AMERICAN MUSEUM Novitates PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 3141, 31 pp., 11 figures, 3 tables June 30, 1995

Orgin of the Greater Antillean Land Mammal Fauna, 1: New Tertiary Fossils from Cuba and Puerto Rico

R. D. E. MAcPHEE' AND MANUEL A. ITURRALDE-VINENT2

ABSTRACT

A new isolobodontine capromyid rodent (Za- Oligocene boundary. Underthe terms ofthis mod- zamys veronicae, new genus and species) and a new el it is not expected that any land mammal faunal platyrrhine primate (unnamed genus and species elements of greater than Late Eocene age will be A) are described from material recovered at Domo found in these islands. It is reasonable to infer that de Zaza, an important late Early Miocene locality Tertiary Antillean mammal groups were widely in south-central Cuba near the town ofSancti Spir- distributed throughout this terrain, and the dis- itus. Also described is a femur believed to repre- covery of a biogeographical "ghost" (Zazamys) sent a small megalonychid sloth (unnamed genus supports this. and species B), from a well-dated Early Oligocene Emergent parts of the present-day Greater An- site near the town ofYauco in southwestern Puerto tilles were in a close-packed array until the com- Rico. mencement of the Neogene. They formed part of These and other discoveries in the Greater An- a long-lasting positive topographic feature of the tilles establish that these islands supported a di- Caribbean Basin, here dubbed GAARlandia. Dur- verse land mammal fauna from the Oligocene on- ing the Late Eocene-Oligocene, this feature in- ward. Pertinent geological evidence is consistent cluded Cuba, northern Hispaniola, Puerto Rico, with the proposition that the initiators of all or Virgin Islands, Aves Rise, and the Venezuelan An- practically all founder clades of Antillean land tillas Menores. Since latest Oligocene, GAARian- mammals were emplaced there near the Eocene- dia has undergone disruption and fragmentation

' Chairman and Curator, Department of Mammalogy, American Museum of Natural History. 2 Curator, Geology and Paleontology Group, Museo Nacional de HistoriaNatural, Capitolio Nacional, CH- 10200, La Habana, Cuba.

into its modem insular components, a process that mammals throughout the later Cenozoic, and this forced vicariance ofthe original fauna. Extinction has important implications for interpreting their has also shaped the distribution of Antillean land historical biogeography.

RESUMEN Un nuevo roedor capromydo isolobodontido estas islas. Es razonable, asimismo, inferir que los (Zazamys veronicae, nuevo genero y especie) y un grupos de mamiferos terrestres estaban amplia- nuevo primate platyrrhinido (genero y especie A, mente distribuidos en las tierras antillanas, y el sin denominar) se describen a partir del material descubrimiento de un "fantasma" biogeografico recobrado en Domo de Zaza, una importante lo- (Zazamys) en el Mioceno Inferior de Cuba, sus- calidad fosilifera del Mioceno Inferior tardio si- tenta esta afirmacion. tuada en la region sur-central de Cuba, cercana a Las partes emergidas de las actuales Antillas la ciudad de Sancti Spiritus. Tambien se describe Mayores formaban entonces un conjunto mas un femur que se cree representa un pequeiio pe- compacto hasta el comienzo del Neogeno. Ellas rezoso megalonychido (genero y especie B, sin de- formaban parte de un elemento topografico po- nominar), de un sitio paleontologico bien datado sitivo muy duradero en la cuenca del Caribe, el de edad Oligoceno Inferior, localizado cerca de la cual aqui se denomina GAARlandia. Durante el ciudad de Yauco, en el suroeste de Puerto Rico. Eoceno Superior-Oligoceno, este elemento incluia Estos descubrimientos, unido a otros recien rea- a Cuba, La Espaniola septentrional, Puerto Rico, lizados en las Antillas Mayores, permiten estable- Islas Virgenes, Cresta de Aves y las Antillas Me- cer que estas islas ya soportaban una variada fauna nores Venezolanas. A partir del Oligoceno tardio, de mamiferos terrestres a partir del Oligocene. Las comenzo la fracturacion y dispersion de GAAR- evidencias geologicas son consistentes con la pro- landia hasta formar sus componentes insulares ac- posicion de que los iniciadores de todos, o prac- tuales, un proceso que forz6o la vicariancia de la ticamente de todos los clados fundadores de la fauna original. Las extinciones que han ocurrido fauna de mamiferos terrestres antillanos, se em- a traves del Cenozoico tardio, tambien es probable plazaron en estas tierras cerca del limite Eoceno- que hayan contribuido a conformar la distribucion Oligoceno. Bajo los terminos de este modelo, es actual de los mamiferos terrestres Antillanos, y de esperar que ningun mamifero terrestre mas estos tiene importantes implicaciones para la in- antiguo que Eoceno Superior sea encontrado en terpretacion de su biogeografia historica.

INTRODUCTION How and when the ancestors of the land Rico. We also extend and modify our model vertebrate fauna ofthe Greater Antilles man- of the land-mammal colonization of the aged to reach those islands are topics per- Greater Antilles (MacPhee and Iturralde-Vi- ennially debated in biogeography (for recent nent, 1994). In the text, the age of stage and examples, see Hedges et al. [1992, 1994] and epoch boundaries cited in millions of years Page and Lydeard [1994]). Although agree- (Ma) follows the framework ofHarland et al. ment concerning the method(s) of coloniza- (1990). tion is evidently not in sight, there is a grow- ing fossil tally (table 1) which provides un- deniable evidence that founder clades were ABBREVIATIONS already well represented in the islands by the mid-Cenozoic. As noted elsewhere (MacPhee AMNH American Museum of Natural His- and Wyss, 1990), fossils are empirical doc- tory uments with temporal as well as phylogenetic AMNH M American Museum of Natural His- are im- tory, Department of Mammalogy implications, and as such critically AMNH VP American Museum of Natural His- portant to the reconstruction ofbiogeograph- tory, Department of Vertebrate Pa- ical history. leontology In this report we announce the discovery e estimate (in measurement, due to of the second and third land mammals of breakage) Tertiary age from Cuba, and bring attention Fm formation to the first such discovery made in Puerto JD Juana Diaz Fm (sensu Frost et al., 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 3

1983), not including "upper carbon- grateful to Prof. Edward Robinson (Depart- ate member" ment of Geology, University of the West In- Ma millions of years (ago) dies) and Lic. Gena Fernandez (CIDP-Cu- MNHNH Museo Nacional de Historia Natural pet, Cuba) for identifying foraminifers in our (La Habana) Yauco samples (see text). MNHNH P Museo Nacional de Historia Natural (La Habana), paleontological collections of Departamento de Colec- EARLY MIOCENE RODENT AND ciones PRIMATE FROM CUBA MNHNH V Museo Nacional de Historia Natural referable to sub- (La Habana), vertebrate collections A new capromyid rodent, of Departamento de Colecciones family Isolobodontinae, is named here on the rev. reversed (as applied to figures, to fa- basis of dental material from the important cilitate comparison) Early Miocene locality of Domo de Zaza in UC "upper carbonate member," in older the province ofSancti Spiritus. A platyrrhine definitions of Juana Diaz Fm monkey is also represented in the Zaza material by a single, well-preserved astragalus. ACKNOWLEDGMENTS Although the astragalus appears to represent Fieldwork at Domo de Zaza was conducted a species new to science, it makes a poor under the auspices ofagreements between the holotype. Accordingly, until specimens ca- Museo Nacional de Historia Natural, the pable ofproviding a secure diagnosis are dis- former Academia de Ciencias de Cuba (pres- covered, we shall simply refer to the new Zaza ent Ministerio de Ciencias, Tecnologia y Me- anthropoid as "unnamed genus and species dio Ambiente) and the American Museum of A (primate)." Locality information is kept Natural History, and was financed by grants purposely brief, since this site and its context from the National Science Foundation (NSF is fully described in another paper (MacPhee BSR 900002) and the AMNH. Fieldwork at et al., in prep.). Yauco was conducted with the assistance of the Departamento de Geologia, Universidad RODENTIA de Puerto Rico (Recinto de Mayagiiez) and CAPROMYIDAE: ISOLOBODONTINAE Victor Gonzalez ofSan Juan, and was funded by a grant from the Caribbean Scientific Ini- Zazamys veronicae, new genus and species tiative ofthe RARE Center for Tropical Con- HOLOTYPE: Left mandibular ?M3 servation. For assistance at Domo de 7aza, (MNHNH P 3072; fig. IA, F-G), found in we extend special thanks to Stephen Diaz lag deposit near south end of Domo de Zaza Franco, Osvaldo Jimenez Vasquez, Luis section by 0. Jimenez in February 1994. Olmo, Reynaldo Rojas Consuegra, and Pavel ETYMoLoGY: Zaza (locality name) + ,uva Valdes Ruiz. In connection with the Puerto (mouse). Species name for Veronica Mahan- Rican expedition, we thank Profs. James ger MacPhee, in gratitude for her unflagging Joyce and Hernan Santos (Universidad de help and sharp editing pencil. Puerto Rico, Mayagiiez) for information con- TYPE LocALiTY AND AGE: Domo de Zaza, cerning the Yauco site, and Veronica a locality in late Early Miocene Lagunitas Fm MacPhee and Clare Flemming for assistance of south-central Cuba. For detailed locality in the field. We also thank Veronica and Clare information and age assessment, see Mac- for assistance in editing the manuscript ver- Phee and Iturralde-Vinent (1994) and sion of this paper. RDEM took the photo- MacPhee et al. (in prep.). graphs used in figures 9 and 11; MIV drew REFERRED MATERIAL: Right mandibular figures 8 and 10; all other illustrations are the Ml or M2 (MNHNH P 3058, fig. 1B, rev.) work of our talented artists Lorraine Meeker and left mandibular Ml or M2 (MNHNH P (figs. 1, 2, 3 and 5) and Clare Flemming (figs. 3071, fig. 1 C), found in lag by (respectively) 4, 6, and 7). Dr. Hans Schellekens (UPR Ma- P. Valdes in November 1993 and R. MacPhee yagiiez) and Dr. Jennifer White (Simpson in February 1994. College) read and commented upon an earlier DiAGNosIs: The type and only species of draft ofthis paper. Finally, we are extremely Zazamys is Z. veronicae-, genus and species 4 AMERICAN MUSEUM NOVITATES NO. 3141

TABLE 1 Reported Tertiary Freshwater and Terrestrial Vertebrates of the Greater Antillesa Island Age/Formation/Locality Reference Cubab Imagocnus zazae (Megalonychidae, E. Miocene; Lagunitas Fm; Domo de MacPhee & Iturralde-Vinent, Xenarthra) Zaza, S. Spinitus 1994 Zazamys veronicae (Capromyidae, E. Miocene; Lagunitas Fm; Domo de This paper Rodentia) Zaza, S. Spiritus Unnamed genus and species A E. Miocene; Lagunitas Fm; Domo de This paper (Platyrrhini, Primates) Zaza, S. Spiritus Unidentified bird E. Miocene; Lagunitas Fm; Domo de MacPhee et al., in prep. Zaza, S. Spiritus Unidentified pelomedusid turtle E. Miocene; Lagunitas Fm; Domo de MacPhee et al., in prep. Zaza, S. Spiritus Hispaniolac.de Cichlasoma woodringi (Cichlidae, ?M. Miocene; Las Cahobas Fm; Mire- Cockerell, 1924 Perciformes) balais, Haiti Anolis dominicanus (Iguanidae, ?Miocene; amber deposits; Dominican Rieppel, 1980 Sauria) Republic Sphaerodactylus sp. (Gekkonidae, ?Oligo-Miocene; amber deposits; Do- Kluge, 1995 Sauria) minican Republic Eleutherodactylus sp. (Leptodactyli- ?Late Eocene; amber deposits; Domin- Poinar & Cannatella, 1987 dae, Anura) ican Republic cf. Nesoctites (Picidae, Aves) ?U. Eocene-E. Miocene; amber depos- Laybourne et al., 1994 its; Dominican Republic Unidentified mammal (?rodent) ?Miocene; amber deposits; Dominican Poinar, 1988 Republic Puerto Rico Boid (Ophidia) E. Miocene; Cibao Fm; Aguadilla MacPhee & Wyss, 1990 ?Iguanid (Sauria) E. Miocene; Cibao Fm; Aguadilla MacPhee & Wyss, 1990 Unnamed genus and species B E. Oligocene; Juana Diaz Fm; Yauco This paper (Tardigrada) Unidentified pelomedusid turtle E. Miocene; Cibao Fm; Aguadilla MacPhee & Wyss, 1990 a This table updates table 2 of MacPhee and Iturralde-Vinent (1994), which should be consulted for additional remarks. b Intact carapaces and plastrons of pelomedusid turtles have been recovered in situ from lagoonal deposits in Lagunitas Fm. This strongly implies that they were freshwater-adapted (cf. MacPhee and Iturralde-Vinent, 1994, table 2); the habitat of Puerto Rico pelomedusids is still ambiguous, but we include them in this paper. c See Grimaldi (in press) for discussion of the general problem of dating some ambers; listed ages follow authors' reports, with question marks expressing level of doubt. d Cichlasoma woodringi may not be distinguishable from extant species (M. L. Smith, personal commun.). e "Sphaerodactylus" dommeli (B6hme, 1984) may be an anole rather than a sphaerodactyl according to Kluge (1995, citing D. Frost, personal commun.).

diagnoses are therefore the same. Among 1989b). Whether this multi-island range is known Antillean taxa, Zazamys veronicae original or the result ofhuman introductions most closely resembles the Quaternary An- has long been debated (cf. Allen, 1942; Mor- tillean genus Isolobodon, representatives of gan and Woods, 1986), but for the purposes which are known from Hispaniola, Puerto of this paper we accept the occurrence of Is- Rico, Mona Island, and several ofthe Virgin olobodon on both Hispaniola and Puerto Rico Islands, but not from Cuba (Anthony, 1926; as natural. Also, in agreement with Woods Reynolds et al., 1953; Varona, 1974; Woods, (1989b) we accept Isolobodon portoricensis 1995 MACPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 5 and L montanus as good species, but follow complete jaws would be needed to confirm Reynolds et al. (1953) in considering L levir these points. Cementum forms noticeable as synonymous with I. portoricensis. ridges only within reentrant folds (fig. 1 F, F'), Occlusal pattern, root shape, and arc of but in MNHNH P 3072 there are barely dis- curvature of the teeth in the Zazamys hy- cernible, quasiparallel lineaments detectable podigm indicate that all are lower molars. on external surfaces (fig. lF, arrows). As far Lower molars of Zazamys have two lingual as may be determined without a histological reentrant folds (mesoflexid and metaflexid) examination, these lineaments are not sepa- and one buccal fold (hypoflexid), as is gen- rately composed of cementum but seem to erally characteristic of capromyids (Woods, lie within the enamel. It is unclear whether 1989b). Zazamys differs from all capro- this condition in Zazamys can be validly re- myines and hexolobodontines sensu Woods garded as antecedent to the highly pro- (1989b, 1992) in that the thick continuous nounced version of cementum ridging seen shell ofcementum encircling the molariforms in Isolobodon (fig. 1E). In capromyines, ce- of taxa in these two subfamilies is absent. mentum within reentrant folds has a much Zazamys resembles the plagiodontines Pla- less organized, frothy texture. giodontia and Rhizoplagiodontia in that ce- Because cementum layers on capromyid mentum is limited to reentrant folds, but dif- teeth are applied to smooth enamel surfaces, fers in that the folds themselves are not they can and do scale off. However, small obliquely oriented (see below). Zazamys re- "islands" of cementum usually remain, ex- sembles the isolobodontines Isolobodon por- cept in specimens that have been severely toricensis and I. montanus in the patterning abraded. The teeth from Zaza are not signif- of occlusal surfaces. Howevei, there are dif- icantly abraded and we therefore conclude ferences which indicate that separation at the that complete sleeves ofcementum were nev- generic level is warranted. The lower molars er present. of Zazamys differ from those of L portori- PHYLOGENETIC RELATIONSHIPS: Isolobo- censis in the following respects: (1) Reenter- dontines and plagiodontines are closely re- ing metaflexid and hypoflexid meet at the lated and are often included in the same sub- center of the occlusal surface, enclosing family (e.g., by Varona, 1974). Woods (1989a) roughly equal areas of dentine between their defined 14 character state differences between arms; in Isolobodon, metaflexid-hypoflexid Isolobodon and Plagiodontia, but ofthese only juncture is lingually displaced, so that me- one (ridged cementum on external surfaces taflexid arms are considerably shorter than oftooth) can be evaluated on the basis ofthe hypoflexid arms and area of dentine enclo- existing Zazamys hypodigm. Within Ca- sure is considerably smaller. (2) Ridged ce- viomorpha, loss or modification of external mentum is essentially restricted to reentrant cementum on molars is not distributed in a folds and is barely discernible on tooth col- way that is consistent with the likeliest phy- umns (see fig. 1); in Isolobodon, ridged ce- logenetic reconstructions of the group; mul- mentum is easily distinguishable outside of tiple convergences across the suborder have reentrant folds. (3) Reentrants are much more therefore occurred. Ridging, however, ap- compressed (also more compressed than in pears to be a unique synapomorphy of iso- L montanus, which is intermediate in this lobodontine capromyids, occurring (as far as regard). is now known) in no other caviomorph clade. DEsCRnPrION: Because these teeth are iso- Possession of a version of this character de- lated finds, their orientation in thejaw during cisively places Zazamys next to known iso- life is not certain. However, judging from the lobodontines, a fact which can be suitably location of intraproximal wear facets and underlined by including the Zaza rodent in other features, they were probably seated in subfamily Isolobodontinae. such a way that the reentrants were more or It is widely agreed that capromyids and less normal to the axis of the jaw. echimyids are closely related, although there The teeth are judged to be hypselodont (ev- is rather less agreement concerning within- ergrowing, with open roots) and probably group relationships and the closeness oftheir hypsodont, although ontogenetic data and affiliation with other caviomorphs (cf. Corbet 6. AMERICAN MUSEUM NOVITATES NO. 3141

Fig. 1. Holotype ofZazamys veronicae (MNHNH P 3072), a left mandibular M3, is seen in occlusal (A), lingual (F), lingual detail (F'), and buccal (G) views, from Early Miocene locality of Domo de Zaza in south-central Cuba. Referred specimens seen in occlusal view are (B), a right (rev.) lower Ml or M2 (MNHNH P 3058); and (C), a left lower Ml or M2 (MNHNH P 3071). Compare to left mandibular dentition ofIsolobodon portoricensis (AMNH M 38409), in occlusal (D) and buccal (E) aspects (anterior is to the left). Note, in F, presence of periodic banding on enamel (arrows), mimicking lines of ridged cementum in Isolobodon; note also, in F' (cast), the organization of interflexid cementum. All scales are 5 mm. 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 7 and Hill, 1980; Reig, 1981; Patterson and distribution in extant platyrrhines as pre- Wood, 1982; Woods, 1989a, 1989b). Woods sented by Meldrum (1990), being exceeded (1989b) has recently considered these issues in this regard only by Ateles, Alouatta, and in some detail, and has concluded that cap- Lagothrix (Brachyteles was not included). romyids are derived from echimyids (or a Comparable dimensions of the astragalus common ancestor of echimyids, myocasto- from Cueva Alta are slightly larger. The head rids, and octodontids). Since capromyids have ofthe Zaza specimen is slightly damaged me- not been identified paleontologically or neon- dially, but the navicular facet was clearly tologically in surrounding mainlands, this ar- ovoid in shape with its long axis (as measured gument carries the implication that capro- by HDW) somewhat obliquely set. The head myids must have evolved wholly within the is missing from the Cueva Alta specimen. Greater Antilles. If Capromyidae as usually The necks of both specimens are compara- defined is a holophyletic group restricted to tively short but robust. The angle described the Greater Antilles and some nearby islands, by the neck on astragalar body is approxi- then based on our reconstruction ofAntillean mately 300 in each specimen (Meldrum's faunal history this clade could not have orig- method). The anterior calcaneal facet of the inated earlier than about 42 Ma- i.e., the be- Zaza anthropoid is extensive proximally, ginning of the Late Eocene, the earliest time showing no tendency to taper in the direction at which it would have been feasible for land ofthe deep astragalar sinus. Distally, it meets mammals to take up permanent residency the navicular facet along a raised margin; a (see Discussion). nonarticular strip between facets is not present. The anterior calcaneal facet on the astragalus assigned to P. varonai is damaged, PRIMATES but there is no doubt that the facet was just ANTHROPOIDEA: PLATYRRHINI as extensive proximally. The posterior calcaneal facet is very large in both specimens. Unnamed Genus and Species A (Primate) "Waisting" in the medial third ofthe facet is MATERIAL: Right astragalus (MNHNH P slight, especially in the Paralouatta astraga- 3059; table 2, fig. 2), found in lag deposit near lus. Indeed, the ratios offacet length to width south end of Domo de Zaza section by 0. (PCL/PCW) in these specimens are consid- Jimenez in November 1993. erably higher than in any ofthe taxa or fossils PRELIMINARY EVALUATION: At the time of sampled by Meldrum (1990). The astragalar its discovery in 1993, MNHNH P 3059 was sinus is much deeper in the Zaza specimen thought to be morphologically distinct from than in the one from Cueva Alta. By contrast, astragali of all known mainland New World trochlear height (TH) is somewhat higher in primates, including near-coevals from Pata- P. varonai. gonia like Carlocebus and Soriacebus In both specimens, the trochlear articular (MacPhee and Iturralde-Vinent, 1995). More surface is high, parallel-sided (not "wedged"), recently, a partial astragalus (MNHNH V 205) and extensive from front to back. The troch- attributable to the Cuban monkey Paralouat- lear spool is moderately deep, and in proxi- ta varonai (fig. 3) has turned up in a faunal mal aspect the medial rim is slightly higher collection from Cueva Alta (for site descrip- than the lateral. The tibial articular surface tion, see Jaimez Salgado et al., 1992). The terminates proximally in a long, fluted shelf two elements are extraordinarily similar in that would have guided the tibial and fibular size and details ofshape, which suggests (un- flexors. In P. varonai this area is broken and surprisingly) that they represent the same abraded. Distally, the tibial articular surface narrow monophyletic group. ends more or less abruptly in both specimens. The following morphological description Medially, a shallow stop for the tibial medial uses the terminology and measurement malleolus encroaches on the neck. The distal methods of Meldrum (1990); table 2 should margin of the lateral trochlear rim projects be consulted for measurement values and ac- beyond the level of the astragalar body in ronyms. In astragalar length, the Zaza spec- both specimens, but especially so in the Cueva imen places near the upper end of the size Alta astragalus. The articular surface for the 8 AMERICAN MUSEUM NOVITATES NO. 3141

Fig. 2. Astragalus (MNHNH P 3059) of an anthropoid (unnamed genus and species A), from Early Miocene locality of Domo de Zaza, south-central Cuba. Views are A, dorsal; B, ventral; C, proximal; D, distal; E, medial; and F, lateral. fibula is flat, but ventrad it spreads out lat- ly higher trochlea, wider posterior calcaneal erally to a marked degree. The posterolateral facet, shallower astragalar sinus in P. varonai) process of Meldrum (1990) seems to have underlines the high probability that the spe- been absent in both specimens, unless the cies which they represent are closely related. tiny eminence in the Cueva Alta specimen In fact, this degree of similarity begs a ques- can be so regarded. tion: if Zaza is confirmably late Early Mio- The fact that there are only minor differ- cene in age (MacPhee et al., in prep.), are the ences between these two astragali (e.g., slight- fossils from Cueva Alta truly as young as 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS99

Fig. 3. Astraglus (MNHNH V 205) attributed to Paralouatta varonai, from Cueva Alta, Sierra de los Organos, western Cuba. Views are A, dorsal; B, ventral; C, proximal; and D, lateral. Shading denotes abraded areas.

Quatemary? Sloth bones from Cueva Alta found in these caves to make U/Th dating (and Cueva del Mono F6sil, to which it is worthwhile. Accordingly, the Quaternary age connected by a fossiliferous fissure) were assessment for the monkey caves has to be found to be beyond the effective range of 14C based on the character of the recovered fau- dating (M. Tamers, personal commun.), and nule, which largely consists of taxa known no usable calcite-bone associations have been from localities of proven or suspected Qua-

TABLE 2 Measurements (in mm) of Zaza and Cueva Alta Astragalia Specimen L HDW HDH ACF PCL PCW TW TH ATW PTW W MNHNH P 3059 20.0 9.7e 7.3e 12.7 9.5 5.5 11.7 10.8 9.2 9.1 16.5 MNHNH V 205 - - - - 10.Oe 6.9 12.0e 11.9 10.5 10.5e 16.9e a For landmarks used in measurement, see Meldrum (1990). Key: ACF anterior calcaneal facet; ATW anterior trochlear width; HDH head height; HDW head width; L length; PCL posterior calcaneal facet length; PCW posterior calcaneal facet width; PTW posterior trochlear width; TH trochlear height; TW trochlear width; W width. The letter e indicates that the measurement is estimated (due to breakage)., 10 AMERICAN MUSEUM NOVITATES NO. 3141

Callicebus Cebus Ateles Alouatta

'N -1I

tA-CH~~ .Th Q' Fig. 4. Astragali of representative extant platyrrhines. Sequence of views (top to bottom of page) is dorsal, ventral, proximal, distal, lateral, and medial. Note individual scales. ternary age elsewhere in Cuba (see Jaimez within these caves (Oscar Arredondo, per- Salgado et al., 1992). However, there are some sonal commun.). Since the fissures from which unique occurrences: in addition to Par- the faunule was collected are not connected alouatta, a previously unknown owl and sloth with the present sedimentary floors of the have also been recovered, from localities caves, and since the cave entrances are sit- 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS

Saimiri Aotus Callithrix Dolichocebus

aXccs

.001)

......

QL_XAX (X ,

Fig. 4. Continued.

uated high above the present valley floor, it (cf. Meldrum, 1990). Several points emerge is distinctly possible that the fissure fillings from a comparison ofthese astragali with the sample a much earlier fauna. Zaza and Cueva Alta specimens. First, the Figure 4 provides schematic views of the Cuban monkeys are not especially similar to astragali of some extant platyrrhines and a any ofthe species illustrated (but see below). specimen usually referred to Dolichocebus This applies in particular to Alouatta, the ge- gaimanensis, an extinct Patagonian species nus originally thought to be the closest living 12 AMERICAN MUSEUM NOVITATES NO. 3141 relative ofParalouatta (Rivero and Arredon- AN EARLY OLIGOCENE SLOTH do, 1991). Among other distinctive features, FROM PUERTO RICO the astragalus ofAlouatta exhibits a "wedged" While engaged in a preliminary survey in trochlea, low trochlear relief, indistinct mar- southern Puerto Rico in March 1994, our gins, vertical medial articular surface for the field party discovered a fossiliferous site near tibia, and a large posteromedial process. The the town of Yauco, in the southwestern part Cueva Alta specimen lacks these features, as of the island. Among the vertebrate fossils does the better-preserved astragalus from we recovered was the proximal part of a Zaza. What, then, is the proper sister group mammalian femur, here referred to the xe- of the Cuban monkeys? narthran clade Tardigrada. This discovery of- In a cladistic analysis of a selection of cra- fers the hope that Puerto Rico can be tied niodental characters, MacPhee et al. (1995) into the paleobiogeographical picture being found that, among sampled Quaternary plat- developed for the other Greater Antilles yrrhines, the extinct Hispaniolan anthropoid (MacPhee and Wyss, 1990; MacPhee and Antillothrix (="Saimiri") bernensis grouped Iturralde-Vinent, 1994). Of equal interest is most closely with Paralouatta varonai. These the apparent age of the Yauco sediments. As Antillean taxa joined Callicebus at the next we note in greater detail below, fossil evi- node. By contrast, no close relationship be- dence indicates that these sediments are Early tween the Antillean monkeys and atelines Oligocene (Globigerina ampliapertura zone, (including Alouatta) was indicted. Consistent or about 33-34 Ma according to the time with the trend of the craniodental evidence, zonation utilized by Berggren et al. [1992]). Callicebus shows more astragalar resem- Yauco may therefore be older than the Zaza blances to the Cuban fossils than atelines do. locality by 15 Ma or more, making it by far In extant titi monkeys, the trochlea is high the oldest land vertebrate site with a well- and only slightly "wedged," trochlear walls constrained age in the West Indies. Oddly are subvertical, and calcaneal facet shape is enough, if the femur is that of a megalony- substantially similar to the Zaza and Cueva chid, as we think is highly probable, it be- Alta specimens. However, correspondences comes the oldest specimen so far referred to to Dolichocebus and other extant platyrrhines that family (cf. Pascual et al., 1990). This such as Saimiri (fig. 4) also exist, raising the means very little by itself, other than to un- distinct possibility that much of this mor- derline the fact that tardigrades had to have phology is simply primitive (see also MacPhee been well diversified and widely distributed and Iturralde-Vinent, 1995). in the Neotropics before the We find that the Zaza and Cueva Alta as- Oligocene. tragali fail to support an affiliation ofthe Cu- ban monkeys with either Alouatta or atelines XENARTHRA in general (cf. Ateles, fig. 4).3 The hypothesis TARDIGRADA: ?MEGALONYCHIDAE that the closest living relative of the Cuban Unnamed Genus and B and Hispaniolan monkeys is Callicebus, as Species (Tardigrade) suggested by craniodental characters MATERIAL: Left proximal femur (AMNH (MacPhee et al., 1995), is plausible but re- VP 129883, fig. 5), preserving head, neck, quires additional evidence and testing. and trochanters. Found by C. Flemming in March 1994. PRELIMINARY EVALUATION: Our criteria for assigning this specimen to Mammalia are as follows: (1) greater and lesser trochanters 3The only biogeographical reason for considering ate- large, projecting, and situated on a marked lines in the context ofplatyrrhine evolution in Cuba has oblique; (2) medullary cavity not densely filled disappeared. The holotype and only known specimen of Ateles (=Montaneia) anthropomorphus-the supposedly with cancellous bone; (3) proximal articular endemic "Cuban spider monkey"-is less than three end ovoid, with well-defined border marking centuries old. It probably represents an imported A. fus- attachment ofjoint capsule; and (4) large fo- ciceps, as has long been suspected on morphological vea in the center of the femoral head. Cor- grounds (see MacPhee and Rivero de la Calle, 1995). porately, these features indicate that AMNH 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 13

VP 129883 represents a mammal, and dis- there is some support for brigading Puerto tinguish it from conditions encountered in Rican Acratocnus, Hispaniolan Synocnus, and chelonians, crocodylians, and squamatans. Cuban Neocnus and Miocnus in one unit Referral to Tardigrada (i.e., the monophy- (equivalent to tribe Acratocnini, Varona letic taxon containing all sloths, living and [1974]). If this grouping is monophyletic, extinct) is considered reasonable because of acratocnins, like some other assemblages of strong resemblances to Quaternary and Re- closely related Antillean taxa, had a multi- cent sloths and because there is no other plau- island distribution; this regularity requires sible candidate among other land mammal explanation (see Discussion). taxa having Antillean representation (i.e., Li- LocALITY AND AGE: The sloth femur was potyphla, Primates, Rodentia). On biogeo- recovered at Locality AMNH PR 1994/1, sit- graphical grounds it is reasonable to believe uated just south of Yauco on Highway 3133 that the Yauco sloth will prove to represent (figs. 7, 8). Sediments in this region are ap- a member ofMegalonychidae, the only sloth portioned to Early Oligocene Juana Diaz Fm family that provably managed to colonize and an overlying but unnamed carbonate unit, West Indian islands (for recent treatments of as described in more detail below. Antillean and mainland members of this The Yauco locality is situated within the family, see White [1993]). Consistent with former "lower clastic member" of the Juana this is the fact that the specimen bears a par- Diaz Fm (Monroe, 1980). The Juana Diaz ticularly close resemblance to the proximal Fm has long been informally divided into two femora of certain living and extinct mega- members, an upper carbonate (UC) unit and lonychids (fig. 6). These resemblances in- a lower clastic unit (Moussa and Seiglie, 1970, clude (1) markedly open angle formed by 1975; Monroe, 1980). Frost et al. (1983) have femoral neck and diaphysis, (2) position and proposed that the Juana Diaz Fm (hereafter orientation of the trochanters, (3) cross-sec- abbreviated as JD) be restricted to the lower tional shape of the shaft, and (4) spherical clastic member, a proposal that we follow in proximal articular surface. A minor differ- this paper. The UC, allocated to a new for- ence from extant Choloepus (and Bradypus) mation as yet unnamed, correlates with the is the presence of a fovea for ligamentum Globorotalia opima opima-Globigerina an- teres; Quaternary Antillean megalonychids gulisuturalis zone and is presumed to be of are usually foveate (Fischer, 1971). The chief Late Oligocene age. The JD correlates with difference is size: the Yauco femur represents the Globigerina ampliapertura zone and is an animal as small as or smaller than living therefore older (Pessagno, 1963; Seiglie and Bradypus, while Neocnus, Acratocnus, and Berm(udez, 1969; Moussa and Seiglie, 1970, Choloepus are considerably larger. Indeed, the 1975; van den Bold, 1971; Todd and Low, Yauco sloth would appear to be one of the 1979; Frost et al., 1983). Abundant plank- smallest fossil tardigradans on record (see tonic forams reported from the JD include White [1993] for evidence ofvery small sloths Globigerina ampliapertura, G. haoi, G. an- in the Santacrucian ofArgentina). In this re- gustiumbilicata, G. ciperoensis, G. rohri, G. gard it stands in marked contrast to Early venezuelana, and Cassigerinella chipolensis Miocene Imagocnus, which appears to have (Seiglie and Berm(udez, 1969). Facies sam- been nearly the size ofMegalocnus (- 200 kg pling very shallow water and terrestrial en- according to Paula Couto [1979]). vironments have no planktonic forams, but The Yauco femur is morphologically closer do yield Amphisteginafloridana, Angulogeri- to Neocnus and Choloepus than it is to Acra- na vicksburgensis, Baggina cojimarensis, tocnus, the only Quaternary sloth known from Pararotalia mexicana and Lepidocyclina un- Puerto Rico. The phylogenetic significance of dosa, all ofwhich are typical Oligocene mark- this (if any) is hard to assess in the absence ers (Todd and Low, 1979). of adequate material. There are several dif- The Yauco section exposes both JD and ferent and mutually exclusive views concern- UC (figs. 8,9). The JD belongs to the "inshore ing the interrelationships of the Antillean shelf sandstones and conglomerates" facies sloths (e.g., Varona, 1974; Paula Couto, 1979; of Frost et al. (1983, their fig. 2). This latter Arredondo, 1988), although we believe that unit, 60-70 m thick, is composed of clays, 14 AMERICAN MUSEUM NOVITATES NO. 3141

Fig. 5. Proximal part of a left femur (AMNH VP 129883) of a megalonychid sloth (unnamed genus and species B), from Early Oligocene Locality AMNH 1994/1, near Yauco, SW Puerto Rico. Views are A, anterior; B, posterior; C, lateral; and D, medial. Scale is 1 cm. sands and gravels with some marly coquina small, laterally overlapping river channels of intercalations. The top of the unit is an ero- a few meters thickness, and are more frequent sional horizon where conglomeratic lime- in the lower half of the exposed section. stones of a few centimeters thickness merge Near the base ofthe outcrop, poorly sorted into coralgal reef limestones of the UC. The coarse-grained conglomerates occur as inter- base of the section was not observed. calations, two to three meters in thickness, Fine-grain sands and sandy clays, with with clasts up to 15 cm in diameter. These abundant oxidized plant remains, comprise conglomerates were probably deposited dur- the most common lithology. These beds form ing flood events. In localities near Yauco, and units tens ofcentimeters thick, distinctly fin- also to the east ofthe locality, conglomerates ing upward and incorporating caliche and at the base ofthe JD reach thicknesses greater other indications of soil-forming processes. than 50 m and are intercalated with red pa- Small concentrations of gypsum, apparently leosols. It is possible that the basal parts of epidiagenetic, occur as fracture fillings in some these conglomerates are as old as latest Eo- of the clays and sandy clays. Such features cene (cf. figs. 9, 10). are consistent with terrestrial conditions and Three marly coquinite horizons, two to an oscillating freshwater table, as might occur three meters thick and located near the base, in relation to swamps and lagoons close to a middle, and top of the section, are formed shoreline. Alluvial beds (figs. 8, 9) occur as mainly by fragmentary or complete tests of 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 15

SI-111a 20

N - Y JD 67° 66° w a b~~~~~~~~~~~~~ Fig. 7. Location ofreferenced localities in Cuba (top) and Puerto Rico (bottom). Key: CA, Cueva Alta; DZ, Domo de Zaza; Y, Yauco; and JD, Juana Diaz.

water vertebrates (sirenians, crocodiles, and turtles) in the same sediments. The coquinite beds probably represent short-lived, shallow- water marine invasions of the coastal plain. Further east, similar beds intercalated with sandy lenses outcrop along the Rio Jacaguas, from which sirenian fossils have been recov- 1SS @( ered (Matthew, 1916). El It is not yet possible to determine whether some horizons at Yauco are more productive of vertebrate remains than others. The sloth Fig. 6. Left femora of (a) Neocnus sp. (Arre- femur was discovered in the upper halfofthe dondo coil.), (b) Acratocnus ondontrigonus (AMNH VP 17716.-i), and (c) Choloepus hoffmanni section, within a lag developed from a clay/ (AMNH M 139772), in anterior (top) and poste- sandy clay/caliche sequence. A number ofar- rior (bottom) views. ticulated sirenian bones, found in the middle ofthe measured section, were associated with sandy clays evidently deposited in brackish Lepidocyclina undosa. Occurring in smaller conditions. A few well-preserved crocodile numbers are shells ofgastropods (turritelids), skull elements and fragments of turtle shell bivalves (pectinids, ostreids, etc.), echino- were found in alluvial channel deposits near derms, and crustaceans, with occasional shark the base of the section. teeth and isolated coral fragments. These In order to test age determinations of the Lepidocyclina coquinites have been inter- JD in the Yauco section, four samples were preted as forereef deep-shelf facies by Frost collected and analyzed for microfossil con- et al. (1983), but at Yauco the coquinite beds tent. Sampling horizons are indicated in fig- are embedded in sediments deposited in a ure 8 as 94/1 6D (coquinite near base of JD), near-shore environment, with an inverte- 94/16C (clays 10 m above), 94/16B (middle brate faunal assemblage characteristic ofvery portion ofJD), and 94/1 6A (base ofUC). The shallow water. Consistent with this interpre- samples yield benthic forams, rare planktonic tation is the discovery offreshwater/brackish forams, fish teeth, echinoderm spines, and 16 AMERICAN MUSEUM NOVITATES NO. 3141

meters arated by an erosional surface associated with '(0) 0- .0 a thin conglomeratic bed (figs. 8, 9). Z

Co DISCUSSION

0 GEOLoGICAL CONSTRAINTS AND LAND NEZoCa Z o MAMMAL BIOGEOGRAPHY OF THE 94/16A 20 - GREATER ANTILLES In an earlier paper (MacPhee and Itur- EN ralde-Vinent 1994), we proposed that both loth femur dispersal and vicariance have occurred in the

c,irenian skeleton phase occurred first and involved the move- ILQ ment ofspecies en bloc out ofnorthern South N E America (whence all known major taxa ex- - co D .C cept Antillean lipotyphlans derive) and onto <: 94/16B subaerial landmasses in the central Carib- z 0 < (D 60 bean. How this was accomplished in detail is ofcourse uncertain. However, unlike most 94/16C other proponents ofdispersal, we assume that some critical event occurred in the early Ce- nozoic which briefly altered the configuration CONGLOMERATES of lands and the amount of open water that E2 SANDSTONES 5 94/16D had to be crossed. That is, dispersals were SANDY CLAYS 80 - not spread out over a long time span, but LIMESTONES crocodilian jaw instead occurred within a very narrow inter- COQUINITES val. The later, vicariant phase involved subdivision offounder clades4 by the erection of Fig. 8. Columnar section of Loccality AMNH water barriers produced by tectonic dismem- 1994/1, near Yauco, SW Puerto RLico. Lenslike berment and reconfiguration of the existing features are alluvial channels (see alsso fig. 9). Also chain of islands. This phase is elsewhere de- indicated are approximate positic of yielding vertebrate fossils (O) and ins scribed as "island-island" vicariance factes (MacPhee and it differs from sapling horizons (0). Tardigradan fA,111minrL& wtYV"reb-1%,- Wyss, 1990); covered from float. 4 Optimally, the number of "founder clades" recognized for Antillean land mammals ought to be equal to ostracods (this last very abundant in sample the number of initiators (propagules) that were phylo- 94/16C). genetically separate from one another (i.e., separate spe- In each horizon the assemblage of forams cies) at the time of original colonization. Our present (table 3) is typical Lower Oligocene, because view is that a minimum of eight initiators has to be the Upper Oligocene markers Miogypsina and inferred to explain the existing land mammal evidence Miogypsinoides are absent. The absence of for the Greater Antilles, distributed as follows: Primates these marker taxa is consistent with the in- (two initiators), Rodentia (three), Insectivora (two), Xe- terpretation that these beds were deposited narthra (one). There could, of course, have been more than one initiator per lower-level monophyletic group, in the Early Oligocene, as indicated by the as some authors have inferred (e.g., Morgan and Woods, planktonic foram data referenced above. No 1986). For the Greater Antilles as geographically defined, large depositional hiatus is evident between accounting for late-evolving oryzomyins (Rodentia) re- the JD and the UC units, because foramin- quires a ninth initiator-but only for Jamaica, which is iferal assemblages are similar (table 3, sam- not considered part of GAARlandia as defined in this ples 94/16A and B), although they are sep- paper. 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 17.

Fig. 9. Top, View of Locality AMNH PR 94/1, near Yauco, southwestern Puerto Rico; section exposing Early Oligocene coastal plain environment ofJuana Diaz Fm (sensu Frost et al., 1983). Bottom, View of ?latest Eocene-Early Oligocene poorly sorted alluvial conglomerates at base ofJuana Diaz Fm, Highway 52 near town of Juana Diaz, southern Puerto Rico; red sandy clays exposed beneath uncon- formity represent a paleosol. Coeval conglomerates onlapping older deposits are found in Puerto Rico, Cuba, and Hispaniola, where they represent subaerial erosion products formed during the phase ofactive late Paleogene uplift discussed in text. 18 AMERICAN MUSEUM NOVITATES NO. 3141

TABLE 3 Locality AMNH PR 1994/1 (Yauco): Identified Foraminifera Samplesa Taxa 94/16A(UC) 94/16B(JD) 94/16C(JD) 94/16D(JD) Baggina, aff. B. xenoulab X Lepidocyclina undosa'c X X X Lepidocyclina gigasc X Lepidocyclina, cf. L. canelleic X Lepidocyclina, cf. L. giraudiac X Lepidocyclina, cf. L. vaughani(?)c X Lepidocyclina, cf. L. canellei or L. vaughanic X Pararotalia mexicanab X X Pararotalia, cf. P. guantanamensisb X Gypsina sp.c X Palaeonummulites diusc X cf. X Amphistegina sp.c X Globigerina, cf. G. angustiumbilicatab X a For location of sampling horizons, see fig. 8. b Identified by Lic. Gena Fernandez (CIDP-Cupet, Cuba). c Identified by Prof. Edward Robinson (University of the West Indies, Mona, Jamaica).

Rosen's (1975, 1985) concept, which as- The simultaneous and extensive develop- sumes that most of the critical vicariantly- ment ofcoarse-grained clastic deposits in the controlled events in Caribbean biogeograph- Greater Antilles and northern South America ical history were "continent-island." New pa- during the latest Eocene-Early Oligocene (fig. leontological discoveries and reconsideration 10) indicates that active uplift took place ofpertinent geological evidence permit some throughout the positive portions of the Ca- additional elaborations, presented below. ribbean, including the Aves Rise, at essen- In our model we specifically proposed that tially the same time. Likewise, recovery of land mammal dispersal to the Greater An- Late Eocene and Early Miocene shallow wa- tilles could have occurred across portions of ter limestones from the walls ofthe Aves Rise the Aves Rise, if (as is often assumed) parts (Fox et al., 1971) confirms the positive char- of this feature were subaerial for a limited acter ofthe Aves during the early and middle time in the mid-Tertiary. Other possibilities, Tertiary. An intriguing feature of the dredge such as movements across the Yucatan samples described by Fox et al. (1971) is that Channel, Nicaragua Rise, or Lesser Antilles, no Oligocene rocks were identified in any of cannot be completely excluded but are much them. Although the evidence is of a negative less plausible for land mammals for one rea- quality, the absence of Oligocene rocks in son or another (briefly summarized below). these samples correlates with the presence of Holcombe and Edgar (1990) considered same-age hiatuses and subaerially deposited fairly elaborate scenarios in which islands of conglomerates elsewhere in the Greater An- various sizes and interconnectivities were re- tilles (fig. 10). Taken together, this evidence constructed for the spine of the Aves Rise, tends to confirm the view that the Aves Rise assuming differing subsidence regimes. They was uplifted around the time of the Eocene- noted in particular that the thin sedimentary Oligocene boundary, although it is to be hoped deposits covering large parts ofthe Aves Rise that some day this inference will be con- (<500 m of Late Cretaceous and Cenozoic firmed by a coring program. sediment combined) imply that the Rise was The late Paleogene uplift event in the Ca- positive with respect to the sea floor during ribbean region, which was coincidental with most of the Cenozoic. Other facts also sup- the worldwide Pyrenean phase of tectogene- port the contention that a major part of the sis (Leonov and Khain, 1987; Brezsnyanszky Aves Rise must have once been emergent. and Iturralde-Vinent, 1978, 1987), resulted 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 19

MID-TERTIARY SEDIMENTARY ENVIRONMENTS AND STRATIGRAPHIC UNITS OF THE GREATER ANTILLES AND SOUTH AMERICA - u 1. : T I T I

ii- I -1 I

Conglomerates: poorly sorted Sands and clays: shallow marine Limestones: shallow marine 1 Deep marine related to active uplift and coastal plain environments shelf environments Fig. 10. Mid-Tertiary sedimentary environments and stratigraphic units for selected areas in the Greater Antilles, Aves Rise, and northern South America (Venezuela). Note (a) correlated deposition of coarse conglomerates in the latest Eocene-Early Oligocene throughout the area, suggesting regional uplift, followed by (b) marine sedimentation from mid-Oligocene through Middle Miocene, indicating regional transgression. Under the dispersal hypothesis favored in this paper, the optimal period for land mammal colonization of the Greater Antilles was at or near the Eocene-Oligocene boundary; by the Early Miocene transgression, the window was closed. Sources: Cuba-Iturralde-Vinent (1969, 1972), Nagy et al. (1983), Brezsnyanszky and Iturralde-Vinent (1978), Cobiella et al. (1977); Hispaniola-Lewis et al. (1990), Mann et al. (1992); Puerto Rico-Monroe (1980), Frost et al. (1983); Aves Rise-Fox et al. (1971), Holcombe et al. (1990); northern South America (Venezuela)-Anonymous (1970), Dallmus and Graves (1972). in the creation of a large positive feature on distributed as they are today, but instead the Caribbean sea floor-GAARlandia, formed a much more compact array, with the formed by the Greater Antilles Ridge and the eastern islands being situated considerably to Aves Rise. Ridge and Rise are also related as the west of their present locations (Malfait members of the original "Great Arc of the and Dinkelman, 1972; Ross and Scotese, Caribbean" formed in the Cretaceous (Burke, 1988; Van Fossen and Channell, 1988; Pin- 1988). At present the chief subaerial com- dell and Barrett, 1990). ponents of GAARlandia are Cuba, Hispan- In addition to general uplift, we originally iola, Puerto Rico, Virgin Islands, Aves Is- posited that dispersals could be more pre- land, and some islands on the northern fringe cisely correlated with the dramatic lowering of South America (e.g., parts of the Vene- of global sea level ca. 29 Ma identified by zuelan Antillas Menores). Prior to the Late Haq et al. (1987), since a major regression Eocene, Cuba, northern Hispaniola, Puerto would have increased the total land area over Rico and the Virgin Islands were not linearly which animals might have dispersed. A nec- 20 AMERICAN MUSEUM NOVITATES NO. 3141 essary corollary of this proposition is that barrier to the circulation ofintermediate and mammal fossils much older than mid-Oli- deep water at -40 Ma. Donnelly noted that gocene time would not be expected to occur nothing more than a discontinuous island arc in Antillean localities. However, the Yauco was implied by this evidence. Whether or not sloth demonstrates that tardigradans were al- surface water circulation was broadly inter- ready in GAARlandia by the Early Oligo- rupted at this time has not been investigated. cene. Evidently, the mid-Oligocene sea level If terrestrial colonizations could have oc- drop of -160 m did not initiate dispersal, at curred in the latest Eocene or Early Oligo- least in the case of sloths. Either the sea level cene, why could they not have occurred ear- drop under discussion is not accurately dated, lier, such as in the Middle Eocene? Earlier or was not global, or for some other reason colonizations cannot be ruled out, but even did not affect GAARlandia in the way orig- if they occurred, it seems unlikely that any inally imagined. colonizing clades of that age could have per- By contrast, the evidence for the wide- sisted into the later Cenozoic. Although sub- spread effect ofuplift correlated with the Pyr- aerial landmasses ofsome sort existed within enean tectogenetic phase is consistent the Caribbean Sea from time to time from throughout the sampled region (fig. 10). In the Cretaceous onward, they would have been Cuba, there is considerable evidence for de- subjected repeatedly to episodes of subsi- positional hiatuses and extensive develop- dence and transgression (most notably in ment of coarse-grained, poorly sorted con- Maastrichtian, Late Paleocene/Early Eocene, glomerates during the Late Eocene-Early Oli- and Middle Eocene [Khudoley and Meyer- gocene (Iturralde-Vinent, 1972; Nagy et al., hoff, 1971; Iturralde-Vinent, 1982; Maur- 1983; Brezsnyanszky and Iturralde-Vinent, rasse 1990]). Our conclusion is that mam- 1978, 1985). The same is true for Hispaniola, mals, at least, could not have successfully col- where coeval uplift produced most of the onized and survived on the nascent islands present Cordillera Central and Cordillera of the Greater Antilles until these bodies be- Oriental. At this time, northern Hispaniola camepermanently subaerial subsequent to the and the eastern end of Cuba were physically beginning ofthe Late Eocene, when the whole connected as subaerial landmasses (Calais et belt was tectonically deformed (Khudoley and al., 1992; Lewis et al. 1990; Mann et al., 1992; Meyerhoff, 1971; Lewis et al., 1990; Itur- Iturralde-Vinent, 1994b). Coarse-grained ralde-Vinent, 1982, 1994a). conglomerates and coastal depositional en- The last or NeoCaribbean tectonic phase vironments certainly appear in Early Oligo- of Greater Antillean evolution, from Mio- cene facies in Puerto Rico (fig. 9; Monroe, cene to Recent, has been described in several 1980), and, as noted above, it is possible that papers (Mann and Burke, 1984; Larue et al., the basal conglomerates of the Juana Diaz 1990; Mann et al., 1990). During this phase, Fm are even older (latest Eocene). These ob- uplift was followed by a general transgression servations demonstrate that correlated uplift (fig. 9). In addition, several pull-apart basins also occurred in the eastern Greater Antilles, and related features opened or expanded along creating subaerial connections between these the northern Caribbean plate boundary, islands as Meyerhoff (1933) originally sug- among them the Cayman Trough southeast gested. One implication of this is that of Cuba (Iturralde-Vinent, 1988, 1991; Ca- GAARlandia might have acted as a barrier lais and Lepinay, 1991), Mona Canyon (La- between Atlantic and Caribbean circulations, rue et al., 1990; Masson and Scanlon, 1991), which would in turn help to explain such ap- and the Sombrero Basin-Anegada Passage parent anomalies as the striking difference system east of the Virgin Islands (Mann et between Oligocene faunal assemblages on the al., 1990; Jany et al., 1990). As a result, deep- north and south sides of Puerto Rico (Bold, water channels and basins now divide the 1971; Kaye, 1959). These observations are blocks comprising the Greater Antilles Ridge consistent with Donnelly's (1989, 1990) in- into several isolated tectonic units, some of ference, from analysis ofbiological silica con- which have large subaerial portions. The tent ofpelagic sediments, that the Aves Ridge marked subsidence ofthe Aves Rise (by 400- (and the nascent Lesser Antilles) acted as a 1400 m from Miocene onward) would have 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 21

Fig. 11. Section of slightly deformed Early-Middle Miocene clastic carbonate deposits, SE Cuba; viewpoint is from east. Sandwiched between subhorizontal beds, a thick horizon composed of large clinoforms dips south (i.e., in direction of the Cayman Trough) at an angle of -50°. prevented this structure from acting as a con- Cuba probably represents a rotation of the nector at any time in the Neogene (Holcombe direction of stresses throughout the Carib- and Edgar, 1990; Holcombe et al., 1990). bean, from ENE-WSW to more or less di- Evidence for the factors that produced the rectly E-W. The changeover correlates well NeoCaribbean tectonic event is well repre- with the Savic tectogenetic event at the Oli- sented in southeastern Cuba. Along the is- gocene-Miocene boundary recorded else- land's southern flank (=north wall ofthe Cay- where in the world (Schwan, 1980; Leonov man Trench), various compressional features and Khain, 1987). A significant consequence are recorded in the rock record, including ofCaribbean stress-field rotation was disrup- overthrusting, reverse faulting, tight folding, tion and dispersion ofthe terranes that formed and strong fracturing of latest Eocene and Hispaniola-Puerto Rico-Virgin Islands (Lewis older rocks (Lewis and Straczek, 1955; Cob- et al., 1990; Mann et al. 1990; Larue and Ryan, iella et al., 1977). By contrast, Miocene and 1990; Larue et al., 1990) along the northern younger deposits are undeformed, showing boundary of the Caribbean Plate. Equivalent only a few degrees of tilting and deepening disruptions occurred along the southern toward the south (Lewis and Straczek, 1955; boundary ofthe Caribbean Plate, affecting Cu- Cobiella et al., 1977), probably in relation to ra,ao, Aruba, and Venezuelan Antillas Me- a transtensional tectonic environment. Low- nores (Dailmus and Graves, 1972; Pindell and er-Middle Miocene clastic rocks outcropping Barrett, 1990; Mann et al., 1990). in fault-bounded grabens in southeastern Finally, note must be made of the severe Cuba display very large south-dipping cli- constraints operating on other apparently noforms, indicating the existence nearby of plausible routings for biotic immigration: a steeply walled deep (fig. 11). The only pos- Central America sible candidate is the Cayman Trough, which (Nicaragua Rise-Cayman Ridge) must therefore have come into existence about this time (Iturralde-Vinent 1991, 1 994b). The The Nicaragua Rise and Cayman Ridge change in tectonic regime from transpres- were evidently a single geological entity be- sional (pre-Miocene) to transtensional (Mio- fore the opening ofthe Cayman Trough; both cene and thereafter) recorded in southeastern have vulcano-plutonic arc rocks oflatest Cre- 22 AMERICAN MUSEUM NOVITATES NO. 3141 taceous-Paleogene age as partial basement have influenced the dispersal ofthe land mam- (Holcombe et al., 1990; Holcombe and Ed- mal groups known from the Greater Antilles. gar, 1990). The Nicaragua Rise is sometimes cited as a possible avenue for mammals into Other Possibilities the Greater Antilles via Jamaica (see discus- Although the Lesser Antilles seem to be sion by Buskirk, 1985), often in connection well situated to act as a discontinuous cor- with an assumed North American origin for ridor across which mammals might have mi- Antillean lipotyphlans. However, emergence grated from South America, the evidence is of the entire Nicaragua Rise-Cayman Ridge against any taxa having done so at a point complex seems unlikely because each com- early enough to comport with the fossil evi- ponent is draped by later Cenozoic marine dence now coming out of the Greater Antil- deposits, implying submergence during the les. Most ofthe evolutionary development of period of interest. These sediments are Eo- this arc occurred subsequent to the latest Eo- cene and younger in the case of the Cayman cene-Oligocene (Maury et al., 1990). Nev- Ridge, Eocene through Quaternary on the ertheless, a very limited suite of mammalian Nicaragua Rise near the Central American taxa got partway up the chain by some mech- mainland, and Eocene through Miocene in anism. This suite includes a sloth and a capy- Jamaica (Holcombe et al., 1990; Maurasse, bara in Grenada during the Pliocene (R. Singer, 1990). Although there are certain faunal sim- personal commun.) and sigmodontines in the ilarities between Jamaica and the rest of the northernmost Lesser Antilles during the Qua- Greater Antilles, there are also substantial ternary (Morgan and Woods, 1986). Sloths, of differences (e.g., absence of insectivores) course, occurred in the Greater Antilles, but which imply that its contacts with other is- sigmodontines did not, except in Jamaica where lands have been limited (Buskirk, 1985; Mor- one species is known from Recent deposits gan, 1993). (Morgan, 1993). Another imponderable is that Yucatan Channel Anguilla and St. Martin, but no other Lesser Offshore seismic and sonar research in the Antillean islands, bore a heptaxodontid rodent western Caribbean Sea (see especially Case et (Amblyrhiza) related to Elasmodontomys from al., 1990; Holcombe et al., 1990) conclusively Puerto Rico. Each ofthese cases represents an demonstrates that a land connection did not interpretative problem, but it needs to be noted exist between western Cuba and the Yucatan also that several ofthese "anomalous" taxa are Peninsula during the Tertiary. The topography very poorly described (or even undescribed) of the basin between westernmost Cuba and and additional information on their phyloge- Cozumel consists of a series of unconnected netic relationships is sorely needed. ridges separated by deep trenches, some ofthem The Beata Rise, south of Hispaniola, does >2 km below sea level. Therefore, sea-level not physically extend as far as South America drops, even ofgreat magnitude, would not have (Case et al., 1990). Its age is uncertain, al- appreciably affected the paleogeography ofthis though Santonian and younger sediments oc- area. In any case, latest Eocene and Oligocene cur on the structure (Holcombe et al., 1990). sinistral strike-slip motion along the Pinar Fault Furthermore, its aspect as a submarine prom- would have acted to widen, not shorten, the ontory of Hispaniola may be a relatively re- separation between Cuba and the Yucatan Pen- cent phenomenon (Mann and Burke, 1984; insula (Iturralde-Vincent, 1994a). It is possible Mann et al., 1990). The Rise has subsided that a much older connection (?late in the Mid- - 1000 m since the Middle Eocene; but since dle Cretaceous) between Central America and most of the Ridge lies at -2500 to -3000 westernmost Cuba once existed, since meta- m, its opportunity for subaerial exposure must morphic rocks recovered from ridges in the have been limited (Holcombe et al., 1990). Yucatan Channel are similar to those outcropping in the Maya Mountains of Belize and in SIGNICANCE OF FOSSIL western Cuba (Pindell and Dewey, 1982). LAND-MAMMAL DIscovERIEs However, even ifthis connection involved sub- Three ofthe four major clades ofAntillean aerial exposure oflands, it was far too early to mammals known from the Quaternary- Tar- 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 23 digrada, Caviomorpha, and Platyrrhini- are How did isolobodontines reach Cuba? One now known, for certain, to have existed on possibility is that Zazamys veronicae repre- landmasses in the central Caribbean Basin sents a unique occurrence of isolobodontines during the mid-Tertiary (table 1). The only in Cuba, its proximate ancestor having been major taxon now missing from the Tertiary rafted in from Hispaniola, the alleged center picture is Lipotyphla, and its absence is surely of isolobodontine diversity (Woods, 1989a). due to the vagaries of fossil discovery, not A difficulty with this interpretation is that actual absence. It is equally important to note Zazamys is morphologically more primitive that, by contrast, nothing of an unexpected than Isolobodon, its apparent sister taxon, nature has yet shown up -no marsupials, no which-reversing the logic implicit in center- carnivores, no condylarths, and so on (cf. of-diversity argumentation -indicates that Perfit and Williams, 1989). In short, to the Hispaniolan isolobodontines are derived from degree that the few fossil discoveries made a Cuban source. to date can be considered representative of A more interesting possibility is that the mid-Tertiary diversity, the Antillean land complementary distribution ofZazamys and mammal fauna was essentially the same then Isolobodon is strictly due to vicariance, i.e., as now in terms of major taxa. A similar the subdivision of populations through the pattern appears to be emerging for other ver- erection ofa barrier, in this instance the sub- tebrate groups, in the sense that specimens division ofland. Island-island vicariance is ro- complete enough to permit confident assign- bust in that it provides a common-cause ex- ment to named lower-level groups have all planation for a number of distributional fea- turned out to be referable to extant clades tures and potential sister-group pairings. As having Antillean representation (table 1). The noted above, there is solid geological support Tertiary arthropod fauna conflicts with this forthe contention that the Windward and Mona neat picture to some degree, for it includes Passages did not exist as such prior to the Mio- several "unexpected" taxa that are not part cene. Thus, during the critical period when land of the modern Neotropical fauna (Wilson, mammals were most likely to have entered the 1988; D. Grimaldi, personal commun.). In Greater Antilles- latest Eocene through Early addition, according to Graham (1990), the Oligocene (fig. 10) -Cuba, northern Hispan- Tertiary flora of the islands has a distinctly iola, and Puerto Rico probably constituted a North American aspect, quite unlike the single landmass (MacPhee and Iturralde-Vi- modern flora which is overwhelmingly South nent, 1994). However, this situation could not American in affinity. However, as Borhidi have prevailed for a lengthy period: Cuba was (1985) notes, in addition to North and South already separated into three large archipelagos American components there are also ele- by the beginning of the Early Miocene (Itur- ments ofAfrican, Pacific, and Tethyan origin. ralde-Vinent, 1969), and eastern Cuba and It is to be hoped that additional discoveries northern Hispaniola would no longer have been will add to the database needed to resolve the in contact by this time (Calais et al., 1992). complex botanical history of these islands. This information suggests that the "orderly" distribution ofmany Quatemnary Antillean taxa ZAZAMYS FROM CUBA may not have been due to dozens ofindividual Zazamys veronicae is a true biogeograph- rafting events among the various islands, cre- ical "ghost"-that is, an extinct member of ating a superficial, ahistorical faunal similarity, a clade that lived outside the clade's supposed but instead to the dissociation of the islands distribution area (and whose existence was themselves as they progressively parted com- unsuspected prior to the discovery of fossil pany with their resident biotas already in place. material). Quaternary representatives of Iso- Some sister-group pairings already recognized lobodontinae are known from Hispaniola, or commented upon in the literature that may Puerto Rico, and some ofthe Virgin Islands, be a result of vicariant events affecting Cuba but because no members of this clade sur- and Hispaniola include Miocnus + Synocnus vived into Pleistocene/Recent time in Cuba, in Megalonychidae (cf. Varona, 1974), Neso- they were assumed never to have occupied phontes micrus + N. paramicrus in Neso- that island (e.g., Simpson, 1956). phontidae (cf. Patterson, 1962), and Par- 24 AMERICAN MUSEUM NOVITATES NO. 3141 alouatta varona + Antillothrix bernensis in Pla- persalist hypothesis that they arrived later. tyrrhini (MacPhee et al., 1995). For the reasons However, this introduces new complications, cited above, island-island vicariance also ac- since in the case of Capromyidae, the non- counts for the existence of Zazamys in Cuba. isolobodontine subfamilies as defined by It can be predicted that other "ghosts" having Woods must have originated as early as or ear- closest relatives on adjacent islands will be lier than Isolobodontinae. Where were th. y, if found, as the Antillean fossil record continues not in the Antilles? Similarly, with the discov- to improve. ery ofan Early Oligocene sloth in Puerto Rico, Zazamys has a direct bearing on another Woods' (1989a: 771) speculation that sloths problem, which is the estimation of the an- may not have entered the Greater Antilles until tiquity ofcertain clades ofAntillean rodents. the Late Miocene can also be rejected. Previously, Woods (1982, 1989a) inferred that the rodent colonization(s) took place in the medial to later Oligocene, which is in reasonable agreement with the evidence pre- ZAZA ANTHROPOID sented here. However, in his most detailed Too little is known of the Zaza primate to analysis of the problem, Woods (1989b: 83) build a picture of its phylogenetic relation- concluded that Antillean capromyids and ships, but its similarity to Paralouatta (and echimyids are derived from an echimyid an- the latter's distinctiveness within Platyrrhini) cestor that raises one issue that needs brieftreatment. In the systematic literature concerning endemic ... invaded the Greater Antilles from South West Indian monkeys, it is frequently as- America either via island-hopping through the sumed that these monkeys have to be the Lesser Antilles or by rafting directly to Puerto little-differentiated, insular representatives of Rico or eastern Hispaniola, which has served Wil- as the center of evolution for capromyids. The Recent mainland clades. Thus although time of this invasion is unknown but all avail- liams and Koopman (1952) recognized the able data suggest it was after the Oligocene and morphological distinctiveness of the holotype probably as recently as the late Miocene. The of Xenothrix mcgregori and were ultimately first radiation gave rise to plagiodontine, hex- unable to fit it within the confines of any ex- olobodontine, and isolobodontine lineages on isting subfamily, they still concluded that it Hispaniola, and later (Pliocene and Pleistocene) could be comfortably regarded as "cebid." In to the capromyine lineage that inhabits Cuba, a similar vein, Rosenberger (1977) and Rosen- Jamaica, and the Bahamas. berger et al. (1990) argued that Xenothrix is a I propose that spiny rats (Heteropsomyinae kind ofpitheciine because ofsimilarities injaw tentatively retained in the Echimyidae), hutias resem- (Capromyidae sensu stricto), and giant hutias shape, despite the absence of detailed (Heptaxodontidae sensu lato) all invaded the blances in molar morphology. Likewise, Ri- Greater Antilles sometime after the middle moli (1977) preferred to deposit the unusual Miocene. monkey maxilla from Cueva Berne within Sai- miri (as S. bernensis) rather than consider the Zazamys proves that the minimum age for more radical alternative that it represented a the start ofthe invasion, however it may have distinct clade. In a rather different way Ford occurred, could have been no later than Early and Morgan (1986) and Ford (1990) have also Miocene, not Middle or Late Miocene, and perpetuated this approach, although they have for geological reasons was assuredly earlier emphasized morphology over systematic affil- still. Furthermore, because it is identifiably iations. isolobodontine, the morphology ofZazamys The assumption that -if we just look long demonstrates that the cladistic split between and hard enough -every Antillean fossil isolobodontines and their sister taxon within monkey can be shoehorned into one or an- Capromyidae must also have occurred before other extant clade of mainland platyrrhines the late Early Miocene. Since there are no at a low categorical level is no longer defen- Tertiary fossils for other groups of Antillean sible in either theory or practice. Discovery rodents, it might still be maintained in a dis- of the Zaza anthropoid demonstrates that 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 25 platyrrhines were certainly present in the because ofthe lack ofpermanent landmasses Greater Antilles at the start of the Neogene, within the Caribbean Sea. Since the emergent and probably considerably earlier. Ifthe Zaza parts of the Greater Antilles Ridge were in a anthropoid is a member of the same mono- close-packed array until the commencement phyletic group as the one recently recognized of the Neogene, it is reasonable to believe for Paralouatta and Antillothrix by MacPhee that initiator populations could have been et al. (1995), the implication is that the cla- well distributed across GAARlandia. With distic event separating this group from its the fragmentation of GAARlandia, progres- closest mainland relatives must have been at sive divergence of vicariated lineages would least as early-making it one of the oldest have begun; also, some island lineages would nodes yet recovered within Platyrrhini. later become extinct (e.g., isolobodontines in Cuba). It is possible that fossil members of founder clades will eventually be discovered SUMMARY in each of the modem Greater Antilles, with Recent discoveries in Cuba and Puerto Rico the possible exception ofJamaica, whose pa- establish that the initiators ofmost (and per- leoposition during the Eocene-Oligocene haps all) founder clades of Antillean land needs to be better constrained in our opinion. mammals were emplaced on subaerial por- Other mammalian clades may have dis- tions of GAARlandia by the mid-Tertiary. persed at the same time as the familiar An- Evidence for a specific horizon in the mid- tillean clades that survived into the Quater- Tertiary is still thin, but in our opinion the nary, but if so they must have died out soon immigration event(s) occurred at or near the after arrival. No such clades have yet been Eocene-Oligocene boundary, probably with- identified. Our model may or may not be in a very narrow interval. We consider it un- generalizable to other Antillean vertebrate likely that successful emplacements could taxa; this will have to be settled by the ap- have occurred much earlier than Late Eocene propriate specialists in these groups.

REFERENCES Allen, G. M. matic and biotic evolution, pp. 29-45. 1942. Extinct and vanishing mammals of the Princeton, NJ: Princeton Univ. Press. western hemisphere, with the marine Biknevicius, A. R., D. A. McFarlane, and R. D. species of all the oceans. New York: E. MacPhee Cooper Square Publishers (1972 re- 1993. Body size in Amblyrhiza inundata print). (Heptaxodontidae, Caviomorpha), an Anonymous extinct megafaunal rodent from the An- 1970. Lexico estratigrafico de Venezuela. Bol. guilla Bank, West Indies: estimates and de Geol., Publ. Espec. 4: 756 pp. implications. Am. Mus. Novitates 3079: Anthony, H. B. 25 pp. 1926. Mammals of Porto Rico, living and ex- Bohme, W. tinct-Rodentia and Edentata. Scientif- 1984. Erstfund eines fossilen Kugelfinger- ic survey of Porto Rico and the Virgin gecko (Sauria: Gekkonidae: Sphaero- Islands 9(2): 97-241. New York: New dactylinae) aus Dominikanischen Bern- York Academy of Sciences. stein der Oligozan von Hispaniola, An- Arredondo, 0. tillen. Salamandra 20: 212-220. 1988. Nueva subfamilia de Megalonychidae Bold, W. A. van den (Mammalia: Edentata). Misc. Zool. 1971. Distribution of ostracodes in the Oli- Acad. Sci. Cuba 39: 2 pp. gomiocene of the northern Caribbean. Berggren, W. A., D. V. Kent, J. D. Obradovich, Trans. 5th Caribbean Geol. Conf., pp. and C. C. Swisher 123-128. New York: Queens College 1992. Toward a revised Paleogene geochro- Press. nology. In D. R. Prothero and W. A. Borhidi, A. Berggren (eds.), Eocene-Oligocene cli- 1985. Phytogeographic survey ofCuba. 1. The 26 AMERICAN MUSEUM NOVITATES NO. 3141

phytogeographic characteristics and Corbet, G. B., and J. E. Hill evolution ofthe flora ofCuba. Acta Bot. 1980. A world list of mammalian species. Hungarica 31: 3-34. London: British Museum (Natural His- Brezsnyanszky, K., and M. A. Iturralde-Vinent tory). 1978. Paleogeographia del Paleogeno de Cuba Dallmus, K. F., and G. R. Graves oriental. Geol. Mijnbouw 57: 123-133. 1972. Tectonic history ofMaracaibo and Bari- 1985. Paleogeografia del Paleogeno de La Ha- nas basins. Trans. 6th Caribbean Geol. bana. In J. Albear et al. (eds.), Contri- Conf., pp. 214-226. Caracas: Escuela de bucion a la geologia de las provincias Geologia y Minas. de La Habana, pp. 100-115. La Ha- Donnelly, T. W. bana: Academia de Ciencias de Cuba. 1989. History of marine barriers and terres- 1987. On two Tertiary tectonic phases in Cuba. trial connections: Caribbean paleogeo- In Yu. G. Leonov and V. E. Khain (eds.), graphic inference from pelagic sediment Global correlation of tectonic move- analysis. In C. A. Woods (ed.), Bioge- ments, pp. 227-230. New York: Wiley. ography of the West Indies, pp. 103- Burke, K. 118. Gainesville, FL: Sandhill Crane 1988. Tectonic evolution of the Caribbean. Press. Annu. Rev. Earth Planet. Sci. 16: 201- 1990. Caribbean biogeography: geological 230. considerations bearing on the problem Burke, K., K. Cooper, J. F. Dewey, P. Mann, and of vicariance vs. dispersal. In A. Az- J. L. Pindell zaroli (ed.), Biogeographical aspects of 1984. Caribbean tectonics and relative plate insularity, Atti Convegni Lincei 85: 595- motions. In W. E. Bonini, R. B. Har- 609. Rome: Accademia Nazionale dei graves, and R. Shagam (eds.), The Ca- Lincei. ribbean-South American plate bound- Fischer, K. ary and regional tectonics. Geol. Soc. 1971. Riesenfaultiere (Megalonychidae, Am. Mem. 162: 31-63. Edentata, Mammalia) aus dem Pleis- Buskirk, R. E. tozan der Pio-Domingo-Hohle in Kuba. 1985. Zoogeographic patterns and tectonic Wiss. Z. Humbolt-Univ. Berlin Math.- history ofJamaica and the northern Ca- Naturwiss. Reihe 20(4/5): 609-673. ribbean. J. Biogeogr. 12: 445-461. Calais, E., and B. M. de Lepinay Ford, S. M. 1991. From transtension to transpression along 1990. Platyrrhine evolution in the West In- the northern Caribbean plate boundary dies. J. Hum. Evol. 19: 237-254. off Cuba: implications for the Recent Ford, S. M., and G. S. Morgan motion of the Caribbean plate. Tec- 1986. A new ceboid femur from the Late Pleis- tonophysics 186(1991): 329-350. tocene of Jamaica. J. Vertebr. Paleon- Calais, E., B. M. de Lepinay, P. Saint-Marc, J. tol. 6: 281-289. Butterlin, and A. Schaaf Fox, P. J., E. Schreiber, and B. C. Heezen 1992. La limite de plaques decrochante nord 1971. The geology ofthe Caribbean crust: Ter- cara1be en Hispaniola: evolution paleo- tiary sediments, granitic and basic rocks geographique et structural cenozo;que. from the Aves Ridge. Tectonophysics Bull. Soc. Geol. France 163: 309-324. 12: 89-109. Case, J. E., W. D. McDonald, and P. J. Fox Frost, S. H. J. L. Harbour, D. K. Beach, M. J. 1990. Caribbean crustal provinces; seismic and Realini, and P. M. Harris gravity evidence. In G. Dengo and J. E. 1983. Oligocene reef track development, Case (eds.), The geology ofNorth Amer- southwestern Puerto Rico. Sedimenta, ica, vol. H, The Caribbean region, pp. vol. 9, pp. 1-144. Miami, FL: Univer- 15-36. Boulder, CO: Geological Society sity of Miami RSMAS. of America. Garrison, L. E., R. G. Martin, H. L. Berryhill, M. Cobiella, J., M. Campos, A. Boiteau, and F. Quin- W. Buell, H. R. Ensminger, and R. K. Perry tas 1972. Preliminary tectonic map ofthe eastern 1977. Geologia del flanco sur de la Sierra del Greater Antilles region. USGS Misc. Purial. Rev. Mineria Cuba 3(1): 54-62, Geol. Inv., Map 1-732, scale 1:500,000. (2) 44-53. Graham, A. Cockerell, T. D. A. 1990. Late Tertiary microfossil flora from the 1924. A fossil cichlid from the Republic of Republic of Haiti. Am. J. Bot. 77: 911- Haiti. Proc. U.S. Nat. Mus. 63(7): 1-2. 926. 1 995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 27

Grimaldi, D. tectonica de Cuba Oriental, pp. 24-27. In press. On the age ofDominican amber. In K La Habana: Instituto de Geografia, ACC. B. Anderson and J. C. Crelling (eds.), Am- 1994a. Cuban geology: A new plate-tectonics ber, resinites, and fossil resins. Washing- synthesis. J. Pet. Geol. 17: 39-70. ton, D.C.: American Chemical Society. 1994b. Tectonostratigraphic correlation of the Haq, B. U., J. Hardenbol, and P. R. Vail NW Caribbean: Dominican Republic. 1987. Chronology offluctuating sea levels since J. Pet. Geol. 17: 243-245. the Triassic. Science 235: 1156-1166. Jaimez Salgado, E., D. Gutierrez Calvache, R. D. Harland, W. B., R. L. Armstrong, L. E. Craig, A. E. MacPhee, and G. C. Gould G. Smith, and D. G. Smith 1992. The monkey caves of Cuba. Cave Sci. 1990. A geologic time scale. Cambridge: Cam- 19: 25-28. bridge Univ. Press. Jany, I., K. M. Scanlon, and A. Mauffret Hedges, S. B., C. A. Hass, and L. R. Maxson 1990. Geological interpretation of combined 1992. Caribbean biogeography: molecular ev- seabeam Gloria and seismic data from idence for dispersal in West Indian ter- Anegada Passage (Virgin Islands, North restrial vertebrates. Proc. Natl. Acad. Caribbean). Mar. Geophys. Res. 12: Sci 89: 1909-1913. 173-196. 1994. Towards a biogeography of the Carib- Kaye, C. A. bean. Cladistics 10: 43-55. 1959. Geology ofIsla Mona, Puerto Rico, and Holcombe, T. L., and N. T. Edgar notes on the age of Mona Passage. U.S. 1990. Late Cretaceous and Cenozoic evolu- Geol. Surv., Prof. Pap. 317-C: 178 pp. tion of Caribbean ridges and rises with Khudoley, K., and A. A. Meyerhoff special reference to paleogeography. In 1971. Paleogeography and geological history A. Azzaroli (ed.), Biogeographical as- of the Greater Antilles. Geol. Soc. Am. pects ofinsularity, Atti Convegni Lincei Mem. 129: 1-199. 85: 611-626. Rome: Accademia Na- Kluge, A. G. zionale dei Lincei. 1995. Cladistic relationships of sphaerodactyl Holcombe, T. L., J. W. Ladd, G. Westbrook, N. lizards. Am Mus. Novitates 3139: 23 pp. T. Edgar, and C. L. Bowland Ladd, J., T. Shih, and C. Tsai 1990. Caribbean marine geology; ridges and 1981. Cenozoic tectonics of Central Hispan- basins ofthe plate interior. In G. Dengo iola and adjacent Caribbean Sea. Am. and J. E. Case (eds.), The geology of Assoc. Pet. Geol. Bull. 65: 466-489. North America, vol. H, The Caribbean Larue, D. K., and H. Ryan region, pp.231-260. Boulder, CO: Geo- 1990. Extensional tectonism in the Mona pas- logical Society of America. sage, Puerto Rico and Hispaniola: A Iturralde-Vinent, M. A. preliminary study. Trans. 12th Carib- 1969. Principal characteristics ofCuban Neo- bean Geol. Conf., pp. 223-230. Miami: gene stratigraphy. Am. Assoc. Pet. Geol. Miami Geological Society. Bull. 53: 1938-1955. Larue, D. K., J. Joyce, and H. Ryan 1972. Principal characteristics of Oligocene 1990. Neotectonics ofthe Puerto Rico Trench: and Lower Miocene stratigraphy of extensional tectonism and forearc sub- Cuba. Am. Assoc. Pet. Geol. Bull. 56: sidence. Trans. 12th Caribbean Geol. 2369-2379. Conf., pp. 231-247. Miami: Miami 1978. Los movimientos tectonicos de la etapa Geological Society. de desarrollo plataf6rmico de Cuba. Laybourne, R. C., D. W. Deedrick, and F. M. Geol. Mijnbouw 57: 205-212. Hueber 1981. Nuevo modelo interpretativo de la evo- 1994. Feather in amber is earliest New World lucion geologica de Cuba. Rev. Cienc. fossil of Picidae. Wilson Bull. 106: 18- Tierra Espac. 3: 51-90. 25. 1982. Aspectos geologicos de la biogeografia Leonov, Yu. G., and V. E. Khain de Cuba. Rev. Cienc. Tierra Espac. 5: 1987. Problems of a systematic approach to 85-100. the study and correlation of tectonic 1988. Naturaleza geologica de Cuba. La Ha- phases. In Yu. G. Leonov and V. E. bana: Editorial Cientifico-Tecnica. Khain (eds.), Global correlation of tec- 1991. Deslizamientos y descensos del terreno tonic movements, pp. 1-12. New York: en el flanco meridional de la Sierra Wiley. Maestra, Cuba sudoriental. In Morfo- Lewis, G. E., and J. A. Straczek 28 AMERICAN MUSEUM NOVITATES NO. 3141

1955. Geology ofsouth central Oriente, Cuba. Caribbean region, pp. 307-338. Boul- Bull. U.S. Geol. Sur. 975-D: 171-336. der, CO: Geological Society ofAmerica. Lewis, J. F., G. Draper, C. Bourdon, C. Bowin, P. Mann, P., G. Draper, and J. F. Lewis Mattson, F. Maurrasse, F. Nagle, and G. Pardo 1992. An overview of the geologic and tec- 1990. Geology and tectonic evolution of tonic development of Hispaniola. Geo- northern Caribbean margin. In G. Den- logic and tectonic development of the go and J. E. Case (eds.), The geology of North America-Caribbean plate North America, vol. H, The Caribbean boundary in Hispaniola, pp. 1-51. region, pp. 77-140. Boulder, CO: Geol. Boulder, CO: Geological Society of Soc. of America. America. MacPhee, R. D. E., and J. G. Fleagle Masson, D G., and K. M. Scanlon 1991. Postcranial remains of Xenothrix 1991. The neotectonic setting of Puerto Rico. mcgregori (Primates, Xenotrichidae) and Geol. Soc. Am. Bull. 103: 144-154. other late Quaternary mammals from Matthew, W. D. Long Mile Cave, Jamaica. Bull. Am. 1916. New sirenian from the Tertiary ofPorto Mus. Nat. Hist. 206: 287-321. Rico, West Indies. Ann. New York MacPhee, R. D. E., and M. A. Iturralde-Vinent Acad. Sci. 27: 23-29. 1994. First Tertiary land mammal from Maurrasse, F. J. M. R. Greater Antilles: an Early Miocene sloth 1990. Stratigraphic correlation for the circum- (Xenarthra, Megalonychidae) from Caribbean region. In G. Dengo and J. Cuba. Am. Mus. Novitates 3094:13 pp. E. Case (eds.), The geology of North 1995. Earliest monkey from Greater Antilles. America, vol. H, The Caribbean region, J. Hum. Evol. 28: 197-200. plates 4, 5a, 5b. Boulder, CO: Geolog- ical Society of America. MacPhee, R. D. E., and A. R. Wyss Maury, R. C., G. K. Westbrook, P. E. Baker, Ph. 1990. Oligo-Miocene vertebrates from Puerto Bouysse, and D. Westercamp Rico, with a catalog of localities. Am. 1990. Geology of the Lesser Antilles. In G. Mus. Novitates 2965: 45 pp. Dengo nd J. E. Case (eds.), The geology MacPhee, R. D. E., and M. Rivero de la Calle of North America, vol. H, The Carib- 1995. AMS 14C age determination for the al- bean region, pp. 141-166. Boulder, CO: leged "Cuban spider monkey," Ateles Geological Society of America. (=Montaneia) anthropomorphus. J. McKenna, M. C. Hum. Evol. 1980. Early history and biogeography ofSouth MacPhee, R. D. E., M. A. Iturralde-Vinent, and America's extinct land mammals. In R. E. S. Gaffney Ciochon and B. Chiarelli (eds.), Evo- In prep. Origin of the Greater Antillean land lutionary biology of New World mon- mammal fauna, 2: Summary ofgeolog- keys and continental drift, pp. 43-78. ical and paleontological investigations New York: Plenum Press. at Domo de Zaza, an Early Miocene lo- Meldrum, J. D. cality in south-central Cuba. 1990. New fossil platyrrhine tali from the Ear- MacPhee, R. D. E., I. Horovitz, 0. Arredondo, ly Miocene of Argentina. Am. J. Phys. and 0. Jimenez Vasquez Anthrop. 83: 403-418. 1995. A new genus for the extinct Hispaniolan Meyerhoff, H. A. monkey Saimiri bernensis (Rimoli, 1933. Geology of Puerto Rico. Puerto Rico 1977), with notes on its systematic po- Univ. Monogr., ser. B, 1: 1-306. sition. Am. Mus. Novitates 3134:21 pp. Monroe, W. H. Malfait, B. T., and M. G. Dinkelman 1980. Geology of the middle Tertiary forma- 1972. Circum-Caribbean tectonics and igne- tions of Puerto Rico. U.S. Geol. Surv. ous activity and the evolution of the Prof. Pap. 953: 93 pp. Caribbean Plate. Geol. Soc. Am. Bull. Morgan, G. S. 83: 251-272. 1993. Quaternary land vertebrates ofJamaica. Mann, P., and K. Burke In R. M. Wright and E. Robinson (eds.), 1984. Neotectonics of the Caribbean. Rev. Biostratigraphy of Jamaica. Geol. Soc. Geophys. Space Phys. 22: 309-362. Am. Mem. 182: 417-442. Mann, P., C. Schubert, and K. Burke Morgan, G. S., and C. A. Woods 1990. Revision of Caribbean neotectonics. In 1986. Extinction and the zoogeography ofWest G. Dengo and J. E. Case (eds.), The ge- Indian land mammals. Biol. J. Linn. ology of North America, vol. H, The Soc., London, 28: 167-203. 1995 MAcPHEE AND ITURRALDE-VINENT: TERTIARY MAMMALS 29

Moussa, M. T., and G. A. Seiglie Pindell, J. L., and J. F. Dewey 1970. Revision of mid-Tertiary stratigraphy 1982. Permo-Triassic reconstruction of west- of southwestern Puerto Rico. Am. As- ern Pangea and the evolution ofthe Gulf soc. Pet. Geol. Bull. 54: 1887-1898. of Mexico/Caribbean region. Tectonics 1975. Stratigraphy and petroleum possibili- 1: 179-212. ties of the middle Tertiary rocks in Poinar, G 0. Puerto Rico (Discussion). Am. Assoc. 1988. Hair in Dominican amber: evidence for Pet. Geol. Bull. 59: 163-168. Tertiary land mammals in the Antilles. Nagy, E., K. Brezsnyanszky, A. Brito, D. Coutin, Experientia 44: 88-89. F. Formell, G. Franco, P. Gyarmati, Gy. Ro- Poinar, G. O., and D. C. Cannatella docz, and P. Jakus 1987. An Upper Eocene frog from the Do- 1983. Contribucion a la geologia de Cuba ori- minican Republic and its implications ental. La Habana: Editorial Cientifico- for Caribbean biogeography. Science Tecnica. 237: 1215-1216. Page, R. D. M., and C. Lydeard Reig, 0. A. 1994. Towards a cladistic biogeography ofthe 1981. Teoria del origen y desarrollo de la fau- Caribbean. Cladistics 10: 21-41. na de mamiferos de America del Sur. Pascual, R., M. G. Vucetich, and G. J. Scillato- Mar del Plata, Argentina: Museo "Lo- Yane renzo Scaglia." 1990. Extinct and Recent South American and Reynolds, T. E., K. F. Koopman, and E. E. Wil- Caribbean Megalonychidae, edentates, liams and hystricomorphous rodents: out- 1953. A cave faunule from western Puerto Rico standing examples of isolation. In A. with a discussion of the genus Isolobo- Azzaroli (ed.), Biogeographical aspects don. Breviora 12: 8 pp. of insularity, Atti Convegni Lincei 85: Rieppel, 0. 627-640. Rome: Accademia Nazionale 1980. Green anole in Dominican amber. Na- dei Lincei. ture 286: 486-487. Patterson, B. Rimoli, R. 1962. An extinct solenodontid insectivore 1977. Una nueva especie de mono (Cebidae: from Hispaniola. Breviora 165: 11 pp. Saimirinae: Saimiri) de la Hispaniola. Patterson, B., and A. E. Wood Cuad. CENDIA, Univ. Aut. Santo Do- 1982. Rodents from the Deseadan Oligocene mingo 242: 5-14. of Bolivia and the relationships of the Rivero, M., and 0. Arredondo Caviomorpha. Bull. Mus. Comp. Zool. 1991. Paralouatta varonai, a new Quaternary 149: 371-543. platyrrhine from Cuba. J. Hum. Evol. Paula Couto, C. de 21: 1-11. 1979. Tratado de paleomastozoologia. Rio de Robinson, E. Janeiro: Academia Brasiliera de Cien- 1994. Jamaica. In S. K. Donovan and T. A. cias. Jackson (eds.), Caribbean geology: an Perfit, M., and E. E. Williams introduction, pp. 111-127. Kingston, 1989. Geological constraints and biological Jamaica: Univ. ofWest Indies Publish- retrodictions in the evolution ofthe Ca- ers Assoc. ribbean Sea and its surroundings. In C. Rosen, D. Woods (ed.), Biogeography of the West 1975. A vicariance model of Caribbean bio- Indies, pp. 47-102. Gainesville, FL: geography. Syst. Zool. 24: 431-464. Sandhill Crane Press. 1985. Geological hierarchies and biogeo- Pessagno, E. A. graphic congruence in the Caribbean. 1963. Planktonic foraminifera from the Juana Ann. Missouri Bot. Gard. 72: 636-659. Diaz Formation, Puerto Rico. Micro- Rosenberger, A. L. paleontology 9: 53-60. 1977. Xenothrix and ceboid phylogeny. J. Pindell, J. L., and S. Barrett Hum. Evol. 6: 461-481. 1990. Geological evolution of the Caribbean Rosenberger, A. L., T. Setoguchi, and N. Shige- region: a plate tectonic perspective. In hara G. Dengo and J. E. Case (eds.), The ge- 1990. The fossil record of callitrichine pri- ology of North America, vol. H, The mates. J. Hum. Evol. 19: 209-236. Caribbean region, pp. 405-432. Boul- Ross, M., and C. R. Scotese der, CO: Geological Society ofAmerica. 1988. A hierarchical tectonic model ofthe Gulf 30 AMERICAN MUSEUM NOVITATES NO. 3141

of Mexico and Caribbean region. Tec- Williams, E. E., and K. F. Koopman tonophysics 155: 139-168. 1952. West Indian fossil monkeys. Am. Mus. Schwan, W. Novitates 1546: 15 pp. 1980. Geodynamic peaks in alpinotype orog- Wilson, E. 0. enies and changes in ocean-floor spread- 1988. The biogeography of the West Indian ing during Late Jurassic-Late Tertiary ants (Hymenoptera: Formicidae). In J. time. Am. Assoc. Pet. Geol. Bull. 64: K. Liebherr (ed.), Zoogeography of Ca- 359-373. ribbean insects, pp. 214-230. Ithaca, Seiglie, G. A., and P. J. Bermuidez NY: Cornell Univ. 1969. Informe preliminar sobre los formini- Woods, C. A. feros del Terciario del Sur de Puerto 1982. The history and classification of South Rico. Caribbean J. Sci. 9: 67-80. American hystricognath rodents: reflec- Simpson, G. G. tions on the far away and long ago. In 1956. Zoogeography of the West Indian land M. A. Mares and H. H. Genoways (eds.), mammals. Am. Mus. Novitates 1759: Mammalian biology in South America, 28 pp. pp. 377-392. Pittsburgh: Univ. Pitts- Todd, R., and D. Low burgh. 1979. Smaller formainifera from deep wells on 1989a. The biogeography of West Indian ro- Puerto Rico and St. Croix. U.S. Geol. dents. In C. A. Woods (ed.), Biogeog- Surv. Prof. Pap. 863: 32 pp. raphy of the West Indies, pp. 741-797. Van Fossen, M., and J. E. T. Channell Gainesville, FL: Sandhill Crane Press. 1988. Paleomagnetism ofLate Cretaceous and 1989b. A new capromyid rodent from Haiti: Eocene limestones and chalks from Hai- the origin, evolution, and extinction of ti: tectonic interpretation. Tectonics 7: West Indian rodents, and their bearing 601-612. on the origin of New World hystricog- Varona, L. naths. In C. C. Black and M. R. Dawson 1974. Catalogo de los mamiferos vivientes y (eds.), Papers on fossil rodents in honor extinguidos de las Antillas. La Habana: of Albert Elmer Wood. Nat Hist. Mus. Academia de Ciencias de Cuba. Los Angeles County Sci. Ser. 33: 59-90. Webb, S. D., and Perrigo, S. 1990. The fossil and Recent land mammals of 1985. New megalonychid sloths from El Sal- the West Indies: an analysis of the ori- vador. In G. G. Montgomery (ed.), The gin, evolution, and extinction of an in- evolution and ecology of armadillos, sular fauna. In A. Azzaroli (ed.), Bio- sloths, and vermilinguas, pp. 113-120. geographical aspects of insularity, Atti Washington, D.C.: Smithsonian Insti- Convegni Lincei 85: 641-680. Rome: tution Press. Accademia Nazionale dei Lincei. White, J. L. 1992. Suborder Hystricognathi. In D. Wilson 1993. Functional and phylogenetic implica- and D. M. Reeder (eds.), Mammal spe- tions ofthe postcranial skeleton offossil cies ofthe world, a taxonomic and geo- sloths for the evolution ofarboreality in graphic reference, 2nd ed, pp. 771-806. tree sloths. Unpubl. Ph.D. diss., State Washington, D.C.: Smithsonian Insti- Univ. of New York at Stony Brook. tution Press.

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates, Bulletin, and Anthropological Papers published during the last five years are available free of charge. Address orders to: American Museum of Natural History Library, Department D, Central Park West at 79th St., New York, N.Y. 10024. TEL: (212) 769-5545. FAX: (212) 769- 5009. E-MAIL: [email protected]

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