirable” ones, that is, of survival value; this assumption is rendered easier if we accept the presence and absence hypothesis; (3) that these dominant “‘desirable” characters add more vigor than they detract from it, and add to vigor to the degree in which they are dominant. This last assump- tion is the critical one; but even that seems very reasonable. KEEBLE and PELLEW? suggested this explanation in 1910, and since then it has had some discussion in the literature. At first statement the theory seems sound, but actually it does not fit the facts. The two chief objections to this theory of dominance may be found in the publications of SHULL, EMERSON, and East (loc. cit.). 1. If hybrid vigor were due to dominance, it would be possible in genera- tions subsequent to the F, to recombine in one race all of the dominant deter- miners. Thus there could be isolated a race that was “100 per cent vigorous,” and since it would be homozygous, its vigor would not be lost by inbreeding. Actually, though, hybrid vigor cannot be fixed in this way; ‘‘all maize varieties lose vigor when inbred.” 2. Experience assures us that the distribution of individuals in the F. generation with reference to hybrid vigor is represented graphically by a sym- metrical curve similar to the normal probabilities curve; the class containing the greatest number of individuals is that which shows the medium amount of hybrid vigor, while on either side of this class the fall in the curve is regular, Teaching its lowest point in the two small extreme classes which show respec- tively greatest hybrid vigor and least hybrid vigor. According to the domi- nance hypothesis, however, the largest class of F, individuals would be that showing the greatest hybrid vigor, while the smallest class would be that show- ing least hybrid vigor. The curve representing such a situation would be unsymmetrical and strikingly different from that which actually occurs. For these two reasons the dominance hypothesis seems to have been discarded. * KEEBLE, F., and PELLEw, C., The mode of inheritance of stature and time of flowering of peas (Pisum sativum). Jour. Genetics 1:47-56. 1910. 72 BOTANICAL GAZETTE [JULY Although it is theoretically attractive, its failure to satisfy these two important details of the hybrid vigor situation has condemned it. Jones’ has ingeniously modified the dominance hypothesis so as to avoid these difficulties. At first consideration his theory seems to be clearly the most reasonable explanation of hybrid vigor that has yet been presented, although in time it may encounter destructive criticism. The argument is essentially the same as that for the old dominance hypothesis, with the following important modification. Assume that one parent contains the dominant determiner A, linked with the recessive c; on another chromosome it contains B linked with d. The total formula may be expressed conveniently as Ac, Bd. The other parent has the formula aC, bD. The hybritl is more vigorous than either parent because it combines all 4 dominant determiners. The attractiveness of this scheme is that it escapes the objections that were made to the older dominance hypothesis: (1) the fact that 100 per cent hybrid vigor cannot be fixed is quite in accordance with JoNnEs’ scheme, for it is obviously impossible to isolate a homozygous race, combining the 4 dominant determiners, A, B, C, and D (unless crossing over occurs); (2) a simple mathematical derssnatention will show that the distribution of F. individuals (with respect to hybrid vigor) is quite what it should be, represented by a symmetrical curve, similar to the curve of probabilities. In fact, this new theory, ‘‘the dominance of linked factors,” seems altogether sound. We should reasonably expect that each chromosome would contain one or more dominant determiners (conducive to vigor) linked with one or more recessives. In this day of factors and determi- ners such a hypothesis is quite appropriate. It may be, however, that in the future such a phenomenon as hybrid vigor may be explained on the basis of the stabilities and reactivities of the constituents of specific protoplasts.— MERLE C. COULTER. Taxonomic notes.—BLAKE‘ has published a fascicle of papers containing descriptions of new species. In the paper dealing with Compositae new species are descri in A poorer’ SP saint aaat: Verbesina, Liabum, and Cirsium. Collections from Vene and Curacao contain new species in the following genera: Ruprechtia @), Metis Bauhinia, Croton (2), Mayltenus, Zizyphus, Vismia, Hecatostemon (a new genus of Flacourtiaceae), Passiflora, Jacquinia, Bumelia, Aspidosperma, Plumeria, Marsdenia (2), Lycium, Tabebuia, Dianthera, Oxycar pha (a new genus of Compositae), Simsia, and Verbesina. The new species from Oaxaca are referred to Iresine (2), Amyris, Guarea, Tri- 3 Jones, D. F., Dominance of linked factors and heterosis. Genetics 2:466-479- ie 4 Brake, S. F., II. Further new or noteworthy Compositae. Contrib. Gray Herb. N.S. no. 53. pp. 23-30. 1918. , New Spermatophytes collected in Venezuela and Curacao by Messrs. aI Curran and : - Pp. 30-55. cree plants from pn Ibid. pp. 55-65. 1918] CURRENT LITERATURE 73 chilia, Comocladia, Astronium, Myginda, Homalium, Schismocarpus (a new genus of Loasaceae), Cuphea, Ardisia, and Bouvardia. Britton’ has described a new Scirpus (S. Congdoni) from California, which is the species from the Pacific states heretofore called S. atrovirens. Miss BurtINGHAM® has described 4 new species of Russu/a from Massa- chusetts. FARWELL’ has described 17 new varieties of Michigan plants, distributed among 9 families; and has also published a list of rare or interesting plants from the state. FERNALD® has described a new species of Littorella (L. americana), one of our rarest plants, and heretofore referred to the European L. uniflora. The same author,? as a result of his study of Epilobium from various regions, has published a number of new varieties and combinations and discussed several critical forms. REENMAN,” in continuation of his monograph of Senecio, has presented TOMENTOsI, recognizing 35 species, 2 of which are new, occurring in California and Colorado. The descriptions are accompanied by a full bibliography and liberal citations of exsiccatae, especially such as occur in American herbaria. JOHNSTON and BRuNER™ have described a new species of Phyllachora (P. Roystoneae) found on the leaves of the royal palm (Roystonea regia) growing in Cuba. It is described as forming “‘ conspicuous = carbonaceous masses Several centimeters long on the midribs of the leaves.’ [AcBRIDE” has described new species in Tricyrtis, Atriplex, Lotus, Loma- lium er and Cirsium, and presented the results of his studies of numerous other fo MovRRIL1,% in continuation of his studies of the Agaricaceae of tropical North America, has begun the presentation of the subtribe Agaricanae, ‘Britton, N. L., An undescribed Scirpus from California. Torreya 18:36. fig. r. 1918. ° BURLINGHAM, GERTRUDE S., New species of Russula from Massachusetts. Mycologia 10:93-96. 1918. * FARWELL, O. A., New species and varieties from Michigan. Mich. Acad. Sci. Rep. 1917. PP. 247-262. * FERNALD, M. L., The North American Littorella. Rhodora 20:61, 62. 1918. : Padre etc. Rhodora 20:1-10, 29-39. 1918. * GRE N, J. M., Monograph of the North and Central American species of the genus Seals Part II. Ann. Mo. Bot. Gard. 5:37-108. pls. 4-6. 1918. * JOHNSTON, J. R., and Bruner, S. C., A Phyllachora of the royal ae Mycolo- Sia 10:43, 44. pl. 2. 1918. ** MACBRIDE, J. Francis, Ne th ts, mostly North Ameri- can Liliaceae and Chenopodisceae.. ‘Coutts. Gray Herb. LN. S. no. 53. pp. 1-22. 1918. ‘* Murritt, Wiiuiam A., The Agaricaceae of tropical North America. VII. Mycologia ro: 15-35. 1918. 74 BOTANICAL GAZETTE [JULY recognizing 14 species, 6 of which are included in the present contril New species are described in Atylospora (11), Psathyrella (5), Psilocybe, and Cam- panularius. In a later paper the same author’ has described 28 new species from the same region in the following genera: Drosophila (8), Hypholoma, Gomphidius, Stropharia (2), Agaricus (13), Coprinus (4). MILispAuGH and SHERFF'S have discovered that the species of X. pene are in great confusion, and have described 5 new species, from Vermont ( sates arpum), New York (X. arcuatum), North Carolina (X. ee and xas (X. crassifolium and X. acutilobum). In the same paper a new species 2 Solidago (S. maa) from Illinois is described. SmitH and SMALL’ have described a new genus (Cavea) of Compositae from India, in the East Fsanlora region, belonging to the Inuloideae. It is an extreme alpine form, its structure associating it with Pluchea, but its appear- ance suggesting Saussurea or Berardia. STEPHANI,” in continuation of his Species Hepaticarum, has completed Meizgeria and presented 25 other genera, ~~ with meneiee ila. A new genus (Kormickia) is described and 121 new species distributed among 9 genera. The largest genus is, Plagiochila with 187 species, 92 of which are new. The remaining 29 new species are distributed among the following genera: Metz- geria (11), Symphyogyna (3), Funicularia, Solenostoma, Jungermannia (2), Jamesoniella (6), Anastrophylium (3), Lophozia (2). Watton® has described a new genus (Eutetramorus) of algae secured from the plankton of a pond on the campus of Ohio State University at Columbus. It belongs to the Coelastraceae (Protococcoideae), the colony consisting of 16 cells. ZELLER and DopcE” have monographed the genus Rhizopogon in North America, recognizing 12 species, 6 of which are described as new. In addition, 15 species are presented which have not as yet been found in North America, but may be discovered later. Among these ‘“extra-limital” species 2 are described as new.—J. M. C. ™ MurrRILL, Witttam A., The Agaricaceae of tropical North America. VIII. Mycologia 10:62-85. 1918. *s Mittspaucs, C. F., and SHerrr, E. E., New species of Xanthium and Solidago. Publ. Field Mus. Nat. Hist. 4:1-7. pls. 1-6. 1918. 6 SMirn, W. W., and SmaLL, JAMEs, Cues a new genus of the Compositae from the East Himalaya. Trans. and Proc. Bot. Soc. Edinburgh 27:119-123. pl. 5. 1917- "1 STEPHANI, FRANZ, Species Hepaticarum 6:49-176. 1917. 1918. - %8 Watton, L. B., Euletramorus globosus, a new genus 6a Pages of algae belong- ing to the Protococnides, Ohio Jour. Sci. 18:126—128. 19 ZELLER, SANFORD M., and DopcGe, CarRoL. W., me in North America. Ann. Mo, Bot. Gard. 5:1-30. pis. 1-3. 1918. 1918] CURRENT LITERATURE 75 Abscission.—Hopcson” and KENDALL” have recently contributed to the literature on the abscission problem, the former having investigated foliar abscission in Citrus and the latter the abscission of flowers and fruits in 10 genera of the Solanaceae, and particularly in Nicotiana. The investigation of an abscission problem may be expected to resolve itself into an effort to deter- mine the following points: (1) the histology of the tissue in which the abscis- sion takes place, and the position of the abscission zone therein; (2) the extent of the abscission zone, its histological differentiation, if any, and its develop- ment, that is, whether performed or not; (3) the position of the separation layer within the abscission zone, and the nature of the actual abscission pro- cess, that is, the method of cell separation; (4) the time of abscission, involving both reaction time and abscission time; and (5) the possibility of inducing abscission experimentally (by poisonous gases, mechanical injury, etc.). The most vital as well as, often, the most obscure of these matters which should receive consideration is the one involving the determination of the method of cell separation. In this connection both Hopcson and KEnpALt found, in the species investigated, that the abscission process conforms to the usual type which involves the separation of cells along the plane of the middle lamella. No cell divisions or elongations were observed to precede or accompany abscis- sion. Hopcson notes a remarkable swelling and gelatinization of the cell walls of the separation layer, which is followed by a dissolution of the gelatinized walls. In this case such cells, after functioning in abscission, resume growth and divide rapidly for a time. The abscission problem i in Citrus is of peculiar interest because of the well known shedding of immature oranges of the Wash- ington navel wiy which annually results in considerable financial loss to orange growers KENDALL’S es contains a more or less satisfactory consideration of all these points noted as of interest, but, as is perhaps inevitable in an attempt to cover so wide a field, no more than a beginning is made in working out some of the more fundamental problems. Thus, he shows that from water extracts of separation zones in which abscission has commenced a decidedly heavier precipitate comes down in 95 per cent alcohol than from those in which abscis- sion has not started. This difference is tentatively ascribed to the presence of pectin in the first case, it being derived from the hydrolysis of pectose during the dissolution of the primary cell membranes in the activated separation cells. This conclusion may or may not be justified, but such experiments indicate * Hopeson, R. W. , An account of the mode of foliar abscission in Citrus. Univ. Calif. Publ. Bot. 6: east 918. * KENDALL, J. N., eRe of flowers and fruits in the Solanaceae with special reference to pee Aa Ibid, 5:347-428. 1918 , R. W., Some abnormal water relations in citrus trees of the arid = Hop Soiathnness: re their possible Rdvancs: Univ. Calif. Publ. Agr. Sci. 3:37-54- 76 BOTANICAL GAZETTE [JULY lines along which future investigation should lie, especially in view of the fact that KENDALL succeeded with lower percentages of alcohol in bringing down a - different type of precipitate. This latter precipitate might be expected to yield cytolytic enzymes. He also finds a reduction in the sugar content of abscission zones following cell separation, and that the normal acidity on Nicotiana pedicels is low and is only slightly reduced during abscission. This latter fact is taken to indicate that the activity of enzymes alone is responsible for the dissolution of the middle lamellae during cell separation. KENDALL reports that illuminating gas and laboratory air will cause abscis- sion in the majority of the species investigated, but that resistance to abscission stimulated in this manner appears suddenly in some species. Tests were also made as to the effect of a variety of mutilations of the flower and pedicel in cing abscission. Relatively slight injuries to the ovary were effective, whereas considerable amounts of tissue had to be removed in the case of other flower parts before abscission was induced. It is interesting to note that mechanical injury was not found to be particularly effective in the tomato, and that the following species rarely or never exhibit floral abscission: Nicotiana Bigelovii (3 varieties), N. quadrivalvis (2 varieties), N. multivalvis, Petunia hybrida, Salpiglossis sinuata, Salpichora rhomboidea, and Lycium australis. A detailed summary of the pertinent literature is included in KENDALL’s paper. —T. H. GoopsPeep. Nitrates in forest soils and forest regeneration.—In an important contri- bution HEssELMAN* has reviewed the present state of our knowledge of the composition of forest soils and finds, among other things, that while from earth containing relatively little humus it has been possible to isolate organic com- pounds of known composition the humus of many soils is composed largely of chemical compounds of undetermined character, but that on the whole the constituents are colloidal in nature and are largely influenced by the amount of mineral salts in the soil and ground water. He distinguishes two types 9 forest humus soils, the ‘‘mild humus” characteristic of deciduous forests, well aerated and containing nitrate-forming as well as denitrifying bacteria, an “raw humus” found in coniferous forests as a series of layers of leaves and litter in various stages of decomposition from which nitrate-forming and denitrifying bacteria are usually absent. Recognizing decomposing litter as one of the principal sources of nitrogen in forest soils, he has investigated the “decay capacity” of various forest types, using several different methods. He has determined the relative abundance of various bacteria, the ere content of trees and sip and has shown that nitrate supply and nitrate formation is at n beech forests and at its minimum in mossy eg stands. Lime in the soil and in solution os oa Seeds Henrik, Studier ver saltpeterbildningen i naturliga jordmaner och dess betydelse i i vaxteekologiskt avseende (with abstract in German). Meddel. fran Statens Skogsférséksanst. Haft. 13-14. 297-527. pls. 7. figs. 30. 1917- 1918] CURRENT LITERATURE 77 in the ground water tends to promote nitrification. He points out that by proper forest management the formation of nitrates =e be accelerated and a decided increase in timber production obtained. In a second article** he investigates the problems of the regeneration of conifer forests, with particular reference to the transformation of nitrogen, for it appears that while trees of pine and spruce often grow in forests where no nitrate formation is taking place, the raw humus developed beneath their dense shade does not prove a good soil for the rapid growth of their seedlings. It seems from experimental evidence that nitrogen transformation in such soils may be initiated and accelerated by the introduction of light through cutting, by burning the surface, or by stirring the surface soil. Decaying timber seems to favor nitrogen transformation, and this may tend to account for the observed abundance of conifer seedlings growing upon fallen logs. In mixed conifer stands, especially where the herbaceous undergrowth is good, nitrate formation is, in contrast, rather active; so much so in many instances as to induce such a rank growth of herb and grass vegetation in clear- ings as to crowd out conifer seedlings. ‘These and other data should help to explain to the ecologist many phenomena of secondary succession, while from the same data the forester should receive guidance for the formulation of a policy of forest management that will favor the formation of the amount of nitrogen best suited to the regeneration of the forest. The value of these excellent papers is increased by an abundance of tabu- os data, ae Maes freely illustrated, and by extensive bibliographies.— Eo. D. Fuiy Mechanics of movement in insectivorous plants.—Two recent papers on this aya . Brown’ and by Hooker,” have supplied some interesting information. Although different plants were used, the results are comparable in many respects. Both investigators find that the bending is accompanied by an extension of the cells on the convex side, which soon becomes fixed by growth; that there is little or no change of size in the cells of the concave side; and that unbending is accompanied by growth on the concave side. Hooker finds the Osmotic pressure of the cells on the convex side of bending tentacles less than that on the concave side, and this decrease is proportional to the increase in the length of the cells. He finds no changes in permeability and concludes that the increased size of the cells is due to decreased elasticity of the cell walls. “ HESSELMAN, Henrik, Om vara ee a pa sa peterbildningen i i marken och dess betydelse f ng (with persis ee ares a fran Statens Se pknin. ‘Haft 13-14. penny pls. 15. gs. - Own, Wa. H., The mechanism of movement and the duration of the effect of 2 cheba in the levies of Dionaea. Amer. Jour. Bot. 3:68-90. 1916. * Hooker, Henry D., Jr., Mechanics of movement in Drosera rotundifolia. Bull. Torr, peg Club 44: 389-403. ted be 78 BOTANICAL GAZETTE [yuLY BROWN reports no determinations of osmotic pressure, but finds that if closed leaves of Dionaea are killed, before the extension of the cells has become fixed, and passed through alcohol to xylene, the leaves reopen, and close again when passed back through alcohol to water. He concludes that the increase in size of the cells is due to increased osmotic pressure. He believes there is no permea- bility change, and thinks changes in the elasticity of the cell walls improbable. It is interesting if, in fact, the mechanics of these two responses, so similar in many respects, are so widely different in another. Geotropic bending of growing organs is similar in many respects to the movements studied. Its comparative slowness should make it somewhat easier to follow, and the results might furnish valuable suggestions as to the mechanics of these more rapid movements. SMALL?’ has found differences in permeability in the two flanks of Vicia Faba, roots bending geotropically—THomas PHILLIPS. Soil moisture studies.—The extensive investigations of Briccs and SHANTZ have shown the importance of the moisture equivalent as a constant that will measure the physical properties of soils. Two recent studies deal with certain phases of the same phenomena. The first* shows that while the addi- tion of various salts does not materially change the moisture equivalent of the soil under investigation, if the same salts are washed from the soil with water it then seems to ss a new and peculiar set of physical properties and its moisture equivalent is markedly increased. This increase varies from 2 to 40 per cent, and is taken to mean that the washing out of the salt has increased the interior surface of the soil. The second article, by SmrrH, reports the investigation of the relationship between the results of mechanical analysis and the moisture equivalent. He concludes that there is at present no formula that gives more than a rough approximation of this relationship, and hence that the moisture equivalent cannot be indirectly determined by mechanical analysis with any degree of accuracy.—GEo. D. FULLER Soil aeration and root growth.—Roots of various plants appear, according to the results of CANNON and FREE,® to respond quite differently to variations in the composition of the soil atmosphere, and this difference in response seems *7 SMALL, JAMES, Geotropism and the Weber-Fechner law. Ann. Botany 31:3137 314. 1917. * Suarp, L. T., and Waynick, D. D., The moisture equivalent determinations of salt-treated soils and their relation to changes in the interior surfaces. Soil Sci. 4:463- » SitH, ALFRED, Relation of the mechanical analysis to the moisture equivalent of soils. Soil Sci. 4:471-476. 1917. %° Cannon, W. A., and Free, E. E., The ecological significance of soil aeration. Science, N.S. 45:178-180. 1917. 1918] CURRENT LITERATURE 79 to be related to the character of the natural habitat of the species in question. Thus Salix sp. (probably nigra) stands at one end of the series and shows no injurious effect even when the oxygen of the atmosphere is entirely replaced by either nitrogen or carbon dioxide. At the opposite end of the series stands Opuntia versicolor, growth of roots ceasing with an atmosphere containing 50 per cent carbon dioxide, while Coleus Blumei is comparable to it, showing injury and ultimate death with the addition of 25 per cent nitrogen to the soil atmosphere. Of the other species at Heliotropium peruvianum was closely comparable to Opuntia, while Nerium oleander and Prosopis velutina prove nearly as resistant as Salix. The net seem to indicate that plants growing naturally in well drained soil are much more sensitive to the composition of the soil atmosphere than those from swamps and poorly drained habitats.— Gro. D. FULLER Embryo of Aucuba.—Patm and RutGERS* have settled the question of apogamy in Aucuba japonica, which has been under suspicion for 40 years. They bagged 300 pistillate flowers and not a single fruit formed, while 600 isolated pistillate flowers produced normal fruit after artificial pollination. It is thought that E1cHLEr’s original suggestion of apogamy probably came from the fruiting of an isolated pistillate plant which had developed staminate flowers, since the authors have repeatedly found staminate flowers on pistillate plants. Staminate plants have also been observed to produce pistillate flowers. The flowers open about the time of megaspore formation, and the embryo sac reaches the fertilization stage about 4 weeks later. The solitary megaspore mother cell becomes deeply placed by the extensive development of parietal tissue. The behavior of the 4 megaspores is usually quite normal, but in one case the 2 megaspores nearest the chalaza were found in division. The develop~ ment of the gametophyte is normal, but stages in endosperm formation were not obtained. The chromosome numbers were determined to be 18 and 36. mS — Disease resistance.—JonEs* has published a summary of his results in securing a race of cabbage resistant to the “yellows.” Some of the funda- mental questions involved in resistance were considered. The difference between susceptible and resistant plants was found not to be due to any super- ficial obstacle, but to the different relations of the interior cells of the host and parasite. ‘The resistant tissues have the ability to restrain the development of the parasite to a greater degree than do the susceptible and so give time for protective cork formation.” It was shown also that resistance is clearly inheritable, not as a single character, but as a complex of a number of heritable * Pata, By., and Rutcers, A. A. L., "The embryology of Aucuba japonica. Rec. Trav. Bot. Néerland. 14:119-126. figs. 12. 1917. * Jones, L. R., Disease resistance in cabbage. Proc. Nat. Acad. Sci. 4:42-46. 18, 80 ; BOTANICAL GAZETTE [JULY factors. Environmental factors were found to have a marked influence upon invasion by the parasite (Fusarium), there being a “‘critical soil temperature”’ (about 17° C.) for such invasion. Below this the plants are not invaded even in the sickest soils—J. M. C Desiccation.—An investigation of the course of desiccation and partial starvation in cacti has been made by MacDovuecat, Lone, and Brown. The principal studies center upon the changing rate of water loss, chemical changes in the food reserves, plasmatic colloids and cell sap, and the morphological changes which occur during long periods of desiccation. In one case a large _ Echinocactus was under observation for 6 years after removal of the plant from he soil. Water loss is rather rapid at first, but proceeds more and more slowly with time. ile 10 per cent of the water was lost the first year in one speci- men, during the sixth year only 5 per cent of the water remaining at the begin- ning of that year was lost. The loss of water is much more rapid of course in the open than in diffuse light, and Echinocactus can withstand desiccation not more than 2 years with free exposure.—GEo. D. FULLER. Aeration of nutrient solutions.—StILEs and JORGENSEN* find that aeration of the nutrient solution increases the rate of growth of barley, as found by various workers, but has no effect on the growth of buckwheat, as found by REE. ey carefully limit their conclusion to the condition under which they experimented, and find themselves unable to explain this specific differ ence. They emphasize the necessity of knowing much more about the physical chemistry of water culture solutions. They also feel that neither the law of the minimum nor the principle of limiting factors gives an adequate expression of the behavior of the plant as a whole—Wwa. CROCKER. Apogamy in ferns.—STEIL* has discovered apogamy in a large number of ferns, the investigation extending over a period of 6 years. It seems that apogamy is of frequent occurrence in Pellaea, Pteris, and Aspidium. The prothallis were sealibe under cultural Sendetorss favorable for the development mbryos in non-apo Many interesting details of embryo pre are given, which mich extend our knowledge of this phenomenon.—J. M 33 MacDoveat, D. T., Lone, E. R., and Brown, J. G., End results of desiccation and respiration in succulent plants. Physiol. Res. 1: 289-325. 1915. 34 Stites, W., and JORGENSEN, I., Observations on the influence of aeration of the nutrient solution in water culture experiments, with some remarks on the water culture methods. New Phytol. 16:182-197. 1917. IL, W. N., Studies of some new cases of apogamy in ferns. Bull. Torr. Bot. Club 45:93-108. pls. 4, §- 1918. VOLUME LXVI NUMBER 2 THE HOTANICAL “GAZE BEE AUGUST 1978 DETERMINATION OF WILTING CONTRIBUTIONS FROM THE HULL BOTANICAL LABORATORY 241 ARTHUR L. BAKKE (WITH FIVE FIGURES) The status of the question of permanent wilting in plants, as described by Briccs and SHANTz (5, 6, 7, 8), CALDWELL (11), SHIVE and Livineston (37), and Atway (1), centers about the determination made by Briccs and SHANrz that a plant is re- garded as having attained a condition of permanent wilting when it does not recover its turgidity in a period of 24 hours when sur- rounded by air saturated with water vapor. The method of employing standardized hygrometric paper (2, 3) 4, 28, 30, 38, 40, 42) in the measurement of the transpiring power in plants consists _ in ascertaining the power of a leaf to give off water and comparing this with the power represented by a saturated blotting paper sur- face at the same time. This is then a measure in both cases of the resistance to the passage of water. The conditions which affect such measurements are internal, but these internal factors are dependent upon external factors. It is obvious, therefore, that data derived will be more or less of a resultant complex of all the forces which have been operative during the history of the plant. The method in principle is the same as has previously been used in investigations upon the foliar transpiring power of plants. In the present studies filter paper circles (Munktell’s Swedish 81 82 BOTANICAL GAZETTE [AUGUST no. oo—11 cm.)' are impregnated with 3 per cent solution of cobalt chloride and are later cut into small squares. Just before using, these squares are heated over a bicycle lamp, or on a granite pie- plate suspended by a clamp over the flame of an alcohol lamp, until they become blue. One of these squares is placed between the jaws of a “transpiration clip,” and as quickly as possible applied to either the upper or lower surface of a leaf. The time required to change the paper square from blue to pink is determined in seconds. The time which it takes to change a similar piece of cobalt paper from -lue to pink when placed over a moist blotting paper surface blanketed by a millimeter of air (2, 4, 28, 30, 34, 42) is recorded. The water apparatus is the same as used by BAKKE and Livincston (4). TRELEASE and LivincsTton (42) have developed the relations of the temperature to vapor tension as first shown by BAKKE (2). These authors have presented a formula whereby the time interval may be ascertained on knowing the temperature. Livincston and SHREVE (30) have recently im- proved and modified this method. The principal improvement is in the adoption of permanent color standards. Instead of the simple square of cobalt chloride paper, a composite slip is em- ployed consisting of a small piece of the hygrometric paper in juxtaposition with two slips having permanent color standards. These provide both an initial and an end point for the color change. For use in the laboratory they advocate and describe a simpler form of standard water evaporating apparatus. These modifica- tions were not used in this study. The possibilities of using the original method of standardized hygrometric paper in determining the extent of wilting and the permanent wilting point was first suggested to me by its author, B. E. Livincston, at the Desert Laboratory of the Carnegie Institu- tion during the summer of 1913. In 1914 the writer (3), working at the Desert Laboratory, performed a series of measurements upon sunflower plants lifted from the soil and later brought into the laboratory to wilt. The results of this series of tests show that ? Livincston and SHREVE in a more recent publication (Improvements in the method for determining the transpiring power of plant surfaces by hygrometric paper. ed Plant World 19: 287-309. 1916) have recommended Whatman’s filter no. 30 (11 cms.) circles as being superior to the Swedish paper. 1918] BAKKE—WILTING 83 wilting occurs at a definite point and that permanent wilting represents the most intense wilting possible, without serious rupture of the water columns of the plant. These studies have been amplified in the present investigation. The experimentation involved in the present study was performed in the greenhouse of the University of Chicago during the summers of 1915 and 1916. The large Russian variety of the common sunflower (Heli- anthus annuus) was used, the seed’ being from W. W. BARNARD of Chicago. The experiments involving the porometer were per- formed in the laboratory of Plant Physiology of Iowa State College. The plants were the same variety, but seed was secured from the Iowa Seed Company of Des Moines, Iowa. Series of 1915 METHOD The seeds used in the tests for 1915 were planted in sheet iron containers 66 inches, on June 31. Germination was forced by placing the containers in a warm house. When the cotyledons had made their appearance, the seedlings were thinned out so that only 3 remained. The cultures were then removed to a cooler place, where the plants were allowed to grow until they were approxi- mately 6 weeks old and about 40 cm. high. The soil used in this Series consisted of 4 parts of compost and 1 part sand. The water- holding capacity was calculated to be 47 per cent. The plants remained in the same containers throughout the entire period of the experiment. They were watered from time to time until the - morning of July 13, when they were heavily watered, and after that no more water was added until the morning of July 16, when the plants were lightly watered and the soil surface covered with Plasticine. Two plants were used as checks in testing out wilting by the Briccs and SHantz method. The values for the indices of foliar transpiring power were obtained according to the original Livingston method; the stand- ard water apparatus was the same as described by BAKKE and Livincston. Throughout the series, cobalt paper squares made from Munktell’s Swedish no. oo filter paper were used. As the work was carried on in the greenhouse, the usual bicycle lamp 84 BOTANICAL GAZETTE [AUGUST for lighting was replaced by electric light. The cobalt paper squares were warmed upon a granite pie-plate, which was adjusted by a clamp over an alcohol flame, so that the paper squares were heated to a temperature sufficient to give them the blue color. EXPERIMENTATION The readings for the 1915 series, begun on August 16, were usually made between the roth and 11th hours and again between the zoth and 21st hours. Two plants were used for the foliar tran- spiring power tests; two additional plants were used for the wilting determinations according to the method of Briccs and SHANTZ. Evaporation was determined at the same time by a standardized Livingston form of cylindrical atmometer.. The readings as recorded in table I show the maximum foliar transpiring power as occurring about the 11th hour, while the minimum usually occurs after sunset. Wherever possible, leaves of different ages were used and were numbered and tagged Ia,, Ia2, Ia;, Ib:, 1., etc., the highest number representing the youngest leaf. In this way the same leaf could be used throughout. The average results of the foliar transpiring power indices, as represented graphically (fig. 1), show a general decline from August 16 to August 20. The maximum index reached on Au- gust 17 possesses a value of 0.89. This index is almost the same as the one obtained earlier by BAKKE and LIVINGSTON. Although the plants were watered a little on the day the experi- ment was begun, they must have given off considerable water during the previous 3-day interval. That the soil moisture con- tent has an appreciable effect upon foliar transpiring power has been proven previously, and from the nature of transpiration it is self-evident. The Helianthus plants of BAKKE and LIVINGSTON were growing in a place where the soil moisture was less than would be regarded as optimum. In all probability the two sets of Helian- thus plants were grown in soil having practically the same amount of moisture. The soil moisture content in both series was below the amount necessary for the production of the greatest growth. For the first half of the series the highest transpiring power occurs during the day, while the lowest transpiring power values 1918] BAKKE—WILTING 85 are at night. The average day values are accordingly 0.72, 0.92, 0.74, 0.38, 0.26, 0.19, 0.32 for one set (Ia); for the other (Ib), 0.61, 0.89, 0.76, 0.30, 0.30, 0.39, 0.42. The average night values for the first series are 0.29, 0.34, 0.24, 0.19, 0.25, 0.44, 0.69; for the second series, 0.31, 0.39, 0.23,0.16,0.50,0.45,0.61. The results obtained by calculating the ratio of the respective day and night values are rather uniform. For August 16 the average ratio is 2.4; for August 17, 2.7; the remaining values for Ia are 3-1, 2.0, 1.0, 0.43, 0.46. The corresponding respective values jor series ID are 2.0, 2.3,°'3.3, “£9, 0.97, 0:87, 0.61. For the first two days, August 16 and 17, the probable normal ratio is between 2 and 3. On the following day there is a slight increase, and after that there is a decrease. Whether the rise in the ratio on the third day presents a normal situation or not cannot at present be stated; at any rate the value is not far from 3. The decrease in foliar transpiring power after August 19 and the resulting decrease in the ratio do not show any definite mathematical relation. For a plant growing in a normal environment, a rise in evaporation will give an increase in transpiring power, but on August 22 there is a high evaporation, a low foliar transpiring power, and a lower day value than night value. Such a status must be looked upon as abnormal for growing plants. Beginning with August 21 there is a rapid ascent. Considerable agreement is present between the graphs in this series and the one for Helianthus (3), where the plants were lifted from the soil. There is a decrease in the foliar transpiring power to a point where there is more or less of a balance, and then again where there is an increase. The time element in the present series is extended over a longer period, and as a result variations which might be masked in the series of short duration would be present. The rupture of the water columns of the plants of the 1915 series is as definite as that presented for the plants lifted from the soil in southern Arizona. The outstanding feature of the curve is the very marked rise on August 20. Upon examination of the rate of evaporation, it will be at once evident that the evaporating power of the air was very low throughout. Two plants of this BOTANICAL GAZETTE [AUGUST 86 61 gto £r'o 61°0 of*o €z‘o | LE°o aBeiway o'% St‘o £1'o gio oto feo | Leo 76s age gir L1-o £1°90 oz'9O of'o zz"o | LE'o ee | | oz 61'0 Lio zz'o gto PEO =| fro aBvleAy ie L1‘o £1'0 1z°O Le-o $€'0 | oF'o yee | gil zz'o 1z'o z°0 SE'o vf°O: | gto 7" tpt vez l9°0 gI‘o g1°o 1z'0 <7" Ore go"o tro SO) | 18 O 4 OO:01 [°° "***"IpT 61 ysn3ny 61 ysnany ee €z'o gio 6z'0 glo zl‘o | 6L'0 a8tiway I’? 1Iz‘o pa = gz‘o 9g°° zg’o 16°90" 7 gz Sz‘o gI‘o zt:o Sg'o zg’o | Lg'o ot ae i tz'o SI'o z£"o blo €g'0 | S$g’o aBviaay Le Sz'o gI'o bE'o £6°0 6g'0 | L6°0 ee. tee bz'o LI‘o ze'o tg'o tg'o | Fg'0 pee oz ¥g'0 zz'O Aare) 6z 0 °* -O0sGr gio St‘o g1‘o | ¥l°0 “ye gi ysnany gi ysnsny £"< 6£'0 Lz'o zS'o 6g°0 1g‘o | 460 a8vioAy : vz obo 6z'0 1$'o L6'0 L6‘0 | L6°0 a2 ie a 4 gto z'o zS‘o og‘o tg'o | L6'0 ees. Lz Eo Sz'o £b'o z6"0 06°0 | 6°0 a3RI0Ay ys ob'o Zt*o gro L6°0 L6'0 | L6'0 ee Q'z £f°0 vz'o zb'o ¥6°0 .| 60 | ¥6'0 ot gz Z1°0 of 0 ozo 6£°0 “OO: 1Z oro Sgo 0g'o | Og'0 io OOOR oc ae 2s wae 21 ysnsny o'z Teo 1z'o Ib'o 19°0 gto | bg‘o aBviaay ee 1f°o zz'o ob'o Loo 6£'0 | ¥6'0 ey. gil of'o 61°90 1b'o SS‘o gf'0 | blo pr aP OO TE [*." 1 ge vs 6z°0 zo Sto zl‘o gS'o | Sg’o aBeioAy ; S*z veo lz‘o Iv‘o 930 tl‘o | g6'0 Be | vz of 0 £z"o gt'o tL’o wS‘o | 900 ee | viz Sto Sz°O gi'o lz‘o & aos 00:02 * s ay 9S ‘0 6b'o to +4b4**O0O°77 eer a “py gI ysnsny QI ysn3ny 4unoH amqUuy jaddQ JOMO’T —- amtqUq qeddy IMO] (3ad_“99 NOLL Preneatb NOLL orivaad nom pine peer smc capone -VAWASAO 40 ee NOLLVTIY | auyis woua| waMOd ONTUIaSNVUL ao xaany | “20H INV AVC | axis noua WAMOd ONTAISNVAL 40 XTANT oe OEE AVG NOILVAOdVAY NOILVAOdVA (SOYIVLI NI VWINIW ‘ddAL GHOVA AIOM NI VWIXVN) S161 ‘cz-gI LsaDAYy WOud ONLLTIM JO HOUVW AAISSAWOOUd ONTANA SLNVId snyjUD1aFT OML JO SAAVAT LNAAAAIIA © AOA ATMOd ONIMIASNVAL AVITOA AO SAOLANT I aTaVvL BAKKE—WILTING 1918] jou SBM JIqUINU JS¥] SIY} ‘xapUl 94} 10} advIdAv 94} Surynduios uy uay} Iz ysn3ny 10.47 “dijo ay3 Jo uoTzeaI[ddv ay} 10y YAnOU aF1e] sea 7 1 72q} OS azIs JUTOYNS “FT PUY "AT S9AVA] 1OJ ABvIVAv 943 SI LL°O fpasn — “OSVIOAL IY] ISVIIOUT A[[BINZLU [[IM SIT, “Sz9yO VY) Jo Aue URY) JeyRIY YOnuL st z JO aWlodaq SvY Jva] MOU ¥B SsazZoId UI UVeq Sey JUOUITIEdxa ay Lo xaput aq} [} OUTS » 000000N0 OHHH OOOO 00000000 69°0 £Pv-o a wy BRaawH RRKRSO 600000 600600 ++* 90:82 zz ysnsny oss s998re Iz ysnany oz ysnsny vE"o : zb'o a8vioay gb'o asviaay €z'0 aBelaay zf'o aBeIvAy of'o asRIIAy = OOLTL 88 O0°TT te ysnsny Iz ysn3ny S I ss Ss = BOTANICAL GAZETTE [AUGUST 88 same series were used for the determination of wilting according to the method of Briccs and SHANtz. The results are given L i ! ! : f J ! ; INDEX OF FOLIAR TRANSPIRING POWER ms vt, oe we ate 2 22 ‘0 Ponom ma a4 19 17 Fre. ¥ a 000 Aug 16,1915 in table II; those obtained giving the residual moisture at the time of wilting agree rather closely for the determinations made In the method of Briccs and according to the two methods. TABLE II Method | Ir | 2 Avecses Briggs and Shantz......... 10.38 7.84 G. 1 seater OE 8.20 8.34 8.27 | Hygrometric paper SHANTZ there is a greater variation than is found to be present where the hygrometric paper is used. 1918] BAKKE—WILTING 89 The breaking point occurring on August 21 is not far from the normal minimum value of the daily march of foliar transpiring power. From previous work upon the march of foliar transpiring power, there is more or less of a definite maximum (usually during the day) as well as a definite minimum (usually during the night). It seems that, in all probability, the minimum in the foliar tran- spiring power indicates approximately the greatest resistance to transpirational water loss. If the water content of the soil coupled with the evaporating power of the air is of such magnitude as to increase the resistance to the passage of water, so that the day maximum has a value as low or lower than the diurnal minimum (at night), the plant is then in a critical condition; at least this has been found to be true for Helianthus. For the entire leaf surface the transpiring power ratios at night are as follows: (1) Ia, —o.23, ©.30, 0.34; (2) Ib, —o0.30, 0.31. On August 20 the respective values are 0.24, 0.26, 0.27 for Ia,, and 0.28 and 0.32 for Ib. The average ratio for the first is 2.2 and the average ratio for the test on August 20iso.91. On August 21 the ratio is less; on the following day it is a little higher. The entire situation as here brought forward centers about the amount of moisture present in the soil during the march of wilting when the index of transpiring power ratio of day to night comes to be represented by unity or less. The duration of this ratio may be an important factor in obtaining data that will give information on the relative drought resistance of plants. Series of 1916 METHOD The method ‘of procedure in the experimentation for 1916 was much the same as for the previous season. The sheet iron con- tainers were a little deeper (7 inches instead of 6). The soil mix- ture was lighter than before, containing 1 part of clean pure sand mixed with 3 parts of rich garden soil, and the variety the same as before (Mammoth Russian). The seeds were planted on June 24, and on July 1 the seedlings were 5 cm. high and were then trans- Planted. Three plants were set deeply in the soil. The cultures: go BOTANICAL GAZETTE [AUGUST were then placed in the greenhouse and were watered from time to time. A Livingston standard atmometer of the cylindrical form was set up in close proximity to measure evaporation. Readings were taken of the atmometer whenever a reading was made of the transpiring power. On July 1g the plants were thoroughly watered and were lightly watered again on July 20. On July 21 the con- tainers were sealed over with plasticine preparatory to making hourly readings of the foliar transpiring power for a period of twenty-four consecutive hours. For the measurements upon the index of foliar transpiring power the same apparatus as employed before was used. On the 18th hour of July 22 the last reading was made for the daily march of foliar transpiring power. Beginning July 24, readings were taken three times during the day: (1) at approximately 10:00 A.M., (2) at the 14th or 15th hour, and (3) at some time during the night. The times chosen really represent the three important periods during the daily march, for the first one gives this value at a time when the transpiring power is near its maximum, the second when evaporation is at its maximum, and the third when the index of transpiring power has its lowest value. The leaves were tagged as before so that the same leaves were used throughout. The soil surface of several additional plants was coated over with plasticine to serve as a comparison or check for the plants used for the determination of foliar transpiring power. In apply- ing the cobalt paper squares from day to day, it became easy to judge the condition of the plant. When plants presenting a physical state such as was in evidence for leaf Ia, on August 3, and for leaf Ib, on August 7, were placed in a moist chamber, they failed to recover. It was then deemed unnecessary to test further. At the time of the beginning of the experiment plant Ia was 25 cm. high, while plant Id was 28 cm. high. INDICES OF FOLIAR TRANSPIRING POWER In using the method of standardized hygrometric paper for the determination of the indices of foliar transpiring power, two separate plants were used. The method of numbering the leaves was the same as for the 1915 series. From plant Ia two leaves 1918] BAKKE—WILTING gI were chosen, Ia, having the dimensions 5 x8 cm., and Ia., 4X6 cm.; from plants Id two leaves were chosen, Ib,, 7X9 cm., and Ib., 34 cm. Whenever a new leaf became sufficiently large for the applica- tion of the clip, approximately 3X4 cm., it was included with the others. The readings were begun on the 18th hour of July 21 and continued at hourly intervals for 24 hours. Readings were ' taken at the same time from a standardized Livingston cylindrical form of atmometer. The results for plant Ia are given in table III. This series shows that the march of the foliar transpiring power is the same as has previously been pointed out (2, 4, 28, 40), in that the maximum transpiring power is attained at a time previous to the greatest evaporation. The highest index occurs usually at the roth and 11th hours, while the evaporation maximum occurs usually on the 14th hour. On account of the larger num- ber of readings it is to be expected that the graphical representation (fig. 2) will show less abruptness than has formerly been presented. Recalling that the leaf represented by Ia; is older than Ia,, it is plain that the index of foliar transpiring power is higher for the younger leaf almost entirely throughout the 24-hour period. The maximum for Ia, is at the 11th hour, when it is 0.93. This value is again in evidence 2 hours later. For Ia, the highest value is at the roth hour, when the transpiring power value is 1.00. This same value is again reached at the 12th hour. It is evident that the younger leaf Ia, reaches its maximum at an earlier period than the older leaf Ia,. This feature substantiates similar conclusions reached by BAKKE and Livincston. Another important feature in connection with the graph showing the march of foliar transpir- ing power is the sudden drop for both leaves. The lowest point (0.59 for Ia, and 0.67 for Ia,) occurs on the 14th hour. At the 15th hour the index values are respectively o.91 and 0.93, while the corresponding values at the 13th hour are 0.93 and 0.83. Although the drop is the feature in the afternoon readings, the recovery occurring at the 15th hour is always below that of the fore- noon maximum. In the present case there is not much difference, being o .91 for Ia,; for the younger leaves there is a greater variation, being ©.93 at 15:00 o’clock and 1.00 for the 10:00 and 13:00 o'clock readings. At the 14th hour the average reading for the g2 BOTANICAL GAZETTE [AUGUST ‘TABLE III DATA FOR MARCH OF INDICES OF FOLIAR TRANSPIRING POWER FOR Helianthus PLANT Ia MAXIMA IN BOLD FACED TYPE, MINIMA IN ITALICS) : . ee ee INDEX OF TRANSPIRING POWER EVAPORATION EAF F NUMBER ceuniihees “ ‘ é = sai penaiimetnee wartecd pa ese hitice Petr apa jeal winieakeay ate. July 21 ES ai i 18:15 31 52 0.77 0.46 0.62 ta 18:20 24 46 I.00 0.52 0.76 564 6 ee 19:10 42 65 0.68 0.44 0.56 0.22 Migs es 19:15 38 55 0.75 0.52 0.64 Meso ciate 20:15 46 59 0.62 °.49 0.56 0.15 fo ee NS 20:20 40 54 9.73 0.54 0.64 Fae 58; 21:10 45 72 0.64 ©.40 0.52 0.16 Dae cowed 25°20 40 50 O78 0.58 0.66 pS eh 22:10 51 a 0.61 0.40 o.51 0.09 (OF aD 22°75 40 65 0.78 0.48 0.63 : We 7 5% 23:10 56 74 0.57 0.43 0.50 0.15 Moe: 23:10 45 69 oO. 71 0.46 0.59 July 22 Migs 24:10 50 90 0.66 0.37 0.52 0.18 Me. 24:15 42 77 0.79 0.43 0.62 ee 776 48 88 0.71 0.38 0.55 0.12 Ble ok ass 15 42 74 0.81 0.46 0.64 Bigs 2°20 57 0.63 0.40 0.52 0.15 CR a Re, aie 49 85 0.74 0.42 0.58 Oe eee 3°10 55 85 0.65 ©.42 0.54 0.15 Me eesia, a:%s 50 75 0.72 0.48 0.60 Mies rye. 4:10 50 85 0.92 0.42 G57 0.15 YF ORES eer 4:15 46 78 0.78 0.46 0.62 Tae 5:10 50 62 0.72 0.58 0.65 0.15 ey 5715 45 57 0.80 0.63 o. 72 jE ae 6:10 39 50 0.87 0.68 0.78 0.15 Fe ee 6:15 36 50 0.904 0.68 0.81 dF ae ae 4230 36 42 o°.81 0.69 0.75 0.15 Ah Se es 1:10 36 a7 0.81 0.78 0.80 Mies 8:15 30 42 0.77 Oss 0.66 0.15 i FE ea 8:20 28 a2 °.82 0.72 0.77 Me 9:05 24 24 0.88 0.88 0.88 0.22 ee VES 9:05 22 22 0.95 0.95 0.95 Meee. 10:05 22 24 0.95 0.88 0.92 0.58 pees 10:05 a1 21 1.00 1.00 1.00 Me Oe. IT:00 21 20 0.91 0.95 OF tics see aes ree II:00 20 20 0.95 0.95 0.95 eee 12:00 21 21 0.86 0.86 C780 Fae hee 12:00 18 18 I.00 I.00 1.00 jE ee Rene 13:10 16 16 0.93 0.93 0.93 1.0 Mae c ee. 13:10 18 18 0.83 0.83 0.83 Rie 14:10 17 yy 0.59 0.59 0.59 1.2 ERS Pate Paras 15 15 °.67 0.67 0.67 A SP ena 15:10 15 16 0.93 0.88 0.91 1.2 as 55's 55s T5:15 15 15 0.93 0.93 0.93 Wi 84s 16:15 17 18 0.88 0.88 0.88 1.6 Toe 16:20 ar 21 0.71 0.71 0.71 Tae 17:00 19 19 0.905 0.95 0.95 I.0 ee eA 17:00 18 20 I.00 ©.90 0.905 ee 18:00 25 35 0.80 0.57 0.69 1.0 Mis cee 18:05 25 1.00 0.80 ©.90 1918] BAKKE—WILTING 93 older leaves (Ia,) is 0.59 and for the younger leaves is 0.67. The differences then in the order given are 0.33 ando.34. The respec- tive values on the 15th hour are 0.91 and 0.93. These give recovery value differences of 0.32 and 0.26. The drop in the afternoon reading is not a new thing, either in foliar transpiring power or in transpiration. No doubt this great resistance to the E ? lez x 4 1 000 Pa NY 938 f aN 875 ; a ‘ a3 ps A j 7508, y ms , P \ INOEX OF FOLIAR TRANSPIRING POWER Mg esd \ / \ ane, F A tare x h.o7 yi es af ow ‘at - 563 emia | sod ee oa 434 376) 313 2 oof 250 sy nso} 1881 F hoo} 125) cunuanes eer ee BEZe — ~ 2 a ~ e S Li T 1 Tt t J wy2ha916 Fic. 2 Passage of water is a condition of incipient drying. It may be that at this period, usually present at about the time of greatest evapora- tion, there is a lack of water, not only in the cells of the leaf, but also in the vessels themselves. SHREVE (40) has submitted evi- dence, at least theoretical, showing that variations in the tran- spiring power are due to variations in the water-holding capacity of the internal tissue. Using the Drxon (16, 17, 18) conception of continuous columns, as well as the results of the experiments of RENNER (34, 35, 36) upon transpiration, there is doubtless a greater tension present upon the water columns. If this is related to 94 | BOTANICAL GAZETTE [AUGUST absorption and incipient drying, an additional force must be present in order to cause the water to be pulled into the cells to a greater degree than before. If this interpretation is correct, the older leaves (on account of their closer proximity to the absorption center) should show a more complete revival. This speculation would necessarily be based upon the readings of the secondary maxima. The comparative values become evident, for Ia has a maximum of 0.93, falls to 0.59, and subsequently returns to 0.91; for Ia, the maximum is 1.00 and comes back from 0.97 to 0.86. This fall and subsequent rise are independent of the evaporation rate. The march of the foliar transpiring power is more or less definite. This is especially true as it bears upon maximum and minimum values. For Ia the maximum value occurs on the 11th hour with an index of 0.95, while the minimum value 0.50 is on the 23d hour. The ratio between maximum and minimum is 1.9. For Ia, the maximum value, 1.00, is in evidence on the roth hour, while the reading 0.58 on the 2d hour of July 22 gives the minimum value. The ratio in the latter case is 1.72. : Another important feature presented by the present series is the high value on the 17th hour of July 22. In the previous experiments which have dealt with foliar transpiring power, there has been a fall in the transpiring power value after the secondary rise. It is noticed that the evaporation during the afternoon is rather intense, being 1.6. The high value of the transpiring power, therefore, is without question due to the high evaporating power of the air. In formulating a graph (fig. 3) from the data presented in the march of foliar transpiring power, the general feature is the high foliar transpiring power before the time of greatest diurnal evapo- ration. For both leaves the maximum is reached at the 11th hour, when the index is 1.00. This value is retained for Id, until the r2th hour, and for Id, until the 13th hour. The minimum value (0.47) for Ib, occurs on the 22d hour, while for Ib, (0.44) it occurs on the 18th hour. The ratio of maximum to minimum or of day value to night value is 2.1 for Ib, and 2.3 for Ib,. The sudden drop in the afternoon reading on the r4th hour is equally as striking as that 1918] BAKKE—WILTING 95 TABLE IV DATA FOR MARCH OF INDICES OF FOLIAR TRANSPIRING POWER FOR Helianthus PLANT b (MAXIMA IN BOLD FACED TYPE, MINIMA IN ITALICS) Ph vw Se pra INDEX OF TRANSPIRING POWER | EVAPORATION Lear TIME OF igi NUMBER OBSERVATION te U Te U Entire poiecuteene ec: serine Rest acne suit ace leaf i speeed July 21 Teor es 18:25 ga 63 0.89 0.38 COA Oty eae: Le 18:25 44 73 0.55 0.33 0.44 ae ER 19:25 40 67 0.72 0.43 0. 58~ 0.22 Mee: 19:30 44 81 0.65 0.35 0.50 Mss . 20:25 41 68 0.71 0.43 0.57- 0.15 Besa ass 20:25 43 78 0.67 0.37 0.52 fe 21:15 90 0.66 0. 32 0.49 0.16 ee 21:20 30 105 0.74 0.28 0.51 WE hess 22:20 52 95 0.60 0.33 0.47- 0.09 Besa. 22:25 50 97 0.62 0.32 0.4 Eee 23:15 47 85 0.68 0.38 0. 53~ 0.15 Se 23:20 43 82 0.74 0.39 0.57 eee 24:20 53 78 0.62 0.42 0.52- °.18 | ee 24:2 40 93 0.67 0.45 0.56 July 22 ee st T:2 51 78 0.67 0.44 0.567 0.12 eas 1:20 47 71 0.72 0.48 0.60 eee 2:20 57 75 0.63 0.48 56- 0.15 ES ay wa 2:26 50 72 o.72 0.50 ° a ea 3:20 46 7° 0.78 0.52 0.65~ 0.15 Mice esse 3:20 41 7° 0.88 0.52 0.79 Pei Ai ais 4:20 40 90 0.73 0.40 o. 57° 0.15 Cis as 4:20 43 75 0.84 0.48 0.66 cs 5:20 48 82 0.75 ° o.60— ere 84 Mee es. 5:25 40 85 ©.90 0.38 0.64 ai. 6:20 42 50 0.81 0.68 ©:75 7° &-45 ERA TAS 6:20 46 55 0.74 0.62 0.68 een ee 7ii5 31 37 0.74 °.62 0.68 0.15 ee 7 36 26 42 0.89 o-5 fee cs 8:25 25 31 0.92 0.74 0.84 0.15 Jar ee 8:25 25 28 0.92. | /0.82 0.87 ae 9:10 25 30 0.84 0.70 o.77 77 0.22 Mere we 9:15 21 22 1.00 ©.95 0.98 a 10:10 25 26 0.84 0.81 0.83 - 0.58 My cas 10:10 23 24 0.91 0.88 0.90 ee IT:10 19 19 1.00 1.00 eg ere Be ele yt TI:I6 19 19 1.00 I.00 I.00 Meee hay I2:10 18 i 1.00 1.00 ba aan ae a ae ahi 12:10 18 18 1.00 1.00 1.00 5 EN T3235 16 16 0.93 0.93 0.93- r.0 Mee. ES 13:20 15 15 1.00 1.00 1.00 cheese 14:15 14 15 0.71 0.67 0.69 1.2 ei be 14:20 17 17 0.59 0.59 0.59 Mie rig ie 15:20 15 15 0.93 0.93 0.93 1.2 Me cae. 15:20 16 17 0.88 | 0.82 | 0.85 A 16:25 16 18 0.94 | 0.83 0.89. 1.6 iwc tt 16:25 17 19 °.8 9.79 0.84 Mew ick a 6 17:10 19 18 0.95 1.00 0.98- ibe Mines 2) T9728 20 23 0.9¢ 0.78 0.84 : cae 18:10 27 30 ° 0.67 | 90.71 1.0 Mae ene s 18:15 20. 33 0.67 0.61 oO. 96 BOTANICAL GAZETTE [AUGUST noted for series Ia. For Id, the drop really begins on the 12th hour and falls from 1.00 to 0.69, giving a difference of 0.31; for Id, the fall is from 1.00 to 0.59, giving a difference of 0.41. The recovery for Ib, is from 0.69 to 0.93, and for Ib, is from 0.59 to 0.85. The difference value is 0.24 in one case and o.26 in the other. It was pointed out for series Ia that the recovery of the older leaf is more marked than that of the younger leaf. This E|? Tb: 1000 \ ae 938] Al; [ SoM, 813] INDEX OF FOLIAR TRANSPIRING POWER Sear Bie 150] 188} a *, x 100] 125 ae a os cape EVAPORATION at Oe oe lewwachosassptoretes i tS jaty 2916 . Tidy > 7 om Ps BNE | IB Bs ese 5 BERS Fic. 3 feature is again borne out in the present series, where the values for Ib; are in excess of those of Id,. For the reason that the leaves of series Id, are very nearly of the same age, the same variation as set forth in the previous season will not be in evidence. The minimum values are slightly lower. As a result the ratios between maxima and minima are respec- tively 2.19 and 2.27. The same high foliar transpiring power is present at the 17th hour. This agrees with the former series. The data submitted in table V give the march of foliar transpiring power during the process of wilting from July 24—August 7. 1918], BAKKE—WILTING TABLE V INDICES OF FOLIAR TRANSPIRING POWER DURING PROGRESS OF WILTING OF Helianthus PLANT Ia Index of t iri r ‘ Se ie ee a t ized atmometer, te ti cc. per hour Or1S ess eas G.51 0.68 rt. TOW PATA OG yo Ge). Fs 0.54 I.23 BOS OO vos °.28 0.39 0.93 G86 is 0.49 0.20 0.24 Jay: S644 O00. 0.27 Oo. 2% 0.71 AL OO. es tas 0.59 0.31 0.66 NE pa teeeaes | 0.51 0.31 o.II Vy 396 444r se. oe: 0.70 0.20 0.QI Bat ES eer oes 0.49 0. 27 0.62 ee 5 Pe nar e 0.43 0.20 0.02 Jy BA AOE i ee. 0.65 o0.18 1.25 7a ae dn a 0.69 0.18 o.51 O00 ora 0.38 o.16 0.12 BO ae POP ET Say: 0.22 0.14 1.12 ns Sig <2 CRP ae 0.36 0.19 0.66 O26 0 ee. 0.33 °.18 0.13 DRY 90 < 14560. os. 0.15 °. STOO 2 oy - 0.13 0.69 G08 ais . peak °. 28 PUY. 2 ROS i488 . 0.12 °o.18 S006. ini iis . o.16 o.9r TOSOS ca ey: . 0.34 °. 26 Poly: at taste: ois... . 0.17 I.1I POE tas . 0.09 0.66 TOrOe. as * 0.17 0.24 Augost 1414:00).25.,.. £ 0.15 0.71 pk eo ne eee ate Se 0.13 o.4I pop fa eg e 0.14 0.24 August. 2/14:10...:...-. * 0.15 1.00 CS | Ce nes - 0.16 0.53 10:00) « 0.14 0.34 August 3 414:00... 02; . . 0.12 1.50 93700. 5a a 0.16 0.37 : een ie 3 0.17 0.22 August 4(14:00........ "8 0.12 1.03 BEOOF ik. a 0.17 1.18 4s oi Co er 0.17 0.33 August 5{14:05........ . +77 0.77 3 feo baat ¥ O.11 6.9% eee * o.14 0.11 August 6414:10........ * o.11 1.49 $1 °06% 0G . 0.13 1.10 Augus oe en * 0.13 O. é hee Vee eves . 0.23 0.74 98 BOTANICAL GAZETTE [AUGUST From the tabulated data of table V, and from graph (fig. 4) of series Ia, it is noticed that there is a marked decrease in foliar tran- = a ~~; “Ww, Me “tay spiring power from July 24 up to the time when the if plant wilts. The transpir- le ing power of leaf Ia, is . very irregular. There is 2 J : no doubt but that the plant Beer has attained its permanent Vg wilting point on July 29, — but because the leaves of H i this series are somewhat a older than Ia,, and as they z tf are located nearer the ab- Fa sorptive center their action BERR will be more or less modi- ? 7 fied by the presence of the a younger leaves at the tip. ba ‘| . For the leaf Ia, there is a Mol g marked decrease in the 4p S P as me \) A] INDEX OF FOLIAR TRANSPIRING POwE az 4 me Fe. *s, at y ly 24 1916 foliar transpiring power from July 24 to July 309, the foliar transpiring power being especially high on tg eT Ge a August 1. This feature is Sina’ PBB ES oe probably in response to the (fi Fae exceedingly high tem- perature at that time. The evaporation from the standardized atmometer bears out this statement. From August 1 to August 7 the index of foliar tran- spiring power proceeds almost in a straight line, except for small dips occur- ring in-the majority of 1918] BAKKE—WILTING ‘ 99 cases when the evaporation was at its highest. This part of the graph conforms with the one obtained when the plants were lifted from the soil. On August 7 the index of foliar transpiring power increases from 0.13 to 0.23, or 77 per cent from the roth hour to the 14th hour. On the previous day it dropped from 0.14 to 0.11, while on the preceding day the two indices were the same. At no other time during the march was there such a great percentage increase. In obtaining the ratio between the day reading and the night reading for 24 consecutive hours, the day readings were usually made between the oth and the roth hours. For the night readings there was no need of selection as only one reading was taken. Beginning with July 24, and continuing until July 29 (time of wilting), the transpiring power indices representing the day read- ings for Ia, are 0.51, 0.49, 0.51, 0.43, 0.38, 0.33, while the corresponding night values are 0.28, 0.59, 0.49, 0.69, 0.36, 0.33. The respective ratio values are 1.82, 0.53, 3.04; 0:62, 2/05. On July 21-22 the ratio between the reading on the oth hour and the reading on the 21st hour is 0.88:0.52, or 1.7. For the entire 24-hour period the maximum and minimum ratio is 1.9. The only normal ratio is the first. It is interesting to note that for leaf Ia, the minimum is normally 0.50. During the progress of wilting the maximum does .not get below this point until July 27; ‘after that it is below the usual minimum. From July 24 to August 7 (time of wilting) the corresponding indices for Ia, are present; for the morning 0.68, 0.29, 0.31, 0.20, 0.16, 0.18, 0.21, 0.34, 0.17, 0.14, 0.14, 0.17, 0.17, 0.14, 0.13; for the night 0.39, 0.31, 0.21, 0.18, 0.19, 0.13, 0.16, 0.09, 0.13, 0.16, 0.16, 0.17, 0.11, 0.13. The ratio of the day (morning) readings to the night readings is respectively 1.74, 0.94, I-11, 0:04, 1.38, 1.3%, 3.78, 1.31; 6.88, 0:88, t.00, 1.54, 1.08. For July 21-22 the ratio of maximum to minimum for Ia, is 1.72. For the corresponding hours the ratio is 0.95:0.66, or 1.44. On this basis, therefore, the readings of the first day are normal, in that the ratio is approximately the same as for the maximum to the Minimum on July 21-22 (1.72). Also the maximum values during the march of wilting, with the exception of the first reading, are all Ioo BOTANICAL GAZETTE [AUGUST below the minimum set during the daily march of foliar beets power for July 21-22. The data presented in table VI give results that harmonize with those of table V. As was stated in connection with the march of TABLE VI INDEX OF FOLIAR TRANSPIRING POWER DURING PROCESS OF WILTING OF Helianthus PLANT 1b. Index of transpirin r ‘ : entire ea — f sean camel, Time of observation ined atinomneter, Ibs Ibs ec. pet hour Pee Os 0.17 Ors 1.00 JU: Sa ete eos 0.18 0.21 1.23 os ie 2 Sa as aa 0.18 o.14 o.75 O2607. 25. 0.12 0.20 0.24 Paly 26 STA GO... ees 0.17 0.14 o.71 PEI 0s cs ©.19 O:3r 0.66 O35 662.4, 0.12 °. 28 O.I1 RLY OO 4 rate ik 0.18 0.12 0.91 Dg td 9 = ar ip a 0.35 °.19 0.62 Orde ce. 0.10 0.13 0.02 JS 9 CET as is 0.19 °.18 1.25, Ee ple © Mpgtaee pets 0.34 ©.20 0.51 Ce a Mentcaene 0.22 0.13 0.12 daly sB i igras. 2 cy: ©.19 °.14 1.12 SE1OO. or. ye 0.12 0,22 0.66 0:30... yess 0.25 O.70 0.13 July 20 (14200. 7, 0.24 9.10 0.80 AE Os 0.14 o.II 0.69 C1 es 0.31 0.14 0.28 JOly 30 (34590. 2. °.14 a1 1.88 BT2ON ot 0.26 0.15 0.91 TODI08 0 osc °. 26 0.14 0. 26 Tey a) (iets, 0.17 0.17 t,42 BEES ac es 0.13 °.10 0.60 FOLIO. i ©.2T 0.24 AugOst 14445005. 5. a a 0.21 O41 aah (0 me i es AE Re o.14 0.41 oe ea eras ie eae 0.17 0.24 August 2 14 Tes i ee, 0.16 1.09 Se 0.12 o.$3 TOSI ie cl ees O16 0.34 August sf PEN Ph Fie 0.20 1.50 foliar transpiring for 24 consecutive hourly selected here were closer together, and considering the relation periods, the leaves 1918} BAKKE—WILTING IOI which is present between the leaves it would be expected that the ‘ variation would not be as great. The data given in table VI show slowly decreasing values; however, the decrease is not marked. The highest foliar transpir- ing power for Ib, is 0.35, while the lowest is 0.10. The highest point as here set forth occurs on the 21st hour on July 26, while on the following day the index at the same time is 0.34. After that there is a slight fall, although this is not true for all the readings, for on July 30 the index iso.31. Even at the time of wilting, the index at the morning hour is 0.26. From July 29 to July 31 the maximum values are approximately the same. This is also true of the minimum values. The last reading for the Ib, series occurs on July 31 and gives an index of 0.13. For leaf Id, the values are strikingly similar to those of the leaf situated just below it upon the stem. The highest transpir- ing power index for the entire time is only 0.28, and occurs at 9: 35, July 26; while the minimum value 0.10 occurs on the 14th hour of July 29 and on the 2rst hour of July 31. The data of table VI, tepresented graphically in fig. 5, show a gradual dropping off of the day maximum values from July 26 to July 30. From July 30 to August 3, with the exception of August 1, the graph of wilting is practically a straight line. On August 3 there is a marked increase, considering that the entire period has given a low index throughout from 0.15 to 0.20, or an increase of 33.3 per cent). Usually during the march of foliar transpiring power a drop is registered at the 14th or 15th hour. On July 27 there is an increase of the index from 0.13 to 0.18 (38.46 per cent) and on July 31 from 0.14 to 0.17 (21.43 per cent). On account of the comparatively small deviation between the maximum and minimum values throughout, the increase in the transpiring power of one-third on August 3 becomes more significant than the graph shows (fig. 5). The ratio between the day indices and night indices is presented as before, the readings of the gth and 21st hours being used. On July 24 the morning reading is 0.17; on the following days the average foliar indices of transpiring power for leaf Ib, areo.12,0.12, ©.10, 0.22, 0.25, 0.31, 0.26. The corresponding night values are 0.18, 0.19, 0.35, 0.34, 0.12, 0.14,0.26,0.13. The corresponding BOTANICAL GAZETTE [AUGUST Aug | Fic. 5 July 24,1916 5 5 4 2.00} 250) 150} 1 1918] BAKKE—WILTING 103 ratios between the day and night readings are respectively 0.95, 0.63, 0.34, 1.83, 1.79, 1.19, 2.00. The ratio between the index of foliar transpiring power for the gth and 21st hours on July 21-22 is 1.57. The ratio between the day maximum and the night mini- mum at that time is 2.11. Taking into consideration that the minimum for Id, during the march of foliar transpiring power on July 21-22 is 0.49, the maxi- mum values during the march of wilting are extremely low from the initial to the final point of wilting. On account of the drop in the maximum and the constant retentive character of the minimum, the index here is larger than recorded previously. The leaf was completely wilted on July 31. The day values for Ib, taken at the same time as before are Q.15, 0.20, 0.28, 0.13, 0.16, 0.14, 0.14, 0.21, 0.17, 0.15;. while the corresponding night values are 0.14, 0.21, 0.19, 0.20, 0.22, O.1I, 0.15, 0.10, 0.14, 0.12. The ratios between these two sets are, in order of their occurrence, 1.07, 0.95, 1.45, 0-65, 0.59, 1-45, 0.93, I.40, 1.50, 1.42. The ratio between the readings for the goth hour and the readings of the 21st hour on July 21-22 for Helian- thus leat Ib, is 0.98:0.51, or 1.92. The ratio between the maxi- mum and the minimum is 2.23. In none of these cases can the proportion be regarded as normal. This plant from the beginning is evidently in a partially wilted state. In comparing the march of foliar transpiring power Gaus the Process of wilting for the two series Ia and Ib, there would natu- rally be some variation. The wilting of series Ia extends over a longer period (to August 7), while that of Ib reaches its permanent wilting point on August 3. In both cases the older leaf wilts long before the younger leaf. However, leaf Ia has a greater range of foliar transpiring power. Leaf Ia, wilts before Ia,; likewise Ib, before Ib,. The ratio between the morning and the night readings of each day gives in the majority of cases a value that is less than the normal. For Ia all the results are either near 1.00 or below it except for the first. As the maximum value is above that of the usual minimum the result cannot be anything but normal. With Ia the ratio on July 31 is extremely high, probably being due to the 104 BOTANICAL GAZETTE [AUGUST extremely high evaporation. Why there should be such a decrease at the 21st hour is not known. The first ratio 1.74 is approxi- mately equal to the normal. The day reading is 0.68. The ratio on August 5 is 1.54, but the 9th hour reading gives a value that is much lower than the usual minimum. With a slight decrease in the minimum, the ratio between the two becomes greater than before. From these data on the basis of the ratio between day and night foliar transpiring power values it is evident that, if the ratio is to be used during the process of wilting, it can only be applicable when the maximum is greater than the usual minimum. Through- out the series of both Ia and Id the ratio does not deviate very far from unity, but in the formation of the ratio there is an evidently greater corresponding decrease in the day value as compared with the night reading. In both cases the extent of daily fluctuation for the younger leaves is very small after the first day. The rate of evaporation throughout fluctuated considerably, but is unusually high for the climate of Chicago. There is neverthe- less no close agreement between evaporation and foliar transpiring power during the march of wilting. Plants similar to the ones used in the experiment were treated like Ia and Ib and were placed in a moist chamber at their respective times of wilting. They behaved in a similar manner and failed to recover in the allotted time. Although the plants were watered at the same time with approximately the same amount of water, figs. 5 and 6 show indirectly that there was much difference in the soil moisture content. The plant designated as Id was larger than Ia, and would be expected to wilt first. This observation is borne out in the experiment. It is also evident from an examination of the two graphs that the soil of Ib was drier at the beginning than that of Ia, as the indices of foliar transpiring power are much smaller. STOMATAL DIFFUSION The index of foliar transpiring power in its very definition is associated with that of vapor tension. The decrease in the index of foliar transpiring power such as is present at night during the daily march represents a great force. A solution may also carry with it just as great a force. Livincston (27), com- 1918] BAKKE—WILTING 105 menting upon Fittinc’s (19) work upon the osmotic pressure present in desert plants, makes the statement that with the lowering of the vapor tension 10 per cent there is represented a pressure of 100 atmospheres. In an examination of the graph in table I there is a reduction in the index from 0.92 to 0.19 during the process of wilting. This therefore represents an approximate pressure of 800 atmospheres. For plant B there would be an approximate pressure of 700 atmospheres. Table V and fig. 4 give leaf Ia, as being able to withstand a pressure of 666 atmos- ~ pheres and leaf Ia, 860 atmospheres. Leaf Id,, with an index o. 10, Suggests a pressure as high as goo atmospheres. At that time the margin of the leaf was sufficiently dry so that the clip could not _ be used without causing injury. This status becomes all the more pertinent when it is compared with the data submitted by SHULL (41), where the force present in air dry seed (Xanthium) is equiva- lent to 1000 atmospheres. During the daily march of foliar transpiring power there is usually considerable variation (figs. 2, 3), even when a plant is supplied with an optimum amount of water. The sudden rise in the foliar transpiring power immediately after sunrise, as set forth by Bakke and Livrincston, gives credence to the view that the stomata open quickly at this time. In using the porometer and Standardized cobalt paper squares simultaneously, TRELEASE and Livincston (42) find that during the daily march there is considerable agreement between the porometer readings and the readings of the foliar transpiring power by the method of stand- ardized cobalt chloride paper. From the results obtained in their investigation they concluded that the porometer gives readings which show the extent of stomatal diffusion. DARWIN (12, 13), using the horn hygroscope and the tempera- ture method, has shown that during wilting there is a temporary Opening of the stomata. Darwin and Prrrz (14), using the porometer, have demonstrated that a similar condition is present during wilting. Lamztow and Knicut (26) in their work upon stomatal behavior during wilting, where they employ a recording porometer, have confirmed the results of Darwin and PERTz, in that the stomata open temporarily during wilting. For Phaseolus 106 BOTANICAL GAZETTE [AUGUST vulgaris the maximum diffusion occurred about 5 minutes after the leaf was severed from the stem, while in the case of the thick leaf of Prunus Laurocerasus nearly 20 minutes elapsed before the maximum stomatal opening was reached. KAMERLING (23) found that when Rhipsalis cassytha had lost 1 to 4 per cent of its normal water supply, the amount of transpiration per unit time increased, and later when the loss in weight had reached a certain point, varying from 6 to ro per cent, the transpiration again diminished. KAMER- LING is of the opinion that the increase in transpiration is due to the opening of the stomata. Ltioyp (31), on the other hand, failed to find this temporary opening. The evidences at hand support the conclusion that the stomata open for a short time during wilting. The time is short, however, and there is no evidence that the stomata ordinarily open up during the early stages of wilting and continue to be open until the plant has attained its permanent wilting point. This point is important in connection with argument presented for the break in the water columns. In order to prove that the stomata open only during the early stages of wilting, the porometer as modified and used by KNIGHT (24) was resorted to. The plant was attached to the aspirator and allowed to remain until the leaves were partially wilted. Tests were made upon plants grown in the greenhouse and later trans- ferred to the laboratory and plants which had been grown con- tinuously in the greenhouse. The plants in the laboratory were kept for 5 days before experimentation was begun. The tests of this series were begun on November 28, 1917, and continued until December 3, 1917; readings were made at three different times of the day. This conforms with the plan adopted in making the readings of the foliar transpiring power. Evaporation was recorded by means of a standardized cylindrical form of atmom- eter of the Livingston type. The data are given in table VI. In an examination of the data presented in table VII, readings were not taken until the 16th hour on November 29, 3 days after all watering had ceased. The time elapsing between two suc- cessive bubbles, as ascertained by means of a stop watch, was found to be 160 seconds. At the 11th hour it took 140 seconds, while 1918] BAKKE—WILTING 107 on December 1 at 10:30 it took 246 seconds. From that time until December 3 there was not much variation either at night or during the day. TABLE VII POROMOTER READINGS DURING PROGRESS OF WILTING OF A Helianthus annuus PLANT GROWING IN LABORATORY Rate cf evapora- | Rate of flow, time 7 Pees tion pied standard-jinterval in seconds deny ap Hen scot between succes- sive bubbles Proves. $9 26200 5 oho oo osc cae ee OG he rk ins ae Pe Oe eh WO Bien \15:30 Pde ee es ee 29 16:00 0.72 160 17: 12 140 30) 14:30 0.79 150 2 1.34 146 {10:30. 0.97 246 December “ae 1.11 266 20: 0.54 259 10:30 i760 276 24 14:30 E.91 260 20: I.41 262 3: 10:30 I.70 269 The data of the series grown continually in the greenhouse, where the evaporation was very low, are given in table VIII. These results, which were obtained at regular intervals on 5 Successive days (November 29 to December 3, 1917), do not show any marked differences. Considering the experimental error which would be present, there is not sufficient difference in any one case to indicate that stomatal movement was present. The plants used in this series were not watered for 3 days before the beginning of the experiment. The stomatal diffusion as cieanived by the porometer was also determined for plants which were grown for the same period, but were not subjected to any extended period of wilting. The general average for the stomatal diffusion as represented by the time interval between successive bubbles of the air intake tube of a porometer was found to be 1o5 seconds. No attempt was 108 BOTANICAL GAZETTE [AUGUST made to ascertain whether or not the stomata were partially closed or whether there was an increased opening after a period of 2 hours, as has been shown for certain plants by Irjrn (20). At any rate, the time interval in the case of intensely wilted Helianthus plants is much smaller. Even if the stomata should be partially TABLE VIII POROMETER READINGS TAKEN OF A WILTING stag PLANT GROWING WHERE EVAPORATING POWE OF LOW AIR IS Rate of evapora- | Rate of flow, tim pret oe pr ea jtion from tia antl gg real in rhino me gro ey between succes- r hour sive F bubbles December 31 16:00. 0200 ee Ui bd a le CAG hee ee cae Oa 12) 16:30... i eee Por eee 10:30... Co ©. MIDIS Tere te pea are eee 13} 14:30.. OOG re Pires een 20:00.. ORG ee eae 10:30 0.37 257 I4} 14:30 0.66 360 20:00 0.34 352 10:30. ©.30 318 I5414:00. 0.36 300 20:00, ©.40 335 16:30: 0.43 357 164 14:30.. 0.43 348 20:00. 0.49 343 10:30. 0.40 341 17} 14:30. 0.44 365 20:00 0.37 343 10 0.34 308 18; 14:30 0.44 389 20 0.39 360 closed, the results obtained from tables VII and VIII show that during the march of wilting, where the plant acquires its perma- nent wilting point, the stomatal opening does not enter in to affect the diffusion or transpirational water loss. This statement is in agreement with that of DARwin, that when the transpiration is high and the supply water insufficient the lack of water is a more important factor than stomatal changes. It would be extremely advantageous, however, to have the stomatal movement question 1918] BAKKE—WILTING 109 settled. Regarding the stomatal diffusion as a minor factor during intense wilting, the problem resolves itself to the point where the resistance to the passage is considered. From the data given in this paper and in a previous publication the resistance is exceed- ingly great. This will give further information, therefore, upon the strength of the evaporating force and that of cohesion. Discussion In comparing the results obtained during the summer of 1915 with those of 1916, considerable additional evidence is set forth which substantiates the argument advanced by BAKKE that wilting occurs at a definite point and is readily determined by the use of standardized hygrometric paper. In the series of 1915 the average of 3 leaves were used in plant Ia and 2 leaves for plant Ib. No effort was made during the 1915 season to obtain the difference in the time of wilting for leaves of different ages. The difference, however, was probably very slight, as the evaporation was exceed- ingly low. At no time during the entire run was the evaporation as high as 0.7 cc. per hour, and usually it was below 0.5 cc. per hour. The temperature of the greenhouse was seldom over 28° C. In contrast, the evaporation during the season of 1916 was high and during the time the experiments were being performed was exceedingly uniform. It may be added that during the progress of the experiment no rain fell. It would then have been preferable to have run the experiments outside, but in the climate of Chicago it is rather difficult to obtain such a continued period of clear weather. The usual feature will then be a low evaporation at night, a higher one during the forenoon, and the maximum at the 14thhour. The high evaporation rate on July 31 is not explainable. It may be well to remark, however, that on July 30 the tempera- ture in the greenhouse was 41.2° C. and at the first hour of July 31 it was 27° C., almost the maximum of the previous year. It is again brought out that for the 1915 and 1916 series a point 1s reached where the foliar transpiring power shows very little fluctuation. In the cases presented, this point can be represented graphically by a line that is almost straight. The ratio values are hot far above unity in the majority of cases, and sometimes are IIo BOTANICAL GAZETTE [AUGUST even lower. It is noted also that the time element of this period varies greatly in the two seasons. In 1915 it is comparatively short, while for both series in 1916 it is extended over a consider- able period. It has been proved by the work of SHREVE (34) that plants grown under different environment not only have different anatomical characters but also have a different rate of transpira- tion. On the basis therefore of a possible change as a result of environment, it can safely be asserted that this is the reason for the short span in 1915 and the long one in 1916. Why or how the plant establishes such an apparent equilibrium cannot be stated. This equilibrium represents the greatest force or tension which can be applied before a plant assumes the condition of permanent wilting. A plant such as Aériplex will necessarily have an extended period when this equilibrium is maintained. The exact wilting will be when there is a serious rupture in the water columns. If this interpretation is correct, the 1915 and 1916 series should exhibit a difference in the foliar transpiring power values during the so-called equilibrium stage. It would be expected that the 1915 series would have a higher minimum than the 1916 series. This is evident, for in 1915 the lowest point reached at any time is never below 0.15, while for the 1916 series it is as low as 0.09 in one case and 0.10 in the other. There would thus seem to be a direct relation between the time of the equilibrium, the lowest point in the index of foliar transpiring power, and the evaporating power of the environment. The point at which wilting occurs is defi- nitely marked out. This point appears graphically to better advantage for the plants of 1915 than for those of 1916; but the plants of 1915 were larger and were grown in smaller containers than those of the following year. For the series of 1915 the permanent wilting occurs on August 21, while for series la (1916) the wilting occurs on August 7, and for series Ib of the same year the wilting occurs on August 4. In this study the same conception of wilting is advanced as before. The present study is really more or less of an elaboration of the former. It is assumed here that Drxon’s (16, 17, 18) con- 1918] BAKKE—WILTING tit ception of continuous water columns is in force. When the force of evaporation becomes sufficient to cause a serious rupture of these water columns, then the plant wilts. Just to what extent a serious rupture can be regarded cannot be stated, but it must be greater than the force of cohesion which holds the water particles together. The extent of this cohesion force has been sufficiently presented and advanced by Drxon (16, 17, 18), RENNER (34, 35, 36), Ur- SPRUNG (43, 44, 45, 46, 47), and others (21, 32), and although the conclusions have been criticized by Jost (22), nevertheless they are substantiated. It is not the province of this article to enter into a critical discussion of these various papers. The approxi- mate point of permanent wilting is readily ascertained from the beginning by taking a series of readings of the foliar transpiring power of the plant in question. Care should be taken to obtain in the series the maximum and minimum. Although there is not any hard and fast relation between the maximum and the minimum, when the moisture in a soil has been reduced to the point where the maximum is below the normal minimum, at a time of the normal maximum, then the water content of that soil has attained what the writer designates as the critical content. From this point it is simply a question of time when the columns break. This then becomes a relatively simple matter. The readings giving the indices of foliar transpiring power taken at hourly intervals present a graph that is similar to graphs set forth previously. The maximum occurs at a time previous to the highest evaporation; the minimum generally occurs somewhere between the 18th hour and the 24th hour. There is a decided drop in the afternoon, which occurs at a time of day when evapora- tion is at its height or nearly so. There is a recovery that is also conspicuous. The cause for this resistance has been advanced by SHREVE (40) as being due to the imbibitional forces of the cell wall and of the colloids of the protoplasm. Although this feature has been noticed wherever the march of foliar transpiring power has been obtained, no one as yet has set forth any evidence as to the length of time necessary for recovery to take place. It is appar- ent that the recovery has been ‘complete before the time of the II2 BOTANICAL GAZETTE [AUGUST beginning of the next reading, which in this case is the next hour. A record of the foliar transpiring power at hourly intervals at Chicago gives results that are similar to those obtained for plants of the same species in southern Arizona. The series of 1916 show conclusively that the older leaves are the first to wilt. In an examination of series Ia the older set of leaves is almost completely dry at the time of the permanent wilting of the plant. On July 29 the edges of the leaf are dry, but at the same time there is a different form of response in the younger leaves, in that the apparent recovery occasioned at the time of permanent wilting does not present itself. The same situation is true for the series Ib, where the older leaves wilt on July 31. That the older leaves are the first to wilt has previously been determined by a number of investigations (15, 33). BAKKE and LIvINGSTON have presented evidence that there is considerable variation in the index of foliar transpiring power of young and old leaves. The fact that the younger leaves wilt later than the older leaves is not necessarily connected with the environment. This is true whether the evaporation is low or whether it is high. The production of the absciss-layer may at least be indirectly formed as a result. PRINGSHEIN (33) previously has shown that young leaves retain their freshness for a longer time than older ones. This he ascribes to a greater osmotic pressure. During the march of wilting it is also noticed that the foliar transpiring power index of the older leaves is always higher, at least than that of the leaves of the tip. The older leaves then give a higher foliar transpiring power throughout. There is also in evidence during the march of wilting not only a low index of foliar transpiring power, but also a gradual increase of the force in opposition to the passage of water. When for a short time there is an evident break or a serious rupture, there is a decrease in the resistance, but an equilibrium with the atmosphere is soon reached. The assumption that there is a temporary open- ing of the stomata may be made at this point. Employing the porometer upon Helianthus plants placed in an environment of high evaporating power and one of low evaporating power, the author failed to find that the stomata are concerned. 1918] BAKKE—WILTING 113 Summary 1. The transpiring power of plants as determined by standard- ized hygrometric paper gives an accurate knowledge of the internal water relations of a plant. The exact wilting point as determined by this method occurs when there is a serious rupture in the water columns. 2. During the daily march of foliar transpiring power obtained by making consecutive hourly readings for 24 hours, the maximum is attained at a time previous to the greatest evaporation. During the time of approximate maximum evaporation there is a marked fall in the foliar transpiring power index, followed shortly by a rise. The ratio between the maximum and the minimum is more or less definite, but not sufficiently so for the formation of any law. When the ratio is reduced to the point where it is in the neighborhood of unity, the plant is in a state of intense incipient drying. When the maximum value does not exceed the usual minimum, the plant is in a soil environment which is critical from the point of water supply, or almost at its wilting coefficient. It is then merely a question of time before the plant wilts. 3. Evaporation plays an important part in the experiment upon transpiration. A high evaporation gives an increased tran- spiring power value, but during the process of wilting the index of foliar transpiring power comes to be independent of evaporation. 4. During the process of the march of wilting an equilibrium point is reached where the indices of foliar transpiring power do not show much variation. It is suggested that the duration of the equilibrium gives a measure of the comparative drought resistance of different plants. Helianthus grown in 1915 during a rainy Season is different from Helianthus grown during 1916, when the ‘S€ason was unusually dry. The equilibrium period of 1915 was much shorter than for 1916. 5. There is a decided difference in the time at which permanent wilting occurs in old and young leaves. The older leaves will wilt long before the younger ones. The time interval varies according to age. 6. Stomatal movements or changes are not important factors when the plant is in an intense state of wilting. II4 BOTANICAL GAZETTE [AUGUST I wish to express my thanks to Dr. H. C. Cow.es and Dr. G. D. FULLER, University of ese who have aided me with suggestions and criticisms. Iowa STATE COLLEGE AmEs, Iowa LITERATURE CITED bal . Atway, A. J., Studies on the relation of the non-available water of the e. to the hygroscopic coefficient. Neb. Agric. Exp. Sta. Research Bull. 1913. a. BAKKE; A. 4... cnet on the transpiring power of plants as indicated by the method « ygrometric paper. Jour. Ecol. 2:145-173. IQT4. , The index of foliar transpiring power as an indicator of permanent wilting in plants. Bor. Gaz. 60:314~319. 1915. . Bakke, A. L., and Livincston, B. E., Further studies on the foliar tran- spiring power in plants. Physiol. Researches 2:51-71. 1916 Briccs, L. J., and SHantz, H. L., Application of wilting coefficient deter- minations in agronomic investigations. Proc. Amer. Soc. Agron. 3:250- 256. 1912. , The bea coefficient and its indirect determination. Bor. Gaz. a 20-37. /<.ne RT wilting coefficients for different plants. Bot. Gaz. baa ade 1912. , The wilting coefficient for different plants and its indirect deter- minations. U.S. Dept. Agric., Bur. Plant Ind. Bull. 230. 1912. " .: Die snes Welkungskoeffizienten verschiedener Pflanzen. Flora 105:224-240. 1 to. BURGERSTEIN, A., Die panate der Pflanzen. Jena. tor4. 11. CALDWELL, J. S., The relation of environmental conditions to the phe- nomenon of permanent wilting in plants. Physiol. Researches 1:1-56. IQI3. 12. Darwin, F., Observations on stomata. Phil. Trans. Roy. Soc. Lond. seein Gar. 1898. a self-recording method applied to the movements of stomata. Bor. Gaz. 37:81-103. 1904. 14. DARwIN, F., and Pertz, D. F. M., New method of estimating the aper- ture of stomata. Proc. Roy. Soc. Lond. B. 84:136-154. 1911 15. DetF, E. Marion, Transpiration in succulent plants. Ann. Botany 26: 4090-442. I9I2 16. Drxon, H. H. , Dranapiration and the rise of sap. Prog. Rei Bot. 3:1-66. 1909. > ¥ 1918] BAKKE—WILTING Its 17. 18. N an N ~I w w bs Dixon, H. H., Transpiration and the ascent of sapin plants. London. 1914. ———., Theoretisches und Experimentelles zur Kohiasiontheorie der Wasserbewegung. Jahrb. Wiss. Bot. 56:617-667. 19 OIs. . Fittinc, Hans, Die’ Wasserversorgung und die cnet Druck- verhaltnisse der Wiistenpflanzen. Zeitschr. Bot. 3 3: 200-75. - Injin,, W. S., Die Probleme des vergleichender Studiums ne Pilandin- Sanat Beih. Bot. Centralbl. 32:15-35. 1914 OLLE, Hans, Untersuchungen oo Welken, Vectiarinns und Wieder- a Een Flora 8:73-126. 1 - Jost, L., Versuche iiber die Wasecricteimg in der Pflanze. Zeitschrift fiir Botanik 8:1-55. 1916. - Kamertine, Z., De Reguliering van de verdamping bij Viscum album in bij Rhipsalis cassytha. Ver. K. Ak. Ned. Amsterdam 22:821-835. Igl4. - Kyicut, R. C., A convenient modification of the porometer. New Phytol. 14:212-216. IgI5. , On the use of the porometer in stomatal investigation. Ann. Botany 30:57-76. 19016 - Larpiow, C. G. P., and Knicurt, R. C., A description of a recording porom- ilti Bo eter and a ro on stomatal behavior during wilting. Ann. tany 747-56. : ine INGSTON, ne E., The relation of the osmotic pressure of the cell sap in plants to arid habitats, Plant World 14:153-164. 1911. ———, The resistance offered by leaves to transpirational water loss. Plant World 15:1-35. 1913. LivinesTon, B. E., and Brown, W. H., Relation of the daily march of transpiration to variation in the water content of leaves. Bor. Gaz. 53: 300-330. Ig12. - Livincston, B. E., and SHReve, Epirx B., Improvements in the method for determining the transpiring power of plant surfaces by hygrometric paper. Plant World 19:287-309. 191 - Lioyn, F. E., The physiology of somata Carnegie Inst. Wash. Publ. 82. 908. Washington. I Norpuausen, M., Uber die Saugkraft transpirierender Sprosse. Ber. Deutsch. Bot. Gesell. 34:619-639. 1916. PRINGSHEIN, E., Wasserbewegung und Turgorregulation in welkenden Pflanzen. Jali: Wiss. Bot. 43:89~144. 1906. Renner, O., Zur Physik der Transpiration. Ber. Deutsch. Bot. Gesell. 29:125-132. IgII : Exnerimentelle Beitrage sur Kenntnis der Wasserbewegung. Flora 103:171-247. 1911 , Versuche zur Mechantk der Wasserversorgung. Der. Druck in den Lacombe von Freilandpflanzen (Vorlaufige Mitteilung). Ber. Deutsch. Bot. Gesell. 30:576-641. 1912 BOTANICAL GAZETTE [AUGUST . SHIVE, J. W., and Livincston, B. E., The relation of atmospheric evaporat- ing power to soil moisture content at permanent wilting in plants. Plant World 17:81-121. 1914. . SHIVE, J. W., and JArtin, Wm. H., The effect of surface films of Bordeaux mixture on the foliar transpiring power in plants. Plant World 20:67-86. 1917. . SHREVE, Epitu B., The daily march of transpiration in a desert perennial. Carnegie Inst. Wash. Pub. 149. Washington. 1914. , An analysis of the cause of variation in the transpiring power of cacti, Physiol. Researches 2: 73-12 IQ . SHULL, a ., Measurement of the se forte’ in soils. Bort. GAz. 6221-31. ee Sik F., and Livrncston, B. E., The daily march of tran- spiring power as inidicated by the porometer and by standardized hygro- metric paper. Jour. Ecol. 4:1-14. 1916 . UrspruNG, A., Zur demonstration der Fliissigkeitskohision. Ber. Deutsch. Bot. Gesell. 31:388-400. 1913. , Uber die Bedeutung der Kohision fiir das saffsteigen. Ber. Deutsch. Bot. Gesell. 31: 401-412. 10913. , Filtration und Hebungskraft. Ber. Deutsch. Bot. Gesell. 33: I12-117. IQI5. , Zweiter Beitrage zur Demonstration der Fliissigkeitskohasion. Ber. Deutsch. Bot. Gesell. 33:253-265. 1915. , Uber die Kohiision der Wassers in Farnannulus. Ber. Deutsch. Bot. Gesell. 33:153-162. 1915. NOTES ON AMERICAN WILLOWS I. THE SPECIES RELATED TO SALIX ARCTICA PALL. CAMILLO SCHNEIDER The more I advance in the study of American willows the more I realize that every species and form needs thorough investigation, and that even the most common and apparently best known species are far from being well understood in their variation and relation- ship to other forms. It will take two years more before I shall be sufficiently acquainted with all the American species hitherto described and preserved in the leading herbaria of this country to undertake their final arrangement in a monograph. At the advice of Professor SARGENT, therefore, I shall prepare, in the course of my studies, a series' of papers dealing with those species and forms which I have had an‘opportunity to investigate as thoroughly as can be done with herbarium material only. In November 1917 I commenced an investigation of the willows treated by RYDBERG in his paper entitled ‘‘Caespitose willows of Arctic America and the Rocky Mountains” (Bull. N.Y. Bot. Gard. 1:257. 1899). I re- ceived from the New York Botanical Garden and from the Her- barium of the Geological Survey of Canada at Ottawa the material that RypDBERG had before him. Besides this I had at my disposal the splendid collections of the Gray Herbarium, the Missouri Botanical Garden, and of course of the Arnold Arbore- tum. Furthermore, I was able to see the Labrador material of the Bebb Herbarium, now in the Herbarium of the Field Museum at Chicago, and also very interesting collections made in Labrador, Greenland, and Alaska from the Herbarium of Cornell University. I take this opportunity to offer my best thanks to the gentlemen in charge of all these herbaria. Unfortunately I have not been able to look over the rich collections of the U.S. National Herbarium at Washington. It would have been of the greatest advantage if I could have seen the material collected by LuNpstROM and used in his ‘‘Kritische * For my first paper see Bor. Gaz. 65:1-41. 1918. ex7] , [Botanical Gazette, vol. 66 118 BOTANICAL GAZETTE [AUGUST Bemerkungen iiber die Weiden Nowaja Semljas und ihren geneti- schen Zusammenhang” (Act. Reg. Soc. Sci. Upsala III. 1877), for * without comparing a good series of specimens of S. arctica and. S. glauca from Northern Asia and Europe it is difficult to get a cor- rect understanding of those forms from North America; but at present it is impossible for me to consult any European herbarium. In this article I shall try to present a critical account. of the species related to S. arctica Pall.; in a following paper I intend to discuss S. glauca L. and the species related to it; while in a third paper a key. will be given containing the species treated in the first two papers and also those of sections RETICULATAE and HERBACEAE (RETUSAE), together with a few other species the systematic posi- tion of which is not yet fully understood, but which are best placed near one or the other of the groups in question. In this key it is intended to indicate briefly the main characters of the species, because full descriptions cannot be given here except of the new species and varieties I wish to propose. The history of most of the species must be explained, I am sorry to say, at considerable length, since otherwise it would be impossible to account for the fact that so many well marked types have been interpreted so differently by various authors. I commence with S. arctica Pall., which is the nucleus of the group of forms I shall try to elucidate. 1. S. ARCTICA Pall., Fl. Ross. 17:86. 1788.—Paxias described this species from the “ plaga arctica muscosa nuda secundum Sinum Obensem et versus glacialem Oceanum”’ in such an unmistakable manner that it could never have been misunderstood had not RosBert Brown, in 1819, proposed a new S. arctica, ignoring alto- gether the older name of Pattas. In Ross, Voy. Expl. Baffin’s Bay (appendix, p. 148, and ed. 2, 2:194, both in 1819), BRowN men- tioned only the name, and a description of his arctica was first given by RICHARDSON in FRANKLIN, Narr. Jour. Polar Sea, Bot. App. 752 (reprint, p. 24). 1823. In the same year Brown published his own description in Chloris Melvilliana, which was issued separately, while Capt. PArry’s Voyage, of which the Chloris is only a part (App. Suppl. pp. 259-305), did not appear until 1824; but in 1823 there also appeared a second edition of FRANKLIN’s book and 1918] SCHNEIDER—AMERICAN WILLOWS 119 RICHARDSON’s Appendix. Later, in the synonymy of S. anglorum Cham., I shall give the full and exact quotations of S. arctica Br. The earlier S. arctica Pall. has also been overlooked by Kocu (1828), who mentioned only Brown’s species. The first author who recognized the two discrepant arctica seems to have been MEYER (De Plant. Labrad. 32. 1830), where in a note to S. arctica Br. he says, “Quid est Salix arctica Pallas (florae rossicae II. pag.170 editionis minoris)? Nullibieam vel ut peculiarem speciem, vel ut synonymon apud botanicos memoratam inveni.” In 1831 CHAMIsso (Linnaea 6:541) proposed the name S. anglorum for S. arctica Br., non Pallas; see under S. anglorum. In 1832 TRAUTVETTER, in his valuable study ‘De Salicibus frigidis Kochii,” described the 3 following species: S. crassijulis Trev., S. diplodictya Trvt., and S. torulosa Led. Of these in 1833 (in LEDEBOouR, Fl. Alt. 4:283) he referred S. crassijulis and S. torulosa as synonyms of S. arctica Pall., which had not been men- tioned by him in 1832. In this year he described and figured only a S. arctica Br., which in 1833, however, he says is nothing but a synonym of S. glauca L. In Mrippenporrr, Reise Sib. 1°:27 (Florul. Taimyr.), TRAUTVETTER again changed his opinion, say- ing, “Sal. arcticam Pall. et Sal. arcticam R. Br. unam eandemque Speciem sistere opinor. Planta, quam in dissertatione de Salicibus frigidis N. 7. tab. VI. sub nomine Sal. arcticae R. Br. proposui, ad Sal. glaucam L. referenda est nec sistit veram Sal. arcticam R. Br., uti e descriptione cel. R. Brownii in Fl. Melv. l.c. elucet.” See also under S. anglorum. In 1849-51 LepEBour (Fl. Ross. 3:619) included under S. arctica Pall. Brown’s species as well as TRAUTVETTER’S 3 species of 1832, and also added to S. arctica such forms as var. minor (S. phlebophylla And.) and var. leiocarpa (S. rotundifolia Trev.). LEDEBOUR seems to have been the first author who mentions S. anglorum Cham. in the synonymy. In a strange way the forms related to S. arctica have been treated by ANDERSSON (DC. Prodr. 167: 285. 1868), who, in 1858, in his previous work on North American willows, only mentioned S. arctica Br. as a “species difficile sane definienda, quasi inter S. myrsinitidem et glaucam prorsus media et formas plures ambiguas 120 BOTANICAL GAZETEE [AUGUST amplectens.”’ ‘To those “‘formas ambiguas”’ belong the 3 varieties (subphylicifolia, subreticulata, and subpolaris) proposed by ANDERS- SON in 1858, which I have not yet been able to interpret correctly - owing to lack of the type material. In the Prodromus, ANDERSSON created a S. Pallasii with the var. crassijulis (Trev.) and var. diplodictya (Trvt.), and mentioned, strange to say, the type of PALLAs under the last variety, while he is using the name S. arctica Pall. to cover a multitude of forms including his var. nervosa, Brownei, groenlandica, petraea, and taimyrensis. He excluded from his S. arctica, therefore, the forms of the true S. arctica Pall., and combined under this name a series of very different things like S. altaica Ldstr. (recte S. torulosa Led.), S. anglorum Cham., S. groenlandica Ldstr., S. petrophila Rydbg., S. taimyrensis Trvt., and others. The first who attempted to clear up the Pallasii-arctica mixture of ANDERSSON was LuNDSTROM in 1877, in his interesting study previously mentioned. He confined S. arctica Pall. to its typical forms, and distinguished besides S. Brownei (And.) Ldstr., for which S. anglorum Cham. is the oldest name, S. groenlandica (And.) Ldstr., and S. altaica Ldstr., in which case he overlooked the priority of S. torulosa Led., a species founded on the same type. Lunpstr6m did not use CHAmIsso’s name because, following ANDERSSON, he referred S. anglorum to S. phlebophylla; see under S.anglorum. Another attempt to interpret properly S. arctica Pall. and S. arctica R. Br. was made by Bess (Bor. Gaz. 14:115. 1889), who, however, did not know Lunpstr6m’s work. Consequently he proposed another S. Brownii which, sensu stricto, corresponds with S. anglorum, a name likewise overlooked by BExBB, who refers some different forms to his Brownii. In 1899 Batt (Trans. Acad. Sci. St. Louis 9:89) mentioned that ‘the methods by which Professor ANDERSSON succeeded in greatly augmenting the then existing con- fusion in regard to S. arctica R. Br. and S. arctica Pall. have been exposed by Mr. Brss,” and stated that Bess had ignored the existence of LUNDsTROm’s earlier homonym; but BALL, in his turn, overlooked the name given by CHAmisso many years before. It was RypBERG who, in 1899, reinstated the name S. anglorum as the oldest correct name for S. arctica. Br., non Pall. 1918] SCHNEIDER—AMERICAN WILLOWS — 121 I have seen but one leaf of the type specimen of Pattas. It does not possess stomata in the epidermis of the upper surface, a character upon which I am inclined to lay considerable stress. It is true that by A. and E.-C. Camus (Class. Saules Europe 2:55. 1905) S. arctica is said to possess ‘“‘stomates ... assez nombreux” in the upper leaf epidermis, but judging by their synonymy these authors include under S. arctica so many widely different forms that they probably did not examine a true arctica at all. So far as I can see, this species is represented in the New World only in Alaska, the Yukon Territory, and the adjacent part of the northwest corner of British Columbia, and in the apparently well marked var. subcordata in southern British Columbia. I am not yet quite sure how far the range of S. arctica extends toward the east, but it seems not to cross 130° W. longitude except in the var. subcordata, of which the geographical distribution is not yet fully known. The specimen collected by BELL on Nottingham Island, Hudson Strait (no. 24623 O.? olim 18825), which is cited by RYDBERG under SS. arctica Pall., belongs certainly to S. anglorum. There seems to be no great difficulty in distinguishing typical forms of S. arctica from those of S. anglorum if one has well devel- oped specimens. Very often, however, it is necessary to deal with mere fragments, and in this case the best character seems to be furnished by the presence or absence of stomata in the upper leaf surface. While they are entirely lacking in what I take for typical S. arctica, they are more or less numerous in all the specimens I have seen of S. anglorum. Generally, S. arctica is a much more robust plant with larger leaves and catkins and thicker branchlets, but when we compare the shape and pubescence of the leaves and the different characters of the flowers and fruits it is rather difficult to express in words those signs that the eye can more or less easily perceive. The best description of the American form of S. arctica is given by Covitte in his excellent study of the “Willows of Alaska”’ (1901), to which is added a good plate. I shall say some- thing more about the differences between S. arctica and S. anglorum = citing herbarium specimens I use the same abbreviations as in my first paper; - Gaz. 65:9. 1918. There are to be added the following: C., Herb. Field Csliens Miscuin: Cor., Herb. Cornell University; O., Herb. Geol. Surv. Canada. 122 BOTANICAL GAZETTE [AUGUST under the latter species; otherwise I refer to the keys that will be given in my third paper. Regarding the variability of S. arctica, COvVILLE said: ‘The large number of specimens examined tends to confirm the idea that the extreme variation in the leaves is chiefly an individual characteristic and does not mark recognizable incipient species. The nearest approach I have found to a subspecific differentiation is in some of the specimens from the Pribilof and St. Matthew Islands in Bering Sea, and the Shumagin Islands. In these specimens the leaves are orbicular, or nearly so, and only about 2-3 cm. in diameter, while the catkins are shorter than usual, about 1.5~3.5 cm. in length.” - These forms represent ANDERSSON’s S. Pallasii a, crassijulis 3 obcordata (1868) (S. Pallasii var. obcordata Turner; S. arctica obcordata Rydb.), who also distinguished f. grandifolia and f. oblongata of his var. crasstjulis. The last two forms are, I believe, without any taxonomic value, while f. obcordata well deserves to be mentioned as a form or even as a variety. It differs chiefly in the characters mentioned by Covitte. In addition to the localities cited by this author, I saw specimens from the Yacutat Bay, Glacier Bay (Muir Glacier), and Unalaska which should be referred to var. obcordata (And.) Rydbg. The following extract from an account given by TuRNER (Contrib. Nat. Hist. Alaska 75. 1886) seems to me worth quoting. S. Pallasii Anders. var. obcordata Anders. This species of willow attains the largest size of any among the Aleutian Islands. The growth is exceedingly crooked, rarely straight for more than a foot, attaining a diameter of 2 to 3 inches, but often decayed within. In all the valleys and wider ravines this species is found in abundance. The roots form an intricate mass, often much exposed, and with the crooked branches and trunks form an impenetrable thicket of considerable area. ... . VemamInoF [a Russian traveler] states that in former years this willow grew to such a size in one of the ravines open- ing on the west side of Captain’s Harbor at Unalaska Island that the Russians and Aleuts procured sufficient of these ee to be used advantageously in making bidaras (open skin boats)... . . visited the locality to find traces of such former growth and found the a to be of but little better size than in other places near by. There is another form which has very glabrescent capsules and may be identical with S. arctica var. glabrata Trautvetter (Act. Hort. 1918] SCHNEIDER—AMERICAN WILLOWS 123 Petrop. 5:107. 1877), of which he remarks: ‘Solum modo ovariis et bracteis parce puberulis a var. typica recedit. Forsan S. arcticae proles hybrida.” Without having seen TRAUTVETTER’S type, which had been collected by CzEKANOWSKI and MUELLER “‘inter fl. Olenek et fl. Lena inferiorem, ad fl. Tyria in tundra,” I am not sure whether the following plants really represent TRAUTVETTER’S variety: Unalaska, Kiuliuk, September 30, 1871, M. W. Harrington (fr.; G.), Dutch Harbor, July 17, 1899, B. E. Fernow (f., fr.; Cor.), Kodiak Island, July 2-4, 1899, B. E. Fernow (f., m.; Cor.), and Yakutat Bay, Disenchantment Bay, August 13, 1892, F. Funston (no. 117 partim, fr.; Cor., M., N.). LepEBour (FI. Ross. 3:6109. 1849-51) has described a variety with entirely glabrous fruits under the name var. lejocarpa, the type having been collected by ERMAN in Kamchatka, in ‘‘ignivomo Schiwelutsch” (Shivelutch). This specimen, which I have not yet been able to compare, was men- tioned by CHamisso (Linnaea 6:541. 1840) as a form of S. arctica Pall., while ANDERSSON (1868) referred it to his S. Pallasii var. diplodictya. The true S. diplodictya Trautv. came from the “insula St. Laurent,’ and its main difference from typical S. arctica is, according to the author’s description and figure, the “folia . subtus pallidiora nec glauca nec glaucescentia, utrinque lucida.’’ I have seen no specimen with such leaves, but CovILLe states that “occasionally specimens are found which lack the glaucousness of the lower leaf surface, a character on which TRAUTVETTER based chiefly his separation of diplodictya.” Ryp- BERG, who kept diplodictya as a species, interpreted it in a very different way, and referred to it certain forms of which I shall speak under S. ovalifolia. The var. subcordata previously mentioned from southern British Columbia is a form that needs further investigation. It has been described by AnpERsson in Ofvers. K. Vet.-Acad. Férh. 15:128 (Bidr. Kanned. Nordam. Pilart.). 1858; in Proc. Amer. Acad. 4:69 (Sal. Bor.-Am. 24). 1858; in Walp., Ann. Bot. 5:754. 1858, from specimens collected by Drummonnp in the “ Rocky Mountains.” In 1890 Bess (Bor. Gaz. 15:55) dealt with this rather obscure plant and stated that “the specimens from which the description of this Supposed new species was drawn are all attached to a single sheet 124 BOTANICAL GAZETTE [AUGUST in the Kew herbarium; they belong to three distinct and well known species.” BrEBB had received, through BAKER, nothing but a drawing and copies of the labels and “‘a few fragments, a capsule or two, to show minute characters.’”’ Through the kindness of Sir DAvip PRAIN the Arnold Arboretum has received (together with a series of photographs of other Salix types of the Hookerian her- barium) an excellent photograph of the type sheet of S. subcordata and also fragments of leaves and flowers. According to this material the fact stands as follows. In the upper left corner there are ‘‘two large specimens of S. arctica Pall.’’ which Bess believed had been labeled ‘‘almost certainly by some mistake”’ as “‘from the Rocky Mts. coll. Drummond” because, as BEBB had explained before, ‘‘nothing approximating in character to S. subcordata And. has been found.” There are before me, however, the excellent specimens mentioned later from the Chilliwack Valley, which, in my opinion, are identical with DRumMoND’s plant. Where this col- lector obtained his material I cannot ascertain, and so far as I know, he did not collect in this part of British Columbia. Beneath these two “‘arctica’’ specimens there are 3 (not 2, as BreBB stated) pieces, of which the one in the left corner is sterile, while the middle one bears female and the right one male flowers. Of both of them the Arboretum received fragments which show that they represent the same species as the older branchlets above them. BEss refers those flowering branchlets to “S. cordifolia Hook.,”’ and he is right in so far as HOOKER included in his cordifolia those Rocky Mountain forms. But ANDERSSON separated in 1858 just those western forms as S. subcordata from the eastern ones, which he then named S. alpestris americana (S. cordifolia Hook., pro parte). The most critical part of the type sheet is the two sterile right hand branchlets which BEBB stated to be “two stunted specimens of S. adenophylla, leaves only, habitat not given.’ To those branchlets refers Dr. BARRATT’s label: “no. 92, S. cordifolia 8 serrulata.” Of those pieces the Arnold Arboretum did not receive fragments, but, so far as I can judge by the photograph and by the corresponding number in Herb. N., they do not at all belong to S. adenophylla sensu BrsB (S. syrticola Fern.), but seem to represent a form of S. Barclayi, which grows together with arctica subcordata, 1918] SCHNEIDER—AMERICAN WILLOWS 125 at least in the Chilliwack Valley. This fact explains the confusion of the two species, and there is to me no great ‘“‘mystery how they came to be placed together on the same Kew sheet.” It is, how- ever, “more inexplicable how so critical a salicologist as ANDERS- SON should have been misled into combining the characters of the two in his S. subcordata.”’ I have also seen the “corresponding num- bers of the Hooker, BARRATT, and Torrey distribution in the Torrey Herbarium” mentioned by Bess. There are 3 sheets before me. One contains a large leaf and 2 sterile young branch- lets, and all 3 pieces belong to var. subcordata. This sheet bears the. following 2 labels: “No. 90. Herb. H.B. [&T., crossed out; instead of it is written beneath “‘fig.’’] S. obovata var. glabra,’ and “88 Barratt, Rocky Mts. Ament leafy at the base about 4 leaves— Smooth and paler beneath.” Underneath the big leaf BEBB, in 1887, has written “S. crassijulis Trev.?, S. subcordata And. in part.” The second sheet contains two flowering branchlets and bears the label ‘“No. 89 Herb. H.B. & T. Rocky Mts.” as well as the statement in BEBB’s handwriting “.S. swbcordata And. in part.” These flowering branchlets are identical with those in Herb. Kew. The third sheet bears the Barclayi form previously mentioned. ANDERSSON apparently had no clear idea of his S. subcordata; in 1858 he stated, “Quoad habitum quasi hybrida a S. cordata (cujus folia habet sed breviora) et S. glauca (amenta!).” In 1868 he referred to it some more specimens collected by BouRGEAU and DE La Pytatr, which I have not yet seen. The material before me looks very much like other robust specimens of S. arctica, but the leaves possess some stomata in the upper surface, at least along the main nerves. I think, therefore, that it is best to keep these forms as a variety of S. arctica, and I use ANDERSSON’S name. S. ARCTICA var. subcordata (And.) nov. var. seems to differ from typical arctica chiefly by the following characters: foliis maximis obovato-ellipticis ad 8:6 cm. vel obovali-oblongis ad 722.5 cm. vel ellipticis ovali-ellipticisve ad 8.5:5 cm. magnis > RypBErG (Fl. Rocky Mts. 167. 1917) uses the name in a different sense. I am not yet quite sure what form is meant by him. 126 BOTANICAL GAZETTE [AUGUST superne stomatiferis; amentis (saltem fructiferis) permagna ad tr cm. longis et 1.5 cm. crassis. : As already stated, the exact locality where DRuMMonpD collected the type is unknown to me. The other material came from British Columbia: Chilliwack Valley, between latitude 49°-49° 10’ and longitude 121° 25’~122°, 1650 m., August 29, 1901, J. M. Macoun (no. 26909 O., fr.; G., N.); Selese Mt., 1290 m., July 25, 1906, W. Spreadborough (no. 79556 O., fr.; Cor., N.; 79557 O., m.; Cor., N.; ‘70558 O,, 4.,.ir.3 Cor: 9959 0., m., fr.;-Cor.); Skeena River, Hazelton Mountains, July 13, 1917, J. M. Macoun (no. 95405 O., f.). There is also a forma incerta foliis oblongo-ellipticis utrinque acuminatis, collected by G. E. Cooley, in Juneau, Alaska, above Silver Bow Basin, August 6, 1891 (m.; G., N.), which has been referred by RyDBERG to S. anglorum. In my opinion it has nothing to do with that species, but may represent a special form of S. arctica, under which species it is cited by CovILte. 2. S. ANGLOoRUM Chamisso in Linnaea 6:541. 1831, exclud. specim. citat.—S. arctica R. Brown in Ross, Voy. Expl. Baffin’s Bay, app. p. 143. 1819, and ed. 2.2:194. 1819, nomen nudum, non Pallas; Chloris Melv. 24. 1823; Capt. Parry’s Voy. App. Suppl. p. 282. 1824; Richardson in Franklin, Narr. Jour. Polar Sea 752 (reprint 24). 1823; ed. 2.765 (reprint 37). 1823.—S. arctica B Brownei Andersson in DC. Prodr. 167:286. 1868, pro parte.— S. Brownei Lundstrém in Nova Act. Reg. Soc. Sci. Upsala III. 1877. 37, pro parte max.—S. Brownii Bebb in Bor. Gaz. 14:115. 1889, pro parte max. As already stated under S. arctica Pall., the existence of this. previous name had apparently been overlooked by BRowN in establishing his new arctica, of which RicHarDson was the first to give a description. It may be that this diagnosis was prepared by Brown, because RICHARDSON in his preface expressly acknowl- edges the great assistance BROWN gave him, but the first edition of RICHARDSON’s Botanical Appendix appeared shortly before the Chloris Melvilliana, in which BRown published an excellent descrip- tion of his species. When in 1831 CHAmisso changed BRrowN’s 1918] SCHNEIDER—AMERICAN WILLOWS : 127 name to S. anglorum after having given a good account of S. arctica Pall., non Br., he said nothing but the following: Salix anglorum N.—S. arctica R. Brown ex Ed. Nees 1. p. 406, supl. to the append. of Cap. Parry’s Voy. p. 282. E. Meyer Lab. p. 32 (non Pallas).— Insula et sinus Sti. Laurentii—Ex insula Chamissonis, capsulis maturis vetustate calvescentibus. These specimens, however, do not belong to S. anglorum, but what in 1839 Hooker (FI. Bor.-Am. 2:153). referred to S. retusa. Hooker, therefore, quotes S. anglorum in his synonymy of this species, for which ANDERSSON (1858) proposed the name S. (retusa*) phlebophylla, and made it a species (S. phlebophylla) in the Prodromus (1868). Here he also quotes S. anglorum in the ' synonymy, having in 1858 ignored entirely CHAMIsso’s name, which has been used by some later authors for phlebophylla instead of anglorum. RYDBERG (1899) was the first to state that CHAMIsso’s name ‘“‘must be regarded as equivalent”’ to S. arctica Br. What, however, is the typical S. arctica Brown? It was first collected by Ross during his exploration of the “Baffin Bay, Lat. 7° 30’ to 76° 12’ on the east side, or at Possession Bay, Lat. 73, on the west side.’”” RICHARDSON probably based his descrip- tion chiefly on his own plants collected on “barren grounds from Point Lake to Arctic Sea” (or, as the explanation in ed. 2 runs, on “barren grounds from Lat. 64° to the Arctic Sea, in Lat. 60°”); while Brown, besides the plants of Ross, mentioned those of Parry’s Expedition from Melville Island, Winter Harbor. I have not yet seen a type specimen, but RicHARDSON’s and BRown’s descriptions are sufficient to furnish us with the following char- acters: Frutex depressus. Rami, beaker 3 floriferi omnes et steri- lium nonnulli adscendentes, adulti glabri. Folia elliptico-obovata vel obovata, integerrima, novella pilis sericeis vestita, adulta utrinque glabra, venis subtus parum eminentibus, venulis ana- stomosantibus. Amenta utriusque sexus.ramos brevissimos foliatos terminantia. Squamae orbiculato-obovatae, saepe retusae, fusco- nigricantes pilis sericeis vestitae. Mascula 8-10 lin. longa, densa. Stamina 2, filamentis distinctis, antheris purpureis. Glandulae 128 BOTANICAL GAZETTE [AUGUST duae. Ovaria sessilia vel brevissime pedicellata, dense griseo- tomentosa. Stylus longitudine varians, nunc stigma aequans, nunc fere dimidio brevior. Glandula unica. Judging by these characters, there seems no doubt what form must be taken for the true S. arctica Br., that is, S. anglorum Cham. According to TRAUTVETTER (see under S. arctica Pall.) there have been distributed by HooKER specimens under the name of S. arctica Br. which do not belong to this species, and TRAuT- VETTER (1847) says that his arctica of 1832 (t. VI.) is not identical with Brown’s plant. However, so far as I can judge by TRAUT- VETTER’S diagnosis and figure, I believe that he had the true S. anglorum before him. Of course, only an inspection of his type can make a final decision regarding its identity possible. Of ANDERSSON’S treatment of S. arctica Br. I have already spoken. Lunpstr6M, who apparently misinterpreted the name anglorum, chose ANDERSSON’S (varietal) name Brownei for what he believed to be S. arctica Br. I strongly suspect that S. Brownei Ldstr. only partly belongs to S. anglorum, and an investigation of LUNDSTROM’S — specimens from Nowaja Semlja is needed to decide what he really understood by his S. Brownei. It seems to me most unlikely that the true S. anglorum should at all occur on Nowaja Semlja or in Arctic Asia or Europe; and the description given by LUNDSTROM, in my opinion, does not fit BRown’s species. There may be in Arctic Asia and Europe similar forms which, however, in reality belong to S. arctica Pall. Bess, as already stated, unfortunately did not know LUND- stROM’s work when proposing a new S. Brownii which comprised S. arctica And. (1868) “excl. var. nervosa.” He created a new mix- ture of forms, including S. groenlandica, S. petrophila, S. taimyrensts, and others. In April 1899 RypBERG said: ‘There is scarcely a species that has been so misunderstood as this [S. arctica Br.]. Even Mr. Brexs, who cleared up somewhat the discrepancy between S. arctica Pall. and S. arctica Br., had a very vague idea about the latter.” Rypserc himself did not interpret correctly BROWN’S species. He quotes as type “Franklin Expedition, Dr. Richard- son,” and cites a specimen of the “Herb. Hooker, Barratt, and Torrey, no. 93,” which I have before me and which bears the label 1918] SCHNEIDER—AMERICAN WILLOWS 129 “S. arctica, Fort Franklin, Mackenzie River.’ It contains male and female branchlets with young flowers and very young, narrowly lanceolate, rather acute leaves. So far as I can judge by the thinly pubescent and distinctly pediceled ovaries, by the oblong bracts, and by the absence of a dorsal gland in the male flowers, the speci- men does not belong to S. anglorum, but may probably be referable to S. groenlandica Ldstr. Furthermore, RYDBERG states that his S. anglorum “‘is characterized by . . . . the exceedingly large cat- kins, which are rather loosely flowered below, and the large conic capsule, which is only moderately hairy.’”’ If we compare this statement and the specimens cited by RypDBERG, his misinterpre- tation of BRown’s species is evident. RypBERG refers to his S. anglorum mostly specimens that in reality belong to S. groenlandica, about which species he certainly had a very wrong idea. - Almost simultaneously with RypBERG (May 1899), BALL pub- lished a statement regarding S. arctica and Begs’s treatment of this species. He knew Lunpstr6m’s study, but overlooked S. anglo- rum Cham.; he said, however, ‘‘I shall not rename the plant now, for I believe the name which has been in use for 80 years (S. arctica R. Br.) can yet do duty until both the numerous variations and the synonymy have been given careful study.” In Brirron and Brown’s Ill. Fl. (ed. 2. 1:605, fig. 1489. 1913) the name S. an- glorum is applied to forms from ‘‘Labrador to Alaska, and-in the Rocky Mountains to Colorado’ in a way I do not understand. Judging by the ample material before me, S. anglorum seems more variable than S. arctica. The habitat of the northeast Ameri- can plant ranges from Northwest Greenland (about Disco Island) and Labrador (where it apparently does not occur south of the 55th parallel) through northern Ungava along the Hudson Strait and the northern shores of the Hudson Bay to the Franklin Bay, reaching, as it seems, its most western point at Cape Bathurst and not ranging beyond 130° W. longitude. There are some forms collected on Herschel Island (coast of Yukon Territory) which might be taken for S. anglorum, but on account of the absence of stomata in the upper surface of the leaves I refer them to S. arctica. Between 130° and 140° W. longitude there may be the meeting ground for the 2 Species, and we need much more and well collected material from 130 BOTANICAL GAZETTE [AUGUST there to get a correct conception of the relationship of those arctic forms. I am not fully convinced that the presence of stomata in the leaf surface of S. anglorum and their absence in S. arctica typica can be regarded as a decisive character in distinguishing certain similar forms, but I think this specific character is of great taxo- nomic value at least in several species. A. and E.-C. Camus lay much stress upon this character in establishing their systematic arrangement according to anatomical features, and an excellent observer like the well known dendrologist E. KOEHNE was always inclined to pay much attention to those characters. In studying willows we should bear in mind the following remarks of the dis- tinguished English salicologist, F. BUCHANAN WHITE (Jour. Linn. Soc. 27:346 [Rev. Brit. Willows]. 1890): Whilst all the parts of the plant are variable, some characters, on which a great deal of reliance has been placed, are so inconstant that they may, in many cases at least, be almost or quite ignored, though in other instances they are really of importance. Familiarity with the species can alone teach the student what are the points on which he can depend. At present it is impossible to interpret properly certain forms because we do not yet know the degree of variation of the species in question. There are, I am convinced, many hybrids, and the fact that has been recognized by all the leading salicologists in Europe “that willows hybridize with the greatest facility adds,” as WHITE (loc. cit., p. 340) says, “immeasurably to the intricacies of the study.” Here in America we are only just beginning to get a better understanding of the taxonomy, variation, and distribution of the numerous willows, and everyone who attempts to further our _knowledge of them ought to be lenient in his criticism of those interested in this study. It is not without hesitation that I propose the following varieties of S. anglorum, but I am encouraged by the fact that such a keen observer as Professor M. L. FERNALD, who has collected most of the material of the new forms and to whom I wish to express My gratitude, agrees with my treatment of them. S. ANGLORUM var. kophophylla,* nov. var.—Frutex prostratus ramis subterraneis ad ultra 1 cm. crassis, ramulis repentibus pl. m. 4 The name is derived from xw¢és, blunt. 1918] SCHNEIDER—AMERICAN WILLOWS I31 elongatis, fructiferis ut videtur tantum ascendentibus; ramuli novelli sparse, rarius subdensius pilosi, vulgo citissime glabres- _ centes, in sicco nigrescentes vel flavescentes vel hornotini autumno ut annotini purpurascentes, ad 2 mm. crassi, annotini biennesque purpurei badiive, interdum ut vetustiores pl. m. pruinosi, vetusti crassiores pl. m. nigrescentes. Gemmae ovatae, obtusae, glabrae, badiae, saepe leviter pruinosae, ad circ. 5 mm. longae, floriferae ut videtur. obovatae, obtusiores. Folia adulta satis chartacea, inferiora minora variabilia, superiora majora vulgo late ovalia, ovato-rotundata, obovata, late elliptica ad orbicularia, apice rotundata vel satis breviter acuta, interdum brevissime plicato- apiculata, basi late cuneata, rotundata ad subcordata, 1.5:1.2 vel 2.3:1.8 ad 3.5:2.5-2.8 cm. magna, interdum ovato-rotunda ad 3.5 cm. longa et 3 cm. lata, margine integerrima, rarius partim sparse subdenticulata, vulgo parce (juniora densius) ciliata, superne ut videtur tantum novella pl. m. sparse villosula et in costa pilosula, cito glabra, saturate et vivide viridia, stomatifera, costa sub- impressa nervis lateralibus subprominulis et etiam graciliter reticu- lata, subtus valde discoloria, glaucescentia, pruinosa, initio magis quam superne sericeo-villosa sed etiam (infimis minoribus exceptis) cito glabra, costa nervisque primariis utrinque 5-8 pl. m. flaves- centibus vel brunnescentibus elevato-nervata et graciliter sed distincte reticulata. Petioli longitudine satis variabiles, superne sulcati, initio pl. m. pilosi, dein glabri, 2-14 mm. longi. Stipulae nullae vel raro evolutae, lineari-lanceolatae, glanduloso-denticulatae, subglabrae, visae vix ultra 2 mm. longae. Amenta satis serotina, ramulos foliatos o.5—2(-2.5) cm. longos pl. m. pilosos terminatia, cylindrica, rhachi villosa; mascula (no. 3232) 1-2 cm. longa et circ. 8 mm. crassa; bracteae obovato-oblongae ad late obovatae, apice obtusae vel retusae, omnino fuscae vel apicem versus atrae (in vivo pi. m. purpurascentes?), utrinque satis longe sericeo-pilosae; stamina 2; filamenta libera, glabra, bracteis demum duplo longiora; antherae parvae, ellipsoideae, ut videtur violaceae; glandulae 2 (vel interdum 1), ventralis ovato-rectangularis, truncata, integra (semper ?), bractea duplo brevior, dorsalis (3232) duplo minor et angustior (in no. 510 nulla); amenta feminea sub anthesi ut vide- tur circ. 1-1.5 cm. longa et 0.6 cm. crassa, satis densiflora, fructifera 132 BOTANICAL GAZETTE [AUGUST ad 3.5:1.4 cm. magna; bracteae ut in floribus masculis; ovaria sub anthesi ovoideo-oblonga ellipticave, sessilia vel subsessilia, albo- vel griseo-villoso-tomentosa; styli distincti, vulgo apice bifidi, rarius subbipartiti, stigmatibus oblongis bifidis 2-23plo longiores; glandula 1 ventralis, anguste ovato-conica, truncata, integra vel ut videtur pleraque bifida bipartitave, bractea subduplo brevior fructus elliptico-conici, circ. 7 mm. longi, fere sessiles, ut ovaria vel minus dense pilosa (interdum in forma porro observanda [nos. 61 et 62] fere glabri), valvis apertis paullo recurvatis. TYPE LOCALITY.—Western New Foundland, Bay of Islands, northeastern region of the Blomidon Mountain _RANGE.—Western New ocastlinid: Bay of Islands and Bonne Bay, and western Gaspé Peninsula, Mt. Albert SPECIMENS EXAMINED.—Western New Foundland, northeastern region of ~ the Blomidon (“‘Blow-me-down”’) Mountains, serpentine tableland, alt. about 550m., July 24, 1910, Fernald and Wiegand (no. 3231,{.; G.; “‘prostrate near melting snow”; no. 3232, m., 3233, fr. type; G.; “prostrate”); Blomidon Range, July 3-5, rorr, C. S. Stewart (no. 29, st.; G.); Bonne Bay, serpentine tableland, alt. about 380 m., August 27, 1910, Fernald and Wiegand (nos. 3227, 3228, fr.; G.)—Gaspé Peninsula, Mt. Albert, deep ravine near snow, July 23, 1881, J. A. Allen (m., f.; G.); north slope of Allen’s ravine, on hornblende schist, July 26, 1906, Fernald and Collins (nos. 501, 503°, f., 507, fr.; G.); on wet serpentine slopes, July 23, 1906, Fernald and Collins (nos. 508, f., 510, M., 514, fr.; G.); brookside near permanent water, alt. 700 m., Kos ci 13, Sicaihay Collins and Fernald (no. 60, fr.; G., N.; partim fructibus sati dry serpentine barrens, sete m. ait August 9, 1905, Collins and Fernald (no. 62, fr.; G.; partim fructibus glabratis ut in no. 60). In its vathine short and dense catkins, at the base not or hardly loosely flowered, this variety approaches typical S. anglorum, but differs in its firmer, more rounded, and soon glabrous leaves and the glabrate twigs, in which characters it comes near to the following varieties. There is also no. 3235 of Fernald and Wiegand, collected at the same time as no. 3231 “from near sea level to serpentine tableland, alt. about 550 m.” as a prostrate shrub. It occurs, according to the specimens in Herb. G., with the rounded leaves of the typical kophophylla, and also with more acute, lanceolate leaves, and both orms seem not to possess stomata in the upper leaf surface. So far as we know at preeent, — so no S. elas pies in the parexion Mountains, =e therefore t cann var. Macou (Rydbg.) m. (see my second article). It Sccgiat needs farther ar onl No. 3234, collected by Fernald and Wiegand in the northeastern region of the Blomidon Range, on serpentine tableland, about 550 m. alt., July 24, 191° 1918] SCHNEIDER—AMERICAN WILLOWS 133 (fr.; G.), much resembles S. cordifolia, but I found stomata in the upper side of the leaves. We do not know enough of the willows of this range to be able to determine this form properly. S. ANGLORUM var. araioclada,> nov. var.—Frutex ut sub var. kophophylla descriptus sed sequentibus signis distinctus: folia adulta satis tenuiter papyracea, minora inferiora obovalia, obovato- oblonga vel ut majora superiora ovalia, elliptica, ovato-elliptica, obovato-elliptica vel rarius obovato-lanceolata, apice vulgo magis obtusa vel rotunda quam acuta, raro retusa vel subito plicato- acutata, basi rotundata ad late cuneata, rarius sensim attenuata, - Margine integerrima, minimis exceptis 1.5:1 vel 3:2 ad 4:2.7 vel 5:2.9 cm. vel angustiora acutiora ad 4:1.8 cm. magna, superiora vulgo abinitio glaberrima; petioli interdum ad 10 mm. longi; stipulae rarissime evolutae, minimae, lineari-lanceolatae, caducae; amenta fructifera ramulos foliatos ad 4 cm. longos terminantia; mascula I.5-2.5:1cm. magna, minus quam in typo sericea; bracteae longe sed satis laxe sericeae; glandulae 2-1; feminea sub anthesi ad 3-5:1cm. magna, fructifera vulgo 3.5-5.5 cm. longa et 1.8 cm. crassa, basim versus vulgo distincte laxiflora; bracteae interdum quam in masculis oblongiores sed saepissime densius sericeae, extus ad apicem interdum partim glabrescentes; ovaria ovoideo-oblonga, griseo-villoso-tomentosa, sessilia vel subsessilia; styli distincti, inte- gri vel apice breviter bifidi, quam stigmata oblonga bifida vix duplo (rarius in floribus valde juvenilibus fere 2}plo) longiores; glandula ut in typo longa, pl. m. anguste conica, bractea duplo brevior; fructus ovato-conici, maturi ad 7-8 mm. longi pedicello subnullo _ Vel brevi glandula 4 ad 2plo breviore excluso, laxius quam ovaria villoso-tomentosi, fulvi. TYPE LocALIry.—Gaspé Peninsula, north slope of Mt. Albert. .—Gaspé Peninsula and the Selkirks and Rocky Mountains in ge Columbia, Asulkan Valley, and in Alberta, near Laggan and Jasper SPECIMENS EXAMINED,—CANADA: Quebec, Gaspé County, Mt. Albert, deep ravine near snow, alt. 810 m., August 2, 1881, J. A. Allen (m.; G.; amentis parvis ovatis, glandula dorsali nulla, forma porro observanda); alt. 750-1050 m., July 26, August 2, 1881, J. A. Allen (m., f.; G. ex herb. Bebb; > Derived from dpatés, slender, and «\ddos, branch. 134 BOTANICAL GAZETTE [AUGUST amentis parvis); head of Allen’s ravine, August 8-15, 1905, Collins and Fernald (fr. im.; G.); north slope of same mountain, on hornblende schist, July 26, 1906, Fernald and Collins (no. 500, m. paratype; G.; so1*, f.; G.; 503, 503%, f. adult., 505, f. type; G.); July 20, 1906, Fernald and Collins (no. 506, fr.; G.); on wet serpentine slopes, July 23, 1906, Fernald and Collins (no. 510%, m.; 514, fr.; G.).—British CoLumsra: Selkirk Mountains, Rogers Pass, alt. 1350 ny July 31, 1890, J. Macoun (no. 18°, f.; N.); Asulkan Valley, Glacier, alt. 1590 m., J. G. Jack, August 14, 1904 (fr.; A., G.); same place and date, A. Rehder (fr.; A.; both specimens identical with those like no. 506 from Gaspé).—ALBERTA: Lake Agnes near Laggan, August 11, 1904, A. Rehder (m., f.; A.; forma incerta quamvis ad S. petrophilam spectans); slopes of ravine on Mt. Aylmer, alt. 2250 m., August 4, 1899, W. C. McCailla (no. 2248, m., f.; Cor.); mountains above Lake Louise, alt. 1800-2400 m., July 21, 1907, F. K. Butters and E. W. D. Holway (no. 262, f.; N.; forma quasi ad S. petro- philam transiens sed foliis magis quam in hac specie discoloribus); Lake Louise, July 22, 1904, J. Macoun (no. 68883 O., fr.; N.); Fitzhugh Mountain, near teapes Park, August 1917, J. M. Wise Tobe: 95379, 95397; 95398, 95401 O., fr., m.). This peculiar variety differs from the type chiefly in its less pubescent, mostly much more elongated, and yellowish twigs, in its almost glabrous young leaves, and in its aments which, on an average, are longer and thinner, at least much more loosely flowered toward the base. It is, apparently, closely connected with var. kophophylla, which as a whole has firmer leaves and denser and shorter catkins, but in its glabrous character comes nearer to var. araioclada than to the typical anglorum. See also my remarks under the following form. S. ANGLORUM var. antiplasta,° nov. var.—Frutex habitu ramu- lisque ut in var. araioclada; folia adulta chartacea, anguste ovalia, elliptico-oblonga, anguste obovato-oblonga, interdum oblanceolata, rarius elliptica vel obovato-elliptica, utrinque pl. m. acuta, raro rotundata, saepe apice breviter plicato-acuminata, vulgo 1.5~ 2.5 cm. longa et vix ultra 1 cm. lata, maxima ad 3:1.3-1-5 (rarius 1.8) cm. magna, integerrima vel interdum basim versus obsolete parce denticulata, superne subtusque ut in var. araioclada sed nervis lateralibus vulgo ut in petrophila angulo acutiore a costa abeuntibus et magis versus apicem currentibus; petioli graciles, 2-8 mm. longi, vulgo sparse pilosi; amenta cylindrica, sub anthesi satis brevia et tenuia, vulgo sublaxiflora, ramulos laterales in mas- culis vix ad 1 cm. longos ceterum ut in araioclada terminantia, § Derived from dvrirdacros, similar. 1918] SCHN EIDER—AMERICAN WILLOWS 135 rhachi parteque nudo pedunculi pl. m. villosa; mascula vix ad I.5:0.7 cm. magna, bracteae et cetera ut in aratoclada, glandula dorsalis (an semper ?) nulla; feminea sub anthesi 1-2:0.5-0.7 cm. magna, fructifera vix ad 3 cm. longa et 1.2 cm. crassa, bracteae ut in masculis; ovaria ovoideo-oblonga, pl. m. sessilia; styli distincti, saepe apice breviter bifidi, stigmatibus brevibus oblongisve paullo vel ad 2.5plo longiores, glandula ut in varietate precedente; fructus ovato-conici, subsessiles, ad 6 mm. longi, laxius quam ovaria villoso-tomentosi vel anni praeteriti subglabrescentes. Type LocaLiry.—Gaspé Peninsula, serpentine slopes of Mt. Albert. RANGE.—As above. SPECIMENS EXAMINED,—Canapa: Quebec, Gaspé Peninsula, Mt. Albert, serpentine slopes, July 23, 1906, M. L. Fernald and J. F. Collins (no. 500, f., fr., type; G.); exposed serpentine barrens, alt. tooo m., August 9, 1905, Collins and Fernald (no. 61, m., f.; G., N., O.); sheltered money knolls, Au- gust 10, 1905, Collins and Fernald (no. 61°, f.; G., N.,O.; a precedente nonnisi petiolis vulgo longioribus differe videtur); on wet serpentine slopes, July 23, 1906, Fernald and Collins (no. 511, fr.; G.; forma gracilis juvenilis, habitu S. petrophilae valde similis); north slope of same mountain, on hornblende schist, July 26, 1906, Fernald and Collins (no. 504, f. defl.; G.; forma satis vegeta, ramulis elongatis, foliis pl. m. plicato-acuminatis). At first sight this variety much resembles S. petrophila in its habit, the shape of the leaves, and the yellowish color of the young twigs, but the leaves are of a deeper green on the upper surface and much paler and glaucescent on the lower surface, and do not differ in this respect from any other form of S. anglorum. It is, however, much easier to distinguish herbarium specimens of both species than to express the differences in exact words. The two species meet each other in the Rockies of Alberta and British Columbia, and there are also certain forms in northern Montana, and even in Wyoming which at present I am at a loss to determine. Some of them may represent hybrids between S. petrophila and other species with which I am not yet sufficiently acquainted. 3. S. PETROPHILA Rydbg., in Bull. N.Y. Bot. Gard. 1:268. 1899, is the species which seems to be nearest related to S. anglorum. It was first described by ANDERSSON (DC., Prodr. 167:287. 1868) as S. arctica petraea from specimens collected by E. Bourgeau “in summo Rocky Mountains.’”’ I have seen a photograph of the type at Kew and a cotype in the Gray herbarium. Both specimens bear the label of Patuiser’s Brit. N. Am. Expl. Expedition, with the printed indication “(Rocky Mountains” and “coll. E. Bourgeau 136 BOTANICAL GAZETTE [AUGUST 1858’; and upon them is written “Salix arctica R. Br. subal pestris And. (forte n. sp.).’’, ANDERSSON apparently changed the varietal name later to petraea. ‘The Kew sheet also bears, in the lower left corner, the inscription ‘‘Salix herbacea. Montagnes rocheuses Palouse prés les Glaciers, 18 aoit 1858.’’ According to MAcouN (Cat. Canad. Pl. preface, p. viii. 1883), BourGEAu “spent some time, in August 1858, in the Bow River Pass and the adjacent . mountains’? in Alberta. S. petrophila differs from S. anglorum chiefly in the color of the rather pale or grayish green leaves, which are not distinctly paler and never whitish beneath. The differences indicated by RypBERG between the two species are of no value, because his S. anglorum is mostly S. groenlandica. As I have already said, there are some forms in the northern habitat of petrophila which I have not yet been able to interpret properly. So far as I can judge by the specimens before me, the species ranges from about 52° N. latitude in southwestern Alberta and southeastern British Columbia through western Montana, north- eastern Wyoming, and central Colorado to the Truchas Peak in northern New Mexico. I have not seen specimens from Washing- ton and it is not mentioned in Pirper’s Flora. In eastern Oregon I know only of two localities. From Utah and Nevada I have seen very little material, and in California it is found in the Sierra Nevada from Sierra County to Tulare County. In western Nevada and the Californian Sierra, S. petrophila is mostly represented by a form which has been described as S. caespi- tosa by KENNEDY (Muhlenbergia 7:135, pl. 9. 1912). Through the kindness of Professor C. W. Lantz I have seen the type, which is preserved in the herbarium of the Agricultural Experiment Station at Reno. It was collected by the author on Mount Rose, Washoe County, Nevada, August 17, 1905 (no. 1173, fr.). It differs from typical petrophila in the more copious pubescence of the upper leaf surface, the acuter leaves, and the very short style. The last character seems very variable, and the type material before me consists only of fruits with withered styles and stigmas. Never- theless, I am inclined to use the name caespitosa for a variety which seems to be the prevailing form in the western part of the range of petrophila, and this var. caespitosa (Kennedy), nov. var., may be 1918] SCHNEIDER—AMERICAN WILLOWS 137 distinguished by its foliis utrinque acutioribus apice subacuminatis superioribus superne (saltem in parte) satis villosis, subtus vulgo glabris ad 3.5:1.3 cm. magnis, amentis femineis (immaturis) inter- dum ad 6:1.3 cm. magnis basi valde laxifloris longe pedunculatis. The most extreme form of this variety has been collected by Hail and Chandler on Mount Goddard, Fresno County, California, July 24-26, 1900 (no. 685, m., f.; G.); and I refer to it also a specimen collected by F. W. Congdon on Mount Dana, Mono County, Cali- fornia, August 27, 1895 (m., fr.; N.). It may be mentioned here that S. cascadensis Cock. (S. tenera And., non A. Br.) is regarded as very closely related to petrophila by RypBER«, or as ‘‘perhaps only a variety”’ of it by Batt. I pre- fer to place it in a different group next to S. phlebophylla, and 1 shall speak of it later. There are three more willows, which, in my opinion, should be. included in the same group with S. arctica, namely S. stolonifera Cov., S. ovalifolia Trautv., and S. groenlandica Ldstr. The first two have been well treated by CovittE (1901), and need only a few remarks, while the history and taxonomy of the last ought to be explained in detail. 4. S. STOLONIFERA Coville, in Proc. Wash. Acad. Sci. 3:333. pl. 41. fig. 1 (Willows of Alaska). 1901, “is a species of eastern Alaska, in the glacier region from Yakutat Bay to Glacier Bay and. Lynn Canal.” RypsBeErc (1899) mentioned this species under the name of S. unalaschensis ‘‘Cham. Linnaea 6:539."’ As COVILLE has explained, CHamisso did not propose such a species, but merely describes a “‘Salix unalaschcensis, multis cum arctica Pall. conve- niens, pluribus ab illa abhorrens, nulli nostrarum propius accedens,” to which he did not give a specific name. His form from Unalaska is the same as S. ovalifolia Trvt., and ANDERSSON has already mentioned in the Prodromus “‘S. unalaschkensis Chamisso”’ among the synonyms of TRAUTVETTER’S species. CovILLE describes the Ovaries as ‘‘smooth or with some traces of pubescence toward the apex,” and he regards the glabrous form as the typical and com- mon one. I think it best to propose a f. subpilosa, f. nov., fructibus pl. m. interdum satis dense pilosis, because such forms resemble somewhat S. arctica, especially when the old fruits have lost the 138 BOTANICAL GAZETTE - [auGusT style. The leaves, so far as I can see, always possess stomata in the upper epidermis, as is the case with typical S. stolonifera, while they are wanting in the leaves of typical S. arctica and S. ovaltfolia. The length of the style and the rather long linear stigmas seem to be the best characters to distinguish S. stolonifera from the other species of this group. “The characteristic of the production of slender leafless, subterranean branches or stolons’ is not always clearly seen on herbarium specimens, and the presence of such stolons may possibly be detected in other related species. 5. S. OVALIFOLIA Trautvetter in Nouv. Mém. Soc. Nat. Mosc. 2:306, pl. 13 (De Salic. Frig. Kochii). 1832.—S. myrtilloides forma 4 Chamisso in Linnaea 6:539. 1831.—S. unalaschkensis Chamisso ex Andersson in Ofv. K. Vet.-Akad. Férh. 15:130. 1858.—S. rotundata Rydberg apud Macoun, List PI. Pribilof Islands in Jordan, Fur Seals N. Pac. 3:571. 1899, non Forbes 1829.—S. cyclophylla Rydberg in Bull. N.Y. Bot. Gard. 1:275. 1899, non Gandoger 1882.—The type locality of the species is Cape Espenberg in the Kotzebue Sound. Its range extends from the Bering Strait, where it is probably also found on the Siberian Coast,’ northward to Point Barrow and Martin Point, where it has been found by F. Johansen, July 30, 1914 (no. 136” or 93484 O., fr.); and southward to the Pribilof and Aleutian Islands and the Alaskan Peninsula, but it has also been collected on Kodiak Island, and to the eastward as far as Yakutat Bay. The typical form has glabrous ovaries and fruits; there are, however, specimens with loosely pubescent capsules collected by Trelease and Saunders, St. Paul Island (no. 3442, fr.; M.), which may represent the var. pubescens And. (DC. Prodr. 167: 291. 1868). This is described as being distinguished by “‘capsulis tenuiter hir- sutis griseo-pubescentibus petiolis et foliis basi longius hirsutis.” As no type is given, I cannot decide whether ANDERSSON’S variety is identical with this specimen. Some other specimens which Covitte has cited as typical S. arctica, while RYDBERG took them for S. diplodictya Trautv., should be discussed. The last species has been described, as I have explained, as having the leaves green and glossy on both sides, and 7Lg. C. Wricut in 1853-56 on Arakam Island. Those specimens are distributed as S. uva-ursi, but agree well with S. ovalifolia except that the fruits are not glaucous. 1918] SCHNEIDER—AMERICAN WILLOWS 139 it has certainly nothing in common with the forms in question. These specimens seem to represent a form somewhat intermediate between typical S. ovalifolia and typical S. arctica. It may be characterized briefly as follows: ab ovalifolia satis diferre videtur foliis amentisque majoribus, floribus masculis tantum (an semper ?) glandula ventrali instructis, ovariis satis pubescentibus etiam fruc- tibus tenuiter vel partim (fere ut in var. pubescente supra) pilosis sed non distincte glaucescentibus; ab arctica praecipue recedit foliis minoribus pl. m. rotundatis vel obovato-rotundis, amentis parvioribus, fructibus minoribus (perfecte maturis non visis) pl. m. glabrescentibus vel partim glabris. I do not want to propose a new name for this form, because it needs further observation, but it is by no means identical either with S. ovalifolia pubescens or with arctica. It may be referred provisionally to S. ovalifolia var. sub- arctica Lundstr6ém in Nov. Act. Roy. Soc. Sci. Upsala III. 1877. p- 41, where the following characters are given: “8, subarctica nob. capsulis pubescentibus; foliis majoribus, subtus parce villosis.”’ As I have said, the forms described by LunpstR6M cannot be fully understood until his type material is examined. There remains another arctic form which I should have regarded as not separable from typical S. ovalifolia but for the fact that I found stomata in the upper leaf epidermis in most of the specimens cited later. So far as I can judge by the rather scanty material before me, this variety, for which I propose the name var. camden- sis, var. nov., seems chiefly to differ from S. ovalifolia in the fol- lowing respects: foliis nondum perfecte evolutis minoribus vel oblongioribus elliptico- vel ovato-oblongis vel oblanceolatis apice acutis vel obtusis basi acutis vel pl. m. rotundatis vix ultra 1.5 cm. longis et 1 cm. latis in epidermide superiore vulgo pl. m. stomati- feris adultis textura tenuiore et subtus minus distincte reticulatis, petiolis saepe quam gemma brevioribus, amentis masculis subma- Jjoribus ad 1.5: cm. magnis, fructiferis subminoribus ad 1.5 cm. longis et 1.2 cm. crassis. I examined the following specimens: Alaska, Camden Bay, Collinson Point, July 17, 1914, F. i (no. 116 or 93482 O., fr., type in O at a 1914, F. Johansen (no. 44° 93807 O.; f.; stomata non visa; no. 44° or 93806 O., m.); Kongenevik, july 1914, F. Johansen (no. 82* or 93805 O., m. I40 BOTANICAL GAZETTE [AUGUST syntype; no. 82° or 93804 O., fr.; stomata superne in foliis non visa); west of Martin Point, July 30, 1914, F. Johansen (no. 136* or 93483 O., st.; folia superne stomatibus numerosis instructa breviter petiolata, forma porro obser- vanda). Dr. Frits JOHANSEN has been so kind as to give me the following informa- tion regarding this variety: “Nos. 44a, b, Collinson Point. This willow grew on more bare, gravelly tundra near the beach (transition region to the latter), in patches of several plants. Its growth was very prostrate and depressed (among stones and vegetation), with the stems and branches lying very close to the ground and spreading widely, so that only the catkins showed up from a little distance. Especially the subterraneous parts (roots and stem parts) were less extensive and spreading than with those found at Kongenevik, Alaska (see below); probably because they did not grow on sand dunes as is the case at the former place.—Nos. 82a, b, Kongenevik. The collecting place was where the seashore (beach) through low sand dunes goes over into the more typical tundra behind. On these sand dunes the vegetation is very char- acteristic and consists almost exclusively of Elymus, Carex, Salix, Chamaerium, etc.; each species spreading (both above and under the ground) over large patches (areas) and dominating more or less to the exclusion of the other species. This Salix seemed to be very prostrate, but the larger part of each plant is buried in the sand, so that only the leaf and catkin-carrying branch parts (outer third) protruded. It was mostly large plants widely spread- ing (both roots and stems); the branches often having the form of long runners’”’ intersecting the sand rhizome-like in all directions. The sand- covered parts of the branches were without leaves or catkins and pale (white-yellow). When growing in less sandy soil the growth is naturally regions at Collinson Point and at Kongenevik, so did also the growth of the Salix in question resemble those of the same species from both of the foregoing places. At the time of collecting the plants had dropped ¢ catkins and had unripe ° catkins.” 6. S. GROENLANDICA Lundstrém, Nov. Act. Reg. Soc. Sci. Upsala III. 1877. p. 36.—S. arctica Liebmann, FI]. Dan. XIV. fasc. 42:7, pl. 2488. 1849, non Pall.—S. arctica , Groenlandica And. in C., Prodr. 16*: 287. ut videtur excl. forma 6 pusilla.—ANDERSSON 1918] SCHNEIDER—AMERICAN WILLOWS I4I based his var. groenlandica on ‘‘S. arctica Fl]. Dan. t. 2488,” and he distinguished 6 forms: (1) hebecarpa, which is nothing but the type; (2) lejocarpa, with glabrous ovaries; (3) Jatifolia, which probably only represents a vigorous form with ‘‘foliis orbiculato- ovalibus’’; (4) angustifolia, a mere form with “‘foliis lanceolatis’”’; (5) macrocarpa, which is nothing but the typical plant with normal big aments; and (6) pusilla, which I cannot interpret because the description (“‘fruticulus vix digitalis, foliis 1-3 lin. longis densis- sime confertis. Salici retusae ser pyllifoliae analoga’’) is insufficient, and ANDERSSON does not cite a type or any locality for it. The description and figure given by LreBMANN are quite sufficient to understand what form is meant, and it is rather surprising that this well marked species could be misunderstood by later authors. Lunpstr6m did not say much about it, because he was dealing with Asiatic and European forms, and only wanted to separate it from the related species. LANGE (Consp. Fl. Groenl. 1:108. 1880), in adding his var. ‘“‘minutifolia And. mscr.’’ to those already described by ANDERSSON (but omitting f. macrocarpa), referred S. arctica Br. (S. Brownei Ldstr.) as a synonym to S. groen- landica, and seems to have misunderstood Brown’s plant. Ryp- BERG, in his turn, as I have said, mixed the real S. groenlandica with his S. anglorum, and gave the name groenlandica to specimens of the latter species and to several forms of different origin. In my opinion the true S. groenlandica may easily be recognized by its glabrous leaves, which are shining dark green and without stomata above and distinctly glaucescent beneath, the margin being entire or often more or less glandular denticulate, by its large aments which measure from 5:1.2 to 10:1.6 cm. in fruit, and by its distinctly pediceled ovaries, which bear a rather thin and short silky pubescence even when young and possess a short and broad gland of about half the length of the pedicel. The shape of the ventral gland, which is the same in both sexes, differs much from that of the other species of this group where, as a rule, it is oblong or ovate-conical and longer than the pedicel. The thin pubescence of the ovaries and fruits, which are often almost glabrate or entirely glabrous in var. lejocarpa (And.) Lange, gives them a different aspect from the tomentose capsules of S. arctica, S. anglorum, or 142 BOTANICAL GAZETTE [AUGUST S. petrophila. In the size of the fruiting aments S. groenlandica is next to vigorous forms of S. arctica and to S. arctica var. sub- cordata. The type of S. groenlandica has been collected by VAHL “in locis humidis Groenlandiae orientalis et occidentalis a limite maris ad alt. 200 pedum.” I have not seen a specimen of VAHL’s and no material from eastern Greenland. Judging by the specimens I have examined, its range extends from Disco Island (70° N. latitude) through the southern part of Baffin’s Land westward to the Bath- urst Inlet (about 109° W. longitude), and southward along the shores of the Hudson Bay through Ungava and Labrador to the western Gaspé Peninsula and the Port 4 Port Bay in western New- foundland. There are also some rather uncertain and fragmentary specimens from the Lancaster and Jones Sound, and probably the habitat of S. groenlandica reaches its northern limit at about the 76th parallel. Other specimens have stomata in the upper epider- mis of their leaves and may represent a different variety or be of hybrid origin; they need further observation. ARNOLD ARBORETUM Jamaica Pratn, Mass. FECUNDATION AND FORMATION OF THE PRIMARY ENDOSPERM NUCLEUS IN CERTAIN LILIACEAE MILDRED NOTHNAGEL (WITH PLATES III-V) Introduction Between 1890 and 1902 many articles appeared on fecundation and double fertilization in the angiosperms. On the whole, the authors have dealt with the entrance of the male nuclei, their behavior within the embryo sac, their form, the union between the egg and male nucleus, and between the male and polar nuclei; but they have not investigated in detail the chromatin changes that occur from the time of contact of these nuclei to the completion of the first division. In 1891 GuiGNArRD (6) described the entrance of the so-called antherozoids into the sac, each accompanied by its centrosomes. One male nucleus became applied to the egg nucleus, each of whi took on the resting condition and remained distinct for some time. While in this state the male nucleus enlarged and both the egg and the sperm nucleus flattened at the surface of contact, but with a distinct line of demarcation remaining between them for some time. Even after the nuclear membranes had disappeared, the contour of the two was traceable at the periphery. Later he dis- tinguished, on opposite sides of the nuclear cavity, two groups of chromatin in the spirem stage. No drawings were made to show this. When the nuclear plate was formed, he asserted that one- half of the chromosomes were contributed by the egg and one-half by the sperm. The process of fertilization in Lilium Martagon and L. candidum was described by Morttier (13) in 1898. In L. Martagon there was no complete fusion of egg and S-shaped sperm, the lack of which resulted in a failure to mature seeds. In the region of the two polar nuclei, which had not fused and which began disintegrating 96 hours after pollination, a nucleus, similar to the nucleus which united with 143] [Botanical Gazette, vol. 66 144 : BOTANICAL GAZETTE [AUGUST the egg, was observed. JL. candidum furnished material for normal fertilization. At the union of the egg and sperm, the latter was about the size of the former and both were in the resting condition, the chromatin being distributed in the form of a fine network. No boundary was observed separating the two elements at the point of contact, and the fusion that took place during resting condition was so complete at the close of fertilization that there was no visible distinction between male and female chromatin. In 1904 Morrrer (14) confirmed his earlier investigations, pointed out the S shape of the male nucleus, the fusion of the sexual nuclei in resting condition, the coming together of the two polar and male nuclei in L. Martagon, and the cause of the non- fusion. Although he stated that the sexual nuclei were in the resting condition at the time of fusion, he called attention to the chromatin of the sperm heing more regular than that of the egg. It was also claimed that the nucleoli fused at fertilization. One of the first reports of double fertilization was made by GUIGNARD (7) in 1899 for Lilium Martagon. In this species he observed the union of one of the male nuclei with one of the polar nuclei, followed by the union with the second polar nucleus. The chromatin of the two male nuclei, on account of being coarser, was distinguishable from that of the egg and polar nuclei with which they had fused. He also stated that he was able to recog- nize the triple origin of the secondary nucleus during the prophases, although no drawings were given. NAWASCHIN (15), in the first report of double fertilization in Lilium Martagon and Fritillaria tenella, noted that the cellulose membrane surrounding the sexual apparatus was absorbed just previous to the entrance of the pollen tube, and that the spiral- shaped male nucleus entered the protoplasm of the sac. He con- cluded that the sperms took on various shapes under various conditions, and, as GUIGNARD had assumed, that they were motile. One sperm was found to enter the egg, the other to unite with the superior polar nucleus, and in both cases a complete fusion occurred after a certain period. The fusion of the superior and inferior polar nuclei took place after the male nucleus had united with the former. The triple fusion was followed in a short time by a division which preceded that of the egg. 1918] NOTHNAGEL—FERTILIZATION 145 In 1900 GUIGNARD (9) found that in some cases the polar nuclei, the upper one of which he said was analogous to the egg, fused before the entrance of the pollen tube, and that when the male nuclei entered the sac, they entered into fusion so quickly that, in some species, one rarely saw them free. In Tulipa he was able to follow the contour of the three nuclei entering into the fusion nucleus some time after their coming together, and even after the membranes had disappeared at their surface of contact. As a result of investigations in various groups of angiosperms by GuIGNARD (6, 7, 8, 9, 10), NAWASCHIN (15), STRASBURGER (17, 18), and others at this time, it was generally concluded that double fecundation was normally found in angiosperms and that the uniting of the male nucleus with the polar nuclei was in the nature of a pseudo-fecundation whose function was to stimulate the formation of ‘‘albumen.”’ Ernst (3), investigating fertilization of Paris quadrifolia and Trillium grandiflorum, found that the two polar nuclei were fused before the male nucleus united with them, and at times a spirem was formed previous to the entrance of the sperm, showing that the male nucleus was not necessary to stimulate division. At other times the spirem was not formed until the three nuclei were fused, in which case he was unable to discern which part of the chromatin was contributed by the various nuclei. He also stated that it was not safe to rely upon the number of nucleoli found in the fused mass to ascertain whether fertilization had taken place or not. In the fecundation of the egg there was a complete blending of the substances, and at cross segmentation he failed to find the arrange- ment of the chromatin into two groups. STRASBURGER (17, 18) used the terms generative and vegetative fertilization, the latter being applied to the triple fusion. The union of the sperm, either with the egg or with the polar nuclei, functioned as a stimulus. In 1911 Coutrer (2), after reviewing the literature on endo- sperm formation, stated that since endosperm may form without the fusion of the sperm or even of the second polar nucleus, these being simply supplementary, there seemed to be no reason why “there should be any hesitation in recognizing the endosperm as gametophyte.”” He concluded that “the product of such fusions is 146 BOTANICAL GAZETTE [AUGUST merely an undifferentiated tissue which practically continues the tissue of the gametophyte, that is, it is simply growth and not organization.” From 1902 to 1913 practically nothing new was published on the subject of fertilization. In 1913 BLACKMAN and WELSFORD (1) reported that the chromatin of the vermiform male nucleus was in a network, although not the network of a resting nucleus, this condi- tion becoming more noticeable later on. At times they also noted that the chromatin of the egg might become threadlike just previous to fusion. The most recent paper on fertilization is by Sax (16) in 1916 on fertilization in Fritillaria pudica, in which he noted that the vermiform sperm lay indented in the egg for some time before the membranes between them disappeared. The chromatin was in more or less of a network and the granules were of various sizes. When the membranes at the surface of contact broke down, the contents of the two nuclei mingled and were not distinguishable from each other. The spirem usually appears after this. Triple fusion was also complete and the resulting nucleus divided before that of the fertilized egg. In none of these cases have the chromatin changes been carefully followed from the time of contact of the nuclei until the completion of the first division, the emphasis previously having been placed upon the actual coming together, the uniting, and the very earliest steps in division. The process of fertilization and distribution of the chromatin contributed by the egg and sperm in Pinus and Abies has been carefully worked out by FERGUSON (4, 5) and HuTCHINSON (12), and it was with the desire that something of this nature should be done for angiosperms that the present investigation was undertaken. Materials and methods For fertilization of the egg, Trillium grandiflorum was used, the material being collected in damp woods along the Des Plaines River, northwest of Evanston, Illinois, from May 3 to May 26, 1916. The first collection was made at the time of pollination, although the pollen tube was not seen in the micropyle until two weeks later, 1918] NOTHNAGEL—FERTILIZATION 147 while in the collection of May 26 dividing endosperm nuclei and second division of the fertilized egg were found for the first time. Lilium Martagon, collected from the garden of Indiana University, Bloomington, Indiana, in May 1916, 96 and 120 hours after pollina- tion, was more favorable for the first division of the endosperm nucleus. The former material was killed and fixed in chrom-osmic- acetic acid 1-2 hours and then in chromo-acetic acid 24-36 hours, washed, dehydrated, and imbedded either from chloroform or xylol. Lilium Martagon ovaries were killed and fixed in chrom-osmic- acetic acid 24-36 hours, washed, dehydrated, and imbedded from chloroform. All sections were cut 12m thick and both modified triple and Heidenhain’s iron-alum-haematoxylin were used for staining, the latter being more satisfactory for most stages, as the chromatin was more sharply differentiated. Formation of the primary endosperm nucleus For the development of the spirem and the first division of the endosperm nucleus, Lilium M artagon was found to be very favor- able, as many dividing primary endosperm nuclei were found in the sacs of material killed 96 and 120 hours after pollination. Activity did not cease at the end of the first divisions, for as many as 12-16 nuclei were found in many sacs of the older material (fig. 29). The sperm comes in contact with the polar nuclei before these two have fused, although they may be in contact or in close proxim- ity (fig. 17). These three nuclei will usually be found in the center of the sac where just previous to the triple fusion the two polar nuclei were to be seen. The chromatin of the egg can scarcely be said to be in a network, but rather to consist of strands which are more or less united (figs. 17, 18), that of the male nucleus being much coarser than that of the polar nuclei. When the sperm reaches the middle of the Sac, it still has its curved or vermiform shape, while the contour of the polar nuclei may vary, sometimes being quite curved before coming together (fig. 17), but at other times only changing to this shape as pressure is exerted by contact. The three nuclei upon uniting may be variously twisted about each other, the male nucleus usually twisting more than the others and recognizable by 148 BOTANICAL GAZETTE [AUGUST its coarser chromatin strands (fig. 18). In fig. 18 the three nuclei as a whole present a more or less globular contour, although the nuclear membranes are still present at the surfaces of contact. In Trillium grandiflorum the three nuclei, which unite to form the primary endosperm nucleus, are all alike in shape, it being impossible to distinguish the male nucleus from the two polar nuclei by its form or size (fig. 16). Since the mass of the three nuclei is so large, it is often impossible to find parts of all three nuclei in one section, and frequently only two will be visible (fig. 19). All three contain nucleoli, sometimes one, while at other times there are many. The chromatin strands thicken until they may be traced for a considerable distance (figs. 19, 20). While in some instances the membranes still separate the nuclei (fig. 20), at other times they are not visible, as in fig. 19; but, nevertheless, where the chromatin contributed by one nucleus leaves off and that of another begins is very easily seen. Up to the period when the spirem has assumed its mature thickness, the separating membranes may not have entirely dis- appeared (fig. 21), and in some cases the three groups of spirems are plainly evident. From fig. 22 it could be concluded that a complete fusion or intermingling of chromatin had previously occurred, but _ such has not happened, for in the next section of this same primary nucleus parts of all three nuclei are seen (fig. 21). Even at this stage, before the complete breaking down of the separating mem- branes, segmentation has begun and spindle fibers are forming about the group. As far back as the coming together of the two polar nuclei and the sperm nucleus, a surrounding complex of fibers could be seen (fig. 18). In fig. 21 the fibers have commenced to radiate out into the cytoplasm, followed after a short period by a complete segmenta- tion of the spirems, resulting in three groups of chromosomes being scattered upon the three arms of the tripolar spindle, respectively (figs. 23, 24). As the tripolar structure gradually assumes the form of a bipolar spindle, the chromosomes, which were previously lying upon the third arm, are pulled into line with the other two groups, thereby forming a typical bipolar spindle (figs. 24, 25): The chromosomes now thicken and are typically arranged into the 1918 NOTHNAGEL—FERTILIZATION 149 equatorial plate of the bipolar spindle; but even yet the third group of chromosomes is recognizable, as can be seen at the left in the group in fig. 25. After this stage the chromosomes contributed by each of the polar nuclei and by the sperm nucleus are no longer distinguishable (fig. 26). No trace of such distinction is seen in early metaphase or later spindle phases. How the various chromosomes finally arrange themselves upon the spindle and their distribution could not be ascertained in this investigation. When the 3x chromo- somes have gathered upon the equatorial plate of the bipolar spindle, each very much elongated chromosome splits longitudinally (fig. 26) preparatory to a typical equational division. No inter- mediate stages between early metaphase and early telophase were found. As the chromosomes reach the poles, they are somewhat shorter than when leaving the equator, and from the count, as seen in fig. 27, 3x chromosomes have passed to each pole. In the third division of the endosperm nuclei of Trillium grandiflorum a peculiarity was noted. In one of the dividing nuclei there were still to be seen the three groups of chromosomes upon the spindle, each group consisting approximately of six chromosomes, or the haploid number. It is easily seen that there is a great simi- larity in appearance between this third division of the endosperm and the first division of the primary endosperm nucleus. A similar stage was observed in the second division of the endosperm nucleus ot Lilium Martagon (fig. 29), showing in the upper dividing nucleus an appearance very similar to that seen in fig. 24. It was not determined how long endosperm division would continue in Lilium Martagon, as nothing older than 120 hours after pollination was collected. Fertilization of the egg and its first division Trillium grandiflorum furnished the best material for this phase of the investigation, as the later stages of the first division of the fertilized egg were not to be found in Lilium Martagon collected 120 hours after pollination. In L. Martagon the chromatin of the egg and the sperm, at the time when the male nucleus lies coiled upon the egg, is similar in 150 BOTANICAL GAZETTE [AUGUST appearance to that described for the polar nuclei and the second sperm nucleus. The chromatin is in strands, that of the sperm being heavier than that of the egg (fig. 1). The sperm fertilizing the egg is very much smaller than the sperm uniting with the polar nuclei at the time of contact and not so vermiform (compare figs. I and 17). Fig. 2 illustrates a typical fertilized egg of Trillium grandiflorum just a little later in development than that of Lilium (fig. 1). In this later stage the chromatin is lumpy, the particles being larger in the sperm than in the egg, and the membranes sepa- rating the two nuclei are becoming very thin, so that it is difficult to distinguish them at all times. After this time these membranes are rarely to be found, although in some instances they persist for a longer period (fig. 4). The chromatin gradually collects into larger groups, forming more or less broken threads connected with each other by fine anastomoses (figs. 3, 4). In many portions of the nucleus of this fertilized egg the parallel nature of some of these strands is quite con- spicuous (figs. 3,4). In some fertilized eggs, as for example in fig. 5, a more or less beaded, although discontinuous, spirem was noted. Even though the nuclear membranes which separated the egg and the sperm have disappeared, the chromatin that has been contrib- uted by each of the two nuclei remains distinct (fig. 3). This con- dition is much more evident in some fertilized eggs than in others. The sperm at the period of union contains a much smaller amount of chromatin than the egg and throughout most of the subsequent stages this condition persists (figs. 3, 4, 9,10, 12). During all this time the fertilized egg is growing in size and increasing the amount of chromatin. When the continuous spirem is first formed, it is quite thin (fig. 6), but as the prophase advances the chromatin thread thickens and shortens until a comparatively thick spirem results (figs. 6-12). Instead of one continuous spirem, two distinct spirems are usually to be seen within the single nuclear cavity, although located in different parts of the cavity (figs. 8-10). In some sections such differentiation is not visible (figs. 6, 7). The dotted lines a—b in figs. 9 and 10 separate the two spirems, one of which was contributed by the sperm, the other by the egg. In spite of the fact that a nucleolus is not seen in the sperm when it unites with the egg, very small ones are found in later 1918] NOTHNAGEL—FERTILIZATION I51 stages (fig. 3), and still later, in the spirem stage, a large nucleolus is frequently observed (fig. 9); but at the beginning of segmentation all traces of these nucleoli have vanished. With the beginning of segmentation, the chromatin threads appear to contract, presenting the appearance of “‘the second con- traction”’ of heterotypic mitosis (figs. 11, 12). Following this, the nuclear membrane surrounding the two groups disappears, leaving the massed segments lying free in the cytoplasm. Even now the two sets of chromosomes are separate (fig. 12), and to all appear- ances a spindle is formed about one group of chromosomes and the other set is pulled into the bipolar spindle, for, as late as in fig. 13, the chromosomes contributed by the sperm are distinct from those contributed by the egg. In each of the two groups (fig. 13) there are approximately six chromosomes, or the haploid number. The writer was unable to determine the arrangement of these two sets upon the equatorial plate, owing to lack of material for later stages. From the number of chromosomes seen at telophase and later divisions, each splits longitudinally at metaphase, so that twelve, the diploid number, pass to each pole. Discussion A very full, detailed account of fertilization in Pinus by FER- GUSON and in Abies by Hutcutnson has been published; but a similar account is not to be found for angiosperms. FERGUSON (5) reports that in Pinus the chromatin of the egg is arranged in an interrupted reticulum, the network consisting of granules of various sizes in a colorless linin. When the contents of the pollen tube have been discharged into the egg, one of the male nuclei takes up a position on the concave side of the egg, this depres- sion having been formed at the approach of the male nucleus. Gradually from each nucleus a spirem is formed from the respective chromatin material, at which period fibers arise in the region of the spirems and the nuclear membrane gradually fades away. At segmentation these two spirems give rise to two groups of chromo- somes, but as they collect on the spindle this distinction is lost. Each chromosome splits longitudinally and each daughter nucleus receives the diploid number. When these daughter nuclei are Preparing for second division, the chromatin collects into two 152 BOTANICAL GAZETTE [AUGUST spirems, the steps being very similar to those of the first division, and it is concluded that in all probability they come from the maternal and paternal source respectively, in spite of the fact that in the formation of the daughter nucleus the chromatin has appeared completely fused. Since the subsequent divisions were not fol- lowed, it could not be determined how long this dual nature per- sisted. The account of fertilization in Abies balsamea by HUTCHINSON (12) varies somewhat from that of FErGuson. The contents of the male nucleus pass into the nucleus of the egg, although the chroma- tin groups remain distinct, and later, when the two sets of spindle fibers are formed, two sets of chromosomes arise from the respective nuclei. These two spindle complexes unite and the chromosomes of the maternal parent pair with the chromosomes of the paternal parent, after which the fibers disappear. The members of each pair twist about each other, bend, and become transversely segmented at the bend so that there are 2x pairs in the fertilized egg. When the second set of fibers appears, the members of the pairs resulting from the transverse segmentation separate for the opposite poles. HEATLEY (11) has described the development of the embryo sac of Trillium cernuum, in which the sac arises from the chalazal daughter nucleus of the megaspore mother cell, two megaspores only being functional. Each functioning megaspore divides twice to form the typically arranged 8-nucleate embryo sac. In the present study it was not considered necessary to work out the development of the sac, and furthermore, no attempt has been made to determine the method of entrance of the sperms into the sac or their passage to the egg and polar nuclei. BLACKMAN ‘and WELsForRD (1), Ernst (3), GUIGNARD (6, 9, 10), MOTTIER (13, 14), SAX (16), and STRASBURGER (17, 18) have reported on the earlier phase of fertilization, and on the whole have agreed. These same authors have also described in detail the coming together of the nuclei, their chromatin condition, and the breaking down of the nuclear membranes separating them, although GuIGNARD (7-9) differs somewhat from the others on the latter point, which will be spoken of later. By comparing these investigations with those of Pinus (FERGUSON 4, 5) and Abies (HUTCHINSON 12), it is apparent 1918] NOTHN AGEL—FERTILIZATION 153 to the writer that the present knowledge of certain phases of fertili- zation in angiosperms is very scanty, especially as to the fate of the maternal and paternal chromatin. ERNST (3) reports for Trillium grandiflorum, and a similar conclusion is reached by various authors for certain other plants, that there is a fusion of the polar nuclei previous to the entrance of the pollen tube; but in not a single case in either Trillium or Lilium has such a condition been found to exist, the polar nuclei always being distinctly separate, although usually in contact (figs. 16-19). Only a few cases of triple fusion were observed in Trillium grandiflorum, although all that were found appeared as illustrated in fig. 16. BLACKMAN and WELSFORD (1), GUIGNARD (7-9), MorTTIER (14), and SAx (16) have noted that the male nucleus can be distinguished from the egg and from the polar nuclei both by its shape and the condition of the chromatin, since this substance is coarser in the male nucleus, and at times assumes almost a spirem condition previous to fusion. The sperms have been found not always to retain their $ or curved form, for in Trillium the male nucleus could not be distinguished from the polar nuclei, either by its size or countour (fig. 16). The three nuclei (superior and inferior polar nuclei and male nucleus) of Lilium Martagon become very much twisted about each other very soon after coming in contact (fig. 18), and even previous to this the polar nuclei may have lost their globular form (fig. 17), although the writer failed to find any mention of this in previous accounts. The fibers appear early about the nuclear complex of Lilium and gradually merge into the cytoplasm, as FERGUSON (4, 5) has reported for the fertilized egg of Pinus. The chromatin is in fine strands and not in a network, as GUIGNARD and Mortter have stated. The number of nucleoli in each nucleus may vary from one to several, and in some specimens there are none. As SAx (16) has said for Fritillaria, at the time of contact the chromatin is threadlike, with large irregular pieces of chromatin scattered throughout. In many instances the separating nuclear membranes are still to be seen when the chromatin has been transformed into a 154 BOTANICAL GAZETTE [AUGUST comparatively heavy spirem (fig. 20), and in some instances at the beginning of segmentation fragments of it still remain (fig. 21). In cases where the separating nuclear membranes do disappear early, the limits of the nuclei are readily followed (fig. 19). GuIG- NARD (7), in his first report of double fecundation, says that the chromatin of the sperm enters into more or less of a spirem before fusion with the two polar nuclei, after which, at times, he is still able to recognize the triple origin of the secondary nucleus of the sac. None of his drawings are later than fig. 18 of the writer, and apparently the chromatin is in the same condition. In a later paper (8) he describes a similar condition in Narcissus. In Fritillaria, Sax (16), after stating that the chromatin of the male nucleus frequently passes into a spirem previous to the breaking down of the separating membranes, and in some few instances observing the beginning spirem in the polar nuclei, con- cludes that there is a complete fusion of the chromatin contributed by the three nuclei, and that this is further proved by finding no incomplete fusions in later stages. In not a single specimen showing the formation of the primary endosperm nucleus was the writer unable to distinguish between the chromatin of the various nuclei that have contributed to this nuclear complex. From the view obtained in fig. 22 it could readily be concluded that a complete intermingling of chromatin has previ- ously occurred, but when the next section of the same primary endosperm nucleus is examined (fig. 21), such a conclusion is seen to be groundless. Whether or not the separating nuclear membranes have entirely broken down by the time of segmentation, the spirems remain dis- tinct, and, following segmentation, three groups of chromosomes collect upon the three arms of the tripolar spindle (figs. 23, 24). In none of the literature examined has such a stage been shown or reported, for, if such had, the idea of complete fusion or inter- mingling of chromatin material could not have been adhered to up to the present time. Since the chromosomes are very long and quite numerous (36), the writer was unable to follow definitely their final arrangement upon the bipolar spindle. From the appearance of fig. 25 it seems 1918] NOTHNAGEL—FERTILIZATION 155 that the group on the right side of the equatorial plate might be the group that has been pulled into line. Soon after this each chromosome splits longitudinally, as FERGUSON (4, 5) has reported for Pinus, and all trace of the individuality of the groups is lost for a time. It has been generally understood that in Lilium Martagon fusion and subsequent divisions did not occur unless the top was cut off from the bulb; but in the plants used in this investigation, in which this was not done, the ovaries showed many dividing endo- sperm nuclei and the sacs were in good condition (fig. 29). In Trillium grandiflorum in a sac of four dividing endosperm nuclei (fig. 28), and in Lilium Martagon in a sac of two dividing nuclei (fig. 29), as described previously, three groups of chromosomes are seen on the spindle. This corresponds to the condition of the second division of the oosphere, as FERGUSON (4, 5) has reported for Pinus, in which she notes that the second division is like the first, there being two spirems. Figs. 28 and 29 distinctly show the three groups, and if such a condition is normal the question arises whether the male and female chromatin remaining distinct is the cause of the mottled appear- ance of some hybrid endosperms as found in Zea Mays. As has been observed by many investigators upon chromosome count in endosperm when it consists of many nuclei, the number varies in the different nuclei, there no longer being the 3x number. If in some of the divisions, when there is not an equal distribution of chromosomes, which is common in endosperm divisions, one group should pass to one pole and two to the other, the chromatin brought in by the sperm would then be in one nucleus by itself, or with one of the polar nuclei, thus causing the mottled appear- ance in the endosperm as seen in Zea Mays. The earlier writers on double fertilization, STRASBURGER (17, 18), Morrrer (14), NAWASCHIN (15), and Ernst (3), and the latest investigator Sax (16), concluded that there was an inter- mingling or a complete fusion of the chromatin contributed by the sperm and two polar nuclei in the formation of the primary endosperm nucleus, and Ernst (3) further stated that he was unable to recognize at segmentation or in spirem the chromatin 156 BOTANICAL GAZETTE [AUGUST that had been contributed by the respective nuclei; but the writer, because of the numerous specimens showing the distinct spirems, the three groups of segments, and the groups on the tripolar spindle, is unable to accept these conclusions for the primary endosperm nucleus of Trillium grandiflorum and Lilium Martagon. What has been said for the chromatin of the primary endosperm nucleus of Lilium Martagon applies for the most part to that of the fertilized egg of Trillium grandiflorum and the early stages of Lilium Martagon and L. philadelphicum. After a careful investigation of L. Martagon and L. candidum, MortIer (13) reported in 1898 that there is a complete fusion of the male nucleus and the egg nucleus in the resting condition, the chromatin being in a fine network. If the subsequent steps are not followed out, such an interpretation could be made for Trillium. Figs. 1 and 2 show fertilized eggs in which the sperm is lying coiled ‘upon the egg, the male chromatin material being in coarser strands thanin the egg. In some instances the nuclear membranes separat- ing the nuclei break down early (fig. 3), while in others they persist for some time (fig. 4). It was the appearance of such stages as fig. 3, and some that will be spoken of later, that caused previous investigators (MotrierR, NAWASCHIN, SAX, STRASBURGER, and others) to conclude that there was a fusion to the extent that the individual components were not recognized. From sections showing the spirems (figs. 6-11) it appears at first sight that the interpretations of figs. 9 and 10 would be differ- ent from those of figs. 6 and 7; for in figs. 9 and 10 two spirems stand out distinctly, while in the other two there appears to be only one. If the reader will consider all the various angles from which the fertilized egg might be cut and all the various positions the two spirems might occupy within the cavity, it will be apparent that frequently the sections might be so cut that the dual nature of the spirems would not be seen. The significance of the contraction, similar to the second contraction of the heterotypic mitosis that occurs just previous to or during segmentation (figs. 11, 12), the writer is unable to interpret. There was no tripolar spindle ob- served in the fertilized egg, and from the appearance of fig. 13 it seems that only a bipolar spindle is formed and the second group 1918] NOTHNAGEL—FERTILIZATION 157 of chromosomes is pulled in upon it. In each group there are approximately six, the haploid number. In 1891 GUIGNARD (6) pointed out in Lilium Martagon that there were two spirems and that one-half of the chromosomes on the spindle were contributed by the male parent and one-half by the female parent. No drawings were made to substantiate these views, and Ernst (3) and Mortrer (13) apparently made it so conclusive that there was a complete fusion or intermingling of chromatin that GuIGNARD’s earlier views were discarded and prac- tically forgotten. In later papers GuIGNARD himself did not place much emphasis upon these earlier views. SAx (16) stated that a spirem was frequently found in the egg and sperm before fusion occurred (fig. 21), but says “the rare appearance of such cases as that of the spirem stage in the egg and male nuclei when their outlines are still distinct, is probably of little significance in this respect. It is probable that these nuclei subsequently fuse completely, because no later stage of incomplete fusion was found.”’ Many writers have looked upon the number of nucleoli present in the fertilized egg as an indication that fertilization has or has not occurred. This, as Ernst (3) has pointed out, is not a safe indi- cator, for, as shown in figs. 5, 6, 9, 19-23, the nuclei have already united and from two to many nucleoli are present. SAX (16) says “fig. 19 shows a stage where the common boundary has disappeared, the contents apparently mingled, and those from the male and female nuclei are not to be distinguished.” In case of the formation of the fertilized egg, as in the formation of the primary endosperm nucleus, the writer is unable to agree with Sax and the earlier writers that there is a complete fusion or inter- mingling of the chromatin of the egg and the sperm; the material so plainly shows that the two remain separate from the time of coming together until the formation of the daughter nuclei. What happens after that does not come within the scope of this paper. Conclusions From the stages that have been found in the fertilized egg leading up to the first division, and in the primary endosperm nucleus up to 158 BOTANICAL GAZETTE [AUGUST the time of the third division of the endosperm, it is evident that - they are analogous to those of Pinus, as reported by FERGUSON (4, 5), although nothing was observed that would correspond to those steps in fertilization, as reported by HutcHtnson (12) for Abies, which differed from those of Pinus. To the writer the finding of the separate, distinct spirems and the separate groups of chromosomes is added evidence that the chromosomes maintain their individuality from one generation to the next. In conclusion the writer wishes to state that, according to her interpretation of the word “fusion” as used by previous writers, there was meant a mingling of the male and female chromatin, so that all trace of the individuality of chromatin and chromosomes contributed by the respective parents was lost by the time of the first division. In this investigation no such fusion was found, but instead, an entrance of two or three masses of chromatin, as the case might be, into a more or less single nuclear cavity, the chromatin contributed by the respective parents remaining dis- tinct throughout the preparation for the first division. The writer is unable to state whether fusion in the sense of com- plete intermingling ever occurs after the completion of the first division. Summary 1. After the male nucleus and two polar nuclei come together, the separating nuclear membranes persist more or less until seg- mentation. 2. Three distinct spirems are formed in the primary endosperm nucleus. 3. A tripolar spindle, each arm with its group of chromosomes, precedes the formation of the bipolar spindle. 4. The three groups of chromosomes maintain their identity, at least until several divisions have occurred. ; 5. The nuclear membranes separating the egg and sperm nuclei disappear earlier than in the preceding case, but the two groups of chromatin remain separate. 1918] NOTHNAGEL—FERTILIZATION 159 6. Two distinct spirems, followed by two groups of chromo- somes, arise from the maternal and paternal chromatin in the fertilized egg. 7. There is no complete intermingling of chromatin at fertiliza- tion. To Professor D. M. Mortier of Indiana University, under whom the greater part of this work was done, I wish to express my appreci- ation for the helpful criticisms and help given; to Professor C. B. ATWELL and Professor W. WoopBuRN of Northwestern University for the courtesies and encouragement given while working in their laboratory; and to Professor C. J. CHAMBERLAIN of the University of Chicago for the suggestion of this problem. INDIANA UNIVERSITY BLoomincton, INp. LITERATURE CITED Lal . BLACKMAN, V. H., and Wetsrorp, E. J., Fertilization in Lilium. Ann. Botany 27:111-115. 1913. . COULTER, JoHN M., The endosperm of angiosperms. Bor. Gaz. 52:380- 386. Iort. - Ernst, A., Chromosomen reduction, Entwickelung des Embryosackes, und Befruchtung bei Paris quadrifolia und Trillium grandiflorum. Flora 91: 1-36. pls. I-6. 1902 4. FerGuson, M. C., The development of ae ig and the fertilization in Pinus Strobus. Rhi Botany 15:435-470. , Contributions to the knowledge of te ite history of Pinus, etc. Proc. Wash. Acad. Sci. 6:1-202. pls. 1-24. 1904. GuicNnarp, L., Nouvelles études sur la Shiantation: Ann. Sci. Nat. Bot. VII. 14:163-296. 1891 7c eon, Sur bes astharoktiides et la double copulation sexuelle chez les végétaux Angiospermes. Compt. Rend. 128:864-871. 1899. 8. , Nouvelles recherches sur la double cnoEe chez les végétaux ‘Aigicmpetnnés: Compt. Rend. 131:153~-160. 9. ———, L’appareil sexuel et la ces fécondation dans les Tulipa. Ann. Sci. Nat. Bot. VIII. 11: 365-387. , La double tdoniilbtion he les Cruciféres. Jour. Botanique 16: ees aes 1902. - Heatitey, Marcaret, A study of the life may of Trillium cernuum L. Bor. Gaz. 61:425-430. 1916. Nv w -% an . Lal - 160 BOTANICAL GAZETTE [AUGUST 12. Hutcurnson, A. H., Fertilization in Abies balsamea. Bort. Gaz. 60:457- 473- 1015. 13. Mortrer, D. M., Uber das Verhalten der Kerne bei der Entwickelung des Embryosackes und die Verginge bei der Befruchtung. Jahrb. Wiss. Bot. 31:125-158. 1898. , Fecundation in plants. Carnegie Inst. Publ. 15 15. NAWASCHIN, S., Neuen Beobachtungen iiber elcehiias a "Pritillaria tenella und Lilium Martagon. Bot. Centralbl. 77:62. 1899. 16. SAX, Kart, a arias in Fritillaria pudica. Bull. Torr. Bot. Club 43 2505-523. ‘ Pa cee "rE, Einige Bemerkungen zur Frage nach der déppelten Befruchtung bei den Angiosperms. Bot. Zeit. 58: 293-316. 1900. , Uber Befruchtung. Bot. Zeit. 59:1-8. 1901 ~ ~J 18. EXPLANATION OF PLATES III-V All figures were drawn with the aid of an Abbe camera lucida, with Bausch & Lomb 1.9 mm. oil immersion, and ocular 6. The magnification of all figures except fig. 29 is X1500; fig. 29 X420. The plates are reduced to two- thirds their original size. PLATE Ill Fic. 1.—Early union of egg and sperm in Lilium M. ee Fic. 2.—A little later stage in Trillium grandifloru Fic. 3.—Early prophase, separating membranes eae egg and sperm having disappeare Fic. 4.—A little later, but separating membranes still present. ie. 5.—An interrupted spirem in fertilized egg, spirem being more or less ’ beaded. Fic. 6.—Early spirem in fertilized egg. Fics. 7, 8.—Development of spirem. Fics. 9, 10.—Two spirems in each nucleus, line a—b separating chromatin contributed by egg and sperm respectively. Fic. 11.—Contraction at time of segmentat G 12.—Later, contraction still evident; ether in two groups, nuclear ‘easter having disappeared. PLATE IV Fic. 13.—Formation of bipolar spindle, second group being pulled into equatorial plate. 1G. 14.—Late telophase of first division. Fic. 15.—Two-celled embryo. Fic. 16.—Trillium grandiflorum; early coming together of three nuclei to form primary endosperm nucleus. BOTANICAL GAZETTE, EXVI PLATE III Z 5 i \ \ y f - 2. a : re \ : . *, 3 Es \ f ree * 4 >. i. | he tf a ae 1 fort -* e 3 wi . % es * a . s t oe iw afer ] mee g . ae ' e ¢ 2 4 7 +p Pi : \ 12 NOTHNAGEL on FERTILIZATION PLATE IV BOTANICAL GAZETTE, LXVI NOTHNAGEL on FERTILIZATION BOTANICAL GAZETTE, LXVI PLATE V OF 98 NOTHNAGEL on FERTILIZATION 1918] NOTHNAGEL—FERTILIZATION 161 Fic. 17.—Lilium Martagon; same stage as fig. 16, the three nuclei being curved, 8 being more curved and chromatin coarser. 1G. 18.—Later; fibers are appearing about nuclear complex. Fic. 19.—Showing two components of primary endosperm nucleus in which the decusatis is collecting in heavier strands and separating nuclear mem- branes are still evident. 1G. 21.—Late spirem; separating nuclear membranes still divide the three spirems; fibers wie sues to races out into cyroplasit. IG. 22.—Next endosperm nucleus as seen in fig. 21. r J PLATE V Fic. 20.—Formation of spirem, the three nuclei still separate. 1G. Loe —Three groups of chromosomes upon a tripolar spindle. Fic. 24.—Slightly later stage; third group is being pulled i in with other two groups to form bipolar spindle. G. 25.—Bipolar spindle of first sili of primary endosperm nucleus, third _— of chromosomes still separa G. 26.—Slightly later stage; all pr ria of groups of chromosomes lost; “htenicaiiies beginning to split longitudinally 1G. 27.—Telophase of first division of primary endosperm nucleus. Fic. 28.—Third division of endosperm nucleus in Trillium grandiflorum, showing 3 groups of chromosomes upon bipolar spindle. Fic. 29.—Second division of endosperm nucleus, upper one showing three groups of chromosomes. FACTORS DETERMINING CHARACTER AND DISTRI- BUTION OF FOOD RESERVE IN WOODY PLANTS EDMUND W. SINNOTT (WITH TWO FIGURES) Introduction The investigations of Russow, FiscHer, and others upon the character and seasonal changes of the food reserves in woody plants, and the considerable attention which this problem has more recently received, have made us familiar with many of the important facts which it involves; but as to the underlying causes which determine the type of reserve food occurring in any cell and which direct its changes i in form and location we are still uncertain. Our present knowledge, derived in greater part from a study of twigs, branches, and small trunks and roots, may be summarized substantially as follows. The major part of the reserves stored up by trees and shrubs during the productive season is evidently composed of starch (fat is also demonstrable, and SaBLon (5) has emphasized the importance of reserve cellulose as a center of storage). At about the beginning of winter there is a decided reduction in the amount of starch, leading to its disappearance in the phloem and cortex of practically all woody plants in our latitude. At the same time the amount of fat seems to increase greatly. In certain forms, called by FiscHEr (2) “‘starch trees,’ there is no further change, the food reserves in the pith and wood persisting in the form of starch throughout the winter. In others, called by him ‘‘fat trees,” the starch vanishes in these portions of the stem as well, and fat appears in abundance, constituting the only visible food reserve during the winter. In all woody plants, late winter or early spring sees a regeneration of starch throughout the tissues of the stem and an apparent diminution in the amount of fat. This regenerated starch is used up in the formation of the spring growth, and it is not until summer that a fresh supply begins to be deposited. It has been supposed that at the seasonal changes Botanical Gazette, vol. 66] [162 1918] SINNOTT—FOOD RESERVE 163 starch was converted directly into fat or fat directly into starch, but as microchemical methods have been employed almost en- tirely this cannot well be proven. The only work involving a quantitative analysis, that of NrkKLEWSKI (3), seems to indicate that changes in the two types of reserve food occur independently of each other. It has been observed that the seasonal changes are most marked in twigs and small branches, less so in main stems, and least of all in roots, where fat is scarce and starch persists practically unaltered throughout the winter. The work of Fas- RICIUS (1) seems to indicate that in the large trunks of spruce conditions may be different from those in small trunks, branches, and twigs, and that starch there may have its maximum in winter and fat its maximum in summer. That temperature is of importance in producing changes in the character of the food reserves is shown by the fact that starch regeneration may be induced in the winter by bringing twigs from out-of-doors into a warm place. That a subjection to cold during the summer will not cause the characteristic winter changes, how- ever, and that these changes will nevertheless occur in the fall, even though the plants remain under a warm environment, indicate that factors other than temperature must be operative. The present paper is an attempt to throw light on this general problem by a careful anatomical study of the storage regions of woody plants with a view to determining the exact distribution of starch and fat there and its change from season to season. It contains the results of nearly three years’ observations on about 300 species of trees and shrubs belonging to over 100 genera and including all the common species of the northeastern United States, together with many exotic ones in the collections of the Arnold Arboretum. With the exception of a little received from the southern states, all the material studied was gathered in Massa- chusetts and Connecticut. Special attention was paid to conditions in twigs and young branches, where seasonal changes are most marked. Thin sections of freshly gathered material were cut on the microtome and treated with iodine and Sudan III to bring out the starch and the fat, respectively. 164 BOTANICAL GAZETTE [AUGUST Observations The results of previous workers as to seasonal changes were substantially confirmed. Although fat is evidently most abundant in the winter months, it is by no means absent during the summer, but at that time it is apt to be masked by the starch. Micro- chemical evidence as to the relative abundance of either starch or fat at different seasons is necessarily unreliable. It is certain, however, that much of the starch which disappears in the fall does not become converted into fat, but changes to glucose or some other non-visible substance, since in many starch trees large numbers of cells are emptied of starch without causing the appearance of fat. The twigs of some trees, notably species of Catalpa, are almost emptied of visible food reserves of all sorts during the winter. There are marked differences between species in their ability to produce fat, as indicated by its abundance in the phloem and cortex. This type of food substance seems to be practically absent in species of Carya and is very small in amount in Fraxinus, Acer. Syringa, and others. It is particularly abundant in such forms as Liriodendron, Populus, and Pinus. In general, fat is less abundant in the phloem and cortex of starch trees than of fat trees. There are certain exceptions to this rule, however; notably Lirio- dendron, a starch tree, but rich in cortical fat; and the soft birches, fat trees, but poor in cortical fat. Fat was universally found to be more abundant in the phloem than anywhere else in the plant. The observations of others that seasonal changes are more marked in twigs than in larger branches and trunks was confirmed. This conservatism is apparently still greater in the roots, where starch was found to be practically unreduced in amount during the winter, a fact recorded by Preston and Puttuips (4). In the root, too, the amount of fat is very much less than in the stem. FapRictus (1) and others have noted the fact that starch trees are predominantly hard-wooded species and fat trees soft-wooded ones. This rule was in general confirmed by the present study, but a number of exceptions were noticed which we shall later find to be significant. The hard pines, for example, are clearly fat trees; and Liriodendron, Magnolia, Ailanthus, and Platanus, all soft- wooded, are clearly starch trees. 1918] SINNOTT—FOOD RESERVE 165 Two other general relations between anatomy and the character of the food reserve were noted. Species with diffuse-pored woods are usually either fat trees or have an abundance of fat; those with ring porous wood are almost always starch trees. Narrow-rayed species may belong to either category, but broad-rayed types are prevailingly starch trees. By no means all the species studied could be classed definitely as starch trees or fat trees. The oaks, ashes, and hickories belong clearly to the former category, and the pines and lindens to the latter; but very many species are intermediate in character, possessing both fat and starch in the wood of the stem. In many instances, also, storage material was noted which was neither starch nor fat but seemed somewhat intermediate in character between the two. The outlines of the original starch grains could sometimes roughly be made out, but the starch content of the cell was appar- ently coalescing into an irregular brownish mass. This was insol- uble in ether and stained neither with iodine nor Sudan III. Its bulky, opaque character indicated that it was actually storage material and not merely the cytoplasm of the cell. It was evident chiefly during the winter, occurring frequently in the cortex as well as in the wood, in cells which had been filled with starch. SurRoz (6) called attention to the existence of such material, but apparently it has not been noted by others. If it is indeed a stage in the transition from starch to fat, its composition might perhaps throw light on the difficult problem of the chemistry of fat production in the cell. Table I presents a rough outline of the character of the food reserve in the pith and wood of the stem (twigs and young branches) of the more common trees and shrubs during the mid- winter months, dividing them into those where fat predomi- nates, those which possess considerable amounts of both starch and fat, and those in which starch predominates. This classifica- tion should not be regarded as rigid, since a considerable varia- tion has been noted in some of the species and genera, but it represents the average condition observed for each. The character of the reserve in phloem and cortex of course is not included in this table. 166 BOTANICAL GAZETTE [AUGUST The most noteworthy facts brought out by these anatomical investigations, however, concern the exact distribution of the reserve foods in the tissues and their changes from season to season. In al TABLE I TYPE OF FOOD RESERVE IN PITH AND WOOD OF STEM (TWIGS AND YOUNG BRANCHES) OF VARIOUS WOODY PLANTS DURING MIDWINTER PREDOMINANTLY FAT BoTH STARCH AND FAT Aescu reed Betula a Gome species) Cat Betula (some < pea Cornus (some species) Chamaecyparis Dir ie ig ae — species) onia pega (some species) Poa at us (some species) Pru Populus (most species) Rhos (cisat species) Pseudotsuga ee binia Rhus (some species) Taxus Sasha Tilia Viburwath (some species) Tsuga Viburnum (some species) PREDOMINANTLY STARCH Acer Itea Ailanthus Jamesia Berberis Kalmia Carpinus Lindera arya Liquidambar Castanea Liriodendron eltis Loni eo ag aa Magnolia Cladras yssa Cornus ae species) Philadelphus rataegus Platanus Deutzia Quercus Diervilla Rhamnus Diospyros Ribes Elaeagnus Evonymus eg agus Styra Fraxinus Symphoriarpos Gleditsia Syri Hamamelis HY eeog Hydrangea Vitis x Xanthoxylum species starch disappears in the fall almost completely from phloem and cortex, and even in the starch trees it is much reduced in the wood as well. The reduction in the wood takes place first and most extensively in the regions immediately around the vessels. In many 1918] SINNOTT—FOOD RESERVE 167 species starch is poorly developed here even in summer. It is in these regions, too, that the “‘transitional’’ material is very apt to appear. Furthermore, even in typical starch trees the wood paren- chyma cells or ray cells which directly adjoin a vessel frequently contain fat; and in species where both starch and fat occur in the wood the fat is conspicuously abundant near the vessels. In the medullary rays of such forms as most of the poplars and willows, for example, fat is found chiefly in those ray cells which touch a vessel and starch in those which adjoin nothing but fibers. This tendency for starch to be absent and fat to be present in the imme- diate vicinity of the vessels is obvious in all woody plants in the midwinter season, and suggests that the character of the food reserve may be related in some way to the water supply. Another anatomical feature which is clearly associated with the kind of food stored in a cell is the character of the cell wall. Wher- ever this is strongly lignified, thick, and provided with few and small pits, starch tends to remain unchanged throughout the winter. When it is thin or provided with many and large pits, starch tends to disappear and fat to be abundant. Thus in the storage cells of phloem and cortex, the walls of which are quite unlignified, starch vanishes early and completely and fat is very common. In the heavily lignified, thick-walled pith cells which occur in so many species starch remains throughout the winter, and in such cells the reserve food is less modified than in almost any other part of the stem. In branch gaps, where such a pith meets the cortex, the line between the starch-containing and the fat-containing cells is abso- lutely sharp and coincides exactly with the line between the lignified and the unlignified tissue. A study of the vertical and ray parenchyma of the wood, the chief seats of food storage in the xylem itself, is particularly instruc- tive in this connection, and furnishes us with a definite anatomical distinction between starch trees and fat trees. Where these cells are thick-walled and have few and small pits, starch predominates; where the walls are thin and well pitted, fat predominates. In hard-wooded species (fig. 1), long noted as starch trees, the paren- chyma shares certain of the characters of the other wood cells and has thick, well lignified, square-cornered, and small-pitted walls. 168 BOTANICAL GAZETTE [AUGUST In soft-wooded species (fig. 2), on the other hand, well known as being prevailingly fat trees, the vertical and ray parenchyma is thinner-walled and less heavily lignified, and the cells of the rays tend consequently to be irregular in shape, with oblique or bulging end walls, an outline quite different from the prevailingly rigid and rectangular one of starch tree parenchyma. They are well provided with pits or are so thin as not to require pitting. Fic. 1.—Nyssa sylvatica, a starch tree: portion of medullary ray of wood seen in radial section as it crosses fibers (at left and right) and a vessel (in center); note thick-walled, squarish ray cells; small pits between ray cells and from ray cells to fibers; and large pits from ray cells to vessel. Of particular significance are the exceptions already noted to the general rule that there is a connection between hardness of wood and type of food reserve. The hard-wooded pines, for instance, are filled with fat, a circumstance evidently related to the fact that their ray parenchyma and resin canal epithelium (the only seats of storage in the wood) are unlignified and very thin-walled. In Liriodendron, Magnolia, Ailanthus, and Platanus, on the other hand, which are soft-wooded but which we have nevertheless observed to contain starch, the rays are made up of thick-walled rectangular 1918] SINNOTT—FOOD RESERVE * 169 cells precisely like those of starch trees. This type of ray cell is here evidently mechanical in its function, since all these species have wide rays which might collapse or be badly crushed were they not built of strong-walled cells. The vertical parenchyma of these soft-wooded starch trees tends to have thinner walls, and in Lirio- dendron, at least, it contains considerable fat. On the basis of these facts we are forced to conclude that the hardness of a wood affects the type of food reserve indirectly, ees ees | S| ae Tl Fic. 2.—Populus grandidentata, a fat tree: medullary ray of wood seen in radial section“as it crosses fibers (at left and right) and a vessel (in center); note thin-walled ray cells, with slanting or rounded ends, and large pits from ray cell to ray cell and from vessel to (marginal) ray cells. through its influence on the walls of the storage cells. In certain cases, however, this effect is evidently more direct. The hard- wooded species of Cornus, such as C. florida, are starch trees, and the soft-wooded ones, such as C. stolonifera, are fat trees, although there is no very striking anatomical difference between the paren- chyma cells of the two groups. The same fact is noticeable in the hard- and soft-wooded species of Viburnum and birches. In these Cases it is fair to assume that the high or low degree of lignification 170 BOTANICAL GAZETTE [AUGUST of the conducting and fibrous cells is shared by the parenchyma, and that it is the actual hardness of the wall rather than its structure and pitting which is related to the character of the stored food. We shall later suggest a cause for this relation. Conditions in starch fibers are of interest here. These are fiber-like, starch-containing cells, and are frequently found in the maples, willows, certain legumes, and other trees. Unlike the parenchyma cells, which are definitely connected with the water supply either by their position next to a vessel or tracheid, or by being linked therewith by other parenchyma cells, these starch fibers usually occur in the midst of non-conducting tissue, being surrounded by fibers of the ordinary type. They possess very small, frequently rudimentary, pits and usually are thick-walled. No instance has been observed by the writer where any reserve food but starch occurs in such cells, and this starch usually stains much more deeply with iodine than that of the ordinary parenchyma, indicating that its water content is lower. The persistent character of the food reserve in these starch fibers is evidently related both to their isolation from channels of water conduction and to their thick and pitless walls. Discussion These two main facts which our anatomical survey brings out, (1) that during the winter starch is commonest in regions remote from centers of conduction (both in xylem and phloem), and in cells with thick, well lignified, and small-pitted walls, and (2) that fat is most abundant close to vessels or tracheids, in the phloem, and in cells with thin or unlignified walls and large pits, at once suggest that the character of the food reserve depends primarily upon the ease with which water or substances carried in water have access to the storage cells. Where access is slow and difficult, the reserve remains in its.summer condition as starch. Where access is easy, it is converted to a greater or less degree into other substances, with the consequent appearance of fat. In storage cells the walls of which are unlignified, as in the phloem and cortex of all species, and the rays and canal epithelium of Pinus, starch is quite absent in the winter and fat is very abun- 1918] SINNOTT—FOOD RESERVE 171 dant. There is evidently no impediment to thorough diffusion in such tissues, and conversion may take place far from any center of water conduction. Where the wall of the storage cell has become heavily lignified, however, even though it is provided with pits (which in such cases are usually very small), diffusion seems to be much impeded, and the reserve food remains unchanged throughout the winter. This is well shown by the terminal cell in the medullary ray of a typical starch tree, which cell, filled with starch and surrounded by its lignified wall, abuts directly upon a starchless, fat-filled cell of the cambial region. There are small pits in the wall between these two cells, but the wall nevertheless seems to serve as an effective barrier between them in preventing rapid diffusion. The same circumstance may often be noted where a ray cell touches a vessel. Here the wall between the two is provided with many large pits (fig. 1), so that communication must be easy. The ray cell in this case is usually without starch in the winter and generally contains some fat. Its neighbors at either end, however, are often full of starch but contain no fat. The tangential ray cell walls, although provided with small pits, seem here also to be permeable with difficulty. This leads us to believe that heavy lignification of the wall is a decided hindrance to the ease of diffusion between cells, and that the pits in such a case are for some reason, perhaps because of their very small size, unable to perform their normal functions. If the wall of the storage cell is less heavily lignified, or is thinner or more abundantly pitted, entrance of water is evidently easier, and we have noted that in these cases starch is more completely converted and fat is more common. In some instances fat may be limited to the cells directly adjacent to a vessel. In others it may extend to adjoining cells, starch occurring only in the more isolated regions. In these cases fat may be observed extending in from the cambial region along the rays for a considerable distance, thus indicating that diffusion takes place between the cells of the phloem region and the ray cells and affording a marked contrast to condi- tions at the ends of the rays in starch trees. In the true fat trees diffusion is evidently still more easy, for fat occurs throughout the 172 BOTANICAL GAZETTE [AUGUST rays and parenchyma, even in regions remote from centers of conduction. . In the cases of Cornus, Viburnum, and others which we have noted, where the degree of lignification of the walls of the storage cells seems to be a factor which determines the character of the food reserve, it probably operates by rendering easy or difficult the diffusion of water into the cell. That the starch in starch fibers is never converted into fat is evidently due to the fact that there can be little or no communica- tion between them and the water-conducting elements. The: willows are illuminating in this connection. Here the phloem, cortex, and medullary rays contain abundant fat in the winter and very little starch, so that the willows have often been regarded as fat trees. In the last annual ring, however, there frequently occur large numbers of starch-filled fibers, so that some writers have included the willows among starch trees. These fibers are almost absolutely pitless. Diffusion of water among them must thus be a slow process, a fact which probably explains this persistence of starch here where it is lost elsewhere in the wood. This hypothesis, that the character of the food reserve in a cell is dependent primarily upon the ease with which water or substances carried by water can pass from cell to cell, thus makes more under- standable the various facts which have been observed as to the type and distribution of reserves in the stems of woody plants from ' season to season, and is the contribution of anatomy to the problem under discussion. With this as a basis we may allow ourselves to speculate a little as to just what factors are operative in causing the seasonal changes in the food reserve. That change of temperature alone is quite insufficient to account for these is shown by the fact that they take place in plants kept over winter in the greenhouse. The writer has also observed their occurrence in trees growing in the frostless area of the Gulf states. The changes are doubtless due in the last analysis to the action of enzymes, presumably diastase and lipase. There are evidently two quite distinct series of pro- cesses, those concerned with changes in starch, and involving the action of diastase, and those concerned with changes in fats, involv- ing the action of lipase. 1918] SINNOTT—FOOD RESERVE 173 Two ways at once suggest themselves through which ease or difficulty of diffusion might affect this enzyme activity. First, the water content of the cells may be modified, those to which water has easy access having a higher content than those which are more isolated or protected. That such a condition actually occurs is indicated by the fact that the starch grains in cells near vessels usually stain more lightly with iodine than the others, showing that they possess a higher proportion of water. Differences in water content doubtless affect the whole physiological activity of the cell and may well determine the type of enzyme action. As to why, on this supposition, there should be such radical seasonal alterations, however, is not clear. There are doubtless changes in the water content of the tissues after leaf-fall and again at the spring awaken- ing, and these changes will of course be felt most by those cells to and from which diffusion is easy. In the case of fat, at least, we know that abundance of water favors lipolytic action, and paucity of water favors the synthesis of fat, facts which probably help to determine the increase or decrease of fat with the seasons. A second suggestion is that the enzymes themselves are carried by the water as it diffuses through the tissues, and thus effect the characteristic changes in the cells which they enter. That these changes take place in the fall we may perhaps ascribe to the presence of enzymes in the sap which is withdrawn into the tissues of the stem from the leaves before the latter are shed. The enzymes would thus be particularly abundant in the phloem, the ordinary channel of conduction from the leaves downward, and they probably would occur in the water of the vessels. They would be progres- sively less common in those parts of the plant farthest from the leaves. This would explain (1) why the changes are most marked in the phloem'and adjacent regions and around the vessels; (2) why they are more marked in twigs than in branches and trunks; and (3) why in roots they are practically absent. _ To determine whether or not lipase is actually present in leaves, a series of experiments was undertaken. The leaves of a number of Species of trees were gathered in the latter part of summer, dried, and finely pulverized, and the leaf powder of each was mixed with olive oil and bottled up, a number of bottles being made for each 174 BOTANICAL GAZETTE [AUGUST species. The proportion by weight of powder and oil was the same in every case. These were kept at room temperature, and once a week a bottle of each species was taken, the oil was removed from the powder by filtration and was titrated against N/1o sodium hydroxide. The fate of increase of acid measured the strength of the lipolytic action and hence, probably, the amount of lipase. Results showed the ferment to be most abundant in the leaves of those species in which fat was commonest in the stems in winter. To determine the exact method by which these seasonal changes in the food reserve are effected, however, is beyond the scope of the present paper, the purpose of which is to emphasize the important part evidently played by the minute anatomy of the stem and root in determining the ease of diffusion of water or solutions through- out their tissues and thus affecting the character, distribution, and seasonal alterations in the stored foods, and doubtless in other functional activities of the plant. We may point out that in any such physiological problem as this one a thorough knowledge of the structures concerned is absolutely essential before sound con- clusions can be reached. Summary 1. Previous observations upon the character, distribution, and seasonal changes of the food reserves of woody plants in temperate regions were in general confirmed by the present investigation and were considerably extended. 2. Astudy of the minute distribution of the food reserves in the tissues of the stem (twigs and young branches) during the winter shows that (1) starch is commonest in regions remote from centers of conduction and in cells with thick, well lignified, small-pitted walls; and (2) fat is most abundant in and near the phloem, close to vessels, and in cells with thin or unlignified walls or large pits. 3. These facts indicate that the character of the food reserve in any cell depends primarily upon the ease with which water or substances carried by water have access to the cell. Where the movement of liquids is apparently slow and difficult, the reserve persists as starch; where such movement is easy, starch disappears at the beginning of winter and fat is produced. 1918] SINNOTT—FOOD RESERVE 175 4. This suggests that differences in the type of food reserve may be due to (1) differences in water content of the various storage cells, resulting in modification of enzyme activity, or (2) differences in the ease with which enzymes have effective access to the storage cells. The writer wishes to thank the authorities of the Agricultural Experiment Stations of Louisiana and of Florida for material from their states, and the authorities of the Arnold Arboretum for speci- mens from their extensive collections of trees and shrubs. For several helpful suggestions he is also indebted to Professor W. J. V. OsTERHOUT of Harvard University. CONNECTICUT AGRICULTURAL COLLEGE SToRRS, CONN. LITERATURE CITED eal - Faprictus, L., Untersuchungen iiber den Starke- und Fettgehalt der Fichte auf den oberbayerischen Hochebene. Naturw. Zeit. Land. u. Forstw. 3: 137. 1905. - Fiscuer, A., Beitrige zur Physiologie der Holzgewichse. Pringsheim’s Jahrb. 22:73. 1890. - NIKLEWSKI, B., Untersuchungen iiber die Umwandlung einiger stickstof- freier Reservestoffe wahrend der Winterperiode der Baume. Beih. Bot. Centralbl. 19:68. 1906. 4. PREsTON, J. F., and Purtuips, F. J., Seasonal variation in the food reserve of trees. Roient ay 9: aga: sede - SABLON, DU, LECLERC, R i les mati d des arbres. Rev. Gen: Botanique 16: 386. 1 mueo2, J. .4,, Oel als Reservestoff der Scie: Beih. Bot. Centralbl. 1:1891. N w on - BRIEFER ARTICLES MODIFIED SAFETY-RAZOR BLADE HOLDER FOR TEMPERATURE CONTROL (WITH ONE FIGURE) The apparatus devised in this laboratory for cutting frozen plant material on the rotary microtome! has been found useful in the cutting of paraffin sections also, especially when a modification of the familiar safety-razor blade holder is employed. For the control of the tempera- ture of the knife in cutting paraffin sections LANp? describes and figures a trough of metal provided with a nipple at either end in which the micro- tome knife is placed and through which water of the desired temperature is made to flow. The use of a Gillette safety-razor blade in a proper holder is apparently becoming rather general among plant cytologists. Certainly the use of such blades with classes in microtechnique is a great saving of time and energy as compared with attempts either to allow the students to sharpen microtome knives themselves or to pro- vide them with such knives properly sharpened. In addition, with such a class, consistently successful results in cutting can only be obtained if it is possible to regulate the temperature of the knife and, in some cases, of the material also. Even at rom with refractory material in 52° paraffin it is difficult for most students always to obtain a smooth ribbon at ordinary laboratory temperatures unless the knife is kept cool. To meet this latter situation a simple modification of the usual type of safety-razor blade holder has been employed with such success in this » laboratory that a brief description of it seems desirable. We have found the original holder made by StricKLER’ the most desirable type of a number at present on the market. To such a holder a small brass tube is attached, as shown in fig. 1. This tube has a bore of 4 mm. and is soldered to the outer leaf of the holder, thus in no way interfering with the separation of the leaves when the safety-razor blade is to be inserted. The tube is extended approximately 6 mm. beyond the holder proper at either end to allow the attaching of small rubber tubes. * Garpner, N, L., A freezing device for the rotary microtome. Bort. GAz. 63: 236-238. 1917. 2LanD, W. J. G., A method of controlling = -— of the paraffin block and microtome knife. Bor. Gaz. 5'7:520-523. 3 CHAMBERLAIN, C. J., Methods in plant dae Chicago. 1915 (p. 9). Botanical Gazette, vol. 66] [x76 1918] BRIEFER ARTICLES 177 For class use, where very thin sections are not ordinarily required, we have found that the temperature of the knife in such a holder is sufficiently low if tap water is allowed to flow through the tube. A very short time is required for the temperature of the water to be com- municated to the knife. A cooling cell such as LAND’s or GARDNER’S also regulated with tap water may be employed in addition, but its use ase ... - 55.8 4.98-9.52 9.39-12.41 Pentose Pentosans pagent ° ae ty eae . 1, AU 965652..55. 0. 36-0. 52 5.19-5.96 O.13~). 72 4.70~ 5.32 I, September i -.++| 0.34-0. 76 4.42-5.90 0.94-1.60 9.98-16.08 IH, October rr... 0.61-0.92 6.21-7.15 1.14-1.93 14. 50-20.99 Information as to the translocation of the sugars was obtained by making separate analyses of the midribs and petioles. At any given picking the amount of sugars and alcohol soluble matter is always greater in midribs than in leaves, greater in top halves of petioles than in midribs, and greater in lower halves than in top halves of petioles. In the first series, the total amount of hexoses and of apparent levulose in the stalks increases very rapidly during the forenoon to reach a maximum at noon, while the corresponding increases in dextrose and saccharose are extremely ny dextrose being actually larger in amount at midnight than at any time during the day. In the bottom halves of the petioles of the series, total hexose, apparent dextrose, ~ apparent levulose run very closely together, reaching a maximum at noon, declining steadily to midnight, and then separating, as levulose continues to acca while the others start upward again. Saccharose rises slightly from 6 A.M. to noon, and then remains stationary for the succeeding 18 hours. In the second series, 1918] CURRENT LITERATURE 281 in which top and bottom halves of petioles were not analyzed separately, saccharose was constant throughout the 24 hours; total hexoses and apparent levulose were least at 11 P.M., increased slowly to 4 A.M., then more rapidly until 4 P.M., when they again declined together until 11 p.m. Apparent dextrose rose from 4 A.M. until 4 P.M., then remained stationary throughout afternoon and night. In the midribs, however, tetal hexoses decreased slowly from 10 A.M. to 4 P.M., then more rapidly through the night, rising again at 4A.M. Apparent levulose decreased, apparent dextrose increased, from 10 A.M. to 4 P.M., after which dextrose rather rapidly fell off while levulose slowly increased until 4 A.M., when both began to increase. Saccharose was stationary from 10 A.M. to 4 P.M., then increased slowly and uniformly through the evening and night, beginning to fall at 4 a.m. There is, therefore, a steady movement of sugars from leaves to midribs, thence through the stalks, the maximum in leaves at 2 A.M. moving onward into the stalks to give a maximum there at 6 A.M., which is succeeded by a minimum 4 hours later, when a large part of the sugar formed during the insolation of the preceding day has passed from stalk to root. The ratios of invert sugar to cane sugar at any given hour of the day, as at 6 A.M., September 10, when it is 1.48 in leaf, 3.32 in midrib, and 5.27 in stalk, are significant, showing as they do that there are progressively more and more hexoses in the stream of sugars as it passes from leaf to root. On August 26 the stalks had at noon 4.25 per cent saccharose and 11.57 per cent hexose; at 10 A.M., September 10, 4.82 per cent saccharose and 20.5 per cent hexose; and at 11 A.M., October 11, 5.29 per cent saccharose and 25.7 per cent hexose. This is strong evidence that hexoses are translocation forms produced by the conversion of cane sugar, as is the fact that cane sugar greatly predominates in the leaves in the early stages of growth, prior to the beginning of storage in the roots. Further evidence is seen in the fact that cane sugar is the predominant sugar in the leaves of the potato, vine, and snowdrop, although these plants store carbohydrate as starch, dextrose, and inulin respectively, and do not store cane sugar. Cane sugar is therefore formed in the mesophyll, transported into the vessels, undergoes progressive inversion as it passes onward throug midribs and stalks, enters the roots as reducing sugars, and these are there transformed once more into saccharose. The authors have not studied the mechanism of this synthesis in the root; invertase was shown to be present in the sieve tubes but was not found in roots by ROBERTSON, IRVINE, and DoBson, and it is believed to be the agent in the inversion occurring during transport. Since the existence of the saccharogenic enzyme of BorDET has not yet been substantiated, and the probability of reversible zymohydrolysis by invertase is contra-indicated by the absence of invertase from the roots, the authors are unable to formulate a theory as to the agent responsible for this synthesis. In the second paper of the series Davis reports the result of a study of the msg ble ratio in the mangold. The determination of these sugars polarimetric methods is falsified by the presence of optically active sub- edhe not a by basic lead acetate.. Glutamine, glutaminic acid, 282 BOTANICAL GAZETTE [SEPTEMBER and aspartic acid give a dextro-rotation which is increased by acids, while asparagine gives either dextro- or laevo-rotation accordingly as the solution is or is not acid. In the mangold and sugar beet there is an apparent excess of dextrose over levulose which is due to the presence of glutamine, while in snow- drop, tropaeolum, and potato the presence of asparagine results in an apparent s transfer of the impurity; in the leaves the dextrose-levulose ratio is in the neighborhood of unity; in the midribs and stalks it ranges from 2.5 to 10.0. The pentoses which are present in the alcoholic extract also affect the readings. The author determined the proportions of the two sugars present in the three series discussed in the preceding paper, using the methods there employed. In the early morning there was found in young leaves a dextro-rotation still greater than that which would be observed if all the sugar present were dextrose. In older leaves there appeared to be a steady formation of a laevo-rotatory substance and a gradual transformation into a compound having still greater laevo-rotation. The author considers that this is manufacture of asparagine and transformation into aspartic acid. Both in the second and the third series there are three rises and three falls in the amount of apparent dextrose in 24 hours, this fact pointing to a regular and rhythmical variation in the rate of production of the optically active impurities. The character of the optically active impurities in the upper portions of the stalks is quite different from that in the lower portions, as shown by the fact that when determinations of the sugars in the lower portions of the stalks are made simultaneously by polarization and reduction methods the polariza- tion results are 4o per cent higher than those obtained by reduction, while on the upper portion of the same stalks the results by polarization are 85 per cent lower than the reduction figures. Hence the optically active substances inter- fering with the polarization are quite different in the two portions of the stalk, suggesting the optical behavior of d- and ]-asparagine and d- and 1-glutamine. Furthermore, there are two different optically active substances at different times during the 24 hours. For all these reasons we can at present obtain no true values for these sugars, and there is at least nothing to disprove the assumption that dextrose and levulose exist in the leaves and stalks as invert sugar, travel in approximately equal amounts to the roots, and there undergo recombination into saccharose. In the third paper of the series Davis and SAWYER have e applied similar yeasts were made. eae the authors consider that BRowN and Morris were 1918] CURRENT LITERATURE 283 incorrect in their conclusion that diastatic formation of maltose and transfer as such occurs in the case of the leaf of Tropaeolum. Brown and Morris unquestionably had maltose present, as shown by the fact that they obtained the osazone and that there was an increase in the reduction of copper after treatment with maltase, but the authors consider that this result may be explained by the fact that the material used by BRown and Morris was sub- jected to preliminary drying in an oven. They regard the leaf as having a mixture of enzymes analogous to that found in Aspergillus oryzae, and that it is therefore able to split maltose rapidly and completely to dextrose. They destroyed all enzymes instantly by dropping the leaves as they were picked into a mixture of boiling alcohol and ammonia. As maltase is easily destroyed by heating to 55°, it was first to go out of action in BRowN and Morris’ oven-dried material, while other more heat-resistant enzymes went on forming maltose which was not split up, hence was found in the analysis. This hypothesis is borne out by the results; Davis and SaAwveEr invariably found more starch in the leaves than did Brown and Morris, the amount always exceeding the sum of starch plus maltose found by the last-named authors. KLuyver employed a biochemical method, using Torula monosa to ferment the hexoses only, 7. dattilla to ferment the cane sugar and hexoses, leaving maltose, and found very small amounts of maltose. The authors consequently believe that starch degradation g down to hexoses with no stop at maltose; that plants must reduce sugars to this form before they can be utilized; and that the fact that the sugar in leaf stalks is largely hexose is thus explained, as is the presence of invertase in almost all plant parts. Dats found maltase wherever starch is found in leaves, but believes it to be an intracellular enzyme occurring in close proximity to diastase, hence never found in the vessels of the stalks. In the leaf saccharose is always greatly in excess of hexoses; in the stalks the reverse is always true. Hence saccharose must be the first product of photosynthesis and hexose a translocation form. The authors are led by unpublished work with a variety of other plants, such as sunflower, grape, and snowdrop, to the conclusion that this is the general situation with all plants regardless of the form in which final storage may occur. Like the potato, the plants just mentioned have two optically active impurities which are formed at different periods in the 24 hours, and hence have apparent large fluctuations in the dextrose-levulose ratio, which it is impossible to measure correctly by reason of their presence. The authors found in the leaves considerable amounts of dextro-rotatory, water-soluble material which was not soluble starch or dextrin, which was greatest in amount between 4 and 8 p.m. Its period of greatest formation synchronizes with the high tide of saccharose and the period of most rapid starch formation, hence it seems to be intermediate between the hexoses and true starch. Starch is very rapidly reduced after sunset, then more slowly with the hexoses rising, while the starch rises again at dawn considerably before ph Sg ew | 284 BOTANICAL GAZETTE [SEPTEMBER the hexoses show increase. The curves for hexose, starch, and this dextro- rotatory material are intimately related and indicate interconvertibility; the last-named substance may be a protein or a gum standing in causal relation to starch synthesis. In the leaves the daily fluctuations of alcohol-soluble substances is through a range almost twice as great as that of total sugars. In the stalks the same is true, in which respect the potato is unlike the mangold. The dextrose-levulose ratio determinations are of little significance because of the presence of laevo- rotatory non-sugars, probably asparagine, but the authors regard them as being a presence of impurities of the same character. Levulose apparently pre- dominates in the leaves and dextrose in the stalks, by reason of the accumula- tion of dextro-rotatory stuffs in the latter, or possibly by reason of an actual excess due to the using up of levulose in tissue building. That this latter alternative is the correct one is indicated by the fact that the determinations of cane sugar by polarization and by reduction are in close agreement.— JoserH S. CALDWELL The Oenothera situation—Three recent papers have cast some light on the perplexing Oenothera situation. One of the most serious objections to the mutation theory has been that mutants which have appeared under observa- tion in artificial cultures have regularly been interfertile, while incipient ash in nature are essentially intersterile. Mertz and BripcEs‘ have shown that mutants may be intersterile, describing two cases in Drosophila, Hig cain two mutants fia either refuse to cross or else give sterile hybri MULLER’ has explained a curious case in Drosophila, w which strikingly resembles the Oenothera situation. A certain race of Drosophila breeds prac- tically true, and yet it isin a heterozygous condition. This paradox is explained by “balanced lethal factors,” a given chromosome and its allelomorph each carrying lethal factors. When one of these factors is present in a zygote it brings death, but when both factors are present they are antagonistic in their action and the zygote develops into a mature individual. Thus the homo- zygotes, which are thrown off every generation, die in infancy, since they con- tain single lethal factors; only the heterozygotes survive, for in them alone are the lethal factors balanced and inactive. The result is that the hetero- zygous race seems to breed true. This balanced race, as we should expect, gives in crosses twin hybrids as in Oenothera crosses, while crossing two such balanced races in Drosophila gives multiple hybrids, as also occurs in Oenothera. 4 Metz, C. W., and Briwces, C. = , Incompatibility of mutant races in Drosophila. Proc. Nat. Acad. Sci. 3:673-678. 1 5 MuLteR, HegMann J., An Pa eS case in Drosophila. Proc. Nat. Acad. Sci. 3:619-626. 1917. 1918] CURRENT LITERATURE 285 Another similarity with the Oenothera situation is that in this Drosophila race there would occasionally appear recessive mutants on one of these two “lethal chromosomes.” These recessive mutants, however, could not become manifest on account of the enforced heterozygosity. They could only become manifest when crossing over occurred and homozygosity was thus made pos- sible. ‘As crossing over occurs with predictable frequencies, those individuals showing characters abnormal to the stock were thrown continually in a definite, very small percentage of cases.” In just such a regular, although small, per- centage of cases does Oenothera Lamarckiana throw its mutants. LLER con- cludes that the Oenothera situation is to be explained by a similar mechanism, “but Pmemy the lethal effect in Oenothera is on the gametes rather than on the zygote A sioiitas idea appears in a paper by Davis,‘ in which we find summarized some of the evidence, old en sak on the suspected hybrid nye of Oenothera Lamarckiana. The regularity with which the same old mutants are thrown and the production a twin hybrids in crosses suggest to this aise the hybrid condition of O. Lamarckiana. The facts that about one-half of both pollen and ovules, in random distribution, are oie and that only 30-40 per cent of the seeds produced are fertile, suggest that only such gametes and zygotes are fertile as will reproduce the hybrid type. The argument is essen- tially similar to that of MuLLER. “If it could be shown that in every group of 4 pollen grains (tetrad) formed as the result of the reduction mitoses only 2 grains are perfect, the conclusion would be justified that pollen sterility was the result of this segregation division.’”” The author regards this as impossible, however, since abortion takes place after the tetrads have lost their identity. On this point we may quote from a review which appeared in this journal’ on some work of GrErts. ‘In Oenothera Lamarckiana 50 per cent of the ovules are found to degenerate and about 50 per cent of the pollen grains, ‘wo from each tetrad of spores.’ It begins to look more and more probable that our classic illustration of q are finally interpreted by the Mendelian system, and there is now much hope that this may soon come to pass.—MERLE C. COULTER Edible and poisonous mushrooms.—Popular interest in the fleshy fungi appears to be growing in many sections of the country. This interest may be attributed to several different causes, chief of which are to be found in the * Davis, B. M., A criticism of the evidence for the mutation theory of De VRIES from the behavior of species of Oenothera in crosses and in selfed lines. Proc. Nat. Acad. Sci. 3:704-710. 1917. 7 Bor. Gaz. 47:481. 1909. 286 BOTANICAL GAZETTE [SEPTEMBER availability of these plants as subjects for nature study and in the desire to add to the dietary a wholesome and palatable food growing without cultivation in forest and field. Here it is almost totally wasted as an article of food, because the comparatively small number of poisonous species cannot be distinguished from the many edible ones, for lack of the elementary knowledge necessary to recognize the more common forms. For this reason, and particu- larly at this time when there is a worthy desire to conserve every available item of food, nutritious or appetizing, it is gratifying that public institutions devoted to research and to the dissemination of useful information are recognizing the growing demand for popular instruction in the identification of edible and poisonous mushrooms. One of the most recent pamphlets =o to this subject is from the Illinois State Laboratory of Natural Histo There is an introductory chapter treating in a simple and clear manner ae the nature, structure, life- Dp described and illustrated by photographs. The arrangement of descriptive text and illustrations is such as to make the work very convenient for practical use by the amateur, and it is to be hoped that the effort of the author will succeed in a further stimulating interest in this group of fungi, often despised under the name of “musheroons,” and in leading its users to the desired knowledge of a satisfactory number of edible and poisonous kinds. Following the introductory chapter, two pages generally are devoted to a single species, one page to the descriptive text, and the opposite page to the photograph. As one reads the text the eye easily turns to the photograph in which most of the specific features can be verified he photographs are in general good, for many of the specific as well as the generic ne siieton as baw in pede To the reviewer, however, they seem to lack th hould be obtained from these plants. Whether this is due in all cases to a lack of care in the original photographs, or to faulty reproduction, is uncertain. The background in a number of cases is unnecessarily spotted, and in general the photographs appear ‘‘flat”’ and not well shaded. It is perhaps a matter of taste in which there may be reasonable differences of opinion, but it would appear preferable that the scale in the photo- graph should not occupy such an obtrusive position as it does in covering up parts of the plants, when it would serve as good a purpose if placed by the side. It appears that a few of the plants are not correctly named. For example, pl. 137 does not appear to be Clavaria cristata; pl. 119 is probably all Craterellus cantharellus; pl. 113 does not resemble Pholiota squarrosa, but rather a Hypho- loma, related we or identical with H. lachrymabundum. ‘The omission of Ama- nita ‘‘muscaria”’,a bi, poisonous species of wide distribution, should be noted. —GEo. F. ATKINSO § McDovcatt, W. B., Some edible and poisonous mushrooms. III. State Lab. Nat. Hist. Bull. 11:413-551. pls. 85-143. 1917. 1918] CURRENT LITERATURE 287 Economic importance of diatoms.—MAnw’ gives an interesting discussion of the uses of diatoms. Among these he enumerates the use of fossil diatoms as abrasives in polishing powders, tooth powders, etc. They have also been used as a food adulterant by mixing with flour, thus increasing the bulk of food but adding nothing to its nutritive value. They were used in this way by the “Earth Eaters.” A later similar use was as an adulterant of candy, but this use is now prohibited by law. They were also formerly used as an absorbent of nitro-glycerine in the manufacture of dynamite. There are beds of diatomite several hundred feet thick on the Pacific coast, and the use of them as a sub- stitute for asbestos in packing steam pipes, as filler for refrigerators, and in the manufacture of pottery is increasing. Another new use in medicine is as a filter forserums. It is suggested that their beautiful designs be used as patterns in the ornamentation of jewelry, wall paper, etc. ince diatoms store their food in the form of oil instead of starch, it is believed that they have been one of the sources of petroleum. On account of their being so minute that living ones may be carried great distances in the ocean, they may be of use in determining the direction of ocean currents. One argument that supports NANSEN’s theory that there is a current passing north- ward from Behring Strait across the north polar regions and down the coast of Greenland and Norway is that the diatoms of these localities are of similar species. Perhaps the one use that is of supreme importance is the furnishing of food either directly or indirectly for aquatic animals. Diatoms are chlorophyll- bearing plants, and are the greatest agency in the water for changing inorganic into organic matter, hence a knowledge of diatoms is fundamental to a study of the food supply of fish and other aquatic animals. Animal life is very abundant on the shores of the Antarctic continent, and in that region there is very little land vegetation. The greater part of the food for all of these animals is supplied originally by the diatoms. The statement that EHRENBERG estimated the number of individuals in a cubic inch of diatomite at 40,000,000 should be 40,000,000,000. The statement is made that diatoms are so minute that roo of them could be placed on the head of a pin. This is well within the facts, for that number of the smallest could find room on the point of a pin. The use mentioned of the diatoms Pleurosigma angulatum and Amphipleura pellucida as test objects for micro- scope objectives has been discontinued. The Bausch and Lomb Company State that the ‘Abbe test saga is now used entirely and is more ecrarate and reliable.—C. J. Erm Addisonia.—The second number of the second volume of this finely illustrated series, issued June 30, contains colored plates and popular descrip- tions of Solidago juncea, Echeveria multicaulis, Catasetum viridiflavum, Sagittaria latifolia, Baccharis halimifolia, Xanth texanum, Secum Bourgaei, Cimicifuga Simplex, Feijoa Sellowianus. and Aster amethystinus. ® MANN, ALBERT, a economic importance of the diatoms. Smiths. Rep. 1916: 377-386. fag I~. 288 BOTANICAL GAZETTE [SEPTEMBER The first number of the third volume of this journal, published by the New York Botanical Garden, contains colored plates and popular descrip- tions of Anonia atropurpurea (Eastern North America), Aster novaeangliae (United States and Canada), .Gymnocalycium multiflorum and G. Mostii (Argentina), Euonymus alata (Eastern Asia), Diospyros virginiana (Eastern United States), Lepadena marginata (Central and Western United States), Maackia amurensis Buergeri (Japan), Hibiscus oculiroscus (Eastern United States), Cornus officinalis (Japan), Opuntia lasiacantha (Mexico).—J. M. C. Morphology of wheat.—JENSEN” has investigated certain strains of wheat and the result is perhaps our fullest account of the morphology of this important plant. The subjects considered are development of spike and flower, of micro- spore and male gametophyte, of megaspore and female gemetophyte, fertili- zation and development of embryo, and endosperm. An interesting record is that fertilization occurred from 32 to 40 hours after pollination—J. M M. C. Intrafascicular cambium in monocotyledons.—Mrs. ARBER"™ has added to ' her previous observations” of rere cambium in monocotyledons other observations which inclu raceae, Dioscoreaceae, Iridaceae, and Potamogetonaceae. Such sdathibeec is now known to occur in “all but two of the cohorts into which ENGLER divides the monocotyledons; the exceptions are the Triuridales and the Synanthae.’”’—J. M. C Seed position and growth.—It has been found that bean seeds planted with the eye up give a somewhat lower degree of germination and growth than when the seed lies flat or is placed eye down.’ This seems to show that the common practice of dropping seeds flat upon the soil when planting gives results that are satisfactory —Gro. D. FULLER. 10 JENSEN, G. H., ee on the morphology of wheat. Bull. 150, State Coll. Washington. pp. 21. pls. 5. 1918. = ARBER, AGNES, sea notes on intrafascicular cambium in monocotyledons. Ann. Botany 32:87-80. figs. 4. 1918. Bot. Gaz. 64:350. 1917. 3 Hatstep, B. D., and Owen, E. J., Environment of seeds and crop production. Plant AWertd 30 20:294-297. 1917. VOLUME LXVI ~° ; NUMBER 4 THE DOTANICAL Gazer OCTOBER 1918 A CONTRIBUTION TO THE LIFE HISTORY OF IMPATIENS SULTANI ALIcEeE M. OTTLEY (WITH PLATES XIV, XV) This paper is based upon a study of slides made through a series of years for class use in the Botany Department of Wellesley College. The material was taken from greenhouse plants of the rose or bright pink variety of Impatiens Sultani Hook. The bright red and light pink varieties also were growing in the greenhouse, but care was taken to collect material from the rose-flowered plants only. Some of the plants from which the flowers were collected were chance seedlings. No attempt was made to deter- mine whether or not these were a pure strain of the rose-colored form. In Battey’s (4) Standard Cyclopedia of Horticulture the original form of J. Sultani is given as a rich scarlet, shades ranging from pink to almost purple being found on hybrids or sports. If this be true, then all the rose-colored forms used for this study are either hybrids or sports. According to BaILey the species J. Sudtani was originally found in Zanzibar and named by Hooker in honor of the Sultan of Zanzibar. In ENGLER and Prantt’s Die Natiirlichen Pflanzen- familien (16) it is cited from Sierra Leone, Western Africa. It is Stated in Gray’s Manual that the Balsaminaceae often contain two kinds of flowers, the large showy ones which rarely ripen 289 290 BOTANICAL GAZETTE [OCTOBER seeds, and small ones which are cleistogamous. J. Sultani is not given by ENGLER and PRANTTI in their list of species containing cleistogamous flowers, and I was unable to find any cleistogamous flowers on the many plants of this species which were investigated. The material fixed ranged from very small buds to young fruits. In preparing the buds for fixing, the smallest ones were put up entire; the sepals and petals were removed from all others; and in the largest buds the pistil and stamens were separated. From the flowers and the fruit only the ovaries were preserved and these were trimmed slightly at the angles to allow more rapid penetra- tion of the fixer, which in all cases was Flemming’s chromo-acetic solution. In general, the material was sectioned longitudinally and stained with Flemming’ s triple stain. Ovary The ovary consists of 5 carpels with axial placentation. There are several ovules in each loculus and the age of the ovules in a given loculus advances from base to apex, the youngest being at the base of the ovary. At the time when the microspore mother cells are in prophase of the heterotypic division the ovules appear as slightly curving outgrowths from the placenta, with or without any indication of the inner integument (fig. 1). It is apparent that the ovules of J. Sultani occur earlier in relation to the develop- ment of the anthers than is the case in many other plants. Miss Buss (8) reports that in Viola the ovule initials cannot be detected when the microspore mother cells are in the prophase of the hetero- typic division, and similar observations have been made by many other investigators. As the inner integument begins to appear, a single hypodermal archesporial cell becomes differentiated at the apex of the nucellus (fig. 2). I. Sultani agrees with I. pallida (Miss Ratt 37) in having but the one archesporial cell, but differs from that species in having no parietal cell cut off from the archesporial cell. According to CouULTER and CHAMBERLAIN (15) there is a general tendency to suppress the parietal tissue among monocotyledons and Archi- chlamydeae. ‘‘The suppression of parietal tissue among Archi- chlamydeae is most extensively displayed by the Ranunculaceae 1918] OTTLEY—IMPATIENS 291 and its allies rather than by the more specialized groups.” Bal- saminaceae, one of the higher groups of the Archichlamydeae, shows complete suppression of parietal tissue in J. Sultani; in I. pallida, however, Miss Rairr (37) describes a parietal cell, but her illustrations are not very conclusive. From the first the megaspore mother cell is the only hypodermal cell at the apex of the ovule. It is surrounded by the epidermis of the nucellus at its sides and apex, and is bounded by several nucellar cells at its base (fig. 2). The cell is slightly longer than wide and extends almost to the plane of insertion of the inner integument. It keeps pace with the growth of the ovule and con- tinues to occupy all of the nucellus within the epidermis except in the chalazal region. As growth continues it changes from a broad cell to a long narrow one (figs. 2-4, 6). The nucleus of the megaspore mother cell contains one nucleolus which at first is separated from the chromatic network by a clear area (fig. 2). As division is initiated, the chromatic reticulum forms a more or less complete spirem, separates from the nuclear membrane, collects about the nucleolus, and enters the synaptic stage (figs. 3, 4). On recovery from synapsis the spirem spreads out into the nuclear cavity and very soon exhibits the second con- traction stage (fig. 5) similar to that which has been described for Lilium and several other angiosperms by ALLEN (2), MoTTIER (32), and others. According to OVERTON (35), this phenomenon does not occur among the majority of the angiosperms. During this stage the spirem is thick, and at places uneven and massed. Here and there, in the less condensed areas, light streaks show, suggesting either a longitudinal splitting of a single spirem or an approximation of two. The contracted spirem is in contact with the nuclear membrane at several points, but the greater amount of the chromatic material is near the nucleolus at the center of the nucleus. At this time the nucleus is near the micro- pylar end of the megaspore mother cell. About midway between it and the chalazal end are numerous fibers with blue staining dots scattered among them. This characteristic has been observed in several megaspore mother cells, but its significance was not deter- mined. These fibers played no apparent réle in the formation 292 BOTANICAL GAZETTE [OCTOBER of the spindle or in the development of the cell wall, since they disappear before the first division. However, they may represent an early assembling of the kinoplasmic substance preparatory to spindle formation. Fig. 6 shows the spirem partially segmented with the nucleolus _ near the center of the cavity and the segments projecting in from the nuclear membrane. Later the chromosomes become very short and group themselves about the nucleolus (fig. 7). At this time they suggest the tetrads described for many animals and closely resemble those of Arisaema triphyllum as figured by ATKIN- son (3). A typical bipolar spindle is formed and the bivalent chromosomes when arranged at the equatorial plate appear as very short X’s, V’s, and Y’s. Two cells separated by a distinct cell wall result. Son the heterotypic division and form the axial row (fig. 8). The micro- pylar cell of this row is smaller than the other, and disintegrates very quickly. The chalazal cell grows and is the mother of the embryo sac. An axial row of two cells is not common among the Archichlamydeae. TrEuB (43) describes an axial row of two cells for Viscum articulatum, and several cases have been reported among the monocotyledons. Miss Raitr (37) states that 4 megaspores are formed in J. pallida, but makes no sketch showing them. In her fig. 1, J, she shows an ovule containing a large cell which she names the functional megaspore. Between it and the epidermis a small disintegrating cell appears. The sketch closely resembles the appearance of the ovules of J. Suliani with an axial . row of 2 cells and throws doubt upon her assertion that there is an axial row of 4 cells. The micropylar cell of the axial row is never large, and is so short-lived that it is easily overlooked, and the embryo sac seems to arise directly from the megaspore mother cell, as in Liliwm. As the micropylar cell disintegrates and the epidermal cells of the nucellus grow, there appears simply a small blue staining cavity between the embryo sac mother cell and the apical region of the epidermis, as cited by CouLTER and CHAMBERLAIN (15) for Clematis, and Helleborus (GUIGNARD 20), and Delphinium (MotTTIER 31). The chalazal cell grows and its nucleus divides, completing the 1918] OTTLEY—IMPATIENS 293 reduction division and giving rise to the 2-nucleate embryo sac (fig. 9). At this stage there is still something of the disintegrating micropylar cell to be seen, but at a slightly later stage it has entirely disappeared (fig. 10). The embryo sac is thus derived from two megaspores as in Viscum articulatum (TREUB 43) of the Archi- chlamydeae and in T: neue (HEATLEY 26) and several other monocotyledons. The two megaspore nuclei move to the opposite soe of the sac and divide (fig. 10). At this stage the sac is vacuolate and continues so until a late 4-nucleate stage (figs. 11-13). The 8-nucleate stage follows rapidly upon the four. Two-, 4-, and 8- nucleate stages have all been found in the same ovary, the 2-nucleate stage being at the base of the loculus. It is very easy, in serial sections, to confuse an early 8-nucleate stage with a late 4- nucleate stage, as the sacs have the same shape and cytoplasmic appearance. In Eriocaulon septangulare (SMitH 40) the central vacuole first appears at the 4-nucleate stage. In J. Sultani the late 2-nucleate sac is vacuolate with large vacuoles between the two nuclei, but there is not one large central vacuole until the female gametophyte has been organized (figs. 1o-14, 17). In an early 4-nucleate stage there are several large vacuoles extending along either side of the row of nuclei (fig. 11). During the 4-nucleate stage the sac enlarges, the cytoplasm becomes more dense, and the large vacuoles decrease (figs. 12, 13). By the time the 8 nuclei are formed the cytoplasm is very dense and contains a large amount of stored food, and the vacuoles have become small and inconspicuous (fig. 14). Very soon, however, a large central vacuole appears (fig. 17). The time of the inception of this central vacuole varies considerably. It may arise while the antipodal polar is at the base of the sac, or it may not appear until the polars are in contact and near the egg (figs. 14, 17, 18). After the organization of the 8 nuclei the egg apparatus soon forms. The egg is more or less pear-shaped, with the larger end extending down below the synergids. The nucleus and greater part of the cytoplasm are in this region and the narrowed part extends up back of the synergids and is vacuolate. It evidently 294 BOTANICAL GAZETTE [OCTOBER resembles the egg of Aster novae-angliae (CHAMBERLAIN 12) and that of many other plants as to shape and relation of nucleus and cytoplasm (fig. 17). The nuclei of the synergids do not always have the same posi- tion (figs. 15-17, 19). In the youngest sac of the series (fig. 15) there is a large vacuole at the base of either synergid, with the nucleus above and near the micropylar end. Fig. 16 illustrates a condition in which the synergid nuclei have moved down halfway, and in one cell there is a vacuole on either side of the nucleus, while in the other there is a large vacuole below it but only a small one above it. In fig. 17 one nucleus has moved entirely below the vacuole and is near the membrane at the end of the cell, whereas the other nucleus is still between two or more vacuoles. In a much older embryo sac (fig. 19) the synergids are longer, the nucleus of each is at its base, and a large vacuole appears above the nucleus. At all stages in the growth of the egg apparatus the synergids contain a fairly dense cytoplasm at the apex. It seems clear that the position of the synergid nucleus varies in relation to the age of the sac; that at first it is near the micropylar end of the synergid and above the vacuole; that later it passes the large vacuole and moves down to the opposite end of the cell. When the egg apparatus is mature the two nuclei are at the base of the cells, near the egg nucleus, and just below the large vacuoles (fig. 19). According to the literature on the subject the position of the synergid nuclei in different plants may vary in relation to the large vacuole of the cell, but I have found no suggestion that the vari- ation in a given species represents different stages in develop- ment. CHAMBERLAIN (12) gives the situation of the nuclei in Aster novae-angliae as varying in position from one end of the cell to the other but most frequently near the middle, the large vacuole being usually at the chalazal end of the synergid. GuIGNARD (19) says that in the Leguminosae the vacuole is usually at the base of the cell and the nucleus is central, but the vacuole may sometimes be above the nucleus. Barnes (6) in Campanula americana, GUIGNARD (24) in Hibiscus Trionum, and STRASBURGER (42) in Wikstroemia indica von Buitenzorg find the nucleus above the 1918] OTTLEY—IMPATIENS 295 vacuole; while PACE (36) finds the synergid vacuoles of Parnassia in various positions. The antipodals are surrounded by denser cytoplasm than is present above them, and they appear rarely as separate cells with delicate walls separating them (fig. 18), or, as is more commonly found, the mass of cytoplasm with the 3 nuclei is more or less cut off from the rest of the sac by a membrane but the cells are not separated. In either case they are but short-lived and dis- appear soon after the egg apparatus is formed. The embryo sac then persists for a long time with but 5 nuclei. Miss Rarrr says that the antipodals in J. pallida cannot be distinguished with certainty and are evidently transitory. This ephemeral nature of the antipodals is common among many of the angiosperms. In Striga lutea (MICHELL 29) the 3 antipodal cells begin to disintegrate before fertilization. In Richardia africana (MICHELL 30) the disintegration is somewhat earlier, evidently more nearly like I. Sultani. In this species the antipedals were never found to increase in size or number and grow into the adjoining tissue, as has been described by CHAMBERLAIN (12) and OpPERMAN (34) for Aster novae-angliae and by others for various plants. Very quickly after the 8 nuclei of the sac have been formed the antipodal polar moves up toward the micropylar polar and the 2 nuclei remain near each other at a short distance below the egg nucleus for some time. The nuclei may or may not be spherical, but they always contain a prominent nucleolus with a small highly refractive spot at the center. The chromatic substance of the polar nuclei is small in amount and forms either a delicate network lying just within the nuclear membrane, or a few strands radiating out from the nucleolus. Most of the food stored in the sac at an earlier period disappears before the female gametophyte reaches maturity and the sac becomes very vacuolate, with only a layer of cytoplasm at its periphery and surrounding the nuclei in the micropylar half of the sac (fig. 37). During the later development of the embryo sac its shape becomes much changed (figs. 17, 19, 37). While the antipodals are still present the sac is a little over three times longer than wide, with the micropylar and antipodal ends both rounded in outline 296 BOTANICAL GAZETTE [OCTOBER (fig. 17). After the antipodal cells disintegrate the basal region of the sac grows down into the chalaza. The growing portion is blunt or more or less triangular in outline, and but little narrower than the sac just above. The antipodal growth continues until the sac is over five times as long as it is wide (fig. 19). Following the development of this antipodal haustorium the sac widens in the region of the polar nuclei and assumes its mature shape (fig. 37). - During the origin and development of the female gametophyte the megasporangium hasbeen undergoing marked changes. The ovule begins to curve before the megaspore mother cell appears, and by the time its nucleus has reached the segmented spirem stage the ovule has attained the anatropous position. At this time the inner integument extends beyond the nucellus and forms a fairly deep micropyle (figs. 1-4, 6). The outer integument arises from the lower two-thirds of the inner integument, appearing as a swelling from the outer part of the latter (fig. 6). This swelling increases greatly in breadth and grows up until its apex is on a line with the tip of the inner integument. It never grows beyond this point to aid in forming the micropyle, and the two integuments become distinct only at the summit (figs. 9, 37). The origin of the outer integument in J. Sultani differs from the majority of plants and from the other species of Impatiens that have been studied. Miss Rarrr (37) in J. pallida and GUIGNARD (22) in J. parviflora both show the outer integument arising from the basal portion of the ovule and remaining through- out its length distinct from the inner. As a result of his study of I. balsamina, BRANDZA (g) states that the Balsaminaceae have but one integument, but BrunorTe (10) disagrees with BRANDzA and gives two integuments for this family. According to LoNGo,’ as cited by Miss Rarrt, this is the rule for the genus Impatiens, but the origin of the outer integument and its extension at the micro- pylar region appear to vary. From my study of the ovule of I. Sultant it can readily be seen how BRAnpza thought there was but one integument if I. balsamina is similar to J. Sultani in having the outer integument an outgrowth of the inner integument with only their tips free. tLonco, B., Richerche su le Impatiens. Annali Bot. 8:65-77. 1909. 1918] OTTLEY—IMPATIENS 207 In De l’ovule, WARMING (45) notes a few exceptions to the usual order of development of the integuments and gives Viola, Ficus, Convallaria, and Orchis as having two integuments which appear to grow as a single organ, and Tropaeolum as having at first two integuments which later appear as one. WARMING quotes STRAS- BURGER as saying that in Delphinium the integuments originate as one and elevate themselves as a unit; later at the summit the two integuments become distinct. Judging from his figure the conditions in Delphinium are much the same as in J. Sultani. As given earlier, the megaspore mother cell when it arises is completely surrounded, except at its base, by the epidermis of the nucellus, and the developing embryo sac also continues to lie in direct contact with the epidermis (figs. 2-4, 6, 8, 9). During the 2-nucleate stage of the embryo sac the epidermis begins to break down, The disintegration first appears as a flattening of the cells and nuclei just below the apex, and then extends gradually to the base of the embryo sac (figs. 9, 10, 14, 37). In Oxalis corniculata, according to HAMMOND (25), the epidermis, which in this case serves as a tapetum, begins to disintegrate before the 2-nucleate embryo sac is formed. The apical cells of the epidermis are often longer-lived than those just below them, for it is quite common to find two or three cells surmounting the embryo sac and connected by only a line with those still persisting about the center of the sac (fig. 12). A somewhat similar appearance has been described by SmiTH (40) for Eriocaulon septangulare, where “the nucellar tissue lateral to the megaspores breaks down and is absorbed by the growing embryo sac. A few of the apical cells of the nucellus persist for a long time and enlarging assume the appearance of a tapetum. These too are ultimately absorbed and the embryo sac abuts directly upon the inner integument and micropyle.”’ In J. Sultani, however, these apical sibaaeilasss cells do not enlarge and seem to have no special function. While the epidermis continues to disappear it leaves but a line around the upper half of the sac (fig. 10). As the disintegration progresses downward, the cells near the base of the sac possess their normal tabular shape, while those nearer the middle of the 298 BOTANICAL GAZETTE [OCTOBER sac are narrowed and pointed. The nuclear substance in the pointed cells is dense, the nuclei are often flattened, and the cell content stains but slightly. The pointed end of the layer often becomes free from the embryo sac (fig. 13). As this layer is dis- appearing it is not always in close contact with the tapetum (fig. 10), and I infer that the disappearance of the epidermis is due, not to its being crushed between the tapetum and the enlarging embryo sac, but rather to the fact that it is being absorbed by the embryo sac. The basal portion of the epidermis continues to exist until after the 8-nucleate sac is formed (fig. 14). Not only is the entire epidermis absorbed eventually but the nucellus beneath the embryo sac also. During the early stages in the development of the embryo sac the cells of this part of the nucellus are similar to those of the interior of the integuments, but beginning with the 4-nucleate stage, or occasionally earlier, they. become stringy in appearance, with their long diameters in line with that of the sac. This strand of cells extends down to the chalaza which is composed of a tissue of regular, compact, densely staining, isodiametric cells (figs. 12, 14, 37). Many of the nuclei of the “‘stringy”’ cells show signs of disintegration. They become dense, lose their rounded outline, and appear elongated. These cells are bounded at the sides by the tapetum (fig. 14). The antipodal region. of the sac absorbs this tissue and pushes down to the nutritive cells of the chalaza, thus completing the absorption of the entire nucellus (fig. 37). J. Sultani agrees with I. parviflora (GUIGNARD 21) and J. amphorata (LONGO 28) in having the embryo sac absorb the nucellus and thus come in contact with the micropyle and the inner integument. These species of [mpatiens, therefore, agree with the Compositae (GotpFLus 17) in the early disappear- ance of the nucellus. As described by GuiGNarp (22), Rarrt (37), Lonco (28), and BRuNOTTE (10) for Impatiens, and by others for various angio- sperms, the epidermis of the inner integument forms the tapetum of regular tabular cells, which, in J. Sultani, extend from the base of the micropyle to a considerable distance below the base of the developing embryo sac (figs. 9, 10, 14). The tapetum loses its uniform character as early as the 2-nucleate stage of the embryo sac, when a densely staining substance appears between the tapetal 1918] OTTLEY—IMPATIENS 299 cells near the base of the micropyle and the contents of these cells stain diffusely. This appearance also extends out laterally and at this time up to the tip of the inner integument (fig. 9). The cells are crowded and their cytoplasm and nucleoplasm stain so diffusely that no attempt was made to represent their appearance in a sketch. As will be seen later, these cells break down during endosperm formation, and this doubtless represents an early stage in their disintegration. This characteristic progresses chalazally in the tapetal cells as the embryo sac develops and reaches the basal cells during early endosperm formation. In an 8-nucleate stage the lower half of the tapetum is still normal and contains a decidedly granular cytoplasm and _nucle- oplasm which suggests the presence of food particles. This granu- lar appearance is not visible in the cytoplasm of the other cells of the inner integument, although they stain more densely than do those of the outer integument. GuIGNARD (22) says that a nitrogenous substance accumulates in the tapetal cells of J. par- viflora and that in all of the Balsaminaceae there is this proteid layer of tabular cells. BILLINGS (7) believes with most students that the tapetal layer, whether from the inner integument or the nucellus, serves a nutritive function, dissolving and absorbing nutriment from the surrounding integument, and that its function is not simply pro- tective, as given by HEGELMAIER (27). VANDENDRIES (44) in a study of the Cruciferae finds the tapetum a part of the inner integu- ment, but believes it plays only a protective function. One reason given is that in the antipodal region, where the tapetum is separated from the sac by a small mass of nucellar cells, it presents the char- acteristic appearance of young and active tissue. This reasoning oes not seem conclusive to me, since it may well be that this was a region of considerable food and that here the active cells of the ta- petum digested and absorbed it and then passed it up to the embryo sac. In J. Sultani the tapetum persists longer at the antipodal end and it is here that growth takes place until the sac reaches the chalaza and passes slightly beyond the end of the tapetum. - _Bauicka-Iwanowska (5) studied certain “Gamopetales” and described the tapetum. In agreement with CHopat (13) and most recent writers, he does not believe that the tapetum is for protection, 300 BOTANICAL GAZETTE [OCTOBER as it is wanting in the vicinity of the haustorium, which does not possess cell walls and would therefore appear in need of pro- tection. He thinks that the tapetal cells possess a ferment in their mucilaginous content and exercise a digestive function, for they persist while the neighboring tissue disintegrates, and they surround the parts which are in the process of rapid growth. Extending through the raphe is a strand of ceJls which is sur- rounded, except at the ends, by a layer of cells with cutinized walls. This strand terminates at the chalaza in the tissue of regular com- pact cells. It is into this that the antipodal region of the sac pushes (fig. 37). In looking at the figure it will be seen that the -outer layer of cutinized cells ends just as it passes this area, and on the inner side the layer ends almost in contact with the antipodal end of the tapetum. It thus forms a protective covering to the conducting tissue as it passes to the chalazal haustorium. No true vessels were ever observed in this strand of conductive cells. As noted earlier, the antipodal nuclei disappear before the haustorium develops, thereby giving rise to the unusual condition of a haustorium unaccompanied by nuclei. In none of the litera- ture studied was I able to find any record of a similar condition. In the cases where a haustorium has developed at the antipodal region before fertilization had occurred, it is usual for the antipodal nuclei to be present in the haustorium formed. An interesting example of this is given by SovEcEs (41) for the Solanaceae, where a pocket is formed at the basal part of the embryo sac and the antipodals take their place in the bottom of this and their digestive juices diffuse into the tissue beneath and dissolve out a cavity. The process of dissolution of the tissue varies among the different mem- bers of the family from a simple disjunction of the digestive layer of cells to a chalazal cavity whose capacity is comparable to that of the embryo sac itself, as in Lycopersicum esculentum. Stamen The flower possesses 5 stamens whose anthers are connivent and form a hood over the pistil. In a cross-section of a bud the sides of two adjacent anthers show as having their cells in contact, and this region appears as solid tissue. Each anther contains 4 micro- 1918] OTTLEY—IMPATIENS 301 sporangia. Fig. 20 shows a cross-section of one-quarter of an anther when the nuclei of the microspore mother cells are in the synaptic state. The walls of the cells of the epidermis are cutinized and thicker than those of the other cells. The wall of the micro- sporangium consists of two distinct regions, the outer irregular portion made up of 1-5 layers of nearly isodiametric cells, and an inner region about 2 cells thick, the cells of which are flattened tangentially. This flattening, doubtless, results from the pressure caused by the growth of the archesporial and tapetal cells. The microspore mother cells are separated from this inner wall by the tapetal cells. The latter, however, are not limited to the peripheral region, but extend into the mass of sporogenous cells and in some cases ramify entirely through the loculus, occupy- . Ing more than one-half of the sporangial cavity. The origin of the tapetum was not definitely determined, but it seems highly probable that it arises from the sterilization of sporogenous tissue rather than from the inner cells of the wall, and that all of the sterile cells within the sporangium are of the same ancestry. The tapetal cells vary in size, many of them being as large as the micro- spore mother cells. They are binucleate and are more vacuolate than are the functional spore mother cells. The two nuclei.of a single cell are usually side by side either at the center of the cell or atoneend. Each nucleus contains one nucleolus or occasionally more. CALDWELL (11) describes a condition for Lemna minor in some respects similar to that just outlined. He finds that during the early stages of the heterotypic division the cells of the tapetum sometimes divide and form groups of cells which project into the mother cell region; that the number of microspore mother cells is not reduced by the presence of the tapetal cells, although only a comparatively few developed spores, the others disorganizing and aiding the tapetum in nourishing the functional mother cells. In I. Sultani only about half the number of microspore mother cells arise that one would expect from the size of the sporangium. As indicated, not all of the tapetal cells arise at the periphery of the sporogenous mass, but many of them originate side by side with the microspore mother cells. The small number of microspore 302 BOTANICAL GAZETTE [OCTOBER mother cells may doubtless be related to the probable hybrid nature of the form of Impatiens Sultani used. The distinction between the functional and non-functional sporogenous tissue can clearly be seen at an early stage, and still shows distinctly when the spore mother cells are in synapsis. The latter are still angular and are only just beginning to separate (fig. 20). As is customary at the time of synapsis, the chromatic substance is massed against the nuclear membrane, with the single large vacuolate nucleolus projecting out from one side. As the prophase of the heterotypic division advances and a delicate spirem fills up the nuclear cavity, the cells become almost entirely free and round off, while the tapetal cells still remain somewhat angular in outline and often contain more than the 2 nuclei of the earlier stage (fig. 21). Their nuclei have become granular, lack a nucleolus, and stain more densely than before. In general, the entire mass of cells has separated from the wall. The ovule at this age shows no sign of the inner integument. As the anther increases in size, the microspore mother cells become entirely rounded off and the spirem takes on a double appearance, whether due to a splitting of a single spirem or an approximation of two was not apparent. At this time the spirem is thicker and less delicate than the spirem immediately following synapsis (figs. 21, 22). The cytoplasm has an obscurely radiate appearance, being densely granular about the nuclear membrane and more vacuolate toward the cell membrane. The spirem thickens, becomes irregular, and segments transversely into bivalent chromosomes. The majority of the segments come to lie against the nuclear membrane and show clearly their double nature in the forms of X’s, Y’s, and V’s. They are rough in outline and are connected here and there by delicate threads (fig. 23), similar to those figured by Mortier (33) for Acer Negundo and Staphylea trifolia. The nucleolus still persists and at this stage there may be two, a large and a small one. The granular area surrounding the nucleus has become still more marked and closely resembles the kinoplasmic region described by ALLEN (1) for the pollen mother cells of Larix and by numerous other writers. The peripheral cytoplasm with its large meshes 1918] OTTLEY—IMPATIENS 303 draws away from the cell membrane at various points. At this stage cellulose begins to be deposited about the cell and when the chromosomes have reached the equatorial plate a broad cell wall of cellulose is formed (fig. 27). As the chromosomes shorten and thicken, they become smooth in outline-and numerous fibers heavier and longer than those mentioned for an earlier stage (fig. 23) appear within the nuclear cavity. These fibers seem to have no apparent relation to the fibers of the kinoplasmic region, although at this time the nuclear membrane has become indistinguishable from the cytoplasm at various places (figs. 24, 25). The fibers are tufted and may extend from a given chromosome across the nuclear cavity to other chromosomes or to the nuclear membrane. At this time no nucleoli are visible within the nuclei and there is the possibility that their substance has assisted in the formation of the ers. After the nuclear membrane has entirely disappeared many fibers appear about the chromosomes (fig. 26). They extend beyond what was the original area of the nucleus and doubtless there has been a union of intra- and extra-nuclear fibers in the formation of the multipolar spindle (fig. 26). By the time meta- phase is reached the spindle has become sharply bipolar and extends across the entire cell (fig. 27). The cytoplasm surround- ing the spindle has lost the dense granular appearance of the early prophase stages and stains less densely than does the peripheral region. While I am not willing to make an unqualified statement regarding the number of bivalent chromosomes, it seems most probable from the study of the heterotypic divisions in the mega- spore and microspore mother cells that the haploid number is 7, as I have been unable to count more than that number. No stages in microsporogenesis were obtained between metaphase of the heterotypic division and the tetrads following the homotypic ivision. When the tetrads are formed the microspores are surrounded by the very thick cellulose wall of the mother cell. Each micro- spore is a little over 3 times longer than wide and possesses a reticulated membrane. At this time its nucleus is not spherical, 304 BOTANICAL GAZETTE [OCTOBER but simulates the outline of the cell and its chromatic material is distributed unevenly throughout the nuclear cavity. Several large masses of chromatin are mingled with chromatic threads which extend out from them in various directions (figs. 28, 29). The loculi containing these tetrads also contain densely stain- ing tapetal cells. These cells still have large vacuoles, but the cytoplasm stains more densely than in earlier stages and their nuclei contain one nucleolus each. Outside the tapetum is the flattened layer of cells, and beyond this the remainder of the wall is still undifferentiated. Later, when the microspores have broken away from the old mother wall, the endothecium with its spirally thickened cell walls forms just beyond the flattened layer. The tapetal cells, in general, -are still in good condition, some of them centrally located having increased very greatly in size and number of nuclei. Compare the tapetal cell (fig. 33) which was magnified 810 times with the microspore (fig. 30) which was taken from a loculus of the same age and magnified 1620 times. The number of nuclei in these tapetal cells may reach as high as 11 or more. These unusually large cells show stages in disintegration, and it is difficult to find one which has not begun to break down. A large number of nuclei, ranging from 6-13 in the tapetal cells of Hepatica acutiloba, has been reported by CouLTrer (14). SCHAFFNER (39) in his descrip- tion of Typha latifolia states that the tapetal cells increase greatly in size while the tetrads are forming, but speaks of but 2 nuclei being formed in each cell. When the microspores escape from the tetrad the chromatin of their nuclei consists of heavy, anastomosing strands which soon give rise to several distinct masses of chromatin connected with each other by more or less delicate threads. In by far the greater number of cases, if not always, these masses correspond in number with the haploid number of chromosomes (fig. 30). No nucleolus is visible at this time. There is evidently no true spirem formed in the division of the nucleus to form the generative and tube cells, but the rather imperfect reticulum of the very young microspores gives rise directly to the chromosomes at a considerable time before the organization of the spindle. It is a very common occurrence 1918] OTTLEY—IMPATIENS 305 to find all the microspores of an anther in an apparently resting stage and showing distinctly 7 chromatic masses. The microspore divides and forms a more or less vacuolate 2-celled pollen grain (figs. 31, 32). The tube nucleus is normally spherical and lies more or less near the center of the developing pollen grain (figs. 32, 34, 35). The generative cell occupies various positions within the cytoplasm of the tube cell, and at all times its nucleus is smaller and its reticulum is much less delicate than that of the tube nucleus. The pollen grain grows and its cytoplasm becomes densely filled with food granules. At this time the genera- tive cell may either be attached to the wall of the pollen grain or lie free in the cytoplasm (figs. 34, 35). When the anther is ready to dehisce, it is impossible to dis- tinguish the cytoplasm of the generative cell, and its nucleus has changed from nearly spherical to a very slender lunate form (fig. 36). At first it was thought to be a sperm nucleus, but while these crescent-shaped nuclei are very characteristic of the developing male gametophyte, no more than one to each pollen grain was ever found. When the nucleus is in this condition it contains several chromatic masses in the center and stains a diffuse yellow at either end. In all cases where the nucleus could be seen clearly throughout its entire length 7 chromatic masses were visible (fig. 36). The mature pollen grains, in longitudinal section, present an almost rectangular appearance and possess 4 germ pores, one at each corner. No instances of the division of the generative cell while still within the anther were discovered, and it is doubtful whether this division takes place until some time after pollination. Ovule after pollination The pollen tube enters the embryo sac after the chalazal haustorium has developed. Many of the embryo sacs of a given ovary were found containing the pollen tube contents in their micropylar region. While no experiments were made to eliminate all chance of cross-pollination, the conditions under which these plants were grown in the greenhouse render it very probable that many of the flowers were self-pollinated. The pollen tube enters the embryo sac at one side of the filiform apparatus and either 306 BOTANICAL GAZETTE [OCTOBER continues down the same side of the sac near to the region of the egg nucleus or crosses over the synergid and extends down the other side (figs. 37, 38). After this has occurred it is difficult to find both of the synergid nuclei. Doubtless one of them soon becomes disorganized, due to the effect of the presence of the pollen tube. SMITH (40) describes the pollen tube of Eriocaulon as either passing through a synergid or between the two without destroying them. In J. Sultani the tube nucleus was usually visible in the embryo sac, but it was often difficult to discover the sperms, due to their small size and also to the presence of many small densely staining bodies which often suggested parts of nuclei but were possibly food particles. The sperms are coiled or spiral in outline as they approach the egg and polar nuclei. In fig. 38 the two sperms are both near the egg nucleus, one is directly over the latter and the other is at its side, still in the dense strand of cytoplasm which marks the path of the pollen tube contents, and doubtless is on its way to the two polar nuclei. No sperm cytoplasm is visible and the nucleus is a spiral body made up of dark and light areas, the former of which are doubtless masses of chromatin. The char- acter of the sperm shown in fig. 39 differs somewhat from those in the preceding figure. Here a sperm nucleus is situated one at either side of the egg nucleus; the one at the right is coiled tightly and shows no distinction between chromatic and clear areas, but stains a clear light blue. It was doubtless on its way through the cytoplasm at the side of the egg to the endosperm nucleus lying directly below the egg. No stages in the actual fusion of the egg and sperm were seen. The fertilized egg differs so slightly from the unfertilized one that it is difficult to decide in a given case whether or not fertilization has occurred. In general, however, the fertilized egg increases slightly in size and its limiting membrane is more conspicuous than it is in earlier stages (fig. 40). Figs. 38 and 40 have the same mag- nification and the increase in size is evident. It seems probable that the polar nuclei unite at an early stage in the fusion of the sexual nuclei. The nuclear membranes of the two polars break down where they come in contact and one of the nucleoli passes over into the other nucleus (figs. 41, 42)- Both 1918] OTTLEY—IMPATIENS 307 nucleoli possess one or more vacuoles. What appears to be a later stage in the fusion of the two polars is given in fig. 43. A dense granular mass, the entering nucleolus, seems to be in vital contact with the nucleolus of the receiving nucleus and gives the impres- sion of giving of its substance to help in the formation of the nucleolus of the resulting endosperm nucleus. The characteris- tically large vacuole of the primary endosperm nucleolus has already appeared. Not all of the entering nucleolus fuses with the receiving nucleolus, however, for coarse strands radiate out from the dense mass of nucleolar substance and appear to be adding to the reticulum of the nucleoplasm. A similar radiating mass has been observed in one of two polars. In this case it might represent a stage in the fusion of the second sperm with one of the polars before the two polar nuclei had united. Similar masses have also been seen in the megaspore mother cell, and here also the nucleolus doubtless contributes to the chromatic substance. The fusion of the two polars has been figured by numer- ous writers for many different plants. In Nicotiana Tabacum (GUIGNARD 23) the two nucleoli remain distinct in the fusion nu- cleus for some time before fusing. VANDENDRIES (44) figures for Cardamine pratensis two nucleoli within the primary endosperm nucleus with a sperm against one side of it. He says that when the pollen tube enters the cavity of the embryo sac the two polar nuclei have begun to fuse but the nucleoli are still distinct. In J. Sultani, however, the fusion of the two polar nuclei begins relatively much later and takes place very quickly. In describing the fusion of the two polars of Arisaema triphyllum, Gow (18) says that the fusion endosperm nucleus frequently contains two nucleoli. It seems highly probable that after the primary endosperm nucleus is formed the second sperm unites with it. By this time the sperm nucleus appears to be larger than it was in earlier stages. In fig. 44 it is just at the point of piercing the nuclear membrane, and in fig. 45 it is within the primary endosperm nucleus and lying either above its nucleolus or within it. It was impossible to deter- mine if the second sperm, in all cases, waited until the primary endosperm nucleus was formed before becoming functional. GUIGNARD (23) is convinced that in the Malvaceae the time of 308 BOTANICAL GAZETTE [OCTOBER formation of the secondary nucleus is constant for a given genus. In Lavatera and others it is formed before fertilization, while it is formed after in Hzbiscus. Following fertilization a micropylar haustorium is formed. The primary endosperm nucleus divides before the division of the fertilized egg. After a few divisions have taken place several of the resulting endosperm nuclei pass up through the micropylar part of the embryo sac (figs. 46, 47) out into the micropyle and form a haustorium which emerges from the micropyle and crosses over the space between the latter and the funiculus (figs. 49a, 6). As the haustorium encounters the funiculus it either extends along the funiculus some little distance before entering it or penetrates it immediately and branches freely within its tissue (fig. 53). In the early stages, as the endosperm nuclei pass through the upper part of the sac, they are surrounded by cytoplasm rendered dense by the presence of a large amount of food substance and consequently the nuclear membrane is entirely obscured (figs. 46, 47). The densely staining filiform apparatus, all that remains of the two synergids, is still present, but it is more widely separated from the enlarged part of the embryo sac than in earlier stages (fig. 47). The narrow micropylar part of the sac where the syner- gids of the mature embryo sac were situated has widened and encroached upon the small disorganizing cells of the inner integu- ment. These cells have now disappeared except for a few remains and large regular cells limit the micropylar cavity into which the sac has pushed (cf. figs. 9, 14, 47).. At this stage the fertilized egg is still undivided and several endosperm nuclei lie below as well as above it. These nuclei are not shown in fig. 47, as they occurred in a different section from the one sketched. In the later stages in the development of the micropylar haus- torium but few nuclei are present in it. These are very large and contain a large nucleolus and stain bright red with safranin (fig. 53)- This haustorium was seen to persist up through the oldest stages studied, namely, those containing embryos with radicle and coty- ledons differentiated. The micropylar haustorium of J. Sultani differs from that of I. amphorata, as described by Lonco (28), in not entering the 1918] OTTLEY—IMPATIENS 309 outer integument. It simply pierces the funiculus and branches extensively within it. As in I. amphorata (LONGO 28), I. Sultani possesses a chalazal haustorium as well as a micropylar one. This haustorium is much less extensive than is the micropylar one. One of the endosperm nuclei at the antipodal region becomes very large, and with its surrounding cytoplasm forms a long cell which pierces through the sac and one end of it enters the chalazal tissue, while the other end remains in contact with numerous normal endosperm cells (figs. 50, 51, 53). It was difficult to secure a vertical section through this haus- torium, as a series of sections which cut through the embryo vertically would section the haustorium somewhat diagonally. Fig. 50 shows the outline of the cell, but no nucleus, while fig. 51 shows the embryo sac portion with a large nucleus, but the chalazal part was cut so that all connection between chalazal tissue and haustorium was lost. This large haustorial cell contains a densely granular cytoplasm and is doubtless active in conveying nutriment from the chalaza to the endosperm. This haustorium does not remain active as long as the micropylar one. By the time the cotyledons of the embryo have become differentiated it is not so prominent as in earlier stages, while the micropylar haustorium still seems very active (fig. 53). The endosperm develops more rapidly at the micropylar and chalazal regions than at the sides of the sac. A layer of cytoplasm with a single row of free nuclei persists at the sides for some time, while at the poles of the sac walls come in early to separate the nuclei (figs. 49a, 50, 51). With the presence of two rows of nuclei at the sides, walls form about the outer layer of nuclei, while those of the inner layer remain free (fig. 48) until some time later, when they are separated by walls,and the endosperm is composed entirely of cells (fig. 53). The size and shape of the embryo sac have under- gone marked changes during endosperm formation. At the time of fertilization the outline of the sac is similar to that shown in fig. 54, with the early antipodal haustorium opposite the micropylar part of the sac. Later, during early endosperm formation, the sac elongates and widens slightly at the micropylar region. Below 310 BOTANICAL GAZETTE [OCTOBER this it enlarges considerably on the side away from the raphe, and the sac becomes asymmetrical (fig. 55). This lateral growth continues until it extends even beyond the original position of the antipodal haustorium, with the result that the chalazal haustorium of endosperm origin comes to lie at the side of the sac rather than at its base, and the embryo sac extends below the chalaza (fig. 53). The embryo develops less quickly than does the endosperm. As stated earlier, the fertilized egg does not divide until after many free endosperm nuclei have been formed. The proembryo becomes differentiated into suspensor and embryo early in its development. By the time the embryo consists of several cells the suspensor, in a longitudinal section, shows but two cells. The cell adjoining the embryo is broader than long and the terminal one is but slightly longer than wide (fig. 49a). The suspensor, doubtless on account of the very effective micropylar haustorium, appears to be but a short-lived organ. It neither elongates nor becomes bladder-like, as is the case for many of the angiosperms, but soon breaks down and disappears. It shows signs of disintegration before the coty- ledons of the embryo appear (fig. 52), and by the time they are differentiated the suspensor has disappeared. The embryo de- velops at the tip of the suspensor and by the time the endosperm consists of two layers of cells about the periphery of the embryo sac the radicle and two cotyledons have become differentiated (figs. 53, 56). The embryo and the chalazal haustorium do not occur in the same vertical plane, therefore it is impossible to secure satisfactory sec- tions of the two from the same embryo sac. In the oldest stage studied a band of endosperm cells lines the sac (figs. 53, 56). The endosperm lies in close contact with the axis of the embryo and the sides of the cotyledons, but it is separated from the chalazal end of the embryo by a large cavity. Unfortunately, through lack of study of the seeds of J. Sultani, I was unable to determine the fate of the endosperm. According to GUIGNARD (22), a thin layer of endosperm remains undigested in the mature seed. BRUNOTTE (10), from his study of the Balsaminaceae, believes that the descrip- tions of the systematists for the mature seed should be changed. 1918] OTTLEY—IMPATIENS 311 Instead of describing the seed as having no endosperm, he would say that there is a small amount of endosperm present. Summary 1. The ovule possesses but one hypodermal archesporial cell. 2. The archesporial cell is also the megaspore mother cell. 3. An axial row of two cells is formed. The chalazal cell is the mother of the embryo sac. 4. A normal 8-nucleate sac is formed. 5. There isa variation in the position af the synergid nuclei, due to their age. 6. The two polar nuclei come in contact directly beneath the egg and do not fuse until after the pollen tube has entered the embryo sac. 7. The 3 antipodals may be either 3 cells or a group of 3 nuclei cut off from the upper region of the sac by a membrane. 8. The antipodals disappear soon after the egg apparatus is formed. 9g. The megaspore mother cell and the early 2-nucleate embryo sac are bounded at the apex and sides by the nucellar epidermis. 10. The embryo sac absorbs the nucellar epidermis and by means of an antipodal haustorium absorbs all of the nucellus between the sac and the chalaza. 11. The tapetum is derived from the inner layer of the inner integument. 12. The outer integument arises from the inner integument. 13. Binucleate tapetal cells surround the microspore mother cells. They also extend into the mass of sporogenous cells and separate the functional mother cells into groups. 14. The nucleus of the generative cell of the pollen grain appar- ently does not divide before pollination. 15. The pollen tube enters the embryo sac along the side of the filiform apparatus and extends down one side ‘of the embryo sac until it is near the egg nucleus. 16. Two coiled sperm nuclei are often seen near the egg nucleus. 17. It seems very probable that triple fusion occurs. ai2 BOTANICAL GAZETTE [OCTOBER 18. An extensive micropylar haustorium and a more simple chalazal one develop from the endosperm. 19. The embryo possesses a short suspensor. 20. The bivalent chromosomes in both megasporogenesis and microsporogenesis show X’s, Y’s, and V’s. 21. A multipolar spindle appears in the prophase of the hetero- typic division in microsporogenesis. 22. The nucleus of the microspore has but a short period of rest; the prophase of division is initiated early and persists for some time. In conclusion, I wish to thank Professor MARGARET C. FERGU- son for her careful reading of this paper and for her very helpful suggestions. WELLESLEY COLLEGE WELLESLEY, Mass. LITERATURE CITED 1. ALLEN, C. E., The early stages of spindle formation . . pollen mother cells of Larix. Ann. Botany 17:281-311. pls. 14, 15. 2. , Nuclear divisions in the pollen mother — of Lilium cana- dense. ‘Kea Botany 19:189-259. pls. 6-9. 1905. . Atkinson, G. F., Studies on reduction in plants. I. Reducing division in Arisaema iriphyliion by ring and tetrad-formation. II. Reducing divi- sion of the chromosomes in Trillium grandiflorum. Bot. Gaz. 28:1-26. pls. 1-6. 1899 4- pia L. i. ‘Standard Cyclopedia of Horticulture. New York. Vol. 3. w& gts. ‘ pose ies G. P., Contribution a l’étude du sac embryonnaire ‘chez certaines Gamopetales. Flora 86:47-71. pls. 3-10. 1899 6. Barnes, C. R., The process of fertilization in Campanula americana. Bor. Gaz. 10:349-354. pl. 1. 1885. 7. Brtuines, F. H., Beitrige zur Kenntniss der Samenentwickelung. Flora 88: 253-318. Igor. 8. Buss, M. C., A contribution to the life history of Viola. Ann. Botany 26:155-163. pls. 17-19. 1912. 9. Branpza, M., Developpement des téguments de la graine. Rev. Gén. Botanique 1:1-32, 71-84, 103-126, 150-165, 227-240. 1891. 10. Brunorre, C., Recherches embryogéniques et, anatomiques sur quelques espéces des genres Impatiens et Tropaeolum. Nancy. 1900. 11. CALDWELL, O. W., On the life history of Lemna minor. Bort. GAZ. 27:37- 66. figs. 59. 1899. 1918] OTTLEY—IMPATIENS 313 pot oO w coal - CHAMBERLAIN, C. J., The embryo sac of Aster Novae-Angliae. Bor. Gaz. 20:205-212. pls. 15, 16. 1895. . Cuopat, R., Principes des botanique. 1907 - CouLTER, J. M., Contribution to the life history of Ranunculus. Bor. GAZ. 25:73-88. pls. 4-7. 1898 . CouLTER, J. M., and CHAMBERLAIN, C. J., Morphology of angiosperms. New York. 1903 . ENGLER, A., and PRANTL, K., Die Natiirlichen Pflanzenfamilien. Leipzig. . GotprLus, M., Sur la structure et les fonctions de l’assise épithéliale et des antipodes chez les Composées. Jour. Botanique 12:374-384. pis. I-6. 1898; 13:87-06. 1 » Gow, J.-E. Eackeoenesy of Arisaema triphyllum. Bot. GAz. 45:38-44. figs. 24. 190 08. . GuicnarD, L., Recherches d’embryogénie végétale comparée. I. Légu- mineuses. Ann. Sci. Nat. Bot. VI. 12:5-166. pls. 1-8. 1881. , Recherches sur le sac embryonnaire des Phanérogames Angio- Sarines, Ann. Sci. Nat. Bot. VI. 13:136—199. pls. 3-7. 1882. I , Note sur l’origine et la structure due tégument séminal chez les Capparidées, Résédacées, Hypéricacées, Balsaminées, et Linacees. Bull. Soc. Bot. France 40:56-60. 1893 , Recherches sur le développement de la graine et en particulier du tégument séminal. Jour. Botanique. 1893. , La double fécondation chez les Solanées. Jour. Botanique 16: 145-167. Sigs. 45. 1902. , La double fécondation chez les Malvacées. Jour. Botanique ‘isageaak. 1904. Hammonp, H. si The embryology of Oxalis corniculata. Ohio Nat 8: 261-264. HEATLEY, M. M. A study of the life history of Trillium cernuum L. Bot. GAZ. 61:425-429. 1916 HEGELMAIER, F., Untersuchungen iiber die Morphologi og Dikotyledonen- Endosperms. Nova Acta Leopoldina 49: 1-104. pls. 5 885. - Lonco, B., Nuove ricerche re nutrizione dell’ edeious sk Reale Accad. Lived 16: 591-504. MICHELL, M. R., The cnt sac —- embryo of Striga lutea. Bor. Gaz. 59:124-135. pls. 8 , The embryo sig ze ‘Richondia africana Kth. Bor. GAz. 61:325- 336. pls. 21-23. 1916. Mortter, D. M., Contributions to the embryology of the Ranunculaceae. Bor. “Pe 20:241-248, 296-304. pls. 17-20. 1805. ——, The development of the heterotypic chromosomes in pollen mother cells. Ann. Botany 21:309-348. pls. 27, 28. 1907. 314 BOTANICAL GAZETTE [OCTOBER 33. Mortier, D. M., Mitosis in the pollen mother cells of Acer Negundo L. and Staphylea wifelia L: Ann. Botany 28:115-133. pls. 9, 10. 1914. 34. OPPERMAN, M., A contribution to the life history of Aster. Bot. GAz. 37:346-352. pls. 14, 15. 1904 35- OverTON, J. B., On the organization of the nuclei in the pollen mother cells of certain plants, with especial reference to the permanence of the chromosomes. Ann. Botany 23:20-61. pis. I-3. 1909 36. Pace, L., Parnassia and some allied genera. Bor Gaz. 54:306-328. IQI2. $9: RAITT, A: B., shea ge: of ovule of Impatiens pallida Nutt. Plant World 19: 195-203. 19 38. Rosrnson, B. L., and Feast, M. L. Aesohan New Manual of Botany. 7thed. New York. 190 39. SCHAFFNER, J. H., The ici of the stamens and carpels of Typha latifolia. Bor. a 24:93-102. pls. 4-6. 1897 40. SMITH, , The floral development and embryogeny of Eriocaulon Chnnue. Bis. GAZ. 49: 281-289. pls. 19, 20. 1910 41. Sources, M. R., Développement et structure du eos séminal chez les Solanacées. i, Sci. Nat. Bot. IX. 6:1-124. 190 42. STRASBURGER, E., Zeitpunkt der Bestimmung des oes Apogamie, Parthenogenesis, und Reduktionsteilung. Hist. Beitr. 7: 1900. 43- TREUB, M., Observations sur les Loranthacées. Ann. Sci. Nat. Bot. 13:250-282. pls. 13-20. 1882; reprinted in Ann. Jard. Bot. Buiten- zorg 3:1-12. pls. I, 2. 1883; 2:54-76. pls. 54-76. 1885. 44. VANDENDRIES, R., Contribution a l’histoire du développement des Cruci- féres. La Cellule 25:415-459. pl. 1. 1909. . Warminc, E., De l’ovule. Ann. ‘Sci, Nat. Bot. VI. 5:177-266. pls. 7-13. 1877. EXPLANATION OF PLATES XIV, XV All figures were drawn with the aid of an Abbé camera lucida and are reduced one-half in reproduction. The number accompanying the description of each figure indicates the magnification before the reduction. The lettering — of the figures is as follows: ii, inner integument; 07, outer integument; Cc, megaspore mother cell; m», nucleolus; f, sepia culus; ma, micropylar cell of axial row; ¢, epidermis of nucellus; ¢, tapetum; s, stringy cells of the nucellus; en, egg nucleus; sm, synergid nucleus; cz, Sanerel vacuole; p, polar nucleus; mm, microspore mother cell; dw, disintegrating air ; tn, tube nucleus; 87, generative nucleus; s', sperm nucleus; S?, sperm n cleus; fa, filiform apparatus, pic, pollen tube contents; rp, receiv sg m, micropyle; ah, antipodal ; ch, chalazal haustorium; SP, > uo rima: ospe u cutinized walls; mrp, nucleolus of receiving polar; edn, endosperm nuclei. 1918] OTTLEY—IMPATIENS 315 PLATE XIV Fig. 1.—Vertical section of ovule before integuments have appeared; X828. Fic. 2.—Vertical section of sh showing origin of inner integument and large megaspore mother cell; 8x Fic. 3.—Vertical section of vale slightly older than fig. 2; 810. Fic. 4.—Vertical section of ovule with nucleus of megaspore mother cell in synapsis; X8ro. Fic. 5.—Vertical section of megaspore mother cell with nucleus in second contraction stage of prophase of heterotypic division; X 1620. Fic. 6.—Vertical section of ovule showing origin of outer integument and devescnted spirem of megaspore mother cell; X 500. Fic. 7.—Vertical section of megaspore mother cell showing tetrad forma- tion in nucleus; X 1300. Fic. 8.—Vertical section of nucellus after 2-celled axial row has been formed; 1300. Fic. 9.—Vertical section of portion of ovule containing 2-nucleate embryo sac; 1000 Fic. to. mae ertical section of 2-nucleate embryo sac with surrounding nucellus and tapetum; nuclei of embryo sac in prophase of division; 810. Fic. 11.—Vertical ain of young 4-nucleate embryo sac; combination of 3 serial sections; Fic. 12.—Vertical pe of slightly older sac with nucellus and re integrating epidermis; 810 Fic. 13.—Vertical ection of still older 4-nucleate embryo sac with basal ion of epidermis still intact; 1000. G. 14.—Vertical section of early 8-nucleate embryo sac with nucellus at base tis instep at either side; G. 15.—Vertical section of early sage in formation of egg apparatus: synergid siche each above a large vacuole; X 1000. Fic. 16.—Vertical section of micropylar region of embryo sac showing 2 synergids; synergid nuclei have reached center of the 2 synergid cells; X 1000. Fic. 17.—Vertical section of 8-nucleate embryo sac before the 2 polar nected have reached their position directly beneath egg; combination of 2 serial sections; X 1000 Fic. 1 #i--Veitical section of antipodal region of 8-nucleate sac before antipodal polar has moved toward micropylar polar; the 3 antipodals show as rial cells; combination of 2 serial sections; 1620. . 19.—Vertical section of embryo sac after antipodals have disappeared and aes haustorium has begun to develop; micropylar tip of sac was covered by inner integument; X 1000. 1G. 20.—Cross-section of microsporangium while spore mother cells are in synaptic stage; 500. 316 BOTANICAL GAZETTE [OCTOBER Fic. 21.—Cross-section of contents of microsporangium soon after chro- matic substance of microspore mother cells has formed a spirem following synapsis; X 500 Fic. 22.—Section of microspore mother cell slightly older than those in preceding aun nucleus contains spirem showing its double nature; 1620. —Section of microspore mother cell soon after spirem has seg- caanteds 2 clea present, small one directly over large one was omitted from sketch; 20. IGS. 25.—Sections of 2 microspore shothice cells after bivalent chromo- somes ise icine! and become smooth in outline and intranuclear fibers have appeared; in fig. 25 a thick wall is beginning to form about the mother cell; 1620 IG. 26. Section of microspore mother cell with multipolar polyarch spindle; heterotypic division; 1750. 1G. 27.—Section of microspore mother cell showing bipolar spindle of heterotypic division; bivalent chromosomes arranged at equatorial plate; 1620. Fic. 28.—Cross-section of tetrad; 1620. Fic. 29.—Vertical section of tetrad with but 2 of the microspores visible; X 1620. Fic. 30.—Vertical section of young microspore; nucleus has passed through a very short resting stage and has entered upon a prolonged prophase; X 1620. Fic. 31.—Cross-section of young pollen grain; large central vacuole present and tube and generative nuclei at one side of pollen grain; 1620. Fic. 32.—Vertical section of slightly older pollen grain than fig. 31; genera- tive cell lies near center of pollen grain; 1620 PLATE XV Fic. 33.—Section of one of large multinucleate tapetal cells which occur within microsporangium after microspores have separated from tetrads; nuclei in various stages of disintegration; 810. IGS. 34, 35-—Vertical sections of nearly mature pollen grains; note change in size and cytoplasmic contents from fig. 32; cytoplasm packed full with food which is doubtless starch grains; in fig. 34 generative cell lies against wall of pollen grain, while in fig. 35 it lies free within cytoplasm of tube cell; 1620. Fic. 36.—Cross-section of mature pollen grain; cytoplasm of generative cell cannot be distinguished from that of tube cell; generative nucleus forms 4 crescent and contains 7 distinct chromatic masses; 1375. Fic. 37.—Vertical section of ovule after pollen tube has entered embryo sac; sketch is largely in outline and was derived from 3 serial sections; 509- Fic. 38.—Vertical section of micropylar region of embryo sac after pollen tube has entered sac; X1000. _ Fic. 39.—Two sperms near egg nucleus; 1620. BOTANICAL GAZETTE, LXVI _ PLATE XIV a & a L\ dram Seo ey OTTLEY on IMPATIENS PLATE XV BOTANICAL GAZETTE, LXVI me ) 8 ~ ¥ OTTLEY on IMPATIENS 1918] OTTLEY—IMPATIENS 317 Fic. 40.—Fertilized egg; 1000. ne 41-43.—Stages in fusion of the 2 polar nuclei; 1620. Fics. 44, 45.—Stages in fusion of second sperm nucleus with endosperm aa fig. 44, X1620; fig. 45, X 1625. 1G. 46.—Vertical section of apex of embryo sac showing 3 endosperm nuclei on their way to form micropylar haustorium; another section shows endosperm nucleus below original position of primary endosperm nucleus; 500. Fic. 47.—Vertical section- of micropylar portion of embryo sac after primary endosperm nucleus has divided several times; sac has enlarged beneath filiform si gaiee X 1000. Fic. 48.—Portion of endosperm present at one side of embryo sac after embryo he formed; X 500. IG. 49a, ioe vertical section through micropyle with micropylar haus- torium, ca through embryo sac at base of micropyle; young embryo with short suSpensor surrounded by endosperm cells; 500; b, vertical section of micropylar haustorium as it extends across space from micropyle to funiculus; a and 6 are sketches of same haustorium, but taken from different sections; X 500 FIG. 50. Ssieegi through chalazal portion of ovule; chalazal haustorium and endosperm cells are filled in while only outline of chalazal tissue is given; sketch is from same ovule as fig. 49a, and is a combination of 2 sections; X 500. : Fic. 51.—Vertical section of chalazal haustorium with surrounding endo- sperm cells; haustorium cut somewhat diagonally; X 500. Fic. 52.—Vertical section through embryo sac after a many-celled embryo has developed; endosperm at base and at one side consists of but one layer of cells, at the other side consists of several layers, and in another section of this same embryo sac chalazal haustorium occurs in this thickened portion of endosperm; 16 Fic. 53 aa. sketch of vertical section of ovule containing embryo of 2 cbt yledions: embryo does not show in section; combination of 2 sections; Fic. 54.—Outline sketch of vertical section of embryo sac at time of fertili- zation or slightly later; growth has taken place at antipodal and at median regions; 162 Fic. 5§.—Outline sketch of vertical section of embryo sac after embryo has been developed; narrow micropylar portion of sac has lengthened and increased somewhat in width; side opposite raphe has pushed out into integu- ment, giving an asymmetrical sac; 162. 1G. 56.—Outline sketch of vertical section of embryo and accompanying endosperm tissue; embryo is differentiated into radicle, hypocotyl, and 2 cotyledons; X 162 NOTES ON AMERICAN WILLOWS II. THE SPECIES RELATED TO SALIX GLAUCA L. CAMILLO SCHNEIDER In my first paper’ I dealt with Salix arctica Pall. and its relatives. These species are mostly united with S. glauca L. and its congeners in one group or section by such American salicologists as P. A. RypBerc and C. R. Baxi. European students of willows like N. J. Anpersson, A. and E.-G. Camus, and O. v. SEEMEN referred the two species to different sections, and I have always thought it best to regard each species as a representative of a distinct group. It is not an easy task to draw a line between the forms of the GLAucaE on the one hand and those of the arctica group on the other, but this is true of most of the sections in a genus like Salix, where it is difficult to define groups of closely related species. As I have already explained in SarcEnt, Pl. Wils. 3:136. 1916, the name ARCTICAE is not available to designate the group of which S. arctica Pall. is the type, because it was first used by ANDERSSON (1858) for a section containing S. Hookeriana Barr., S. speciosa Hook. and Arn., non Host. (S. alaxensis Cov.), etc., which in 1868 ANDERSSON included in his sect. NIvEAE B. VrttosaE; therefore I (I.c. 140) proposed the name DretopictyaeE for this group, but at this time I also kept the OvatiroLtar of RYDBERG as a separate unit, expressing, however, a doubt “whether the species united by RypBERG in this section really belong in the same group.” At present I believe that S. ovalifolia should be placed in the same section with S. arctica, and consequently the name OVALIFOLIAE must be adopted for this group. To distinguish between those two sections the color and pubescence of the bracts (or scales) seems to afford a rather reliable character. In the OvALIFOLIAE the bracts are usually more or less bicolor, being pale at base and dark brown, fuscous, or even blackish toward the apex, while the forms of the GLAUCAE mostly have uniformly yellowish, light brown, or straw- colored bracts, which sometimes (especially in the upper part of * Bot. Gaz. 117-142. 1918. Botanical Gazette, vol. 66] [338 1918] SCHNEIDER—AMERICAN WILLOWS 319 the ament) are reddish or purplish toward the apex, but never become really fuscous or blackish. Furthermore, the pubescence of the 2 kinds of bracts is of a different character. In the first group it usually consists of rather long, straight, silky hairs, of which at least the uppermost are about the same length as the bract, which mostly does not bear many short hairs on its surface, and often becomes nearly glabrous. In the second group the hairs are comparatively shorter, less straight, and rarely distinctly silky, but are softer and sometimes a little curly. As a whole the bracts are more or less covered with pubescence, and are rarely distinctly ciliated at apex with long silky hairs. These characters are usually more easily detected in the female than in the male specimens, which are often more similar in the two groups. It takes some time for the student to become familiar with these peculiarities, which are by no means clearly recognizable in every specimen. There are of course exceptions also, but in such cases we find other characters to determine the real affinity of a certain form. Many so-called intermediate forms are of hybrid origin, or should receive closer observation in the field before defining their taxonomic position. This is what I have to say at present regard- ing the separation of the sections GLAUCAE and OVALIFOLIAE. Later I hope to have the opportunity to discuss in detail the sys- tematic arrangement of the American species of Salix. As I now understand them, the following species belong to the section GLAUCAE: S. anamesa Schn., S. brachycarpa Nutt., S. chlo- rolepis Fern., S. cordifolia Pursh, S. desertorum Rich., S. fuller- tonensis Schn., S. glauca L., S. lingulata And., S. niphoclada Rydbg., and S. pseudolapponum v. Seem. I do not include in this group S. chlorophylla And., S. McCalliana Row., S. Nelsonii Ball, S. sas- katchavana v. Seem.,and S.idah is (Ball) Rydbg., which RYDBERG places in his section ARCTICAE (FI. Rocky Mts. 190. 1917), which seems to me an unnatural mixture of species of different affinities. 1. S. etauca L., Sp. Pl. 2:1019. 1753.—Before we can decide whether any American forms or which of them are to be referred to this species it seems necessary to discuss the characters of the typical S. glauca L. It is founded on “363. Salix foliis integris subtus tenuissime villosis ovatis. Tab. VIII. fig. p. and Tab. VIL. 320 BOTANICAL GAZETTE [OCTOBER fig. 5°’ in Linnaeus’ FI. Lapp. 290. 1737. The description and figure given by the author in this place give a rather good idea of histype. Furthermore, ENANDER (Stud. Salic. Linnés Herb. pp. 51, 54, 59. 1917) describes the female and male specimens of S. glauca genuina, typica or vera in Linnaeus’ herbarium. F ollowing LInNAEUs and ENANDER, I find the following characters.of what has to be called the typical S. glauca: Frutex bi- vel tripedalis. Rami rubescentes, glabri; ramuli novelli villosi. Folia ovata, lanceolata, ovali-lanceolata vel ovato-oblonga, utrinque fere aequaliter attenuata vel inferiora apice obtusa, superiora magis oblonga acutiora, integerrima, 15:8 vel 40:12 ad 60:15 vel 55: 20 mm. magna, utrinque (subtus tamen densius) villosi vel superne pilis parcius obsita vel fere glabra, non vero nitida, subtus pallidiora, “villis oblongis raris hirsuta’”’ vel ‘“pilis albicantibus vestita”’; petioli 8-10 mm. longi, villosi; amenta pedunculis ad 3 cm. longis foliolis circ. 4 ceteris similibus instructis suffulta; flores masculi bracteis pallidis pilosis, filamentis basi pilis crispatis ornatis antheris testacei coloris instructis; feminei bracteis similibus, ovariis capsulisve tomentosis sessilibus, stigmatibus stylisque quasi semipalmato-alcicornibus. Not having sufficient herbarium material from Lapland at my disposal, I think it best to add the following characters given by ANDERSSON in his Salic. Lap. 73, fig. 22. 1845: Amenta serotina ramulos breves crassiusculos tomentosos foltis ceteris Vix minoribus 3-6 vestitos terminantia, iisque plerumque longiora, cylindrica, obtusa, erecta, demum sublaxa; mascula 1-2 uncialia, densa, squamis oblongis, obtusis, fulvis apice roseis, itis albis longis rectis villosissimis, stam. 2, fila- feminea subdensiflora, abrupta, obtusa, 1-3 uncialia, primum rigida, demum laxa, squamis fulvis apice roseis, oblongis, obtusis, ventrem capsulae superan- tibus, albo-villosis; capsulae pedicello nectario lato quadrangulari pl.m. pro- unde partito, dimidio breviori, brevissime pedicellatae, ovales vel conicae, obtusae, lana alba densissime tomentosae, stylis aut omnino geminis aut fere usque ad basin bipartitis (ut eorum pars, quando adest, per sub lana pars, q capsulae lateat), stigmatibusque linearibus divaricato-bipartitis, rufo-fulvis terminatae. Judging by these characters we have to decide, I believe, whether there are in America forms identical or closely related to the typical 1918] SCHNEIDER—AMERICAN WILLOWS 321 S. glauca from Northern Europe. According to the leading Euro- pean salicologists S. glauca is rather variable, but I fail to find a good arrangement of the different variations already known from Europe and Northern Asia in the existing literature. It is impos- sible to judge the American forms correctly without having a clear understanding of the Asiatic forms already described, because it is to be expected that the forms from Eastern Asia will have the closest affinity with the American forms. No mention is made of S. glauca by Pursu (1814) or MicHaux (1820), or even by HooKER (1839). It was ANDERSSON who in 1858 first mentioned S. glauca as occurring “in provinciis septen- trionalibus et arcticis Americae borealis.”’ He further said: Haec species . . . . in arcticis regionibus Americae habitu externo vix nostrae similis exstat. Specimina tamen a Seemann in parte occidentali et a Lyall in Disco Island lecta, nec non e “Rocky Mountains” reportata cum nostris tamen ita congruunt, ut de identitate non dubitare liceat. Folia nunc utrinque molliter villosa et incana, nunc denudata subviridia, amenta semper foliato-pedunculata, capsulae brevius pedicellatae. Huic certissime ut forma tantum associanda—villosa. S. villosa (D. Don) Hook. l.c. p. 144, no. 3. This S. villosa Barratt apud HooKER has to be ascertained before anything else can be done to determine which forms may be refer- able to S. glauca. Hooker (FI. Bor.-Am. 2:145. 1839) said: “that ‘Dr. Barratt considers it to be the same as S. villosa of D. Don, in Pursh, Herb. Canad.’”’ I have never seen specimens from a “Herb. Canad.” of Pursu, nor do I know whether Pursu ever dis- tributed such a collection.2, Neither he nor D. Don published a S. villosa; there are, however, two species bearing this name, one of SCHLEICHER (Cat. Pl. Helv. ed. 3.26. 1815) which is a nomen nudum and was probably first mentioned in the first edition of the Catalogue in 1809; the other of Forses (Salict. Wob. 183. pl. 92. 1829) representing a sterile specimen of unknown origin. Thus the 2There is, however, a specimen, consisting of 3 leaves only, in herb. N. labeled “Salix Micha D. Don, in Pursh’s Canadian Herb. (collected in Lord Selkirk’s Exped. from Mr. Lambert’s Herb.).” Down did not publish a species S. leucodendron, and I am not yet sure to which species these 3 leaves belong. PuRsx’s herba- in possession of LAMBERT (see Gard. Chron. 1842, p. 439), but Purs# him- self did not collect in Canada at all (see HarsHperceEr, Bot. Philad. 115. 1899). I have not been able to get any information on ‘‘Lord Selkirk’s Exped.” 322 BOTANICAL GAZETTE [OCTOBER name S. villosa Barr. cannot be used even if the form described by Hooker under this name could be recognized as a good species. I have been fortunate in seeing photographs and fragments of the types of S. villosa preserved at Kew, and also the corresponding specimens of BARRATT’s collection in herb. N., and I am convinced that Hooker included different forms under his S. villosa. At first glance his diagnosis fits well the forms described by RYDBERG as S. Seemannii (see later) and the material before me from the Yukon Territory, but the character given by HooxeEr in the following phrase: “‘rami foliisque junioribus lana arachnoidea villosis’’ seems | peculiar tome. I cannot detect traces of a “lana arachnoidea”’ on the specimens before me, and furthermore, the specimen collected by Drummond (no. 7. Herb. H. and B.) which is regarded as the “type” is not characterized by “‘foliis lato-lanceolatis.”’ There is, however, a specimen in Herb. Torrey (N.) labeled “‘no. 6. Herb. H.B. and T.”’ and “an S. villosa D. Don” in which the lower sur- faces of the lanceolate leaves are covered when young with a “‘lana arachnoidea,” the prominent rib being nearly glabrous, while the lateral nerves are almost hidden by the pubescence. Later the leaves become more or less glabrous, and the first or lowermost leaves show nothing but a few scattered long silky hairs. The petioles are nearly glabrous, and the stipules are very small, hardly a fourth of the length of the petiole, very glabrescent, semiovate, and denticulate. This does not agree with Hooxer’s statement: “stipulis semicordatis petiolo sublongioribus,” which is the case in no. 7; and HoOKER’s diagnosis seems to me only explicable if we presume that he mixed two different forms. On the same sheet with no. 6 is also an old fruiting catkin with a leafy peduncle which is identical with those of no. 7. I am not yet sure to what species the sterile branch of no. 6 really belongs. HOOKER also described a var. “8. acutifolia; foliis magis acutis vel subacuminatis,” collected by RicHARDSON at Fort Franklin on the Mackenzie River, of which a photograph of the type “no. 70. Hb. H.B. and T.” is before me. It consists of 3 pieces of young female flowering branchlets. I also saw a sheet with the label “no. 58. Hb. H. B. & T.” ex herb. Torrey (N.) marked ‘‘Fort Franklin, Richardson,” which contains 2 fruiting and 1 sterile 1918] SCHNEIDER—AMERICAN WILLOWS 323 branchlet of var. acutifolia named by RypBERG S. villosa and marked no. 2. Besides these there are 2 sterile branchlets which may belong to var. glabrescens; the upper middle one was referred by RypBErG to S. villosa, while the small one at the left corner of the lower label is without number. Furthermore, there are 3 ster- ile branchlets numbered 1 and named S. chlorophylla by RYDBERG which indeed look very much like this species or may be refer- able to S. pulchra Cham. All the specimens are referred (by Barratt?) to S. planifolia Pursh, a very uncertain species which may be identical with S. chlorophylla And. As already mentioned, ANDERSSON regarded HooxeEr’s S. villosa in 1858 as only a variety of S. glauca L. He said: “Haec forma speciei maxime vegeta videtur,”’ and in his short description (in Ofv. K. Vet.-Akad. Férh. 15:109) we read: “foliis tenuioribus, supra (sic!) glaucis, sparse pilosis, elevato-venosis, stipulis subper- sistentibus lanceolato-linearibus; amentis sat longis erectis laxius- culis, subrarifloris . . . .” The same diagnosis is repeated in Sal. Bor.-Am. 22, and in Walp., Ann. Bot. 5:753. 1858. The statement “foliis supra glaucis”’ is certainly a misprint for “subtus glaucis,” or it may be that a whole sentence has been omitted. Unfortu- nately, ANDERSSON did not cite a specimen, but his description scarcely fits the Rocky Mountain material collected by Drum- MOND. Ten years later ANDERSSON (in DC. Prodr. 16°:281) pro- posed a new hybrid S. glaucops,3 which he placed without a number between 107. S. glauca and 108. S. desertorum, and of which he describes 2 ‘‘modifications,’ namely, var. villosa, being identical with his former S. glauca var. villosa, and var. glabrescens, which he based on specimens collected by BourcEAv in the Rocky Moun- tains, and with which I shall deal later. ANDERSSON referred to his S. glaucops villosa not only HOOKER’s villosa and his own S. glauca villosa, but also S. villosa Seemann (“Voy. of Herald. p. 39.54’’) and “S. cordifolia Hook. Fl. Boreal. 3It ought to be mentioned that this species has been entirely misunderstood by M. J. Jones, Willow Fam. Great Plat. 16. 1908. The author says of his study: ‘This work, on western willows, is put forth tentatively in order to clear up doubts. ....” But he certainly succeeded in greatly augmenting the existing confusion in regard to many species. There are scarcely 2 willows better ae eer than S. glaucops And. and S. subcoerulea Pip., which Jones makes synon 324 BOTANICAL GAZETTE [OCTOBER amer. p. 152. p.p. (non Pursh).”” SEEMANN’s plants came from western Eskimaux-Land (Northwestern Alaska from Norton Sound to Point Barrow), while HooKER’s specimens to which ANDERSSON alludes were collected in Labrador. The Rocky Mountain speci- mens mentioned by HooKER are not included, as they had already been described by ANDERSSON as S. subcordata (S. arctica var. sub- cordata Schn., see my first paper). ANDERSSON’s main description of S. glaucops fits best Seemann’s specimens, and such forms as S. villosa acutifolia Hook., of which ANDERSSON made no mention at all either in 1858 or in 1868. According to the rules of nomen- clature the name S. glaucops has to be applied to the S. glauca of Alaska, the Yukon, and the Mackenzie district if further investigations should prove that these forms can be regarded as a distinct species. Unfortunately, this name has been used by RypBerG (1899) and Batt (1909) to designate a more southern form of the Rockies, for which I use the name S. pseudolapponum v. Seem. (see later). When Batt first treated this form in 1899 (Trans. Acad. Sci. St. Louis 9:88) he expressly said: “Our Rocky Mt. form was included under S. glauca villosa by Mr. BEBB, but it is certainly not the S. villosa Don described by HOOKER (Fl. Bor.- Am. 2:144) and later published by ANDERSSON as S. glauca villosa (Sal. Bor.-Am. 22). That had long leaves and thick aments 2-3 inches long, being thus more closely related to the European S. glauca,” and (J.c. 89) Batt designates S. glauca var. villosa And. (S. villosa Barr., S. glaucops And.) as a form of which “full discus- sion must be deferred until more abundant material is accessible.” He adds that “HaNseEn’s no. 800, Fl. Sequoia Reg., 1892, is a plant which nearly answers the original description,’ but in my opinion HANSEN’s specimen differs widely from it, and belongs to S. cal?- fornica Bebb, a fact suggested by BALL himself. According to RYDBERG (1899), S. glauca is “apparently rare in Am«rica, and probably confined to the extreme northeast portion.” Nevertheless, he cites, besides specimens from western Greenland and Labrador, “Alaska: Nurkagak, 1881. McKay,” meaning Nushagak in the Bristol Bay. This specimen is referred by Covitte to S. glauca. Furthermore, in 1901, RypBERG described a S. Seemannii, the type of which had been ‘collected at Dawson 1918] SCHNEIDER—AMERICAN WILLOWS 325 by R. S. WrxttAMs, June 11, 1899, a more mature specimen June 12. Also collected by SEEMANN on Chamisso Island, 1851, no. 1873, and Kotzebue Sound and Norton Sound, 1849, no. 1423.” He accompanied his description with the following remarks: ‘“SEE- MANN’S specimens, cited below, were named by HooxeEr S. glauca var. macrocarpa, but the plant is neither S. macrocarpa of TRaut- VETTER nor that of NUTTALL; it is related to the former, but not to the latter. S. macrocarpa Trautv. (S. glauca macrocarpa Ledeb.) is described as having sessile stigmas and fuscous bracts; it prob- ably does not occur in America.” In the original dcscription of S. macrocarpa Ledeb. apud TRAUTVETTER (in Nouv. Mém. Soc. Nat. Mosc. 2:292. 1832) I fail to find the statement that the stigmas are sessile; this part of the diagnosis runs: ‘“‘stylo basin usque bipartito, stigmatibus bifidis.””, TRAUTVETTER compares in detail S. glauca and S. macrocarpa, and attributes to the latter the follow- ing characters: “frutex pedalis prostratus,”’ ‘‘folia majora, acumi- nata, juniora jam fere prorsus glabra,’”’ and “pedicellus interdum fere longitudine ovarii.” The same statement is given for S. glauca 8 macrocarpa Trautvetter in LepErBour, FI. Alt. 4:281. 1833. Judging by those characters the American form in question cannot be identified with this var. macrocarpa, but better agrees with what TRAUTVETTER (1832) regarded as typical S. glauca, and named S. glauca var. microcarpa Ledeb. in 1833. Here TRAut- VETTER says, after having given an ample description of the speci- mens from the Altai, “‘exemplaria altaicis simillima Cl. Eschscholtz legit ad Cap. Espenberg.”’ As already stated, it is difficult to decide at the present status of our knowledge of the Old World forms of S. glauca whether some of them are identical with the American forms. So far as I can judge by TrRauTvETTER’s descriptions and the material I have seen from Asia, I am not convinced that the forms of Northwestern America can be regarded as representing the typical S. glauca or one of TRAUTVETTER’s varieties. A keen and careful observer like COvILLE, in 1901, said: “There is a tendency among American willow students to exclude Salix glauca from the North American flora, but our Alaskan specimens show so close an agreement with some European material of this species that I am unwilling to 326 BOTANICAL GAZETTE [OCTOBER separate them.”’ He adds that he is not “able to find in the descrip- tion [of Seemannii] a record of any characters that serve to dis- tinguish the specimens assigned to the latter species from forms of glauca found in America and Europe.” I agree with CovILLE that the North American forms are very similar to those of S. glauca, but they are in my opinion not fully identical with the typical S. glauca L. s. str., the characters of which I have already indicated. In looking over the copious and well collected American specimens before me, I hesitate to designate them as typical S. glauca, nor am I willing to regard them as a separate species until a closer study of this circumpolar willow has convinced me of one fact or the other. Those specimens exhibit a great degree of variability in the shape and size of the leaves, in the amount of pubescence, in the length of the aments, and in the characters of the flowers. As a whole they seem to differ from the typical .S. glauca by the usually well developed stipules, by the longer pedicels of the fruits which nor- mally are from one-half to twice longer than the gland, and by the tendency of the filaments to become almost glabrous. Judging by Covittr’s statement with regard to S. reticulata, that “in all the other Alaskan willows the filaments are glabrous throughout,” I supposed that this fact might furnish a good character to separate specifically the American S. glauca from the European-Asiatic species, but a close investigation of all the male specimens at my . disposal convinced me that the filaments are always more or less hairy at their base. Specimens like nos. 3369 or 3373 of TRELEASE ~ and SAUNDERS, which apparently have entirely glabrous filaments, do not seem to be pure S. glauca, and the American form probably hybridizes with other willows as freely as does the European one. If we regard the American S. glauca as a distinct variety, we have unfortunately to use the varietal name acutifolia given by Hooker to his variety of S. villosa, because it antedates ANDERS- son’s S. glauca villosa by almost 30 years, and apparently represents a rather extreme form with narrowly lanceolate leaves. I regar my determinations as rather provisional, and I am not convinced that my present limitation of the Northeastern American forms of S. glauca can be taken as a definite solution of this difficult question. 1918] SCHNEIDER—AMERICAN WILLOWS 327 As previously stated, ANDERSSON also described a S. glaucops var. glabrescens from specimens collected by BourGEAU in 1858, probably in Alberta, Rocky Mountains district, near the Bow River Pass. The description runs thus: “8, glabrescens, amentis crassis vulgo multifloris, foliis rigidioribus supra sparse pilosis demum glabris subtus sat intense glaucis.’’ Furthermore he said: ‘‘quan) autem glabrescentem appelavi longius distat et S. chlorophyllam (e typo S. phylicifoliae) non parum revocat, foliis glabris supra lucidis et nervosis subtus glaucis reticulatis, amentis multo brevius pedun- culatis et capsulis distinctius pedicellatis. Ad S. desertorum mani- festissimum praebet transitum.’’ There is a specimen of Bour- GEAU’S in Herb. G. which, in my opinion, represents a co-type of ANDERSSON’s variety. It bears the label of PALLISER’s Expedition and is named “S. glauca L. x pallida glabrata And.’’ ANDERSSON not infrequently changed a name in his publication after having marked the herbarium sheets in a different way. The specimen consists of 2 pieces of well fruiting branchlets. The aments meas- ure up to 5:1.5 cm. and do not differ from typical var. acutifolza (syn. var. villosa) as collected by BourGEAU and Drummonp in the Rockies. The co-type before me apparently represents a less gla- brescent form, and it approaches much the other variety, with which it seems connected by a whole series of intermediate forms. The most glabrous ones I have seen were collected in the vicinity of Dawson, Yukon Territory. I deem it best to give the following characteristics and synonomy of the two varieties: 1. S. GLAUCA var. acutifolia, comb. nov.—S. villosa Barratt apud Hooker, Fl. Bor.-Am. 2:144. 1839, p.p.; SEEMANN, Bot. Voy. Herald 39 (Fl. W. Eskimaux Land). 1852.—5S. villosa var. acutt- folia Hook., Fl. I.c.—S. glauca var. villosa And. in Ofv. K. Vet.- Akad. Férh. 15:127. 1858, p.p.—S. glaucops a villosa And, in DC. Prodr. 16?:281. 1868, p.p.—S.. glauca Richardson in FRANKLIN, Narr. Jour. Polar Sea, Bot. App. 753. 1833, non L.; CovILLE in Proc. Wash. Acad. 3:321. pl. 39. 1901.—?S. glauca subarctica Kjellman, Fanerog. Vest-Eskim. Land 51. 1883, in Nordenskiéld, Vega Exp. Vet. Takttag. 2:51. 1883, non Ldstr.—S. Seemannii Rydbg. in Bull. N.Y. Bot. Gard. 2:164. 1901.—S. glauca, var. Seemanii Ostenfeld in Vid.-Selsk. Skrift. I. Math.-Nat. Kl. 1909. 328 BOTANICAL GAZETTE [OCTOBER no. 8:34 (Vasc. Pl. Arc. N. Am. Gjéa Exp. 1904-6). 1910.— Frutex erectus, 0.5-1.5 m. altus; ramuli novelli dense albo- sericeo-villosi vel villoso-tomentosi, hornotini vix vel paullo glabres- centes, annotini biennesque pl. m. purpureo-brunnei vel epidermide secedente flavo-cinereo-brunnei, sparse vel partim villosulo- tomentosi vel glabrati, interdum pl. m. nitiduli, circ. 2-3 mm. crassi, vetustiores similes, glabri, saepe castanei; gemmae ut videtur ovatae, ventre pl. m. applanatae, apice saepe rostratae, obtusae, initio dense pilosae, demum fere glabrae, purpureo-brunneae, 4-5 mm. longae. Folia adulta pl. m. chartacea vel papyracea, inferiora minora (vel pedunculorum) elliptica, obovato-oblonga, oblanceolata, vel obovalia, basi obtusa vel vulgo cuneata, apice obtusa ad subacuta vel subrotunda et apiculata, integerrima, minimis margine saepe tenuiter glanduloso-denticulatis exceptis circ. 2:0.8 ad 3.5-4:1.5 cm. magna, superiora majora late lanceo- lata vel oblanceolata, elliptico- vel obovato-oblonga, ovato- vel obovato-elliptica, apice obtusa vel acuta vel fere breviter subacu- minata, basi obtusa vel vulgo subito vel sensim cuneata, 5:15 vel 4.5:2 ad 7:2.3 cm. magna, surculorum saepe late ovato- vel obovato-oblonga vel late elliptica, 7:3 ad 12:6.5 cm. magna, maxima interdum margine distincte breviter denticulata, superne novella adpresse sericeo-villosa, etiam adulta pl. m. villosula vel vulgo glabrescentia, ad costam marginemque tantum distinctius villosula, estomatifera, costa nervisque lateralibus planis vel subim- pressis, nervillis satis indistinctis, subtus novella densius quam superne sericeo-villosa, etiam adulta pl. m. -_adpresse sericea vel villosula, valde discoloria, albescentia vel glaucescentia, costa flaves- cente nervisque lateralibus utrinque circ. 6-10 pl. m. elevatis, ner- villis saltem in foliis adultis tenuiter prominulis. Petioli 3-12 mm., in surculis ad 2 cm., longi, superne sulcati, ut ramuli pilosi, vel dein glabrati, flavescentes. Stipulae ut videtur semper evolutae, in ramulis vegetis maximae, pl. m. lineari- ad semi-cordato-lanceolatae, rarius semiovato-rotundae, satis distincte glanduloso-denticulatae, ut folia pilosa, petiolo duplo breviores ad } vel fere duplo longiores, maximae surculorum ad 3:1 cm. magnae. Amenta coetanea vel subserotina, pedunculos ut rami pilosos foliatos terminatia, satis longe cylindrica, sub anthesi vulgo densiflora, rhachide villosa; 1918] SCHNEIDER—AMERICAN WILLOWS 329 mascula pedunculis 0.5 ad 1.5 cm. longis exceptis (2—-)2.5-4 cm. longa et o.8-1 cm. crassa; bracteae oblongae vel obovato-oblongae, apice obtusae vel rotundae, rarius acutiusculae, concolores et strami- neae vel ut videtur saepius pl. m. bicolores, apice purpurascentes vel fuscescentes, utrinque villosae et apice pl]. m. sericeo-villosae; stamina 2, filamenta libera (rariter ad basim paullo connata), basi vel fere ad medium pl. m. pilosa (an interdum glabra ?), adulta bracteis duplo ad 2 . splo longiora; antherae ut videtur ellipsoideae, flavae vel initio roseae; glandulae vulgo 2, interdum dorsalis non visa; ventralis late ovato-rectangularis vel oblongo-conica, apice truncata vel apice incisa vel bi(—3)fida, quam bractea 2-2. 5plo brevior; dorsalis vulgo distincte minor et angustior, integra; feminea sub anthesi vulgo 2-4:0.8-1 cm. magna, fructifera laxiora satis elongata 3.5 ad 7 cm. longa et circ. 1.5 cm. crassa, pedunculis 1-3 cm. longis exclusis; bracteae oblongae, obtusae, iis florum mascul. similes; ovaria sub anthesi ovoideo-oblonga, subsessilia vel pedicello glandula breviore suffulta, dense albo- vel griseo-villoso- tomentosa; styli distincti, subcrassi, integri vel apice bifidi vel fere bipartiti brachiis saepe divaricatis quam stigmata bifida oblonga vix vel circ. $plo longiores; glandula 1, ventralis, pl. m. late ovato- rectangularis et integra vel bifida ad bipartita, bractea circ. duplo brevior; fructus maturi ellipsoideo-conici, ut ovaria vel laxius vil- losi, pedicello vulgo distincto glandulam interdum 3 (rarius 2plo) superante excluso (6-)7—-8(—9) mm. longi, valvis apertis paullo recurvatis. his variety seems closely connected with the following one by inter- mediate forms, although the extremes look rather different. S. GLAUCA var. glabrescens, nov. comb.—S. glaucops, var. gla- brescens And. in DC. Prodr. 167: 281. 1868.—S. Austinae Rydbg., Fl. Rocky Mts. 198. 1917, non Bebb, pro parte.—Frutex ut in var. acuti- folia descriptus, ab ea signis sequentibus praecipue differt: ramuli novelli vulgo minus dense villosi, hornotini pl. m. glabrescentes, rarius ab initio fere glabri vel citissime glabri, annotini biennesque glabri vel sparse (saltem partim) pilosi, olivaceo-purpurascentes, pl. m. nitiduli; folia apice saepe magis acuta, margine (sal- tem inferiora) saepius sed satis obsolete denticulata, majora 330 BOTANICAL GAZETTE [OCTOBER obovato-elliptico-oblonga vel late elliptico-lanceolata ad 6-7:2- 2.5 cm. magna, superne saepe ab initio glabra, vividius colorata, subtus novella tantum pl. m. dense villosa, citius quam in acutifolia glabrescentia, adulta fere glabra vel parce pilosa, albescentia; fila- menta basi sparsius pilosa, bracteae florum satis glabrescentes; amenta fructifera in co-typo ad 5 cm. longa et 1.5 cm. crassa, basi vix laxiflora, sed vulgo satis variabilia et basi pl. m. laxiflora; fructus pl. m. glabriores vel basi glabri. So far as I can see, the range of var. glabrescens extends through the Rockies of Alberta and British Columbia to the northwest corner of this state and adjacent Alaska northward into the Yukon Territory, in the vicinity of Dawson. It probably occurs also in Alaska and the eastern Northwest Territories together with var. acutifolia. As mentioned in the synonymy, RypBERG has used the name S. Austinae Bebb as the specific designation for S. glaucops glabrescens And., but he also determined forms of a different origin as S. Austinae. This species had been proposed by Bess in Watson, Bot. Calif. 2:88. 1879, but BeBB himself stated (in Bor. Gaz. 16:106. 1891) that it forms a mixture of 3 species, including S. Lemmonii Bebb and S. lasiolepis Bth. The female piece only represented an apparently new willow, and it is described as having “sessile aments appear- ing before the leaves, with small early deciduous bracts, dark scales, clothed with silky hairs.” I fail to see how the name given to such a different form can be applied to S. glauca glabrescens even if we raise this variety to a specific rank. Before it is possible to define correctly a variety like glabrescens we have to become much better acquainted with S. pseudolapponum, the so-called S. glaucops of the Rocky Mountain floras. A few words must be said about the “glauca”’ of northeastern arctic America and of Greenland. I have to take into considera- tion the willows of Greenland because the flora of (at least western) Greenland is essentially American, and the forms of Labrador and northeastern arctic Canada cannot be properly understood without elucidating those of Greenland. The best enumeration of Green- land’s Salix has hitherto been. given by LANGE in his Conspectus Fl. Green. pt. 1. 1880 and pt. 2. 1887. In 1880 he cites not less than 5 varieties under S. glauca, which I cannot interpret correctly without comparing the specimens LANGE had before him, which are preserved in the herbarium at Copenhagen. So far as I can judge by the figures and quotations cited by LANGE, none of those varieties 1918] SCHNEIDER—AMERICAN WILLOWS 331 seems to be identical with the typical S. glauca or any of the forms of Northeastern Canada. The specimens from Labrador and Greenland referred to S. glauca by RyDBERG do not belong to it or are at least very uncertain in their relationship. There is only one form before me which seems to be closely connected with the true S. glauca, and of this I shall say something under S. anamesa, after having discussed the types and relatives of S. desertorum, S. pseudolapponum, and S. cordifolia. 2. S. DESERTORUM Richardson, Bot. App. in Franklin, Narr. Jour. Polar Sea 753 (reprint p. 25). 1833; ed. 2.765 (reprint p. 37). 1833; Hooker, FI. Bor.-Am. 2:151. 1839, pro parte; ANDERSSON in DC. Prodr. 16:281. 1868, excl. var.; RYDBERG in Bull. N.Y. Bot. Gard. 1:272. 1899; excl. specim. Drummond.; BALL in Trans. St. Louis Acad. Sci. 9:85. 1899, pro parte.—S. glauca *S. deser- torum And. in Ofv. K. Vet.-Akad. Foérh. 127. 1858.—This is one of the most misunderstood willows, and I am sorry to say that I have not yet been able to explain it sufficiently. The type was collected by RicHARDSON at old Fort Franklin on the Mackenzie iver. I have before me a photograph and fragments of the type material preserved in the Hookerian Herbarium at Kew, which show that the specimens distributed by BARRAtTT under no. 70 are identical with it. Unfortunately all the specimens have only young flowers and leaves except a few fragments of a fruiting catkin of the previous year in the Kew specimen. Hooker (1839) referred to S. desertorum also specimens collected by DrumMoND, and BEBB (apud Rorurock in Wheeler, Rep. U.S. Geol. Surv. West of rooth merid. 6:Bot. 241. 1878) apparently took DRuMMOND’S no. 657 for the typical S. desertorum, as did BALL (1899) on BEBB’s author- ity. RYDBERG (1899) said: “It is evident that Mr. Bess did not exactly know the true S. desertorum,’’ and he stated that it is Drummonp’s no. 658 that “matches RICHARDSON’S specimens exactly.’’ Both of DrumMonp’s specimens are before me. There is no doubt that no. 657 belongs to S. brachycarpa Nutt. (S. stricta Rydbg.), but I am likewise convinced that no. 658 is not identical with RicHARDSON’s type. This number consists of two young male and female branchlets, and it differs chiefly by the pubescence of the young parts (the lower surface of the leaves, etc.) which is 332 BOTANICAL GAZETTE [ocTOBER mixed with minute fulvous hairs, by the rather long pedicel of the young ovaries which is about twice as long as the gland, and by the absence of a dorsal gland in the male flowers. At present | am unable to determine this plant correctly, as we do not know much of the Salix of the regions where DrumMMonpD’‘ collected. ANDERSSON first mentioned S. desertorum quasi as a subspecies of S. glauca, and he said: “‘Insignis sane est forma, in orbe vetere quantum scio, non crescens. ... Transitus vero ad normalem S. glaucam non nunquam reperti; videtur itaque hujus modificatio frigida.”’ In 1868 he kept S. desertorum as a species, and added the following varieties: a, elata, 8, stricta, and y, fruticulosa. The last two are, in my opinion, nothing but S. brachycarpa Nutt. The first — is based on specimens collected by DRuMMonD in the Rockies, but no number is given. It is described as “‘frutex 4—-5-pedalis, ramis subsimplicibus crassis rufescentibus, foliis basi subangustatis supra glabris venis modice impressis subtus demum_ glabrescentibus amentis semipollicem longis.” This description rather fits the male pieces of DRuMMOND’s no. 660 in herb. G., while the female piece of this number can hardly be distinguished from S. brachy- carpa. This male no. 660 is the only one of DRUMMOND’S specimens I have seen that may belong to the true desertorum. This seems to be a species confined to the northern parts of Alberta and the North- west Territories, but the young types are not sufficient to give a correct idea of the species. There is another specimen, however, preserved in the Torrey herbarium at New York and labeled ‘‘Salix desertorum Fl. Bor. Am.” It consists of 4 pieces; a fruiting branchlet in the upper left corner of the sheet, a female one under- neath it, and 2 male pieces at the right hand. The fruiting branch- let is undoubtedly S. brachycarpa, while the male and female material may be identical with the true S. desertorum. The male branchlet seems to represent a late flowering stage, and it bears * According to J. Macoun (Cat. Can. Pl. I. preface p. viii. 1883), DrumMOND “explored the whole country from the Red and een Rivers by the North Sas- katchewan and Athabasca to the Rocky Mountains.” He also “collected in the main range of the Rocky Mountains, between lat. oe and particularly in the part about the head of the Smoky River, a tributary of the Peace.” Dr. J. M. MAcouN is spending the summer of this year in these regions and will probably bring back many of the forms collected by Drummonp from their original localities. 1918] SCHNEIDER—AMERICAN WILLOWS 333 rather far advanced leaves, the largest of which measure up to 4 cm. in length and 12 mm. in width. They are narrowly elliptical, acute at both ends, finely puberulous on the midrib and on some of the veins above, and glaucescent and almost wholly glabrous beneath, with an entire, ciliated margin. To the true S. desertorum may also belong (at least partly) the following 2 specimens collected by J. W. Tyrutt: Hudson Bay, west of Chesterfield Inlet, Septem- ber 2, 1893 (no. 1711 O., f., fr.), and between Lake Athabaska and Chesterfield Inlet, July-August 1893 (no. 1712 O., m., f.). Both numbers consist of several small pieces apparently taken from different plants, and it is impossible to judge them properly. I think it best to give the following description of the type material, because I shall not be able to insert S. desertorum into the key. Frutex erectus sesquipedalis (fide RicHarpson), habitu ut videtur S. pseudolapponum non absimilis; ramuli novelli satis dense, rarius laxius albo-sericeo-villosi, hornotini paullo glabrescentes, annotini subglabriores vel tantum partim pilosi, brunnescentes vel interdum ut vetustiores vulgo nondum perfecte glabri purpura- scentes et nitiduli, adulti flavo-brunnei, epidermide griseo secedente obtecti; rami cinereo-badii vel nigrescentes; folia speciminum typi- corum valde juvenilia vel vix semi-evoluta membranacea, elliptico- lanceolata vel obovato-lanceolata (ex auctore “‘exacte elliptica’’), apice obtusa vel vulgo pl. m. acuta, basi subito vel sensim in petiolum angustata, superiora integerrima vel tantum versus basim parce et obsolete denticulata, infima fere circumcirca tenuiter glanduloso-subdenticulata, maxime evoluta ad 3:0.9 cm. magna; superne ab initio glabra vel sparse (ad costam densius) villosula, in sicco subnigrescentia, valde indistincte subinciso-nervata et reticu- lata, stomatibus (an semper ?) instructa, subtus pl. m. discoloria, glaucescentia, pruinosa, magis (saltem inferiora) sericeo-villosa sed cito satis glabrescentia pilis adpressis difficile recognoscentibus (tardius evoluta ut videtur distinctius pilosa); petioli nondum satis evoluti, vix ad 5 mm. longi, laxe sericei; stipulae nullae vel vulgo pl.m. evolutae, minimae vel parvae, maximae lanceolatae, glanduloso-denticulatae, parce pilosae, petiolis circ. }— 2 breviores; amenta coetanea vel subserotina, cylindrica, ramulos breves foliatos sub anthesi vix ad 1 cm. longos terminantia, rhachide villosa; 334 BOTANICAL GAZETTE [OCTOBER | mascula 2-3.5 cm. longa et circ. 8 mm. crassa, densiflora; bracteae oblongae ad obovatae, apice obtusae rotundataeve, stramineae vel flavo-brunneae (vix fuscae), pl.m. laxe praesertim ad apicem sericeo-villosae vel distinctius sericeae, extus saepe glabrescentes; stamina 2 filamentis liberis basi pl.m. pilosis dein bracteis duplo superantibus; antherae flavae( ?), ellipsoideae, satis parvae; glan- dulae 2, ventralis ovato-conica, apice truncata, interdum pl.m. bifida, bractea 2.5—-3plo brevior, dorsalis minor, angustior, vulgo integra; feminea sub anthesi 1.5-3:0.7 cm. magna; bracteae ut in masculis, sed brevius villosae, vix sericeae; ovaria oblongo- ellipsoidea, albo-villoso-tomentosa, sessilia vel subsessilia, styli sub anthesi breves, stigmatibus brevibus oblongis bifidis vix longiores, integri vel apice breviter bifidi; glandula 1, elongato- conica, apice truncata vel subretusa, interdum leviter incrassata, bractea duplo brevior; fructus tantum pauci anni praeteriti ex herbario Kewensi visi ellipsoideo-conici, subrostrati, circ. 6.5 mm. longi, satis glabrescentes vel tantum basi pedicelloque quam glan- dula subduplo breviore pilosi. 3. S. PSEUDOLAPPONUM v. Seemen in Bot. Jahrb. 29. Beibl. 65: 28. 1900; RypBERG, Fl. Rocky Mts. 197. 1917.—S. glauca villosa Andersson in Ofv. K. Vet.-Akad. Forh. 15:127. 1858, pro parte; BEB in Coult., Man. Bot. Rocky Mts. 338. 1885, pro parte max.— S. glaucops Rydberg in Bull. N.Y. Bot. Gard. 1:270. 1899, p-p-™-; BALL in Coult. and Nels., New Man. Rocky Mts. Bot. 135. 1909, p.p.m.—S. desertorum Ball in Trans. St. Louis Acad. Sci. 9:85. 1899, pro parte.—S. glauca var.? Ball, l.c. 88, p.p.—S. Wolfii var. pseudo- lapponum Jones, Willow Fam. 17. 1908, prob. tantum ex parte.— To understand this species it is necessary to compare the explana- tions given under S. brachycarpa and S. desertorum. As I have already explained under S. glauca, the names S. glaucops and S. glauca villosa And. cannot be used for those forms which are named S. glaucops by Batt and also by RypBERG, who keeps S. pseudolapponum as a different species. In his Flora Colorado 93. 1906, he distinguished them in the key by the following charac- ters: “Leaf-blades oblong or linear-oblong; bracts obovate; shrub depressed,” S. pseudolapponum, and “‘leaf-blades oblanceolate or obovate-lanceolate; bracts oblong; shrub not depressed,” 1918] SCHNEIDER—AMERICAN WILLOWS 335 S. glaucops. In his Fl. Rocky Mts. 190. 1917, he says the same and adds that the leaves are 2-3 cm. long in the first, while they measure 3-6 cm. in length in the second species. The largest leaves of S. pseudolapponum I have seen measured up to 5-5:1.8 cm., but usually they are not longer than 4-4.5 cm., and from about 1.5 to 2.2 cm. wide. RYDBERG apparently refers to his glaucops some forms which I do not regard as belonging to it, giving as the range ‘Alta —N.M.—Utah—Calif—Yukon,”’ while he restricts S. pseudolapponum to Colorado. The type of this species (Baker, Earle, and Tracy, no. 300%, male) came from Mount Hesperus in the La Plata Mountains in southwestern Colorado, and represents a young flowering stage which naturally looks rather different from a fully developed specimen with old fruits. After having compared an extensive series of well collected specimens, I fail to see how it is possible to separate specifically this southern - Colorado plant from the other forms in Colorado, where the species seems to have its headquarters, but the typical S. pseudolapponum may represent a dwarfed more alpine form of the so-called S. glaucops, which, therefore, should be distinguished as a new variety of S. pseudolapponum. There are several forms which otherwise seem to be identical but do not have stomata in the upper leaf epidermis, with which the typical S. pseudolapponum is always provided, differing in this respect from S. brachycarpa (see later). So far as I can judge by the copious material before me, these two - Species seem to hybridize rather freely, and I cannot explain certain forms in any other way. We need, however, a much more careful study of these forms in the field to decide the question whether these hybrids are common. ' From New Mexico | know S. pseudo- lapponum only in a somewhat uncertain sterile form from Taos County, Costilla Valley (leg. E. O. Wooton, September 4, 1914), and from Wyoming I saw no specimen but Nelson’s no. 7831 from the Medicine Bow Mountains in Albany County. From farther northward I saw specimens from Teton County, Montana (leg. C. S. Sargent in 1883), and from Alberta, Sulphur Mountain, near Banff (leg. A. Rehder, August 8, 1904). Specimens from Lake County, Utah, need further observation, and I have seen nothing from Nevada, California, Oregon, or Washington which I can refer 336 BOTANICAL GAZETTE [OCTOBER to this species. A more intimate acquaintance with the Salix flora of these regions may lead me to a different opinion, but I hesitate to refer any doubtful forms to a certain species as long as I do not yet know all the other willows that may occur in the locality. Different species may sometimes look very similar at a certain stage of their development, and it needs a long time and the most scrupulous observation to become familiar with the variation of such polymorphic plants as the willows usually are. 4. S. BRACHYCARPA Nutt., North Am. Sylva 1:69. 1843; RypBERG, Fl. Colorado, 95. 1906; Fl. Rocky Mts. 197. 1917; BALL in Coult. and Nelson, New Man. R. Mt. Bot. 135. 1909.— S. desertorum Andersson in DC. Prodr. 167:281. 1868, saltem var. 8 et y, non Richardson; Bess apud RorHRock in Wheeler, Rep. U.S. Geog. Surv. west rooth Merid. 6: Bot. 241. 1878; in Coulter, Man. Bot. R. Mts. 338. 1885, excl. var.; BALL in Trans. Acad. Sci. St. Louis 9:85. 1899, pro parte.—S. stricta Rydbg. in Bull. N.Y. Bot. Gard. 1:273. 1899; in Mem. N.Y. Bot. Gard. 1:114 (Cat. Fl. Mont.). 1900.—The type of this graceful and well marked species was collected by Nurra.t in August 1818 “‘in the Rocky Mountain range, on the borders of the Bear River, a clear rapid brook cutting its way through basaltic dykes to the curious lake of Timpanagos, in New Mexico”’ (now the Great Salt Lake of Utah). No type specimen seems to be in existence, neither have I seen a plant from the type locality, but NuTraLt’s ample and vivid description leaves no doubt as to the form of which he is speaking. ANDERSSON entirely misunderstood this species when (in 1867 and 1868) he added NutTaLt’s name with ? as a synonym to his S. longi- folia argyrophylla angustissima. Rowtee (in Bull. Torr. Bot. Club 27:248. 1900) seems to have been the first who reinstated Nvttatt’s name for S. stricla (And.) Rydbg. As already stated, S. brachycarpa is apparently connected with S. pseudolapponum by intermediate forms, and in 1899, through his investigation of the Rocky Mountain material, BALL was led “to the conclusion that no rigid line can be drawn between the species as they are represented in that region.”” The extreme forms, he said, are widely divergent, but the numerous intermediates present an almost perfect gradation between these extremes. After all, this 1918] SCHNEIDER—AMERICAN WILLOWS 337 is true only to a certain degree, and in my opinion the difficulty might be settled by regarding the intermediate forms as hybrids. Compared with each other, S. brachycarpa is distinguished by the denser and shorter, almost tomentose pubescence, the absence of stomata in the upper leaf epidermis, the shorter petioles, and the denser and shorter aments, especially the staminate with their minute globose anthers; while S. pseudolapponum seems to be well marked by the looser, almost a little silky-villose pubes- cence, the relatively longer petioles, the presence of more or less numerous stomata in the upper leaf surface, and by the somewhat looser male aments with rather stiff filaments and larger, more ellipsoid anthers. In the female aments the differences are often less obvious, and the differences given by BALL (1909) and by RYDBERG (1917) seem to me not borne out in fact. I have seen no material of S. brachycarpa from Utah where the type had been collected. The species seems to be abundant in central Colorado from the Culebra Range in the south to the Medi- cine Bow Mountains in the north, and southern Wyoming, where it is frequently met with in the western part of the state and in or near the region of the Yellowstone Park, including northeastern Idaho and southern Montana. There is also a specimen before me from the Wallowa Mountains in southeastern Oregon (Cusick, no. 2298). From. northern Montana its range extends in the Rockies to about 59° N. lat. and about 122° W. long., while in Alberta it occurs east of the Athabasca River through Saskatchewan to about 59° N. lat. I also have before me specimens from Churchill on the western shore of the Hudson Bay in Manitoba, and from the Gaspé Peninsula, which I am unable to separate even as a variety. At first sight the eastern forms seem to differ by the relatively shorter and broader leaves, the somewhat longer styles, and the longer ventral glands, but the same variations can be observed in western specimens. The form from Churchill (J. M. Macoun, no. 79156 O.), however, needs further observation. An uncertain form is represented by no. 74. Hb. H.B. and T. (fr.; N.), named S. desertorum var. acutifolia. It differs from the type by foliis sub- acutis ad 32:9 mm. magnis et praecipue amentis fructiferis satis laxifloris ad 3.5:1 cm. magnis. 338 BOTANICAL GAZETTE [OCTOBER Professor J. M. Macoun, to whom I am indebted for much help, has collected in company with M. O. Motte a very interesting variety at Jasper Park, Alberta, on the low point running into the Athabasca River on the west side of the discharge of Beauvert Lake, July 30, 1917 (no. 95374, fr.; O.), which has glabrous or almost glabrous ovaries and fruits. It resembles S. chlorolepis, but the leaves of this species possess stomata in the upper surface which soon becomes glabrous, while in the western form the leaves are without stomata as in typical S. brachycarpa and have the same kind of pubescence. I also received a male specimen collected by J. M. Macown at the same place as the female type on July 23, 1918, and I am giving the following description of this variety for which I propose the name: S. BRACHYCARPA var. glabellicarpa, nov. var —Frutex ut videtur parvus, dense et breviter ut in var. fypica ramosus; ramuli novelli vetustioresque ut in illa; folia conferta, anguste lanceolata, oblan- ceolata vel anguste elliptico-lanceolata, apice acuta vel subito apiculata, basi cuneata ad subrotunda, 7:2 ad 28:8 mm. magna, integerrima, sed infima pl.m. dense tenuiter glanduloso-denticulata, superne infimis exceptis pl.m. laxe villosula, vivide ( ?) viridia, esto- matifera, costa rubescente vel flavescente subimpressa nervis vix visibilibus, subtus discoloria, glaucescentia, densius villosula vel inferiora initio magis sericea demum glabrescentia, costa promi- nente, nervis lateralibus utrinque ad 8 angulo acuto a costa abeun- tibus vix vel paullo prominulis; petioli 1-3 mm. longi gemmis (an satis evolutis?) ad subduplo longiores; amenta pedunculo ad 5 mm. longo normaliter foliato suffulta; mascula circ. 8:5 mm. magna ceterum a typo non diversa; fructifera circ. 1:1 cm. magna, subglobosa; ovaria sessilia vel subsessilia, glabra vel ad apicem parce villosa, stylo integro, stigmatibus siccis parvis bifidis ad 2. 5plo longiore coronata; glandula 1, ventralis, anguste ovato- conica, quam bractea obovata flavescens vel apice straminea utrinque laxe villosa subduplo brevior; fructus ovoideo-conici, 4-5 mm. longi, ut ovaria glabra vel apice sparse pilosa. 5. S. CHLOROLEPIS Fernald in Rhodora 7:186. 1905, is a species peculiar to the Gaspé Peninsula, where it was detected in 1905 at the headquarters of Ruisseau du Diable on the famous Mount 1918] SCHNEIDER—AMERICAN WILLOWS 339 Albert by Fernald and Collins (no. 59, m., f.; G., type). As FER- NALD has already pointed out, it closely simulates in habit, bark, and foliage S. brachycarpa, but differs from it by its glabrous cap- sules and glabrous green bracts. There are, however, pubescent forms which look rather intermediate between S. brachycarpa and S. chlorolepis, and which have been taken for hybrids by FERNALD. The main difference between the two species is, in my opinion, found in the glabrousness of the filaments in chlorolepis, which are more or less pilose in brachycarpa, and in the presence of numerous stomata in the upper leaf epidermis of S. chlorolepis, while S. brachy- carpa is entirely destitute of them. The pubescent form agrees well with typical S. chlorolepis in this respect, and cannot therefore be regarded as of hybrid origin; consequently I propose the follow- ing variety: S. CHLOROLEPIS var. antimima,’ var. nov.—S. desertorum Fer- nald in schedis, non Richardson.—A var. typica nonnisi differt ramulis foliisque novellis bracteis vulgo extus et ovariis omnino vel parte superiore pl.m. breviter cinereo-villosulis, foliis vulgo oblongioribus ad 3:1 cm. magnis etiam adultioribus subtus saepe sparse pilosis. The following specimens have been examined: Quebec: Gaspé Peninsula, Mt. Albert, on wet serpentine slopes, July 23, 1906, Fernald and Collins (nos. 512, 512%, f., 512°, f., type, 512, f., 512°, fr., 512‘, m.; G.); July 21, 1906, Fernald and Collins (nos. 518, m. paratype, 519, {; G.; no. 519 forma inter- media inter var. typicam et var. antimimam videtur et ab cl. FERNALD sub nomine chlorolepis X desertorum distributa est); ravine of cold brook, local, alt. goo m., August 12, 1905, Collins and Fernald (no. 64,m.,f.; A., N.; “ascending shrubs 3-6 dm. high”). There are indeed also some forms which have to be regarded as true hybrids between S. chlorolepis and S. brachycarpa. Ishall deal with them on a later occasion. 6. S. NIPHOCLADA Rydberg in Bull. N.Y. Bot. Gard. 1:272. 1899; CovILLE in Proc. Wash. Acad. Sc. 3:322. fig. 20. 1901.— This species is still very little known. Its type was collected in 1892 by Miss E. Taytor in the Northwest Territories on the “Mackenzie River, at a point 30 miles north of the Arctic Circle.” I did not see the type specimen, but the specimens mentioned by CoviLte (F. Funston, no. 185 and E. A. and A. E. Prebble no. 26), 5 Derived from dvrtuiuos, closely resembling. 340 BOTANICAL GAZETTE [OCTOBER who identified them with the type. The first came from the mouth of the Porcupine River in eastern Alaska, while the second was found near the mouth of the Seal River, 40 miles northwest of Fort Churchill on the Hudson Bay. Through the kindness of Professor J. M. Macown I saw also a small specimen collected by F. Johansen at Icy Reef in northeastern Alaska in 1914 (no. 164 or 93794 O.), which agrees well with Funston no. 185. S. niphoclada is “ appar- ently nearest related to S. stricta’ (S. brachycarpa) as stated by RyDBERG, while CovILLE was of the opinion that “the nearest relative to the species among American willows is S. glauca.” In some respects S. niphoclada seems to approach S. desertorum, which, however, is still too insufficiently known. The statement in Ryb- BERG’S description, “style 5 mm. long,” is clearly a misprint for 0.5mm. Owing to the lack of more copious material I am unable to elucidate the genetic relations between S. desertorum, S. nipho- clada, and S. brachycarpa, nor can I properly define the taxonomic characters of the first 2 species. The most significant character of S. niphoclada seems to me the dense white silky-villose pubes- cence of the first season’s shoots combined with the very short and densely silky petioles, which apparently do not exceed 2 mm. in length, while they are about twice as long and more obvious in S. desertorum. 1 am not inclined, therefore, to refer Seton and Prebble’s no. 79 (no. 78300 O.) from the Mackenzie district, Artillery Lake, Last woods, to S. niphoclada, as it has been determined by BALL, as it seems to me more closely related to S. desertorum. We know, however, almost nothing of the Salix flora of the woodland region of the Northwest Territories, which must be an Eldorado for willows. The following species which I propose is likewise characterized by the very short petioles, but it has an entirely different prostrate habit. 7. S. fullertonensis, nov. spec.—Frutex humilis depressus ramis ramulisque vulgo satis elongatis repentibus, floriferis ut videtur tantum adscendentibus. Ramuli novelli pl.m. villosuli vel breviter sericeo-villosuli, hornotini pl.m. glabrescentes, purpureo-brunnes- centes, annotini fere glabri vel partim tomentelli, intense brun- nescentes vel fere castanei, interdum subnitiduli, vix ultra 2 mm. 1918] SCHNEIDER—AMERICAN WILLOWS 341 crassi, vetustiores epidermide secedente griseo obtecti; rami pl.m. cinereo-brunnei. Gemmae parvae, oblongae, obtusae, ellipsoideae vel fere ovato-globosae, flavescentes vel purpurascentes, initio pilosae, ut videtur vix ultra 2.5 mm. longae. Folia satis parva, adulta sub-chartacea, lanceolata, ovato- vel elliptico-oblonga, inter- dum anguste ovato-elliptica, elliptica, ovalia vel obovato-oblonga, apice vulgo acuta, rarius obtusa, basi pleraque rotundata, interdum late cuneata, margine integerrima vel rarius basim versus dentibus distantibus minimis glanduliferis paucis instructa, 1:0.4 ad 2.5: o.9 cm. magna vel (in no. 79161) ad 3 cm. longa et ad 1.1 cm. lata, superne novella pl.m. villosula vel etiam adulta nondum glabra, rarius fere ab initio glabra, ut videtur intense sed satis obscure viridia, costa paulo impressa nervis lateralibus subplanis, epider- mide (an semper?) stomatifera, margine villosulo-ciliata, subtus discoloria, albescentia vel glaucescentia, pruinosa, novella et etiam adulta ut superne sericeo-villosula vel demum fere glabra, costa flavescente elevata nervisque lateralibus utrinque 5-8 prominulis ceterum satis indistincte tenuiter reticulata; petioli brevissimi, gemmis duplo breviores ad aequilongi, superne sulcati, pilosi, basi dilatati, vix ultra 2 mm. longi; stipulae vulgo evolutae, semi- cordatae vel semiovato-lanceolatae, acutae, pl.m. glanduloso- denticulatae, pilosae, 1-3 mm. longae. Amenta tantum feminea saepius fructifera visa, pedunculis (o.5—)1-2 cm. longis foliatis suffulta, cylindrica, sublaxiflora, sub anthesi circ. 1.2-1.5:0.5 cm., fructifera 2:1 ad 4:1.3 cm. magna; ovaria ovoideo-conica, dense griseo-villoso-tomentosa, sessilia, stylo brevi semipartito vel integro quam stigmata oblonga subbreviore ad sublongiore coronata; brac- teae anguste oblongae, obtusae (in no. 79161 obovali-oblongae), ‘brunnescentes, villosulae vel sericeo-villosulae, extus ad apicem interdum glabrescentes; glandula 1, ventralis, anguste ovato- conica, apice truncata, integra vel pl.m. bifida bipartitave, quam bractea-circ. duplo brevior, in no. 79161 interdum glandula dorsalis parva visa; fructus anguste ovoideo-conici, ut ovaria vel minus dense tomentosi, sessiles vel subsessiles, 4-6 (vel ad 7) mm. longi. Type Locatiry: Eastern Canada, Hudson Bay, Fullerton, lat. 63°57’. SPECIMENS EXAMINED: Canada: Hudson Bay, Fullerton, September 4, 1910, J. M. Macoun (79164, fr.; type; G., N., O.); July 10, 1904, E. L. Borden 342 BOTANICAL GAZETTE [OCTOBER (no. 63043, f.; N., O.; a young oe stage); Ranken Inlet, lat. 62°45’, August 30, 1910, J. M. Macoun (nos. 79163, 79165, 79166, fr.; Cor., N., O.; identical with type); Bathurst Tnlet: Arctic Sound, lat. 67° ms 68° N., long. tog’ to 111° W., August 25, 1915, R. M. Anderson (no. 467 or 93776 O., fr. im.; amentis satis laxifloris); Cape Eskimo, lat. 61° 05’, August 26, 1910, J. M Macoun (no. 79161, fr.; Cor., N., O.; forma foliis fructibusque majoribus, saltem in specim. in O., stomata superne in foliis ut videtur deficientibus) ; Mansfield Island, September 1884, R. Bell (no. 24622, fr.; O.; specimen mancum incertum). This is an interesting willow, and well marked in its typical form by the very short petioles of the small leaves, which are normally provided with stomata in the upper surface. It seems to be an entirely prostrate shrub with very slender creeping branches. Some of the forms I regard as S. fullertonensis or nearly related to it have been referred by RypDBERG to his S. Macounit, the type of which represents a very different plant, which I shall discuss under S. cordifolia. The following specimens look to me more or less like forms that might be taken for S. fullertonensis XS. groenlandica. They seem to differ from S. fullertonensis in the following characters: gemmis majoribus ad 5:3 mm. magnis, foliis latioribus ovato- vel obovato- ellipticis ovalibus vel obovato-oblongis apice saepe plicato, acutis basi rotundis ad late cuneatis adultis margine sparse ciliato excepto glabris superne magis nitidulo-viridibus (stomatiferis) subtus paullo distinctius nervatis reticulatisque maximis ad 2.8:1.5 cm. magnis; petiolis ad 4 mm. longis sed gemmas bene evolutas non superantibus; amentis fructiferis fructibusque vix diversis, bracteis late obovatis pl.m. longius et magis sericeo-pilosis; fructibus sessilibus vel pedi- cello distincto glandulam interdum superante suffultis, circ. 7 mm. longis. Hudson Bay: lat. 55-56°, barren shores, August 1886, J. M. Macoun (no. 18822, fr.; O.; ovariis sessilibus, bracteis sericeis, foliis distincte petiolatis, stomatiferis); Fullerton, September 4, 1910, J. M. Macoun (no. 79148 ir.; O-; 79167, fr.; Cor., G., N., O.; forma foliorum ut in fullertonensi sed petioli longiores, stomata desunt, ovaria subsessilia, glandulae saepe 2, bracteae sericeae); Ranken Inlet, lat. 62°45’, August 30, 1910; J. M. Macoun (no. 79162, fr.; Cor., G., N.,O.; S. groenlandicae satis similis); Nottingham Island, 1884, R. "Bell (no. 18820" Cine 54358 O., fr. juv.; satis ad anglorum spectans sed sine stomata); Digges Island, 1884, R. Bell (no. 188203 olim,= 54359 9.5 fragmentum, ut praecedens); Mansfield Island, 1884, R. Bell (no. 188205 olim, = 54360 O.; fr.; probabiliter ut praecedens); James Bay mouth of Albany 1918] SCHNEIDER—AMERICAN WILLOWS 343 River, July 25, 1904, W. Spreadborough (no. 62618, fr.; O.; magis ad groen- landicam spectat); Bathurst Inlet, Katur Point, lat. 67° to 68° N., long. 109° to 111° W., August 22, 1915, R. M. Anderson (no. 456 or 93775 O., f.; specimen mancum). 7. 5S. CORDIFOLIA Pursh, Fl. Am. Sept. 2:611. 1814; TRaut- VETTER in Nouv. Mém. Soc. Imp. Nat. Mosc. 2:208. pl. 9 (De Salic. frig. Kochii). 1832; Hooxer, FI. Bor.-Am. 2:152. 1839, exclud. specim. Drummond.—S. callicarpaea Trautv., Ic. 295, pl. 7; RypBERG in Bull. N.Y. Bot. Gard. 1:270. 1899, quoad specim. labrad.—S. planifolia Hook., l.c. 150, quoad specim. labrad. saltem ex parte, probabiliter non Pursh.—S. alpestris c) americana Andersson in Ofv. K. Vet.-Akad. Foérh. 15:129. 1858.—S. arctica 8 Brownei 3° fumosa And. in DC. Prodr. 167:287. 1868, quoad pl. labr.—S. glauca Rydbg. in Bull. /.c. 271, quoad pl. labr.—S. Wag- hornet Rydbg., l.c., pro parte; Britton and Browy, IIl. FI. ed. 2. 1:604. fig. 1486. 1913.—S. labradorica Rydbg., l.c. 274, pro parte max.—PuRsH’s description of this species is very short and runs as follows: “S. depressa; foliis ovalibus subacutis basi cordatis integerrimis reticulato-venosis supra glabris, subtus pallidis nervo margineque pilosis, stipulis semicordatis.’’ It was taken from a sterile plant cultivated ‘‘in Hort. Andersson.” PursH adds “in general habit it resembles S. myrsinites.”’ Unfortunately there is .no type left by Pursu, but a specimen from ANDERSSON’S garden is preserved at Kew, of which I have not yet seen a photograph, but only a rough outline sketch in herb. G. The plant is next mentioned by Fores (Salict. Wob. 277. fig. 143. 1829), who only translated PursH’s diagnosis. The leaf represented in fig. 143 clearly shows a finely denticulate margin, and it looks much more like a leaf of S. calcicola Fern. Iam unable to ascertain its identity. HOOKER said: ‘‘The plant named for me by Mr. Borrer, who is probably acquainted with the original plant cultivated by ANDERS- SON, little deserves the appellation of cordifolia, its leaves being more frequently acute than retuse at the base. Many of the speci- mens approach very near the following” (S. arctica R. Br.). Ihave not yet seen the Labrador type of HookEr’s cordifolia collected by KOHLMEISTER. Hooker also referred to this species specimens collected in the Rockies by DrumMonp which represent S. arctica 344 BOTANICAL GAZETTE |OCTOBER subcordata (And.) Schn. (see my first paper). In the synonymy he mentioned S. obovata Pursh with a ?, but this species is described with ‘‘amentis subcoetaneis sessilibus’”’? and does not apparently belong to our species. Furthermore, HooKeEr’s S. planifolia is the same as S. cordifolia as to Miss BRENTON’s specimens from Labra- dor, of which I have a photograph and fragments before me. The sheet in herb. K. contains 6 specimens with fruits and adult female flowers of which only one (the middle piece at the left hand side) seems to belong to a different form on account of the presence of stomata in the upper leaf surface which are wanting in typical S. cordifolia. Judging by the ample descriptions and the figures, TRAUTVET- TER’S.S. cordifolia and S. callicarpaea seem to represent nothing but two different stages of one species. His S. cordifolia is a poor speci- men of a female plant with young flowers, while the figure of his callicarpaea shows a fruiting specimen collected by HERZBERG at Okak. Of Rypperre’s S. callicarpaea I have only seen BELL'S Labrador specimen from ‘“‘ Nachhak” (Nachvak), a rather poor and sterile one (no. 18819, O.) which I cannot distinguish from typical S. cordifolia. The other specimen cited by RypBERG from Mt. Gaspé (prob- ably meaning Mount Albert, Gaspé Peninsula), collected in 1882 by Macoun (no. 18826 O.), has not been available to me; it may: belong to S. anglorum var. kophophylla Schn. ANDERSSON (1858) divided S. cordifolia Hook. in his S. subcor- data and S. alpestris americana, the latter representing the Labrador plant. In the Prodromus (1868) no mention is made of his alpestris and its 3 varieties of 1858, but only of the older S. alpestris Wulfen, which has nothing to do with it. S. cordifolia is cited under S. arctica 8 Brownei f. 1. obovata in the following sentence: ‘ Huc S. cordifolia Pursh fl. 2. p. 611; Hook. fl. boreali-amer. 2. 152; Trautv. /.c. p. 298 t. 9 ex Labrador forsan etiam pertinet”’; while on the following page under f. 3. fwmosa of the same variety he says “Nonne haec potissimum: S. cordifolia Pursh fl. Amer. syt. 2.611. ?, Trautv. l.c. p. 298 (quae tamen stylo longissimo insignis videtur!) ”, and S. callicarpaea Trautv. is mentioned as a quasi-synonym under the last form. Besides this ANDERSSON says under S. pyrenaica: 1918] SCHNEIDER—AMERICAN WILLOWS 345 “S. cordifolia americana, quam olim S. Pyrenaicae forma credidimus, vix a formis foliis tenuioribus nigricantibus S. villosae est distin- guenda.” This is a most curious statement, because he never referred S. cordifolia (or part of it) to S. pyrenaica, but he did pro- pose (1858) aS. alpestris a pyrenaica besides his alpestris americana. Furthermore, under S. glaucops var. villosa ANDERSSON (1868) quotes “S. cordifolia Hook. Fl. Boreal.-amer. p. 152 p.p. (non Pursh).” These statements convey the impression that ANDERS- SON was unable to interpret properly HOOKER’s species. RYDBERG (1899) proposed the new name S. Waghornei for S. cordifolia Hook., not Pursh, without explaining why both are not identical, and without mentioning the fact that Hooker in his cor- difolia also included specimens of DRuMMOND from the “high parts of the Rocky Mountains.” He says “Type in Herb. Torrey (‘FI. Am. Bor.’),’’ which is a poor and almost valueless fragment con- sisting of one piece with a few remnants of fruits and another small one with undeveloped rather abnormal male catkins. The leaves of both have stomata in the upper epidermis, and the specimen looks more like a hybrid between S. cordifolia and S. anglorum than like S. cordifolia, which is certainly not identical with this ‘type.”’ I am inclined therefore to use the name S. Waghornei for this sup- posed hybrid. RYDBERG (1899) proposed 2 more species: S. aira and S. labra- dorica. Judging by the type before me, S. aéra represents nothing but a form of S. cordifolia, of which I shall speak later, while S. labradorica is still a rather uncertain form because the female type (Waghorn’s no. 36, 1892) as well as the male syntype (Waghorn’s no. 31, 1892) differ from typical S. cordifolia by the presence of stomata in the upper leaf epidermis. The plants are too young to afford sufficient characters to recognize their real affinity. Accord- ing to RvpBEro’s key, S. labradorica differs from the other species by its broadly ovate leaves “with white, villous almost permanent hairs, spreading in all directions,” while in S. Waghornei and S. atra “the leaves are somewhat hairy when young, but the long white hairs are, as in S. glauca, appressed and parallel to the midrib.” This kind of silky pubescence may be seen on the lower surface of the first (lowermost) leaves of almost all the forms in question, 346 BOTANICAL GAZETTE ; [ocToBER while the later (superior) leaves bear more or less villous hairs “spreading in all directions,” especially on the upper surface if the latter is not glabrous even when young, as is mostly the case with the young (first) leaves of the flowering branchlets. I have been unable to distinguish different forms by the amount or the character of the pubescence, and it is often difficult to determine properly young flowering specimens in the herbarium. S. cordifolia is a widely distributed and variable species, its range extending from southern Greenland (about the 67th parallel) and Labrador (from the vicinity of Nachvak southward to the Strait of Belle Isle) westward to the western shores of the Hudson Bay (in var. aira) and southward to the Mingan Islands and the western Gaspé Peninsula,’ northwestern Newfoundland, and in var. Macounii to the Bonne Bay region in western Newfoundland, but it is not yet reported from the Bay of Islands or the Blomidon range there. The forms of Greenland which I take for S. cordifolia are discussed under S. anamesa. In Labrador it is often represented by the f. atra (Rydbg.), nov. forma, which seems to differ from the type only in its more oblong leaves which are acuter at both ends. The “‘turning black in dry- ing”’ of the leaves mentioned by RYDBERG seems to me no character of taxonomic value because it is too often only a result of neglect in the press. I shall give an enumeration of the specimens referable to f. atra in my final book. At present I wish to draw the attention of collectors to another form for which I propose the name f. hypo- prionota’ noy. forma, because it chiefly differs from the type by its “‘foliis ex parte pl.m. serrato-denticulatis”; otherwise it seems to vary in the same manner as the type, being sometimes more or less prostrate, sometimes an erect shrub up to 1 m. in height. I refer to it the following specimens: LaBRADOR: Straits of Belle Isle: Blanc Sablon, limestone and calcareous sandstone terraces, by brook, August 1, 1910, Fernald and Wiegand (nos. 7 Derived from 474, somewhat, and rpiovwrés, serrated. 1918] SCHN EIDER—AMERICAN WILLOWS 347 3224,{f.; G.; foliis elliptico-oblongis paullo ad f. atram spectans; 3226 fr. type; b . high”; foli A: Sohn ig iis obovato-ellipticis ad 5.8 magnis superne magis quam cubtus laxe adpresse villosis vel inferioribus minoribus m e ciliato excepto glabris); Forteau, springy ks and damp hillsides, July 10, 1910, Fernald and gand (nos. 3210, 3220, fr.; arbor, near Battle Harbor, September 15, 1891, Waghorne (no. 11°, fr.; Cor.); Ungava, along a river, July preadborough (no. 13687* O.; Cor.); Queb I , Mingan Islands, Ile St. Généviéve, July 1, 1915, H. St. John (no. 90840 O., BU ts ; Island of Anticosti, Baie Sainte Claire, August 17-18, 1917, M. Fictoots ‘bon. 4340, St., 4350, St., 4351, fr.; A. A distinct variety seems to be represented by the typical 5S. Macounti Rydbg., which came from Ellis Bay on Anticosti Island. RyDBERG referred to this species forms of different origin, but mostly those related to S. fullertonensis and S. groenlandica. It may be briefly characterized as follows: 7b. S. CORDIFOLIA var. Macounii, nov. var.—S. Macounii Rydbg. in Bull. N.Y. Bot. Gard. 1:269. 1899, quoad specim. typic. —S. Rydbergi® Heller, Cat. N. Am. Pl. ed. 2. 4. 1900.—S. vaccini- formis Rydbg. in Brirron, Man. FI. N. St. Can. 319. 1901.—A typo praecipue differt foliis etiam adultis minoribus vix ultra 3:1.5 cm. magnis vulgo satis exacte ellipticis utrinque pl.m. acutis interdum ' Margine pl.m. denticulatis adultis glaberrimis sed novellis pl.m. (saltem superne!) ut in typo villosis; amentis fructiferis vix ultra 3:1 cm. magnis. Type Locatity: Island of Anticosti, Ellis Bay. RANGE: Anticosti and northwestern Newfoundland, possibly also in Labrador and northern Ungava. SPECIMENS EXAMINED: ee Anticosti, Ellis Bay, September 7, 1883, J. Macoun (no. 18830 O., type).—Newfoundland: Ingornachoix Bay, damp rocky limestone eae near the sea level, August 4, 1910, Fernald and Wiegand (nos. 3203, f., fr., 3207, fr.; G.); dry rocky limestone barrens, near sea level, August 1, 1910, Fernald aa Wiegand (no. 3218, fr.; G.; pros- trate); August 2, 1910, Fernald and Wiegand (no. 3221, f., fr.; G.); Bonne Bay, barrens at the base of the serpentine table lands, August 27, 1910, Fernald and Wiegand (no. 3229, f.; G.); serpentine table land, alt. about 380 m., same date and collectors (no. 3230, fr.; G.). here is no reason according to the international rules or the Philadelphia code to Hie the name Macounii on account of the previous S. Richardsonii var. as HELLER in November 1900 and RypBeERG a few months later did, the latter not kiowing of HELLER’s name. 348 BOTANICAL GAZETTE [OCTOBER This variety needs further observation. It seems to be the prevailing one on Anticosti Island and in northwestern Newfoundland. Some more vigorous forms from Blanc Sablon and Forteau with more distinctly denticulate leaves might also be referable to it. RypBERG’s type is a very glabrous specimen collected in September. Forms from Hopedale in Labrador (Sornborger, no. extra 1) and northern Ungava (A. P. Low, no. 24769 O.) are rather uncer- tain. Specimens like no. 3207 have the mature leaves entirely glabrous (except a few hairs on the margin), as in the type, while the young parts show a more copious pubescence similar to that of S. cordifolia typica. There are other specimens which I cannot determine properly and which are worth further observation: Newfoundland: Ingornachoix Bay, Pointe Riche, limestone barrens near sea level, August 4, 1910, Fernald and Wiegand (no. 3204, fr.; G.), forma foliis pl.m. orbicularibus ‘vel elliptico- rotundis satis ad var. Macounii spectans, fere ut in var. typica pilosa, sed floribus femineis glandula satis lata (fere ut in groen- landica) instructis et fructibus pedicello quam glandula sublongiore suffultis laxe puberulis, stylis brevibus stigmatibus brevibus bifidis vix longioribus, bracteis obovatis substramineis breviter pilosis.— Quebec: Saguenay County, Archipel Ouapitagone, Ile Matchiatik, sprawling on ledge, July 15, 1915, H. St. John (no. 90841 O., f.; G.), praecedente non absimilis. A very uncertain form has been found by St. John on the Mingan Islands, Ile au Marteau (Eskimo Island), top of limestone shingle, July 28, 1915 (no. 90837 O., m., f.; G.): ramulis novellis perspicue dense albo-tomentosis, foliis semi-evolutis obovato-ellipticis ad 5:2.5 cm. magnis costa ex parte petioloque excepto glabris superne in epidermide stomatiferis inferioribus ut in f. hypoprionota den- ticulatis, stipulis semiovatis denticulatis glabris, floribus ut in S. cordifolia sed bracteis apice interdum leviter fuscis. Lastly, there remains to be discussed a willow from western Greenland which seems most closely related to S. cordifolia, but which also considerably resembles S. anglorum, and has apparently been referred by most of the authors to S. glauca. I cannot include it among any of the species previously mentioned, but deem it best to propose a new species. g. S. anamesa,’ spec. nova.—Frutex ut videtur habitu vanatill ut in S. cordifolia; ramuli novelli dense sericeo-villosi, hornotini » Derived from dvdyecos, intermediate. 1918] SCHNEIDER—AMERICAN WILLOWS 349 pl.m. glabrescentes, autumno ut annotini vulgo partim pilosi, badii vel purpurascentes, etiam vetustiores saepe vix omnino glabri, dein nigro-purpurascentes vel epidermide secedente pl.m. cinereo flaves- centes, ad circ. 5 mm. crassi. Gemmae ovato-oblongae, obtusius- culae, initio dense pilosae, dein glabrescentes, purpurascentes, petiolis duplo breviores. Folia adulta ut videtur papyracea, ellip- tica, elliptico-oblonga, ovali-elliptica vel elliptico-obovata, minima interdum anguste elliptico-lanceolata vel oblanceolata, margine integerrima vel rarius parva dentibus minimis sparsis glandulosis sub pilis occultis instructa (in no. 156 etiam majora distinctius sparse denticulata), maxima nondum perfecte evoluta ramulorum typi ad 2.5:1 cm. magna, in speciminibus a cl. Hartz in Augusto lectis ad 3.5:1.5 cm. magna et in ramulo vegeto (in no. 156) ad 4.8:2.3 cm. vel in forma satis incerta a Disco Island ad 5:2 cm. magna, superne ut videtur obscure viridia, in sicco vulgo pl.m. nigricantia, novella inferiora adpresse sericea, superiora pl.m. (prae- sertim versus marginem) villosula, adulta satis glabrescentia sed in costa pl.m. pilosula et margine ciliato-villosa, in epidermide pl.m. (saltem secundum nervos) stomatifera, subtus valde discoloria, ~ glaucescentia, inferiora et novella dense sericea vel sericeo-villosa (pilis adpressis albis vel paullo flavescentibus), demum glabriora et adulta interdum tantum sparse pilosa, costa nervisque lateralibus 6-10 prominulis flavescentibus et laxe tenuiter reticulata. Petioli initio dense, dein sparse sericeo-villosi, superne sulcati, 2—5(—6) mm. longi. Stipulae breviter ovatae vel ovato-lanceolatae, acutae, den- ticulatae, ut folia colorata et pilosa, 1-3 mm. longa vel nulla (punctiformia). Amenta coetanea, ramulos breves dense sericeo- villosos foliatos sub anthesi vix ultra 12 mm. longos terminantia, cylindrica, rhachide sericeo-villosa; mascula 1.2—3:1 cm. magna, basi saepe sublaxiflora; bracteae oblongae, obtusae vel subobtusae, stramineae vel apice paullo fuscescentes, omnino sericeo-villosae et apice magis sericeae; stamina 2, filamenta libera, circ. 3 pilosa, bracteis dein duplo longiora; antherae ellipsoideae, mediocres, violaceae (tantum juvenilia?), glandulae 2, ventralis anguste conica, apice truncata, integra vel pl.m. bipartita, dorsalis minor, angustior; feminea 1-2:0.8-o.9 cm., fructifera ut videtur ad 4:1.5 cm. (Hartz, Holstenborg) magna, basi vix distincte laxiflora; bracteae ut in masculis, saepe brevius pilosae, omnino stramineae; 350 BOTANICAL GAZETTE [OCTOBER ovaria ovoideo-oblonga, dense albo-villoso-tomentosa, subsessilia; styli distincti, bifidi vel bipartiti (brachiis saepe divaricatis) stig- matibus brevibus oblongis bifidis haud vel ad duplo breviores; glandula 1, ventralis, ut in masculis, bractea circ. duplo brevior. Fructus ovoideo-conici, ut videtur ad 8-9 mm. longi, minus dense quam ovaria villosi, subsessiles. TYPE LOCALITY: South Greenland, Ilua, lat. bor. 59°55’. RANGE: Southern and western Greenland. SPECIMENS EXAMINED: Greenland: Ilua, lat. bor. 59°55’, May 15-31, 1889, E. L. Lundholm (m., f., type; M.); Sermiliarsuk, circ. 61°30’, August 3, 1889, N. Hartz (ir.; N.); Kingua Kuanersok, circ. 62°, July 12, 1880, NV. Harts (m.; N.); Kvanefjord S. f. Frederickshaab, 1886, L. K. Rosenvinge (no. 18873 Q., fr.; needs further observation); Godthaabs district, Kobbefjord, June 28, 1884, Warming and Holm (m., f.; G.); Holstenborg, June 14, 1889, magi observanda); Godthavn, August 2, 1896, Cornell Party (m., f.; Cor.; forma ut videtur prostrata aspectu S. Gislerin non absimilis sed characteribus florum ab S. anamesa typica non diversa); Nugsuak Peninsula, Camp 2, August 10, 1896, Cornell Party (fr.; Cor.; forma porro observanda, bracteis magis ohevatie fructibus breviter pedicellatis, foliis apice saepe subito breviter plicato-acuminatis); Wilcox Head, August 15, 1896, Cornell Party (f., fr.; Cor.; forma porro observanda, amentis fructiferis ad 4.5:1.4 cm. magnis, fructibus pedicello quam glandula vix breviore suffultis, foliis ad 4:2 cm. —— pl. m. obovato-ellipticis); Camp 3, August 20, 1896, Cornell Party ; Cor.; forma foliis satis breviter petiolatis ceterum paullo ad S. anglorum accedens); Upernivik, 72°47’, July 18, 1886, L. K. Rosenvinge (no. 24514 O., f.; ooks very much like S. cordifolia but the pubescence reminds more of S. anglorum; on July 24 the same collector found a specimen at Proven which I cannot distinguish from S. anglorum); Cape York, July 23, 1804 , E. Wetherill (no. 214; G.; specimen mancum dubium tantum amentis reactivate adultis praeditum habitu valde ad S. anglorum spectans sed bracteis breviter villosis oblongis); Omenak (Umanak) Fjord, Omenak Island, August 9, 1897, D. White and Ch. Schuchert (no. 156, fr.; N.; forma porro observanda, foliis ad 4.5:2 magnis, amentis fructiferis ad 3.5 cm. longis et 1.6 cm. crassis). As already said, this species is certainly most closely related to S. cordifolia, from which it a. differs by the presence of stomata in the upper leaf surface. I should have treated it as a variety of this species were it not for the fact that there are a number of quasi intermediate forms between it and S. anglorum. On the other 1918] SCHNEIDER—AMERICAN WILLOWS 351 hand, S. anamesa is not identical with S. Waghornei, which I take for a hybrid between S. anglorum and S. cordifolia. Ihave not yet seen any S. anglorum south of Disco Island in Greenland, and the Greenland material which I am inclined to refer to S. cordifolia is very scanty and needs further observation. From S. anglorum the new species may at once be distinguished by its hairy filaments and its narrowly oblong, light brown bracts, which have the rather short and villous pubescence of the cordifolia type. It seems to me that S. anamesa represents the plant commonly called S. glauca by LaNncE, Hartz, and other authors, but I am not sufficiently acquainted with the Salix of Greenland, owing to the scarcity of material from there in American herbaria, to give a more proper definition of the so-called S. glauca and the numerous varieties of it described by ANDERSSON, LANGE, and others. I do not find in the existing literature a name I could apply to S. anamesa. The Salix of Greenland seem always to have been compared only with those of Europe, while in fact the material before me indicates a much closer relationship with the species from Northeastern America. If we glance at the varieties of S. glauca mentioned from Greenland, we find the following in LANGE’s Consp. Fl. Groenland. I:110. 1880, and 2:279. 1887: S. glauca var. sericea And., the type of which is S. sericea Vill., Hist. Pl. Dauph. 1:382. 1786, nom. nud.; 3:782. pl. 51. fig. 27. 1789, and which ANDERSSON refers to his f. 3. Janceolata. Accord- ing to LANGE (1880) this var. sericea and also var. appendiculata (Vahl) Wahlb. are ‘‘tolerably common on some moist places.” The latter variety is well figured by VAHL, FI. Dan. 6. fasc. 18:6. pl. 1056. 1792. Neither of these varieties seems to me identical with the forms I refer to S. anamesa. LANGE’s third variety, var. ovalifolia Lge., Fl. Dan. 17, fasc. 50:11. pl. 2981. 1880 (S. glauca a sericea 2 ovalifolia And.; ?S. glauca var. Brown in Trans. Bot. Soc. Edinbgh. 9:450. 1868) pro parte, may be represented by the following 2 specimens before me: Disco Island, September 1854, Lyall (fr.; N., ex Herb. Hook.), and ‘“Gebiet des Umanakfjordes (7o-71° N.Br.),” August 18, 1892, E. Vanhéffen (no. 89[220], fr.; N.). The broad-elliptic or oval leaves which measure up to 3.5:2.3 or 5: 2.2 cm., and are more or less villous, especially on the rib of the 352 BOTANICAL GAZETTE [OCTOBER upper surface and on the margin, do not have stomata in the upper epidermis, and their villous petioles are hardly 5 mm. long. Some of the leaves, especially in no. 89, show a few fine distant teeth toward the base. The branchlets of the season are covered with rather long villous hairs, while the older ones become glabrous and of a shining dark purplish color. The fruiting aments measure up to 4.5 by 1.5 cm.,and the capsules are about 10 mm. long, including the very short pedicels. The habit of the plant cannot clearly be recognized, but there is another very similar fruiting specimen col- lected by H. E. Wetherill, at Netiulene, Whale Sound, North Green- land, August 13, 1894 (no. 226; G.), which certainly is taken from a prostrate plant. This number is enumerated by RyDBERG (1899) under S. anglorum, but it lacks the stomata in the upper epidermis, and seems more closely connected with the var. ovalifolia, being however a little more glabrescent than the other 2 specimens men- tioned. The sessile capsules are about 8 mm. long, and the bracts somewhat darker. The var. angustifolia Lange, Fl. Dan. 17. fasc. 50:11. pl. 2982. 1880, is a very striking narrow-leaved form, the type of which came from Iceland (“prope Myvatu Islandiae legit cl. Lundgren”). I much doubt if it is the same as S. glauca a sericea 4 angustifolia And. (1868). LANGE (Consp. Fl. Gr. 1:110) refers to it specimens from western and eastern Greenland which I have had no opportunity to compare. The only specimen I saw which somewhat resembles Lance's plate is Wetherill’s no. 225 from the north side of the Jones Sound, August 1894 (f.; G.), but here the leaves have stomata 1n the upper epidermis and the rather silky pubescence of the dark bracts points more to S. anglorum, of which it may be a narrow- leaved form. I have seen rather similar specimens of S. anglorum from southwestern Victoria Land (R. M. Anderson) and north- eastern Greenland (A. Lundager). : LANGE’s last var. alpina (not S. glauca 6 alpina And., which is the same as S. glauca B macrocarpa Ledeb.) is described as a “iro- ticulus humilis, repens vel prostratus, ramis adscendentibus, foliis minutis, raro ultra 4 poll. longis,” and as the type there has to be taken a specimen collected by R. Brown (of Campster) in 1867 at Jakobshavn in western Greenland (S. glauca Brown in Trans. Bot. 1918] SCHNEIDER—AMERICAN WILLOWS 353 Soc. Edinbgh 9:430. 1868, pro parte). I have seen nothing identi- cal with this variety; there is only one specimen before me from the ‘Kvanefjord S. f. Frederickshaab,”’ collected in 1886 by L. K. Rosenvinge (no. 18873 O., fr.) which I should take for a small- leaved form of S. anamesa, the narrowly elliptical leaves measuring up to 21:9 mm. I can only repeat that we have to make a much closer investi- gation of the so-called S. glauca of Greenland in order to decide which of the forms can really be referred to the European species. They are certainly not identical with the var. acutifolia and var. glabrescens previously mentioned. I strongly believe that the true S. glauca is entirely absent from Eastern North America, and here represented by S. cordifolia and its varieties. It is the main pur- pose of these explanations to call attention to what is still unknown of the difficult forms of this group of willows, of which the following remains to be discussed. 10. S. LINGULATA Andersson in DC. Prodr.: 167:281. 1868; Herder in Act. Hort. Bot. Petrop. 11:437. 1891.—This is a very poorly known Alaskan species not mentioned by CovILLe. ANDERSSON described it from specimens collected by Kostalsky ‘ad Alaxa”’ as a low shrub resembling in habit a small S. arbuscula. There are a few fragments in herb. N. ex Herb. Fischer which agree well with ANDERSSON’s description (except that the leaves are not quite glabrous above), but are much too scanty to give a distinct impression of this species. The flowers, etc., suggest those of S. desertorum, and S. lingulata is certainly closely connected with the species of the GLAUCAE with pilose filaments, but has nothing in common with S. reticulata, to which it is said by ANDERSSON “capsulis globoso-ovalibus . . . . sat evidenter referrens.” ARNOLD ARBORETUM Jamaica Prat, Mass. THE SPORANGIA OF THISMIA AMERICANA NorMA E,. PFEIFFER (WITH PLATE XVI) Of the investigations among Burmanniaceae, the morphologi- cal studies of TreuB, Jonow, and Ernst and BERNARD are prominent. These studies included both chlorophyllous and de- pendent forms, although the latter are better represented. The accounts vary considerably in completeness, since in the earlier ones close stages are sometimes lacking. That there is variety within the family in development up to the mature seed is evidenced in the widely different accounts for those forms in which there is no evidence of fertilization, as com- pared with those where this process undoubtedly occurs. The net product seems to be approximately the same, that is, a small mass of endosperm cells about an embryo of from 2-10 or more cells, usually with no differentiation. A striking exception occurs in Thismia clandestina, which has a 3-celled suspensor and a spherical body differentiated into 2 layers. As in Orchidaceae (12), how- ever, the preliminaries to this vary. Division of the megaspore mother cell may produce a row of 2 cells (as Burmannia candida, 5), in which the inner cell gives rise to the embryo sac; or a row of 3 cells, the innermost of which, a true megaspore, functions in pro- ducing the female gametophyte; or the usual tetrad of angiosperms, of which the innermost megaspore is functional. a In the production of these cells the mother cell may go through a reduction division (as Burmannia Championii, 5), in which case fertilization is the rule; or it may divide by an ordinary mitotic division, so that the progeny have the double number of chromo- somes rather than the reduced number (Burmannia coelestis, 2). In all cases the embryo sac mother cell, whether a megaspore or the result of a single division. of the archesporial cell, develops by 3 consecutive divisions to produce the 8-celled stage. Polarity is early evident, and the egg apparatus is organized, with small Botanical Gazette, vol. 66] [354 1918] PFEIFFER—THISMIA 355 antipodal cells at the opposite end of the sac, while the 2 polar nuclei usually meet near the center, sometimes nearer the chalazal or micropylar end (as Burmannia Championii, 5). When the egg is mature, in some cases there is evidence of the entrance of a pollen tube with the discharge of two male cells, one of which fuses with the egg, the other with the polar nuclei, as B. candida. That the latter fusion is not a complete one is held by ERNstT and BERNARD, who see in a 3-parted nucleus with 3 nucleoli evidence against entire merging, at least in the first divisions of this endosperm nucleus. The fusion of the egg nucleus, however, is slower here than that of the 3 nuclei in the center of the sac. When there are 2~4 cells in the endosperm, the sex nuclei still remain dis- tinct in B. candida (5). Later the fertilized egg gives rise to an embryo of 2 or more cells, varying with the form studied. In cases where no fertilization has been observed there was development of seeds as indicated, except that no fusion save that of the polar nuclei occurred. Thismia javanica (3) and Burmannia coelestis (2), examples in which this condition holds, show no reduc- tion division in the formation of the ‘‘megaspore.”” This condition is the one to be expected from such work as has been done in par- thenogenetic angiosperms. The development of the seed is first evidenced in B. coelestis by the division of the endosperm nucleus, which usually results from the fusion of 2 polar nuclei; occasionally there are more than the two concerned, as 3-5, probably through the functioning of synergids or antipodal cells. Thereafter the development seems much as in sacs where fertilization has taken place. Ernst and BERNARD in their series of studies of Burman- niaceae report for B. coelestis, B. candida, B. Championii, Thismia clandestina, T. Versteegii, and T. javanica, practically the same sort of development in the endosperm region, regardless of the intro- duction of a male cell. The first division of the fusion nucleus gives rise to 2 nuclei, the lower of which is cut off by a wall. The cell thus formed is designated as the “basal apparat’’ or haustorium cell. The other nucleus, however, continues to go through succes- sive divisions in which no cell plate is formed, with the result that there are a number of free nuclei in the endosperm region. Walls then develop in this region at approximately the same time or a 356 BOTANICAL GAZETTE [OCTOBER little before the beginning of nuclear division in the embryo cell proper. The extent of tissue development in B. Championtt may be judged by Ernst and BERNARD’s statement that there are 6-8 cells in the median longitudinal line in the mature sac, and that B. coelestis has about 30 endosperm cells at maturity. The antipodal cells, never conspicuous, usually appear in a little V-shaped region below the haustorium region, sometimes as a row of cells, more frequently as two cells above one. The cell giving rise to the embryo, whether after fertilization or not, goes through at least one nuclear division, and usually more. Gonyanthes candida, as reported by TREUB (13), develops a 2-celled embryo; as reported by Jonow, and again by Ernst and BERNARD (as B. candida), it has a 3-celled embryo. Jonow (8) found in Gymnosiphon tenellus a 3-celled situation similar to B. candida, and in Dictyostegia orobanchioides and A pteria setacea a 4-celled embryo, comparable to that found in B. javanica by TREUB (13). Gymmno- siphon trinitatis (8) and Thismia javanica (3) show slightly greater development in a 6 or more-celled embryo, whereas Thismia clan- destina (4) shows the greatest differentiation in a structure con- sisting of a 3-celled suspensor and a spherical body in which a single outermost layer of cells is differentiated from the inner mass. There is a striking similarity to Orchidaceae (12) so far as extent of development of the embryo is concerned. The contrast in the mature seed, on the other hand, due to failure of endosperm devel- opment in Orchidaceae, is equally noticeable. Jouow, and later Ernst and BERNARD, have described the development of a small “nucellus polster’”’ above the embryo sac, and an even more con- spicuous tissue at the chalazal end. The possibility of the func- tioning of the latter at the time of germination of the seed as a region of water transfer (the rest of the tissue shows great cutiniza- tion) has been suggested, although no evidence of experimental character has been forthcoming. In contrast to the striking nucellus tissue at the ends, there is very evident degeneration of the cells in the middle zone or ring, as in Gymnosiphon, Burmannia candida, and Thismia clandestina. In comparison with the thorough work done on embryo sacs, the scant attention paid to the pollen situation brings forth prac- Tgr8}] PFEIFFER—THISMIA 247 tically only the method of pollination and fertilization where this process occurs. This has been reported by several workers: MIERS (9) in Dictyostegia orobanchioides, WARMING on Brazilian forms and in A pleria lilacina, and ErNst and BERNARD in Burmannia candida and B. Championii. In all these forms germination of pollen occurs in the pollen sacs, so that the tufts or bundles of pollen tubes issue from the anthers and penetrate the:stigma. Mzrers remarked that the identity of these pollen tubes is clear with the use of a common lens, while the cottony mass of threads is evident, sup- posedly to the naked eye. He distinctly stated that this is not true, however, in Myostoma and Ophiomeris (9), and took this as evidence, in his early time, that thereby ‘‘the theory of the appli- cation of pollen tubes for the fertilization of its ovules is distinctly disproved.”” Ernst and BERNARD were unable to discover this method of pollination in Thismia javanica or T. clandestina, although aware of its presence in other.forms and so alert for indications here. So far the evidence goes to show that such early germination of pollen and subsequent growth occur only in Euburmanniae, where the structure of the flower is different from that in Thismia. There seems to be a general conclusion, however, that forms are self- pollinated, through evidence such as given by SCHLECHTER in Thismia Winkleri (1, 11), where little diptera were found in the base of the flower where the pollen must fall. Investigation The material upon which the present study is based is that of Thismia americana, collected by the writer in Chicago, Illinois, dur- ing the summers of 1913 and 1914. The relationships of this form and a description of its structure, etc., were given in a previous paper (10). In very young stages the stamen set appears to be distinct earlier than the ovary parts. Each stamen, of which there are 6, produces the usual 4 microsporangia, all of which are directed away from the central axis of the flower. Thus the surface of the anther toward the center is quite flat or slightly concave, while the oppo- site one is marked by the 4 lobes, in 2 pairs, which represent the rudiments of the microsporangia. At this stage usually the 358 BOTANICAL GAZETTE [ocTOBER connectives have not become so broadened as later, so that the individual stamens appear more distinct than the tube shows at maturity. The youngest stage where differentiation appears is indicated in fig. 14, where hypodermal masses of meristematic cells, separated from each other by a double layer of sterile cells, appear beneath a distinct, large-celled epidermal layer. At this time the ovary chamber is just beginning to show distinctly with the 3 pla- centae, which later give rise to the ovules projecting inward. Later the individual sporangia show the parietal layer to be but 2 layers thick, within which there is a conspicuous tapetum, while outside of it is the epidermal layer (fig. 15). The tapetum shows dark irregular bodies which may represent waste or reserve material. At this stage it is evident in many preparations that not all of the tissue originally differentiated as ‘“‘spotogenous”’ is fertile. A number of the spores abort, so that in any one section only a few appear normal (fig. 16). Often between adjacent cells small oil globules appear as extraneous matter, possibly released through changes due to degeneration of the spores. The microspores are shed from the stamens through a longitu- dinal dehiscence of the anther. At the time of shedding one division of the microspore nucleus has taken place in such as appear functional. The tube and generative nuclei can be distinguished quite readily, although often other bits of dark staining material are present. Germination of pollen grains with formation of fine pollen tubes has been observed. By dissection of the style several tubes were traced through to the ovary cavity. At this time practically all the pollen had been shed from the stamens of the flowers under consideration. It seems likely that there is self-pollination as in other forms. The contrast with the Euburmannia forms reported lies in failure of development of the mass of pollen tubes from the microsporangia to the stigma, as reported by Miers and Ernst and | BERNARD. The structure would practically bar such a possibility, since the greatly developed stamen tube arising from the connec- tives usually extends below the level of the stigma. The dehiscence of the microsporangia occurs on the face away from the central region in which the style is erected, and the pollen falling from the 1918] PFEIFFER—THISMIA 350 sacs would naturally drop to the floor of the cavity, that is, the roof of the ovary. In this fall it is obvious that the grains cannot come in contact with the stigma, which is separated by the stamen tube, although grains have been observed along the style. The cells of the inner surface of the stamen tube are often glandular in nature (fig. 16), although this would seem to have no special. significance except in connection with the entrance of insects. It seems likely that the latter are necessary agents in pollination because of the mechanics involved. The placentae which appear in the ovary during the development of the microsporangia give rise after a time to the primordia of the ovules (fig. 1). The surface of the placentae first becomes uneven through the appearance of the little lobes marking the rudiments. Soon the inner integument appears, and finally, as the ovule assumes the anatropous orientation, the outer integument is quite distinct except on the side where the funiculus appears. Meanwhile the hypodermal archesporial cell has become differentiated (fig. 2). The condition of mother cells usually occurs in the stamens at the same time that this archesporium appears in the ovule (cf. figs. 2 and 15). This cell represents the megaspore mother cell directly, since no parietal cells are developed here. It enlarges noticeably, and at length undergoes nuclear division, during which the chro- matic material becomes massed at one side of the nucleus in synap- sis (fig. 2). After division two cells separated by a thin wall are evident (fig. 3). At the same time the whole ovule is developing rapidly, as shown by the spindles in the tissue about the megaspore mother cell or its progeny. The two daughter cells divide further. The spindle in the outermost cell is oriented at right angles to the long axis of the ovule, that of the inner parallel to this axis. The result is that there is a pair of megaspores side by side which fre- quently are so crushed together in later stages that they lie obliquely (fig. 5) or appear finally as one (fig. 10). Sooner or later these cells disorganize, as does the sister cell of the functional megaspore, which lies innermost in the series of four. The pressure of develop- ment usually shows first on the outermost megaspores (fig. 5), but sometimes the third non-functional one is crushed first (fig. 6). 369 BOTANICAL GAZETTE [OCTOBER At the time of the first division of the megaspore developing the gametophyte, the abortive cells are dark staining, often wholly dis- integrated masses of material. The binucleate stage shows nothing unusual, with its tendency toward polarity with the appearance of a central vacuole (figs. 8, 9). This stage is followed by the usual 4-nucleate situation arising from the division of each of the nuclei (fig. 10). The 4-nucleate phase must give rise very soon after for- mation to the 8-nucleate, since it represents a difficult stage to find. The early 8-nucleate stages (fig. 11) show 4 free nuclei at each pole, with a large central vacuole. This is followed by great enlargement and the organization into an embryo sac of the typical form of angiosperms, the egg apparatus at the micropylar end con- sisting of 2 large synergids in contact with the egg, 3 smaller free antipodal nuclei in the narrower, more pointed chalazal end of the sac, and 2 polar nuclei, usually coming in contact with each other near the micropylar rather than the chalazal end (12). Stages both before and after the fusion of these polar nuclei have been found. The peculiar lobed effect reported by Ernst and BERNARD in Burmannia candida and interpreted there as incomplete fusion is sometimes evident here. That there is any special significance here seems doubtful. At this time the cells surrounding the embryo sac stain more deeply and stand out more sharply than in younger stages. So far fertilization stages have not been observed. Contrary to. Ernst’s report of development in Burmannia coelestis, it seems altogether likely that fertilization does occur, since pollen tubes are developed. The development of the seed has not been followed in detail. At one time the larger portion of the sac is filled with the free nuclei resulting from the division of the endosperm nucleus. Soon walls come in, forming large cells. At about this time the egg cell undergoes division, so that a 2-celled proembryo is present imbedded in the conspicuous endosperm tissue. Further division occurs in the proembryo cells, and in the oldest material obtained (presum- ably mature seeds, although not so proved by germination) the embryo consists of many cells in a globular mass with a short suspensor region (fig. 13). The situation is much like that in 1918] _ PFEIFFER—THISMIA 361 Thismia clandestina. The endosperm is packed with reserve material at this time, and stains very deeply as a result. The development of the nucellus and integument into peculiar layers has been noted under the literature of other forms. In Thismia americana there is also at maturity a distinct mass of irregular small cells at the base connecting by means of a dark staining nucellar layer with a cap of peculiar cells at the micropylar end. The nucellar layer next to the endosperm shows fungal hyphae and many oil bodies as part of the contents. Gelatinization of the walls at the chalazal end begins early, and is responsible to some extent for the prominence of the mass of cells at that end. Enough material has not been available to try a satisfactorily large range of germination experiments. Those which have been tested have given negative results. In all probability, as in orchids, the fungus plays a réle in the early development of the plant. Summary 1. In the microsporangia the sporogenous cells develop from hypodermal masses, 4 in number, in the usual fashion. 2. At maturity the innermost parietal layer appears crushed by the large tapetal cells. 3. There is marked abortion of sporogenous cells in the micro- sporangia. 4. The division of the megaspore mother cells gives rise to 4 megaspores, the outer 2 oriented at right angles to the long axis of the ovule. 5. The 3 outer megaspores degenerate very soon, disappearing entirely after a short time. 6. The functional megaspore divides in the usual way, so that eventually an embryo sac of 8 nuclei is produced. 7. Presence of pollen tubes makes fertilization seem likely. 8. The well developed embryo is imbedded in large endosperm cells which are conspicuous in storage contents. g. In the seed the nucellus makes a conspicuous layer, develop- ing into a cap of tissue at each end. University or Nortu DaKkotTa Granp Forks, N.D. iS) w BOTANICAL GAZETTE [OCTOBER LITERATURE CITED . Encier, A., Thismia Winkleri Engl., eine neue afrikanische Burmannia- ceae. Bot. Tab: 38:89-01. fig. I 1907. . Ernst, A., Apogamie bei Raapusie coelestis Don. Ber. Deutsch. Bot. Gesells. 27:157-168. pl. 7. 1909. . Ernst, A., and BERNARD, CHARLES, Beitriige zur Kenntniss der Saprophy- ten Javas. III. Embryologie von Thismia javanica J.J.Sm. Ann. Jard. Bot. Buitenzorg 8:48-61. pls. 14-17. 1909. , VI. Beitriige zur Embryologie von Thismia clandestina Mig. und a: Veseiesit ¥. J. Sm. bid. 9:71-78. pls. 12, 13. 1911 I7. 1912 6. “J , [X. Entwicklungsgeschichte des Porbiownackes und des Embryos von B. candida Engl. und B. Championii Thw. Ibid. 10:161-188. pls. 13- _XIL. Entwicklungsgeschichte des Embryosackes, des Embryos, und pees Endosperms von B. coelestis Don. Ibid. 11:234-257. pls. 19-22. 12. 19 . JoHow, Fr., Die chlorophyllfreien tia hence Westindiens. Jahrb. Wiss. Bot. 5: 415-449. pls. 16-18. 1885. , Die chlorophyllfreien Humusbewohner. Jahrb. Wiss. Bot. 20: Ves: pls. 19-22. 1880. . Mrers, Joun, On a new genus of plants of the family Burmanniaceae. Trans. Linn. Soc. 20:373-382. pl. 15. 1851. PretrFeR, Norma E., Morphology of Thismia americana. Bot. GAZ. 57: 122-135. pls. 7-II. 1914. . SCHLECHTER, R., Burmanniaceae Africanae. Bot. Jahrb. 38:137-143- 1907. . SHarp, Lester W., The orchid embryo sac. Bot. Gaz. 54:372-385- pls. 21-23. 1912. . TrEvB, M., Notes sur l’embryon, le sac embryonnaire et l’ovule. Ann. Jard. Bot. Buitenzorg 3:120-128. pls. 18-19. 1 EXPLANATION OF PLATE XVI res were drawn with the aid of the camera lucida, and show mag- — A nifications as follows: figs. 1, 3, 5, 10, X840; 13, 14, 15, 16, X500; 17, X 260; 2, 7, 8, 9, 11, X916; 4, 6, X784; 12, X1651. F G. 1.—Primordium of ovule. Fic. 2.—Synapsis in megaspore mother cell. Fic. 3.—Daughter cells of megaspore mother cell. Fic. 4.—Four megaspores. Fic. 5.—Four megaspores, two outer cells already disorganized. Fic. 6.—Same stage, but sister cell to functional megaspore crushed. Fics. 7-9.—Binucleate embryo sacs. - BOTANICAL GAZETTE, LXVI- 1918] PFEIFFER—THISMIA 363 Fic. 10.—Four-nucleate embryo sac; non-functional megaspores disor- Fic. 11.—Eight-nucleate embryo sac. Fic. 12.—Embryo sac at maturity; chalazal walls conspicuously gela- Fic. 13.—Embryo. Fic. 14.—Young anther, showing 2 of 4 microsporangia. Fic. 15.—Microsporangium with mother cells in synapsis, longitudinal section. Fic. 16.—Microsporangium, showing large number of sterile pollen grains, tapetum disorganizing. Fic. 17.—Portion of stamen tube, showing glandular cells of inner surface (nearest style). ROOT VARIATIONS INDUCED BY CARBON DIOXIDE S ADDITIONS TO SOIL’ HA NOVEsS i: Bo Trost ana: Li YouEs (WITH NINE FIGURES) Under discussions of tropisms in plants it has been customary to include statements relative to the tropic influences of gases on plant roots. Primary investigations on this subject were made by Motiscu,’ using seedlings of Pisum sativum and Zea Mays. Gases were caused to flow past the roots of the plants, and tropic curvatures were reported for all the gases employed. BENNETT? repeated these experiments and concluded that the results obtained were hydrotropic. BENNETT made further studies with Zea Mays, Raphanus sativus, Cucurbita Pepo, Pisum sativum, and Lupinus albus, both in artificial and in so-called natural media. Studies made with the seedling roots in air gave no indication of aéro- tropism. Studies made in earth, when the sprouted seedlings were placed between blotting papers in pots of moist earth and then subjected to streams of carbon dioxide gas for periods varying from 24 to 60 hours, gave no definite curvatures. Cannon and Free,‘ after working with Prosopis veluntia, Opuntia versicolor, Fouquieria, Coleus Blumei, Heliotropium peru- vianum, Nerium oleander, and Salix (probably S. nigra), concluded that ‘it seems probable that soil aération must be added as a factor of no less importance than temperature and water,”’ for these plants were found to have different responses to carbon dioxide added to soil. The — quotation is self-explanatory. t Contribution f R h Cl id Bacteriology Laboratories of Depart- ment = Horticulture, 7 irene Agrivultaral Experiment Station, Lafayette, India cu, H., Uber die Ablenkung der Wurzeln von ihrer normalen Wach- Gucethane durch Gase (Aérotropismus). Sitzungsber. Akad. Wiss. Wien. 3 BENNETT, Mary E., Are roots aérotropic? Bot. GAZ. 37: 241-259- 1904- 4 Cannon, W. A., and Free, E. E., The ecological significance of soil aération. Science N.S. 45: 1917. Botanical Gazette, vol. 66] [364 1918] NOYES, TROST, & YODER—ROOT VARIATIONS 365 ti) Ne ecological bearing of these facts is manifest. Although deficiency in aération has frequently been suggested as an agricultural difficulty, or as the. reason why certain species do not grow upon soils of heavy texture, it does not appear that this suggestion has had any exact experimental basis, nor does it seem to have been appreciated that different species may have great differences in the oxygen requirement of their roots and widely variant responses to differ- ences in soil aération, responses which appear to be quite as specific and significant as the responses to temperature and to available water which forms the present basis of ecological classification. One of the writers’ reported 2 preliminary experiments with Zea Mays and Lycopersicum esculentum. Flower pots containing these species were kept surrounded by an atmosphere of carbon dioxide. Practically all the aérial portions of the plants were in normal atmosphere. The plants responded differently to the gas during: and subsequent to the 2 weeks’ treatment given. This paper is a report of experiments in which carbon dioxide gas was introduced subterraneously into soil in Wagner pots. Experiments will be reported following up the work of CANNON and FREE, in which the plants will be grown in soil sealed away from the air, so that there is no chance for the oxygen of the air to diffuse down into the soil. Studies on the effects of aération on bacterial activities have convinced the writers that unless the soil worked with was sterile (which would be unnatural) or contained known organisms of known antagonisms and activities, the responses to changed conditions of aération might be due to a cessation of certain necessary biological activities, or to the occurrence of certain detrimental biological activities. Adding carbon dioxide gas to the soil was expected to change the biochemical activities of the soil, but by having the atmosphere come in direct contact with the surface, it was believed that necessary biochemical activ- ities could exist, although perhaps closer to the surface than normally. The surface of the soil of all pots was left normal (dust mulch), so that all conditions might more nearly approximate those present when the carbon dioxide content of the soil was increased by natural means. Differences in amount, nature, and type of root growth were thus to be attributed to the carbon 5 Noyes, H. A., The effect on plant growth of saturating a soil with carbon dioxide. Science N.S. 40: 1914. . 366 BOTANICAL GAZETTE [OCTOBER dioxide gas added in equal amounts and in the same manner to all pots receiving gas treatments. Equal weights of thoroughly mixed soil were put in paraifined Wagner pots of the most approved type (fig. 1). The soil in all pots was compacted by dropping each pot on the cement floor an equal number of times. Distilled water was added through the tubes to bring the moisture content up to one-half saturation, where it was kept by successive oS additions of water throughout the Bg EEE, periods of investigation. The relative position of. the pots was changed at regular intervals to : correct differences in exposure and temperature in the greenhouse. Experiment A The. Christmas pepper (Caps7- cum annuum abbreviatum) was the first plant used. Plants were started in November 1915 in the greenhouse from seed, and trans- planted February 1 into the Wagner pots. The soil used was Sioux silt loam. The plants were about 1.5 inches high and carbon dioxide treatments were commenced after the plants became estab- lished. Three pots containing 4 plants each received no applications of carbon dioxide, 3 others received carbon dioxide applications 8 hours each day, and yet another set of 3 pots received carbon dioxide applications constantly. The gas was applied at the rate of approximately 650 cc. per hour, and fig. 2 shows the method of getting the gas from the pipe line to the individual pot. The wash bottles served as a means of equalizing the flow of gas into each pot. Fig. 2 shows the Christmas pepper plants after 4 months’ treatment. At first the carbon dioxide treatment retarded growth, but by the time the picture was taken there was no great difference in size between the treated and untreated plants. Fig. 3 Fic. 1.—Wagner pot ysieey sub- irrigation tubes in place 1918] NOYES, TROST, & YODER—ROOT VARIATIONS 367 shows representative roots where no carbon dioxide gas was applied. The roots were uniformly long and fibrous and extended to the bottoms of the culture pots. Representative roots grown where the carbon dioxide treatment was 8 hours per day are shown in fig. 4. These roots did not penetrate to a depth lower than 7 inches. They were clumped and coarser when compared with those to which no carbon dioxide treatment was given. Aérial roots were quite prominent, and the main root was very thickly Fic. 2.—Capsicum annuum abbreviatum 4 months after carbon dioxide gas treatments were started: row of pots fronted by no. 11 received constant carbon dioxide treatment; row fronted by no. 8 received 8 hours’ aakeyy dioxide treatment daily; row fronted by no. 3 received no carbon dioxide treatmen set with branching roots at a depth of about 3 inches. The roots shown in fig. 5 are representative of those that grew when the carbon dioxide treatment was constant. They compare unfavor- ably with those obtained under no treatment and under intermittent treatment. Aérial roots are many and prominent. The main roots are dwarfed and coarse and irregular. No roots were found at a depth lower than 5 inches. The carbon dioxide gas added to soil growing the Christmas pepper caused abnormal root devel- opments. The gas had a much greater effect on the root devel- opment of the pepper plant than was apparent in the aérial portions. 368 BOTANICAL GAZETTE [OCTOBER The soil used for experiments B, C, and D was a fine sand, which has been classified by the Bureau of Soils as Wabash sandy Joam This soil was chosen because of its excellent physical condition and low organic matter content. Sos be e 2 Z 3 : Fic. 3.—Representative roots of Christmas pepper plants which received no carbon dioxide gas treatments. Experiment B Head lettuce plants (Lactuca sativa) about 2.5 inches in diameter were transplanted into the pots in March 1917 treatments were started at once. Fig. 6 shows the best of each ol the 3 triplicates. paw Carbon dioxide It is noted that carbon dioxide gas appears to have benefited the plants receiving treatment. These plants retained their relative sizes until harvested about 3 weeks later. 1918] NOYES, TROST, & YODER—ROOT VARIATIONS 360 Fig. 7 shows the roots from the 2 most representative of each set of 3 plants grown under the different treatments. Carbon dioxide has affected the roots of these plants, although not to the extent that it did those of the Christmas peppers. — 1 Root deve a & . : B 3 © s g 5 B t Fic. 4.—Representative roots of Christmas pepper plants which received 8 hours’ treatment of carbon dioxide daily opment departs from normal with increased carbon dioxide appli- cations. Experiment C Radishes (Raphanus sativus) of the variety “Rapid Red” were sown in Wagner pots in March 1917. None of the plants were disturbed after the seed was sown. At the time of harvest the series of plants receiving no carbon dioxide gas applications had straight tap roots, while the roots of those receiving the gas showed 370 BOTANICAL GAZETTE [OCTOBER a perceptible tendency to horizontal growth. Large numbers of small roots were growing from the base of the bulbs in approxi- mately horizontal directions. No photographs were taken of this experiment. Fic. 5.—Representative roots of Christmas pepper plants which received con- stant treatment of carbon dioxide. Experiment D Burpee’s stringless green pod bean (Phaseolus vulgaris) was grown from seed without and with the 2 carbon dioxide gas treat- ments. The plants were harvested just after blossoming ceased. Fig. 8 shows the plants growing in the best of each set of triplicate pots. The difference between the plants growing in the 3 pots is small. Fig. 9 shows the roots of the plants appearing in fig. 8. z 1918] NOYES, TROST, & YODER—ROOT VARIATIONS 371 Carbon dioxide gas additions to the soil did not prevent the roots from penetrating deeply, for in all pots the roots penetrated to the bottom. It was noted that roots grew to very near the openings ore — pea Fic. 6.—Best 3 Lactuca sativa of 9 under comparison: pot at left received no carbon dioxide, one in middle 8 hours daily, one at right constant treatment of carbon dioxide 2 at left no carbon dioxide treat- 7, 2 at right constant treat- Fic. 7,—Representativ e roots of Lactuca sativi ments, 2 in middle 8 hours’ carbon dioxide treatment daily, ment with carbon dioxide. where the carbon dioxide gas entered the pots. The gas had an effect on the development of the roots of the bean plant that was different from that observed with any other plant tested. The 372 BOTANICAL GAZETTE [OCTOBER intermittent carbon dioxide treatment was apparently about optimum for the development of the roots of this plant. Fic. 8.—Phaseolus vulgaris subjected to different carbon dioxide treatments: pot at left received no carbon dioxide treatment, one in middle received 8 hours’ treatment daily, one at right constant carbon dioxide mabe, Fic. 9.—Roots of plants shown in fig, 8: 2 at left no carbon dioxide treatments, > 2 in middle 8 hours’ carbon dioxide treatment daily, and 2 at right constant carbon dioxide treatment. 1918] NOYES, TROST, & YODER—ROOT VARIATIONS 373 Summary 1. Plants respond differently to carbon dioxide gas added to the soil in which they are grown. 2. The roots of the Christmas pepper, head lettuce, radish, and string bean were all found to be affected Bes additions of carbon dioxide gas to the soil. 3. The effects of carbon dioxide on root development were greater than those on the aérial portions of the plants. 4. The intermittent and constant applications of the carbon dioxide gas did not affect the roots of all the plants to the same extent. 5. The effect of the gas was not the same for the different plants used, although a constant treatment of 650 cc. of carbon dioxide gas per hour was apparently preventative of normal root develop- ment. 6. Decaying organic matter is held to be beneficial to growing plants. Cases have been cited by others where turning under immense amounts of green material has hurt the land temporarily; therefore the results obtained in these experiments lead to the belief that the carbon dioxide content of garden soils is sometimes detrimental to the root development of some plants growing in the garden. 7. The conclusion of CANNON and FREE that soil aération must be a factor of no less importance in plant growth than water and _ temperature is supported. PURDUE UNIVERSITY LAFAYETTE, IND ABSORPTION OF SODIUM AND CALCIUM BY WHEAT : SEEDLINGS’ HowarRp 8S. REED (WITH ONE FIGURE) Sea water, mammalian blood, and certain artificial solutions in which living cells are immersed are capable of continuing the life of those cells for considerable periods of time. These so-called “balanced solutions” may contain different ions which, separately, have a marked deleterious effect upon the cell, but which, when present in certain proportions, “balance” or “antagonize’’ each other. The result is that organisms live normally in such solutions. There are two ways in which the mixture of ions or molecules in a balanced solution may overcome cytolytic factors: (1) by “antagonizing”? each other, that is, by opposing and mutually excluding each other at the surface of the plasmatic bodies or other units of living structure; (2) by producing in the organism such a state of “tolerance,” that is, by producing effects on the intracellular complexes, either alone or in conjunction with each other, that the harmful effects of single ions or molecules are eliminated. Or, in other words, the antagonism of ions may be either peripheral or internal. Until recently the majority of physiologists were inclined to the former view, a view which was clearly stated in TRAUBE’S “sieve theory of permeability’ and in Overton’s “‘lipoid-solubility”’ theory. Of the many objections to these two theories and to their various modifications, none was more cogent than that based on the fact that, even in a balanced solution, ions do slowly enter the cell. Indeed, if such were not the case, it would be impossible for the cell to obtain the salts necessary for its existence. The objections to the former ideas of “antagonism” and “tolerance” have largely been met by a theory of antagonism pro- * Paper 47, from the University of California, Graduate School of Tropical Agri- culture and Citrus Experiment Station, Riverside, California. Botanical Gazette, vol. 66] (374 1918] REED—WHEAT SEEDLINGS 375 posed by OsTERHOUT,? which holds that the slow penetration of salts may produce effects on the cell quite different from those produced by rapid penetration, and that their action is on the life processes rather than on the manner or rate of penetration. “From this point of view we regard the slow penetration of salts in bal- anced solutions not as the cause but as the result of antagonism, or rather we may regard the slow penetration and the increased length of life (or growth, etc.), by which we measure antagonism, as the results of certain life processes which are directly acted on by the antagonistic substances.”’ OsTERHOUT found that the theory was satisfactorily supported when the penetration of certain known mixtures of NaCl and CaCl, into living cells was studied. He makes, however, a seeming exception in the case of solutions of lower concentration, stating: “Below the saturation point’ the relative proportions of the salts will be of less importance than their total concentration; this is the case at low concentrations in the region of the so-called ‘nutritive effects.’”’ SHIVE and TorrincHaM, on the other hand, not to mention others, have rather definitely shown that there are certain dis- tinctly favorable ratios in nutrient solutions of equivalent con- centration. In view of OsteRHouT’s rather sweeping exclusion of nutrient solutions of low concentration, it seemed profitable to the writer to investigate the effect of some of OsTERHOUT’S proportions in weak solutions, coupled with analyses of the plants to determine the amounts of solute taken up. It is a pleasure to acknowledge my indebtedness to Mr. J. F. BREAZEALE for the cultures and analyses upon which this work is based. The experiments to be reported were conducted on wheat seed- lings grown on disks of perforated aluminum buoyed by glass bulbs ? OsTERHOUT, W. J. V., ne: penetration of balanced solutions and the theory of antagonism. Science, N.S. 16. A dyanshical cn, ofantagonism. Proc. Amer. Phil. Soc. 55:533. 1916. 3 The term “saturation point” as used is taken to mean the point at which the surface of a plasmatic structure is saturated with the antagonizing salts. Beyond this point an increase in their concentration in the solution produces no effect on their concentration in the surface 376 BOTANICAL GAZETTE [OCTOBER in such a way that the aluminum disk floated at the surface of the solution. Each disk was floated on about 3 liters of solution in an agate enameled pan. The seeds were germinated in a solution of the same composition as that designed for the experiment. None but sprouted seeds were used. Each disk originally held about rooo seedlings, but careful selection brought the number down to about 200. In attempting to study the effect of a small amount of calcium, special precautions are necessary, owing to the abundance of cal- cium compounds in our environment. The seeds and apparatus used must be washed in dilute HC] and rinsed with distilled water of undoubted purity. The cultures must be carefully protected from dust, especially dust from plastered walls, or from cement floors, which might carry salts of calcium, since 1 part per million of calcium may produce distinct effects. It is necessary to work in somewhat the same way as one works with cultures of bacteria. The antagonism of calcium and sodium has been a matter of record in connection with OsterHouT’s data. The case is illus- trated by the cultures shown in fig. 1, which show the toxic action of 4000 p.p.m. NaC] and the antidoting action of CaSO,, CaO, and Mg(HCO,),. The wheat seedlings shown in the figure were similar at the outset and grew 7 days in the respective cultures. The concentration of 4000 p.p.m. NaCl is about the toxic limit for wheat seedlings under these conditions, yet 30 p.p.m. of a calcium salt antagonized completely the toxicity. Magnesium bicarbonate was not so successful in overcoming the bad effects of sodium chloride. A further illustration of the antagonistic action of calcium is shown in table I, which gives data pertaining to wheat plants grown in solutions of sodium chloride with and without the addition of CaO. It will be seen that (1) measured by ash content and by dry weight of plants the addition of 30 p.p.m. CaO was beneficial to growth; (2) the amount of NaCl absorbed by the plants was not decreased when CaO was added. In the case of cultures 4 and 5 of table I the ratio of NaCl to total ash is 1:1.9 where only NaCl was present in the solution and 1:2.3 where both NaCl and CaO were present. From this it would seem that the calcium salt 1918] REED—WHEAT SEEDLINGS 377 has not benefited the plant by excluding sodium, especially in view of the amount found in the tops, but rather has rendered it harm- less within the plant. A second set of cultures was made in which the ratios of sodium to calcium were identical with some of those employed by OsTEr- HOUT. The pure NaCl and CaCl, solutions were each 0.004M. Fic. 1.—Effect of calcium and magnesium salts upon toxicity of sodium chloride to wheat plants: culture solutions were jar 1, 4000 p.p.m. NaCl; jar 2, 4000 p.p.m. NaCl plus 30 p.p.m. CaSO,; jar 3, 4000 p.p.m. NaCl plus 30 p.p.m. CaO; jar 4, 4000 p.p.m. NaCl plus 30 p.p.m. Mg(HCQ;),. This is much less than the concentration of NaCl employed in the first series, being 230 p.p.m. of NaCl instead of 4000 p.p.m. This series of cultures was continued for 16 days, at which time 100 rep- resentative plants were withdrawn from each culture, weighed, and analyzed, giving the data shown in table II. 378 BOTANICAL GAZETTE [OCTOBER It seems quite evident from these results that one of the ratios of Na:Ca in which OsterHouTt found the greatest amount of antagonism was the one most favorable for growth in this series. TABLE I EFFECT OF CALCIUM OXIDE ON GROWTH AND ABSORPTION OF SODIUM CHLORIDE BY WHEAT PLANTS re Se ANALYSIS OF 190 ENTIRE PLANTS PLANTS GROWN IN Days : Green | Dry Ash NaCl weight | weight Ash CaO | NaCl 1; — Waterco: 8 jo. 0.0462 TPAC Ge 5 BOF. is aes trace 2; oe. NaCl oS 8 0775|0.03201...... L098 es fe 0.0553 3 Same plus 3p pm. CaOl 8 j6 ogatio o3s4l:. eee Weis oh eae .0505 ‘ p-p.m. NaCl...... pe ieee peer an (eee seas .6 | £.74010.0940|0.0006)0.0485 Ss. Same plus 30 ppm. Cae AG tyes 12.2 | 1.820/0.1240|0.0030/0.0543 6. ofan ie pine ee Ora ee ee ea 0.0570|0.0020]...... 7. 4000 p.p.m lag te Bee Be ie ew Pg ee er, ©.0660/0.0009}..---- 8. . Same 3op.p.m CAROL Ae Galore iow eau ©.O0910/0.0017]..---- Plants grown in the solution containing 98Na:2Ca attained the greatest dry weight and were larger than any others in the series. From this solution the greatest amount of ash constituents was TABLE II WHEAT PLANTS GROWN 16 DAYS IN SOLUTION CULTURES; 100 PLANTS WITH SEEDS ATTACHED tice Ni oe COMPOSITION OF 100 PLANTS cei ov Mb SOLUTION TO ASH : Dry weight Ash Ca Na oe eee t. 100Na: -0Ca. : 2.10 ©.1950 °. 0.017 I:11.5 a. Gane, Sa 2.74 0.2250 ©,0102 0.022 1:10.2 4, SgNar aoa 52. 2.37 0.1730 0106 0.019 1: 9-% 4, Gal aCe &. 2.41 0.1842 0.0132 0.016 55555 5: ONG: 62Ca... 2.27 0.1930 | 0.0204 0.012 1:16.1 6. ONS: ioeta. 2.33 ©. 2040 0.0226 ° 1:25-5 absorbed and also the greatest amount of sodium. None of the other solutions appeared to contain as favorable a ratio of sodium to calcium, although the dry weight of plants in the last 4 solutions does not vary enough to offset the experimental error. The 1918] REED—WHEAT SEEDLINGS 379 Na:ash ratio in the first 4 sets of plants does not show any real difference. In the last 2 the relative amounts of sodium are less. The amounts of sodium and calcium actually absorbed and retained by the plants may be of more interest to consider because me hos show what the plants in the various cultures actually The results shown in table III are taken from table II, - TABLE III AMOUNTS OF NA AND CA ABSORBED FROM SOLUTIONS BY WHEAT PLANTS IN 16 DAYS AMOUNTS ABSORBED RATIOS FURNISHED Na Ca ¥, SOONG? OUR. cg 0.009 0.000 2: GENS: 20a oslo. 0.014 0.0042 Boo) BSNS T5OB io ce 0.011 0.0046 A> Ghee See ak a fore) 0.0072 BEWas Gete e sr: 0.004 0.01 6 : BS Maa ge ° 0.0166 and show the amounts of sodium and calcium after deducting what was found in plants grown in solutions containing none of the ele- ment in question. These results show that the greatest amount of sodium was absorbed from the solution containing the ratio 98Na:2Ca. Less sodium was absorbed from any other combination, even from pure sodium chloride. At the same time they also show no such selec- tive absorption in the case of calcium. The amounts of calcium absorbed increase steadily as the amount in the solution increases, reaching their maximum in pure calcium chloride. There appears to be a “‘preferred ratio” of calcium for sodium, but none of sodium for calcium, although more cultures employing smaller amounts of sodium chloride should have been tried. Summary The antagonism of calcium and sodium which has been found by other workers exists also in more dilute solutions and may be shown by chemical analyses of the plants grown therein. 380 BOTANICAL GAZETTE [OCTOBER Concentrations of sodium chloride which were strongly toxic to wheat seedlings were antidoted by 30 parts per million of calcium oxide. The most successful antagonism in the concentrations employed was found when the Na:Ca ratio was 98:2. At this ratio the calcium was not found to exclude sodium from the plant, but to render it harmless after entrance. The antagonism appears to be internal rather than peripheral. UNIVERSITY OF CALIFORNIA CitRUS EXPERIMENT STATION RIVERSIDE, CAL. BRIEFER ARTICLES METHOD OF REPLACING PARAFFIN SOLVENT ITH PARAFFIN Some years ago LAND’ proposed a method oft insuring the gradual saturation with paraffin of the xylol now almost universally used as the clearing medium and paraffin solvent in the paraffin method. He suggested the placing of a section of fine wire screening below the level of the xylol in the shell vial upon which the paraffin ordinarily grated into the vial would be held. Such an arrangement obviates the danger of allowing the paraffin fragments to fall to the bottom of the container, there to be in contact with the material and to surround it almost at once with a high percentage of dissolved paraffin. For some years previous to the publication of LAND’s suggestion in this matter we had been employing, in this laboratory, similar devices to insure the gradual saturation of the paraffin solvent. Some time ago, however, we replaced this-‘method for most material with another which is more simple and has given consistently good results. In this scheme melted paraffin is carefully poured on the surface of the xylol in a shell vial until a plug of the desired thickness is formed. A hot needle run around the inside of the vial will loosen the plug of paraffin, which is then pushed down below the level of the xylol. We have never found an instance in which the paraffin plug slipped down to the bottom of the vial or indeed changed its original position appreciably. The entire paraffin plug becomes rather rapidly saturated with xylol and a layer of xylol-paraffin soon forms at the lower surface of contact. If the vial is not shaken, a very gradual saturation of the xylol takes place, and at the end of 4-6 days the xylol has taken up its maximum quantity of paraffin. In practice we pour a plug of paraffin which will weigh 4-6 gr. into a vial containing 10-15 cc. of xylol. The increase in time required according to this scheme in the paraffin method seems justified by the rather ideally slow replacing of the paraffin solvent by paraffin. It is often desirable to cool the vial containing the *Lanp, W. J. G., Microchemical methods, an improved method of replacing the paraffin solvent with ‘pareifin. Bort. GAZ. 59: 397-1915. 381] [Botanical Gazette, vol. 66 382 BOTANICAL GAZETTE [OCTOBER xylol under the tap before pouring in the paraffin. The latter step requires only slight practice to be successful, and indeed the only effect of a too rapid pouring in of the melted paraffin seems to be the formation of strings of paraffin reaching down into the xylol. If the paraffin plug needs to be removed at any time, this can be accomplished readily by forcing through it a hot needle, the tip of which has been bent at right angles. The needle after cooling for an instant may be turned slightly and the plug pulled from the vial—T. H. Goopsprep, University of California. ADAPTATION AND NATURAL SELECTION I wish to correct a false impression which my paper on the agency of fire in the propagation of the longleaf pines (Bor. Gaz. 64:497-508. 1917) has left in the minds of some of my correspondents, to whom it seems that the conclusions there reached might lead to the absurd economic paradox that forest fires should be encouraged for the con- servation of our pine lumber supply. As a matter of fact, all the evi- dence produced in that paper was intended to show that it is their adaptation for resistance to fire which insures the survival of this species. The action of natural selection in this case, as in practically all others that have come under my observation, is negative and indirect. It preserves not by selection of the fit, but by elimination of the unfit, thus giving the best adapted a free hand in the struggle for existence. In calling attention, therefore, to the peculiar relation between the longleaf pine and fire, there was no thought of suggesting that we should imitate the method of nature; but having learned that a clean forest floor and plenty of sunshine are essential conditions for the propagation of the longleaf pine, these conditions may be secured by other means than fire, such as judicious cutting and thinning, and a periodic cleaning up of the forest floor. Whether a well guarded ground fire at the proper season might not be a useful aid in accomplishing this last purpose is a ques- tion which must be left for the practical forester to decide—E. F. ANDREWS, Rome, Ga. CURRENT LITERATURE MINOR NOTICES Mosses and ferns.—A third edition of CAMPBELL’s well known textbook! has appeared. The body of the text is the same as in the second edition of 1905, the new material being added in the form of an appendix, under the corresponding chapter headings. In addition to numerous contributions by other investigators, the appendix contains noteworthy results of the author in his investigations of tropical liverworts and ferns. The bibliography is com- pletely recast, including 772 titles, distributed among 336 authors. The author has rendered an important service to morphologists in bringing up to date, and in convenient form, our knowledge of these great groups.—J. M. C. NOTES FOR STUDENTS Mitochondria.—GUILLIERMOND has published a number of short reports dealing with the results of his investigations on the nature and function of mitochondria. In a paper? dealing with the origin of chromoplasts and pig- ments, he finds that chromoplasts are formed from mitochondria, more espe- cially from the elongated forms called chondriocontes; and that pigments of the xanthophyll and carotin groups are elaborated either (1) directly by the mito- chondria, or (2) by chromoplasts which arose from mitochondria, or (3) by chromoplasts resulting from a metamorphosis of chloroplasts which in turn arose from mitochondria. Added interest is given because of the fact that in a great many plants the process can be observed in the living material under the microscope. Both granular and crystalline pigments have the same origin. Epidermal cells from petals of Iris germanica, Tulipa suaveolens, Tropaecolum majus, and young fruits of Arum maculatum, Asparagus officinalis, and numer- ous others furnished the material for this study. In a later paper’ dealing with the chondrium of Tulipa, he reports that the mitochondria are easily visible in the living material under the oil immersion lens. Epidermal cells from petals are used. The mitochondria are long, thin, and undulate, although smaller granular and rod mitochondria are present. Material is at its best just about the time the flowers first open. The reviewer has verified these observations with the yellow-flowered variety, but was not * CampBett, D. H., The structure and development of mosses and ferns (Arche- goniatae). 3d ed. 8vo, pp. 708. figs. 322. New York: Macmillan Co, 1918. $4.50. 2 GUILLIERMOND, A., Compt. Rend. Acad. Sci. 164:232-235. 1917. , Loc. cit. 1642407-409. 1917. 383 3 384 BOTANICAL GAZETTE [ocToBER successful with the white ones, and found further that the eS begin to disintegrate in the course of 30 minutes after the mount is m n a third paper‘ dealing with alterations of the rete hs finds that “the mitochondria are the most fragile elements of the cell, and it is through them that the first signs of degeneration and the first symptoms of trouble due to osmotic changes are manifested.” The alteration consists of the transforma- tion of the mitochondria into vesicles having the aspect of vacuoles and giving the cytoplasm an alveolar appearance. This, the author remarks, is interesting when we think of BUrscut1’s alveolar structures. In a fourth paper’ dealing with the fixation of cytoplasm, he finds that from the point of view of their action on the chondrium fixing agents may be grouped in three classes: (1) Gen Mann’s, Zenker’s, and Carnoy’s fluids all disturb the structure of the cytoplasm and destroy the mitochondria; (2) picric acid, mercuric chloride, Co and strong Flemming’s generally Regaud’s, and Flemming’s with only a trace of acetic acid are the fixing agents commonly used for fixing mitochondria and the cytoplasm as nearly like living as possible. In general it is those reagents which contain alcohol or acetic acid which alter the mitochondria most Mortter® has published a valuable contribution to the study of mito- chondria, not only in the new facts he has revealed, but more especially in the account of his methods, which will enable workers much less qualified to take up studies in this interesting field. In the main he used Flemming’s fluid, with very much reduced amounts of acetic acid for a fixative and iron haema- toxylin and crystal violet for stains. For material he used root tips of Pisum sativum, Zea Mays, and Adiantum pedatum; the thallus of Marchantia poly- morpha, Anthoceros, and Pallavacinia; seedlings of Pinus Banksiana; leaves of Elodea canadensis; and certain algae He finds that root tips of Piswm furnish excellent material for a study of the primordia of plastids and their transformations. Mitochondria-like structures are very numerous, such as rods of various lengths and thicknesses, straight, variously curved and bent, and also numerous smaller granules and slender delicate rods. Leucoplasts develop from the larger structures, but the smaller ones do not form plastids. Although these structures all give the same histochemical reactions, the term chondriosome (mitochondrium) is reserved for those smaller structures which do not form plastids. The former he calls “plastid primordia.” Zea Mays is similar in all essentials to Pisum. In Marchantia the “plastid primordia” are more readily distinguished from the 4 GUILLERMOND, A., Loc. cit. 164:609-612. 1917. 5 , Loc. cit. 164:643-646. 1917. ® Mortier, D. A., Chondriosomes es the primordia of chloroplasts and leuco- plasts. Pine ‘Bot tany ya: gI-114. pl. 1 1918] CURRENT LITERATURE 385 mitochondria in that they are larger and rounded, while the mitochondria are very small granular or rod forms. Cells from the thallus of Amthoceros were studied because they have each a single chloroplast and hence furnish favorable objects for determining whether mitochondria are merely disorganized chloro- plasts. This -saention is answered in the negative. In Adiantum pedatum finds tl lria are small, granular, and rod-shaped. Discussion is here confined to root tips, and we are promised a subsequent paper dealing with other parts. The ‘“‘plastid primordia” are rounded, lenticular, and rod- shaped, but much larger than the ern = sstelerasetees “ primordia ” which do not develop into leucoplasts ‘‘ contin long-drawn-out threads and finally disappear.” In the younger growing parts of the stem of Pinus Banksiana numerous small rounded bodies with colorless centers (plastid primordia) and densely staining granules (mitochondria) were found. In the older parts these bodies with the colorless centers form the plastids, while the granular mitochondria have become larger or formed rod mitochondria. In the leaves of Elodea canadensis the primordia are rod-shaped and can easily be traced in their transformation into plastids. The mitochondria are very numerous, and in cells with fully developed chloroplasts they are globular and even rod-shaped, differing from the primordia only in size. We are promised a later paper dealing with his results on Hydrodictyon. GUILLIERMOND includes under the term mitochondria all those structures which give the same histochemical reactions, regardless of their functions; while Mortter, on the other hand, considers only those structures which do not develop into plastids to be included under the term. Both, however, agree that these structures are ‘‘morphological units of the cell with the same rank as the nucleus.” Mortrer goes farther and asks, ‘‘ What characteristics are transmitted solely by the nucleus, and what by the primordia of plastid and by the chondriosomes? There are many transmissible characteristics which cannot as yet be definitely expressed in any Mendelian ratio c that certain phenomena of fluctuating variations and other numerous charac- teristics, Mendelian or otherwise, owe their appearance and transmission to the primordia of plastids and chondriosomes may be a daring hypothesis, but if, as there is good ground to believe, tiene bodies are permanent organs, there is no escape from some such assumption.’’—Ray C. FRIESNER. Units of vegetation and their classification.—With the advance of the sci- ence of ecology there has been a gradual evolution of opinion as to the units most suitable for the analysis and study of vegetation. The earlier stages of this evolution have been well diseussed by Moss,’ who also advanced the devel- opmental concept of the plant formation. The half decade following this paper passed without a further notable contribution to the subject, but recently three 7 Moss, C. E., The fundamental units of vegetation. New Phytol. 9:18—53. Igio, 386 BOTANICAL GAZETTE [OCTOBER articles have appeared that are notable, not only for the divergence of the views expressed, but also for the decided advance they have made in providing a logical system of classified units for the use of students of vegetation. Greason® embodies in his article an individualistic concept of ecology, contending that all phenomena of vegetation depend upon the phenomena of the individual plant. The plant association he conceives to be an area of uni- form vegetation developed by similar environmental selection from the immi- grants from the surrounding population. is position, while extreme, will prove most useful if it serves to focus attention upon the intensive study of some of the most important species of a vegetation so as to discover their reactions to various environments and to the factors which limit their invasion and establishment in plant communities. e other extreme is seen in the work of CLEMENTS,’ as expressed in what doubtless is the most notable of recent contributions to ecological literature. Without attempting to review or criticize his book as a whole, it may be pointed out that he selects the formation as the fundamental unit and regards. this plant community as an organic entity exhibiting origin, growth, maturity, and death. As an organism it is able to reproduce itself and possesses a life history which is a complex but definite process. The climax community is the adult organism of which all initial and medial stages are but stages of development. Thus CLEMENTS would limit the term formation to the climax community, while the successional series leading up to the climax formation he calls a “sere.” He has provided a complete system of subordinate units for the analysis of both formation and the sere, the former being divided successively into asso- ciations, consociations, 7 and clans; the latter into associes, consocies, socies, colonies, and families. This recognition of a plant community as an entity comparable in some extent at least to an organism seems strictly in accord with the views of most ecological workers, and if the relationship be regarded as one of close analogy rather than homology it will probably prove the most stimulating and anGatactory atetale: It Aprents, however, that CLEMENTS’ system of subordinate ore elaborate than is required to meet the needs of most investigators. A somewhat simpler system, introducing but few new concepts or terms, recently organized by NicHots,” commends itself to the reviewer as including those units and terms dssaienly in. _ past have aig an satisfactory, and which now for the first time h combined system. NICHOLS * GLE , H. A., The structure and — of the plant association. Bull. ans ha Club 44:463-481. 1917. 9 CLEMENTS, F. E., Plant succession, Carn, Inst. Wash, Pub. 242. pp. xilit-51!- pls. 61. 1916. * Nicwots, Geo. E., The interpretation and application of certain terms and concepts in the ecological classification of plant communities. Plant World 20:305~ 319, 341-353. 1917. 1918] CURRENT LITERATURE 387 himself claims that the scheme is the outgrowth of the classification originally presented by Cow Les,” and by his selection of the association as the funda- mental unit of vegetation he recognizes the tendency of ecologists as a whole to become more and more agreed upon the use of the term “plant association,” even while differing somewhat as to the content of the term. He defines the association as any community of plants, taken in its entirety, which occupies a common habitat, or in other terms, any stage in a given successional series. The “‘habitat,’”’ thus made the criterion of the association, is understood to be a unit area with an essentially uniform environment made up of a complex of climatic, edaphic, and biotic factors which determine the ecological aspect of the vegetation. The subdivisions of the association agree with those of CLEMENTs in being consociation and society, but differ in that ‘‘association”’ (and its subdivisions) is applied to both the climax and the seral units N. h ecologists recognizing this situation have preferred to regard such associations as belonging to a “‘temporary climax,” postulating the oo although much delayed dominance of a climax limited by climate o Grouping plant associations upon a developmental basis, the plant com- munity of the next higher order is termed an “edaphic formation” and defined as “‘an association-complex which is related to a specific phys: iographic unit concepts to include the developmental idea. The various climatic formations long to various “climatic formation-types,”’ several of which may form the “climatic formation-complex”’ of a continent or other large unit area. Nicuots has further demonstrated the utility of his excellent scheme of classification by applying it to the analysis of the vegetation of northern Cape * Cowes, H. -C., se a ecology of Chicago and vicinity. Bor. Gaz. 31: ne, 145-235. 388 BOTANICAL GAZETTE [OCTOBER Breton Island, appending various explanatory remarks which should prove useful to students attempting to make similar applications to other regions.— Gro. D. FULLER Permeability——Several interesting contributions to our knowledge of protoplasmic permeability have appeared recently. DEt¥F has investigated the influence of temperature on the permeability of protoplasm to water by the tissue shrinkage method, using sections of onion leaves and dandelion scapes in subtonic solutions of cane sugar. The curve of contraction at different temperatures was measured by means of an optical lever which greatly mag- nified the shrinkage, and from this curve the rate of contraction. at the time when 30, 50, and 70 per cent of the shrinkage had occurred, was measured by the tangents to the curves at these points. From the rates the values for Q:. were obtained. This value increases as the icmperetur rises. In the onion leaf the value of Q® at 10-20° C. is 1.5, at 20-30° C. is 2.6, and at 30- 40° C. is 3. o. In the dandelion scape the greatest value of Q,. was obtained at 20-30°C., at which temperatures it was 3.8. Above and below those temperatures the value falls. Contrary to the results of VAN RYSSELBERGHE, who found very little increase in permeability above 20° C., DEtF finds that permeability of the protoplasm to water continues to increase rapidly up to the highest temperature investigated, 42°C. The methods used by VAN RYSSEL- BERGHE are justly criticized, particularly with reference to the means of deriving a temperature relation from his data. The strength of solutions used by VAN RYSSELBERGHE may also have led to serious errors. Miss Hinp* has studied the absorption of acids by living plant tissues, using electrical conductivity methods, and electrometrical measurement of the cate ion concentration in acid solutions which were in contact with living potato disks and roots of Vicia Faba. She found that the hydrogen ion is aR absorbed from dilute acid solutions by living tissues, and concluded that the anion, particularly in organic acids, plays a large part in determining the effects of the acid on protoplasm. In the case of the mineral acids, HCl, HNO,, and H.SO,, the stronger solutions can penetrate the cells for a time without causing much injury as measured by exudation of electrolytes; but organic acids like formic and acetic cause very rapid increase in conductivity, due to exosmosis of electrolytes from the cell. With these two acids there is first a decrease and then after a few hours a very noticeable increase in’ H+ ion concentratién. This is thought to be due possibly to the production of acids within the tissues which diffuse out through the altered plasmatic membrane. - , E. Marion, Studies of protoplasmic permeability by measurement of rate of i. of turgid tissues. I. The influence of temperature on the per- meability of protoplasm to water. Ann. Botany 30:283-310. 1916. s Hinp, Mitprep, Studies in permeability. III. The absorption of acids by plant tissue. Ann. Botany 30:223-238. 1916. 1918] CURRENT LITERATURE 389 As to the mechanism of absorption, a few experiments furnish evidence favoring the idea that the plasmatic proteins rather than the lipoids are active in the acid absorption. ; very important paper by STILES and JorGENSEN™ challenges not only the theory of permeability proposed by CzAPEK some years ago, the surface tension theory, but also all the facts and assumptions upon which that theory was founded. . Because of its greater exactness and more general applicability to a study of all kinds of plant tissue, the Kohlrausch electrical conductivity method of estimating osmosis of electrolytes was used as a means of measuring changed permeability. Disks of potato were placed in non-electrolytic reagents of such strength as to produce irreversible changes in the protoplasm. Exos- mosis of electrolytes was measured in the presence of a number of homologous monohydric alcohols, chloroform, chloral hydrate, ether, urethane, acetone, aniline, and pyridine. In all cases corrections for the depression of con- ductivity caused by the presence of the non-electrolyte in the external solution were made. In every case the rate of exosmosis was found to depend upon the concentration of the substance in solution in contact with the disks. The higher the concentration the more rapid the exosmosis, and CzAPEK’s observa- tion that any member of the homologous series of primary alcohols has a greater effect on osmosis than a lower member of the series, if of equimolecular concen- tration, is confirmed. No such thing as a critical concentration, however, below which exosmosis did not occur and above which it did occur, could be found. Exosmose of electrolytes occurred in all concentrations used, down to mere fractions of the critical concentrations for exosmosis found by CzaPEK’s crude methods. The rate of diffusion of electrolytes was found not to be a function of surface tension alone. If the critical concentration of isobutyl alcohol were to be taken as 0.3 M. and the other alcohols compared with it as to equal exosmosis in a given time, the surface tensions of the alcohols do not agree at 0.68 of the surface tension of water, as CZAPEK stated, but vary from 0.79 in methyl alcohol to 0.59 in isoamyl. The higher the alcohol the greater the lowering of the surface tension required to produce a given amount of exosmosis in a given time. Each item of evidence and the whole tissue of assumptions upon which CzaPek built his theory of the plasmatic membrane is considered in detailed fashion and without gloves. The authors reject each int and assumption as untenable. In their own words, ‘‘from this review of the details of Czapek’s work on the plasma membrane, it is clear that neither the experimental evidence nor any part of the theory based upon it can be accepted.” They have sought to apply the law of mass action to the rate of osmosis in cases of permeability involving irreversible changes in the proto- plasm, and a mathematical expression has been deduced connecting the time 4 Stices, WALTER, and JéRGENSEN, INGvaR, Studies in permeability. IV. The action of various organic substances on the permeability of the plant cell, and its bearing on CzapeEx’s theory of the plasma membrane. Ann. Botany 31:47-76. 1917. 399 BOTANICAL GAZETTE [OCTOBER element with exosmosis. Curves representing the equation derived thus on theoretical grounds resemble in type those obtained in actual experiments. The methods used in this work seem admirably adapted to a crucial test of CzarEk’s theory, which seems entirely untenable in view of the evidence submitted.—Cuart-es A. SHULL. Desiccation.—An investigation of the course of desiccation and partial starvation in cacti has been made by MacDovucat, Lone, and Brown.’ The principal studies center upon the changing rate of water loss, chemical changes in the food reserves, plasmatic colloids, and cell sap, and the morphological changes which occur during long periods of desiccation. In one case a large Echinocactus was under observation for 6 years after removal of the plant from the soil. Water loss is rather rapid at first, but proceeds more and more slowly with time. While ro per cent of the water was lost the first year in one specimen, during the sixth year only 5 per cent of the water remaining at the beginning of that year was lost. The loss of water is much more rapid of course in the open than in diffuse light and Echinocactus can withstand desiccation not more than 2 years with free exposure. The chief chemical changes noted during the starving period concern the carbohydrates. The density of the cell sap decreases, due to disintegration of the carbohydrates, and the reducing sugars are found mainly in the inner part of the cortex in desiccated specimens rather than near the surface as in normal plants. The total amount of reducing sugars decreases during desiccation, while non-reducing sugars are increased noticeably in the cell sap. Reduction of the amount of sugars leads to reduction of acidity if the light intensity is sufficient for photolysis of the acid. In weak light even, if the sugars run low, the acids may accumulate because of the absence of photolysis. Differences in acidity are thought to be partially responsible for differences in the colloid hydration and swelling of tissues when placed in water. The main morphological changes consist in thickening of the cuticle, thin- ning of the anterior walls of the guard cells, partial destruction of the plasmatic colloids, shrinkage in the size of the nucleus, and especially the development of cortical lacunae through hydrolysis of the cell walls of this region of the stem, The vascular tissues are not affected, and the medullary cells much less than the cortical cells —CHartes A. SHULL. The vegetation of Michigan.—From the data obtained during a few wee eks in Michigan, Harper” has listed the principal plants in the order of their abundance and has discussed certain features of the environment. He recog- nizes but two types of succession, the one from the filling up of lakes and other a MacDoveas, D.T., Lona, &. R., and Brown, J. G., End results of desicca- tion and respiration in siocsiieat plants. Physiol. Res. 1:289-325. 1015. Harper, R. M., The plant population - saben lower Michigan and its environment. Bull. Torr. Bot. Club 45:23-42. 1918] ; CURRENT LITERATURE 391 depressions, and the other that following fire. In connection with the former, he distinguishes the usually recognized types of marsh and bog vegetation and states that the main distinction between the two is in the rate of growth, the slow rate of growth in bog plants being largely explained upon the basis of a dearth of mineral plant food in the substratum, which is also supposed to account for the presence of the same species upon the uplands in colder climates. o experimental evidence is given in support of this explanation. It is also rather surprising to be told that bog vegetation is ‘“‘sometimes erroneously called xerophytic,” after the almost endless discussion of bog xerophytes. eficiency of mineral plant food is also given as an explanation of the slow progress toward mesophytism of the pine forests upon sandy uplands. Leaching is supposed to prevent the accumulation of any considerable amount of plant food near the surface of the ground. This may possibly hold for the sandy plains, but if so it is difficult to see why it should not also apply to the pure sand of the dunes, where mesophytic forests develop rather quickly and where the conifers are soon largely replaced by deciduous species In discussing the influence of fire upon forest establishment, the error is made of stating that the cones of Pinus Banksiana remain closed and attached to the tree for many years, opening and discharging their seed after burning. Closer observation would have shown that the cones that remain for several years upon this pine open and discharge their seed very promptly upon ripen- ing, and that the tree is in no wise dependent upon fire for its seeding —GEo. D. Futter. Fairy rings and their effect on vegetation.—Of more than ordinary interest is a recent paper on fairy rings by SHANTz and PreMEIsEL."” Before taking up their own researches, they present an excellent summary of past studies and theories ning them, as well as a table of the fungi that have been reported as being responsible for rings. Some fungi, as Agaricus tabularis, are very destructive to grass and other vegetation; some, as Calvatia and Lycoperdon, are beneficial; and some, as Lepiota, have little effect of any sort. in conclusions are given relative to the age of rings. The conditions in eastern Colorado are not very favorable, either for spore germination or mycelial advance; in favorable years there may be a mycelial advance from the ring center of 30-60 cm., as compared with almost no advance at all in dry years. Some of the rings are very large, and from the growth measurements that have been made, a few are estimated to be 400-600 years old. Where vegetation is stimulated, it was concluded from careful study that this is due to the reduction ' of nitrogenous organic matter to available nitrates and ammonia salts, and to the subsequent decay of the fungous filaments. Deterioration or death of vege- tation are attributed mainly to drought, caused by the prevention of water 1 SHantz, H. L., and Premetset, R. L., Fungous fairy rings in eastern Colorado and their effect on vegetation. Jour. Agric. Research r1:191-246. pls. 21. figs. 5. 1917. 392 BOTANICAL GAZETTE ; [OCTOBER penetration by the masses of fungal filaments. Vegetation thus destroyed is replaced, after the death of the fungous, first by weeds, then by short-lived grasses, and eventually by the original short-grass cover.—H. C. Cowles. Foreign pollen on Cycas.—It is well known that in some cycads the ovules reach the maximum size for the species whether pollination has occurred or not; while in others the ovules, if not pollinated, soon disorganize. Cycas Rumphii belongs to the latter category.* Female plants of this species are very abundant in Ceylon, but no male plants have been observed for several years. In local- ities where male cones of Encephalartos and Macrozamia are abundant, the pollen of these species germinates in the pollen chamber of Cycas Rumphii and causes the ovule to develop to the full size. Since the pollen of cycads germinates readily in artificial solutions, it is not strange that pollen of one species should germinate in the pollen chamber of another. In this case, however, no fertilization takes place, and mature seeds, which should show the embryo in an advanced stage of development, showed no trace of an embryo. A few years ago the reviewer pollinated Stangeria with Zamia and obtained three large seeds, which were planted but failed to germinate. It is possible that the pollen stimulated growth but failed to fertilize the egg, so that, as in Cycas Rumphii, no embryo was produced.—CHARLES J. CHAMBERLAIN. Water culture.—In a critical discussion of the water culture method of studying growth phenomena, STILES" calls attention to the limitations of the method. He points out the great complexity of the factors involved, and applies BLACKMAN’S idea of limiting factors. The difficulty of analyzing the results of such experiments, due to the interaction of so large a complex of factors, few of which, even those whose action is under investigation, can be controlled, is made clear. Some factors, as for instance the influence of the respiratory activity of the roots on the culture solutions, have been neglected in all water culture work. The variability of individual plants is so great that a large amount of labor is required to secure results even with a low degree of accuracy. Nevertheless, for certain kinds of problems it may be the only method available——Cuartes A. SHULL Grasses of Illinois.—Miss MosHEr” has published a manual of the grasses of Illinois, recognizing 204 species in 63 genera, over one-fifth of the species being recorded for the first time as occurring in Illinois. The analytical keys, escriptions, and numerous text age make the bulletin very useful in the recognition of the grass flora.—J. M * Le Goc, M. J., Effect of foreign pollination on Cycas Rumphii. Ann. Roy. Bot. Gard. Peradeniya 6:187-194. pl. 13. 1917. ‘9 STrLES, WALTER, On the interpretation of the results of water culture experi- ments. Ann. Botany 30:427-426. 1916. , Epna, The grasses of Illinois. Univ. Ill. Agric. Exper. Sta. Bull. 205. pp. Paty Sigs. 287. 1918. VOLUME LXVI * NUMBER 5 THE BOTAN IGAL GeAZETFE NOVEMBER 1978 MORPHOLOGY OF RUMEX CRISPUS CONTRIBUTIONS FROM THE HULL BOTANICAL LABORATORY 244 WINFIELD DUDGEON (WITH PLATES XVII-XIX AND TWENTY-ONE FIGURES) Introduction A chance examination of a stem of Rumex crispus L. showed the presence of well developed internal bundles. Since this char- acter is to be regarded as advanced, Dr. W. J. G. LAND suggested that it might be of interest to investigate the morphology of the entire plant. This paper is concerned only with an account of the morphology of the floral structures; a study of the vascular situation is already under way. Aside from monographs, the genus Rumex, and indeed the entire family Polygonaceae and order Polygonales, have received little attention. Frvx (3) made a study of the ovular structures in R. verticillatus L. and R. mexicanus Meisn. (R. salicifolius Man.), which develop very similarly. The archesporial cell cuts off a primary wall cell, then forms a linear tetrad, the innermost mega- spore of which functions. He found approximately 24 chromosomes in the spindle of the first division of the megaspore mother cell, but was not certain whether this was a true reduction division. A regu- lar 8-nucleate embryo sac is formed, and pollen tubes enter, although actual fusion of the gametes was observed but once, and that in an unfavorable preparation. Rots (8) investigated several European 393 304 ; BOTANICAL GAZETTE [NOVEMBER species of Rumex, among them R. crispus. He found the haploid number of chromosomes in the microspore mother cells to be 8 in R. Acetosa, R. hispanica, R. arifolius, and R. nivalis; 16 in R. Acetosella; and probably 4o in R. cordifolius. R. Acetosa appar- ently undergoes reduction in the megaspore mother cell, although he saw no indication of subsequent fertilization. The embryo sac very frequently degenerates. He found evidence of apogamy in some species, and thinks it probable that, at least in the group Acetosa, this has been the result of dioecism. STRASBURGER (9) early investigated Polygonum divaricatum, and his figures of the origin and development of the embryo sac have become a classic example of normal behavior. Material and methods Inflorescences and individual flowers of various ages were collected along street borders near the University of Chicago, and in a flood plain pasture near Mineral Springs, Indiana, during the summers of 1915 and 1916. They were killed in chrom-acetic, imbedded in paraffin, and cut in the usual manner. Both iron alum-hematoxylin and safranin-gentian violet were employed for staining. Some of the most important features of the morphology became apparent in the course of the investigation during the winter of. 1916-17, and as no more material could be obtained then, there remain some points the solution of which requires the collection of more flowers and careful observations on the growing plants. Normal development ORGANOGENY.—The young inflorescence of Rumex crispus is closely invested by the sheaths of successive bracts. It is pro- fusely branched, and the branches bear flower buds of considerable size before they emerge from the protecting sheaths. All the young parts are covered with a mucilaginous secretion, rendering the penetration of reagents slow. The flowers are borne in clusters at the nodes. The oldest are nearest the main stem, while the successively younger arise outside these, on the upper surface of the enlarged projecting nodes. 1918] DUDGEON—RUMEX CRISPUS 3905 The sequence of development of the floral organs is centripetal, although the petals and carpels are somewhat delayed (figs. 1-7). The sepals appear as 3 thick prominences, and rapidly grow up over the young flower. The 3 petals appear almost simultaneously with the stamens, which they closely resemble at first (figs. 3, 8, 9). The 6 stamens appear in pairs, the 2 of each pair arising so close together that their bases are joined to each other and to the sepal opposite which they lie (fig. 9@). The carpels first appear as a thick ring about the base of the nucellus, but the latter develops much more rapidly and is not inclosed by them until the megaspore mother cell is considerably enlarged (figs. 4-7). The carpels develop as a continuous ring led by 3 growing points (fig. 8), until the ovarial cavity is inclosed, when the 3 points continue separately to form the styles and stigmas (figs. 6, 7). The styles are reflexed so that the much branched stigmas are finally placed between the bases of the anthers. The inner integument appears during the prophase of the first reduction division (fig. 6), and the outer appears with the homoio- typic division (fig. 7). Both are 2-layered from the beginning. The inner grows up beyond the nucellus, and turns inward to close up and form the micropyle. The outer integument never extends much beyond the tip of the nucellus. During the development of the embryo sac, the cells of the outer layer of the outer integument and the epidermis of the ovary thicken, lose their contents, and form a continuous impervious layer (fig. 7; see also figs. 13, 14). This process is significant, because it leaves only the chalazal region of the ovule as a point of intake for nutrient materials. MEGASPOROGENESIS.—The terminal cell of a definite axial row in the nucellus enlarges as the archespore (fig. 22). It soon divides to form the primary parietal cell and the megaspore mother cell (fig. 23). The parietal cell divides twice by anticlinal walls to form a cap of 4 cells (figs. 24, 28). Occasionally any or all of these cells may divide by a periclinal wall (figs. 25, 32). Less frequently, a cell or two of the adjacent epidermis may also divide periclinally (figs. 33, 36, 44). Occasionally there are two archesporial cells, and in one ovule there was a mass of probably 7 archesporial cells, a few of which had undergone the first division. In the most 396 BOTANICAL GAZETTE stamens to each other and to sepal (a); figs. 1-7, 175; figs. 8,9, X155- [NOVEMBER 1918] DUDGEON—RUMEX CRISPUS 307 advanced case observed, the 2 megaspore mother cells had fully enlarged and were in prophase of the heterotypic mitosis (fig. 24). They were separated by a crushed cell of equal length, which may have been a megaspore mother cell, but more probably was only a vegetative cell that became so crushed that it could not divide, and was forced to elongate with the enlarging mother cells. The megaspore mother cell enlarges and elongates considerably, then undergoes two successive divisions to form a tetrad of cells (fig. 28). Apparently this division is a true reduction, for all the stages seem to be normal, and at diakinesis there are 32 pairs of chromosomes (fig. 26). While an accurate count of the chromo- somes could not be made on the spindle, careful estimation indicates that the number still is 32 (fig. 27). In the vegetative cells the spindle is shorter, and proportionally much broader, and while the chromosomes are too small and too densely massed to be definitely counted at any stage, they clearly are more numerous than in the megaspore mother cell; I could only estimate that there are about twice as many, that is, 64. Wall formation follows each of the reduction mitoses. The first wall usually divides the mother cell a little above the center (fig. 30). The second wall is usually near the outer end of the inner cell (figs. 29, 30), although in the best preparation found (fig. 28) the cell is nearly equally divided. The wall in the outer cell is always longitudinal, instead of transverse (fig. 28). There is some irregularity in the sequence of the homoiotypic division; usually the inner cell divides first (fig. 30); sometimes the divisions are simultaneous (fig. 28); and in one case the outer cell was the first to divide. The inner megaspore functions, and the others quickly degener- ate (fig. 29). The third megaspore rarely forms a normal cell, and is usually the first to degenerate (figs. 31, 33). The outer cell may degenerate before it has a chance to divide, or the 2 mega- spores may degenerate before they are separated by a wall (fig. 31). Empryo sac.—The functioning megaspore rapidly elongates and develops a large vacuole at each end, with the nucleus centrally placed (figs. 33-35). The daughter nuclei migrate to the poles, where two more mitoses produce 8 nuclei (figs. 38, 42, 44). The 398 BOTANICAL GAZETTE [NOVEMBER mature sac is of the usual organization (fig. 46). The outer end _ enlarges greatly, crushing the parietal and adjacent nucellar cells, and comes to lie in contact with the epidermis of the nucellus. The extreme inner end remains small through this and subsequent development, and in it the 3 antipodal cells are cut off by walls. While they persist until the embryo is of considerable size, they never manifest any activity (fig. 47). The polar nuclei fuse early (figs. 44-47), and the fusion nucleus lies well toward the outer end of the sac. The egg apparatus is typical. DEVELOPMENT OF .STAMENS.—The stamen primordia are at first oval in cross-section (fig. 8), but early differentiation of the archesporial cells makes them somewhat rectangular, and sets off the anther region from the filament. The archespores are single rows of cells (fig. 48); each soon divides by a periclinal wall to form the primary parietal cell and the primary microsporogenous cell (fig. 49), then both divide anticlinally. Frequently the anticlinal division precedes the periclinal (fig. 50). There are 2 periclinal divisions in the primary parietal cell, and an appropriate number of transverse divisions to keep pace with the rapidly elongating anther. The first (fig. 54) sets off a layer that finally differentiates into a well marked endothecium with characteristic spiral thicken- ings; the second (fig. 57) forms the middle wall layer and the tape- tum, and takes place about the same time as the last division in the sporogenous tissue. The middle layer is soon crushed and obliter- ated (fig. 58). The tapetal cells appear to behave in all possible ways. Accord- ing to BonnETT (1) the nucleus usually divides twice by normal mitoses, after which there may occur a great variety of nuclear fusions and abnormal mitoses. In some cases there is but one division. Usually in Rumex crispus the tapetum becomes bi- nucleate; often it is multinucleate; and now and then there may appear large irregular nuclei with many nucleoli, as if formed by the fusion of a number of small nuclei (fig. 61). Apparently the normal condition is for the tapetum to persist as a functioning nutritive layer up to the liberation of the microspores from the tetrad (fig. 63); but in keeping with the widespread degenerations occurring throughout the flower, it may begin to degenerate while 1918] DUDGEON—RUMEX CRISPUS 399 the microspore mother cells are entering reduction (fig. 58). The cytoplasm becomes vacuolate, the nucleus stains very deeply, and shortly the entire protoplast collapses. Before the microspore mother cell stage, the epidermis over most of the anther begins to enlarge and thicken, and the pro- toplasts to disorganize. This thickening extends to the cells of the connective, so that each loculus becomes inclosed by an impervious layer, except in the stomial region (fig. 10). Here a plate of the ae iw Fic. 4.—Part of periderm of young juvenile leaf —— May gl p, phloem, with eetilins rays (m) crowded with resin drops; rmal wood; s, inverse wood in transfusion sheath, staining orange with Sudan TI, e, pales Mahe of a well defined bundle of inverse wood, which may often extend from the ventral face of the protoxylem to the endodermis (figs. §=10)/ 2. RELATION OF PINES TO FERNS.—Other tumor-infected needles show phloem differentiating on the dorsal side of the inverted 2, Van TIEGHEM considered the transfusion sheath on the ventral side of normal wood in pine needles to be the homologue of the “inverse wood”’ in the petiole of cycads. Following Takepa (19), we found this untenable. 3 This Coccus has been recorded by DurrENoy from stem tumors of pines, but the name was omitted from the note (13). 4Stomatal anatomy also emphasizes the origin of a. and pines from a com- mon stock (REHFOUS 18). 442 BOTANICAL GAZETTE [NOVEMBER bundle of xylem, resulting in a fernlike state, comprising two bundles of xylem facing each other, with phloem outside (fig. 7). The relation of pines to the fern stock is further emphasized by the occurrence (in the wood or periderm of tumor needles) of all transitional forms, from normal pitted elements to scalariform cells, such as are present in ferns, and restricted to endo- dermal cells of normal needles.® 3. RELATION OF PINES TO GNETALES.— Scalariform cells in pine .—Tumor of Coccus resinifians on pine Fic. 5 needle (collected at Arcachon, dunes of Abatilles, June 1917): part of periderm (schematic); e, endo- are much more abundant in tumor than in sound neighboring tissue); d‘, lignified cells of hyperplasia due to infection by Coccus; b, epidermis and hypo- derm; x, normal wood; hloem; x', inverse wood (staining red with eosin and green with methy] green) and of vessels staining orange with Sudan III; this inverse wood develops from ventral face of protoxylem to endodermis, and is homologous with inverse wood in cycads. needles may be compared to the tracheae in Gnetales, and suggest the origin of both from a common fern stock. Fic. 6.—Part of fig. 5: ', 4. RELATION OF PINES TO EqutsE- Phloem; w, lignified vessels 23 ° normal w same in inverse TALES.—The origin and evolution of the wood; j, lignified cells with protoxylem is strikingly similar in pine hyperplasia. ood; 2, ’ The perforations in the scalariform cells of pines may be explained as derived from the fusion of enlarged bordered pits, as claimed by Tomson for Gnetales; but the reverse is probably true, bordered pits being derived from ancestral simple incom- plete perforations by acquiring highly specialized characters. 1918] DUFRENOY—PINE NEEDLES 443 needles and in the stems of Eguisetum, as it differentiates in both centripetally, from ‘‘péles ligneux,’’ and partially dissolves into lacunae. The reappearance of polystelic organs, where the stele shows two vascular strands with opposed “xylem and _ peripheral phloem, is fundamental in indi- cating the origin of conifers from a fern stock. The cycad stele may be derived from the fern stele by suppression of phloem in one of the two vascular “inverted wood.” If this in- Fic. 7.—Half of vascular strand in verted wood itself almost entirely tumor needle (collected November 1917): disappears then the normal state €, endodermis; /, periderm; w, normal nee ‘ : wood; p, normal phloem; z, inverse f the pine needle is obtained. wood; g, phloem. } JAG... Development REJUVENESCENCE AND JUVENILE LEAVES.—When- ever a resting organ grows Z again, rejuvenescence must take place, and this is always observed in pines, either at the germination of the seed, or when lateral, dormant buds are caused to develop pathologically. When the pine seed germi- © ) nates, cotyledons develop Fic. 8.—Proliferating spur shoots, springing on the young shoot, and up between the two geminate needles (a); j, tien single juvenile loaves. juvenile leaf with spur shoot in axil. These are smaller the higher up the shoot they develop, and at a certain height they are mere scale leaves. It is at the base of | , 444 BOTANICAL GAZETTE [NOVEMBER these scales that the spur shoot of the needles arises. After it has developed a few scale leaves and a tuft of needles, this spur shoot generally remains dormant. If, however, the terminal bud of a branch happens to die; these lateral shoots may grow into a normal branch, bearing at first isolated juvenile green leaves and then scale leaves with spur shoots in their axils (fig. 8). VARIATIONS IN THE NUMBER OF NEEDLES.—The number of needles on the spur shoot of each species is considered constant enough to be used as a character for classification; still, on rejuve- nated or infected twigs, shoots are found which bear an unusual number of needles. Boopie (1) makes the following statement: “In Pinus monophylla the spur shoots as a rule bear each a single needle, but two are occasionally present. Masters found by studying early stages that two leaf rudiments are always produced, but that one of them generally becomes arrested at an early stage.” Single needles have been observed by BooptE on P. Laricio, and we found some on twigs of P. maritima that were infected by the larva of a xylophage insect (Hylesinus piniperda). These single needles are roughly cylindrical; in many cases a groove is present on one side of the leaf. On following it downward, it is found to contain two papillae, one of which is the apex of the spur shoot, the other the rudiment of the second needle. Variation in the number of needles in this case is due to arrest in the develop- ment of one of them. It is a rare occurrence in P. maritima and P. Laricio, but it has become the rule in P. monophylla. The multiplication of needles on the spur shoot has often been recorded on wounded, infected (13), or vigorous (20) shoots, and it has been regarded as a reversion toward ancestral, many-leaved gymno- sperms (3). _ Although these variations are somatic in origin, we have proved that they comply with Mendel’s law, in that the proportion of bud mutations on the shoot is precisely that of F, recessives in the case of hybrids. Shoots of vigorous P. maritima or those infected with Coccus resinifians have been observed to yield 75 per cent normal 2-needled spur shoots, and 25 per cent 3-needled spur shoots. Proliferating spur shoots on P. virginica in Arcachon often yield 75 per cent normal 3-needled spur shoots, and 25 per cent abnormal 1918] DUFRENOY—PINE NEEDLES 445 2-needled spur shoots, or vice versa (14). These bud mutations, like proliferating spur shoots,’ are due to modifications in the normal nutrition of the pine, caused by environmental factors, traumatisms, and chiefly parasites, and they result in increase of Osmotic pressure in differentiating tissues. That osmotic pressure is the ultimate determining factor is demonstrated by the fact that we were able to force 7-year old P. maritima to produce 3-needled spur shoots or proliferating spur shoots by watering abundantly, which of course increased tur- Fic. 9 Fics. 9-10.—Fig. 9, protoxylem in first needle appearing on juvenile pine: w, cells whose walls begin to show lignification and stain red with phloroglucin; fig. 10, protoxylem in very young needle of juvenile pine: section near apex; only one xylem pole gescence of cells. Whether needles or juvenile leaves develop depends upon the relative supply of soluble osmotic material to the cells. Needles, or all adult organs in general, develop from material obtained from the earth and atmosphere, by gradual assimilation. Juvenile leaves, or all juvenile organs in general, develop from material stored in the reserve tissue (9). TRANSITION FROM JUVENILE LEAVES TO NEEDLES.—As needles and juvenile leaves are but different responses of the same organism to environmental factors, they may be assumed to show transitional ° Development of lateral long shoots is exaggerated on pine seedlings exposed to sea wind, and results in “buissornement,” like that recorded by Devaux (6) for Erica on ocean dunes 446 BOTANICAL GAZETTE [NOVEMBER forms (contrary to the opinion of DaGuILLoN 5). In fact, transi- tions are observed. The first needle to appear on juvenile pines shows anatomical features of juvenile leaves (figs. 9-17); and J] Fic. 12 Fic, 11 Fries. 11-13.—Fig. 11, transition from one vascular bundle to two semi-bundles in successive transverse sections from apex to base of needle: primary wood at apex (I) grouped into one bundle and like wood in juvenile leaves or in leaves of ancestral hypodermal cell present except below stomatal cells; letters as in fig. 13; fig. 13, St0- matal and epide cells of needle on adult pine: a, cutin, staining orange with Sudan IIT; /, lignin, staining red with phloroglucin; }, thickening of epidermal cells, staining green with cotton blue; c, thickening of hypodermal cells; /, hypoderm; e*, epidermis; s', stomatal cell (note local thinning of lignified wall, forming hinges). transitions from one vascular bundle (as shown in juvenile leaves) to two semi-bundles (as typical of needles) is observed in successive sections of needles from the apex downward (fig. 11).7 Needles _’ The anatomy of needles varies so much from apex to base as to make all com- parison worthless unless distance of the transverse section from the apex be clearly stated. The same statement applies to the age of the needle and the season when matérial is collected. 1918] DUFRENOY—PINE NEEDLES 447 CS C7 j f KUT Fic. 16 Fics. 14-17.—Fig. 14, periderm of very young juvenile leaf (collected April 25, 1918): ” ; p, phloem; y, cells of future wood beginning to lignify; g, procambium; rrows indicate direction of differentiation; wood first differentiates centripetally : : of ; sho j es; ft, foliar vascular bundles; proliferating spur shoot: bearing juvenile leaves: c', cauline vascular bu ] between, also resin canals Al ashy ae} et NP n x, xylem; p, phloem; fig. 16, course of vascular bundles in xylem inside (crossed lines), phloem outside, and cambium and medullary rays; f', foliar bundle consisting of 2 semi-b y defined medullary rays; fig. 17, part of foliar bundle (f*) in needle of proliferating spur shoot (collected May 3, 1918), part of ft of fig. 16: ¢, cambium active and pro- ducing spring wood above (x) and phloem (/) below; /*, wood vessels; m*, medullary rays; note that revegetation begins sooner in proliferating shoots than in normal spur shoots, which are still dormant at this season (cf. fig. 1). 448 BOTANICAL GAZETTE [NOVEMBER may also be derived from juvenile leaves through such transitional forms as bilobed juvenile leaves (figs. needles® (fig. 21). fossil. Fic. 18.—Proliferating logi . ally, shoot bearing double 18 tissues (bilobed) juvenile leaves. being rec- ognized from color reactions (see table I). Physiology Living cells must excrete poisonous materials which result from the disassimilation process (12). In the cells of pines these materials are resinous drops, which must be gotten rid of. In the primitive organs, resin probably filtered through the epidermis, and the epidermal cells were & This is still the also secreting cells. "IGS. 20-21.—Fi contains 2 semi-bu ndles note hyp (s) on Métal d side of protoxyle se concrescent cade ae been considered the homologue of the double 8 The —- of Sciadopit itys s shown by arrows in fig. , as in typical needles; fig. 21, concrescent or do s mal resin canals cite as are typical of juvenile leaves, also inverse WO 18-20) and concrescent Histology The chemical nature of the cell walls may afford good data for the comparison of pine needles with juvenile leaves or with phyllodes of other gymnosperms, living and Pine needles appear to be more differentiated histologically than morpho- Fic. 19.—Transverse sec tion of double hy S usually occur only in needles; dermal resin canal typical of Vike: leaf; ai deer poesincat obj. 3 Fic. at g. 20, schematic view of fig. 19: note that each vascular strand uble needle: bored may be homologized with bilobed juvenile leaves bent DUFRENOY—PINE NEEDLES 449 1918] *(Z) ,assnig ap anjq,, uleys asoqoed ‘H+ wmnissejod jo apruréd -O119} YIIM p2}va1] pure ‘19zVM paljlystp Ul paysea waqy ‘aplsoyyD snosay ul peureys suorIegt “UuoOIyNJos auUTAVIATZ puv joyosy} “UorNTos jousydozse'T 4 MOTIPA ef eeeceele nee ses eotesesers ee . ee Jo a8urio an{q é see SF Se . ee de ee ee eS ee ee Se et a ee ee ee ee or sdoip utsay aZuvio Neer aera i eels Pe Sara ae per Yat] qu Bty sae ees eee ee as hee are . juvenile leaf erie dle storage of reserves > cotyledonary needle (7) self-protecting > scale A needle which has specialized in the assimilation of carbon is itself a sort of assimilating organ; leaves of angiosperms are another. Needles are the physiological leaves of pines. They differ from leaves in that they are perennial and are much less fragile. Typical leaves are temporary, delicate, perfectly shaped for intense assimilation, but unable to stand bad weather. Pine needles last several seasons. They have efficient xerophytic adaptation and can stand the roughest weather on arid lands, windy mountain tops, or storm beaten coasts (15, 16, 17). In conclusion, thanks are extended to Professor FRon for his valuable encouragement, and to Dr. F. Latesque, Honorary President of the Station Biologique d’Arcachon, for his ‘many kindnesses and valuable documents.’ STATION BIOLOGIQUE D’ARCACHON LITERATURE CITED 1. Boopte, L. A., Solitary and concrescent leaves of Pinus. 14:19-26. 1915. 2. BouyGuEs, a, Contribution 4 l’étude du systéme libero-ligneux des Ceyptoghnits siancilaires. Act. Soc. Linn. Bordeaux 9: 1905. ° We are also indebted to The Bureau of Plant Industry, Professor R. T. Baker, Dr. Fracoso, and Professors onp and Marre for most valuable papers; while part of the expenses involved in these researches was defrayed by a grant from the Association Francaise pour l’Avancement des Sciences. New Phytol. 454 BOTANICAL GAZETTE [NOVEMBER 3. Burtincame, L. L., The origin and relationships of the Araucarians. Bor. GAz. 60:1—26, 89-114 . Coutter, J. M., Evolutionary tendency among Gymnosperms. Bor. Gaz. 48:81-97. 1909 . DAGUILLON, A., Bonhirches morphologiques sur les feuilles des Coniféres. Rev. Gén. Botanique 2: 1890. . Devaux H., Le buissonnement du Prunus spinosa au bord de la mer. Rev. Gén. Botanique 27:225. 1915. 7. Durrenoy, J., Contribution des feuilles voisines dans le développement de bourgeons qui normalement restent rudimentaires. Compt. Rend. Soc. Biol. January 9, IQ17 ‘ : cations srorluiles par le vent marin sur des spun es males de P. maritima. Compt. Rend. Soc. Biol. February 17, 1 9- , Les données et les problémes actuels de la shvioseenti bie Rev. Gén. Sci. May 30, 1917. , L’origine des espéces des Graminées. Rev. Gén. Sci. September 30, 1917. 11. ———, A new case of De doy variation in grasses and its significance. Jour. Wash. Acad. Sci. 7:535. 12. ———, La signification Set des essences et des pigments. Rev. Gén. Sci. October 31, 1917. , Sur les tumeurs du pin maritime. Compt. Rend. 166:355. 1918. 14. ———,, Les mutations gemmaires. Rev. Gén. Sci. June 15, 1918. 15. GAIN, E., Les régions florales. Nancy. 1908 16. LALESQUE, F., Monographie scientifique et médicale d’une ville de santé. Arcachon (in press). 17. Marre, R., La végétation des montagnes du Sud-Oranais. Bull. Soc. Hist. Nat. ‘de P Afrique du Nord. 27:210-292. pls. 4-16. 1914. 18. Reurous, L., Etudes sur les stomates. Thése Geis 1917; abstract in Rev. Gén. Sci. July 30, 1918. 19. TAKEDA, H., A theory of “transfusion tissue.” Ann. Botany 27:359-393- fe uu an 20. THomson, R. B., The spur shoot of the pines. Bot. Gaz. 57:362-385- pls. 20-23. 1914. 21. VAN TIEGHEM, Pu., Botanique 1:290; 2:236. Paris. 1916. 22. WEIR 7 e » and Houses E. E., Forest disease surveys. U.S. Dept. Agric. Bull. 6 BRIEFER ARTICLES JOSEPH YOUNG BERGEN (WITH PORTRAIT) There are many ways of advancing science, and hardly less signifi- _ cant than the investigator is he who makes men wish to investigate. Unquestionably no small number of those who have advanced botany have come to it with an inclination formed before university days, and he who set their compass was often one of those wise enthusiasts who guided their first steps in science. If we should take into account this service alone, American botany would acknowledge its debt to JosepH YouNG BERGEN, who died at his home in Cambridge, Massachusetts, on October 10, 1917. He was born February 22, 1851, at Rye Beach, Maine, his family moving in 1855 to Peoria, Illinois, where for some years the family home was beautifully situated on the bluffs outside the city. Here the nature- loving parents were accustomed to take their children on pleasant country trips to gather flowers, fruits, or nuts, according to the season. This home influence was strengthened for our future botanist by an intimate acquaintance with Dr. Stewarp, an old-time physician of Peoria, who took the lad, on many of his professional drives into the surrounding country. This amateur botanist watched the progress of growing things along the roadside, and new or especially interesting plants found their way into the doctor’s buggy for more careful inspec- tion at his leisure. ; Although the boy was prepared for college chiefly by home study, he had some time in the grammar and high schools in Peoria and two years in the old academy at Pembroke, New Hampshire. In due course he went to Antioch College in southern Ohio, that small but memorable institution whose first president was Horace Mann, of well known influence in the educational world. It is probable that at Antioch he received that bent toward geology which led to his first scientific work, done in connection with the Ohio State Geological Survey. Later he made practical application of his geological and chemical training in dealing with the problems of lead and zinc mining at Joplin, Missouri. 455] [Botanical Gazette, vol. 66 456 BOTANICAL GAZETTE [NOVEMBER In 1876 he married FanNy Dickerson, also of Antioch College, in collaboration with whom in 1890 he published “A Primer of Darwinism and Organic Evolution.” Mrs. BERGEN’s interests turned later to American folk lore, to which she has made a significant contribution. In 1878, not long after his marriage, Mr. BERGEN returned to New England and began his long career as a teacher by becoming prin- cipal of the high school at Deerfield, Massachusetts. Three years later he ac- cepted an appointment as professor of the physical and biological sciences at Lombard College, a posi- tion which he relinquished after 2 years. In 1887 he became teacher of phys- ics in the Boston Latin school. Physics as taught in the high schools of the time was more often an exercise in textbook study than one of application of principles to laboratory practice. Doubtless to one of Mr. BERGEN’S broad experience and keen perception of real values the lack of adequate pres- entation came home with unusual force. In 1891, fessor E. M. H ee in collaboration with Pro- rE. M. Harr of Harvard University, Mr. BERGEN brought out the well known textbook in high school physics which had a far-reaching and permanent influence on the teaching of this science in America. Although his chief interest was later transferred to botany, he main- tained an active connection with the teaching of physics by acting for 10 years as instructor in this branch in the Harvard summer school. eS ‘eines ne cng mica to the Boston English high school as a oe Acta e cme 12 years, during the remainder - Here again the need of a new presentation 1918] BRIEFER ARTICLES 457 of his subject for high school work led to the writing of his ‘‘ Elements of Botany”’ in 1896. The practicable way in which the main features of the newer botany with its greater emphasis on the physiological aspects of the subject were brought out in text instruction and directions for laboratory study went far to make the book an important influence in turning botanical instruction in secondary schools away from the rather dry descriptions of form, to the more interesting and equally valuable study of the activities of life. This book and its successors, the “Foundations of Botany” (1901) with keys to the commoner plants of the great divisions of the country prepared by Miss ALicE Eastwoop, Professor S. M. Tracy, and by himself, the “Principles of Botany” written in collaboration with Dr. BrapLey M. Davis in 1906, “ Essen- tials of Botany” (1908), “Practical Botany” in collaboration with Dr. Otts W. CALDWELL in 1911, and “Introduction to Botany” by the same authors in 1914, have provided a series of elementary texts which have kept abreast of the newer movements in botanical development and have served to induct a vast host of young Americans into the study of plants. . The success of these books brings sufficient evidence of a wise choice of material and of clearness and adequacy of presentation. While Mr. BERGEN is perhaps most widely known as a teacher and writer of books, he was also a genuine investigator. Both by early training and by inclination a man of out-of-doors, he found his instincts for the field leading him toward the problems of ecology, and his per- haps equally strong inclination toward the precision of the laboratory investigator led him when opportunity presented itself to a fruitful application of laboratory methods to the study of plants in their en- vironment. His opportunity came when in rgor he retired from teach- ing and went to southern Italy, where in the neighborhood of Naples he spent some 4 years. Here he made use of the rich facilities of the Biological Station and made the valued acquaintance of FEDERICO Detpino, Professor of Botany at Naples University, and of other members of the botanical faculty. He found great delight in tramping with Professor MATTEI, now of Palermo, who at that time was mapping the flora on the Solfatara, the partially active volcano near Pozzuoli. After a midday dinner at the Bergen residence they would “tramp off over that wonderful phlegrain plain, perhaps through a basaltic paved Greek lane, perhaps passing some wonderful ruined Greek temple or haunt of Horace or Virgil on their way out into the country.” The results of this happy time found their way to the botanical world chiefly through short articles printed in the Boranicat Gazette and in Plant 458 BOTANICAL GAZETTE [NOVEMBER World between 1903 and 1909. The transpiration problems of the xero- phytes of the Neapolitan region, drought tolerance, reactions to light, and the behavior of strand halophytes were among the subjects dealt with in at least a dozen articles. One article on his friend DELPINO (Science 21:996) recalls the personal relations of those days. Of course to one whose life had been given largely to teaching, peda- gogical matters would necessarily present their claim, and here Mr. BERGEN’Ss broad experience and sympathetic common sense always con- tributed genuine substance to the discussion. After his return to Cam- bridge from Italy, Mr. BERGEN’s time was for the most part spent on his series of textbooks. Although Mr. BERGEN took but little part in the work of scientific societies, the circle of botanists and zodlogists who in their Cambridge days found the Bergen home a place of sincere hospitality and of helpful appreciation and encouragement would of itself form a very respect- able society. The direct searching comment, the enthusiastic cheering- on, and the sympathetic and straightforward honesty met there were tonic and corrective and stimulant all in one. There are many of us who feel that we owe him a never-to-be-forgotten debt for these and for still more precious gifts. I am permitted to add an incident told immediately after Mr. BERGEN’s death by the gentleman to whom it happened. A few years ago a western botanist, visiting the Harvard Botanic Garden, noticed a , tall, spare man of distinguished appearance deeply absorbed in some observations he was making among the flower beds. The visitor asked one of the old gardeners near by if he could tell him the gentleman’s name. The old man replied “We call him Saint Joseph.” I believe that in every one of the wide circle of those who called Mr. BERGEN friend this incident will find an echo. In remembering him we value the botanist and the teacher, we respect the far-reaching penetration and creative work of the scientist, and we acknowledge and revere the rigor, the force and moral fervor, the patience and exceed- ing gentleness of the saint—Ropnry H. True, Bureau of Plant Industry, Washington, D.C. CURRENT LITERATURE NOTES FOR STUDENTS Development in gymnocarpous Agaricaceae.—In a recent paper Miss Dovetas' describes the development of 7 species of gymnocarpous Agaricaceae. She studied one species of Mycena (M. subalcalina), 3 of Hygrophorus (H. miniatus, H. nitidus, and H. borealis), and 3 of Entoloma (E. flavifolium, E. grayanum, and E. cuspidatum). The general course of development is alike in all the species, the variations presented relating to specific or generic features. The fundament of the fruit body just before the differentiation of the stipe and pileus primordia is cone-shaped, homogeneous in structure, the hyphae more or less interlaced and branched, extending in general parallel with the central axis of the cone. The surface is more or less floccose from the ends of single hyphae, or minute tufts, which diverge slightly. The young fundament of Entoloma cuspidatum differs from that of the others in being greatly elongated in proportion to its diameter, being nearly cylindrical, or even slightly clavate, with a conoid apex. The slender, elongate fundament appears to bear a direct relation to the slender form of the mature basidiocarp, and also to the very moist habitat of the species. The specimens studied were growing in sphag- num. The rapid elongation of the fundament serves to bring the growing points out of the watery environment in which they originate at the apex of slender rhizomorphs. The growing point for the formation of new tissue is apical, while elonga- tion occurs in the older hyphae. The first evidence of pileus formation is a great increase in the apical hyphae which begin to diverge, thus giving to the young fundament a sheaflike form. The pileus and stipe fundaments are thus differentiated. While apical growth of the basidiocarp continues, the most active seat of new tissue formation is now shifted from the apex to the annual furrow between pileus and stem primordia, and later to the under surface and extreme margin of the pileus. This marks the origin of the hymenophore. It begins at once in the 3 species of Hygrophorus, but is delayed for a short time after differentiation of the stipe and pileus fundaments in Mycena subalcalina and in the 3 species of Entoloma. It is recognized by the rich protoplasmic content of the hyphae, which usually react more strongly to stains, and thus become more deeply colored. The growth direction of these hyphae of the ‘Dovuctas, Gertrupe E., The development of some exogenous species of agarics. Amer. Jour. Bot. 5:36-54. pls. 1-7. 1018. 4590 460 BOTANICAL GAZETTE [NOVEMBER hymenophore primordium is perpendicular to the point of their origin, whether over the upper end of the stem, in the angle between stem and pileus, or on the undersurface of the pileus. Their course is parallel, although in the angle of the furrow there is more or less of a convergence in their growth direction. At first these hyphae are very slender and terete, but later they become stouter blunt. From the time of their origin they form a palisade layer whose surface is, in general, level until gill formation begins. In the majority of the species, the ends of the hyphae soon reach the same level. Their “register” is even, and the surface compact; but in Hygrophorus miniatus and H. nitidus the palisade for some time is not compact and the hyphae do not register evenly. The even register of the palisade hyphae is delayed in these species for some time after the origin of the gill salients During the early stages of aciooiiei of the hymenophore there is a strong epinastic growth of the pileus margin, causing it to curve downward and inward. This is particularly strong in most of the species, less so in Entoloma ais Didkoheork and less so in Hygrophorus nitidus. The gill salients are formed by the more rapid downward growth and extension of the subjacent tissue in regularly spaced radial areas. The development advances in a peripheral direction from the stem toward the margin of the pileus.. The growth direction is perpendicular to the morphological undersurface of the pileus, and the situa- tion from this oo can readily be understood when the pileus margin is wae incurv 4 cables layer, which eventually becomes the hymenium, the ele- in are multiplied by branching of the subhymenial elements. In the species of Hygrophorus in particular, and to some extent also in Entoloma ninhsine ei the pressure of the i i palisade loosens up the elements of a zone of less density. This peculiarity is well shown also i in Siphaks chrysophylla and Clitocybe cerussata studied by BLIZzZARD.? In both these papers dealing with gymnocarpous forms, it is shown that the origin and the general course of development of the hymenophore corre- sponds with that of angiocarpous forms of the Agaricus type. It is further shown that there is a tendency in the early stages of development for a supet- ficial zone of the pileus, here of quite limited extent, to be arrested in growth, sometimes quite regularly and normally. The regular course of development being thus shifted to a slightly interior zone presages the later evolutionary _ type of development presented by the angiocarpous forms, where the origin and differentiation of stipe and pileus primordia are shifted permanently to the interior of the young basidiocarp primordium, with a more or less well marked external zone, the blematogen.—Gro. F. ATKINSON * BuizzarpD, A. W., The development of some species of agarics. Amer. Joe Bot. 4:221-240. pls. 6-11. 1917. 1918] CURRENT LITERATURE 461 Self-sterility——East and Park} have recently published the results of some extensive experiments on 4 self-sterile species of Nicotiana, and have pro- posed an explanation. In the past there have been several attempts to inter- pret self-sterility as a response to environmental factors, notably humidity. Such interpretations may have been quite true in some cases of self-sterility, where only a single race of plants has been involved, but highly unsatisfactory in explaining cases where pollen fails on own stigmas and functions on stigmas of another race. It is such a situation that the authors have dealt with, and they have shown conclusively for their material that self-sterility is inherited. Normal seasonal changes at times induced “pseudo self-fertility ” in their self- sterile races, but ‘other a fatiors appeared to have little or no influence on self-fertilit As to the shyeloicated nature of self-sterility, the authors state that it is involved with rate of pollen tube growth. This in itself suggests that self- sterility behaves as a sporophytic character. The fact is more definitely demonstrated, however, “by the behavior of reciprocal matings, pairs of _ Teciprocals always giving like results either when fertile or sterile.” Going further, the authors state “that modern discoveries tend more and more to show that the sole function of the gametophyte of the angiosperms is to produce sporophytes. The characters which they possess appear to be wholly sporppny ts, ithe aceenae which they carry functioning only after fer- tilization.” Thi ms directed at such theories as that of BELLING, who has given us a striking scislaiaiion of “semi-sterility” in beans, on the basis of the direct influence of the germinal equipment of gametophytes upon the gametophytes themselves. It is quite probable, however, that the two cases are involved with distinctly different phenomena, since BELLING’s material showed degeneration and sometimes complete abortion in pollen and embryo sacs, while the Nicotianas of East and Park were self-sterile merely because of the failure of pollen tubes. The hereditary mechanism of the two cases must be quite different. To explain the hereditary behavior of their Nicotianas, the authors have assumed a mechanism involving multiple allelomorphs and crossing over. If two plants differ in but one of a number of effective factors, they are fertile in intercrosses. “Intrasterile classes” are composed of individuals which differ in none of the effective factors. Anything like a thorough appreciation of this theory can be obtained only from the original article. This explanation seems sufficiently accurate in interpreting the results of the authors, as well as the results of some of the earlier investigators. From a practical point of view, however, it seems rather unsatisfactory, since it con- siders only the behavior of self-sterile plants when bred inter se. The authors 3 East, E. M., and Park, J. B., Studies on self-sterility. I. The bebavior of self- sterile plants. Genetics 2:525-609. 1917 402 BOTANICAL GAZETTE [NOVEMBER state that ‘‘all questions connected with the relation between true self-fertility and self-sterility have been omitted designedly as pertaining to a distinct prob- lem.” Are we unreasonable in asking for a single theory to explain both self- fertility and self-sterility ? Are we wrong in thinking that the significance of self-sterility lies in its relation to self-fertility? Such a general theory, no doubt, will be provided by the authors in their later reports; the present pub- lication evidently represents merely the first of a series on the general subject of self-sterility he explanation also has another theoretical shortcoming, similar to that. which applied to East’s “heterozygosis.” In heterozygosis East stated that hybrids are vigorous because of their heterozygous sets. This virtually amounted to saying that hybrids are vigorous because they are hybrids. “Heterozygosis”’ was a more accurate and scientific statement of the fact of hybrid vigor, but it was not an explanation. Now East states that pollen will not function on stigmas of a plant of which the germinal constitution is the same as that of the plant which produced the pollen. Couched ina terminology involving multiple allelomorphs and crossing over, this may well be a more accurate and scientific statement of the facts of self-sterility and its behavior in inheritance, but it is not an explanation. Such scientific restatements are very valuable in helping to organize facts, and “‘ heterozygosis”’ unquestionably had such a value. The present theory, however, seems at first sight a muc less valuable one, since it is so elastic as to be confusing. But whether the theoretical argument of the authors is destined to stand or fall, they have done an exemplary piece of research. This seems to have COULTER. Buffer processes in succulents.—JENNY HEMPEL‘ has made a very impor- tant addition to our rather limited knowledge of actual reaction in plants. Succulents were used in this work, since, with their well known wide and rapid variations in acid content, they might be expected to supply especially interesting material for such astudy. Determinations by the use of the hydrogen electrode were made on the juices of numerous specimens of the plants studied, after they had been exposed to varying conditions. The values found range from Pu=3.9to Py=5.7. Higher acidity than the more acid of these values is re- corded in the same work in lemon juice (Py=2.19); and by Haas’ in citrus fruits (Py=2.22-3.8), in cranberries (Py=2.4), and by a less exact method in the petals of certain flowers (Py=about 3). ‘HempeL, Jenny, Buffer processes in the metabolism of succulent plants.. Compt. Rend. Trav. Lab. Carlsberg 13:1-129. 1917. ’ Haas, A. R. The reaction of plant protoplasm. Bor. Gaz. 63:232-235- 1917- ——,, The acidity of plant cells as shown by natural indicators. Jour. Biol- Chem. 27:233-241. 1916. 1918] CURRENT LITERATURE 463 It would seem that such marked changes as were found in the reaction of the juices of active tissue must affect considerably the metabolic processes, as well as the physical condition of the tissue. CROCKER’ has suggested that these changes may be important in the regulation of transpiration by succu- lents. The lower Values are of the same order as those reported in the same work by HEMPEL and also quoted from WacGNneER for non-succulent plants. Such P, values range from 5.4 to somewhat above 6. Slightly alkaline juices are reported by Haas (/oc cit.) in the petals of certain flowers; he finds, however, that blue pigments by no means always indicate an alkaline reaction. As the title suggests, the principal object of the work was to gain some information as to the substances in the plant juices which act as buffers, or regulators of their reaction. On the acid side of the neutral point the following data were obtained for this study: (1) titration to the litmus end point (Pg=6.8) compared with the original Py value; (2) qualitative tests to determine the organic and inorganic acid radicals present; (3) ash analyses to determine the total base present; (4) studies of the reaction and titration values of malic acid salts, and such mixtures of them as appear likely to occur in the plant. The data are most complete for the juices of Rochea falcata, Cotyledon obvallata, and C. linguaefolia. The conclusion is reached that in these plants, and prob- ably in all succulents, the concentration of hydrogen ions is determined by the relation between the quantities of acid and normal malate present. On the alkaline side of the litmus end point the data may be grouped as follows: (x) titration from the litmus end point to that of phenolphthalein (Pu= phone 9.2); @) determination of nitrogen and in some cases phosphorus; (3 ) titr ith alumin m malate; (4) titratio with wn and verable cous ea ak at the Sere OER point. It is concluded that aluminum malate and the unknown substances men- tioned are the principal buffers in this region. The nitrogen and phosphorous compounds have very little effect. The Hivatioti to the phe erRT end point is admitted to be very unreliable. It seems unfortunate that as con- siderable quantities of the juice were available ~ slectromerc method of titration was not used. Such results would have co uch to the com- pleteness and accuracy of the data—THomas G. PHILLIPs. Mutationists and selectionists.—JENNINGS® has attempted to reconcile the views of the “mutationists” and the “selectionists.”” The latter, headed by Caste, have claimed that selection can modify unit characters, and presented striking evidence on the point. The mutationists have then demon- strated that these data may also be interpreted by assuming that there is but 7 Crocker, Wa., Rev. Bor. Gaz. 64:526-527- * JENNIN se > Modifying factors and oul allelomorphs in relation to the results 5 we Amer. Nat. 51: 301-306. 1 Sok chanel in hereditary Tra hea in relation to evolution. Jour. Wash. Acad: Sei 7:281-301. 1917. 464 BOTANICAL GAZETTE [NOVEMBER one basic invariable unit determining the presence or absence of a character, plus numerous modifying factors; the number of the latter present in a given case determines the degree of expression of the character. The author admits that such modifying factors have been demonstrated in Drosophila, but goes on to show how “the objections raised by the mutationists to gradual change through selection are breaking down as a result of the thoroughness of the mutationists’ own studies.” For in Drosophila there have gradually been dis- covered not only 7 modifying factors for eye color, located on different regions of chromatin from the basic factor for eye color, but also 7 grades of the basic factor itself, that is, different conditions of the same unit. “What more does the selectionist want ? Is not the controversy at an end?” ere still remains, however, a fundamental difference between the two views. The selectionists claim that these changes (in unit characters) are continuous, and in a definite direction determined by the standard of selection. The mutationists, on the contrary, claim that these changes occur in distinct steps (mutations), and do not occur in any definite order or direction as the result of selection. JENNINGS takes exception to this last claim of the muta- tionists, and presents some of his work on protozoa, to show the effectiveness of selection in a series of asiaeorene generations. There is much to be d din sucha iliation between the two schools, but more evidence must come in before there can be much hope of bringing it about. At present the views of the mutationists seem to be in better favor, chiefly because they give a much more definite basis for description of the phenomena of inheritance. “If one creates a hypothetical unit by which to describe phenomena and this unit varies, he really has no basis for description (East).”—MERLE C. CouLter. Narcotic plants and stimulants.—Sarrorp® has published a very instruc- tive account of plants used by the “ancient Americans” as sources of narcotics and stimulants long before the discovery of America. He indicates 13 such plants as chiefly in use, among them Nicotiana, Datura Stramonium (a source of atropine), Erythroxylon Coca (a source of cocaine). Other plants of minor importance are also noted. In concluding the summary, the following state- ment is made: ‘In view of the shortage of medicinal alkaloids resulting from the present war, it is suggested that investigations be made to determine the nature of the properties of these less-known narcotics, with a view to their utilization as substitutes for others now recognized in the standard phar- macopoeias.””— < 9 Sine FFORD, W. E., Narcotic plants rag = of the ancient Americans. Smithson. Rep. 1916. pp. 387-424. pls. 1 VOLUME LXVI NUMBER 6 THe BOTANICAL GSAZETTE DECEMBER 1918 LIMITING FACTORS IN RELATION TO SPECIFIC RANGES OF TOLERANCE OF FOREST TREES A. H. HUTCHINSON (WITH SEVEN FIGURES) The conclusions recorded in this paper are drawn from a study of the forests throughout the Province of Ontario, particularly along the shores of Lake Ontario, Lake Simcoe, the Kawartha Lakes, and Rideau Lakes; in Algonquin Park and in Mattagami and Timagami Forest Reserves. Observations, with notes, have been made during more than 6000 miles of travel by canoe and over- land through the forest country of northern Ontario, especially along the streams and lakes forming the headwaters of the Muskoka, Maganatawan, Petewawa, and Madawaska rivers; also of the Montreal, Sturgeon, Wanapitei, Vermilion, Mattagami, and Abitibi rivers. While the greater part of the discussion has particular reference to Ontario, the conclusions are made in the light of some personal knowledge of the forests southward to the Gulf of Mexico and westward to the Pacific. The data regarding the limits of forest trees recorded in the accompanying maps have been obtained principally from accounts of the explorations of BELL (2), Macoun (19, 20), and Low (18). The records of isotherms and precipitation areas have been copied from the Geological Atlas of Canada, 1915. So far as the writer 465 466 BOTANICAL GAZETTE [DECEMBER has been able to observe, the records of the explorers mentioned have been even more accurate than has generally been conceded. Although the specific limits of forest species have been rather definitely outlined, there seems to be no agreement regarding the part played in determining these limits by the various factors affect- ing forest growth. In this paper an attempt has been made to relate the limiting factors to the specific range of tolerance of forest trees, and in this way to account for the respective distributions of some of the species dominating the forests of Ontario. SCHIMPER (21), as a result of his extensive studies in plant geog- raphy, concludes that ‘‘the differentiation of the earth’s vegetation is thus controlled by 3 factors: heat, atmospheric precipitation (including winds), soil. Heat determines the flora, climatic humid- ity the vegetation; the soil as a rule merely picks out and blends the materials supplied by these two climatic factors, and on its own account adds a few details.” Investigators have mentioned many factors which affect the composition of forests. Drawing his conclusions from the explora- tion of Labrador, Low (18) says “the distribution of forest areas and the range of the various trees depend upon several factors, among which may be mentioned position as regards latitude, height above the sea coast, and the character of the soil.” BOwMAN (3 in the light of his physiographic studies says as follows: The distribution of forests is controlled largely by rainfall, although the distribution of species within each region is also controlled by insolation, temperature, wind velocity, water supply, and geographic relation to post- glacial centers of distribution. When more detailed statements are attempted many difficulties are encountered in the form of apparent inconsistencies. . Some species appear to find their appropriate conditions in different latitudes by a change in their habitat; for example, the larch, balsam fir, and white birch which in the north grow freely on dry or hilly ground, toward the southern limits seek the cold ground in swamps. The white cedar and white pine in some places manifest the same tendency. FROTHINGHAM (12) in his report on hardwood forests sums UP the situation as follows: “How moisture and temperature affect the different species in the complexity of forest environmen is still so little known that no positive information can be given.” 1918] HUTCHINSON—FOREST TREES 467 Temperature factor The northern limits of many tree species are undoubtedly the result of low temperatures. WARMING (24) states, “It is clear that conditions as regards heat determine the boundaries of the dis- tribution of species on the earth.” The effect of temperature is emphasized by the fact that “the appropriate temperature for the growth of a number of species, such as Picea and Abies, is carried far to the south of their normal latitudes along the elevated parts of the continent, especially the Alleghanies and Rocky Mountains” (BELL 2). In such regions the tree species are in most cases identical with those found farther north. However, it is more diffi- cult to account for the southern limits of trees on a basis of minimum temperature. Bray (5) finds difficulty in explaining the occurrence of boreal (Picea, Abies) associations in the bogs of regions domi- nantly austral. ‘‘The question arises as to whether the factor of temperature plays a réle in the occurrence of these bogs,” again, “the extremely irregular boundary between the boreal co- nifer forests and the temperate hardwood forests of New Eng- land, for example, can hardly be explained by temperature alone”’ (itkeees 14). The lines representing the limits of Picea nigra, Larix americana, and Betula papyrifera follow yearly isotherms very closely from the mouth of the Mackenzie River across the continent until they reach the coast of Labrador, where they swing southward, here following a course almost parallel with the coast line. There is reason to believe that temperature is the limiting factor throughout a great area, while a second factor is active along the Labrador coast. From the fact that the same order in the limitation of these species is retained, even in the Labrador region, it would seem that the limiting factors are similar throughout. Excessive loss of heat energy due to the air currents so prevalent in this region has the same effect as the loss of heat energy due to excessively low tempera- tures. Similarly in southern Ontario, where latitude and lake influence together result in a region of a relatively high yearly temperature average, the limits of trees such as Juglans nigra and Castanea dentata are parallel with isotherms. Here also the evi- dence would indicate that temperature is the limiting factor with 468 BOTANICAL GAZETTE [DECEMBER respect to such species. The general conclusion that temperature is usually, if not universally, the determinant of northern limits has resulted from making general statements based upon selected and favorable instances which are specific rather than general. There is abundant evidence that while temperature acts as a limiting factor in many instances, it is by no means the only factor controlling even the northern limits of tree species. This is amply demonstrated by the data recorded on the accompanying map (fig. 1). Many of the lines indicating the northern limits of tree species intersect; this cannot be accounted for on a temperature basis. Isotherms do not intersect nor do lines indicating the length of the growing season. The northern limit of Pinus Banksi- ana at 100° W. long. traverses a region the yearly isotherm of which is 25° F.; at 80° W. long. the isotherm which the northern limit traverses is 32.5° F.; at 75° W. long. it is 20° F.; and at 7o° W. long. it reaches the 32.5° F. isotherm. The isotherms correspond- ing to the northern limits of Ulmus americana at various regions are at 100° W. long. 27.5° F.; at 95° W. long. 32.5° F.; at 80° W. long. 30 F.; at 75° W. long. 40° F.;"and at 70° W. long. 32.5° F., a remarkable range of variation. The looping of the lines repre- senting the limits of such species as Picea canadensis, Populus bal- samifera, and Populus tremuloides, as shown in the Labrador region (northern Quebec), is significant, particularly in the case of Picea canadensis, in contrast with the closely related Picea mariana. The northward deviation of the limits for Betula lutea, Acer sac- charum, Tsuga canadensis, and Quercus rubra at 80° W. long., 4 point where the isotherm swings southward, cannot be explained on a temperature basis. The anomalous tree distribution in the Saugenay region is another case in point. Any idea of the possi- bility of explaining the western limits on a temperature basis has long been discarded. It is evident that in the instances men- tioned something other than temperature must be the limiting factor. Water factor Water as a factor in the determination of tree distribution has received considerable recognition. Cowzes (8) says, “On the whole there has been a general tendency to overestimate the influence of NM/D Fic. 1.—Northern limits of tree species represented by lines having identification numbers and marks: 0, original; r, revised; isobars outlined by connected dots and average yearly temperature indicated in degrees; 1, Picea mariana; 2, P. canadensis; 3, Larix americana; 4, Populus balsamifera; 5, Betula papyrifera; 6, Populus tremu- loides 7, Pinus Banksiana; 8, P. Banksiana (other data and outliers); 9, Abies canadensis; 10, Thuja occidentalis; 1oA, Ulmus americana; 11, Pinus Strebus; 12, Betula lutea; 13, Acer Saccharum; 14, Quercus rubra; 15, Tsuga canadensis; 16, Fagus americana; 17, Juglans cinerea; 18, Carya amara; 19, Juglans nigra; 20, Castanca dentata. 1918] HUTCHINSON—FOREST TREES , 469 temperature as an ecological factor. The trend of nearly all experi- ment has been to show that water is of vastly greater importance.” BELL (2) states, “A great difference in the moisture of the air of two regions otherwise resembling each other in climatic conditions has also a powerful effect upon the growth of forests; and the dry- ness of the air in the western prairie and arid regions is, no doubt, the chief cause of the absence of timber.’’ There has been much recent research concerned with the water relation of plants (9). The greater number of investigators have selected the region of the great plains, a region where water is domi- nantly the limiting factor, as their field of investigation. TRAN- SEAU (22, 23), LivincsTon (16, 17), and FULLER (13) have shown that the water factor may be regarded as a complex depending primarily upon the amount of soil water available for the plant and the rate of evaporation. The amount of available soil water is dependent upon precipitation during the growing season and the physical properties of the soil, while evaporation depends chiefly upon the humidity, air currents, and temperature of the atmos- phere. Briccs and SHANTz (4) in their work on the wilting co- efficient have emphasized the specificity of tolerance in plants with respect to minimum soil water. The valuable experimental data recorded in these papers demonstrate that water frequently acts as a limiting factor; the converse, that there is a considerable range for any given species wherein water does not factor in a limiting capacity, is less frequently emphasized. It is significant that the limits of tree species such as Picea mariana, P. canadensis, Larix americana, Populus balsamifera, Abies canadensis, and Betula papyrifera, which extend north of the arid plains, are not deflected southward’ in the Manitoba-Minne- sota region, while almost invariably the more southern species are deflected when they come in contact with the region of diminished precipitation. The evidence would indicate that the water factor limits the westward extension of such species as Acer saccharum, Tsuga canadensis, Fagus americana, Thuja occidentalis, and Ulmus americana. It may be noted, also, that when this deflection takes place the order in which the tree limits occur is changed in many instances. 470 BOTANICAL GAZETTE (DECEMBER Thuja occidentalis, which extends northward far beyond Acer saccharum, does not reach the western limits of the latter, while Quercus rubra extends westward beyond the limits of many species which are to be found beyond its northern limit. Such phenomena doubtless are the result of the fact that different limiting factors dominate the several regions forming the boundaries of distribu- tion. In southeastern Canada it is difficult to find any parallelism between the limits of tree species and the boundaries of precipita- tion areas. For instance, Juglans cinerea extends throughout areas whose respective yearly precipitations are 30-35 in., 35-40 in., 30-35 in., 40-45 in., and 45-50 in., with no apparent hctiencbiche of the lines bounding its growth area. The limits of Fagus americana pass through similar areas, with the addition of an area where the minimum yearly precipitation is 25 in. This range is of par- ticular significance in consideration of the fact that Fagus is gener- ally regarded as mesophytic. The absence of a parallelism between precipitation and forest limits in the Ontario section is so obvious from a consideration of the accompanying map (fig. 2) that further emphasis would be superfluous. Other conditions being favorable, there is sufficient rainfall for forest growth; or, in other words, precipitation does not enter here as a limiting factor. WARMING (24) has based his ecological classification of plants upon the premise that ‘‘the most potent and decisive factor is the amount of water in the soil.” ‘‘ The soil upon which the coniferous forest occurs varies widely, yet so far as reliable information is available it is always physically or physiologically dry.” “The cold winter is a physiologically dry season against which trees can protect themselves by defoliation or by xerophytic structure.” This may account for the xerophytic structure of the coniferous leaf; it is difficult to understand what bearing it has in connection with the contention that conifers as a class grow in dry soil condi- tions, the winter, even for conifers, being a period of comparative dormancy. Deciduous trees are protected against the ‘physio- logical dryness” of winter by leaf fall; conifers by having leaves of xerophytic structure. Some deciduous trees, for example Quercus, are comparatively xerophytic, while others are decidedly meso- “ ff Jags erat — ‘ —_ = wm iO. 109. Say , f A ~ titre ian Aa ¢ a + ea ioe, ee f nite esonacr Saar el xs ae, g oe ~ : : oe Ss er oe ‘4 ae ce es KS wee ” ey 2 9S : \ r Ay dis re ons e Ke; RAS + FS arts SA” Oy “al vw; " Sd ma | \ oy -"F mei ion 4 yt 2 D G aa ENTS. I C eeey 2 tes das). See 5 RT SE ih wo Ls Seeds, Was. : As > Jf Ke ei LW { if we / / fs x. - 1 HS Lal 4 15 { Q AN \\ it - ae EGRET «* > Me AS Fic. 2.—Similar to fig. 1 except that a precipitation rather than a temperature map is superimposed upon the chart of tree limits; numbers indicate precipitat 1918] HUTCHINSON—FOREST TREES 471 phytic, as Fagus and Acer. It seems possible that there might be a similar range amongst conifers. Zon (27) states, “Balsam fir attains its best growth and largest size on flats the soil of which is usually a moist, deep sand-loam.”” An abundance of available soil water is not the factor which so often excludes Abies balsamea from such soils, particularly in the more temperate regions. During the summer of 1914 a series of experiments were con- ducted in Algonquin Park to discover the relation of seedling growth to atmospheric humidity. Atmometers of the Lrvincston design were set up at a number of stations, including those where seedlings of Acer saccharum, Abies balsamea, and Picea mariana were abundant. The readings for the months of July, August, and part of September proved that in this region there is no appreciable difference in the rates of evaporation at the stations mentioned, and that in each case the humidity was in excess of that which FULLER (13) regards as characteristic of a mesophytic forest. Moreover, Acer grows on the more exposed ridges; Abies and Picea on the less exposed lowlands or slopes. Experiment has shown that such conditions hold generally for the “lake country,” where in many cases one-tenth of the total area is covered by water, and the greatest distance of any point from bodies of water seldom exceeds 2 miles. Three of the limiting factors most frequently emphasized, temperature, atmospheric humidity, and precipitation, are eliminated, as such, under conditions prevailing, and still there is a marked segregation of forest associations. Soil factor The problem regarding the extent to which soil cpmposition may act as a limiting factor in the determination of forest dis- tribution has been variously answered. FROTHINGHAM (12) states “The soils of the northern hardwood forest are as a rule loamy sands, the results of the decay of granite, quartzites, and siliceous gneisses, also the water assorted loams and clays or the unassorted morainal tills, rich in clay; but they also thrive on light sandy soil in localities subject to moist winds.” In connection with the forests of Michigan, BEAL and WHEELER (1) state that “The best wheat lands are usually found on uplands near 472 BOTANICAL GAZETTE [DECEMBER streams, where the oak timber gradually shades into beech and maples.” On the other hand, “evergreen’trees, whether conifer- ous or broad-leaved, seem to be just as characteristic of poor soil as of any particular kind of climate” (14). Bowman (3) draws attention to the limitations of soil composition as a determining factor. Cow es has shown that the composition of the rock from which any soil may be derived seldom acts in a limiting capacity with respect to the species which that soil may support. It is only in exceptional cases that a soil, newly weathered, is defi- cient in the mineral constituents necessary for plant growth. This generalization is particularly applicable in Ontario, where the soil, whether it be glacial drift toward the south, or the weathered deposits and exposed rocks farther north, is derived from the dominantly granitic rock of the Laurentian Plateau. The original composition of the soil is seldom a limiting factor, at least in so far as the forests of Ontario are concerned. Humus factor It is scarcely necessary to emphasize the importance of the humus content of the soil as an ecological factor; its significance as a limiting factor with respect to the forests of Ontario is our chief concern. In forest regions the humus content of the soil increases the water retaining capacity; increases the porosity, and hence the aeration of the soil. Mineral salts are retained by the adsorp- tive properties of humus, and incidentally, conditions are made more favorable for soil bacteria, which are essential for the growth of such species as Fagus. Cows (8) states, “Although bare sand supports a.xerophytic flora, the accumulation of a thin humus layer is sufficient for forest development, and the Michigan dunes show that the most mesophytic of our forests can grow on a sand dune if there is present a-humus layer a few centimeters in thickness.” In the Algonquin Park region to which reference has been made, the Acer or Acer-Fagus forest occupies the ridges, while the Abies- Picea forest occupies the lower slopes and lowlands. On the slopes where the exposure of the rocks, due to drainage of glacial lakes, has been comparatively recent, only a small amount of rock soil has accumulated; this is covered by a humus layer, but the two are not intimately intermingled by weathering processes. The 1918] HUTCHINSON—FOREST TREES 473 humus content of the soil proper is low. In the lowlands a similar condition maintains. This is especially applicable to the peat bog, where humus is most abundant. There has been no movement, however, of the particles of the contiguous strata of the rock soil and the overlying humus; they are distinct, hence the otherwise valuable humus is practically useless in so far as the improvement of soil properties is concerned. On the ridges which were exposed first by the subsidence of glacial ice and water there is much deeper soil, and the humus: accumulated from antecedent vegetation has become intimately associated with the rock soil by weathering. The soil proper has a high humus content and is able to support such trees as Acer and Fagus. It will be remembered that the - temperature, precipitation, humidity, and original soil composi- tion may be regarded as constant; the varying factors are those associated with the accumulation of humus. The humus content of the soil is at least a local factor in the determination of tree distribution. : It has been maintained that differences in the composition of soil have only a local effect. It does not seem clear why a factor which is potent locally should not be potent throughout greater areas: The gradients of soil changes are usually greater when limited areas are considered; hence also those of the associated floral changes. The Laurentian Plateau is a great area dominated by the coniferous forest, while the contiguous region of glacial drift is dominated by the deciduous hardwood forest. The marked differences in forest species prevailing in the regions north and south of the Kawartha Lakes, respectively, is strikingly in accord with soil differences. Moreover, the line separating the dominantly coniferous region from the dominantly deciduous hardwood region does not follow any isotherm or the boundaries of any precipitation area, but rather the outlines of the Laurentian Plateau, roughly from the southeast part of Georgian Bay to Lake Simcoe along the Kawartha Lakes, southeastward to the Thousand Islands, northward to Ottawa, and again eastward along the northern limits of the Ottawa and St. Lawrence valleys. In the coniferous region there are oases of deciduous hardwoods of considerable area, such as that at Renfrew, or of limited area, such as the ridges already mentioned; in fact, wherever the soil is similar to that found in the 474 BOTANICAL GAZETTE [DECEMBER characteristically deciduous hardwood area. It is true that these broad outlines have been obscured in many places by large tracts being covered with pioneer forms, such as Populus and Betula, as a result of ‘‘burns”’ (10, 11, 15). These regional forest limitations cannot be explained except upon some basis of soil differences, such as have been described as determining the local limitations of the forest types of Algonquin Park. The evidence clearly indicates that the slowly weathering rock of the Laurentian Plateau has been a barrier against’ migration of the hardwood forest, which, however, has been able to establish outposts where favorable soil conditions have been found. In brief, the development of a soil, particularly with reference to its humus content, may act as a limiting factor regionally as well as locally. Light factor It is generally accepted that seedlings of some tree species grow only where there is abundance of light, while others grow best under shade conditions (26). FROTHINGHAM (12) has classified the trees of the northern hardwood forest upon the basis of light tolerance. The seedlings of pioneer species are necessarily light tolerant in contrast with those species forming the climax forest, which are shade tolerant; seedlings of Pinus Banksiana and P. Sirobus thrive only in direct sunlight, which is also the case with seedlings of Abies balsamea and Picea canadensis, although to a less marked degree. On the other hand, the seedlings of Acer and Fagus grow best in the dense shade of mature trees; Tsuga canadensis is an example of a conifer which is similar in this respect. Because of the specificity of the range of tree species with respect to intensity of light, certain forms cannot be pioneers, while others are eliminated from forests which have been well established, except where destructive agencies such as cause windfalls and erosion are at work. To this extent the intensity of light may act as a limiting factor in tree distribution. Time factor The time factor deserves a most important place in any Con- sideration of the distribution of forest trees, and it is of particular _ significance in connection with the forests of Ontario. Time as 4 1918] _ HUTCHINSON—FOREST TREES 475 factor in limiting the distribution of forest species is an expression of the rate of change in ecological conditions and of the specific _ rates of migration of the various species. Ordinarily conditions change so slowly that migration keeps pace; when there are more rapid changes migration lags behind. The time factor, therefore, must be considered in relation to the rate of change of such condi- tions as temperature, water, soil, light intensity, and secondarily in relation to methods of distribution. WARMING (24) states, ‘‘Changes in the physical relationship of the soil are everywhere and always taking place, and in close corre- lation with this plant communities also undergo modification, but it does not seem possible to use development as the fundamental basis of classification of plant societies.” (CowzrEs (7), while recognizing the same factors, attaches more importance to develop- ment of successional associations. The forests of Ontario have been made possible only by the retreat of glacial ice and water and the establishment of condi- tions permitting the growth of trees. It is evident that modifica- tions of temperature have been prerequisite for the northward migration of tree species; by many it is regarded as the only factor. ADAMS suggests that the northern migration following the retreat- ing glacier would comprise 3 great waves of life: (1) a wave of glacial or arctic vegetation, of which there are remnants in New York and Mount Marcy and two or three other high peaks; (2) a wave comprising the northernmost species of trees, stunted willows, birches, alders, and the coniferous forest spruces, hemlock, and pines; (3) a wave embracing the temperate zone deciduous trees. HARSHBERGER records a similar conclusion: ‘“‘Several great waves of plant migration may be recognized, namely, glacial vegetation, tundra coniferous forests, and a migration of the deciduous forest elements from the southeastern center.’ If forest migration has kept pace with temperature changes, it might be expected that the limits of forest species would conform in outline with respec- tive isotherms. It has already been demonstrated that in many instances this is not the case. The conclusion that in many places — the migration of such species as Tsuga canadensis, Acer saccharum, and Fagus americana has lagged behind temperature changes is 476 BOTANICAL GAZETTE [DECEMBER made necessary. To a greater or less extent this is true of all the species forming the forests of Ontario. Under existing tempera- tures any further migration is dependent upon changes in the . conditions now acting as limiting factors, as water, soil, and light. The rate of migration, and hence the distribution of forest trees, is dependent, primarily, upon the rate of change in temperature; however, migration may be restricted by other factors. There is reason to believe that the yearly precipitation has gradually decreased since the glacial epoch. The data regarding the exact extent of these changes are limited. There can be no doubt, however, that the westward migration made possible by temperature changes has been checked by the water factor; also the irregularity of the limits of Pinus Banksiana may be explained by the fact that although temperature conditions have so changed that this species has migrated to 56° N. lat. in the highlands of northern Quebec, it has been limited in its northward progress by the low lying lands south and westward from James Bay. The incon- sistencies of data regarding the northward distribution of Pinus Banksiana are doubtless due to the presence of certain outliers which might be expected when available soil moisture and other » soil conditions act as the limiting factor, but which would be most improbable were temperature the determining factor of distribu- tion. In regions where water is a limiting factor the rate of migration is dependent upon the rate of change in water condi- tions, in other words, upon the time factor. Time factor in relation to soil development It has been demonstrated that soil development, particularly with reference to the humus content, is a potent factor in deter- mining the boundaries separating the Acer-Fagus and the Abies- Picea forests of Ontario. Since the Acer-Fagus forest demands the most highly developed soil, we are forced to the conclusion that in a forest succession the deciduous hardwood forest is the climax type. Over a vast area this climax type of forest has been excluded by soil conditions rather than by temperature. North- ward migration of the deciduous hardwood forest has been limited by the rate of soil changes rather than by the rate of temperature 1918] HUTCHINSON—FOREST TREES 477 changes; with respect to the latter migration has lagged behind. Upon such a basis the “anomalous” separation of the deciduous hardwood forest and the coniferous forest is readily explained. The granitic rock of the Laurentian Plateau has weathered slowly, humus has accumulated slowly; in brief, the soil has developed slowly, hence the migration of the climax forest has been checked. This principle applies regionally as well as locally. In the region of glacial moraines the deep soil has made possible a rapid accumu- lation of humus, as well as a thorough intermingling of rock soil and humus. There has been a rapid development of the soil, consequently the Acer-Fagus forest has been permitted to invade such regions. The time factor as an expression of the rate of soil development has limited the rate of migration and hence forest distributfon also. The time factor in relation to soil development explains both the numerous northerly outliers of Acer associations and also the outliers of the Picea-Abies forest. The northward deviation of the limits of such species as Acer saccharum, Fagus americana, and Tsuga canadensis has been noted. It is significant that this deviation coincides with a great depression extending to the height of land in which highly developed soil deposits are present. The deciduous hardwoods occur as outliers and are always found on the better soils. Although the writer has not been able to study the Saugenay basin personally, it may be ventured that the north- ward migration of Acer at this point is also to be explained on the basis of soil development. Bray (5) has found difficulty in explain- ing the occurrence of such trees as Abies and Picea in the swamps of New York. The soil in such localities is in a primitive stage of development. Although much humus has accumulated, there has been little or no intermingling of rock soil and humus, and the degree of aeration is low. The soil has been protected from the action of atmospheric agencies and running water, hence its undeveloped condition. The result is the same as when the Slowly weathering Laurentian rock resists the agencies which promote soil formation. The fact that the same forest species are present in both places emphasizes the potency of the rate of soil : development as a factor in the determination of tree distribution. 478 BOTANICAL GAZETTE [DECEMBER The evidence submitted is regarded as sufficient to prove that throughout a great region of Ontario dominated by Picea and Abies these genera are not permanent or climax forms, since they are replaced by Acer when soil conditions become favorable. WuitT- FORD (25), after studying the forests of Michigan, states that soils are improved by coniferous trees, and when sufficient humus soil has accumulated the deciduous species establish themselves. BRAY (5) also implies this relation. CoopEer (6), however, after a study of Isle Royal, comes to the conclusion that “this type (Adies- Picea-Betula forest) is the climax forest of that portion of the northeastern conifer region under consideration; in other words, upon Isle Royal it is the final and permanent vegetative stage toward the establishment of which all other plant societies are successive steps. It is the climatic forest of the region, permanent while the climate remains essentially as now.” The same paper records stands of Acer on certain ridges of Isle Royal and in other places where soil conditions seem particularly favorable. It seems probable that the occurrence of these stands might be a on the basis of soil development. It is evident from a study of the forests of northern Ontario that. the deciduous hardwood forest is encroaching upon the coniferous . forest region, and that the progress of this encroachment has lagged behind temperature changes, being now dependent principally upon, the rate of soil development. The relation of shade to the time factor of distribution is in accord with the specific tolerance of a given species with respect to light. The Acer-Picea forest provides shade which is essential for Acer seedlings, while detrimental to Picea or Abies seedlings. The encroachment of the deciduous hardwood forest upon the coniferous forest, made possible by changes in temperature and soil development, is also promoted, and the result made more permanent by decreasing light intensity due to shade condi- tions. The importance of methods of seed dispersal as an element of the time factor of distribution is obvious. Where changes in condi- tions are slow, for instance yearly temperature modifications, even the trees whose methods of dispersal allow them to migrate slowly 1918] HUTCHINSON—FOREST TREES 479 are able to keep pace, and where ecological changes are more rapid species having the best methods of seed dispersal naturally migrate most rapidly. The rapid invasion of a burned area by the Populus- Betula association is due primarily to the widespread dispersal of the seeds of these species. In contrast, Pinus takes its. place among the trees which appear later, largely because it has a less efficient method of scattering seeds. A number of examples of the limitations of seed dispersal have been noted. In several cases where a burn had left only one or two pines upon an island the usual Populus-Betula association was unable to gain a foothold because of the distance from the mainland; hence these species were superseded by numerous pine seedlings. Doubtless the same prin- ciples may be applied to the relation between seed dispersal and tree migration even over greater areas. The time factor of distri- bution may be an expression of the rate of migration as it is deter- ‘mined by the method of seed dispersal. The time factor of distribution is an expression of the rate of migration. The rate of migration is dependent upon such primary conditions as temperature, water supply, soil properties, light in- tensity, and methods of distribution. Time, as a condition of change in environmental factors, becomes itself of great importance in any consideration of the factors of forest distribution. Competition factor Competition results in the survival of the fittest. The fittest is that species or individual whose specific range of tolerance is best related to the environmental condition acting as a limiting factor toward other species; hence temperature, water supply, soil, or light may act as the basis of competition. Time may also act as a basis of competition, since it changes conditions in environmental factors. In order that competition may act as a distributional factor, conditions must be favorable for one or more species and unfavorable for others. While the time factor is an expression of the rate of change of the environmental factor acting in a limiting Capacity, the competition factor is an expression of the relation between the ranges of tolerance of the forms in question toward the limiting environmental factor. 480. BOTANICAL GAZETTE [DECEMBER The encroachment of the deciduous hardwood forest of Ontario upon the coniferous forest is accompanied by competition. The progressive changes in such conditions as humus content of the soil and light intensity are such as to increasingly favor the former asso- ciation to the detriment of the latter. Abies, for instance, grows readily on good soil, but it cannot tolerate the shade of an Acer forest. The competition becomes too great; in other words, the changes in environmental factors have been such that the mean of the range of tolerance of Acer more closely approximates existing conditions than that of Abies. The factor of competition plays its chief réle in the so-called transition areas, where the specific ranges of tolerance of the species concerned all include existing conditions although unequally. That species dominates, other things being equal, whose mean of tolerance more nearly approximates environ- mental conditions. Specific ranges of tolerance The specific ranges of tolerance of some of the dominant forest species of Ontario, together with their relation to limiting factors, will be considered. Many of the data are represented diagram- matically in the accompanying diagrams (figs. 3-6). These dia- grams summarize data collected regarding the specific ranges of tolerance of a number of forest species. In preparing the tempera- ture diagrams (fig. 3), for example, other factors have been elimi- nated by selecting data respecting localities where other conditions have been favorable; in this way the maxima and minima have been determined. The diagrams are relative rather than quanti- tative, hence they suggest a field of research which would supply absolute numbers. When the maxima and minima have been — determined, the means are represented by the mid-points of the lines joining these extremes. The radii of the circles of which the lines joining the extremes are diameters represent the magnitudes of the specific ranges of tolerance. The comparative areas of dis- tribution as determined by the several limiting factors are repre- sented, theoretically, by circles whose centers are the means of their ranges of tolerance and whose radii are the lines representing those ranges. 1918] HUTCHINSON—FOREST TREES 481 Picea mariana Picea canadensis Pinus Banksiana Abies balsamea Populus balsamifera Betula papyrifera Thuja occidentalis Tsuga canadensis Pinus Strobus Acer saccharum | Fagus americana | Quercus rubra Carya amara Castanea dentata Juglans cinerea Fic. 3.—Forest trees: specific ranges of tolerance with respect to temperature BOTANICAL GAZETTE [DECEMBER | Alnus incana Ulmus americana Picea mariana Larix americana | Fagus americana Tsuga canadensis Abies balsamea Betula papyrifera Pinus Strobus Populus tremuloides | Pinus Banksiana Fic. 4.—Fo: ‘ i 4 rest trees: specific ranges of tolerance with respect to water 1918] HUTCHINSON—FOREST TREES | Fagus americana Acer saccharum | Tsuga canadensis Abies balsamea Betula lutea Picea mariana Betula papyrifera Pinus Strobus (h (Zuercus rubr Populus balsamifera Thuja occidentalis Larix americana Fic. 5.—Forest trees: specific ranges of tolerance with 4384 BOTANICAL GAZETTE [DECEMBER | Populus tremuloides : Betula papyrifera | Pinus Strobus Abies balsamea Quercus rubra Picea mariana a Pony Betula lutea Pyrus americana Wiss Lo | Tilia americana Acer saccharum Wo Tuga canadensis Fagus americana Fic. 6.— : i : i 1G. 6.—Forest trees: specific ranges of tolerance with respect to intensity of light 1918] HUTCHINSON—FOREST TREES 485 ABIES BALSAMEA.—Physical factors‘ Moisture and tempera- ture are the main factors influencing the distribution of Abies balsamea; it requires a cold climate and a constant supply of mois- ture at its roots; a mean annual temperature not exceeding 40° F. with an average summer temperature of not more than 70° F. and a mean precipitation not less than 25 inches evenly distributed throughout the year are the necessary conditions for its growth” ZON (27). The maximum of its range of temperature tolerance is high, very closely approximating that of Picea; the minimum is lower than has generally been conceded, other factors having prac- tically eliminated it from the warmer regions of its normal tempera- ture range. While Abies balsamea normally has a wide water range, it seldom thrives except in a moist soil because this hin- ders the growth of a fungus which in a drier soil’attacks the roots (“ground rot’). Southward Abies balsamea “attains its best growth and largest sizes on flats the soil of which is usually a moist deep sand loam” (Zon 27), while ‘“‘southwest of Hudson Bay it grows only in the warmest and best soils and is entirely wanting in the cold swampy tracts” (Low 18). A. balsamea demands com- paratively high light intensity; seedlings are seldom found except in clearings caused by windfall, or otherwise. Generally, Abies has a wide range of tolerance. Competition factor—Northward Picea is the chief competitor of Abies; their ranges of tolerance are similar, the maxima and minima of Picea generally being more extreme; consequently Abies under most conditions would be secondary were it not for the fact that near the mean of their ranges Abies grows more rapidly than Picea. Southward the chief competitors are Acer and Tsuga. These forms have the advantage of being more shade tolerant, and hence they gradually encroach upon and finally exterminate Abies, which demands greater light intensity (fig. 7). Time factor —The northward migration of Abies balsamea is conditioned by temperature, and since the magnitude of tempera- ture changes is dependent upon time, it is evident that the time factor has a bearing upon distribution. The distribution south- ward is affected by competition of such forms as Acer and Tsuga. Time is necessary for the environmental changes which produce ¥ 486 BOTANICAL GAZETTE [DECEMBER AP Acer—soil | \ Abies—light Abies—soil Abies—temperature ay EEN NOE NIN OND Ne ~ 8 rs Kt Acer—water LS] N AKPREDLS . S Arce. . i) SES] A cer—temperature os be L +1 Acer—light iS Ne de \ ue ‘ Cae eet a ~*~ sy re 85 Cte ub ah ; ieee Fic. 7 impo: light; Wack I areas barred, the bars being tolerance; barred areas represent areas of competition. d Acer saccharum shown by super- cca areas of tolerance toward factors of temperature, water, similar to radii representing sp 58 soil, and of 1918] HUTCHINSON—FOREST TREES 487 conditions more nearly approximating the mean of the range of tolerance of these competitors, thereby contributing toward the elimination of Abies. The problem which arises by the appearance of Abies in swamps ‘ south of its “normal” range may be explained by regarding soil rather than temperature as the limiting factor. Soil changes have been slower in the undisturbed humus and rock soil layers of the swamp than on the weathered uplands. Soil conditions approxi- mating the mean of the range for Abies have been maintained, hence this form has persisted; also, the presence of abundant water inhibits the attacks of parasitic fungi, thereby permitting the growth of Abies. PICEA MARIANA AND P. CANADENSIS.—Although these species are closely related morphologically, they are quite different eco- logically; in this respect P. canadensis is quite closely associated with Abies balsamea. P. mariana has a wider range of tolerance than either, Low states, “In Labrador (and northern Quebec) the white spruce grows on rich intervale grounds or near the shores of lakes and rivers. The black spruce is found on hills and in cold swamps. The two kinds have the same geographical range north- ward.”’ Soil development and soil water frequently become limiting factors, separating these species. The status of tempera- ture as a factor in distribution is demonstrated by the differences existing between Picea mariana and P. canadensis. Although they have practically the same temperature range, the latter is not found throughout a vast area of the region lying between Hudson Bay and Labrador. Available accounts and the evidence given by its habitat in other regions indicate that soil development is the limit- ing factor. In this respect Abies balsamea takes a position inter- mediate between these two species of Picea. Picea has previously been referred to as the chief competitor of A dies. LARIX AMERICANA.—This species has a very wide range of tolerance toward temperature, water, and soil conditions. BELL (2) states ‘That it has an equally thrifty growth in the country , to the south of James Bay and westward toward Lake Winnipeg. In this great region it attains its greatest perfection in the dry up- lands and in good soil near the rivers, but smaller trees with small 488 BOTANICAL GAZETTE [DECEMBER black spruces grow everywhere on the low swampy grounds. South of the Ottawa River it grows principally on low and level land.” Low states, ‘‘ Larix is probably the hardiest tree of the subarctic forest belt. Throughout the interior it is found in all the cold swamps and is always the largest tree in the vicinity. Along the northern margin of the forest the larch continues a tree to the very edge where the black spruce is dwarfed to a mere shrub. Larix demonstrates the principle that a tree which has a wide range of tolerance does not flourish in competition with species of smaller range, but is crowded into situations where conditions exclude competitors. Such a form is usually of slow growth compared with forms which are more specialized. Larix cannot be called a xerophyte, a hydrophyte, or a mesophyte, since it may be any of the three. Although it is usually found under extreme conditions, it grows best under mean conditions, provided competitors have been eliminated. The distribution of Larix is accounted for by its wide range of tolerance, together with its low status in the com- petition scale.” THUJA OCCIDENTALIS.—The “anomalous” distribution of Thuja occidentalis defies explanation by regarding temperature, water, or soil as the limiting factors (figs. 1, 2). ‘It is absent in New- foundland, Cape Breton, Nova Scotia, and the east half of Prince Edward Island, but unusually large and fine in New Brunswick and the Gaspé peninsula, in which the climate, soil, etc., are the same as in the adjacent regions, where no trace of the species is to be found.”” Bett (2) also states that ‘there is a remarkable outlier of white cedar brushwood around Cedar Lake on the upper part of the Saskatchewan River at a distance of 190 miles to the north- west of the nearest point of the main area covered by this species.” Moreover, it is notable that throughout great areas, for instance the Temagami region, Thuja is unknown, while in the surrounding country it is abundant. T. occidentalis has a wide range of toler- ance toward environmental conditions. The presence of out- liers’’ where conditions are similar to those prevailing in other regions where it ordinarily occurs indicates that the general area of its distribution does not extend to its ecological limit, in many instances at least. The northern area of its distribution is 1918] HUTCHINSON—FOREST TREES 489 roughly outlined by a semicircle, a fact which contributes evi- dence that Thuja has migrated radiately from a limited area. The method of reproduction is such that it does not migrate rapidly; that a great proportion of seeds fail to develop is of importance in this connection. It would seem that the migration of this form has lagged behind changes in ecological conditions. With respect to its range of tolerance and its position in the scale of competitors under mean conditions, Thuja resembles Larix. These characters, together with the limiting action of time, account for most of the facts of its distribution. Pinus BANKsIANA.—The tolerance of this form toward low temperatures, dry conditions, and soil poor in humus, together with its limited range toward the other extremes, place it in a unique place among the trees of the northwestern region. It is practically eliminated from the low lying lands to the south and west of Hudson Bay and James Bay, water being the limiting factor. The inconsistencies in accounts of its northward distribution in this region are the result of its occastonal presence where there are higher lands between rivers. It extends northward to 56° N. lat. on the dry uplands east of Hudson Bay. Farther south, also, it is to be found only on dry rocky or sandy soil containing little humus. It is one of the pioneer forms and survives where it can avoid competition by enduring severe conditions. Pinus Strospus.—This species is also a pioneer among the coni- fers. Seedlings are seldom found except where there is a high light intensity and well drained soil. Its ranges of tolerance with respect to temperature and water do not include the extremes which characterize P. Banksiana. The northern limit follows a yearly isotherm (33° F.) very closely. It would seem that in this case temperature acts as a limiting factor. Because of its longevity and its towering height individuals or groves of mature trees often persist in a region where seedlings have long been eliminated by other forms which are higher in the competition scale. The pine forest is normally succeeded by such forms as Tsuga or Acer whose seedlings tolerated shade, the time factor; hence its perpetuation depends upon the maintenance of or reversion to pioneer condi- tions. 490 BOTANICAL GAZETTE [DECEMBER TSUGA CANADENSIS.—This species is among conifers what Acer and Fagus are among deciduous trees; it isa climax form. In fact, its ranges of tolerance are almost identical with those of the decidu- ous forms already mentioned. TJ. canadensis, when contrasted with such species as are represented by Pinus Banksiana, serves to emphasize the ecological diversity of conifers. BELL (2) states that ‘‘this tree maintains a good size to the verge of its range and always appears to terminate abruptly.”’ Stands of mature trees are to be found as “‘outliers”’ beyond the general area of its dis- tribution. This evidence confirms the belief that Tsuga is still migrating; that in many instances it has been limited by the time factor rather than by environmental factors. PoPULUS BALSAMIFERA AND P. TREMULOIDES.—Although P. balsamifera generally extends farther north than P. tremuloides, having a greater temperature reins of tolerance, its northern limit passes south of the latter at 71° W. long.; soil becomes the limiting factor in this region. P. balsamifera “appears to confine itself to heavy clay soil of the river valleys on the modified drift of the Cambrian areas’”’ (Low 18), while ‘“‘P. tremuloides is most plenti- ful on the unmodified glacial till of the drift ridges.’’ The seedlings of both require a high degree of light intensity, and as such are pioneer forms. Southward they occur only where fire and other destructive agencies have restored pioneer conditions. The abundance of these species of Populus northward, especially south and west from Hudson Bay, would indicate that this region is biologically young. ACER SACCHARUM.—South of the Laurentian Plateau A. sac- charum dominates, except in the undrained lowlands. Its range of tolerance is limited to a mature soil (that is, well drained, well aerated, and containing a relatively large amount of humus inti- mately mixed with the rock soil) and low light intensity. The humidity of the atmosphere in Ontario is such that it is doubtful whether it ever acts as a limiting factor, other conditions being favorable. It is evident that the distribution of Acer is chiefly an expression of the time factor; the time required to give rise to 4 deep, well drained humus soil and to shade conditions, and in addi- tion the time which is necessary to crowd out those forms which 1918] HUTCHINSON—FOREST TREES 491 have been instrumental in providing such conditions. As men- tioned before (under time factor), the time rate of change has been less in the lowlands and upon the rock outcrops of the Laurentian Plateau than upon the highland and the weathered glacial moraine. There is abundant evidence that Acer is migrating northward, its progress being contingent upon the time rate of soil development. FaGUS AMERICANA.—This species has a range of tolerance toward soil conditions which is even more limited than that of Acer. What has already been said for Acer applies equally for Fagus, since the latter is closely associated with the former. ULMus AMERICANA.—This species is another form whose dis- tribution defies explanation by considering either temperature or rainfall as limiting factors. The limit extends well into the plains and northward beyond Lake Winnipeg; it swings south- ward, then northward again in the region south of James Bay; then abruptly southward and again northward with no appar- ent dependence upon temperature or precipitation conditions. Even within its general limits it is found only where there is a clay, imperfectly drained soil; over large areas, especially throughout the Laurentian Plateau, it has not been found. ‘On the Mis- sinabi or west branch of the Moose River the white elm reappears 130 miles north of its general boundary on descending to a sufh- ciently low elevation above the sea” (BELL 2). Soil conditions are the chief limiting factors; on the clay soil of the lowlands, where there is poor drainage, is its favorite habitat; for this reason it is intermittingly distributed. Its reappearance north of the height of land, its occurrence in the lowlands about Lake Winnipeg, as well as many other eccentricities of this species, may be explained upon this basis. BETULA LUTEA.—This species may be associated with pioneer forms such as B. papyrifera or climax forms such as Acer. “Yellow birch is the most abundant hardwood in New England” (12), while in the lake region it is seldom seen; it becomes more abun- dant in the Laurentian region. “It grows in forests of widely different composition and shares to some extent the habits of paper birch, appearing on burns in small even-aged stands”’ (12). The seedlings are pioneer, yet, because of its comparative longevity, 492 BOTANICAL GAZETTE [DECEMBER the species persists after such forms as Picea, Abies, and even Acer have established themselves. Among such forms only mature trees are to be found. These characters explain its distribution. In New England and the Laurentian region the time rate of change has been small. The forms which will eventually succeed this species have not had time to eliminate it. On the glacial moraine of the lake region this process of crowding out has in most places reached completion. UNIVERSITY OF BritIsH COLUMBIA VANCOUVER, B.C LITERATURE CITED 1. BEAL, W. J., and WHEELER, C. F., Michigan flora. 1892. 2. Bett, Rosert, Report on explorations of rivers draining into Hudson Bay. Geog. Survey Canada. 1879-1881. Bowman, IsataAu, Forest physiography. 1914. 4. Briccs, L. J., and SHanrz, H. L., The wilting coefficient of different plants and its indirect determination. U.S. Bur. Pl. Ind. Bull. 2 and 3. Ig12. 5. Bray, W. L., The development of the vegetation of New York State. New York State Coll. Forestry, Tech. Pub. 2. 1915. 6. Cooper, W. S., The climax forest of Isle Royal, Lake Superior, and its development. ‘Bor. GAZ. 55:1-44, I§5-140, 189-235. 1913. Cow tes, H. C., The causes of vegetative cycles. Bor. Gaz. 51:161-183. w rs 7 IQII # , The physiographic ecology of Chicago and vicinity: a study of the origin, development, and classification of plant societies. Bot. GAz. 31%:73-108, 145-182. rgor. 9. DeForest, H., Recent ecological investigations. Proc. Soc. Amer. For- esters 60: I9gT4. : 10. FERNOw, B. E., An analysis of Canada’s timber wealth. Forestry Quart. 6: 337-353. 1908. ti. , Forest problems and forest resources of Canada. Proc. Soc. 12. FROTHINGHAM, E. H., The northern hardwood - its composition, growth, and snaewenaelt U.S. Dept. Agric. Bull. 13. FuLtER, Gro. D., Evaporation and soil moisture in ‘santo to the suc- cession of plant associations. Bor. Gaz. 58:193-234. 1914 14. Harper, R. M., Car window notes on the vegetation of the upper peninsula of Michigan. Rep. Mich. Acad. Sci. 1914: 193-108. 1918] HUTCHINSON—FOREST TREES 493 . Hows, C. D., and Wurre,’ J. H., Commission of Conservation, Canada. Forest protection in Canada, 1913-1 . Livincston, B. E., Temperature — in ‘plant geography and climatology. Bor. Bad 56: 340-375. , Climatic areas of the oe States as related to plant growth. Proc. ae Phil. Soc. 52:257-275. . Low, A. P., Report on eepiadia in he Labrador peninsula along the East Main, Koksoak, Hamilton, Manicugan, and assis of other rivers in 1892, 1893, 1894, 1895. Geol. Survey Can. 8:30-40. Macovwn, Joun, The forests of Canada and their SERN with notes on the more interesting species. Trans. Roy. Soc. Canada 4:3-20. 1894. , Ontario emer of ‘ee Lands. Rep. Survey and Explora- tion of N. Ontario of 1900. . Scuimper, A. F. W., Plant cae upon a physiological basis. Engl. ed. 1903. TRANSEAU, E. N., Forest centers of eastern North America. hee, Nat. 39:875-889. 1905. ————, Climatic centers and centers of plant distribution. Rep. Mich. Acad. Sci. 1905: 73-75. - WaRMING, E., Oecology of plants. Engl. ed. 1909 Wuitrorp, H. N., The genetic development of the forests of northern Michigan. Bor. Gaz. 31: 289-325. 1901 ZEDERBAUER, E., The light requirements of forest trees and methods of measuring light. Forestry Quart. 6:253-262. 1908. 27. ZON, RAPHAEL, Balsam fir. U.S. Dept. Agric. Bull. 55. 1914. NOTES ON NORTH AMERICAN TREES. III. TILIA. I C.- So SARGCENT 8. TILIA NEGLECTA Spach, Ann. Sci. Nat. II. 2: 140, ¢. 15. 1834; Hist. Vég. 4:29. 1835.—Tilia americana Curtis, Rep. Geol. Surv. N. Car. 3:79 (not Linnaeus). 1860; Tilia pubescens Watson and Coulter, Gray’s Man. ed. 6, 71 (in so far as relates to Long Island) (not Aiton). 1889; Sargent, Silva N. Am. 1:55 (in so far as relates toLong Island). 1891; Robinson, Gray Syn. Fl. 1*:343 (in so far as relates to Long Island). 1897; Britton and Brown, Ill. Fl. 2:414 (in so far as relates to Long Island). 1897; Tilia Michauxit Sargent, Man. 673. fig. 549 (not Nuttall). 1903; Robinson and Fernald, Gray’s Man. ed. 7. 565. 1908; Britton and Brown, Ill. Fl. ed. 2, 513 (probably in part). 1913.—Leaves thick and firm, acute or abruptly narrowed and long-pointed at apex, obliquely concave or unsymmetrically cordate at base, coarsely serrate with straight apiculate teeth pointing forward, dark green, smooth, glabrous and lustrous above, covered below except on the midribs and veins more or less thickly with short gray pubescence often slightly tinged with brown, and furnished with conspicuous tufts of axillary hairs, usually 11-14 cm. long and 6-11 cm. wide; petioles stout, glabrous, 3-6 cm. in length. Flowers about 1 cm. long, on pubescent or: nearly glabrous pedicels, in long-branched, slender, glabrous, mostly 5-15- flowered corymbs; peduncles slender, glabrous, the free portion 3-4 cm. in length, the bract nearly sessile or raised on a stalk up to 1.5 cm. in length, gradually narrowed and cuneate or unsymmetri- cally cuneate or rounded at base, rounded at apex, glabrous, 1-2 cm. wide and 7-15 cm. long, longer than the peduncle; sepals broadly ovate, acute, ciliate on the margins, glabrous on the outer surface, covered on the inner surface with long white hairs, about half as long as the lanceolate petals, rounded and notched at apex an rather longer than the spathulate staminodia; stamens included; style villose toward the base. Fruit ellipsoidal, ovoid, obovoid, or depressed-globose; rounded or acute or rarely gradually narrowed Botanical Gazette, vol. 66] ye | [404 1918] SARGENT—TILIA 495 and acuminate at apex, rarely 5-angled, covered with rusty or pale pubescence, usually 8-10 cm. in diameter. A tree 25~30 m. high, with a trunk sometimes 1 m. in diameter, smooth, often pendulous, branches forming a broad round head, and slender glabrous branchlets. Winter-buds ovoid, rounded at narrowed apex, about 5 mm. long, with glabrous, red-brown or light brown scales. Bark of the trunk about 2.5 cm. thick, deeply furrowed, pale reddish brown and covered with small thin scales. Flowers at the north in July and southward about a month earlier. Fruit ripens in September. Rich moist soil, Province of Quebec, near Montreal, to the coast of Massa- chusetts and New York, through the middle states to the valley of the Poto- mac River and along the Appalachian Mountains to those of North Carolina, and to Iuka, Tishimingo County, Mississippi, and from central and western New York to northern and southwestern Missouri (B. F. Bush, Noel, May 27 and October 8, 1909, nos. 5765, 59083; E. J. Palmer, Elk Springs, McDonald County, no. 4285; limestone cliffs, Current River, Van Buren County, July 4, 1914, no. 6180). Although I have not seen a type specimen of Spacn’s T. neglecta, his description can only refer to this tree, which seems to have been understood only by Spacu, whose description was made from trees cultivated in France. The younger Micuavux must have seen it in western New York, where he found what he called T. americana between Batavia and New Amsterdam forming two-thirds of the forest growth. In western New York, however, T. neglecta is a much more common tree than T. glabra. Gray, too, must have been familiar with T. neglecta, for it is common in central New York where as a young man he did most of his field work, and in his descriptions of T. americana he always says “essentially glabrous,” which would indicate that it might not be always glabrous. It was mistaken for T. glabra by Curtis as it seems to replace that species south of Maryland. Specimens of a tree of T. neglecta growing near Wading River, Long Island, have been referred by many authors to T. pubescens Aiton, and other authors have followed me in considering the tree which I now consider 7. neglecta to have been the T. Michauxii of NUTTALL, which is the T. argentea of MIcHAUX. In the shape and serration of the leaves and in the size and structure of the flowers and fruit there is little by which 7. neglecta can be distinguished from T. glabra, but as the absence or presence of pubescence or tomentum on Ameri- can species of Tilia is so important in distinguishing species, and as the pubes- cence on the lower surface of the leaves of 7. neglecta is so constant and so Persistent throughout the season, it seems best to consider it a species rather than a pubescent form of T. glabra. The base of the style of 7. neglecta is furnished with long hairs and that of 7. glabra appears to be quite glabrous. I find a slight pubescence on a branchlet from the upper part of a tree collected by Curtis and Corr near Ithaca, New York. Space describes the fruit of 496 BOTANICAL GAZETTE [DECEMBER his species as subpentagynous, and his figure represents a fruit with 5 distinct ridges. I have not seen such fruits on any specimens of wild trees, but they occur on two specimens of cultivated trees in the herbarium of the Arboretum, one from Germany and the other from Rochester, New York. On a tree culti- vated in Goldsboro, North Carolina, the fruit is ellipsoidal and borne in unusually long-branched clusters. g. TILIA CAROLINIANA Miller, Dict. ed. 8. 1758.—Tilia pubes- cens Aiton, Hort. Kew. 2:229. 1789; Ventenat, An. Hist. Nat. 2:68. 1800; Mém. Acad. Sci. Paris 4:10. #. 3. 1802; Elliott, Sk. 2:3. 1824; Tilia multiflora, Hort. ex Ventenat in An. Hist. Nat. 2:64. 1800; Tilia pubescens var. leptophylla Ventenat, l. c.; Tilia lepto- phylla Small, Fl. Southern States 762 (in part?). 1911.—Leaves ovate, oblique and truncate or cordate at base, abruptly long- pointed at apex, coarsely dentate with broad apiculate glandular teeth pointing forward, and coated below with a rusty or pale easily detached pubescence of fascicled hairs; when they unfold coated with hoary tomentum, soon glabrous on the upper surface, and at maturity dark yellow-green and lustrous above, 7-15 cm. long and 6-12 cm. wide; petioles stout, glabrous, 2.5-4 cm. in length. Flowers 6-7 mm. long, on slender pubescent pedicels, in small stout-branched, pubescent, mostly 8-15-flowered corymbs; peduncle slender, pubescent, the free portion 2-3 cm. long, the bract nearly sessile, oblong-obovate, cuneate at base, rounded or acute at apex, when it first appears nearly glabrous on the upper surface, pubescent becoming glabrous or almost glabrous below, 2 cm. wide, longer or shorter than the peduncle; sepals ovate, acuminate, ciliate on the margins, brown and covered with pale pubescence on the outer - surface, coated on the inner surface with long white hairs; petals lanceolate, acuminate, a third longer than the sepals; staminodia oblong-obovate, rounded at apex, rather shorter than the petals; style tomentose at base or glabrous. Fruit subglobose, ellipsoidal or obovoid, 7-9 mm. in diameter. A large tree with slender, red-brown, glabrous or slightly pubescent branchlets. Winter-buds ovate, acute, glabrous or rarely sacs a 5-6 mm. ng. NortH CaRoLina.—Wrightsville Beach, New Hanover cei, Ww. W. Ashe (no. 261); Wilmington, New Hanover County, T. G. Harbison, June aI, 1915, wae 2, 1916 (nos. 6, 8, 11, 12). 1918] SARGENT—TILIA 497 SouTH CaRoLInA.—Near Charleston and on James Island, T. G. Harbison, June 17 and 18 and September 6, rors (nos. 1, 7, 8, 13, 14, 1@), September 4 and 5, 1916 (nos. 15, 17, 18), May 1, ror7, June 3, 1918 (nos. 45, 46, 47, 48); Calhoun Falls, Abbeville County, May 26, 1918 (no. 17). Grorc1a.—Colonel’s Island, near Dunham, Liberty County, T. G. Harbi- son, September 8 and 9, 1916 (nos. 3, 7), June 19, 1917 (no. 18), Miss Julia King, crete 1QI7. UISIANA.—Avery Island, Iberia Parish, R. S. Cocks, October 18, 1910 (no. 6), May 24, 1914, May 209, July 28, 1916 (nos. 4040, 4054), Miss McIihenny, June 1915; Welsh, Jeff Davis Parish, E. J. Palmer, May 17, September 10, 1915 (nos. 7675, Sani Opelousas, St. Landry Parish, C. S. Sargent, March 2s, 1917; Little Bayou Téche, east of Opelousas, R. S. Cocks, April 3, July 24, 1916 (nos. 4012, 4016, 4018); rich woods near Winnfield, C. S: Sargent, April 6, 1913; Lake Charles, Calcasieu Parish, C. S. Sargent, April 10-13, 1915, R. S. ee May 21, June 1, 1915 (no. 2530); Natchitoches Parish, Natchitoches, - S. Cocks, April 15 and 27, 1911, E. J. Palmer, April 17 and 23, May 3, roi to and 14, July 10, September 30, rors (nos. 7397, 7474, 7946, 7952, 8013, 8021, 8747, 9416); Creston, E. J. Palmer, April 28, 1915 (no. 7420); Chopin, May 6, rors (no. 7554). ARKANSAS.—Fulton, Hempstead County, B. F. Bush, October 4, 1909 (no. 5926); Gum Springs, Clark County, E. J. Palmer, June 20, 1915 (no. 8074). EXAS.—Palestine, Anderson County, E. J. Palmer, September 21, 1917 (no. 12816); Marshall, Harrison County, June 8 and September 26, 1915 (nos. 7910, 8673), March 29, 1918 (no. 1320); Groesbeck, Limestone County, June 1, Pb 5 (no. 7934); Jacksonville, Cherokee County, June 4, 1915 (no. 7871); Larissa, Cherokee County, April 7, 1916 (nos. 9374, 9381); Houston, Harris County, September 15, 1917 (no. 12759); San Augustine, San Augustine County, April 19 and September 8, 1916 (nos. 9498, 10637); near Pledger, Matagorda County, May 8, 1916 (nos. 9698, 9704); Dayton, Liberty County, _ May 25, r915 (no. 7767); Blanco, Blanco County, June 4, 1917 (no. 12165); near Boerne, Kendall County, S. H. Hastings, 1911, C. A. Schatienberg, 1915, C. S. Sargent, 1915, E. J. Palmer, September 29, 1916 (no. 10866), April 20, 1917 (no. 10866). ivgsio.- fogs “998 Juni 55, Orizaba” (in Herb. Kew), Orizaba, 63, 1869 (in Herb. Kew), Pr. el Chica, C. Erenberg (in Herb. Kew, with slightly pubescent branchlets and winter-buds). MILLER’s specimen of his T. caroliniana from a tree cultivated in England, where it had been introduced from Carolina by CaTEsBY, is eine in the British Museum, the name being written on the sheet in MILLER’s hand This specimen is also the type of Arron’s T. pubescens, that name also sigbar: ing on the sheet in Arron’s or DRYANDER’s handwriting. This specimen has glabrous branchlets, coarsely serrate leaves, very oblique and truncate at base, and covered below, like the corymbs, with rusty pubescence. The leaves are rather smaller than those of the trees now growing about Charleston Harbor, 498 BOTANICAL GAZETTE [DECEMBER as might be expected in the case of a tree from the southern states cultivated in England. There is no other linden in the South Carolina region which at all agrees with MILLErR’s specimen, and his name can properly be taken up for thistree. T.caroliniana has usually been considered a synonym of T. americana Linnaeus, and T. pubescens has been adopted for one of the southern coast species. This name, however, except as a synonym of T. caroliniana, must now disappear. The leaves of the specimens collected west of the Mississippi River which are here referred to T. caroliniana are certainly not thinner than those from the Carolina coast region, and I can find no characters by which the eastern and western trees can be distinguished. As here understood the range of T. caroliniana is remarkable, as there is no evidence that it occurs between the coast of Georgia and western Louisiana. TILIA CAROLINIANA var. rhoophila, n. var.—Tilia pubescens Torrey and Gray, Fl. N. Am. 1:240 (insomuch as relates to Texas). 1842; Tilia pubescens Sargent, Silva N. Am. 1:55 (insomuch as relates to Louisiana and Texas). 1891, and later authors; T#lia pubescens var. a Aitonii, V. Engler, Monog. Tilia, 128 (insomuch as relates to Texas specimens). 1909.—Differing from the type in its pubescent branchlets and winter-buds, its usually larger leaves, and in its tomentose corymbs of more numerous flowers. Leaves broadly ovate, oblique and truncate or cordate at base, abruptly short-pointed and acuminate at apex, coarsely serrate with broad apiculate teeth pointing forward, dark green and lustrous on the upper surface, pale and thickly covered on the lower surface with persistent white or brownish pubescence, 10-12 cm. long and 7-12 cm. wide, with slender midribs and primary veins pubescent on the lower side and small i axillary tufts of pale hairs; petioles stout, thickly coated with pubescence, 2.5—4 cm. in length; on vigorous shoots leaves often 16 cm. long and 14 cm. wide, and occasionally 24 cm. long and 18 cm. wide. Flowers 5—6 mm. long, on short, hoary tomentose pedicels in wide, thin-branched, pubescent, many-flowered (sometimes 50) corymbs; peduncle thickly covered with fascicled hairs, the free portion 3.5-5 cm. long, the bract oblong, unequally rounded at base, rounded at apex, glabrous on the upper, pubescent on the lower surface, 1.5—2 cm. wide, usually shorter than the peduncle; sepals acuminate, coated on the outer surface with pale or slightly rusty pubescence, villose and furnished at base on the inner surface with tufts of long hairs; petals lan- 1918] SARGENT—TILIA 499 ceolate, acuminate and ciliate at apex, about a third longer than the sepals; staminodia spatulate, acute, about half the length of the petals; style coated at base with long white hairs. Fruit sub- globose, covered with rusty tomentum, 7-8 mm. in diameter. A tree with slender branchlets thickly coated ‘during their first year with pale pubescence, dark red-brown or gray and puberulous during their second season. Winter-buds covered with pale pubescence. ARKANSAS.—Fulton, Hempstead County, E. J. Palmer, June 17, 1915 (no. 8023); Gum Springs, Clark County, June 21, 1915 (no. 8074) Lovutstana.—Bank of the Calcasieu River, Lake Charles, Calcasieu Parish, R. S. Cocks, May 21, 1916 (no. baci C. S. Sargent, March 23, 1917; low woods, Welsh, Jeff Davis Parish, E. J. Palmer, May 17, June 21, and September ro, rgr5 (nos. 7674, 8074, 8500). Trxas.—Houston, Harris County, F. Lindheimer, 1842 (no. 10830 in Herb. Missouri Bot. Gard.), E. J. Palmer, May 24 and 26 and September 17, 1915 (nos. 7758, 7776 type for flowers, 8578), April 29, 1916 (no. 9613), April 2, May 16, 17, 18 and September 15, 18, 1917 (nos. 11142, 11443, 11448, 11451, 11454, IIQII, I1QI2, 11913, IIQI4, 11916, 11917, 11918, 11933, 11934, 11946, 11964, 12755, 12756, 12758, 12762, 12788), March 19, 29, 1918 (nos. 13114, 13115); Harrisburg, Harris County, E. J. Palmer, May 17, 1917 (no. 11933); Morgan’s Point, Harris County, E. J. Palmer, May 20, 1917 (no. 11957); near Pledger, Matagorda County, E. J. Palmer, May 8, 1916 (no. 9695) ; Dayton, Lib- erty County, E. J. Palmer, May 25 and September 16, 1915 (nos. 7672, 7767, 7770, 8548, 8564, 8566), April 28, 1916 (nos. 9603, 9604, 9605, 9607), April 3,. May 21, and September 17, 1917 (nos. 11457, 11458, 11460, 11465, 11466, 11975, 11976, 11982, 11984, 12776, 12777, 12778, 12779 with bracts of the peduncles ro-rz cm. long and 3.5 cm. wide); Palestine, Anderson County, E. J. Palmer, May 29, 1917 (no. 12086); Marshall; Harrison County, B. F, Bush, August 9, 1901 (no. 659), E. J. Palmer, June 8, 1915 (no. 7922); College Station, Brazos County, E. J. Palmer, April 28, 1917 (nos. 11720, 11721); Bryan, Brazos County, EZ. J. Palmer, April 28, 1917 (no. 11721); Liberty, Liberty County, E. J. Palmer, May 22, 1915 (no. 7735), April 28, 1916 (no. 9594); Livingston, Polk County, September 12, 1916 “ 10697), September 19, 1917 (no. 12798); New Braunfels, Comal County, FP. Lindheimer, 1842 (no. 10839 in Herb. Missouri, Bot. Gard.); rocky banks of the Guadalupe River, Kerrville, Kerr County, E. J. Palmer, April 29, 1916 (nos. 9931, 9934). Growing usually on the margins of sandy bogs and on moist sandy hillsides, this tree varies, according to the moisture it obtains, in the size of the leaves and in the amount of the pubescence on the branchlets. The bark on trees growing in wet situations is smooth and pale, but on trees in dry soil or higher on the hillsides it is dark and Siete the leaves are smaller and the branchlets are less pubescent. 500 BOTANICAL GAZETTE [DECEMBER 10. Tilia texana, n. sp.—Tuilia pubescens var. B Ventenatii V. Engler Monog. Tilia 129 (in part). 1909.—-Leaves thin, oblong- ovate, abruptly contracted into long slender acuminate points, cordate or obliquely cordate at base, finely dentate with broad apiculate teeth; early in the season pubescent above with scattered fascicled hairs and covered below with brownish, slightly attached pubescence, and in the autumn light yellow-green, lustrous and nearly glabrous on the upper surface, slightly pubescent on the lower surface, 10-14 cm. long and 8-10 cm. wide, with slender mid- ribs and primary veins sparingly villose on the upper side and nearly glabrous on the lower side, and small axillary tufts of brownish hairs; petioles slender, pubescent with fascicled hairs, 2.5-4 cm. in length; leaves on vigorous shoots often furnished with one or two large, lateral, acuminate, serrate lobes, more coarsely dentate and more thickly covered on the lower surface with pubescence, often 13-15 cm. long and 9-15 cm. wide. Flowers 6-7 mm. long on slender tomentose pedicels in small, villose-pubescent, mostly 7—10- flowered corymbs; peduncle slender, slightly villose-pubescent, the free portion 3—3.5 cm. in length, the bract oblong-ovate to slightly obovate, unsymmetrically cuneate at base, rounded and occa- sionally lobed at apex, glabrous on the upper surface, densely pubescent early in the season, later becoming nearly glabrous on the lower surface, longer or shorter than the peduncle; sepals ovate, acute, pale pubescent on the outer surface, covered on the inner surface with white hairs longer and more abundant near the base; petals lanceolate, acuminate, a third longer than the sepals; staminodia linear-lanceolate, acuminate; style hoary tomentose at the base. Fruit ellipsoidal, covered with rusty brown tomentum, 8-9 mm. long and 5-6 mm. in diameter. A small tree with slender branchlets thickly covered during their first season with close pale pubescence, and pale and puberulous or glabrous in their second year. Winter-buds ovate, obtusely pointed, thickly covered with pale pubescence, 4-5 mm. long. On vigorous terminal branchlets the pubes- cence is thicker ahd light rusty brown. Trexas.—Columbia, Brazos County, B. F. Bush, October 18, 1900 (no. 1570), September 25, rgor (no. 914); Houston, Harris County, E. J. Palmer, September 17, rors (no. 8578), April 29, 1916 (nos. 9608, 9610); Larissa, Chero- kee County, June 3 and September 22, 191 5 (nos. 7845, 8620); on Spring Creek, 1918] SARGENT—TILIA 501 near Boerne, Kendall County, E. J. Palmer, April 7 and September 29, 1917 (nos. 11485, 12899); along the southwest bank of the Guadalupe River on the rocky talus in a canyon at the foot of a limestone bluff at Kerrville, Kerr County, E. J. Palmer, October 2, 1916 (nos. 10887, 10888 type for fruit), April 8 and June 9, 1917 (nos. 11495, 11501, 11502, 12212, 12213 type for flowers, 12214). I have not seen leaves and coepens bracts with lateral lobes on any other American linden. TILIA TEXANA var. grosseserrata, n. var.—Differing from the type in the coarse serration of the leaves, in the absence of lateral lobes on the leaves and on the bracts of the peduncles, and in the constantly pale, never rusty pubescence of the branchlets and winter-buds. A small tree with several stems 7~9 m. high, the bark dark gray and rough near the ground and smooth and pale above, ih rocky soil at the foot of a lime- stone bluff by a small stream forming the head of the Sabinal River, near Utopia, Uvalde County, Texas, E. J. Palmer, June 17, 1916 (no. 10227 type), April 10 and October 6, 1917 (nos. 11522, 12937). At the end of their first winter the branchlets of this tree are pale pubescent, puberulous or nearly glabrous, and the winter-buds are reddish or pale brown and glabrous. This linden is interesting as the most western representative of the genus in the United States. 11. Tilia phanera, n. sp.—Leaves semiorbicular to broadly ovate, deeply and usually symmetrically cordate at base, abruptly short-pointed at apex, finely dentate with straight or incurved apiculate teeth; when they unfold glabrous above with the excep- tion of a few Sindee on the midribs and veins, and thickly coated below \with hoary tomentum, and at maturity thin, blue-green, smooth and lustrous on the upper surface, paler and often brownish and coated with a floccose easily detached pubescence of fascicled hairs on the lower surface, 5—9 cm. wide and usually rather broader than long, with slender midribs and primary veins pubescent on the lower: surface, and small axillary clusters of rusty brown hairs; petioles slender, coated when they first appear with hoary tomen- tum, glabrous or slightly pubescent in the autumn, 2.5-4 cm. in length. Flowers 5—6 mm. long, on tomentose pedicels in compact, villose, mostly 16-20-flowered corymbs; peduncle villose, the free portion 1.2-1.5 cm. in length, the bract longer than the peduncle, 502 BOTANICAL GAZETTE [DECEMBER short-stalked, obovate, cuneate at base, broad and rounded at apex, floccose pubescent on the lower surface, nearly glabrous on the upper surface; sepals acuminate, pale pubescent on the outer surface, villose along the margins and furnished at the base on the inner surface with a tuft of long white hairs, broader and shorter than the lanceolate acuminate petals; staminodia oblong-obovate, rounded at apex, style glabrous except at the base. Fruit ellip- soidal, covered with rusty tomentum, 8-10 mm. long and 6-7 mm. wide, on stout, densely floccose-pubescent pedicels. A tree with slender, light gray-brown, often zigzag branchlets covered when they first appear with fascicled hairs, deciduous during their first summer. Winter-buds ovate, obtusely pointed, ie reddish brown, glabrous, 4-5 mm. long. Flowers the middle of June. Fruit ripens the end of September. Banks of Spring Creek, near Boerne, Kendall County, Texas, E. J. Palmer, September 27, 1916 (no. 10825 type); April 7 and 11 and June 13, 1917 (nos. 11486, 11593, 12242). TILIA PHANERA var. scabrida, n. var.—Tuilia pubescens var. a Aitonii f. gymnophylia V. Engler, Monog. Tilia 130 (in part). 1909. —Differing from the type in the scabrate lower surface of the leaves. Leaves broadly ovate, cordate at base, abruptly short- pointed at apex; when they unfold pubescent above with scattered straight white hairs and hoary tomentose below, and at maturity thin, yellow-green and glabrous above and roughened below by the persistent bases of fascicled hairs, 10 cm. long and broad; petioles 2-2.5 cm. in length. Flowers not collected. Fruit on tomentose pedicels, ovoid to subglobose, covered with pale reddish tomentum. A small tree with dark deeply ridged bark and glabrous branchlets. On a low limestone bluff of the Blanco River, near Blanco, Blanco County, Texas, J. Reverchon, July 1885 (no. 1500 type), E. J. Palmer, April 16 and September 24, 1917 (nos. 11565, 12858); College Station, Brazos County, Texas, B. F. Bush, July 4, 1900 (nos. 1015, i Velasco, Brazoria County, Texas, E. J. Palmer, March 21, 1918 (no. 13 12. Tilia lasioclada, n. sp.—Leaves ovate, abruptly catia at apex into short acuminate points, oblique and truncate or on weak branchlets, often nearly symmetrical and deeply cordate at base, and finely serrate with straight apiculate teeth; when they unfold covered above with soft caducous hairs, pubescent below, and at 1918] SARGENT—TILIA 503 maturity thick, bright green, smooth and lustrous on the upper surface, pale and covered on the lower surface with a thick floccose, easily detached pubescence of fascicled hairs, pale on those of lower leaves and often rufous on those of upper branches, 10-15 cm. long and 8-12 cm. wide, the slender midribs and veins covered below with straight hairs mixed with fascicled hairs, and small conspicuous axillary tufts; petioles covered when they first appear with straight hairs mixed with fascicled hairs, soon glabrous, usually 3-4 cm. in length, those of the leaves of weak branchlets very slender and often 5-6 cm. long. Flowers 5-6 mm. long, on stout villose pedicels, in long-branched, mostly 1o~15-flowered corymbs more or less thickly covered with straight white hairs; peduncle covered with long white hairs, the free portion 2.5-—3 cm. in length, the bract nearly sessile, rounded and unsymmetrical or acute at base, rounded or acute at apex, the midrib more or less thickly covered on the lower side with straight hairs, otherwise glabrous, 2-5 cm. wide; sepals narrow, acute, pubescent on the outer surface, villose on the inner surface, about one-third as long as the lanceolate acuminate petals; staminodia spathulate, rounded and often lobed at apex, about as long as the sepals; style slightly villose at base. Fruit globose or depressed-globose, covered with rusty tomentum, about 1 cm. in diameter. A tree sometimes 20 m. high with a trunk 30-60 cm. in diameter, stout branches forming a broad round-topped weitte and stout neloheve branchlets sometimes glabrous in early summer less thickly during their first and second seasons with long straight hairs. SoutH Carotmna.—Calhoun Falls, Abbeville County; upland woods, Anderson County, T. G. Harbison, May 21, 1918; rich wooded slopes near the Savannah River, three miles below Augusta, T. G. Harbison, June 17 and August 23, 1916 (no. 8 type), June 17, 1917 (no. 9); Beach Island, a rich wooded slope rising from the north bank of the Savannah River a few miles below Augusta, R. C. Berckmans, June 12, 1914. Grorci1a.—Shell Bluff on Savannah River 30 miles below Augusta, Rich- mond County, C. S. Sargent, April 6, 1914; steep rocky bluff at the Locks above Augusta, T. G. Harbison, May 13, 1913 (nos. 1162, 1163), May 27 and October 6, r914 (no. 9), April 6, 1916 (no. 6), August 23, 1916 (nos. 12, 13). Brickyard near the Berckman’s Nursery west of Augusta, October 5, 1914 (no. 7), August 23, 1916 (no. 16), June 29, 1917 (no. 16), May 31, 1918 it 31), May 30, 1918 (nos. 18, 20, 21 type). 504 BOTANICAL GAZETTE [DECEMBER FiLorma.—River Junction, Gadsden County, T. G. Harbison, April 25 and September 21, 1914 (no. 1479), April 19 and June 25, 1917 (nos. 116, 119). From all other American lindens this species differs in the straight hairs on the lower side of the midribs and veins of the leaves, on the peduncle and branches of the inflorescence and on the branchlets, and similar to those of the European Tilia platyphyllos Scopoli. The number of these hairs varies on different individuals, and on some trees the branchlets become nearly glabrous by the middle of June, while on others the hairs are present for 2 or 3 years. They are longer and more abundant on the trees growing on the Savannah River at the Locks above Augusta than on trees from other localities, and do not entirely disappear until their third season. 13. TILIA HETEROPHYLLA Ventenat.—Different plants have been referred to this species and it is still by no means clear what should be taken as the type. VENTENAT gives the locality for his tree as ‘‘la basse Caroline”? where it was discovered by MICHAUX and Fraser. ‘Basse Caroline” may mean the coast region or the whole state east of the mountains. There is no Tilia in the South Carolina coast region which at all agrees with VENTENAT’S descrip- tion and figure, but near Augusta and in Columbia County, Georgia, and in the neighborhood of Walhalla in Oconee County, South Carolina, on the eastern foothills of the Blue Ridge, a linden is common which in the shape of the leaves agrees better with those figured by VENTENAT than any I have seen. MIcHAvx in his journeys from Charleston to the high Carolina mountains went up the valley of the Savannah River and passed by Augusta and through Oconee County, South Carolina. VENTENAT describes the leaves of T. heterophylla as snow white on the lower surface. On the Georgia and Walhalla trees the tomentum on the lower surface of some of the leaves is white and on others, especially from upper branches, it is rusty brown, a peculiarity of this tree which is common in other parts of the country. WENTENAT describes the fruit of his tree as globose and 5-ribbed. The fruit which he figured, however, is ellipsoidal and shows no trace of ribs. If the Walhalla trees, as I believe, are to be considered typical of T. heterophylla Ventenat, that species may be described as follows: TILIA HETEROPHYLLA Ventenat, An. Hist. Nat. 2:63. 1800; Mém. Inst. Paris 4:16. ¢. 5.—Leaves ovate, obliquely truncate or rarely slightly cordate at base, gradually narrowed and acuminate 1918] SARGENT—TILIA 505 at apex, finely dentate with apiculate gland-tipped teeth; when they unfold pubescent on the upper surface with caducous fascicled hairs, and at maturity dark green and glabrous on the upper sur- face, covered below with thick, firmly attached, white or on upper branches often brownish tomentum, and usually furnished with small axillary tufts of rusty brown hairs, 8-13 cm. long and 6-10 cm. wide; petioles slender, glabrous, 3.5-4 cm. in length. Flowers 6-7 mm. long on pedicels pubescent with fascicled hairs, in wide mostly 10-20-flowered pubescent corymbs; peduncle glabrous, the free portion 2-4 cm. in length, the bract narrowed and rounded at apex, unsymmetrically cuneate at base, pubescent on the upper, tomentose on the lower surface when it first appears, becoming glabrous, nearly sessile or raised on a stalk up to 1 cm. in length; sepals acuminate, pale-pubescent on the outer surface, villose on the inner surface and furnished at base with a tuft of long white hairs; petals lanceolate, acuminate, a third longer than the sepals; staminodia oblong-ovate, acute, sometimes notched at apex; style villose at base with long white hairs. Fruit ellipsoidal, apiculate at apex, covered with rusty brown tomentum, 7-10 mm. long. A large tree, with slender, glabrous, reddish or yellowish brown branchlets and oblong-ovate, slightly flattened, glabrous winter-buds 5-7 mm. in length, the outer scales slightly ciliate at apex. Nortu Carortna.—Falls of the Yadkin River, Stanley County, J. K. Small, August 1892; near Newbern, Craven County, 7. G. Harbison, June 5, 1918 (nos. 42, 44 with styles villose to the middle). SoutH Carortina.—Walhalla, Oconee County, 7. G. Harbison, June 4 and 22,1915, March and October 11, 1917; Russell, Oconee County, T. G. Harbison, May 5 and June 29, 1917 (nos. 3, 4), July 7, 1917 (nos. 18, 20). Grorc1a.—Cornelia, Habersham County, T. G. Harbison, July 7, 1917 (nos. 18, 20); Toccoa, Stevens County, 7. G. Harbison, June 15, 1918 (no. 9); banks of Flint River, Albany, Dougherty County, J. K. Small, May 24-28, 1806, T. G. Harbison, June 25, 1915 (no. 3); near Zluguenin, Sumter County, R. M. Harper, July 11, 1901 (no. 1049); banks of Savannah River, Germain’s Island, Columbia County, R. M. Harper, June 1, 1902 (no. 1302). Fiorma.—Tallahassee, Leon County, 7. G. Harbison, June and Septem- ber 1915 (nos. 1-6); River Junction, Gadsden County, June rors and 1916 (nos. 1, 8, 14, 17, 18, 19, 20, 22, 23, 28, 30, 34, 36, 37, 38, 62); Rotk Cave, Jackson County, R. M. Harper, April 28, 1910; near Marianna, Jackson County, T. G. Harbison, September 18, 1916 (nos. 2, 4), May 26, 1917 (no. 21). 506 BOTANICAL GAZETTE [DECEMBER ALABAMA.—Berlin, Dallas County, R. S. Cocks, June 4, 15, 1915 (nos. 780, 782), July 20, 26, 1916 (nos. 962, 970, 1012), June 18, July 25, 1918 (nos. 790, 792); near Selma, Dallas County, T. G. Harbison, April 20, 1915 (no. 22). West Vircinta.—White Sulphur Springs, Greenbrier County, Kenneth | Mackensie, no. 7532 in Herb. Mo. Bot. Gard. (T. heterophylla var. microdonta V. Engler, Monog. Tilia, 135). INDIANA.—Nea ened Switzerland County, C. C. Deam, July 25, 1913, June 19 and September 8, ro15 (nos. 13808, 16159, 18806); near the Ohio River, Jefferson County, September 8, 1915 (no. 16219). On the Florida trees the clusters of hairs at the base of the inner surface of the sepals and the hairs at the base of the style are sometimes wanting; and the fruit is subglobose, sometimes longer than broad or a little broader than long. Like the trees at Walhalla, the tomentum on the under surface of the leaves of the upper branches is usually rusty brown and silvery white on those of the lower branches. This linden is the common species in the neighborhood of Tallahassee and River Junction, and it appears to have been usually confounded in recent years with a tree of the higher Appalachian Mountains to which I have given the name of T. monticola. In the size and shape of the leaves this mountain tree resembles those of T. heterophylla, but the tomentum on the lower surface is thicker and whiter and never brown; the petioles are longer and the flowers are nearly twice as large; the branches are red, not yellowish brown, and the winter-buds are larger, more compressed, and bright red. TILIA HETEROPHYLLA, var. Michauxii, n. var.—Tilia alba Michaux f. Hist. Arb. Am. 3:315, f. 2 (not Linnaeus). 1813; Tilia heterophylla Nuttall, Silva 1:90, t. 23. 1842, and of many authors insomuch as relates to the Northern States; Tilia M ichauxtt Nuttall, Silva 1:92. 1842; Britton and Shafer, North Am. Trees 688 (in part). 1908; Britton and Brown, Ill. Fl. ed. 2, 2:513 (in part), fig. 2846. 1913; Tilia eburnea Ashe, Bot. GAZ. 332230. 19025 Tilia apposita Ashe, Bull. Charleston Mus. 13:27. 1917; Tilia tenera Ashe, /.c. 1917.—Differing from the type in the usually cordate, rarely obliquely truncate, more coarsely serrate leaves, broader and more abruptly acuminate at apex, and always white or grayish white, not brownish, tomentose below. This is one of the most widely distributed of the American lindens, ranging from the valley of the Susquehanna River in Pennsylvania, where it was first noticed by the younger Micuavx in Lancaster County, to southern and western New York, through southern Ohio and Indiana to northeastern Missouri (Ilasco, Ralls County, John Davis, September 30, 1914 (no. 3164), southwestern Mis- souri (Eagle Rock, Barry County, E. J. Palmer, July 16, 1914, no. 6287); 1918] SARGENT—TILIA 507 and northwestern Arkansas (Eureka Springs, Carroll County, E. J. Palmer, September 21, 1913, no. 4412, Cotter, Marion County, September 1, 191s, no. 8405). Southward it ranges through eastern Kentucky and Tennessee to northeastern Mississippi, along the Appalachian Mountains and their foothills to northern Georgia and to southern Georgia and Dallas County, Alabama. I have not seen specimens of this linden from Illinois, although it may be expected to occur in ravines near the Ohio River in the southern part of the state. : TILIA HETEROPHYLLA var. nivea, n. var.—Differing from the type in the whiter tomentum on the lower surface of the leaves, the’ glabrous styles, in the tomentum on the lower side of the bract of the peduncle at the time the flowers open, the slightly pubescent gray or pale reddish brown branches, and in the puberulous winter- uds. FLorwwa.—In deep woods, River Junction, Gadsden County, T. G. Harbi- son, April 19 and June 25, 1917 (no. 29 type), June 7, 1915, and June 25, 1917 (no. 27), A. H. Curtiss, June 4 and September 13, 1897 (no. 5875). TILIA HETEROPHYLLA var. amphiloba, n. var.—Difiering from the type in the fascicled hairs on the upper surface,of the young leaves and in the often pubescent branchlets. Leaves broadly - Ovate, sometimes broader than long, abruptly short-pointed or gradually narrowed and acuminate, or occasionally rounded at apex, symmetrically or obliquely cordate or obliquely truncate at base, finely serrate with apiculate teeth; when they unfold hoary tomentose below and covered above with fascicled hairs, and at maturity thin, dark yellow-green, smooth and Justrous on the upper surface, pale green or brownish and covered below with thick, white, somewhat loose tomentum, on lateral branchlets 4-6 cm. long and 5-7 cm. wide, and on leading shoots 9-10 cm. long and 7-8 cm. wide, the midrib and primary veins covered below with fascicled hairs; axillary hairs rusty brown in small inconspicuous tufts, often wanting; petioles slender, sparingly pubescent when they first appear, becoming glabrous, 2-2.5 cm. in length. Flowers 4-5 mm. long, on stout tomentose pedicels, in broad, thin-branched, slightly pubescent, 7—25-flowered corymbs; peduncle slender, pubescent, the free portion 3-5 cm. in length, the bract oblong to oblong-obovate, cuneate at base, rounded at apex, short-stalked 508 BOTANICAL GAZETTE [DECEMBER or nearly sessile, 7 mm.—2.5 cm. in width, thickly covered when it first appears with hoary tomentum, and at maturity tomentose on the upper and pubescent on the lower surface; sepals acuminate, densely pubescent on the outer surface, villose near the margins on the inner surface, about as long as the lanceolate acuminate petals; staminodia oblong-obovate, rounded at apex, about as long as the sepals; style slightly villose at base. Fruit ellipsoidal, covered with rusty brown tomentum, 7-8 mm. long and 5-6 mm. in diameter. A tree 20 m. high with slender red-brown or orange-brown branchlets glabrous or sometimes covered early in their first season with fascicled hairs. Winter-buds terete, glabrous or when first formed sparingly villose, 2-3 mm. in length. Flowers at the end of June and at River Junction later than the other species with which it is associated. Fruit ripens the middle of September. Fiorma.—In woods in sandy soil, River Junction, Gadsden County, T. G. Harbison, April 26 and September 21, 1914, April 19 and June 25, 1917 (no. 1484 type), September 21, 1914 (no. 1), June 7 and 28 and September 14, I915 (nos. 12, 13, 34, 34a, 36, 36a). ALABAMA.—Valley Head, Dekalb County, T. G. Harbison, June 26, 1918 (nos. 42, 42). I once believed that these trees could be specifically separated from 7. heterophylla, but their close connection with that species is shown by a tree of T. heterophylla var. Michauxii which was growing near Tiptop, Tazewell County, Virginia, in May 1914 (T. G. Harbison, no. 1616). ‘The upper surface of the leaves of this tree were then covered with fascicled hairs and the branch- lets were glabrous. When I visited Tiptop in September of the same year this tree had been cut down, but had produced shoots from the stump which were thickly covered with fascicled hairs and bore large leaves densely pubes- cent on the upper surface. 14. Tilia monticola, n. sp.—Tilia heterophylla, Sargent, Silva N. Am., 1:59 (in part, not Ventenat). #. 27. 1891; Man. 674 (in part). fig. 550; Robinson in Gray Syn. Fl. 17:344 (in part). 1908; Small, Fl. S. States 761 (in part). 1903; Robinson and Fernald, Gray’s Man. ed. 7, 566 (in part). 1908; Britton and Shafer, N. Am. Trees 686 (in part). 1908; Britton and Brown, Ill. Fl. ed. 2, 2:51? (in part). 1913.—Leaves thin, ovate to oblong-ovate, very oblique and truncate or obliquely cordate at base, gradually narrowed and acuminate at apex, finely serrate with straight or incurved apiculate teeth, smooth, dark green and lustrous on the upper surface, thickly 1918] SARGENT—TILIA 509 coated on the lower surface with hoary tomentum, 10-17 cm. long and 8-12 cm. wide; petioles slender, glabrous, 4-7 cm. in length. Flowers 10-12 mm. long, on stout sparingly pubescent pedicels in mostly 7—10-flowered, thin-branched, glabrous corymbs; peduncle slender, glabrous, the free portion 3.5~4 cm. in length, the bract gradually narrowed and cuneate or rounded at base, narrowed and rounded at apex, glabrous, 10-14 cm. long and 2-2.5 cm. wide, its stalk varying in length from 1 to 2.5 mm.; sepals ovate, acute, ciliate on the margins, covered on the outer surface with short pale pubescence and with silky white hairs on the inner surface; petals lanceolate, acuminate, twice longer than the sepals; stami- nodia oblong-lanceolate, rounded at the narrowed apex, as long or nearly as long as the petals; style clothed at the base with long white hairs. Fruit ovate to ellipsoidal, covered with pale rusty tomentum, 7-8 mm. long and 6-7 mm. in diameter. A tree rarely exceeding 20 m. in height with a trunk 1-1.10 m. in diameter, slender branches forming a narrow rather pyramidal head, and stout glabrous branchlets usually bright red during their first year, becoming brown in their second season. Winter-buds compressed, ovate, acute or rounded at apex, light red, covered with a glaucous bloom, 7-10 mm. long. Bark of the trunk 1.5 cm. in thickness, deeply furrowed, the surface broken into small, thin, light brown scales. Flowers from July 12 to July 25. Fruit ripens in September. Nort Carorina.—Highlands, Macon County, at an altitude of about 600 m., 7. G. Harbison (many specimens), June, July, and September rors; Busbee Mountain, near Biltmore, July 5 and September 16, 1897 (ex herb. Biltmore 1030 B). TENNESSEE.—Johnson City, Washington County, Gray, Sargent, Redfield, and Canby, June 21, 1877. ViRGINIA.—Farmer Mountain, on New River, Cornell County, J. K. Small, July 12, 1892, “‘altitude 2200 feet.” This tree has long been confounded with T. heterophylla and its variety Michauxii. From these trees it differs in its larger leaves generally more ob- lique at base, covered below with a denser, always silvery white, tomentum, its longer petioles, its fewer flowered corymbs and in its larger flowers which are larger than those of the other American lindens. It differs, too, in its stouter branchlets, and in the winter-buds which are red, compressed, and much larger than those of other American lindens. At Highlands, North Carolina, where this and T. heterophylla var. Michauxii are common at altitudes between 400 - and 600 m., T. monticola flowers 10 or 12 days later than the other tree. The specimen from Johnson City, Tennessee, although it is from a much lower 510 BOTANICAL GAZETTE [DECEMBER altitude than the others, is typical of the species, with leaves very oblique at base and up to 17 cm. long on the flowering branches; the petioles vary from 6 to 8cm.in length. The pedunculate bract is1t.5 cm.inlength. At this low altitude the trees naturally bloom earlier than at Highlands. T. monticola, with its large leaves snowy white on the lower surface and drooping gracefully on their long petioles, and its large flowers, is the showiest of the American lindens 15. Tilia georgiana, n. sp.—Tilia pubescens Ventenat, Ann. Hist. Nat. 2:62. 1800; Mém. Acad. Sci. 4:10. #. 3 (not Aiton). 1802.—Leaves ovate, slightly unsymmetrical at base and usually cordate on lateral branches and often oblique or truncate on leading branches, abruptly short-pointed at apex, and finely dentate, with glandular teeth pointing forward; when they unfold deeply tinged with red, covered above by fascicled hairs and tomentose below; when the flowers open dark yellow-green, dull and scabrate above and covered below with a thick coat of tomentum, pale on those of the lower branches and tinged with brown on those from the top of the tree, conspicuously reticulate-venulose, and at maturity thick, dull yellow-green, pubescent or glabrous above, rusty or pale tomen- tose below, sometimes becoming nearly glabrous in the autumn, 6-10 cm. long and 5-8 cm. wide; petioles slender, tomentose, 2-4 cm. in length. Flowers 6-7 mm. long, on slender pubescent pedicels in compact, slender-branched, pubescent, mostly 10-15- flowered corymbs; peduncle slender, pubescent on the lower, nearly glabrous on the upper, surface, the free portion 2. 5-3 cm. in length; sepals ovate, acuminate, coated on the outer surface with pale pubescence and on the inner surface with pale hairs longest and most abundant at the base, not more than one-half the length of the lanceolate acuminate, narrow petals; staminodia oblong- obovate to spathulate, acute, about two-thirds as long as the petals; style glabrous or furnished with a few hairs at the very base. Fruit on pubescent pedicels, depressed-globose, occasionally slightly grooved and ridged, covered with thick rusty tomentum, 5-6 mm. in diameter. A tree with slender branchlets thickly coated during their first season with - pale tomentum, and dark red-brown or brown and puberulous in their second year. Winter-buds covered with rusty brown pubescence, 6-7 mm. long. Flowers the middle of June. Fruit ripens early in September. 3 1918] SARGENT—TILIA 511 SoutH Carorina.—Near Charleston, T. G. Harbison, September 4, 1916 (no. 16). Grorc1A.—Colonel’s Island, near Dunham, Liberty County, T. G. Harbi- son, September 9, 1916 (nos. 4, 5, 8, 9), June 19, 1917 (no. 19); Brunswick, Glynn County, T. G. Harbison, May 24, June 19; September 2 and 3, 1916 (nos. 6, 7 type, 10, 11, 13, 15). FLorma.—San Mateo, Putnam County, A. H. Curtiss (no. 401a), Gains- ville, Alachua County, T. G. Harbison, June 10 and September 10, 1915, June 21 and September 14 and 15, 1916, April 24 and 25 and June 15, 1917; Lake City, Columbia County, G. V. Nash, July 11-19, 1805, T. G. Harbison, June 14, 1915, September 16, 1916, April 22 and June 23, 1917; Sumner, Levy County, TI. G. Harbison, June 12, 1915, September 12, 1916, April 25, June 15 and Sep- tember 25, 1917; Tallahassee, Leon County, T. G. Harbison, April.14, 1916; Crawfordville, Wakulla County, R. M. Harper, June 19, 1914 (no. 211); Mari- anna, Jackson County, T. G. Harbison, September 19, 1916 (no. 8), April 20 and May 26, 1017. What is perhaps best considered a variety of this species may be de- scribed as— TILIA GEORGIANA var. crinita, n. var.—Tilia pubescens Sargent, Silva N. Am. 155, ¢. 26 (in so far as relates to South Carolina, not Aiton); Man. 675. fig. 55. 1905.—Differing from the type in the longer and more matted, usually rusty brown hairs of the pubes- cence, usually less closely attached to the under surface of the leaves and often very conspicuous on the young branchlets. SoutH CaRoLina.—Sandy woods, Bluffton, Beaufort County, J. H. Melli- champ, May 28, 1887; near Charleston, T. G. Harbison, September 6, 1915 (no. 13). Grorc1A.—Colonel’s Island, near Dunham, Liberty County, Miss Julia King, July 1915, T. G. Harbison, September 8, 1916 (nos. 1, 2). This linden has a general resemblance to T. Houghii Rose, which differs in its rather looser pubescence and large and conspicuous tufts of hairs in the axils of the veins. Moreover, it hardly seems possible that a tree known only at a few stations on the coast of South Carolina and Georgia should also grow south of the City of Mexico, and so far as is now known nowhere else. ARNOLD ARBORETUM Jamaica Prarn, Mass. THE PURPLE HYACINTH BEAN GEoRGE F. FREEMAN , (WITH SEVEN FIGURES) Students of the genus Dolichos are now somewhat perplexed concerning the identity of D. lignosus Linn. and D. Lablab Linn. According to accepted usage, the former name applies to the small- leaved perennial vine sparingly grown as a greenhouse climber in northern climates and for arbors and trellises in warmer countries. Curtis (Bot. Mag. 1797, p. 380) states that it is perennial in Eng- land. D. Lablab is generally understood to refer to the common hyacinth bean or Bonavist, which has large purple leaves and racemes of showy purple flowers and seeds which are mottled mahogany brown to black. There are a number of varieties of this species, some of which have white flowers, white seeds, and green leaves. The size of the seed and the length and compactness of the racemes also vary strongly in the different kinds. This plant is strictly annual in the United States. It is used as an ornamental climber for porches, summerhouses, etc. Now Prain (Jour. Asiatic Soc. Bengal 667:429-430. 1897) reverses the incidence of these names and makes D. Lablab refer to the perennial species and D. lignosus to the annual hyacinth bean. In the 1895 edition of Index Kewensis, JACKSON does not recog- nize Dolichos lignosus as a valid species, but makes D. lignosus Jacq. Select. Am. 205 equal D. Jacquini DC. Prod. 2:397, Ind. occ.; and he makes D. lignosus Linn. Sp. Pl. 726 equal D. Lablab Linn. Sp. Pl. 725, Reg. trop. Again, Prrer (U.S. Dept. Agric. Bull. 318. 1915. p. 5), evidently following Pratn and JACKSON, accepts the validity of D. Jacquini DC. Prod. 2:397 and assigns to this species the small perennial variety of Dolichos formerly grown in various parts of the world as D. lignosus Linn. These references to LINNAEUS are to the edition of 1753, which is now the recognized beginning date of the binomial nomenclature. Evidently Linnaeus considered these two species as distinct. If Botanical Gazette, vol. 66] [$12 1918] FREEMAN—HYACINTH BEAN 513 we are to follow the Index Kewensis in this matter we must assume that LinNAEus was mistaken and that he had only two divergent forms out of the many varieties into which we now know D. Lablab to be subdivided. The evidence available, however, does not support this view, but indicates rather that the plants from which these two species were described were really specifically distinct. The original descriptions from Species Plantarum (pp. 725, 726) are aS follows: t. DOLICHOS leguminibus ovato-acinaciformibus, seminibus ovatis. Lablab hilo arcuato versus alteram extremitatem. Roy. lugdb. 368. Hort. ups. 214 Phaseolus aegyptius, nigro semine, Bauh. pin. 341. Phaseolus niger Lablab. Alp. aegypt. 74. t. 75. Vest. aegypt. 27. Habitat in Aegypto. : Legumina dorso scabra. Caules ramique ¢erefes, retrorsum scabri. Pedunculi semiverticillati. DOLICHOS caule perenni, pedunculis capitatis, leguminibus lignosus strictis linearibus. Dolichos caule perenni lignoso. Hort. cliff. 360. t. 20. Phaseolus indicus perennis, floribus purpurascentibus. Eichr. carol. 36. Habitat . . . Of the identity of D. Lablab L. we have no doubt. The original description corresponds exactly with the plant grown today under the name hyacinth bean. This is further confirmed by the presence in the herbarium of the Linnean Society of London of a specimen of.this plant identified and written up by Luynazvs himself. In fig. 1 is given a tracing of this specimen which was very kindly furnished by the general secretary of the Linnean Society.' The identity of this plant with the common hyacinth bean is evident. Compare fig. 2, which is a photograph of a specimen grown by the writer. The references given by Linnarus to “Bauh. pin. 341” and “Alp. Aegypt. 74 t. 75’’ have been examined and leave no doubt but that they refer to the common hyacinth bean. The reference to “Roy lugdb. 368, Hort. ups. 214” has not been avail- able. - * This tracing was kindly obtained for the writer by Dr. Oakes Ames of Har- vard University. 514 BOTANICAL GAZETTE [DECEMBER The confusion has evidently arisen on account of the uncer- tainty as to the identity of the plant described by LINNAEUS as D. lignosus. There is no cut accompanying this description, but LINNAEUS refers to an earlier publication by himself (Hort. Cliff. Gr Macecte tant At Rime Reden fe Fic. 1.—Tracing of specimen of Dolichos may identified and described by LrynaeEvs; from specimen in herb. Linnean Societ 360. t. 20) and to “Eichr. Carol. 36.” This latter publication has not been available, but the former has been examined carefully. Fig. 3 is a reproduction of a photostat copy of LINNAEUS’ figure in Hort. Cliff. 360 t. 20. The description accompanying this plate was much more full and complete than that in the Species Plantarum of 1753, and moreover was evidently made from a 1918] FREEMAN—HYACINTH BEAN 515 . 2.—Purple hyacinth bean (Dolichos Lablab): photograph of specimen grown at Hilec oe of Arizona, 1914. 516 BOTANICAL GAZETTE [DECEMBER fresh specimen before him. This description may be quoted as follows: 2. Dolichos caule perenni lignoso. Vide Tab. Phaseolus indicus perennis, floribus purpurascentibus. Hort. Carolsrb. 36. Crescit in America. Ante accessum nostrum enata fuit planta frutescens arcte scandens, plus quam homanae altitudinis; caule tereti, contorto, vix striato, ramis plurimis tenuibus. Folia ad ramorum -exortum ternata, petiolo communi insidentia, quorum quod intermedium ovato-cordatum, acuminatum, latitu- dine pollicis, glabrum, petiolo proprio quaduplo reliquorum productiori insidens; lateralia latere exteriori magis dilatata, interiori vero dimidio angustiora. Flores in pedunculo pauci, corolla rubra seu purpurea. Abso- luta florescentia absque fructu periit. It should be noted in this description that although the plant was 6 ft. or more, the leaves were but 1 inch wide and are recorded as being smooth. This plant, which was probably the only speci- men of D. lignosus actually seen by Linnarus, bloomed freely but did not set seed. Liynarus therefore probably never saw the seeds or pods of this species, but quoted the descriptions of these organs in his later publications from descriptions by other authors of plants which he assumed to be of the same species. Here in all probability lies the source of confusion. LinNAEus had observed that his D. Lablab was an annual in Europe and did not know that in warmer countries this same species may persist as an herbaceous perennial. When therefore he met with a Dolichos which was described as perennial, he would naturally be inclined to associate it with a species of Dolichos which he knew to be perennial, that is, his own D. lignosus. Thus he made the error in the 1763 edition of Species Plantarum (p. 1022) of citing Phaseolus perennis of Riimph. Amb. 5, pp. 378 t. 136, as a synonym of his D. lignosus, although this species (which is clearly D. Lablab) is described by Rtmputvs, in the publication named, as having leaves 3-4 inches long and nearly as broad, with racemes 1 ft. long and bearing many flowers. Ritmpuivs’ plate, a reproduction of which is given in fig. 4, could scarcely be assumed to represent the same plant as that of D. lignosus in Hort. Cliff. 360. ‘The only point of similarity is in N—HYACINTH BEAN Fic. 3.—Dolichos lignosus Linn.: from photostat copy of Livnagus’ figure in Hortus Cliffortianus. 518 . BOTANICAL GAZETTE [DECEMBER the flower clusters, and here the descriptions show these to be | entirely distinct (see legend to fig. 4). Turning to the descriptions and plates left by other botanists of the immediately succeeding decades we find in JACQUIN Selec- tarum stir pinum Americanarum historia (1763) a plant described as D. lignosus: Citations to the D. lignhosus of Linn. Sp. Pl. 726 and Linn. Hort. Cliff. 360. t. 20 are given with question marks, indicat- ing that the author doubted that the plant he described was the same as that described by LinnaEus. An examination of his description makes it clear that the two plants were entirely differ- ent, for the plant of Jacquin had pilose stems and pods, peduncles shotter than the scabrous leaves, white flowers, and pods 3-4 inches long, containing about 18 seeds, whereas, as we shall see, the plant described by LinnaEus had nearly smooth stems, pods, and leaves, peduncles longer than the leaves, purple flowers, pods 1-2 inches long with 7 or 8 seeds at the most. Arron (Hort. Kew. 3:31, 33. 1789) recognizes both D. Lablab L. and D. lignosus L. and gives practically the same descriptions as are given by Linnarus. He states that D. lignosus was introduced into England in 1776 by Monr. THOUIN. In 1792 Situ (Spicilegium Bot. no. 2, Gleanings of Botany, pp. 19 and pl. 21) describes and pictures a plant which he calls D. lignosus. Smrru’s plate is here reproduced in fig. 5 and his description is so clear and concise that it is quoted in full as follows: TABLE XXI . Dolichos lignosus. Purple woody Dolichos. Diadelphia Decandria. Stigma downy. GEN. cHAR. Standard marked at its base with two parallel oblong tubercles, compressing the under side of the wings. Section 1. Climbers SPEC. CHAR. Stem climbing, perennial. Flowers in little heads. Pods straight, linear. Syn. Dolichos lignowns Linn. Sp. Pl. 1022. Hort. Cliff. t. 20. Ait. Hort. Kew. V. 3.33. A native of the East Indies. Root woody, perennial. Stem woody, supple, climbing, much branched, roundish, striated, smooth; branches alternate, very long and slender, but FREEMAN—HYACINTH BEAN Fic. 4.—Phaseolus perennis Riimph. Amb. 5. 378. t. 136: the description accom- panying this plate states that the racemes are 1 ft. long and many-flowered; plant here represented is undoubtedly D. Ladlab Linn, 520 BOTANICAL GAZETTE [DECEMBER little subdivided, round, striated, somewhat downy, leafy, many-flowered. Leaves alternate, on long footstalks, ternate, or rather binate with an odd one. ee ge & 9 4 Fic. 5.—Dolichos lignosus: reproduction of pl. 21, J. E. Sura, Spicilegium Bot. no. 2, -Gleanings of Botany. 1792. 1918] FREEMAN—HYACINTH BEAN 521 Common footstalk roundish, channeled above, swelling and purplish at the base; partial ones very short, swelled, incurved. Leaflets rhomboid, elongated, acute, entire, obsoletely 3-nerved; bright green and shining above; glaucous beneath. Stipulae entire, sharp, somewhat triangular, downy on the margin, dark purple at the base; of which 2 larger ones are placed at the bottom of the common footstalks, and 2 smaller, lanceolate, at the insertion of the partial footstalks. Clusters axillary, solitary, erect, each having from 3 to 6 flowers in a little head. Common flowerstalk simple, very long, striated, angular in the upper part; partial ones generally 2 together, short, downy, single-flowered. ¥ Fic. 6.—Dolichos lignosus: from photograph of colored plate in Curtis’ Bot. Mag. 2:380. 1797 Bracteae lanceolate, acute, hairy. Flowers somewhat eA rose hee sen with a purplish keel. Calyx smooth, thickly ciliated in the m Pod an inch long, a little recurved, brownish, smooth. Seeds black. According to ArTon, this beautiful plant was introduced from the French gardens to our own in 1776. It is easily propagated by seed, and in a stove produces abundance of flowers during the summer. A little study of Smrrn’s plate and descriptions shows that it agrees very closely with the plate and description of D. lignosus of LINNAEUs and that it cannot possibly be the D. lignosus of JAcQUIN. Five years later, in Curtis’ Bot. Mag. 11:380. 1797, is found a description and plate of D. lignosus, which is reproduced in fig. 6. 522 BOTANICAL GAZETTE [DECEMBER Comparing the plate and description with those of SmirH and LINNAEUS, there is no doubt that they all had the same plant. Finally, fig. 7 is made from a fresh specimen of plants grown at the University of California. This plant agrees perfectly with Sm1TH’s description and with every essential part of the description and +s Fic. 7.—Dolichos lignosus: from photograph of fresh specimen furnished by Professor GREGG, University of California, January 1914. plate by Linnarvs except that in which he states that the pods are straight. When we remember that LINNAEUS probably never saw the pods of this species, such a discrepancy is not surprising. Misled by the error of Livnarus in ascribing straight pods to his D. lignosus, DECANDOLLE (Prod. 2: 397. 1825) makes the D. lignosus of Curtis (Bot. Mag. t. 382?) a variety of D. lignosus Linn. He moreover corrects the error made by Jacquin in assign- * This is an error by DECANDOLLE, and should read 380. 1918] FREEMAN—HYACINTH BEAN 523 ing his plant (described in Select. Stirp. Amer. Hist. 1763, p. 205) to D. lignosus L. by calling this plant D. J ‘acguint. To emphasize the justice of this disposition of the 2 species by DECANDOLLE, the original description furnished by JAcQuIN may be quoted as follows: Planta perennis, volubilis, tota pilosa; praecipue vero rami inferiores lignosi, and legumina, pilis hispida sunt. Foliola sunt ovata, acuta, scabrius- cula, duos pollices longa, lateralibus interne obliteratis. Stipulae ex lanceolato Ovatae, acuminatae, basi emarginatae. Pedunculi umbellati, foliis breviores, pauciflori. Flores albidi. Legumina tres quatuorve pollices longa, acuminata, interne nivea. Semina circiter rere nitida, atra cum hilo albido, parva, compressiuscula, ex oblongo reniformi abitat in Caribaearum iecade: It would be difficult to harmonize this description with that of either SmirH (Spic. Bot. no. 2, p. 19) or LinnaEus (Hort. Cliff. 360. t. 20). We must couclane. therefore, with DECANDOLLE, that it is a distinct species and follow him in calling it D. Jacquini DC. Prod. 2:397. In the opinion of the writer, the evidence presented herewith is sufficient to show that the plants described as D. lignosus by LinnaEvs (Sp. Plant. ed. 1, 1753, p. 726), and more fully in his earlier work (Hort. Cliff. 360. t. 20. ian by J. E. Smaru (Spic. Bot. no. 2, p. 19, pl. 21. 1792), by Curtis (Bot. Mag. 11:380. 1797), and the plant now grown in various parts of the world as D. lignosus and shown in fig. 7 are all one and the same species, which is distinct from D. Lablab L. We are therefore unable to follow either Jackson (Index Kewensis 1895) in making D. lignosus L. a synonym of D. Lablab L; Pratn (Jour. Asiatic Soc. Bengal 667: 429-430. 1897) in reversing the incidence of the original Linnaean names by making D. Labdlab L. refer to the perennial species and D. lignosus L. to the annual hyacinth bean; or PIPER (U.S. Dept. Bull. Agr. 318. 1915, p. 5), in assigning the plant commonly grown as D. lignosus L. to D. Jacquini DC. On the other hand, we must hold to the original Linnaean designation of the common annual (frequently perennial in tropical countries) hyacinth bean (and its many varieties, fig. 2) as D. Lablab L., and the more slender perennial greenhouse (in northern climates) climber shown in fig. 7 as D. lignosus L. : AGRICULTURAL EXPERIMENT STATION U; SITY OF ARIZONA A MORPHOLOGICAL STUDY OF PALLAVICINIA LYELLII CONTRIBUTIONS FROM THE HULL BOTANICAL LABORATORY 245 ARTHUR W. Haupt (WITH PLATES XX—XXIV) Pallavicinia, according to SCHIFFNER’S (12) census, is repre- sented by 21 species, most of which are tropical. Later, SCHIFF- NER (13) added 2 European species to his former list, thus making 5 species indigenous to the Old World. Pallavicinia Lyellii is found in the more humid parts of both the Northern and Southern Hemi- spheres; it grows near Chicago in a peat bog at Mineral Springs, Indiana. Pallavicinia, Symphyogyna, and Monoclea are included in the family Leptotheceae. The affinities of the Japanese genus Makinoa, described by MrvaxeE (9), seem to place it in this family, as is done by Cavers. The disposition of Monoclea is a matter of great difference of opinion, some, as JOHNSON (6), placing it with the Marchantiales. There can be-no doubt, however, as to the closeness of relationship between Pallavicinia and Symphyogyna, regardless of the classification of the other genera of the family. CAVERS (2) divides Pallavicinia into the two genera of GOTTSCHE: Blyttia and Mérckia. According to the Vienna code, the older name Pallavicinia must be retained; if ScHIFFNER’s subgenus Morckia (Gott.) is to be elevated to ponkiie rank, it must not be done at the expense of the name Pallavicinia. STEPHANI (14) sepa- rates the genus into the 2 sections PROCUMBENTES and DENDROI- DEAE; SCHIMPER, into the subgenera Eupallavicinia, Morckia, and Mittenia. Pallavicinia Lyellii belongs to the PROCUMBENTES OF upallavicinia division. Material Most of the material studied was collected by Mr. R. P. MASON, at Columbiana, Alabama, to whom the writer is greatly indebted. Additional material was obtained by Dr. W. J. G. Lanp, at Mineral Springs, Indiana. Most of the slides illustrating the antheridial, Botanical Gazette, vol. 66] [524 1918] HAUPT—PALLAVICINIA 525 series were made by Dr. LAND and Mr. Mason, while those showing the development of the archegonium and the sporophyte were pre- pared by the writer. Thallus The vegetative body of Pallavicinia Lyellii consists of a creep- ing, prostrate thallus 4-5 mm. wide, composed of a midrib with thin, one-layered, lateral wings, and bearing rhizoids. The margin is somewhat undulate, with no indications of hooked appendages as in P. longispina, P. xiphioides, or P. Zollengeri. The midrib con- sists of pitted conducting cells with thickened walls, which become differentiated directly behind the apical cell (fig. 45); about 70 to 80 may be seen in cross-section (fig. 44). ANSLEY and CHIcK (15) made a careful study of these cells and showed by eosin solutions that they conduct water. Miss McCormick (8) demonstrated that in Symphyogyna aspera they are composed of pectose. These con- ducting cells are also found in Hymenophyton. Growth of the thallus is by means of a dolabrate (zweischneidig) apical cell (fig. 43). This feature seems to have first been observed by LerrcEs (7), who discusses at considerable length apical growth and the development of the thallus body. Two-celled mucilage hairs arise both dorsally and ventrally in connection with the apical cell, strongly resembling sex organ initials. Branching is of two kinds: apical, from the apical cell; and endogenous, from ventral adventitious shoots. Material showing the origin of the latter was lacking and hence LeITcEB’s statement, that the conducting tissue of the ventral branch is not continuous in origin with the central cells of the main thallus, could not be verified. Sex organs The gametophytes of Pallavicinia Lyellii are strictly dioecious, the male plants being slightly more slender than the female. Both antheridia and archegonia are dorsal, the former lying in 2 parallel rows on each side of the midrib, and the latter remaining directly above the midrib, slightly raised on a pad. Two involucres are present, the outer one corresponding to that of Symphyogyna and Monoclea; the inner one, or perianth, is characteristic of Pallavi- cinia, Podomitrium, and Calycularia. 526 BOTANICAL GAZETTE [DECEMBER ANTHERIDIA The antheridia originate in close proximity to the apical cell, arising in acropetal succession on the dorsal side of the thallus, usually singly, but occasionally two or three together. With further apical divisions they come to lie in 2 parallel rows on each side of the midrib, slightly sunken in the thallus by the develop- ment, from behind, of an involucral upgrowth. The mature antheridia are spherical and short-stalked and point diagonally outward and upward, each one being separated from the one pre- ceding it by sterile tissue (fig. 1): An antheridium initial appears as a papillate projection above the surface of the thallus, resembling closely one of the mucilage hairs with which it is associated. A transverse wall appears, dividing the initial into 2 nearly equal segments, the basal one remaining in the thallus and the outer one projecting above the surface of the thallus (fig. 2). The outer cell divides transversely into equal segments, forming a primary stalk cell and a primary antheridial cell (fig. 3). With further increase in size, the latter. divides by a median vertical wall, followed rapidly by a similar division in the stalk cell (fig. 4). One or two further transverse divisions complete the stalk, while a periclinal wall cuts off a pe- ripheral cell on one side of the antheridium, intersecting the first vertical wall near the top (fig. 5). A corresponding periclinal wall appears on the other side, followed by 2 more walls at right angles to the first two, intersecting both these and the first median division. As a result, 4 primary wall cells inclose 2 central cells, the entire structure being bisected by the original vertical wall. At this stage the involucre appears as an upgrowth of the thallus behind the young antheridium (fig. 9). It is built up by basal growth, and by the time the antheridium is mature, it consists of a scalelike sheath, 6-10 cells in length. Whether these coverings are to be regarded as the beginnings of true foliar structures or merely as dorsal upgrowths of the thallus seems to be entirely a matter of opinion. If the complete involucre of Pellia be taken as representing the initial stage, a failure of the forward portion to develop would result in the precise condition found in Pallavicinia. Sphaerocarpus, pet- haps, represents an intermediate stage, as here the development of 1918] HAUPT—PALLAVICINIA 527 the forward portion is slightly arrested, resulting in greater protec- tion from behind. It seems a perfectly logical step from the antheridial condition in Pallavicinia to that of one of the simpler acrogynous Jungermanniales, such as Porella, in which case the coverings are longer and more leaflike in appearance. Further growth of the antheridium corresponds to that of the other anacrogynous Jungermanniales. During the 2 or 3 mitoses preceding the formation of the sperm mother cells, the cell walls of the spermatogenous mass gradually disappear and abundant muci- lage surrounds the dividing protoplasts. Walls around the sperm mother cells were evident, but it could not be determined whether they had been laid down by the mother cell protoplasts, or rep- resented the remaining cellulose which had not become mucilage. * The sperm mother cells produce two sperms, each with little cyto- plasm, and separated by a very thin wall. The nuclei were so small that it was not possible to study the details of spermatogenesis. The development is probably the same as that of Pallavicinia Zol- lengeri, described by CAMPBELL and WILLIAMs (1). ARCHEGONIA The earliest stages in the development of the archegonial group were not present in the material studied. A group of initials seems to arise a short distance back of the apical cell, directly above the midrib on the dorsal side of the thallus. This group presently becomes surrounded by an annular upgrowth of the thallus, which becomes the involucre. The apical cell is not checked by the development of the archegonia, but continues the growth of the thallus, so that often 2 or 3 groups may be produced along the mid- rib, separated by sterile areas. The archegonial group continues to produce archegonia up to the time of fertilization, many young sex organs frequently being found with mature ones. Two-celled mucilage hairs are abundantly produced. Twenty to 30 arche- gonia usually occur in a group. The archegonium, like the antheridium, arises as a papillate projection from one of the cells inclosed by the involucre. A trans- verse wall cuts off a basal cell, which remains within the thallus, and an outer cell, which is freely exposed (fig. 10). The latter 528 BOTANICAL GAZETTE [DECEMBER undergoes 2 transverse divisions, the sequence of which could not be determined; the lower 2 cells form the stalk and the upper one the archegonium proper, agreeing in this repsect with Pallavicinia radiculosa, described by CAMPBELL and WILLIAMS (figs. 11,~12). Three vertical divisions occur in the archegonium proper, accord- ing to the manner of all anacrogynous Jungermanniales, resulting in the formation of an inner cell surrounded by 3 primary wall cells, 2 of which can be seen in a longitudinal section (fig. 12). A trans- verse division in the upper part of the inner cell results in the formation of a central cell and a cap cell (fig. 13), which later under- goes further division, contributing to the development of the neck. Following the formation of the cap cell, the central cell divides into two nearly equal cells (fig. 14), the upper being the primary neck canal cell, and the lower the primary ventral cell. The devel- opment of the axial row usually precedes the division of the primary ventral cell, although frequently mitoses can be seen in the neck cells after the formation of the ventral canal cell and egg (fig. 22). In most cases about 10 neck canal cells were seen; sometimes, how- ever, as many as 18 are formed (fig. 25). The primary ventral cell, by a transverse division, produces a ventral canal cell and egg which are almost equal in size (fig. 18). The neck canal cells are surrounded by a jacket of 5 cells, although frequently one or more of these may divide (fig. 24). Very soon after the division of the ventral cell the ventral canal cell becomes mucilaginous and finally the entire axial row is broken down (figs. 19, 20). The egg nucleus at this stage is very promi- nent, the dense nucleolus being surrounded by extremely light nucleoplasm. With the preparation of the egg for fertilization, the wall of the venter becomes 2-layered, the first divisions occurring as the ventral canal cell begins to disorganize. The mature arche- gonium is characterized by a rather long slender stalk, a narrow venter, and a long twisted neck; it closely resembles an arche- gonium of Symphyogyna. Just before the older archegonia in a group mature the charac- teristic perianth appears immediately within the involucre. It attains a height of several cells (fig. 21), but as soon as fertilization is effected, it is greatly stimulated, and develops, by basal inter- 1918] HAUPT—PALLAVICINIA 529 calary divisions, much in excess of the young sporophyte. The perianth keeps pace with the elongation of the embryo, reaching a maximum height of about 5 mm. At the time immediately pre- ceding spore disperal, the seta shows remarkable growth, becoming 25-27 mm. long. The perianth attains a thickness of 3 or 4 cells, as seen in cross-section, and becomes fringed around the top. The involucre, on the other hand, does not at any time exceed the height of the archegonia and it is related to their protection in the same way that the perianth is associated with the protection of the sporo- phyte. With the development of the perianth, following fertiliza- tion, the involucre becomes flaring and denticulate around the top. After fertilization the egg cytoplasm becomes denser and a heavy wall is laid down around the protoplast, thus making it independent of the tissue of the archegonium. Sporophyte The youngest sporophyte which was observed consisted of a tier of 4 cells (fig. 26). The first division is followed by a transverse wall in the lower segment and then by a similar wall in the upper segment. A vertical division then occurs in the upper half of the embryo (figs. 27, 28), followed by vertical and transverse walls. The lower half of the embryo usually undergoes one vertical divi- sion, but contributes nothing to the development of the foot, seta, or capsule (fig. 29). Half of the potentially sporogenous tissue derived from the fertilized egg thus is diverted for haustorial pur- poses. A similar situation has been observed by Miss Ciapp (3) in Aneura pinguis, and by CAMPBELL and WILLIAMs in Pallavicinia Zollengeri. The relation between the early divisions in the embryo and the development of the 3 regions of the sporophyte could not be ascertained, material being wanting. The differentiation of the sporogenous tissue, however, occurs relatively late. According to FARMER (4), the young embryo of Pallavicinia decipiens consists of a tier of 3 cells, the upper segment forming the capsule, the middle segment the seta and part of the foot, and the lower segment the rest of the foot. With the growth of the embryo, the venter of the archegonium becomes a calyptra 4 or 5 cells in thickness, which grows in length 530 BOTANICAL GAZETTE [DECEMBER with the sporophyte. The calyptra is notably smaller than in Symphyogyna and Monoclea, presumably because its protective function is performed by the perianth. The non-functioning arche- gonia are carried up with the tissue of the calyptra but do not per- sist long. Only one embryo was seen developing in an archegonial group, although it is possible that more than one may be formed, as in Symphyogyna aspera. The differentiation of the spores and elaters occurs late in the development of the sporophyte, and follows precisely the method of Symphyogyna aspera, as described by Miss McCormicx (figs. 33- 38). Material showing the reduction division in the formation of spores was entirely absent in the material studied. FARMER (5), in his study of this process in Pellia epiphylla and Pallavicinia decipiens, noted the presence of a quadripolar spindle in the spore mother cell. Moore (10, 11), however, working with Pallavicinia Lyellii, failed to find such a condition, but observed that the two divisions take place in very rapid sequence, giving the appearance of such a spindle as FARMER describes. The mature capsule is cylindrical, is inclosed by a sterile wall one cell thick, and bears spiral thickenings (figs. 32, 39). The sterile cap at the apex of the mature capsule is not so prominent as in Symphyogyna, being only 5 or 6 cells thick, and bears no rela- tion to the elaters. The mature sporophyte reaches a length of 40 mm., the capsule being about 3.5 mm. long. Dehiscence is by means of 4 longitudinal slits which remain attached at the top. The foot is wedge-shaped as in Symphyogyna, but it occasionally shows a resemblance to the anchor-like foot of Marchantia (figs. 41, 42). The mature elaters reach a length of nearly o.3 mm., and are furnished with a double spiral band. The spores are about se 5 mm. in diameter, the wall being conspicuously reticulate fig. 38). Summary 1. Pallavicinia Lyellii belongs to the subgenus Eupallavicinia, _ the vegetative body consisting of a single prostrate portion. 2. The apical cell is of the dolabrate type. Branching is both apical and adventitious. , 1918] HAUPT—PALLAVICINIA 531 3. Pallavicinia Lyellii, like the other species of the genus, is strictly dioecious. 4. The antheridia occur in 2 parallel rows on each side of the midrib, and are protected from behind by an involucral upgrowth. Their development, with minor variations, follows the type for the anacrogynous Jungermanniales. 5. The archegonia are in dorsal groups and are surrounded by an involucre and a perianth, the latter remaining inconspicuous until after fertilization. 6. The young archegonial stalk consists of 2 cells. The egg is small and the neck long and twist 7. The lower half of the fentilind egg becomes a haustorial organ and contributes nothing to the development of the foot, seta, or capsule. 8. The calyptra is 4 or 5 cells in thickness, in this respect differ- ing from that of Symphyogyna. 9. The differentiation of the spores and elaters occurs relatively late in the development of the sporophyte, and follows the method of Symphyogyna. 10. Asterile cap is present at the apex of the capsule and remains intact in dehiscence, which is accomplished by means of 4 longi- tudinal slits. To Dr. W. J. G. Lanp, under whose direction the study was undertaken, the writer is indebted for many helpful suggestions and criticisms. University oF CHICAGO LITERATURE CITED 1. CAMPBELL, D. H., and Wrrt1ams, FitoreNnce, A morphological study of some members of the aps ees Leland Stanford Junior Univ. Pub., Univ. Series, pp. 44 2. Cavers. F., The Ss eas of the ran III. Anacrogynous Jurigserinscitialae: New Phytol. 9:197-234. 3- CLapp, Grace L., The life history of Aneura peasy Bot. GAZ. 54:177- 4. FARMER, I. BRETLAND, Studies in peetenes on Pallavicinia decipiens Mitten. Ann. Botany 8:35-52. 1894. 532 5. 6. 7: ag al fe pal on BOTANICAL GAZETTE [DECEMBER FARMER, J. BERTRAND, The cise spindle in the spore mother cell of Pellia epiphylla, Ibid. 15:431-433. Jounson, D. S., The development ma velationihin of Monoclea. Bor. GAZ. 38: {8e-a0% 1904. LertceB, HusBert, Untersuchungen iiber die Lebermoose, “ae 3, Die frondosen Jungermannien. Leipzig. 1877. cCorMICK, FLoRENCE A., A study of Symphyogyna aspera. Bor. Gaz. 58:401-418. 1914. Mrvake, K., Makinoa, a new genus of Hepaticae. Bot. Mag. Tokyo 13: 21-24. 1890. Moore, A. C., The mitoses in the spore mother cell of Pallavicinia. Bot. Gaz. 36: 384-388. 1 1903. , Sporogenesis in Pallavicinia. Ibid. 40:81-96. 1905. ScurreNReE, V. ERG in ENGLER and PRANTL’s Natiirlichen Pflanzen- familien 1:1-144. -Daiereadienas iiber Mérckia Flotowiana und iiber das Verhaltnis der Gatiunees Mérckia Gott. und Calycularia Mitt. zu einander. Oesterr. Bot. Zeitschr. 51:41-51. 1 . STEPHANI, F., Species Hepaticarum. Memoires de L’Herbier Boissier. 1900. . TANSLEY, A. G., and Cuicx, Ep1tu, Notes on the conducting tissue system in Bryophyta. Ann. Botany 15:1-38. 1901. EXPLANATION OF PLATES XX-XXIV Pallavicinia Lyellit Fic. 1.—Thallus with antheridia; X1o. Fics. 2~7.—Stages in development of antheridium; 630. Fic. 8.—Older antheridium; X62. Fic. 9.—Young antheridium with involucre; 46. Fics. 10-20.—Stages in development of archegonium; X 520. Fic. 21.—Perianth before fertilization; Fic. 22.—Archegonium showing elongation of neck; X 50. Fic. 23.—Median longitudinal section of archegonial group; X25. Fic. 24.—Cross-section of neck of archegonium; X 520. 1G. 25.—Mature archegonium showing 18 — — des: X 260. Fics. 26-28.—Stages in development of emb: Fic. 29.—Older embryo showing haustorial ral nit differentiation of Sporogenous tissue; 50; sketch of same stage; X8. Fic. 30 i Deubowas of sporogenous tissue; X50; sketch of same stage; <8. Fic. 31.—Sketch of ‘idee sporophyte; <6. Fic. 32.—Sporophyte containing mature spores and elaters; <6. Fics. 33-37.—Stages in development of spore mother cells; X 520.- BOTANICAL GAZETTE, LXVI PLATE XX PLATE XXI BOTANICAL GAZETTE, LXVI DUT a ITY HAUPT on PALLAVICINIA BOTANICAL GAZETTE, LXVI PLATE XXII HAUPT on PALLAVICINIA PLATE XXIII BOTANICAL GAZETTE, LXVI ag tt ii as eo) TTR Puc aiy ‘e - a) aoa *, a if 39 40 HAUPT on PALLAVICINIA BOTANICAL GAZETTE, LXVI ; PLATE XXIV 42 HAUPT on PALLAVICINIA 1918] HAUPT—PALLAVICINIA 533 Fic. 38.—Mature spores and elater; 630. Fic. 39.—Wall of mature capsule and calyptra; X30 Fic. 40.—Median longitudinal section of sterile cap of capsule in mother cell stage; X50. Fic. 41.—Typical foot of mature sporophyte; X30. Fic. 42.—Anchor-like foot of sporophyte; X15. Fic. 43.—Median longitudinal section of apical cell; X50. Fic. 44.—Cross-section of cells in conducting bindandl X 520 Fic. 45.—Median longitudinal section of region directly bark of apical cell showing differentiation of conducting cells; 520. G. 46.—Conducting cells showing pits; 520. Fic. 47.—Part of same; 1040. BRIEFER ARTICLES MODIFICATION OF HAND MICROTOME (WITH FIVE FIGURES) Figs. 1-5 represent a simple modification of the familiar hand micro- tome, and one which has been found to be a decided improvement over the original instrument from which it was derived. In the ordinary type, when cutting sections of woody stems or more delicate material held in pith, it is always difficult to be certain of obtaining the necessary pressure for holding the material at the proper point. The steel rod which moves in or out upon the turning of the single pressure screw will usually hold the material firmly at its lower end but not so firmly at its upper end, with the result that the material has a tendency to wobble when the knife begins to cut the section. On the other hand, when this difficulty does not arise it is often almost impossible to screw up the material for the next section because of the pressure of the material against the walls of the tube or well. To obviate these rather commonly encountered difficulties in the ordinary type of hand microtome the modification of it shown in the figures was devised. Figs. 4 and 5 give two views of an inner “material holder.” It consists of two pieces of curved steel which are long enough to reach to the bottom of the tube or well (just below cc in fig. 2). This inner material holder is provided with a spreading spring at ed which surrounds a small steel bar cc. Each curved piece of steel has a hole at aa (fig. 4) through which project the ends of the two pressure screws 66 (fig. 2). The manipulation of the apparatus is as follows where, for example, cross-sections of a woody stem are to be cut. The pressure screws 6b are turned out until their ends at aa are pulled out of the holes in the material holder. The microtome is inverted and the material holder falls out. The stem or a portion of it is placed between the leaves of the material holder and properly oriented and, if necessary, a rubber band is bound around the material holder just above aa. The material holder containing the stem is now pushed down into the well _ or tube of the microtome and oriented so that the holes are opposite the ends aa of the turned-back pressure screws. These screws are a Gazette, vol. 66] «(534 1918] BRIEFER ARTICLES 535 turned in, their ends pass into the holes in the material holder, and pressure is finally exerted on both sides. As the pressure becomes VA. Beeraiz Frc. 1.—Modification of hand microtome * 536 BOTANICAL GAZETTE [DECEMBER greater, the spring at de prevents the upper ends of the material holder from spreading and insures maximum pressure against the material at these upper ends. Finally the stem is held firmly in the center of the tube or well of the microtome between the leaves of the material holder. The pressure screws are free to move up or down in their openings because no appreciable pressure is exerted upon the walls of the tube or well. In a similar manner material held in pith is very conveniently arranged in this apparatus. The possibility of arranging and orienting such material held in pith in the material holder outside the microtome is an obvious advantage. Longitudinal sections of small woody stems are readily cut in this modified hand microtome, whereas their small diameter makes it very difficult to secure them firmly in the original apparatus. As may be seen, it is possible to orient material to obtain all angles in the case of sections to be cut obliquely or in the case of unsymmetrical material. This modified hand microtome was devised to meet a special need and has admirably served its original purpose. This description of it is presented primarily because it illustrates the possibility of modifying an apparatus in a relatively simple and inexpensive manner to increase greatly its convenience and the range of its usefulness. The original modification from which the drawing was made has been somewhat improved recently. The knurled heads 6b should be much larger than those illustrated, and for woody stems the leaves of the material holder should be thicker and their inner surfaces more nearly flat.—T. H. GoopsPEED, University of California. CURRENT LIFERATURE MINOR NOTICES Fungous diseases and insect pests.—In a small volume issued as one of the Cambridge Farm Institute Series, PETHERBRIDGE' gives a popular account of the more common fungous diseases and insect pests of farm crops. The book is designed to be helpful to farmers and others who wish to acquire a knowledge of such things. The treatment is very elementary, but sufficiently extensive to give the uninitiated some idea of the nature of fungi and insects and their relation to agricultural crops. The text is nearly equally apportioned between the two main divisions of the subject-matter. The first division deals with fungi and fungous diseases, and the second with insect pests. ch division is introduced by a general chapter giving in each case a brief descrip- tion of fungi, their mode of life, and the part they play in crop economy; and in the second part a general account of the structure, life histories, and habits of insects. In the special chapters the plan is followed of describing in detail some of the representative types of fungi and insects, as for instance, Erysiphe graminis as an example of the mildews, and grouping around them others of similar nature. An idea of the scope of the work can best be gained from the chapter headings, as follows: Introduction to fungi; Potato diseases and allied diseases; Finger and toe, and wart disease; Mildews; Ergot and clover sick- ness; Rusts; Smuts; Introduction to insects; Butterflies and moths; Beetles; Flies; Aphids and sawflies; Eelworms. The book is written in a clear style and it will undoubtedly prove useful to the farmers of England in enabling them to identify the common insect and fungous diseases, and to find means of combating them. In the more exten- sive and diversified agriculture of the United States, where a vast special litera- ture dealing with each particular condition is already available to the farmer, the book would find little application.—H. HassELBRING. Flora of the Northern Territory of Australia—Ewart and Davies’ have published a flora of the large area known as the Northern Territory of Australia, not merely as a contribution to taxonomy, but also as an indication of “the fertility of the soil, the moisture conditions, and the fodder or other values of t Per , F. R., Fungoid and insect pests of the farm. Cambridge. 1918. 2 Ewart, S hereies ve on Davies, OLIvE B., The flora of the Northern Terri- - tory. ag wis. pls. 27. Melbourne. 1917. $37 538 BOTANICAL GAZETTE [DECEMBER the natural vegetation.” There are _ of plants of fodder value, valuable woods medicinal plants. Four appendices also deal with Cyperaceae, pHs nicoeot Eucalyptus), Eucalyptus, and Acacia. Four new genera are established by Ewart as follows: Spathia and Setosa (Gramineae), Rossittia (Rutaceae), and Carpentea (Convolvulaceae); and in addition 30 new species are described.—J. M. C NOTES FOR STUDENTS Phenomena of parasitism.—In a summary of his researches on the processes involved in the attacks of plant tissues by Botrytis cinerea, BRowN’ gives a review of the work already published and a forecast of investigations now in progress. The published work has already been noted in this journal,s and we need only allude to the author’s speculation on the question whether the effects produced by the fungous extract on the cell wall and on the proto- plasm are attributable to the same or to different substances. In the absence of any means of disentangling the mixture of substances occurring in plant ex- tracts or of excluding the action of all but one, it seems futile to speculate on the specificity of action of any of the substances. Future work as outlined by the author is to cover such problems as the germinating capacity of spores in water and in nutrient solutions, the diffusion of substances from age cells into water placed on the cuticle, and the physics of cuticular resistance The fourth contribution to this series’ deals with some £ the factors influencing the production of cytase in cultures of Botrytis cinerea. In the first paper of the series it was shown that very active cytolytic extracts could be obtained from young germ tubes of the spores of the fungus. As might be expected, therefore, the activity of the enzyme extracted from cultures of differ- ent ages is proportional to the quantity of actively growing mycelium. Con- sequently, with respect to enzymatic activity, a growing culture soon reaches a maximum, and thereafter the enzyme content rapidly diminishes. The enzyme content of the culture fluid follows a course in general parallel to that of the mycelium. Dilution of the enzyme extract by a similar extract deacti- vated by exposure to a temperature of 65° has the same effect as dilution by caused by the development of inhibiting substances. As might appear self- evident, cultures thickly sown with spores gave stronger enzyme extracts than cultures thinly sown. The experiments confirm the former conclusions that Acast * ton ka tc. ol ing ds of the hyphae.—H. HASSEL- BRING. 3 Brown, W., On the physiology of parasitism. New a hy tal. 16:109-126, 1917. 4 Rev. Bor. Gaz. 61:80. 1916; 63:240. 1917. 5’ Brown, W., Studies in the physiology of parasitism. -IV. On the distribution of cytase in cultures of Botrytis cinerea. Ann, Botany 31: ao-uek: 1917. : 1918} CURRENT LITERATURE 539 Phytogeography of South Africa.—The very diverse vegetational types of South Africa have been classified and mapped by Evans* in such a manner as to give a good idea of the ecological divisions of the southern part of that continent. The woodland has been subdivided into forest, scrub, bushveld, and palmveld. The first of these, which is mostly evergreen, is dominated by species of Podocarpus, while the scrub is a type of Sclerophyllous Shrub, in which the Proteaceae, Ericaceae, and Restionaceae contribute the dominant forms. From this the bushveld differs in its deciduous character and also in its more parklike aspect and its floristic composition. Bushveld is widely distributed, and while dominated by Acacia spp., such genera as Tamarix, Combretum, Ficus, Zizyphus, and Rhus are of common occurrence. The belt comprises a littoral strip on the southeast in which such palms as Mimusops caffra, Phoenix reclinata, Raphia vinifera, and Cocos nucifera mingle with succu- lents from the genera Aloe and Euphorbia. The grassland covers the greater portion of the country with transitions to scrub and desert. That of the Kalahari region occupies much of the central portion of South Africa with an open formation, short, low, wiry grasses, species of Aristida and Eragrostis, occurring in isolated tufts. This and the other grasslands show transitions to the desert toward the west. Four distinct desert types are briefly characterized and mapped, perhaps the most remarkable being the southern portion, a vast shallow basin, the Karroo, sparsely populated by succulent, tuberous, and bulbous plants. Prominent genera are Crassula, M pase on setae Cotyledon, Euphorbia, Aloe, Stapelia, Senecio, Encephalartos, and Eucl More important perhaps than the text, at nae for the American botanist, — are the excellent plates, enabling one to visualize the different types, and the map showing their distribution. Gro. D. FULLER. Pigment production in Penicillium.—BRENNER,’ investigating the pro- duction of pigment in cultures of Penicillium, finds that in the absence of magnesium in the culture medium, or in the presence of ammonium salts whose utilization leads to an acid reaction of the culture fluid, no red, but only yellow pigment is produced. The red pigment is produced only in neutral media or in media developing an alkaline reaction. Iron apparently is not necessary for the formation of the red color. The author further reports a few preliminary experiments on the extraction and chemical reactions of the pig- ment which is insoluble in ether, chloroform, toluene, and similar organic solvents, but soluble in alcohol and dilute alkalies or ammonia. On account of the acid nature of the pigment the author attributes to it the physiological function of maintaining the neutrality of the medium ‘ Evans, F. B. Pore, The plant geography of South Africa. mn % Agric. Union of sree ia Official Year Book. 1917. pp. 8. pls. 24. map. 1918. R, W., Die :epenhcetie bei Penicillium purpurogenum. Svensk. Bot. Takes 12: gI~1o2. 1918. 540 BOTANICAL GAZETTE [DECEMBER Many investigations have been made on the so-called influence of various environmental factors on the production of pigments by fungi, but a survey of the facts seems to indicate that between the absorption of an elementary nutrient and the production of a complex pigment two processes intervene to permit of the establishment of a direct relation between stages at the extreme ends of the series. A much better knowledge than is at hand at present of the nature and structure of fungous pigments is necessary before their physio- logical status can be determined. Different colors may often be due to the modification of the same pigment, depending on different reactions of the medium.—H. HASSELBRING. Origin and goal of geobotany.—Rtset’ has issued a compact and useful paper, dealing with the main phases of the development of geobotany and with the aims of its various subdivisions. Geobotany he regards as embracing all interrelations between plants and the earth, including much of ecology, cho- rology, chronology, and genetics; thus it includes all of phytogeography in the widest sense, and more. The historical presentation deals especially with the work of THEOPHRASTUS, TOURNEFORT, LINNAEUS, HALLER, SOULAVIE, WILLDE- now, HumBoLtpt, WAHLENBERG, and ScHouw. Geobotany may be either floristic or vegetational, each of which subdivisions may consider the problems of space (distribution), habitat (ecology), or change (genetics). Thus RUBEL recognizes 6 fields of geobotany: autochorology, or floristics; synchorology, or the distribution of plant associations; autecology, or the relation between the individual and the habitat; synecology, or the relation between the plant asso- _ ciation and the habitat; ra or the change of floras; and syngenetics, or the change of plant associations. It appears to the reviewer that this is the most logical classification of dese fields of study with which he is familiar. As a matter of practice, however, it is unlikely that investigators will increas- ingly recognize such subdivisions. A treatise dealing only with synchorology was fairly satisfactory in times gone by, but in these days it would seem sterile, except as livened up with ecology and genetics.—H. C. Cow Les. ontinuous variation.—Stout and Boas,’ as the result of their extensive statistical studies of variation in Cichorium, recommend that critical study of species variation should be based upon intensive studies of partial (existing among the parts of a single individual) and individual (characteristics of plants as wholes based on their entire record) variabilities. They suggest that failure to appreciate this necessity has allowed considerable error to creep into the work of a number of investigators. For example, hereditary studies of such * Riser, Epuarp, Anfinge und Ziele der Geobotanik. Vierteljahrsschrift der naturforschenden Gesellschaft in Ziirich 62:629-650. 1917. 9 Stour, A. B., and Boas, HELENE M., Statistical studies of flower number per head in Cichorium Intybus: kinds of variability, heredity, and effects of selection. Mem. Torr. Bot. Club 17:334-458. pls. 10-13. 1918. 1918] CURRENT LITERATURE 541 characters as the size of flowers should be prefaced by an accurate knowledge of how such characters vary with relative place position on the plant or relative time position in the total period of bloom. he authors have been able to isolate and maintain a number of races, but further state that “within each race there are further variations, continuous in gradation and of the same nature as those appearing in a more mixed popu- lation, which are unmistakable evidences of the instability of characters and hereditary units.”—MERLE C. CouLTER New-place effect.—CoLiins” has performed a rather unusual experiment with maize, testing the immediate effect of transferring various races to new habitats. We have abundant testimony that it is unwise to go very far from home for seed corn, and have generally concluded that local corn has become the best adapted to local conditions as the result mainly of artificial selection, whether conscious or unconscious. In accordance with this we should naturally suppose that to transfer seed would depress its yield (for a few generations at least). Coxtins, however, shows that while Texas seed of a given strain, planted side by side in Maryland with Maryland seed of the same strain, exceeds the latter in yield by 8 per cent; when the two are grown in Texas the Texas seed exceeds in yield the Maryland seed by only 2 per cent. It seems that the transfer of Maryland seed has acted as a stimulus to relatively greater yield. This phenomenon is termed “new-place” effect. It adds a further complication to the already perplexing problem of vigor in maize.— MERLE C. CouLter. Dominance and parasitism.—JoNEs"™ finds support of his theory” that dominance accounts for hybrid vigor, from observations on susceptibility to parasitism in maize. It has hitherto been demonstrated by several investiga- tors that resistance to aaa behaves as a definite heritable factor. JONES shows that inbreeding corn serves to isolate certain homozygous races which are susceptible to smut ai leaf blight while the more heterozygous ancestors are resistant.. He concludes that “‘as in so many other cases, those factors which enable an organism to attain the best development tend to dominate.” Thus, in general, the most heterozygous corn, which therefore shows the greatest hybrid vigor, will be the most resistant. A difficulty arises here, since certain diseases are known to thrive best in the most vigorous plants. It might be possible to account for this difference on the ground that certain diseases are immediately destructive to the host while others are not; although if this were true, Jones’s leaf blight disease and smut should behave differently —MERLE C. COULTER. %” Cottins, G. N., New-place effect in maize. Jour. Agric. Research 12:231~243. 8. 1 Jones, Donan F., Segregation of susceptibility to parasit Amer. Jour. Bot. 5: 295-300. 1918. ™ Rev. Bot. Gaz. 66:70. 1918. Igl 542 BOTANICAL GAZETTE [DECEMBER Lichen growth.—As the results of experiments and observations extending over a period of 8 years, Finx"3 has determined the rate of growth in certain crustose and foliose lichens, as determined by measurements of the diameter of the thallus, to vary from increases of 0.36 cm. per year for Umbilicaria pus- tulata, and 0.42 cm. for Physica pulverulenta, to 1.3 cm. per year for Parmelia Borreri and P. caperata, and 1.75 cm. for Peltigera canina. Some of the inter- mediate annual increments were 0. 2-0.75 cm. for Graphis scripta, 0.6 cm. for Verrucaria muralis, and 1.16 cm. for Parmelia conspersa. In these measure- ments FINK has given us practically the only definite data we possess relative to the increase in size of these pioneer plants. With regard to migration, FINK declines to indulge in speculations regarding possible methods, and says “nothing is definitely known further than seeing parts of Cladonia thalli lying on some of the quadrats in early stages of ecesis.”—Gro. D. FULLER. Vegetation studies in Natal—Brws continues his interesting studies of the vegetation of Natal, his latest paper dealing with the ecology of the rakensberg.5 These mountains exhibit picturesque and even stupendous scenery, the highest peaks being more than 11,000 ft. above the sea. The most extensive formation, as elsewhere in Natal, is the veld or grassland. The alpine veld is composed more of tussock grasses than is the lowland veld, and the growth forms are more xerophytic. An interesting formation is the Protea _veld, dominated by various species of small trees of the genus Protea. The climax formation is the bush, dominated by species of Podocarpus, and occupy- ing the more protected situations. The mountain top vegetation is markedly xerophytic, and is dominated by composites (as Helichrysum) and heathers (as Erica). The last section of the paper deals with successions and inter- relations.—H. C. Cow es. Tree growth in lowa.—In presenting data upon tree growth in the vicinity of Grinnell, Iowa, Conarp* brings out several interesting facts in addition to the average annual increment of several species. There seems to be conclusive evidence that trees are encroaching upon the grasslands, and this is ascribed to the elimination of prairie fires during the past half century. While this accounts for the present increase of forested areas, it is not regarded as explain- ing the presence of grasslands which constituted’ the natural vegetation upon the best soils in the region. These richer soils are very favorable to tree growth and the increments are sufficiently large to indicate that timber would * Fink, Bruce, The rate of growth and ecesis in lichens. Mycologia 9:138-158- 1917. ™ Bor. Gaz. 64:85-86. 1917. *s Bews, J. W., The plant ecology of the Drakensberg Range. Annals Natal Museum 3:511-565. pls. 4. figs. 3. 1917. © ConarD, H. S., Tree growth in the vicinity of Grinnell, Iowa. Jour. Forestry 16:100-106. 1918. 1918] CURRENT LITERATURE 543 prove a profitable crop. Some typical average annual increments are Carya ovata 0.22 inch, Quercus macrocarpa 0.30 inch, Q. velutina 0.29 inch, Acer saccharinum o.63 inch, and Juglans nigra 0.34 inch—Gero. D. FULLER. Inheritance of height in peas.—According to MENDEL’s original classic experiment with peas, the cross tall dwarf gives a simple monohybrid ratio, with tallness dominant. The work of a number of recent investigators, how- ever, has indicated that height in peas is a much more complex character, and that Mendel’s 3:1 ratio by no means states the whole truth. Wutre” has made a critical examination of these investigations and has added some of his own. He concludes that there are at least 5 genetic factors involved, 2 for internode length, and 3 for number of nodes. He points out, however, that the same genetic pea material that MENDEL used will still give the 3:1 ratio. ‘The inheritance of height in peas has become complex only because of studies on new or distinctly different material, the characters of which, there is reason to believe, are due to distinct mutations.”—MERLE C. CouLTER. Intercellular canals—ReEcorpD® has investigated the occurrence of inter- cellular canals in dicotyledonous woods, and has discovered 16 families in which they occur, mostly tropical. In some cases they are a normal feature of the wood, while in other cases they develop as a result of injury. They vary in direction and origin, in certain features resembling those of gymno- sperms, but in many important features quite distinct. The secretions exhibit a wide range of variation, being resinous, oily, gummy, or tanniferous, as contrasted with conifers, in which the secretions are wholly resinous. ReEcorD concludes that the presence of intercellular canals in wood is a valu- able diagnostic feature, and it was with this primarily in view that the investi- gation was made.—J. M. Inheritance in Pisum.— WHITE” has presented a very significant paper on the interrelation of the genetic factors of Pisum. He has collected a mass of data of his own and also of earlier investigators of Pisum. He distinguishes 35 factors and discusses 5 linkage groups. A model section appears under the title “Modification of the expression of Pisum factors by different environ- ments and by each other.” This is one of the first successful attempts to make an intensive study of inheritance in plants, such as has been so well made on the fruit fly. Another such study, on corn, is now maturing at Cornell under the direction of Dr. R. A. EMERSON.—MERLE C, COULTER. "7 Waite, Ortanp E., Inheritance studies in Pisum. UI. The inheritance of height in peas. Mem. Torr. Bot. Club 17:316-322. fig. r. 1918. 8 RecorD, S. J., Intercellular canals in dicotyledonous woods. Jour. Forestry as: —_ 1918. 9 Waite, ORLAND E., Inheritance studies in Pisum. IV. Interrelation of the — tase factors of Pisum. doen Agric. Research 11:167-190. 1917. 544 BOTANICAL GAZETTE [DECEMBER Rusts of Oregon.—JACKSON” has published an annotated list of the rusts of Oregon, which brings together for the first time the rust flora of a state on the Pacific coast. All of the grain rusts recorded for North America (except Puceinia Sorghi) are known to occur in the state, and also all of the rusts of greenhouse crops. In addition to these, the Pacific coast rust of pears and quinces is vnipe to be of ee = ee importance; and of course the forest-tree dofinvestigation. The list includes 220 species of rusts ents on about 500 pretnn hosts, 8 of the species being described as new.—J. M. C. Practical breeding.—CoLiins and Krempron* have given an excellent example of the effective application of the principles of pure science to the solution of a practical problem. The production of a race of sweet corn resist- ant, to the earworm has been a strictly practical problem, and introduces no new phenomena or theories of inheritance. The authors, however, have established statistically the correlation between the amount of damage done by the earworm and certain superficial plant characters, and have followed this by selective breeding for those significant characters MERLE C. COULTER. The morning glory in genetics.—BARKER” has found that the morning- glory is very favorable material for work in genetics. The almost innumerable combinations of floral colors are beautifully explained by the enzyme theo “Each epistatic type is due to the addition of one or more gener probably enzymatic in nature, which are not present in the hypostatic type.” —MERL C. CouLTER Rusts of Cuba.—ArtHur and JouNston® have brought together all collections of Cuban rusts as a “basis for a thoroughly scientific and economic exploration of the island.” The list includes 140 species, 12 of which are described as new, 15 are new to the North American flora, and ro are exclusively Cuban. a .c. ts oi Plein: H.S., The Uredinales of Oregon. Mem. Brooklyn Bot. Gard. 1:198~ 297. 1918. * CoLuins, G. N., and Kempton, J. H., Breeding sweet corn resistant to the corn earworm. aie A zig Research 12:549-572. 1917. ARKER, E. E., Hereditary studies in the mere cage’ oS ais purpurea). Cornell Univ. “i ‘Rae Sta. Bull. no. 392. pp. 38. pis. 3 3 ARTHUR, J. C., and Jounston, J. R., Uredinales of pa ne Torr. Bot. Club 17:97-175. 1918. GENERAL INDEX Classified entries will be found under Contributors and Reviewers. New names and names of new genera, species, and varieties are printed in bold face type; synonyms in italics A Abscission 75, in Coleus Adaptation a a Ariss steotion 382 ddisonia 2 Aération of coriout solutions 80 Agaricaceae, a ma ua in 459 8 Andrews, E. Apog: y in Arber, Agnes, work of r, J. C., work of 182, Aucuba, embryo of 7 Aus tralia, flora of the Northern Terri- tory 537 B ere. Lis St, ee * toe ork o j My Begonia, endemic, of Hawaii Young, Ye aphesl sketch of 45 Bews, J. W., work of 542 Blake, S.k. ’ work of 72 ryophy Buchholz, J. T. 185 Burlingham, Gertrude S., work of 73 C erie 45 t be 277 Cam H., ‘‘Mosses and ferns” 545 Cannon, W. A., work of 78 Carbohydrates, ‘etinaton of ag forma- tion and translocation of 2 6, alba anomala 238, formis 231, Dunbarii 254, 247) § glabra 242, glabra megacarpa 244, neyi 253, leiodermis hamberlain C1] egoeooeoe > oe = # ZB. =) ollins, G. N. onard, H. S., work of 542 ‘ontinuous variation 540 ontributo Andrews, 382; tkinson, G. F. 285, 459; Bakke, A. L 81; Bliss, Mary C. 54; Bro ds G 269; Buchholz, J. T. 185 Caldwell, J. S. 178, 277; Chambe 3. 2 Coulter, J. M. 72, 79, 80, 182, 183, 184, 287, 383, 392, 464, 537, 543, 544 Coulter, . 70, 284, 461, 463, 549, 541, 543, 544; Cowles, H. C. 391, $40, ae r, W: 80, 5 g- , B. 272; dgeo 393; 364; Ottley, Norma E. 354; Phillips, 1... 37, 546 462; SOROS ig J. 61; Reed, H. S. 743 Coulter a, i. a 183, 184, 287, 383, 392, 464," 537) $43; 544 Coulte Seg Ze 70, 284, 461, 463, 540, 541, a Cowles, a eM 301, 540, 542, Sr of 387 Crocker, W. 80, 184; work of 4) Cuba, rusts of ae bon einen B.3 Cycas, foreign poliew on 392 D Daish, A. J., ~~ of 178, 180, 181, 277 Davis, B. M., ork of 2 Davis, Olive B., Flora of the Northern Territory” 53 Davis, he For work of 178, 179, 180, 181, or E. ML. seg of 388 Desivcation E East, E. ai — of 70, 461 Elmore ys 287 Elymus Pedant aa — of Pinus 185; of Aucuba 7 Embryo sac of Oenothera . Rissesern | nucleus in Liliaceae Eutetramorus 74 Evans, A. W., I Evans, * B. ae we ork of 539 Ewart, A. J., oe of the Northern Territory” 5 F Fairy rings 3 Farwell, O. x. work of 73 Fernald, M. ‘ie as of 73, 183 ers, apogamy in 80 Fertilization in Liliaceae 143 Filicales, Phylogeny of 183 INDEX TO VOLUME LXVI : abe soils and for [DECEMBER Fink, B., work of 5 Food reserve in estate plants 162 est depeorcatib 76; competition factor 479; hum ried 4723 i pa factor 474; soil factine 471; tem fac 467; tim mse 4743 ue fates 408; range 465 Pines E E., work dys 78 Freeman, G. F. Friesner, R. C. ies: spr , G. D. 78, 78, 288, 385, 390, 539, hon of Ceylon 183 G Genetics, morning glory in 544 Gleaso n, H. ns work of 386 Good T. H. ns 176, 381, 534 is 3 in, ye, Week of 73 Guillermond, A. , work of 383 H Haas, A. R., work of 462 Halsted, B. D. , work of Fire ork of 184 . 524 waii, endemic Begonia of 273 Hayes, ‘i. K., work of 70 utchin Hybrid vigor 70 i Illinois, grasses of 3 Impatiens sultani, Tite history - bcd Insectivorous plants, mechanics of move- ment in 77 Insect Intercellular canals $ 543 owa, tree growth in 542 Tron i = nutrient solutions 184 I a, M., work of 184 J Jackson, H. S., work of 544 ennings, H. .. work of 463 Jensen, G. H., work of 288 © 1918] Jones, L. R., work of 79 J@rgensen, {, work of 80, 389 K Kluyver, A ay work of 179 Kormickia L LeGoc, M. J., work of 392 ilium, fert ilization in 259 Livingston, B. E., work of 183 Loeb, J. 60 Long, E. R., “work of 80, 390 M Macbride, J. F., work of 73 MacCaughey, V.2 273 Mac bE D. T., work of 80, 390 Mann, A., work of 28 Metz, C. be? These f 284 Michigan, the vegetation of 3 Microtome, caodtieation ot oe 534 Millspaugh, C. F., work of 74 oo experimental investigations Misachondvis 4 Monocotsledons, intrafascicular cambi- in Morning ae in genetics 544 Mosher, Edna, a 392 I Muller, H. J., wor Murmill, W.A — of 73, 7, poisonous 285 463 ushrooms, Mutations and sie Hovlats N Narcotic plants and stimulants 464 tal, vegetation studies in 542 Natural selection and adaptation 382 New-place effect 541 Nichols, G. E., work of gd Nothnagel, , Mildred 143 N Sutin Sie a aération of 80 INDEX TO VOLUME LXVI 547 ie embryo sac of 184; the situa- 2 shoe Alice M. 2 a) am Be oak a 288 r Pallavicinia Lyellii ay Palm, B., work of 7 Paraffin solvent aa with paraffin 381 Parasitism and dominance 541; phe- Pellew, C., wor! enicillium, arin peotaction in 539 Permeability 88 Petch, T.., ork of 1 Petherbridge Me fe er ‘ pungoid and insect pests Pfeiffer, Norma re » Phillips, T 1.6.7 eee, tle chit parasitism in Piemeisel, 2 wee of 391 needles, th OE and his- m, bis seer cones ak eg um, inheritance in 543 of Asclepias cryptoceras 177 pubeeavors 184 Practical breeding 544 R gn map experimental investiga- ns I econ: 5. 5 61; work of 543 and ferns”’ No a $ “Flora of gic 537 Robertson, C. 1 mS ob cad ‘soll aération 78; varia- tia 538 oe E., work of 540 R crispus, morphology of 393 Rusts, a Cuba 544; of 544. Rut AL AcL., work of 79 548 S Safety-razor pie Lape) AN Safford, W. E., 130; arctica 118; ar 128; brachycarpa 336; brachycarpa ellicarpa 338; a 3383 chlorolepis antimima 339; cordifolia ; cordifolia Macounii 347; deser- ora 43%: fi beer gd 340; glauca uca acutifolia 32 gltbrescens Fares g lingulata 353; ee 339; ovali- folia 138; petrophila 1 pseudolap- sce 3345 | stolonifera es Shantz, it. L., work of 391 Sharp, L. T., work of 78 work of 74 Soil aération = root growth 78; mois- ture studies South Afi fecal of 539 ptm 53 irogyra, abnormal — in 269; cross-conjugation in Sporangia of Stout, A. B., work of 540 Succulents, buffer process in 462 T Taxineae, sac phepeni of 54 Thismia americana, sporangia of 354 INDEX TO VOLUME LXVI [DECEMBER 1918 Tilia 421; caroliniana 496; caroliniana thoophila 498; Cocksii 437; creno- serrata 430; floridana 431; ~~ australis c ana 510; Mic 506; hete rophylla. ni georgiana 502; texana Eat — 428; venulosa multi- Tracheids, significance of resinous 61 sty on: True, R. H. - G U Units of Me sae 385 Uredineae W Walton, L. B., work of 74 Water culture © 302 hie a , D. D., work of 78 Weniger, Wanda Whea’ itn a Seg of 288; absorption of atice and calcium by seedlings 374 White, O. E., work of 543 Wiegand, K. M., work ic 8 + Yoder, L. 364 oe A Zeller, S. M., work of 74