AMERICAN MUSEUM

NorntattesPUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORYCENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024Number 3094, 13 pp., 4 figures, 2 tables May 19, 1994

First Tertiary Land Mammal fromGreater Antilles: An Early Miocene Sloth(Xenarthra, Megalonychidae) From Cuba

R. D. E. MACPHEE' AND M. A. ITURRALDE-VINENT2

ABSTRACT

Parts of the Greater Antilles have been contin-uously subaerial since the beginning of the LateEocene, and therefore have been potentially opento colonization by land vertebrates for more than40 million years. However, documentation for thetiming of Antillean colonization(s) is very poor,partly because of the paucity of fossils. Here wereport the first discovery, in the Greater Antilles,ofskeletal remains ofa land mammal in a securelydated Tertiary context. The remains are those ofa heretofore unknown megalonychid sloth, fromLagunitas Formation sediments at the locality ofDomo de Zaza, province of Sancti Spiritus, south-central Cuba. Lagunitas sediments were depositedin a coastal environmental setting late in the LowerMiocene, perhaps around 17.5-18.5 mya.

This and other recent discoveries establish thatthe Greater Antilles possessed a terrestrial verte-brate fauna by early Neogene, but how this faunawas emplaced (by vicariance or by dispersal) re-mains controversial. One possibility is that em-placement occurred during the Oligocene, whenlowered sea level and the paleogeographical con-figuration of the Greater Antilles may have pro-vided conditions particularly favorable for dis-persal across then-subaerial parts ofthe Aves Ridge.By contrast, patterns of species distribution in theQuaternary could have a strictly vicariant histor-ical explanation, if, as seems likely, they are theresult of Neogene tectonic fragmentation of theGreater Antilles Ridge.

' Chairman and Curator, Department of Mammalogy, American Museum of Natural History; Adjunct Professor,Department of Biology, City College, City University of New York; Adjunct Associate Professor, Department ofAnthropology, State University of New York, Stony Brook.

2 Curator, Museo Nacional de Historia Natural, Academia de Ciencias de Cuba, Capitolio Nacional, Ciudad deLa Habana; Profesor Adjunto, Departmento de Geofisica, Instituto Superior Tecnol6gico "Jose A. Echevarria,"Ciudad de La Habana, Cuba.

Copyright © American Museum of Natural History 1994 ISSN 0003-0082 / Price $2. 10

AMERICAN MUSEUM NOVITATES

RESUMEN

Una parte de las Antillas Mayores ha perma-necido emergida desde el comienzo del EocenoSuperior y por lo tanto ha estado potencialmenteabierta a la colonizacion por los vertebrados te-rrestres durante mas de 40 millones de afios. Sinembargo, la verificacion del fechado de la, o las,colonizaciones es muy pobre, debido a la escasezde restos f6siles. Aqui se reporta el primer des-cubrimiento en las Antillas Mayores, de restos es-queleticos de un mamifero terrestre en un contextoTerciario bien datado. Los restos f6siles corres-ponden con un perezoso megalonychido antes des-conocido, encontrado en los dep6sitos del Mio-ceno Inferior (alrededor de 17.5-18.5 Ma) en lalocalidad que aqui denominamos Domo de Zaza,situada en Cuba sur central (Provincia de SanctiSpiritus).

Este descubrimiento, junto a otros muy recien-

tes, permiten afirmar que las Antillas Mayores po-seian su propia fauna de vertebrados terrestres enel Mioceno Inferior. Pero la manera en que estafauna se establecio (si por vicarianza o dispersion)es un asunto aun controvertido. Una posibilidades que el establecimiento haya ocurrido duranteel Oligoceno, cuando el bajo nivel del mar y laconfiguraci6n geologica de las Antillas Mayorespudieron ofrecer condiciones mas favorables parala dispersion a traves de partes entonces emergidasde la Dorsal de Aves. En contraste con la anterior,los patrones de distribucion de las especies en elCuaternario pueden tener una explicacion estric-tamente por vicarianza historica, si, como todoparece indicar, ellos son el resultado de una frag-mentacion tectonica de las Antillas Mayores du-rante el Mioceno.

INTRODUCTION

The Greater Antilles have had complicatedtectonic, paleogeographic, and faunal histo-ries, the outlines ofwhich are still poorly un-derstood and the subject ofmuch controversy(e.g., Matthew, 1915; Barbour, 1916; Dar-lington, 1938; Rosen, 19753, 1985; Mac-Fadden, 1980; Borhidi and Muiiiz, 1980;Borhidi, 1985; Pregill, 1981; Sykes et al.,1982; Iturralde-Vinent, 1982, 1988; Burke etal., 1984; Buskirk, 1985; Guyer and Savage,1986; Pindell et al., 1988; Kluge, 1988; Wil-liams, 1989; MacPhee and Wyss, 1990;Hedges et al., 1992). One of the least under-stood aspects of biotic history concerns howand when Cenozoic land vertebrates-par-ticularly mammals-became part of the fau-na of these islands. Biogeographers havetended to concentrate on the question ofhowthis occurred because of its relevance to as-sessing whether dispersal or vicariance wasthe major mode of faunal formation in theWest Indies. Knowing the timing of faunalformation, however, is also critically impor-

3 Rosen's paper is frequently cited as "Rosen, 1976"because this is the date that appears at the head of theabstract and agrees with the actual distribution date (6April 1976) ofthe "December 1975" number ofvolume24 of Systematic Zoology. However, for information re-trieval purposes the use of actual time of distributionrather than volume/number is problematic and ought tobe avoided.

tant because time supplies a constraint forlikelihood models of Caribbean biogeogra-phy (see discussion by MacPhee and Wyss,1990). Fossils supply temporal information,but at present the Tertiary paleontologicalrecord for the Greater Antilles is woefullyinadequate for testing purposes (Williams,1989). Any substantial amplification of thisrecord is therefore of great potential impor-tance. In this report we describe the first landmammal recovered so far from a Tertiarycontext within the Greater Antilles-a lateEarly Miocene megalonychid sloth fromDomo de Zaza, south-central Cuba (SanctiSpiritus).

ACKNOWLEDGMENTS

Fieldwork at Domo de Zaza was conductedunder the auspices of an agreement betweenthe Museo Nacional de Historia Natural (LaHabana) and the American Museum of Nat-ural History, and was financed by the PhilipMcKenna Foundation and National ScienceFoundation (BSR 900002). We thank theMNHNH and the Academia de Ciencias deCuba for permission to collect and describethe specimens discussed in this paper.We are grateful to all of the individuals,

Cuban and American, who made our jointventure possible. Our special thanks go to

2 NO. 3094

MAcPHEE AND ITURRALDE-VINENT: TERTIARY CUBAN SLOTH

Gilberto Silva Taboada (MNHNH) and Mi-chael L. Smith (Center for Marine Conser-vation), who labored hardest to create theMNHNH-AMNH agreement; and to Rey-naldo Rojas (MNHNH), Luis Olmo (then ofthe Museo de Historia Natural, Sancti Spi-ritus), Alejandro Emperador (Grupo "Sama,"Sociedad Espeleologica de Cuba, Sancti Spi-ritus), and Michael L. Smith, all of whomhelped with the fieldwork. We also thank A.Areces-Mallea for pollen analysis of Lagu-nitas sediments; S. Algar, 0. Arredondo, T.W. Donnelly, D. Grimaldi, V. MacPhee, R.Rojas, M. L. Smith, S. D. Webb, and E. E.Williams for helpful discussions concerningissues raised in this paper; L. Meeker, E. Heck,and C. Tarka for illustrations; and C. Flem-ming for specimen preparation.

ABBREVIATIONS

a 1/a2/a3 first/second/third molariform (al-veoli)

AMNH American Museum of Natural His-tory

f fossa in advance of alveolus of firstmolariform

Fm formationims intermaxillary sutureMNHNH Museo Nacional de Historia Natural

(La Habana)MNHNH P Paleontological Collection,

MNHNHmya million years agopc palatine canal(s)tims torus on intermaxillary suture

SYSTEMATIC PALEONTOLOGY

Genus and species diagnoses for the newsloth are the same and are therefore efficientlypresented together, as follows:

Order XenarthraInfraorder PhyllophagaFamily Megalonychidae

IMAGOCNUS ZAZAE,NEW GENUS AND SPECIES

HOLOTYPE: Edentulous partial palate,MNHNH P 3014 (fig. 1), found in lag de-posits.ETYMOLOGY: Roots are imago, "image,

likeness" (Latin) + ocnus, "delay, slowness"

(latinized Greek); gender of combination ismasculine. Root "-ocnus" is traditional com-bining form used in propounding names forAntillean megalonychids. Species name is de-rived from Domo de Zaza, the type locality.REFERRED MATERIAL: Molariform tooth,

MNHNH P 3013 (fig. 2); caudal vertebra,MNHNH P 3050; partial pelvis (uppermostpart of sacrum and adjoining ilia), MNHNHP 3021. The pelvis was found in situ withinlagoonal clay resting on sand; the other spec-imens were found in lag deposits.TYPE LOCALITY AND LOCAL FAUNA: Ho-

lotype and referred specimens were collectedat Domo de Zaza (approximately 21°45'N,79030'W, 18 km SE of the town of SanctiSpiritus) in lagoonal/near-shore sediments ofLagunitas Fm (see Study Area). Associatedvertebrate fauna, as currently known, in-cludes an unidentified bird, pelomedusid tur-tles, crocodiles, sharks, sirenians, and a smallodontocete. These fossils will be described ina separate report (MacPhee et al., in prep.).AGE: Late Early Miocene, based on inver-

tebrate assemblage recovered from relatedmarine facies (see Study Area).

DIAGNOSIS: The type species (and onlyknown member) of genus Imagocnus is I.zazae. The holotype palate ofL zazae is dis-tinctively slothlike, being narrow but thick inprofile, and bearing subparallel rows of large,closely spaced alveoli (fig. 1). Subparallel, an-teriorly diverging cheektooth rows are con-sidered derived within Phyllophaga by us,but are not exclusive to Megalonychidae (cf.Fischer, 1971; Engelmann, 1985). However,attribution to this family is warranted be-cause the partly preserved fossae in front ofthe first molariform are steep-walled and in-wardly concave, unlike anything seen inmegatheriids and mylodontids. (These com-ments assume that these conventional slothfamilies are monophyletic; on this point seeEngelmann, 1985.) We do not wish to baseany part of this diagnosis on inferential evi-dence, but for the record we note that we alsorecovered what appears to be a natural castof a megalonychid upper caniniform, in abiocalcarenite near the top of the sequenceat Domo de Zaza. The "tooth" is quasitrian-gular in section-an earmark of megalony-chid caniniforms (Webb and Perrigo, 1985).However, the cast (if that is what it is) is

31 994

AMERICAN MUSEUM NOVITATES

Fig. 1. The holotype palate (MNHNH P 3014) of Imagocnus zazae in dorsal (A), right lateral (B),ventral (C), and rostral (D) aspects. Same scale as in figure 2.

imperfect on one side, and, as no part of theoriginal object remains, we cannot respon-sibly include this specimen in the hypodigm.(However, for future reference the specimenis cataloged as MNHNH P 3015.)

Imagocnus zazae differs from all adequate-ly described Antillean sloths except speciesofAcratocnus and Mesocnus in lacking large,paired palatal palatine foramina at the trans-verse level of the first molariform teeth. (Pal-

4 NO. 3094

MACPHEE AND ITURRALDE-VINENT: TERTIARY CUBAN SLOTH

atal palatine foramina may have opened at amore rostral transverse level, or, ifthere wereno discrete apertures, vasculature and nervespassed through the host of "nutrient" foram-ina perforating the ventral surface.) Holotypepalate differs from equivalent portion of pal-ate ofAcratocnus odontrigonus (and A. major,if distinct) in being much larger, possessingdistinct midsagittal torus on intermaxillarysuture, and lacking sharp medial borders onalveoli. Differs from palate of Mesocnusbrowni (and M. torrei, if distinct) in beinglarger and in having greater interalveolarbreadth at the level of first molariform dueto slight rostral divergence of toothrows.

Additional information concerning thesystematic position of Imagocnus can begleaned from MNHNH P 3013 (fig. 2). Thisreferred tooth is fragmentary, and both theocclusal surface and root end are missing.However, it is a molariform of a megalony-chid sloth because of its shape (reniform insection) and composition (thick core of vas-cular dentine successively enwrapped bymuch thinner shells of compact dentine andcementum). In mylodontids, tooth shape ismore lobular and the layer of compact den-tine tends to be relatively thicker, whilemegatheriids typically have lophodont, dis-tinctly prismatic molariforms (Fischer, 1971;Englemann, 1978, 1985). In morphology thetooth closely resembles mandibular molari-forms of the Quaternary Cuban species Mes-ocnus torrei, but in size it is intermediate be-tween this species and Megalocnus rodens,the largest Antillean Quaternary sloth (Fi-scher, 1971). Other Zaza specimens referableto I. zazae-a caudal vertebra and a frag-mentary pelvis-are seriously incomplete andin any case do not provide characters usefulfor differential diagnosis.

DISCUSSION

STUDY AREA

Rio Zaza (fig. 3) drains the eastern andsouthern flanks ofhighland masses in centralCuba (Alturas de Santa Clara, del Nordestede Las Villas, and de Sancti Spiritus). In re-cent years outflow through the lower Zazabasin has been substantially modified throughthe construction of a dam (Presa de Zaza) onthis river and diversion of water through anewly built artificial channel, the Canal de

Fig. 2. Presumed lower molariform (MNHNHP 3013; 17.0 x 10.4 mm in section) referred toImagocnus zazae, in cross-sectional (A) and lateral(B) aspects. Occlusal surface and root end are bothmissing. In morphology the tooth closely resem-bles mandibular molariforms of the QuaternaryCuban species Mesocnus torrei, but in size it isintermediate between this species and Megalocnusrodens, the largest Antillean Quatemary sloth.

Zaza. In order to direct the canal toward theirrigation works that it was designed to serve,it was necessary to transect a low, domelikehill (Domo de Zaza) located about 1 km S ofthe dam site and 18 km SE of the city ofSancti Spiritus. This transection provides oneof the very few glimpses of the underlyingstratigraphy ofthis region, which is otherwisealmost devoid of relief and thickly coveredby late Pleistocene-Recent sediments (Mac-Phee et al., in prep.). Unexpectedly, this cutrevealed the first mid-Tertiary terrestrial de-positional environments so far identified incentral Cuba, the first fossil land vertebratefrom which is described in this report. Thelocality consists of approximately 3 km ofexposures, on both sides ofthe water channel,with a maximum exposure of 10-15 m.The gravels, sands, clays, and rare lime-

stone intercalations exposed on channel side-walls at Domo de Zaza are referable to theLagunitas Fm, a unit defined by E. Popov in

1 994 5

AMERICAN MUSEUM NOVITATES

b~~~~~eCarl sea

0 3 6 9Km

Fig. 3. Location map for Domo de Zaza, western part of Lagunitas Fm (shaded), and place namesmentioned in text.

an unpublished geological survey conductedas part of the 1:250,000 scale mapping pro-ject of the former Las Villas province(Kantchev et al., MS). This formation, thewestern part of which is depicted in figure 3,will be more fully described in a related paper(MacPhee et al., in prep.). Here it is sufficientto note that the unit was deposited in a coastalplain environment, where conditions variedfrom terrestrial to shallow-water marine. Thesection exposed at Domo de Zaza is laterallyand vertically complex, showing evidence ofat least four different depositional regimeswithin the limited area of exposure. The se-quence as a whole is composed of severaltransgressive cycles, beginning with lagoonalclays at base with thin limestones locally, andending with a thick and ubiquitous limestonecap. However, within cycles there were clear-ly short phases of uplift. Alluvial deposits,soil development, and land-plant pollen (A.Areces-Mallea, personal commun.) occur interrestrial facies, providing clear evidence ofthe periodic presence of emergent land.The youngest lithostratigraphic unit over-

lain by Lagunitas Fm is Oligocene Las Cue-

vas Fm, in the type area near Trinidad (fig.3). Lagunitas is in turn overlain by MiddleMiocene Giiines Fm and younger sediment.We note that the invertebrate faunal assem-blage recovered from marine facies of La-gunitas Fm (table 1) corresponds in generalcontent to the Lower Miocene Soritiidae-Miogypsina zone of Iturralde-Vinent (1969,1971), because both nominate taxa ofthe zoneare present in association with Heterosteginaantillea. However, the earliest part of theMiocene is evidently not represented at La-gunitas, because the marker taxon Lepido-cyclina is absent while Sorites and Archaiasare present (last two taxa are never found nearOligocene/Miocene boundary or in associa-tion with Lepidocyclina). Additionally, thepresence of Procythereis cf. deformis indi-cates that the Lagunitas is younger than zoneN5 (Catapsidrax dissimilis zone), and prob-ably fits biochronologically within N6-N7(after last appearance of C. dissimilis, beforefirst appearance of Orbulina). This infor-mation is consistent with Burdigalian age (ap-proximately 16.9-19 mya). Furthermore, be-cause Lagunitas is located at the base of a

NO. 30946

MACPHEE AND ITURRALDE-VINENT: TERTIARY CUBAN SLOTH

thick Lower through Middle Miocene section(as represented by Lagunitas-Giiines-ArabosFms) in a transgressive interval, correlationwith the 18.5-17.5 mya coastal onlap eventofHaq et al. (1987) is highly likely. Elsewherein Cuba, age and lithological partial equiv-alents of Lagunitas Fm have been describedas Paso Real, Arabos, Rosario, and Magan-tilla Fms (Iturralde-Vinent 1969; Iturralde-Vinent et al., 1987).For much of the Tertiary, Cuba consisted

of three main archipelagos distributed alongan approximately east-west axis (Iturralde-Vinent, 1988: fig. 4). The middle group ofislands was centered on the Cordillera delEscambray, existing as a high massif at leastsince the Late Paleogene (Iturralde-Vinent,1982, 1988). We envisage the Escambray, inEarly Miocene, as having formed a small buthigh island that was ringed by periodicallyinundated keys in shallow seas. The riversdraining the highlands would have formedextensive deltaic deposits in their lowerreaches (future Lagunitas and related for-mations). During major regressions (lateMiddle Eocene, Late Oligocene, Late Mio-cene, and Plio-Pleistocene) these archipela-gos were temporarily united; they becamepermanently so at the end of the Pliocene(Iturralde-Vinent, 1988). From a biogeo-graphical standpoint, these periods ofmarkedregression are of critical significance, becausethey would have both increased the total landarea available for colonization and reduceddistances between islands (see fig. 4).

SYSTEMATIC POSITION OF IMAGOCNUS

Although the material available for the di-agnosis of Imagocnus zazae is limited, thebasis for its allocation to Phyllophaga seemsindisputable. A fragmentary, toothless palateof a very large caviomorph might be mo-mentarily confused with that of a sloth, be-cause caviomorphs also possess narrow in-teralveolar spaces defined by subparallel rowsof closely spaced alveoli. However, cavio-morph palates are never as thick as those ofsloths, and moreover they lack fossae in ad-vance of the anteriormost tooth.

Allocation to Megalonychidae also seemsbeyond question, for biogeographical as wellas morphological reasons. Megalonychids are

TABLE 1Invertebrate Fauna, Lagunitas Formationa

Foraminifera Ammonia beccariiAmphistegina aff. A. taberanaAmphistegina angulataArchaias aduncusArchaias angulatusDiscorbis sp.Elphidium discoidaleElphidium nautiloideumElphidium sagreiHeterostegina antilleaMiogypsina antilleaPyrgo sp.Sorites marginalis

Ostracoda Aurila aff. A. cicatricosaBairdia antilleaBairdia bradyiCaudites sp.Costa (Rectotrachileberis) sp.Cytheropteron sp.Haplocytheridea cubensisHaplocytheridea sp.Loxoconcha antillea nodosaLoxoconcha cubensisProcythereis cf. deformisQuadricythereis antillea

Mollusca Crassostrea vaughani insularisGlycymeris acuticostatus aff. G. a.guaica

Glycymeris jamaicanusa As reported for type section by Popov (in Kantshev

et al., MS); for biochronological significance for Zaza ver-tebrate fauna, see text.

the only sloths known to have entered theWest Indies, where they were widespread anddiverse prior to their extinction in the lateQuaternary (e.g., Arredondo, 1961; Mayo,1969; Fischer, 1971). Their known distri-bution included Cuba, Hispaniola, PuertoRico, and small islands close to northernSouth America (Anthony, 1918; Hoffstetter,1955; Hooijer, 1962; White, 1992). A newPliocene record for Antillean sloths, for Gre-nada, has not yet been formally published butseems secure (R. Singer, personal commun.).

Megalonychids, both continental and An-tillean, are in need of cladistic analysis (cf.Webb and Perrigo, 1985) and there is littlepoint in trying to specify the sister taxon ofImagocnus until revisionary treatments ofthisputatively monophyletic family have beenundertaken. For the present it may be notedthat strong torus development and mesialplacement of the palatine foramina are alsoseen in various Miocene Patagonian mega-

1994 7

AMERICAN MUSEUM NOVITATES

Fig. 4. Paleogeographic maps of Caribbeanarea, 38-16 mya, based mainly on Ross and Scotese(1988) with various modifications (cf. Iturralde-Vinent, 1988). Latest Eocene: Cuba (designated asC) has been approximately stationary since "dock-ing" against Bahamas Platform (BP) in late MiddleEocene. For much of the Tertiary, Cuba consistedof several archipelagos rather than a single large

lonychids with suspected ties to better-knownAcratocnus and Mesocnus (Pascual et al.,1990). On the basis of postcranial features,White (1992, 1993) has recently argued thatAntillean sloths may be most closely relatedto Santacrucian Eucholoeops. Megalonychidsentered North America well before the com-pletion of the Panamanian landbridge ca. 3mya (Webb and Perrigo, 1985), but at presentthere is no indication that they did so via the

landmass. To the southeast, the Greater Antillesmagmatic arc (future Hispaniola [H], Puerto Rico[PR], and Virgin Islands [VI]) is continued south-ward by another arc, the Aves Ridge (AR), in col-lision with South America. Yucatan Basin (YB) isalready in existence, as is Straits of Florida (SF).Oligocene: Grenada Basin, probably formed byintra-arc spreading, began to split old Aves Ridgefrom new Lesser Antilles Arc (LSA) after the closeof the Eocene. Two important events occurred af-ter termination of island arc volcanic activity atend of Eocene: (1) substantial tectonic uplift tookplace along the Greater Antilles Ridge and AvesRidge, and (2) sea level dropped - 160 m at - 30mya. Substantial emergence of land along theseridges, indicated by latest Eocene-Oligocene de-positional hiatuses in many West Indian strati-graphic sequences (Iturralde-Vinent, 1988; Maur-rasse, 1990), was probably sufficient to createcontinuous land or groups of very closely spacedislands from western Cuba through to northernSouth America. However, they were insufficientto create connectors between the Greater Antillesand the Beata Rise (BR) or Cozumel-Yucatan Ridge(CY). Conditions for overland dispersion to theGreater Antilles from South America were lesspropitious by Late Oligocene (27 mya) because ofrising sea level. Early/Middle Miocene: Topo-graphic highs in various parts offuture Cuba, His-paniola, and Puerto Rico remained subaerial forrest of Neogene. Aves Ridge eventually subsides600-1200 m (Holcombe and Edgar, 1990), leavingAves Island (Al) and Venezuelan Leeward Islands(LEA) as sole subaerial portions. Volcanic activitycontinues in parts ofLesser Antilles Arc until pres-ent. Reorganization of stress fields within Carib-bean Plate, beginning in late Paleogene, had bythis time caused significant disruption of formerlycompact Greater Antilles Ridge along left lateralfaults. Lateral displacements of major blocks andcreation of concomitant deep water gaps (e.g.,Windward Passage between Cuba and Hispaniola,along rejuvenated Oriental Fault [OF]) led to vi-cariant separations of formerly continuous pop-ulations.

Greater Antilles, or that Antillean sloths sharea more recent ancestry with North Americanthan with South American taxa. White (1993)presented some evidence for the propositionthat two different clades of megalonychidscolonized the northern Greater Antilles, al-though the difference in tree length betweenthis solution and slightly less parsimoniousones providing for a single initiator lineageis small.

8 NO. 3094

MACPHEE AND ITURRALDE-VINENT: TERTIARY CUBAN SLOTH

TABLE 2Tertiary Freshwater and Terrestrial Vertebrates of the Greater Antillesa

Island Age/formation/locality Reference

CubabImagocnus zazae (Megalonychidae, Xe- E. Miocene; Lagunitas Fm, Domo de Zaza, S. This paper

narthra) SpiritusUnidentified bird E. Miocene; Lagunitas Fm, Domo de Zaza, S. MacPhee et al., in

Spiritus prep.

Hispaniolac,dCichlasoma woodringi (Cichlidae, Perci- ?M. Miocene; Las Cahobas Fm, Mirebalais Cockerell, 1924

formes)Anolis dominicanus (Iguanidae, Sauria) ?Miocene; amber deposits, Dominican Rieppel, 1980

RepublicSphaerodactylus dommeli (Gekkonidae, ?Miocene; amber deposits, Dominican Re- B6hme, 1984

Sauria) publicEleutherodactylus sp. (Leptodactylidae, ?Late Eocene; amber deposits, Dominican Poinar and Cannatel-

Anura) Republic la, 1987Unidentified mammal (?rodent) ?Miocene; amber deposits, Dominican Re- Poinar, 1988

public

Puerto Ricob,eBoid (Ophida) E. Miocene; Cibao Fm, Aguadilla MacPhee and Wyss,

1990?Iguanid (Sauria) E. Miocene; Cibao Fm, Aguadilla MacPhee and Wyss,

1990a Sloth and caviomorph remains of Pliocene age from Grenada (R. Singer, personal commun.) are the only known

Tertiary land vertebrate fossils from the Lesser Antilles.b Miocene pelomedusid turtles from the Greater Antilles, known from Puerto Rico (Wood and Gaffney, 1989) and

Cuba (MacPhee et al., in prep.) are possibly but not certainly freshwater. They are therefore omitted from this list.c Some reported ages of Dominican amber deposits are controversial; listed ages follow authors' reports.d Additional vertebrate inclusions in Dominican amber have been discovered in recent years (including additional

examples of embedded hair, feathers) and are in the process of being described by various workers.e Age of echimyid rodent Puertoricomys corozalus from Corozal is undetermined (MacPhee and Wyss, 1990),

although Woods (1989) regarded it as possibly Tertiary.

BIOGEOGRAPHICAL SIGNIFICANCE OFIMAGOCNUS

The discovery of Imagocnus zazae marksan important milestone in efforts to recoverthe enigmatic biotic history of the GreaterAntilles, but its significance needs to bejudgedin relation to other discoveries. Like Ima-gocnus, most ofthe small number ofTertiaryland vertebrates known from the Greater An-tilles are dated to Miocene (table 2). How-ever, parts ofCuba and Hispaniola have beencontinuously subaerial since the beginning ofthe Late Eocene (Iturralde-Vinent, 1982,1988), and palynological evidence (Grahamand Jarzen, 1969) indicates that terrestrialenvironments with moderate plant diversitywere already in existence in Puerto Rico by

Late Oligocene. Land vertebrates couldtherefore be older than the Early Mioceneminimum date supplied by the majority offossils. (The earliest attributed date for a ver-tebrate, ?Late Eocene for a frog referable tothe extant Neotropical genus Eleutherodac-tylus [Poinar and Cannatella, 1987], is olderthan generally accepted dates for Dominicanamber and needs to be confirmed.)Some recent authors (e.g., Poinar and Can-

natella, 1987) have argued that proof of theexistence of land vertebrates in the GreaterAntilles during the mid-Tertiary tends tosupport vicariist rather than dispersalist in-terpretations of Caribbean historical bioge-ography. However, all taxonomically iden-tifiable land vertebrate taxa (includingImagocnus) can be comfortably nested, at

1 994 9

AMERICAN MUSEUM NOVITATES

relatively low hierarchical levels, withinclades that persisted into the Quaternary.(Specimens for which no reliable lower-levelplacement has yet been made, like the ?ro-dent hair reported by Poinar [1988] or the?iguanid vertebra described by MacPhee andWyss [1990], have no bearing on this obser-vation.) In lacking any "unexpected" verte-brate taxa-that is, groups representing ma-jor clades not known to have existed on theislands in the Quaternary (MacPhee andWyss, 1990)-the paleontological record canbe interpreted to mean that Antillean diver-sity has always been limited to a few groups(Williams, 1989; Woods, 1989). Yet the fail-ure to find "unexpected" taxa may simplymean that an originally more diverse faunahad already been reduced by the start of theNeogene, when the fossil record begins, togroups that would survive into the Quater-nary. Or, alternatively, it may indicate thatcollection efforts have been inadequate to givea fair picture of previous diversity (Mayerand Lazell, 1988; Williams, 1989; MacPheeand Wyss, 1990). Several examples of "un-expected" terrestrial arthropods have beenrecovered from ?Miocene Dominican amber(Wilson, 1988), proving that mid-Tertiary di-versity for some groups was greater than (orat least different from) their Quaternary di-versity.

Despite arguments to the contrary, hy-potheses of vicariance in the strict sense pro-posed by Rosen (1975, 1985) are probablyno more readily falsifiable than ones invokingdispersal as historical explanations for bioticdistributions. At this stage of knowledge it ispossibly more productive to ask whether therewas any sort of "common cause" that couldhave made Antillean colonization (howeveraccomplished) easier for land vertebrates atone time rather than another. Various linesofevidence indicate that a possible optimumfor colonization occurred between the end ofthe Middle Eocene and the beginning of theLate Oligocene (fig. 4). During this intervalthere was diachronous compression along thenorthern and southern Caribbean Plateboundaries (Pindell and Barrett, 1990), sub-stantial terminal Palaeogene uplift along theaxes ofthe Greater Antillean and Aves Ridg-

es (Holcombe et al., 1990; Maurrasse, 1990),and a drop in eustatic sea-level of 160 m, ca.30 mya (Haq et al., 1987; but see Hallam,1992). According to Algar and Erikson (1992),progressive Palaeogene overthrusting by theeastward-moving Caribbean Plate created aforebulge in northern South America close tothe southernmost part of the Aves Ridge arc(future Venezuelan Antillas Menores). Thisis consistent with the evidence discussed bySpeed et al. (1993) for an Oligocene collisionbetween the South American continentalslope and the Paleogene magmatic arc of thesouthern Lesser Antilles, followed by upliftin Early Miocene. This event may have pro-duced the continental terminus of a con-nector (realized as closely spaced islands, longpeninsula, or actual landbridge) involving theAves Ridge and thence the Greater Antilles(cf. Holcombe and Edgar, 1990), especiallyduring times of low sea level. How long con-ditions favorable for the movement of landvertebrates might have persisted is of courseunknown, but it was surely not very long.By contrast, vergence between Cuba and

Florida never closed the Straits of Florida,and the Yucatan Channel has existed in someform since the Cretaceous. Long persistenceofthese marine barriers offers an explanationfor the low number of apparently NorthAmerican elements in the mammalian fau-na-just two taxa, the insectivores Solenodonand Nesophontes (MacFadden, 1980). (Sig-nificant presence of Central American ele-ments in the herpetofauna [Guyer and Sav-age, 1986] may imply a different temporaloptimum for this group.) Immunological dis-tance (ID) measurements for a number ofreptile and amphibian groups display widescattering when calibrated against a molec-ular clock model (Hedges et al., 1992). Scat-tering, with estimated divergences as late asthe Pleistocene in some cases, is said to falsifyRosen's (1975) Caribbean vicariance hypoth-esis while favoring dispersal as the overallmethod of faunal formation in the West In-dies. However, these results do not precludethe scenario offered here, because we inferthat faunal emplacement was a mid-Tertiaryevent, not terminal Cretaceous as envisagedby Rosen. (For a discussion of the inappro-priateness of ID measurements for correlat-

NO. 309410

MAcPHEE AND ITURRALDE-VINENT: TERTIARY CUBAN SLOTH

ing events in earth history and phylogeny, seePage and Lydeard, in press.)Congruence among Quaternary faunal

suites ofthe major islands suggests the impactof another common cause. We assume thatancestors of the known Quaternary rodents,primates, sloths, and insectivores were inplace on landmasses in the Caribbean by themid-Tertiary. It was at this time that reor-ganization of stress fields within the Carib-bean Plate began to progressively disrupt theoriginally more compact Greater AntillesRidges by horizontally displacing large blocks(islands) along left lateral faults (Mann andBurke, 1984; Pindell and Barrett, 1990; Hol-combe et al., 1990; Ross and Scotese, 1988;

Iturralde-Vinent, 1988). Concomitant crea-tion of deep water gaps between islands (e.g.,Windward Passage between Cuba and His-paniola) could have produced true vicariantseparation of populations that were formerlycontinuous. Any compositional differencesbetween individual island faunas (e.g., ap-parent absence of heptaxodontid rodents inCuba, platyrrhine monkeys in Puerto Rico)may thus be evidence of postvicariance localextinction, not failed dispersal. Losses, iftheyoccurred, are potentially demonstrable withfossil evidence, illustrating how paleontologymay play yet another role in resolving leadingproblems in Caribbean biogeography.

REFERENCES

Algar, S., and J. Erikson1992. Correlation of the pre-Miocene stratig-

raphies of Trinidad and eastern Vene-zuela: Cretaceous passive margin to Oli-go-Miocene foreland basin. Progr.Abstr., 13th Carib. Geol. Conf., p. 64.

Anthony, H. E.1918. The indigenous land mammals ofPorto

Rico, living and extinct. Mem. Am. Mus.Nat. Hist. 2: 329-435.

Arredondo, 0.1961. Descripciones preliminares de dos nue-

vos generos y especies de edentados delPleistoceno cubano. Bol. Grupo Explor.Cien. 1: 19-40.

Barbour, T.1916. Some remarks upon Matthew's "Cli-

mate and Evolution." Ann. New YorkAcad. Sci. 27: 1-15.

B6hme, W.1984. Erstfund eines fossilen Kugelfinger-

gecko (Sauria: Gekkonidae: Sphaero-dactylinae) aus Dominikanischen Bern-stein der Oligozan von Hispaniola,Antillen. Salamandra 20: 212-220.

Borhidi, A.1985. Phytogeographic survey ofCuba. 1. The

phytogeographic characteristics andevolution ofthe flora ofCuba. Acta Bot.Hungarica 31: 3-34.

Borhidi, A., and 0. Muniiz1980. Die vegetationskarte von Kuba. Acta

Bot. Hungarica 26: 25-53.Burke, K., C. Cooper, J. F. Dewey, P. Mann, and

J. L. Pindell1984. Caribbean tectonics and relative plate

motions. Mem. Geol. Soc. Am. 162: 31-63.

Buskirk, R. E.1985. Zoogeographic patterns and tectonic

history ofJamaica and the northern Ca-ribbean. J. Biogeogr. 12: 445-461.

Cockerell, T. D. A.1924. A fossil cichlid from the Republic of

Haiti. Proc. U.S. Nat. Mus. 63(7): 1-2.Darlington, P. J.

1938. The origin of the fauna of the GreaterAntilles, with discussion of dispersal ofanimals over water and through the air.Q. Rev. Biol. 13: 274-300.

Engelmann, G.1978. The logic of phylogenetic analysis and

the phylogeny of the Xenarthra (Mam-malia). Unpubl. Ph.D. diss., ColumbiaUniv., New York, 329 pp.

1985. The phylogeny of the Xenarthra. In G.G. Montgomery (ed.), The evolution andecology of armadillos, sloths, and ver-milinguas, pp. 51-64. Washington:Smithsonian Institution.

Fischer, K.1971. Reisenfaultiere (Megalonychidae,

Edentata, Mammalia) aus dem Pleis-toziin der Pio-Domingo-Hohle in Kuba.Wiss. Z. Humboldt-Univ. Berlin, Math.-Nat. R. 20: 609-674.

Graham, A., and D. M. Jarzen1969. Studies in neotropical paleobotany. 1.

The Oligocene communities of PuertoRico. Ann. Missouri Bot. Gard. 56: 308-357.

1994 I1I

AMERICAN MUSEUM NOVITATES

Guyer, G., and J. M. Savage1986. Cladistic relationships among anoles

(Sauria: Iguanidae). Syst. Zool. 35: 509-531.

Hallam, A.1992. Phanerozoic sea-level changes. New

York: Columbia Univ. Press.Haq, B. U., J. Hardenbol, and P. R. Vail

1987. Chronology offluctuating sea levels sincethe Triassic. Science 235: 1156-1166.

Hedges, S. B., C. A. Hass, L. R. Maxson1992. Caribbean biogeography: molecular ev-

idence for dispersal in West Indian ter-restrial vertebrates. Proc. Natl. Acad.Sci. 89: 1909-1913.

Hoffstetter, R.1955. Un megalonychide (edente gravigrade)

fossile de Saint-Domingue (Ile d'Haiti).Bull. Mus. Nat. Hist. Nat., Paris (ser. 2)27: 100-104.

Holcombe, T. L., and N. T. Edgar1990. Late Cretaceous and Cenozoic evolu-

tion of Caribbean ridges and rises withspecial reference to paleogeography. InA. Azzaroli (ed.), Biogeographical as-pects ofinsularity, Atti Convegni Lincei85: 611-626. Rome: Accademia Na-zionale dei Lincei.

Holcombe, T. L., J. W. Ladd, G. Westbrook, N.T. Edgar, and C. L. Bowland1990. Caribbean marine geology; ridges and

basins ofthe plate interior. In G. Dengoand J. E. Case (eds.), The geology ofNorth America, vol. H, The Caribbeanregion, pp.231-260. Boulder, CO: Geo-logical Society of America.

Hooijer, D. A.1962. A fossil ground sloth from Cura9ao,

Netherlands Antilles. Proc. K. Nederl.Akad. Wetensch. (ser. B) 65: 46-60.

Iturralde-Vinent, M. A.1969. Principal characteristics ofCuban Neo-

gene stratigraphy. Bull. Am. Assoc. Pet-rol. Geol. 53: 1938-1955.

1971. Correlacion estratigrafica del Neogenode Cuba. Revista Tecnol6gica 9(1): 15-19.

1982. Aspectos geol6gicos de la biogeografiade Cuba. Rev. Cien. Tierra Espac. 5: 85-100.

1988. Naturaleza geologica de Cuba. C. de LaHabana: Editorial Cientifico-Tecnica,248 pp.

Iturralde-Vinent, M. A., A. De la Torre, and M.Quintana1987. Rocas magnesianas sedimentarias en el

Mioceno de Camagiiey. Metodos deprospeccion. Rev. Tecnol. 18(3): 4-9.

Kantchev, I., et al. (eds.)MS Resultados del levantamiento geologico

y las investigaciones cientificas en laprovincia de Las Villas. Unpublishedreport with maps on file with Academiade Ciencias de Cuba, C. de La Habana,1976.

Kluge, A.1988. Parsimony in vicariance biogeography:

a quantitative method and a GreaterAntillean example. Syst. Zool. 37: 315-328.

MacFadden, B. J.1980. Rafting mammals or drifting islands?:

biogeography of the Greater Antilleaninsectivores Nesophontes and Soleno-don. J. Biogeogr. 7: 11-22.

MacPhee, R. D. E., and A. R. Wyss1990. Oligo-Miocene vertebrates from Puerto

Rico, with a catalog of localities. Am.Mus. Novitates 2965: 45 pp.

Mann, P., and K. Burke1984. Neotectonics of the Caribbean. Rev.

Geophys. Space Phys. 22: 309-362.Matthew, W. D.

1915. Climate and evolution. Ann. New YorkAcad. Sci. 24: 171-213.

Maurrasse, F. J. M. R.1990. Stratigraphic correlation for the circum-

Caribbean region. In G. Dengo and J.E. Case (eds.), The geology of NorthAmerica, vol. H, The Caribbean region,plates 4, Sa, 5b. Boulder, CO: Geolog-ical Society of America.

Mayer, G. C., and J. D. Lazell1988. Significance of frog in amber. Science

239: 1477-1478.Mayo, N. A.

1969. Nueva especie de Megalonychidae ydescripcion de los depositos cuaterna-rios de la cueva del Vaho, Boca de Ja-ruco, La Habana. Mem. Fac. Cien. Univ.Habana (ser. Cien. Biol.) 3: 1-58.

Page, R. D. M., and C. LydeardIn press. Towards a cladistic biogeography of

the Caribbean. Cladistics.Pascual, R., M. G. Vucetich, and G. J. Scillato-Yane1990. Extinct and Recent South American and

Caribbean Megalonychidae, edentates,and hystricomorphous rodents: out-standing examples of isolation. In A.Azzaroli (ed.), Biogeographical aspectsof insularity, Atti Convegni Lincei 85:627-640. Rome: Accademia Nazionaledei Lincei.

Pindell, J. L., and S. F. Barrett1990. Geological evolution of the Caribbean

12 NO. 3094

MACPHEE AND ITURRALDE-VINENT: TERTIARY CUBAN SLOTH

region: a plate-tectonic perspective. InG. Dengo and J. E. Case (eds.), The ge-ology of North America, vol. H, TheCaribbean region, pp. 405-432. Boul-der, CO: Geological Society ofAmerica.

Pindell, J. L., S. C. Cande, W. C. Pitman, D. R.Rowley, J. F. Dewey, J. Labreque, and W. Hax-by1988. A plate-tectonic framework for models

of Caribbean evolution. Tectonophys-ics 155: 121-138.

Poinar, G. 0.1988. Hair in Dominican amber: evidence for

Tertiary land mammals in the Antilles.Experientia 44: 88-89.

Poinar, G. O., and D. C. Cannatella1987. An Upper Eocene frog from the Do-

minican Republic and its implicationsfor Caribbean biogeography. Science237: 1215-1216.

Pregill, G. K.1981. An appraisal of the vicariance hypoth-

esis of Caribbean biogeography and itsapplication to West Indian terrestrialvertebrates. Syst. Zool. 30: 147-155.

Rieppel, 0.1980. Green anole in Dominican amber. Na-

ture 286: 486-487.Rosen, D. E.

1975. A vicariance model of Caribbean bio-geography. Syst. Zool. 24: 431-464.

1985. Geological hierarchies and biogeo-graphical congruence in the Caribbean.Ann. Missouri Bot. Gard. 72: 636-659.

Ross, M. I., and C. R. Scotese1988. A hierarchical tectonic model ofthe Gulf

of Mexico and Caribbean region. Tec-tonophysics 155: 139-168.

Speed, R. C., P. L. Smith-Horowitz, K. Perch-Nielsen, J. B. Saunders, and A. B. Sanfilippo1993. Southern Lesser Antilles arc platform:

pre-Late Miocene stratigraphy, struc-ture, and tectonic evolution. Geol. Soc.Am. Spec. Pap. 277: 1-98.

Sykes, L. R., W. R. McCann, and A. L. Kafka1982. Motion of Caribbean plate during last

7 million years and implications for ear-lier Cenozoic movements. J. Geophys.Res. 87: 10,656-10,676.

Webb, S. D., and S. Perrigo1985. New megalonychid sloths from El Sal-

vador. In G. G. Montgomery (ed.), Theevolution and ecology of armadillos,sloths, and vermilinguas, pp. 113-120.Washington: Smithsonian Institution.

White, J. L.1992. Functional and phylogenetic implica-

tions ofthe postcranial skeleton offossilsloths for the evolution ofarboreality intree sloths. Unpubl. Ph.D. diss., StateUniv. New York, Stony Brook.

1993. Phylogenetic implications of the post-cranial skeleton in sloths. J. Vert. Pa-leontol. 13(Suppl.): 62A.

Williams, E. E.1989. Old problems and new opportunities in

West Indian biogeography. In C. A.Woods (ed.), Biogeography of the WestIndies, pp. 1-46. Gainesville, FL: Sand-hill Crane Press.

Wilson, E. 0.1988. The biogeography of the West Indian

ants (Hymenoptera: Formicidae). In J.K. Leibherr (ed.), Zoogeography of Ca-ribbean insects, pp. 214-230. Ithaca,NY: Cornell.

Wood, R. C., and E. S. Gaffney1989. New fossil pelomedusid (side-necked)

turtle remains from the Oligocene ofPuerto Rico. J. Vert. Paleontol.9(Suppl.): 46A.

Woods, C. A.1989. A new capromyid rodent from Haiti:

the origin, evolution, and extinction ofWest Indian rodents and their bearingon the origin of New World hystricog-naths. In C. C. Black and M. R. Dawson(eds.), Papers on fossil rodents in honorof Albert Elmer Wood, pp. 59-90. LosAngeles: Natural History Museum ofLos Angeles County.

1 994 13

0 This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of theNovitates, Bulletin, and Anthropological Papers published during the last five years are availablefree of charge. Address orders to: American Museum of Natural History Library, Department D,Central Park West at 79th St., New York, N.Y. 10024.