MEHMOTLIRS

NATIONAL ACADEMY OF SCIENCES

Volume XI

ee,

NATIONAL ACADEMY OF SCIENCES

Volume XI

AGAVE IN THE WEST INDIES

BY

WILLIAM TRELEASE

ey

CONDE LS:

Page. Page.

LNW GIRL A cocndedasebaccecsn doopsegenpEcecscosss 7 | Antillanae—Continued.

GEABERONALACUCTE nan. Haas ios cmos eemeen 11 Aw racleulanisse een. ister aaa ea oe asses secineiies 34 MRta rT See ate re ays es cee eyo ec ie crease th IATA TSISG aN armen see sme ey tees utp Athans coal aye 34 WOR VOSS tec os ts 5ale So Hie Sa ee es Se eae 12 Atreniscarclreniier sana ease eee ee oceee eerie 35 HH OTesCENCe Sarma ie cee ee eaclmmiSecrse 13 AlN PTISeaLO DESISp INA =tse cies e ee eases alaciele ers 35 HED ANRC SRE eye eer Sia etincmeie eefenie 13 PAWS Shiaitonimme nr were sara ee eee eearnto 35 IER Peete chee era nine einscnine ae cine 14 Avalbenrellignasenereas secre eas ccc ae cece oes 35 TEXSUISTISS 36 Ges ED EECr eae AS Serene se enn 4 Averell lian aybreviil oraeeeeeere ne eeeen cee eer 36

Paxonomic application... 2.0. soss-s-ees 15 ASS ON PIPES NS ape ee eet ne ys wetness = 36

SILO DE Gh ssang cae se Pecan EB eee ae CROCE EE BEBEee ee 16 Amano malaeaancr te sen nee sona seeenia tee 36

Ce ERE Metare tee iafa ee, Najarni= i= wie ~)afsis ain -mae sins eine e cle 17 AGAUnderwoodlbesseeaseeeee eee bee nee seer 37 2 TESTE VEN oF 5 See ee a eee 18 GILT SSI ONT fee sees eye R seat ni dee aie awic sea ON Jig VWIGHIE cnadesbococs ob eeeooe Sab ecenne rede 19 IAM POLLONIGEM SIS eae ere see Sa er ieee ee ra 38 JAN; CHOU 4 -ceacumye ao seue Une COU bb ere coroEeE WOM eBahamanaceas came voters eine na sega ive 38 Ais PERE EE Cea Spe neeacorese seleinyajeicisies === cl/si 20 ING AVC TBIACCAND a ici- als ajo eis ines Hees cin ese sc 40 Mev osAG CNS S sen et tsetse note en scmiecrnateeie arate 20 IN ont onGhi tae sees cacao ceeee ne neL an AO eeOOs 40 Amabvol dine han aye eyyyoe cee e< elses ace eie oem 21 Ave Mill spau chiieeeermrs tae acest iercc aan 41

Waripaeicett eae Se ie honce ee PO Orc a AIOE 21 IN ENGOV ASIC oie so he NS ee ee ae EE re 41 AEGNI@ URANO so sooo ceececsgedsooscasacpesas 23 INA EY AICO Ae Saas oe nj oeeer ia ase eeOe see 41 Lu. Welt CRORS tide a ses enaseonenseaGaseosoosss 24 Avinda gaboruliersn see eens a eee ericrss 42 LN. TIGRIS Secanaa poe a eenenoemaaoRete oc Cuose 24a Ambilllaresae eer severe crac a eagie ic Se sete ct reves erste 42 PRMeN ON ESCITALCN SIS! ce ela eine [= = eric els = 24 Noavies Will dinette ss sastsasese toe essences =e 42 PNURTINC OL ORAM Dera ehateoyas viata cisco eve oie Se cverettins slate 24 IAC al bGSCEMS Ste cheers = aac e emcees en sin me 44 As. GHEWAGHOR, bobo deisedonoocesacoserossocguae 25 AGE DAD YLOCALP Ae ee tener sec ateee ie teenie 44 PRES CHOU CLM ADIAN Ae aj.5.<1)5 am oo: cje)~ niciejciin's sale 25 IN) Bim OMib ih eeatpocaasa te eeoee cae eai ee aaa e 44 FOU GUC basse aaah cen ssa saicsic cise see aes Acinic = 25 /\.. BYRNE) OEE MANS sscacassanososoneane 45 Ji, UBRN aG PSE so ccobecenoses oneeo ose boduade 26 JAE OU Nike cea ace ae ena is sei et oe ee ero ae 45 /\.. MUSSB oct os bo epeseeecsosacorces suosaer 2G elnaeu ensesoee sae sees ane os dames coe co Sess 45 Ao. COMME GIGOE cococecacaoocacccoccocszemscns 27 NeavesNashitize oe at indo .ae -etiswinyarepricteo eae 45 PNMPBIOPErSiAM ase cis scs is ciesiais ct aaie | arespeeiewe 28 ASsIMapUeNSISs ees ooh oe oon cee cteeiernaeeets ace 47 Jd TRY SEC GIR: Sone ses aoe aan eeesoe sass Sener ZOule Sisdlanae se veepereeeiiien eee tae se ree eerie 47 A, (ing CUIR ee eestocoscooogde be oonTaosceonTe 29 Npavealpustitoliae.seeeeene sean 47 PAE CNGUIN-VCLSAita ss). inci Jae ees amieacne 29 IAS TOUNGCKOY CeSt yas seer ee seiner nies e yer 48

PNET ATI ACR epee affeyersi=t\e as a arcre |= isi SASS SiS wisie mie Sia 29 IN, GISMAN So oneodee pedssemsccacasdEedouSase 49 / wane eminent. -ooncecocosugocoedseoaceuace 31 AA, GRAMERCY, Sec goSssbeoceanoodooseecde 49 Ak, TVG ERPs So Scb oocecosocusoseeeooedane 32 | Occurrence of species by islands................--. 50 Ai, GO SO NIGTEe-scoscoe dase noeasusecuedasuEnore 321 eCollectorsiandicollectionss=e—-sse-se= => eee eee 51 INQMOVPISU se yas 2 cjaje ane Sa Se seieei deseo s esis 36} || Comimnein mens cacc2acscoodoesnadecossHenonoce 53 Jig ii brat papasaaneeoacecn SeAoe ae eaGnea ec 34 | Index to species described and figured........... 55

fe ws A ELON Ss:

- Page. Page. Ficure 1. Distribution of Viviparae ........-.-.. st Puiatse A. The earliest-known West Indian Agave. 58 2. Distribution of Caribaeae .......-..--- 22 B. Spine agreements and contrasts... ..---. 60 3. Distribution of Antillanae ........-...- 30 C. Differences in bulbil plants ....-..---. 62 4. Distribution of Bahamanae...........- 39 D and E. Agave and the Antillean Bridge. 64, 66 5. Distribution of Antillares............- 43 1-116. Species of Agave (see p. 55 for index) . 67-298 6. Distribution of Inaguenses..........-- 46 7. Distribution of Sisalanae ............-- 48

akan Le

my ORG

i paver”

Freee cute ae

pis gee tigh are

mn” urea

l

--f paoraey :VIGTAZANTA

“Leaded aovyor, “fI—>LISLe ~--eqniy Beige” ovdeing Soo eee ollemo g cae | Seal hazel CS Rel (Ree eee e3n410, weele-[---[---]---]/---[------ - ~eULIVsIV IW, : “spuBisy svoog 7 (tas ess |e | ea | beat pepruny, peal .|---|---]---]-----++=="oggqour, ‘SGNVIST GUVMGaT

BSG aia onl end ao epeualy sourpeualy jUsdUTA “4S ee Bony “4S SI Se ead Os a esa tC sopeqieg ed ees eee ee =e. -onbiuyie yy 56 laeeel ete) [se = = BotUrt0 eles os - + sone sary BSA seclioaa sae -oqUR[BH oUeyy Beal ee aa "== opelsed Ae |e Z(cee ---adnojepeny Sosa vats cise atc etckerecis ye.110S} WOW ee ee ees ee ees ey ensyuy

Eee SHIM 9S "7 *snigeqsng “4$ - MOUIOTOYIBE “4S ee aos uy ey, “4S

Per ieees ooeqy “spueysy Alleq

Bish eo ioe ~*-|"90uepIAOIgd MON, pase yseiie ABO 189 **BUInNX ere: 300, ---* £9 tangy ~ ----qorde0009

ooo o Bure M

z ““Spuvysy Yn pita | Sear cic *senseuy ISVAVHVGE

---B1qe[NO Sciyppeessisiais sonberA. Be panto” oory OVEN

eee CLIN Bed Hones @ATBUOX) peo tees 2304101, eugene dO ~--="-sueurAeo

ISHTTIINY UalvVaup

= | iS 2 = >

elalale|"|si4 alls re callie = >| 4 Ale

4 ais NIA elo Sy s)s8 8 ayy lat) cath SN tsp Spit Be | eleaial: =P S| 2) L/S} ail | EPI Sy 7) an SIZISISl(SISIEISIGISIS EI FICS Sl slslelelelAlclel ZiSlelelclSlSlSlelSlelslelslclelelslZlZislSiclelsieleisiele S/2IS/2IB/SIP/SISISISIS/SIZ/SIS(SIS/s Si Slel alae S|EIS|EIE| 3/5 SSS slEls Zl alsi/Slel@iSisisisisisieye ‘SU'VILLNV tl earl) eal deb Slr Ilol els |e) s Ble = lo a BIliels] @ a9 N54 Ble a; 20 6/215) 8 Slololes |EI EIS) a] Sle SISISITIPISIBIEIS|SIZISIBISISISIS/ZISIFICISISISIBIGISIE/SISIB/S/Z/S/S1 s/s sl SlSISISlSIBISIE/ Sl? lelelai Slee GH NI Ge) 2) BI| JP lElOCl le |S (el/B/Sleleral| ale | Solal- |] BIE SIS/PlSeIS ISI SISI/SiSIBI Bl 21Slelalel] wl e2i sy: A oN aad ot IT eel to] AAVDY z. : oe 4/8) a/8/ So] s] > B)B|s/8 =| 6 a = Blalsi-|° |B) Se)! |S)s)Si° le = # Salas aoe 5 ele|2i/5/5/5 Pelee] (Bis! |: el] JEP ls Pls) (2) |2leiel | Fl) ie 5 © | 6 ae ° So) = 5 oe id a =P : | S xa : i 2

an eel oF {Ps

AGAVE IN THE WEST INDIES.

By Witt1am TRELEASE. ANALYSIS.

Conspicuous and strikingly dissimilar to anything then known in Europe, the American aloes, as they have come to be called, must have attracted the attention of Columbus and his companions, when, in the Bahamas and Antilles, they discovered the New World in the autumn of 1492. It appears to have been use rather than form, however, which first excited comment, and the earliest written mention of them is found in Chanca’s account ? of the aloes of Haiti (1493) and Oviedo’s (1535)? commentary on the maguei of Haiti and the Arayan mainland.

As plants, the West Indian agaves were first described and pictured in Munting’s Aloe americana minor (1680) (pl. A) ascribed by Boerhaave to Curagao, Hermann’s Aloe americana sobolifera (1687) (pl. 44) from Jamaica, and Dillenius’s Aloe barbadensis, etc. (1732) (pl. 34) from Barbados. Their nomenclature under Agave, apart from deviations in spelling, comprises A. viviparA (1753), A. Karatto (1768), A. americana (1774), A. spicata (1802), A. ANTILLARUM (1827), A. soBoLIFERA (1834), A. Offoyana (1864), A. LEGRELLIANA (1866), A. coccinea (1876), A. CARIBAEA (1877), A. Morrisi (1887), A. polyacantha (1888), A. mexicana (1889), A. RIGIDA (1890), A. Witipiner (1891), A. dominicensis (1893), and A. Wieuti (1907), the date in each case marking the first connection of the name with an Antillean plant. The five names in italics properly belong to continental species that do not enter into the West Indian flora; those in lower case Roman type are synonyms of two of the remaining 10 species which, under these or different names, belong to that flora. A. spicata, described at Madrid from a plant supposedly native at Havana, is now regarded as of continental origin and has no close allies in the West Indies.

Until within recent years herbarium material of Agave, and especially of its West Indian representatives, has been both scanty and questionable. My attention was diverted from a general study of the genus in 1905 by the receipt of specimens of a strikingly xerophytic new form (pls. 101-103) collected on Inagua by Mr. George V. Nash of the New York Botanical Garden, and no small part of the time that could be given to the genus for the past five years has gone to an examination of its West Indian representatives, of which extensive collections brought together by the several energetic collectors of the New York Botanical Garden have been supplemented by specimens contributed by correspondents resident or traveling in the archipelago and by my own gleanings during one field trip on which nine of the islands were visited. To the collectors who are mentioned in connection with the specimens cited, and espe- cially to Professor N. L. Britton and his associates, is due the possibility of bringing together the following synopsis of species. In this, gaps still remain to be filled and errors of judgment doubtless occur, but otherwise it is believed to contain a fairly complete presentation of the West Indian agaves, except that the large and little-explored islands of Cuba and especially Haiti may be expected to yield additional species of their characteristic groups, and collections are still to be seen from several minor islands.

The plants occur more or less locally on the several islands and usually affect rocky places or arid exposures (pls. 4, 9, 84, 95, 101), sometimes in association with cacti (pl. 74). Appar-

1 Presented in abstract before the National Academy of Sciences, Nov. 8, 1910; analysis of geographical distribution and probable mode of introduction previously communicated to the National Academy of Sciences, Apr. 19, 1910, and to the Academy of Sciences of St. Louis, May 16, 1910.

2 See Smithsonian Misc. Coll., vol. 48, 1907, p. 455.

3 See Rept. Missouri Bot. Gard., vol. 18, 1907, p. 32. 5

8 MEMOIRS NATIONAL ACADEMY OF SCIENCES, VOL. XI.

ently they are to be sought with confidence on the few islands of any size for which none are yet known. As arule, except for the larger islands and the Leeward group of the continental shelf, only one form is found on an island, and each is confined to a single island or to islands rising from acommon bank. The impression produced is that each main division of the archipelago possesses a distinctive species of large-flowered Agave differentiated into minor forms on its several islands. The coordination of these forms, however, reveals such manifold agreements and differences in characters as to result in the conviction that they themselves constitute species, in a conservative sense, and that the more comprehensive and widely distributed types are really groups of species rather even than superspecies; and they are so treated here.

Of the 50 species here recognized for the West Indian flora, 3 (pls. 106—-115)—Agave angustifolia, A. fourcroydes, and A. sisalana—are clearly introductions, and represent a group of the Yucatan region. The other species are all endemic. Five of them are confined to islands on the continental shelf adjacent to the Venezuelan coast, where a closely allied species occurs on the mainland; and, though the affinity is not very close, these appear to be related in a manner to the as yet uncharacterized Costa Rican Agave Werckler. None of the others shows close relationship with any form of the North American table-land, on which, especially in Mexico, the genus centers. - ;

Even on superficial examination, the indigenous species fall into three groups: A small- flowered rigid-leaved very xerophytic type (pls. 101-105), confined to the Inaguas; a small- flowered fleshy-leaved type (pls. 93-100), of Cuba and its islets; and a medium or large-flowered fleshy-leaved type (pls. 1-92), ranging through the entire chain of islands and reaching the Venezuelan coast. Closer acquaintance with it shows that this third type—the representatives of which are mostly stately plants, rivaling in size the well-known century plant (A. americana) with which they have often been confounded—treally consists of four separable groups: (a) An ample-panicled capsule-bearing type with usually green leaves ending in a long grooved spine and often repand and bordered with rather large prickles (pls. B and 41-82), of the Greater Antilles; (b) a similar but rather smaller-panicled type with mostly gray leaves ending in a long flat or grooved spine and bordered with moderate prickles (pls. B and 83-92) of the Bahamas; (c) a rather narrow-panicled sometimes exclusively bulbiferous type with usually green, straight and rosy edged leaves ending in a little-grooved short spine or mucro, below which, however, the leaf tip hardens into a heavy involute base, and bordered with minute or close-set prickles (pls. B and 14-40), of the Caribbees and Leeward Islands; and (d) a smaller if anything more succulent type with green or glaucous rather repand leaves ending in a long and usually slender grooved or involute spine and bordered with rather slender prickles (pls. B and 1-13), which is peculiar to the Leeward Islands and the adjacent Venezuelan mainland.

The small-flowered forms comprise two not intimately if at all allied groups: (a) Inaguenses (2 species), limited to the Inagua and Caicos islands; and (b) Antillares (5 species and one minor form), of the Cuban region. The large-flowered forms, on the other hand, to which the Antillares appear to be allied, include four related much differentiated groups: (a) Antillanae (14 species and three minor forms) of the Greater Antilles, (b) Bahamanae (6 species) of the Bahamas, (c) Caribaeae (15 species), of the Caribbees, and (d) Viviparae of the Leeward Islands (5 species) and of northern Venezuela (1 species). None of the species into which these groups are divisible, so far as now known crosses a 100-fathom barrier, however narrow, except as they might be understood to do so when heading the labyrynthine channels of the Bahamian banks.

The characters and affinities of the groups are shown more clearly in a contrast of the species taken as typical of each than in any other form of analysis, as follows:

AGAVE IN THE WEST INDIES—TRELEASE. i)

Leaves broad or curved, fleshy. (Native.) Plants moderately large, suckering, bulbiferous. Leaves green, transiently glaucous; spine acicular, involutely erooved; prickles rather close together. Flowers and seeds medium-sized. (VIVIPARAE)....- Agave vivipara. Plants large, not suckering, bulbiferous. Flowers and seeds rather large. Spine obliquely mucronate from an involutely slit tumid base; prickles rather close together. Leaves green.

(CARIB ATAE) Sseeae Soe ea as cr ya ea ISE SIS cialalare ayelelelcieie s)e,cisicimsaiciclavela(ale vautasegeas A. Karatto.

Spine conical; prickles rather distant. berestecen, (UNNm WON) son ceenecaoeuec t EF eae aodbcs 5b6 so eoe Cosa TaeossoSsEnseeaaar A. antillarum. ILGRMES ERR, ((BVNTARININE)\ = ooo beseteo sp oocne gece escdcec cu de doceecpmpeasecduceépeagse A. bahamana. Plants moderately large, not suckering or bulbiferous. Leaves green or slightly glaucous; spine conical, narrowly grooved; pricklesrather distant; flowersand seedssmall. (ANTILLARES)..............--.-.--- A. Willdingii.

Leaves narrow, erect, firm. (Native.) Plants small, suckering, not bulbiferous. Leavesfew, gray; spine conical; pricklesrather close together. Flowers SO MEPEeCGsinmall gs (INA GUISNGES) pe <2 .ete(eeie slarsinis sone oie eee naively le Sie) = o1s.o nya sinPne aa esse A. Nashii. Leaves narrow, spreading, firm. (Introduced.) Plants moderately large, suckering, bulbiferous. Leaves numerous, grayish; spine conical; prickles rather distant. lowers andseeds large. | (SISATANAE) ®: 05.) 2.2 isan cose lee cece seceec sees A. angustifolia.

Inferences from examination of the plants and from analysis of their geographic distri- bution are: (a) That even narrow sea channels here constitute all but insuperable natural barriers to their dissemination, notwithstanding that the century plant is known to be cast up occasionally along the Mediterranean in viable condition; (b) that the significance of a channel barrier lies rather in depth of submergence, corresponding to lapse of time, than in width; (c) that the narrow but very deep channel separating the Greater Antilles from the Caribbean chain has afforded a longer isolation than the chasm between Jamaica and Haiti or the channels that part the Caribbees; and (d) that the Bahamas have had more recent interchange with the Greater Antilles than has existed between these and the Caribbees, though less recent than that between adjacent islands of each group.

The closer relationship between the large-flowered insular groups of species than with any focal group of the mainland suggests that they have been derived through one rather than more than one continental immigration, and that they have been isolated from the ancestral source for a period longer than that of their separation from one another. The small-flowered Antillares may have originated locally in the mountains of Cuba from the large-flowered Antil- lanae; whether they furnished the parent stock of the Inaguenses, modified under unusually arid conditions, is a matter for less founded conjecture; but this latter group can hardly be accounted for otherwise, except by the assumption of an additional immigration from the mainland of an ancestral form which left no recognizable traces in the islands over which it must have passed.

There can scarcely be a doubt that the parent stock of the West Indian agaves, which are all paniculate, was derived from North America rather than from South America. Apart from the Viviparae common to the Leeward Islands and the coast region of Venezuela, the genus is not known as native to the southern continent except for one minute form (Agave pumila) reputed to occur in the Andes of Colombia, belonging to the spicate subgenus Littaea, and four as yet unnamed very glaucous small-prickled forms noted by Werckle! as occurring toward the headwaters of the Magdalena River of Colombia, which may be allied to A. Wercklei. North of the Isthmus, Agave is represented by many greatly diversified species of both Littaca and the paniculate subgenus Huagave.

It is hard to resist a belief (a) that perhaps, though not probably, with exception of the Viviparae, the genus penetrated from Yucatan to and through what are now the West Indies at a time when they formed a continuation of the Central American mainland; (b) that the Bahamian group derived its stock through later and transient union with the Greater Antilles; (c) that the Caribbean volcanic peaks were united by land at one time during the insular history of the genus, though they are now divided by deep-sea channels; and (d) that the peculiar species

1 Monatsschr. f. Kakteenkunde, vol. 17, 1907, p. 122.

10 MEMOIRS NATIONAL ACADEMY OF SCIENCES, VOL. XI.

of Inagua entered by way of the Greater Antilles and not through the northern Bahamas (pl. E). The following diagram expresses these conclusions graphically:

sitet CARIBAEAE. _-7 BAHAMANAE. ieee ee inugta ae Costaricenses ? ?------- LS VIPAR At: o teredees , ooo 2% lS aes OR ere

Even between islands rising from a common bank covered by water of considerable depth, the local forms of Agave have undergone differentiation for the most part; but the small extent of differentiation between these forms, or even between the groups of species characteristic of the larger island groups, points to rather slow plasticity in the genus during the time which must have elapsed since their isolation (pls. D and E)—probably at an early period in modern geological time. The scant occurrence of Agave in South America may also indicate that this period antedates its passage of the Isthmus and supports the conclusion that the Antillanae rather than the Viviparae stand nearest to the parent stock.

This analysis does not require a discussion of the union and disruption of the continent before Tertiary time, or perhaps even during any part of that period. It is significant that Agave is so slightly represented in South America when compared with such Tertiary genera as Fagus. Geologists and physiographers, though they differ in willingness to admit elevations and depressions corresponding to the deepest present channels, find reason for believing in unions and disruptions of some sort; and a Pacific faunal element, found in the Gulf and Caribbean deposits, shows that the waters from the west must have entered when these were laid down. The West Indian agaves do not controvert Spencer’s conclusion! that the deep channels separating these islands are to be interpreted as erosion canyons formed above the level of the sea in a land elevated after the Matanzas limestones and the Lafayette loams had been silted into still earlier canyon deeps at the end of the Tertiary period.

The mental image that may be formed of the Caribbean region in early Quaternary time is that of a rather low tableland stretching from what is now Yucatan and Central America to South America; bordered by a volcanic range the peaks of which now project some thousands of feet above the water, though the present difference of level between their summits and the intervening deeps may perhaps be due to crumpling as well as general elevation and attendant erosion; and including extensive lagoons corresponding to the present Caribbean deeps, with the Gulf of Mexico to the north constituting a great dead sea. Its climate, as Spencer surmises, may have been comparable with that of the Mexican plateau to-day, with torrential rainfall on its limiting mountains. No sounding data, as they are now obtainable, may be depended on to outline this connection of the continents, because elevations and subsidences can hardly be assumed to have been uniform throughout, but a restoration of the contours now marking 1,200 or 1,500 fathoms of depth gives an approximate suggestion of the extent of the land at this time (pl. D).

Onto this Pleistocene plateau, already sinking and soon cut off from Yucatan by the opening of a connection between the Gulf and the Caribbean Sea (pl. E), may be pictured the dry-shod passage of the primal form from which are descended the dominant and characteristic fleshy-leaved agaves of the West Indies, if not also of another form from which the now unique and localized xerophytic Inaguan group has sprung. These forms, making their way to the east and south around the widening seas and heading the fiords into which the settling canyons were passing, were undifferentiated, apart from local variation, until the filling by the sea of what is now the Anegada channel divided them into stocks from which have been derived the Antillanae and subsequently the Antillares to the north and the Caribaeae and probably

1 Bull. Geol. Sac. America, vol. 6, 1895, p. 128, etc.

AGAVE IN THE WEST INDIES—TRELEASE. 11

the Viviparae to the south. To the newer Bahamian table a transient connection with the Cuban land mass seems to have given the parent form of the group of species peculiar to the islands into which this intricately canyoned plateau has now settled. The segregation of the Viviparae from the dominant Caribbean form appears to have been effected early under conti- nental influences, if they be not of Central American origin; that of the Antillares from the dominant Antillean form may be ascribed to the diversified physiography and attendant climatic differences of the great island, Cuba.

Successive disruption of each of the principal land masses, as subsidence continued (pls. D and E), has clearly afforded progressive isolation, with opportunity for local’ evolution, until each natural division of the archipelago now possesses its exclusive fleshy-leaved type, with subtypes on the great islands and next the South American continent; and each island or group of islands on a common bank possesses its distinctive species or group of species in which foliage, flowers, and fruit show diverse and intricate combinations of the group characters. The genus, therefore, here epitomizes unusually well the facts of insular evolution, a process which is fairly capable of division into three periods represented by: (1) The major groups, (2) such minor groups as extend through islands on a common bank, like the chain from Anguilla to Dominica, and (3) the species of which these groups consist to-day (see the table facing p. 1). Their source is so unquestionably North American and their limitation by water barriers is so marked as to leave little room for doubt as to the former existence of the much disputed Antillean union between the continents, whether or not in the exact form here outlined.!

SPECIFIC CHARACTERS.

Because so commonly described from garden specimens, frequently young, the species of Agave that are currently listed are often hard to recognize in nature and it is very important to understand those differences which may be called characteristic. The constancy of agree- ments and differences is so marked in such well-known and commonly grown species as A. americana, A. picta, A. Victoriae Reginae, A. filifera, A. fourcroydes, etc., as to warrant confi- dence in the general stability of species on the one hand, while on the other the changes that plants undergo from youth to age and the differences that appear between seedlings of a common parentage as well as between plants ascribed to a single species indicate the necessity of caution in either stating or accepting characters as of absolute value, and suggest mutational plasticity.

TRUNK.?

In general, the trunk or caudex possesses little taxonomic significance in Agave, for, though many species are conveniently spoken of as acaulescent, the leaf bases as a rule cover a trunk which, denuded, is seen to be of measurable length. Such differences as the so-called acaules- cent and subacaulescent species show (cf. A. viipara, pl. 1, and A. evadens, pl. 9) are chiefly due to the number and relative thickness of the leaves and the degree of their persistence until the maturity of the plant. Marked exceptions only are of appreciable diagnostic value; such, for example, as those offered by A. fourcroydes (pl. 110), A. attenuata, and especially the truly arborescent A. Karwinsku,*? which have a distinct and much elongated trunk finally denuded or covered with dried leaves below.

1 Apart from the many analyses of the boreal and austral elements in the general flora and fauna of the West Indies, reference may be made here to the newly published Phytogeographic Survey of North America by Harshberger (Engler and Drude, Vegetation der Erde, vol. 13, 1879), with its bibliography, as containing important data on this subject. The marked division in the flora corresponding to the Anegada Passage is noted by Eggers in his account of the Flora of St. Croix and the Virgin Islands (Bull. U. S. Nat. Mus., No. 13, 1879, pp. 13-16, etc.), and Hitch- cock (Rept. Missouri Bot. Gard., vol. 4, 1893, p. 158, etc.) devotes several pages to figures that are very suggestive. The closer biological connection of Jamaica with Haiti than with Cuba is indicated by Barbour (Bull. Mus. Comp. Zool., vol. 52, 1910, p. 277), and a parallel to the entrance of Agave from Yucatan into Cuba rather than from Central America into Jamaica is indicated on a map illustrating the migrations of South American fresh-water fishes which accompanies Eigenmann’s Catalogue of the Fresh-water Fishes of Tropical and South-Temperate America (Scott, Rept. Princeton Univ. Exped. to Patagonia, vol. 3, 1910).

2Engelmann, Trans. Acad. Sci. St. Louis, vol. 3, 1875, p. 293; Bot. Works, 1887, p. 301.

3 Rept. Missouri Bot. Gard., vol. 18, 1907, pl. 30.

12 MEMOIRS NATIONAL ACADEMY OF SCIENCES, VOL. XI.

LEAVES.1

Among the leaves of such well-known cultivated species as have been named, ratio between length and breadth, basal and apical narrowing, concavity, stiffness, and direction of growth are fairly constant. In other species there may be great differences (compare plates 36, 37, and 107—Agave barbadensis, plates 45 and 46—A. sobolifera, plates 67 and 68—A. Under- woodii, and the plants contrasted on plates 76 and 79—A. portoricensis), though reason has scarcely been found for seeing more than individual or racial differences in either case. In these respects, too, hopeless contradictions and confusions were exhibited in the foliage of very young plants of the Bahamian species when their study was first taken up, and it may be that within this group apparently characteristic differences in the width of mature leaves, as shown by material now accessible, will prove inconstant. Within limits, smoothness or rough- ness of the leaf surface, as well as its coloration and the presence or absence of glaucousness, are indicative characters; but though some well-known species are persistently smooth, and others as persistently granulated, there are others in which a characteristic slight roughening varies into either extreme.

Very significant and constant, in the cultivated species referred to, are the characters pre- sented by the apical spine (pl. B) and the sinuation and marginal arming of the leaves. In these features, indeed, seems to lie great, if not the greatest, stability in the leaf characters of the genus. Here, too, however, there are differences which must be understood and taken into consideration, for in repandness the leaf margins of spontaneous plants are found to differ to a very consid- erable extent, and the size and the form of basal widening of the prickles vary according to the amount of marginal tissue that has sclerified in their development. Contrasts of this sort are afforded by Agave vwipara and A. petiolata (pls. 2 and 8), different individuals of A. sobolifera, or the prickles of a single leaf of A. Boldinghiana (pl. 12) or of A. Legrelliana (pl. 59), ete.

The terminal spines of a species often show differences in stoutness which correspond with variations in the general form of the leaf tip; and the partial or complete evolution of the spine, sometimes depending on the chance of environment, may lead to marked differences in its structure and appearance. In this respect the groups of Agave sometimes differ greatly. It is in those forms with a determinate or clearly limited spine that this becomes most charac- teristic, for a definite part of the leaf tip is here devoted to spine formation, and its cells become hardened and colored early. Among native West Indian species indeterminate spines are the rule. In cultivation, species with an indeterminate spine often fail to develop more than a very short apical point of hard tissue, and A. sobolifera is often little pungent in gardens, while its wild representatives possess a strong though variable spine. No really sharp line may be drawn, however, between determinate and indeterminate spines, and in the species producing the latter a time usually comes, if they are well developed, when continued intrusion into the fleshy leaf tissue ceases, and the lower part, already dried, becomes nearly or quite as hard and deep in color as the apical part, or as an entire determinate spine.

The often very characteristic ventral grooving of the spine is greatly dependent on the chances attending the development of an indeterminate spine, for while a prompt basal hardening insures its perfection, the failure of such a hardening often results in its being pinched into a V-like form, a frequent occurrence in the most typical Antillanae and Bahamanae, though sometimes without uniformity of result. It is because of indeterminate development that the spines of the Caribaeae become so greatly and unequally thickened at the base; and the marked contrasts between slit-grooved and open-grooved spines in some forms of A. Underwoodii (pls. 69 and 70) and A. portoricensis (pls. 78 and 80) are apparently to be explained in no other way.

In addition to variations in the straightness of the leaf margin between the prickles, in its repandness, or in its abruptly hummocky appearance when such species as Agave Karatto (pl. 14) and A. Harrisia (pl. 50) or A. Legrelliana (pl. 59) are compared, the margin itself affords characters of considerable constancy and value for the recognition of species. In the subgenus Tittaea a convenient basis of classification is found in the connection of the prickles by a horny

1 Engelmann, Trans. Acad. Sci. St. Louis, vol. 3, 1875, p. 294; Bot. Works, 1887, p. 302.

AGAVE IN THE WEST INDIES—TRELEASE. 13

detachable border in species like A. lophantha, or the fibrous shredding away of the margin in such species as A. filifera; and a considerable number of euagaves, such as A. atrovirens, are marked by a partial hardening of the edge, so that the prickles are connected by it, at least near the end of the leaf. The Antillean groups do not present any of these characters, but a large part of the species described in the present paper have the leaf margin of such a texture that in maturity it.dries into a parchment-like border, either for a short distance below the end spine or throughout. In some species, especially of the Caribaeae, this border is conspicu- ously red during the early life of the leaf, and the original leaf description of A. Karatto has been misunderstood because it included mention of such a marginal coloration.

INFLORESCENCE.,!

Apart from the very obvious difference between the spicate inflorescence of Littaea and the paniculate inflorescence of Euagave, obscured in such representatives of the former subgenus as Agave utahensis? and A. Engelmanni, the relative elongation of the scape below the flower cluster shown when the long-stalked panicle of A. Cocui (pl. 6) is compared with the nearly sessile panicle of A. Legrelliana (pl. 60) or A. Underwoodia (pl. 68) is generally, if not constantly, characteristic. This is also true, with reasonable limitation, of (a) the crowding or separation of the scape bracts, dependent on the greater or less lengthening of internodes; (b) their breadth of form (A. Trankeera, pl. 25, and A. Eggersiana, pl. 31), a feature which is not always the result of difference in diameter in the scape and attendant width of their bases; and (c) their erect or spreading direction. Subject also to limitation, but diagnostic, are the differences in the length of the panicle branches and the consequent shape of the inflorescence shown by A. Underwoodii (pl. 68) and A. missionum (pl. 74). The most characteristic differences in the inflorescence, however, lie in the greater or less branching and extension of the panicle branches, which is evident when such forms as A. Willdingu or A. Nashii (pls. 93 and 101) and A. Legrel- liana (pl. 60) are contrasted. The relative length of the pedicels in which these branches end either accentuates (A. Legrelliana, pls. 60 and 61) or diminishes (A. Willdingvi, pls. 93 and 94) the separation of the flowers, and can in general be depended on in the segregation of species.

FLOWERS.‘

In the flowers themselves, which are much frequented by insects and birds of several groups, differences in color and odor are often very marked, as when the greenish fetid flowers of Agave sisalana are compared with the yellow pumpkin-scented flowers of A. Eggersiana or the orange-throated flowers of A. Legrelliana or A. antillarum. Unfortunately such characters as odor are not preserved in herbarium specimens, and the color of the flowers may be lost in drying them. Generally, size, shape, and proportion of the flowers of a plant are counted on as being fairly constant. Perhaps this is as true in Agave as in most genera, and such differ- ences sufficiently distinguish the small, short-stamened flowers of A. Willdingii (pl. 94), the rather large, long-stamened flowers of A. fourcroydes (pl. 112) with urceolately contracted tube, and the very large conical-tubed flowers of A. Legrelliana (pl. 61). Neither size nor proportion is satisfactorily limited, however, and in flowers with an inferior ovary, such as these, the form and measurements of the germen and its proportion to other parts of the flower have to be accepted with caution, as possibly influenced by the age of the flower or the time that has elapsed since fertilization was effected. Measurements of the flower are further dependent very largely on its freshness or the manner in which it has been preserved, and though well- pressed flowers may give measurements closely comparable with those of fresh flowers, informa- tion of this kind derived from withered flowers like those shown on plate 40 is of only very general value.

One of the most useful floral characters in Agave is afforded by the height at which the filaments are inserted in its perianth tube. Engelmann, who first applied this character com- 1 Engelmann, Trans. Acad. Sei. St. Louis, vol. 3, 1875, p. 295; Bot. Works, 1887, p. 302.

2 Pop. Sci. Mo., vol. 28, 1886, p. 11, fig. 9.

3 Rept. Missouri Bot. Gard., vol. 3, 1892, pl. 55. 4 Engelmann, Trans. Acad. Sci, 8t. Louis, vol. 3, 1875, p. 297; Bot. Works, 1887, p. 303.

14 MEMOIRS NATIONAL ACADEMY OF SCIENCES, VOL. XI.

paratively, made effective use of it in differentiating the agaves of the United States. This difference in insertion of the filaments is very marked when a species like Agave barbadensis (pl. 38), where they are nearly in the throat, is compared with one like A. sisalana (pl. 114), where they become free at about the middle of the tube; but the stamens of Agave are in two series, and the two sets are commonly inserted at somewhat different heights in the tube, so that in closely related forms the fractional differences that may be observed are not readily specified or measured with trustworthy precision. In proportion as the bottom of the tube is flattened, also, the insertion of deep-seated filaments may appear to be lowered from the middle to the base. The absolute length of the filaments and their ratio to the length of the segments of the perianth seem to be fairly characteristic but subject to wide fluctuation; on the other hand, the style, which at first is short but, even after the maturity of its stigma, continues to grow indefinitely, can be relied on to a very slight extent only as affording specific characters. The anthers of related species are usually much alike in general and rather variable

in particular. FRUIT.1

As in other groups of plants, the fruit and seeds afford characters of a fairly trustworthy character. Marked differences separate the large, broad, rather thick-walled capsules of Agave Karatto (pl. 15), the elongated, equally thick-walled capsules of A. Dussiana (pl. 29), the smaller, nearly round, thinner-walled capsules of A. portoricensis (pl. 78), and the small, round, almost papery capsules of A. papyrocarpa (pl. 97). Among the West Indian agaves there may be also marked differences in the shape of the base of the capsules (e. g., A. portoricensis, pl. 78, and A, papyrocarpa, pl. 97) and of their distal ends (e. g., A. Dussiana, pl. 29, and A. indagatorum, pl. 92), even when in size and general proportions they are comparable. Abundant material usually shows the constancy of such differences, but starveling fruits (A. anomala, pl. 66) here as elsewhere must be taken as uncharacteristic. The seed dimensions of such fruits are not to be depended on any more than the measurements of seeds from the extreme top or bottom of a normal capsule, though those from the middle of well developed fruits average up pretty uniformly ; and the seeds may differ measurably in unrelated species (e. g., A. Brittoniana, pl. 98, and A. angustifolia, pl. 109). It must be questioned, until further material is secured, whether the small capsules of A. Mhilspaughii (pl. 88) or the narrow capsules of A. cacozela (pl. 90) and A. Harrisi (pl. 51), of which few and not obviously normal fruits have been observed, are as representative as those of some other species are known to be.

BULBILS.

The agaves first known to be bulbiferous, Agave vivipara and A. sobolifera, received their specific names because of this fact, and a good deal of confusion has crept into the literature of Agave through a misconception as to the constancy of bulbil formation in a given species and its prevalence through the genus. The Sisalanae are markedly bulbiferous, and yet A. angus- tifolia, which typically shares this trait with its near relatives, sometimes fails to produce bulbils when capsules are found abundantly (pl. 107), or defers their production to a period subsequent to the ripening of the seed. On the other hand, a few littaeas of the marginate and filiferous groups are known to be proliferously viviparous at the end of the spike above the latest flowers; Agave attenuata has been shown very recently to be freely bulbiferous sometimes; the lower part of the scape of regularly bulbiferous species like A. barbadensis may bear, in the axils of the bracts, dense masses of unusually shaped bulbils which develop into as unusual plants (pl. C); similar plantlets are often borne on the trunk of polycarpic littaeas or of the magueys when they have been prevented from flowering normally; and it may be unwise to consider the absence of inflorescence bulbils as of constant diagnostic value even in species like A. atrovirens and A. picta, which have not yet been reported to bear them.

When produced, the bulbils of one species are usually somewhat different from those of another, indeed those of widely distinct groups differ greatly; but their differences have been little utilized as yet for the characterization of species. However constant it may be, the

1 Engelmann, Trans. Acad. Sci. St. Louis, vol. 3, 1875, p. 299; Bot. Works, 1887, p- 304.

AGAVE IN THE WEST INDIES—TRELEASE. 15

marginal arming of their leaves is not even indicative of the condition that will prevail on the mature foliage, the essentially smooth-edged and the most