paleogeography of the caribbean region: implications...

PALEOGEOGRAPHY OF THE

CARIBBEAN REGION:

IMPLICATIONS FOR

CENOZOIC BIOGEOGRAPHY

MANUEL A. ITURRALDE-VINENT

Research Associate, Department of Mammalogy American Museum of Natural History

Curator, Geology and Paleontology Group Museo Nacional de Historia Natural

Obispo #61, Plaza de Armas, CH-10100, Cuba

R.D.E. MA~PHEE

Chairman and Curator, Department of Mammalogy American Museum of Natural History

BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY

Number 238, 95 pages, 22 figures, 2 appendices

Issued April 28, 1999

Price: $10.60 a copy

Copyright O American Museum of Natural History 1999 ISSN 0003-0090

2

CONTENTS

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Resumen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4Resumo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Statement of Problem and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9Paleogeography of the Caribbean Region: Evidence and Analysis . . . . . . . . . . . . . . . . . . 18

Early Middle Jurassic to Late Eocene Paleogeography . . . . . . . . . . . . . . . . . . . . . . . . . . 18Latest Eocene to Middle Miocene Paleogeography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Eocene-Oligocene Transition (35–33 Ma) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Late Oligocene (27–25 Ma) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31Early Middle Miocene (16–14 Ma) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Biogeographical Hypotheses and Caribbean Paleogeography . . . . . . . . . . . . . . . . . . . . . . . 35Continent-Island Vicariance: Model of Rosen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Passive Overwater Dispersal: Model of Hedges and Co-workers . . . . . . . . . . . . . . . . . 40

Preliminary Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40Sources of Error in Estimating Times of Lineage Origins . . . . . . . . . . . . . . . . . . . . . 43Passive Transport and Cenozoic Surface-Current Patterns . . . . . . . . . . . . . . . . . . . . . . 45

Surface-Current Patterns and Flotsam Dispersal . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Surface-Current Patterns and Paleogeography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Surface-Current Patterns and Proxy Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48Other Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

GAARlandia Landspan and Island–Island Vicariance: Model of MacPheeand Iturralde-Vinent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Landspans, Vicariance, and Diversity Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59Appendix 1: Reconstructing Caribbean Paleogeography: An Analytical Guide . . . . . . . . 72

Yucatan Peninsula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72Northern Central America, Nicaragua Rise, and Western Jamaica . . . . . . . . . . . . . . . . . 73Southern Central America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75Northwestern South America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75Aruba/Tobago Belt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75Greater Antilles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Blue Mountains Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80Aves Ridge, Lesser Antilles, and Grenada Basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84Beata Ridge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87Cayman Islands and Cayman Ridge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

Appendix 2: A Plate Tectonic Model of the Caribbean from Latest Eocene to MiddleMiocene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

1999 3ITURRALDE-VINENT AND MACPHEE: CARIBBEAN PALEOGEOGRAPHY

ABSTRACT

This paper* presents a series of detailed paleo-geographical analyses of the Caribbean region,beginning with the opening of the Caribbean ba-sin in the Middle Jurassic and running to the endof the Middle Miocene. Three intervals within theCenozoic are given special treatment: Eocene–Ol-igocene transition (35–33 Ma), Late Oligocene(27–25 Ma), and early Middle Miocene (16–14Ma). While land mammals and other terrestrialvertebrates may have occupied landmasses in theCaribbean basin at any time, according to the in-terpretation presented here the existing GreaterAntillean islands, as islands, are no older thanMiddle Eocene. Earlier islands must have existed,but it is not likely that they remained as such (i.e.,as subaerial entities) due to repeated transgres-sions, subsidence, and (not incidentally) the K/Tbolide impact and associated mega-tsunamis. Ac-cordingly, we infer that the on-island lineagesforming the existing (i.e., Quaternary) Antilleanfauna must all be younger than Middle Eocene.The fossil record, although still very poor, is con-sistent with the observation that most land mam-mal lineages entered the Greater Antilles aroundthe Eocene–Oligocene transition.

Western Laurasia (North America) and westernGondwana (South America) were physically con-nected as continental areas until the mid-Jurassic,ca. 170 Ma. Terrestrial connections between thesecontinental areas since then can only have oc-curred via landbridges. In the Cretaceous, threemajor uplift events, recorded as regional uncon-formities, may have produced intercontinentallandbridges involving the Cretaceous Antillean is-land arc. The Late Campanian/Early Maastrich-tian uplift event is the one most likely to haveresulted in a landbridge, as it would have beencoeval with uplift of the dying Cretaceous arc.However, evidence is too limited for any certaintyon this point. The existing landbridge (Panaman-ian isthmus) was completed in the Pliocene; evi-dence for a precursor bridge late in the MiddleMiocene is ambiguous.

We marshal extensive geological evidence toshow that, during the Eocene–Oligocene transi-tion, the developing northern Greater Antilles andnorthwestern South America were briefly con-

* Contribution 2 to the series ‘‘Origin of the GreaterAntillean Land Mammal Fauna.’’

nected by a ‘‘landspan’’ (i.e., a subaerial connec-tion between a continent and one or more off-shelf islands) centered on the emergent AvesRidge. This structure (Greater Antilles 1 AvesRidge) is dubbed GAARlandia. The massive up-lift event that apparently permitted these connec-tions was spent by 32 Ma; a general subsidencefollowed, ending the GAARlandia landspanphase. Thereafter, Caribbean neotectonism result-ed in the subdivision of existing land areas.

The GAARlandia hypothesis has great signifi-cance for understanding the history of the Antil-lean biota. Typically, the historical biogeographyof the Greater Antilles is discussed in terms ofwhether the fauna was largely shaped by strictdispersal or strict continent–island vicariance. TheGAARlandia hypothesis involves elements ofboth. Continent–island vicariance sensu Rosen ap-pears to be excludable for any time period sincethe mid-Jurassic. Even if vicariance occurred atthat time, its relevance for understanding the ori-gin of the modern Antillean biota is minimal.Hedges and co-workers have strongly espousedover-water dispersal as the major and perhapsonly method of vertebrate faunal formation in theCaribbean region. However, surface-current dis-persal of propagules is inadequate as an expla-nation of observed distribution patterns of terres-trial faunas in the Greater Antilles. Even thoughthere is a general tendency for Caribbean surfacecurrents to flow northward with respect to theSouth American coastline, experimental evidenceindicates that the final depositional sites of pas-sively floating objects is highly unpredictable.Crucially, prior to the Pliocene, regional pale-oceanography was such that current-flow patternsfrom major rivers would have delivered SouthAmerican waifs to the Central American coast,not to the Greater or Lesser Antilles. Since at leastthree (capromyid rodents, pitheciine primates, andmegalonychid sloths) and possibly four (neso-phontid insectivores) lineages of Antillean mam-mals were already on one or more of the GreaterAntilles by the Early Miocene, Hedges’ inferenceas to the primacy of over-water dispersal appearsto be at odds with the facts. By contrast, the land-span model is consistent with most aspects of An-tillean land-mammal biogeography as currentlyknown; whether it is consistent with the bioge-ography of other groups remains to be seen.

4 NO. 238BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY

RESUMEN

El proposito de este trabajo es presentar unaserie de analisis paleogeograficos detallados de laregion del Caribe, comenzando con la apertura dela cuenca del Caribe en el Jurasico Medio y ex-tendiendose hasta el Mioceno Medio. Tres inter-valos del Cenozoico reciben un tratamiento es-pecial: la transicion Eoceno-Oligoceno (35–33Ma), el Oligoceno Tardıo (27–25 Ma), y el Mio-ceno Medio temprano (16–14 Ma). Aunque losmamıferos terrestres pudieron haber ocupado ma-sas de tierra en la cuenca del Caribe en cualquiermomento de su historia, de acuerdo con la inter-pretacion que se presenta en este trabajo, las ac-tuales Antillas Mayores, como islas, son no masantiguas que Eoceno Medio. Islas mas antiguasdeben haber existido, pero no es probable que el-las hayan permanecido como tales (es decir, comoentidades subaereas) debido a las repetidas trans-gresiones, subsidencia, y (no incidentalmente) alimpacto del lımite K/T y el megatsunami asociadoal mismo. De acuerdo con esto nosotros inferimosque los linajes insulares que forman la fauna an-tillana actual (y cuaternaria en general) deben sermas jovenes que el Eoceno Medio. El registro fos-il, a pesar de ser muy pobre aun, es consistentecon la observacion de que la mayorıa de los li-najes de mamıferos llegaron a las Antillas Ma-yores alrededor del lımite Eoceno–Oligoceno.

El oeste de Laurasia (America del Norte) y elde Gondwana (America del Sur) estuvieron fısi-camente conectados como areas continentaleshasta el intervalo Bajociano al Oxfordiano (178–160 Ma) cuando se comenzo a formar la cuencaoceanica del Caribe. Coneccciones terrestres entredichas areas continentales a partir de entoncessolo pudieron ocurrir mediante puentes naturalesde terreno. En el Cretacico tres eventos principa-les de levantamiento, coincidentes con inconfor-midades regionales, pudieran haber producidopuentes intercontinentales que involucraron alarco de islas volcanicas de las Antillas. El lev-antamiento ocurrido en el Campaniano tardıo aMaastrichtiano temprano es el que tiene las ma-yores posibilidades de haber producido un puentenatural, ya que este coincidio en el tiempo con laextincion del vulcanismo cretacico. No obstante,la evidencia es muy limitada para tener algunaseguridad sobre este asunto. El puente natural queexiste actualmente (itsmo de Panama) se completoen el Plioceno; pero la evidencia para un puenteanterior en el Mioceno Medio tardıo es ambigua.

Aquı se presenta extensa evidencia geologicapara mostrar que, durante la transicion entre elEoceno y el Oligoceno, las tierras antillanas y laporcion noroccidental de America del Sur estu-

vieron brevemenmte conectadas por una ‘‘land-span’’ (proyeccion de terreno) (es decir, por unaconeccion subaerea entre un continente y una omas islas situadas fuera del lımite de la plataformacontinental), coneccion que estuvo centrada en laentonces emergida Cresta de Aves. Esta estructura(Crestas de las Antillas Mayores y de Aves) sedenomino GAARlandia. El evento de levanta-miento masivo que aparentemente permitio estaconeccion termino hace unos 32 millones de anos;debido a una subsidencia general que termino conla fase de ‘‘landspan’’ de GAARlandia. Posterior-mente la etapa neotectonica caribena resulto en lasubdivision de las tierras existentes.

La hipotesis GAARlandia tiene un gran signi-ficado para comprender la historia de la biota An-tillana. Tıpicamente, la biogeografıa historica delas Antillas Mayores se discute en terminos de sila fauna fue principalmente formada por disper-sion estricta o por estricta vicariancia continente–isla. La hipotesis GAARlandia comprende ele-mentos de ambas. Pero la vicariancia continente–isla al estilo de Rosen puede ser excluida paracualquier momento desde el Jurasico Medio. In-cluso si la vicariancia hubiese ocurrido en aquellaepoca, su relevancia para comprender el origen dela biota antillana moderna es mınima. Hedges ysus colaboradores han propuesto con enfasis ladispersion por agua como el principal, sino el un-ico, metodo de formacion de la fauna de verte-brados en la region del Caribe. Sin embargo, ladispersion de propagulos mediante las corrientesmarinas superficiales es inefectiva para explicarlos patrones de distribucion actual de la fauna ter-restre en las Antillas Mayores. Incluso aunque ex-iste una tendencia general de las corrientes super-ficiales del Caribe a fluir hacia el norte con res-pecto a la costa sudamericana, las evidencias ex-perimentales indican que es practicamenteimpredecible donde seran finalmente depositadoslos objetos flotantes acarreados por dichas cor-rientes. Al respecto, es crucial el hecho de que,antes del Plioceno, la paleoceanografıa regionalfue tal que los patrones de corrientes de los rıossudamericanos debieron acarrear los objetos a laderiva provenientes de America del sur hacia lascostas de America Central, o hacia el Pacıfico, nohacia las Antillas Mayores. Dado que al menostres linajes (roedores capromidos, primates pite-cinos y perezosos megalonıchidos) y posible-mente cuatro (insectıvoros nesofontidos) de losmamıferos antillanos se encontraban ya en las An-tillas Mayores a comienzos del Mioceno, las in-ferencias de Hedges respecto al dominio de la


dispersion por agua como el modo de migracionde esta fauna esta en desacuerdo con los hechos.En contraste, el modelo de ‘‘landspan’’ es consis-tente con muchos de los aspectos de la biogeo-

grafıa historica de los mamıferos terrestres antil-lanos tal como se conoce hoy dıa; aunque si tam-bien es consistente con la biogeografıa de otrosgrupos es algo que esta aun por definir.

RESUMO

Este trabalho apresenta uma serie de analisespaleogeograficas detalhadas da regiao do Caribe,comecando com a abertura da bacia do Caribe noJurassico Medio e seguindo ate o fim do MiocenoMedio. Tres intervalos do Cenozoico receberamespecial atencao: a transicao Eoceno-Oligoceno(35–33 Ma), o final do Oligoceno (27–25 Ma) eo inıcio do Mioceno Medio (16–14 Ma). Aindaque mamıferos e outros vertebrados terrestres pos-sam ter ocupado massas de terra na bacia do Ca-ribe em qualquer momento de sua historia, segun-do o presente estudo, a existencia das GrandesAntilhas, como ilhas, nao e mais antiga do que oMioceno Medio. Ilhas mais antigas podem ter ex-istido, no entanto, e pouco provavel que tenhampermanecido como tais por longos perıodos. As-sim sendo, inferimos que todas as linhagens queformam a fauna antilhana atual (ou seja, Quater-naria) devam ser mais recentes que o Eoceno Me-dio. O registro fossil, apesar de ser bastante pobre,e consistente com a observacao de que a maioriadas linhagens de mamıferos terrestres chegaramas Grandes Antilhas por volta da transicaoEoceno-Oligoceno.

O oeste da Laurasia (America do Norte) e o daGondwana (America do Sul) estiveram fisica-mente conectados como areas continentais ate oJurassico Medio. Tres principais eventos de soer-guimento no Cretaceo, registrados por inconfor-midades regionais, podem ter produzido pontesintercontinentais envolvendo o arco de ilhas an-tilhanas do Cretaceo. O soerguimento ocorrido noCampaniano superior/Maastrichtiano inferior pa-rece ser o mais relacionado com a formacao deuma ponte de conexao, uma vez que este coincidecom o soerguimento ocorrido durante o desapa-recimento do arco do Cretaceo. A conexao atual(istmo do Panama) completou-se no Plioceno;evidencias de uma ponte anterior no fim do Mio-ceno Medio sao ambıguas.

Nos buscamos extensivas evidencias geologicaspara demonstrar que durante a transicao Eoceno-Oligoceno, a parte norte das Grandes Antilhas(entao em desenvolvimento) e o noroeste daAmerica do Sul estiveram brevemente conectadospor uma ‘‘landspan’’ (i.e., uma conexao sub-aerea

entre um continente e uma ou mais ilhas oceani-cas). Esta ‘‘landspan’’ estaria centrada na, entaoemergente, Cadeia de Aves. Denominou-se estaestrutura (Cadeia das Grandes Antilhas 1 Cadeiade Aves) como GAARlandia. O soerguimentomassivo que aparentemente permitiu estas conex-oes ocorreu a cerca de 32 milhoes de anos, sendoseguido por uma subsidencia geral que terminoucom a fase de ‘‘landspan’’ da GAARlandia. Pos-teriormente, o neotectonismo caribenho resultouna subdivisao das terras existentes.

Tradicionalmente, a historia biogeografica dasGrandes Antilhas e discutida em termos de for-macao da fauna estritamente por dispersao ou porvicariancia continente-ilha. A hipotese da GAAR-landia envolve elementos de ambas as correntes,ainda que os modelos de vicariancia continente-ilha sensu Rosen possam ser excluıdos para qu-alquer perıodo desde o Jurassico Medio. Hedgese colaboradores tem veementemente sugerido adispersao atraves d’agua como o principal, senaoo unico, meio pelo qual teria ocorrido a formacaoda fauna de vertebrados do Caribe. Entretanto, aproposta de dispersao de propagulos por correntesmarinhas superficiais e inadequada para explicaros padroes de distribuicao de fauna terrestre ob-servados nas Grandes Antilhas. Enfatiza-se que,antes do Plioceno, a paleoceanografia da regiaoera tal que o padrao de fluxo de correntes dosprincipais rios sulamericanos deviam carrear ob-jetos para a costa da America Central e nao paraas Grandes e Pequenas Antilhas. No mınimo tres(roedores capromiıdeos, primatas piteciıneos epreguicas megaloniquideas) e talvez quatro (in-setıvoros nesofondideos) linhagens de mamıferosantilhanos ja ocorriam em uma ou mais ilhas dasGrandes Antilhas no inıcio do Mioceno, indican-do que as propostas de Hedge quanto a primaziada dispersao aquatica da fauna nao estao de acor-do com os fatos. Neste sentido, o modelo de‘‘landspan’’ e consistente com a maioria das pro-postas atualmente aceitas para a biogeografia dosmamıferos terrestres antilhanos. Permanece emaberto se este modelo esta de acordo com as pro-postas biogeograficas para outros grupos.


It is this independence of biological from geological data that makes thecomparison of the two so interesting because it is hard to imagine howcongruence between the two could be the result of anything but a causalhistory in which geology acts as the independent variable providing op-portunities for change in the dependent biological world.

— Donn E. Rosen (1985: 637)

INTRODUCTION

During the past century, a number of hy-potheses have been offered as partial or com-plete explanations for the origins of Antilleanterrestrial vertebrate faunas.1 Three mecha-nisms have been discussed extensively in theliterature: (1) dispersal over water barriers(e.g., Matthew, 1918; Darlington, 1938;Woods, 1989; Hedges et al., 1992, 1994;Hedges, 1996a, 1996b); (2) dispersal overshort-lived landbridges and landspans2 (e.g.,Fernandez de Castro, 1884; De La Torre,1910; Gayet et al., 1992; MacPhee and Itur-ralde-Vinent, 1994, 1995); and (3) vicari-ance, i.e., splitting or division of a biota ortaxon through the development of a naturalbarrier (e.g., Rosen, 1975, 1985; Guyer andSavage, 1987; MacPhee and Wyss, 1990).Although these mechanisms are sometimespresented as though they were discrete, mu-tually exclusive alternatives, depending onthe time, place, and taxon under discussion,any or all of them may have been involvedin Antillean faunal formation. In fact, to pre-view the chief conclusion of this paper, it

1 There are several current biogeographical definitionsof the ‘‘Antilles,’’ ‘‘West Indies,’’ ‘‘Caribbean Islands,’’‘‘insular Neotropics’’ and their various subdivisions. Inthis paper, Greater and Lesser Antilles will have theirusual meanings; ‘‘Antillean’’ as an adjective refers toanything having to do with these islands, and is used inpreference to ‘‘West Indian’’ (which, as generally used,covers other, non-Antillean islands such as Bahamas).The Caribbean region, which we newly define as a pa-leogeographical concept, consists (at any stage of its de-velopment) of the Caribbean Sea and all of its contents,plus the facing continental margins of North, South, andCentral America. Thus the Caribbean region is largerthan the Caribbean Plate, although the structures on thatplate comprise most of the entities of interest here. It isalso larger than ‘‘West Indies’’ as defined by Hedges(1996a, 1996b).

2 For definition of ‘‘landspan’’ see section entitledGAARlandia Landspan and Island–Island Vicariance.

seems inescapable that all three were in-volved in the formation of the Antillean land-mammal fauna, although not necessarily ineither the manner or the degree envisaged byother authors.

Our present purpose is to present a freshperspective on the ‘‘geography’’ part of bio-geography, as it relates to the Caribbean re-gion, and to examine how this may offernovel insights into historical processes offaunal formation. (For a listing of most ofthe features and localities mentioned in thetext, see figure 1). Although our specific con-cept of Antillean paleogeographical historydiffers in various ways from those of otherauthors (see Biogeographical Hypothesesand Caribbean Paleogeography), we havemade a particular effort to document andevaluate other views.

It is widely recognized that hypothesesconcerning Antillean historical biogeographyare critically dependent on specific recon-structions of regional paleogeography, paleo-ceanography, tectonics, and other bodies ofdata. However, most discussions of this sub-ject by life scientists have tended to empha-size biological evidence over geological ev-idence. This bias should not be viewed asbeing merely reflective of biologists’ under-standable preference for their own kinds ofdata, because several issues are involved.

First, although much of the general geo-logical and tectonic literature is significantfor understanding the biogeographical histo-ry of the Caribbean region, virtually none ofit was written with the needs of biologists inmind. Accordingly, biologists hoping to in-tegrate geological information into theirwork are faced with the daunting tasks ofhaving to compile evidence from many dif-ferent sources, judge as best they can the ac-curacy of age assignments and other primary


Fig

.1.

Ref

eren

cem

apof

Car

ibbe

anre

gion

illu

stra

ting

mos

tof

the

geog

raph

ical

and

geol

ogic

alfe

atur

esre

ferr

edto

inte

xt.


data, and, frankly, recognize when geologicaltheory outstrips fact. Lack of familiarity withthe subject matter and the methods of geo-logical argumentation may lead to simplefactual errors, or, possibly worse, encouragethe uncritical acceptance of insufficientlytested geological scenarios primarily becausethey appear to support certain biological hy-potheses. It is partly for this reason that someauthors validly question whether biologistsshould place themselves in the very vulner-able position of relying on the revealed truthsof geologists to explain biogeographical pat-ternings (Henderson, 1991: 61; see also Crawand Weston, 1984). Knowledge, we suggest,is the best antidote to vulnerability.

Second, despite their apparent elegance,plate tectonic models (e.g., Malfait and Din-kelmann, 1972; Duncan and Hargraves,1984; Leclere and Stephan, 1985; Ross andScotese, 1988; Donnelly, 1989a; Pindell andBarrett, 1990; Mann et al., 1995; Hay andWold, 1996; Iturralde-Vinent, 1996a, 1997b)vary widely in their comprehensiveness andtestability (Rull and Schubert, 1989; Perfitand Williams, 1989). For example, agree-ment is still lacking regarding the numberand fit of plates and microplates in the Ca-ribbean Region—a basic issue of fact (cf.Donnelly, 1985; Ross and Scotese, 1988;Pindell, 1994; Hay and Wold, 1996). Fur-thermore, plate tectonic models do not nec-essarily provide the kinds of information thatbiologists are most interested in. Typically,such models focus on reconstructing histor-ical positions of specific geologic units com-monly denoted as plates, terranes, blocks,volcanic arcs, and ridges (e.g., Malfait andDinkelmann, 1972; Duncan and Hargrave,1984; Leclere and Stephan, 1985; Donnelly,1985; Ross and Scotese, 1988; Pindell andBarrett, 1990; Pindell, 1994; Mann et al.,1995; Hay and Wold, 1996). They are not atall, or are only incidentally, concerned withcreating well-constrained paleogeographicalmaps that portray the physical geography ofsuch units (or parts thereof) through time.(For further discussion of these and otherconcepts, see Paleogeography of the Carib-bean Region: Evidence and Analysis.) Withthe purely tectonic literature as the soleguide, one cannot derive any consistent pic-ture concerning how many times subaerial

land masses existed in the Caribbean, orwhen or how many times these land masseswere connected to nearby continents, or thenature of the relief they exhibited. Thus,Hedges et al. (1992) explored structural re-lationships among Caribbean land massesand nearby continents on the basis of Pindelland Barrett’s (1990) tectonic reconstructionwhich, in fact, contains no information onsuch relationships. (The latter authors discussonly the position of geological units, whichis not paleogeography as we define it.)

In this paper we offer the first comprehen-sive paleogeographical and paleoceanograph-ical reconstructions of the Caribbean basin,from latest Eocene to Middle Miocene, aninterval selected for reasons explained in de-tail in succeeding sections. We also brieflyreview paleogeographical scenarios for Ju-rassic through Late Eocene time, for the pur-pose of evaluating evidence for early landconnections and island permanency. Reflect-ing our own interests, we concentrate on thepaleogeography of the central part of the Ca-ribbean basin and portions of Central Amer-ica and northwestern South America. In themain, the biogeographical implications thatwe have pursued in this investigation arethose most closely tied in with geography.Understanding the phylogeny of the Antil-lean biota is equally important and interest-ing; however, this topic is reserved for a sub-sequent paper in this series (MacPhee andIturralde-Vinent, in prep.).

Because of the large quantity of ancillarydocumentation required to support a study ofthis sort, for efficiency in presentation muchof the basic geological, paleogeographical,and paleontological information is presentedin the form of appendices, tables, and figures.The text summarizes this information in dis-cursive form, but its main function is to dis-cuss problems of interpretation and expla-nation. Readers wishing to utilize the chiefresults of our investigations may profitablyconsult the main text, but those requiring agreater level of detail should refer to the ap-pendices throughout.

ACKNOWLEDGMENTS

The manuscript of the present paper wasawarded the 1997 Premio Anual de la Cien-


cia by the Academia de Ciencias de Cuba foroutstanding scientific investigations. We aregrateful to the Academy, and thus the peopleof Cuba, for this honor.

We extend special thanks to Gregory Mayer(University of Wisconsin–Parkside), PeterMattson (Queens College, City University ofNew York), Gary Morgan (New Mexico Mu-seum of Natural History), and Marcelo San-chez Villagra (Universitat Tubingen) for re-viewing the manuscript, to Lisa Gahagan (In-stitute of Geophysics, University of Texas atAustin) for programming assistance in devel-oping the tectonic model presented in appendix2, and to Clare Flemming (AMNH) for helpwith editing the text. We also acknowledge ourgreat debt to Clare Flemming and Ines Horo-vitz (AMNH) and to Stephen Dıaz Franco andReinaldo Rojas Consuegra (MNHNH) for their

many contributions to our field program. Re-search related to this paper was partly sup-ported by a Kalbfleisch postdoctoral fellowshipat the AMNH (to MIV) and NSF 902002 (toRDEM). Figures were drafted by MIV, withthe exception of figures 10 and 11 (by PatriciaJ. Wynne) and figures 9 and 13 (by RDEM).Fernando Sicuro kindly supplied the Portu-guese version of the abstract.

ABBREVIATIONS

AMNH American Museum of Natural HistoryID immunological distanceMa millions of years (ago)MNHNH Museo Nacional de Historia Natural,

La HabanaNWSA Northwestern South America (micro-

continent or microplate)Fm formation

STATEMENT OF PROBLEM AND METHODS

We agree with Hedges (1996b: 166) thatthe aspect of Caribbean geological history ofgreatest interest to biogeographers, the rela-tionships of emergent land areas, is unfortu-nately the one that is most poorly under-stood. This point is best explored by notingsome illustrative examples:Not all geological maps contain recoverablepaleogeographical information.

Eva and MacFarlane’s (1985: figs. 4–12)study of carbonate development in Jamaicaincludes a series of nine illustrations that de-pict the development of various features(e.g., subaerial land surfaces, subaerial andmarine volcanoes, shallow and deep sea)from Paleocene to Pliocene times. However,they are not paleogeographical maps in anyliteral sense, because all features are shownas evolving within the present-day perimeterof the island, with no attempt to restore fold-ed and faulted rock units to their original rel-ative geographical positions. Because someof Jamaica’s basement rocks were stronglyfolded and faulted during the Middle Eoceneand late Neogene (Lewis et al., 1990; Rob-inson, 1994), any effort to capture paleogeo-graphical reality would require the use ofpalinspastic methods to reconstruct displace-ments of blocks along faults and restore orig-inal surface areas of deformed formations.

Eva and MacFarlane’s (1985) investigationdid not require this kind of reconstruction,and they simply portrayed the different geo-logic units forming the Jamaican basement(ophiolites, metamorphic rocks, Cretaceous/Paleogene island arc suites, and late Cam-panian to Holocene sedimentary formations)as though they have had the same relativepositions and areal dispositions since the Pa-leocene. The message is that nonpalinspasticreconstructions—by far the most common‘‘paleogeographical’’ representations in thegeological literature (e.g., Khudoley andMeyerhoff, 1971; Maurrasse, 1982; Salvador,1987; Smith et al., 1994)—may be of negli-gible value for biogeographical investigationsbecause they are not intended to be paleogeo-graphically accurate. A simplified example ofpalinspastic reconstruction for the Greater An-tilles is illustrated in figures 2 and 3.Paleogeographical information is difficult totreat in a uniform way.

Just as there is no such thing as a perfectbiological classification, there is no suchthing as a perfect map. Often, maps are de-signed to depict only one body of data ac-curately; thus, even if they are intentionallypaleogeographical in nature, they may notshow different categories of information withequal degrees of precision. For example, as


Fig. 2. Simplified palinspastic reconstruction of Greater Antilles Foldbelt along a cross section pass-ing through eastern Cuba and western Hispaniola, to illustrate methodological points. Note that thefoldbelt consists of a series of tectonically superimposed units that have been foreshortened by defor-mation. As a result, the current width of the set of geological units transected by the section amountsto only a fraction of the units’ original width (for additional explanation, see text). See figure 3 forlocation of cross section.


Fig

.3.

Pal

insp

asti

cre

cons

truc

tion

ofea

ster

npa

rtof

Gre

ater

Ant

ille

sF

oldb

elt,

for

peri

odco

rres

pond

ing

toL

ate

Eoc

ene

thro

ugh

mid

-Oli

goce

ne(c

a.37

–30

Ma)

.T

his

fram

ewor

kha

sbe

engr

eatl

ydi

srup

ted

sinc

eth

em

id-O

ligo

cene

bysi

nist

ral

stri

ke-s

lip

mov

emen

tsan

dde

form

atio

nsal

ong

the

nort

hern

mar

gin

ofth

eC

arib

bean

plat

e,ea

stof

Hol

guın

.A

and

A9

are

refe

renc

epo

ints

for

cros

sse

ctio

nde

pict

edin

figu

re2.

(Inf

orm

atio

nco

mpi

led

from

man

yso

urce

s;se

eap

pend

ix1.

)


its title implies, the Atlas of Mesozoic andCenozoic Coastlines (Smith et al., 1994) isconcerned with depicting coastline informa-tion, on a global scale, for the past 245 Ma.Because its intended scope is the entire plan-et, regional details are often lost or portrayedinaccurately. As the authors point out, smalldiscrepancies do not affect the big picture,but they do complicate the use of the mapsfor other purposes. Of interest here is the factthat the Greater Antilles are depicted in theircurrent sizes and positions relative to NorthAmerica during all relevant time periods. Be-cause the dimensions of the Caribbean Seahave changed over time, the Greater Antillesare forced by the mapping program into po-sitions they could never have occupied.Thus, these islands overlap northern SouthAmerica at 170 Ma, project east of Trinidadinto the South Atlantic (!) at 155 Ma, andfinally end up at their present-day position at80–0 Ma. Furthermore, subaerial exposureson these islands are depicted on Oligoceneand later maps, but not on earlier ones. Thissuggests that land areas did not exist as suchprior to the mid-Cenozoic, which is not ac-curate. Similarly, the Chortis Block is de-picted as overlapping southern CentralAmerica at 170 Ma, then acting as a bridgebetween southern Central America and NorthAmerica at 155 Ma, then coming into contactwith the Maya (Yucatan) Block as both wereuplifted around 148 Ma, and finally achiev-ing its present-day position with respect toNorth America by 80–0 Ma. The YucatanPeninsula is successively depicted as (1)overlapping South America at 170 Ma, (2)uplifted and in contact with northern SouthAmerica by 155 Ma, and (3) situated near itspresent-day position with respect to NorthAmerica in the Tithonian (148 Ma). Al-though these implied motions can be dis-missed as minor artifacts of mapping pro-grams designed to show large segments ofthe geode, the point is that they seriouslyconflict with all available models of Carib-bean plate tectonics (Malfait and Dinkel-mann, 1972; Donnelly 1985; Pindell andBarrett, 1990; Pindell, 1994; Mann et al.,1995). Fine-scale interpretation of the paleo-geography of small areas is untenable withmaps of this sort, although they have been

used for that purpose (e.g., Hedges, 1996a,1996b).

Island geology and island paleogeographyare not isomorphous.

Conventionally, a geographical island isdefined by its shoreline (i.e., its subaerial pe-rimeter), although other criteria are of coursepossible (e.g., 2100 m isobath). Shorelinescan be affected by rise or fall in sea level,deposition or erosion, and uplift or subsi-dence of the geological unit that constitutesthe islands basement. Shorelines are there-fore exceptionally dynamic at virtually all di-mensional and temporal scales. Geologicalunits, by contrast, are only indirectly affectedby surficial processes such as prograding ordegrading shorelines; they reflect a deeperstructure, and deep structure is rarely coter-minous with conventional geography. Theusual objective of geological study is to in-terpret the history of tectonic elements interms of their formation, evolution and sub-sequent transformation into other elements,utilizing the imprints that such processesleave in the rocks themselves. Shoreline re-construction is therefore a difficult task, be-cause the evidence needed to make paleo-geographical reconstructions is almost inev-itably destroyed or modified substantiallyover geologically long periods of time.

Smith et al. (1994) pointed out that coast-line reconstruction is additionally complica-ted by the fact that different datasets, osten-sibly for the same interval, may yield quitedifferent paleogeographical results. This canhappen when stratigraphic data are collectedfor specific purposes, such as documenting ageneral transgression or sealevel drop. Forexample, reconstructions of Maastrichtiancoastlines for the same area may look quitedifferent from one another, depending onwhat part of the interval and which eventsindividual authors intended to depict.

Another point about island geology versusisland paleogeography can be made by ref-erence to Hedges’ (1996b) claim that someterrestrial environments in the Caribbean Seamay have been in existence since the end ofthe Cretaceous. Hedges (1996b: 166) statedthat there is ‘‘no place in the West Indies thatis known by the presence of a continuous


sequence of sediments to have been emer-gent since the late Cretaceous, although someareas of Cuba, northern Hispaniola, and pos-sibly Puerto Rico, may have been.’’3 It is un-clear what Hedges meant to convey by thephrase ‘‘continuous sequence of sediments,’’since the only conceivable contexts in whichunbroken sedimentary accumulations mighthave occurred throughout the last 65 Ma aredeep oceanic basins situated on comparative-ly ancient sea floor (e.g., eastern Pacific).Furthermore, his paleogeographical obser-vation has significance only if it is addition-ally inferred that the hypothesized emergentlands of the late Mesozoic and early Ceno-zoic were incorporated while still subaerialinto the developing Greater Antilles. Weknow of no geological evidence that supportsthis inference; the little evidence that doesexist indicates that no terrestrial contextsfrom these earlier periods survived as suchinto the Late Eocene (see Paleogeography ofthe Caribbean Region: Evidence and Analy-sis).

Tectonic modeling and paleogeographicalreconstruction are not the same.

This point can be conveniently illustratedby reference to Pindell’s (1994) frequentlycited work. As part of a general tectonic re-construction of the Caribbean region, Pindell(1994: fig. 2.6a–n; see also Pindell and Bar-rett, 1990) attempted to depict the paleogeo-graphical history of certain physical features(subaerial land, deep and shallow water, vol-canic arcs). Although Pindell’s model utilizesan extensive database, it contains a few in-accuracies and unverifiable conjectures thathave both tectonic and paleogeographical im-plications. For example, Florida and Baha-mas are shown as comprising a single car-bonate platform from late Jurassic throughLate Miocene, which contradicts recent evi-dence for their long-term separation (Austinet al., 1988; Droxler et al., 1989; Hine, 1997;Denny et al., 1994; Iturralde-Vinent et al.,1996a). The Nicaragua Rise is drawn as a

3 ‘‘West Indies’’ is defined by Hedges (1996b) to in-clude the Greater and Lesser Antilles, Bahamas, and anumber of small islands that lie immediately off the con-tinental shelves of Central and South America. It doesnot include the southern Netherlands Antilles or Trinidadand Tobago (cf. footnote 1).

shallow-water promontory of the ChortisBlock from Middle Eocene to Late Miocenetime, contra Sigurdsson et al. (1997) who de-pict it more correctly as a series of isolatedcarbonate banks. Cuba is shown as upliftedat 59, 21, and 10 Ma, but as completely sub-merged at 49 and 35 Ma, in conflict with thedata analyzed by Iturralde-Vinent (1969,1972, 1988a).

Pindell’s (1994) maps are particularlyproblematic in their depiction of island arcconnections. North and South America areshown as being completely linked by islandarcs in the Valanginian, late Albian, Cam-panian, Maastrichtian, and Middle Eocene,and as nearly linked in the Barremian, Tu-ronian, and Paleocene. Although these vol-canic arcs certainly existed, reconstructingthem accurately as paleogeographical entitiesrequires detailed scrutiny of relevant evi-dence. It is obvious from present-day geog-raphy that volcanic arcs may form continu-ous subaerial entities (e.g., Kamchatka Pen-insula, presently sutured to Chukotka; isth-mus region of Central America) as well asisland chains (e.g., Lesser Antilles, Kuriles).The life-span of an island qua island cannotbe predicted from first principles: islands atthe position of Krakatau and Aldabra, to citetwo quite different examples, have appearedand disappeared more than once in the lateQuaternary (Stoddart et al., 1971; Nunn,1994). Without careful appraisal, contradic-tory conclusions may be reached on the basisof the same evidence. For example, Gayet etal. (1992: fig. 1) argued on the basis of Pin-dell’s model (see Pindell and Barrett, 1990)that ‘‘the terrestrial bridge that linked Northand South America by latest Cretaceous andPaleocene times probably comprised theGreater Antilles and the Aves Ridge whichconsisted of a magmatic [arc] submitted touplift and deformation. . . .’’ By contrast,Hedges (1996a), also citing Pindell and Bar-rett (1990), claimed that any possible con-nection between North and South Americavia the developing ‘‘proto-Antilles’’ (or‘‘proto-Greater Antilles’’) was sundered inthe Late Cretaceous (70–80 Ma). However,in actuality Pindell and Barrett (1990; seealso Pindell, 1994) took no position on theexistence of land connections, as this issuewas ancillary to the topics they were consid-


ering. In any case, if numerous long-termconnections between North and South Amer-ica had indeed existed (as might be inferredfrom a literal reading of Pindell’s [1994]maps), evidence of this fact would surelyhave been found in the vertebrate fossil rec-ord; but it has not (Gingerich, 1985; Alva-rado, 1988; Webb, 1985; see also Biogeo-graphical Hypotheses and Caribbean Paleo-geography).

Similar difficulties attend the use of Perfitand Williams’ (1989) tectonic reconstruc-tions as paleogeographical evidence. In thisuseful and incisive paper—in several waysthe intellectual precursor of the presentmonograph—maps were intended to supporta critical discussion of Antillean biogeogra-phy and paleogeography. However, their re-constructions actually present little in theway of physical geography, as land and seaare the only items discriminated. More prob-lematically, their maps (but not their text)give the impression that most of the GreaterAntillean islands originated as such in thelate Cretaceous, and merely grew larger asthey were tectonically transported to theircurrent relative plate position. In reality, theGreater Antilles in their current guises arerelatively young geographical features (seeappendix 1; Iturralde-Vinent, 1978, 1982,1988a, 1994a).

From this brief review it is evident that therecent Caribbean geological literature is not(and was never intended to be) a source ofready-made, easily interpreted paleogeo-graphical maps—that is, maps specificallydesigned to trace the physical and positionalhistory of particular geographical entities.Appropriate design features for such mapsvary with the nature of the entities beingtraced. In the case of terrestrial environ-ments, which constitute our special concern,reliable paleogeographical maps of the Ca-ribbean region would help to both constrainand enrich discussion of a range of signifi-cant inquiries, among which are: (1) Whendid specific terrestrial environments(‘‘lands’’) exist in the Caribbean region, andfor how long? (2) Where were these landsgeodesically located, at any given time pe-riod? (3) What was the nature of physicalconnections between and among differentlands? (4) How long did such connections

last? (5) What were the surface areas of in-dividual lands, and how did their sizes alterover time? As has already been made clear,such questions cannot be answered merelyby examining geological maps or tectonic re-constructions.

Our paleogeographical maps provide in-formation on four contexts: high-elevationand low-elevation terrestrial environments(hereafter, ‘‘highlands’’ and ‘‘lowlands’’),and shallow-water and deep-water marineenvironments (‘‘shallow marine’’ and ‘‘deepmarine’’). These environments are distin-guished by certain diagnostic features,among which positive or negative elevationrelative to ambient sea level is the most sig-nificant (see also appendix 1). Although pre-cise measures of elevation are not possible,plausible benchmark estimates can be madewithin 61 order of magnitude. We definehighlands as environments that existed atpositive elevations greater than ca. 200 m;lowlands were less than 200 m. Althoughwith good faunal evidence it is possible todistinguish marine environments very finely,we reconstruct only two—shallow marine,covering shelf conditions to 2100 m; anddeep marine, embracing all sea-floor settingsdeeper than 2100 m.

We emphasize that the contact line be-tween terrestrial and marine environments onany given map should be thought of as a me-dian value for coastline position during theinterval being depicted. The accuracy of anypaleocoastline delimitation is, in any case, afunction of sedimentary exposure: in general,paleocoastline positions can be traced moreaccurately within uplifted areas than in onesthat are currently below sea level.

Three quasi-independent parameters wereused in map construction: (1) geological con-stitution, (2) geographical positioning, and(3) physical paleogeography. Geologicalconstitution is the sum of those attributes ofa particular geological unit that are definedby its composition, boundaries, and positionwith respect to other such units in the Carib-bean area or elsewhere. A geological unit canbe thought of as a time-bounded suite ofrocks that were formed by a particular set ofgeodynamic processes operating at a desig-nated location in the lithosphere. Typical ex-amples of such units might include a specific


section of oceanic or continental crust, a vol-canic arc, a set of genetically related sedi-mentary basins, a foldbelt, or a block-terrane.Each geological unit has its own ontogeny,spanning its origin, evolution, and possibletransformation into other units. Units arenamed for convenience and book-keepingpurposes, but the reader should be aware thatnames do not necessarily imply identity withgeographical elements. For example, the Ca-ribbean Mountains of northern South Amer-ica, which now form an area of high relief,contain fragments of oceanic crust and islandarc. Both Beata and Aves are identified as‘‘ridges,’’ but geologically they are quite dif-ferent. The Beata Ridge is a thick oceaniccrustal unit, while the Aves Ridge is part ofan extinct volcanic arc (Holcombe et al.,1990). In the Caribbean region, many largerfeatures, such as terranes, blocks, ridges, andarcs, are minutely subdivided by faults. Faultboundaries permit the delimitation of smallerentities that can be considered to have hadsemi-independent histories since faulting oc-curred (appendix 1, fig. 1).

Geological units can be superimposed in asingle stack (i.e., in the same crustal posi-tion), as the following simple example illus-trates. A section of early Mesozoic oceaniccrust might evolve into a Cretaceous volca-nic arc; later, both of these units might bedeformed into a series of mountains and ba-sins. The fate of each unit implies a differentpaleogeographical setting (in this examplethere are three: ocean floor, volcanic island-arc, and fully terrestrial conditions with com-plex relief). In this example only the last geo-logic unit can be said to be ‘‘active’’ or stillin its original geomorphological form. Theother two units are no longer resolvable aseither ocean floor or volcanic arc; they arepart of the basement of the last unit.

A more dynamic example would illustratethe point that, over time, segments of theearth’s crust can change their relationship,form, and position, thereby affecting paleo-geography. Along compressional plateboundaries, for example, oceanic crust andridges may undergo subduction, thereby los-ing their original geological and geomorpho-logical character. Similarly, ocean crust orvolcanic arc suites overthrusting continentalmargins along a collisional suture will create

a new stack, one having the distinctive com-posite nature of a foldbelt (a geological unitin its own right, consisting of several amal-gamated ‘‘fossil’’ precursors). Thus a newgeography may arise ontogenetically fromthe old.

In creating reliable paleogeographicalmaps, it has long been recognized that to de-termine the successive sizes of a geologicalunit, the effects of movement and deforma-tion have to be figuratively undone. How‘‘reverse ontogeny’’ can be understood andworked out in a particular case of interest isillustrated in figure 2. The two cross sectionsin figure 2A represent the structure of thefoldbelt in eastern Cuba and western Hispan-iola as seen today. In figure 2B, these crosssections are restored to their relative posi-tions before separation caused by sinistralmovements along fault A9–A (steps 1 and 2).This requires the deletion of entities thathave been intercalated as the result of move-ment along A9–A (Cayman Trench andsouthern and northern Hispaniolan blocks).With these omitted, it can be seen that thetwo cross sections can be precisely lined upalong their volcanic arc sequences (step 2).In step 3, we simplify the present-day rela-tive position and width of the geologicalunits found in the foldbelt (carbonate conti-nental margin, ophiolites, volcanic arcs, andsedimentary basins), depicting them as a se-ries of superimposed bars. In steps 4 and 5we sequentially remove the effects of over-thrusting and shortening due to internal de-formation within the units themselves, there-by resolving the original width of the fold-belt. Although this example is schematic, itmakes the point that the geography of thepresent may differ radically from the geog-raphy of the past, even for the ‘‘same’’ landmass.

For each geological unit under discussion,positional coordinates in time (in Ma) andspace (in degrees of latitude and longitude)are provided in appendix 2. Unit positionswere constrained per time-slice using varioussources of information, including maximumpossible amplitude of strike-slip movements,continuity of geological structures across dif-ferent block-terranes or arc segments, pres-ence of correlatable rock complexes ofknown age in two or more distinct units, and


so forth (fig. 3). Plate motions provide an-other, more general form of constraint (ap-pendix 2).

The third variable used to build paleogeo-graphic maps is physical geography, whichwe define as information (indicators) regard-ing positive or negative relief of contiguousgeological units (or their subcomponents)across time. Relevant information is widelybut thinly scattered in the geological litera-ture. For effective use, this information hadto be refined in various ways, including re-interpretation of the age of late Tertiary sec-tions according to current paleontologicaland stratigraphical criteria, reanalysis (if re-quired) of depositional environment, deter-mination of topographic indicators, and soforth.

A land indicator provides evidence of theexistence of subaerial conditions within ageological unit at a specific time in its on-togeny. However, land indicators have to beinterpreted carefully, as some are much betterthan others. Thus, although unconformitiesand hiatuses both represent gaps in the geo-logical record, their significance is not thesame. An unconformity is a surface that sep-arates two superimposed sets of strata. Un-conformities are erosional surfaces and mayprovide clear evidence of land emergence ifassociated with long-lasting hiatuses (but seebelow). Although also a gap in the rock re-cord, a hiatus is defined as time not repre-sented by strata (i.e., an interval of nonde-position, erosion, or both). Uplift and non-deposition of sediments during n1 Ma canproduce a hiatus, but uplift also causes ero-sion of preexisting rocks, thereby addition-ally enlarging the gap in the record by n2Ma. The total gap (n1 1 n2) is therefore pro-duced by erosion as well as nondeposition.Owing to the effects of bottom currents orrising sea floor, hiatuses can also be producedin the absence of subaerial exposure by sub-marine erosion, nondeposition, or both ofthese processes. Thus, a hiatus by itself doesnot imply land emergence, nor does the ac-tual gap (in millions of years) necessarilymirror the time during which the area wasuplifted as land.

Usually, indicators other than gaps per seare required in order to reach reliable con-clusions about land emergence and coastline

position. These may include (1) sedimentsdeposited in terrestrial environments (e.g.,red beds, alluvium), (2) weathering surfacesor weathering products (e.g., paleosols,‘‘coated’’ pebbles), (3) land-derived sedi-ments in contiguous marine basins (e.g., con-glomerates, sandstones), (4) lagoonal depos-its representing fresh- or brackish-water near-shore environments, (5) coastal sediments(e.g., beach sands, dunes), and (6) remains ofterrestrial organisms preserved in marinesediments (e.g., lignites, terrestrial plants,pollen, spores).

Among marine indicators, different rocktypes and their fossil inclusions are of greatvalue because they are highly correlated withwater depth at the time of original deposi-tion. Lack of marine sediments in a particularsection may be the result of erosion (withina hiatus), so whether a transgression actuallyoccurred has to be resolved by examining thecomposition and environment of depositionof rocks in surrounding basins. This is whysome geological units are represented as sub-marine environments in the paleogeographi-cal maps, even though marine rocks of ap-propriate age are not known within them. Forexample, although several hiatuses are re-corded on the Beata Ridge, unequivocal in-dicators of subaerial conditions have notbeen found. Accordingly, we assume thatthese hiatuses are due to submarine erosionand nondeposition, and portray the ridge asa submarine feature from Late Oligocene toRecent (see fig. 4).

Other phenomena may have paleogeo-graphical significance if they provide insightsinto specific conditions or occurrences in thepast. For present purposes, the most impor-tant of these is termination of volcanic activ-ity in island arcs. This phenomenon, usuallydue to arc–arc, arc–ridge or arc–continentalcollision (Hamilton, 1988), is known to havea profound effect on uplift. The mechanismof uplift is related to the emplacement ofhuge intrusive bodies coincidental with arcextinction, causing widespread isostatic ad-justment (uplift) along the arc axis (Iturralde-Vinent, 1988a, 1994a). Uplift is then fol-lowed by subsidence within a period of onlya few million years. Although postmagmatic-phase uplift is well substantiated, its perti-nence to Caribbean paleogeography has not


Fig. 4. Beata Ridge: top left, Current topography and approximate thickness of late Tertiary sedi-ments (deep to basement); top right, stratigraphic columns (in the case of core sites, gaps in shadingsignify presence of hiatuses); and bottom, cross section (compiled from many sources; see appendices1 and 2). Chronostratigraphic reference column in this and all other section diagrams after Berggren etal. (1995), at scale. For details on biozones and chronology consult figure 20.


been widely appreciated (but see Early Mid-dle Jurassic to Late Eocene Paleogeography).As discussed here in relation to the arcs as-sociated with the evolution of the CaribbeanPlate, termination of arc magmatism in theAptian, Late Campanian/Early Maastrichtian,

and early Middle Eocene consistently pro-duced substantial uplift followed by deeperosion of the extinct volcanic arc edifices(Iturralde-Vinent, 1988a, 1994a; MacPheeand Iturralde-Vinent, 1994, 1995).

PALEOGEOGRAPHY OF THE CARIBBEAN REGION:EVIDENCE AND ANALYSIS

This section provides an analysis of theseveral categories of basic geological infor-mation presented in the tables, figures, andappendices. Appendices 1 and 2 and tables1–4 should be consulted throughout for sup-porting evidence and additional literature notdirectly referenced in the following para-graphs.

EARLY MIDDLE JURASSIC TO LATEEOCENE PALEOGEOGRAPHY

This long stage in the evolution of the Ca-ribbean region may be considered to have be-gun with the creation of a basin, the embry-onic Caribbean Sea, coincident with thebreak-up of Pangaea and the separation ofLaurasia from Gondwana (fig. 5; table 1).During the late Triassic/middle Jurassic, anepicontinental siliciclastic basin developedbetween the cratonic areas of South andNorth America in reaction to the eastwardmigration of the Tethys (Anderson andSchmidt, 1983; Burke et al., 1984; Bartok,1993). This epicontinental sea should not bethought of as the Caribbean basin as pres-ently configured, but as a precursor situatedwithin western Pangaea (see Pindell, 1994:fig. 2.6a). Its epicontinental nature is dem-onstrated conclusively by associated marineinvertebrate faunas and sediment composi-tion (Salvador, 1987, 1991; Pszczolkowski,1987).

The early Caribbean basin began as a nar-row seaway between the Pacific and Tethys,probably during the Bajocian/Bathonian(Bartok et al., 1985) as oceanic crust was be-ing formed between western Laurasia (NorthAmerica) and western Gondwana (SouthAmerica) (fig. 5; Pindell, 1994: fig. 2.6b, c).The existence of lands closely bordering thisseaway is indicated by evidence of coastal-

type vegetation at several localities of ?Early/Middle Jurassic to early Late Jurassic age inMexico and western Cuba (Areces-Mallea,1990).

In the North American portion of Laurasiaan important marine transgression took placeduring early to middle Oxfordian time. Vo-lant terrestrial and shallow-water marine ver-tebrates, indicative of proximate land envi-ronments, make an appearance at this time atlocalities on the Guaniguanico terrane thatforms westernmost Cuba (Iturralde-Vinentand Norell, 1996). The occurrence of thisOxfordian faunule in what is now Cuba con-stitutes an example of ‘‘Viking funeral ship’’emplacement (McKenna, 1973), becausethese taxa were extinct before Guaniguanicoreached western Cuba long after detachingfrom the Yucatan borderland early in the Ter-tiary (Iturralde-Vinent, 1994a, 1996a; Bra-lower and Iturralde-Vinent, 1997). In theSouth American portion of Gondwana thepaleogeographical context was different, be-cause the main transgression across thenorthern continental margin took place later,during the Early Cretaceous. With the Ox-fordian transgression and widening marinegap between Laurasia (North America) andGondwana (South America), any possibilityof direct, overland dispersal between thesecontinental areas ended (fig. 5).

The developing Caribbean seaway (‘‘His-panic Corridor’’ of Bartok et al., 1985) un-derwent widening from the Middle Jurassicto the Early Cretaceous as a consequence ofsea floor spreading (Pindell, 1994: fig. 2.6c,d). Oceanic crust and sediments formed atthat time are now represented in part by de-formed ophiolite bodies and thrust beltsaround the margins of the Caribbean region(Guatemala, Greater Antilles, Aruba/Tobago


Belt, Caribbean Mountains, and Colombian–Venezuelan Andes) (Dengo and Case, 1990).Oceanic basalts associated with radiolariancherts and carbonate rocks of Jurassicthrough Cretaceous age have been foundwithin these allochthonous crustal bodies(Bartok et al., 1985; Montgomery et al.,1994; Iturralde-Vinent, 1996a). In some plac-es these vulcano-sedimentary rocks consti-tute segments of a continuous section (e.g.,northwestern Cuba, southwestern PuertoRico), but most frequently they consist ofisolated cobble- to boulder-sized bodieswithin highly deformed belts. Although itcannot be determined from existing infor-mation whether the age gaps record actualhiatuses, the absence of any evidence of ter-restrial environments in these contextsstrongly suggests prevailing deep-water con-ditions from their origin until their incorpo-ration into the foldbelts fringing the Carib-bean region.

Indications of lands or shallow seas (orboth) within the confines of the early Carib-bean sea are found in rocks of the Cretaceousvolcanic arc. As a geological unit, the Cre-taceous arc is defined by a particular set ofigneous, sedimentary, and metamorphicrocks of Neocomian through late Campanian/early Maastrichtian age (Dengo and Case,1990; Iturralde-Vinent, 1994a, 1994c, 1996a,1996b). Today, elements of this arc are wide-ly distributed in the foldbelts found withinthe Caribbean region (fig. 5). The paleogeo-graphical position of the Cretaceous arc inrelation to North and South America remainsthe subject of debate (see Leclere and Ste-phan, 1985; Ross and Scotese, 1988; Don-nelly, 1989a; Pindell, 1994, Mann et al.,1995; Hay and Wold, 1996; Iturralde-Vinent,1996a, 1997b).

Shallow marine environments are markedby the occurrence of rudist limestones of dif-ferent ages, occurring as isolated, lenticularintercalations within marine-deposited vol-canic sediments. Although the presence ofrudists is not diagnostic of nearby emergentland, it is consistent with the existence ofatoll-like islands similar to those seen todayon shallowly submerged volcanoes. Moresubstantive indications of land developmentin the Cretaceous island arc are Neocomianplant remains reported from the Los Ranchos

Fm, a volcanic arc section in Hispaniola(Smiley, MS, cited by Kesler et al., 1991b).This assemblage (including Gleichenites,Zamites, Phoenicopsis, Yuccites, Podozami-tes, and other taxa) is thought to have grownin a warm, open, seasonally dry habitat ad-jacent to the shallow-water marine environ-ment in which the remains were deposited(Kesler et al., 1991b). Other geological in-dicators of emergence during the evolutionof the Cretaceous arc include terrestrially de-posited volcanic rocks and several major un-conformities occurring within volcanic arcsections, usually in association with hiatusesand basal conglomerates.

Unfortunately, the paleogeographical in-formation content of these indicators is lim-ited and cannot be reliably used to provide adetailed assessment of areal extensivenessand orographic relief in the Cretaceous arc.On the other hand, such information doesprovide some indication of the temporal suc-cession of environments in specific geologi-cal units. In Hispaniola, the Neocomianplant-bearing rocks are overlain by marinelimestones of the Albian Rio Husillo Fm(Kesler et al., 1991b; Iturralde-Vinent,1997a), implying that a transgression oblit-erated previously existing terrestrial environ-ments. This sequence of events is rathercommon in the Cretaceous arc sections, andour interpretation is substantiated by twopieces of evidence. First, volcanic and non-volcanic marine rocks drape all of the un-conformities recorded in the Cretaceous arc,indicating that hiatuses were succeeded by anew phase of marine deposition (Nagy et al.,1983; Lewis et al., 1991; Rojas et al., 1995;Iturralde-Vinent, 1995, 1997a; Beccaluva etal., 1996). Second, the rudist limestones oc-cur in the form of lenses intercalated withinother marine sediments and laterally transi-tional with isochronous deeper water beds(Rojas et al., 1995). This indicates that anyintra-Caribbean land environments existingon the volcanic arc at that time would havebeen limited in size, and therefore suscepti-ble to rapid obliteration during transgressivephases. We conclude from this that, whileland environments certainly existed in theCaribbean Basin during the Cretaceous, theywere short-lived, probably winking in and


out of existence within periods of only a fewmillion years.

A good example of such evanescence canbe seen in the section along the Canal PasoBonito a Cruces, northwest of Sierra de Es-cambray, Cuba (unpubl. obs.). Here, Creta-ceous volcanic breccias and agglomerates arepatchily overlain by slope and alluvial sandsand gravels which have been weathered topaleosols that exhibit caliches, root casts, andother indications of subaerial exposure.These rocks are in turn succeeded by un-weathered, well-bedded tuffs and rare marinelimestones. Thus, however long this islandmay have existed, it did not last as land intolater epochs.

Tracking the paleoposition (i.e., changes inlatitude and longitude through time) of theCretaceous volcanic arc is another issue ofgreat importance. There appears to be a con-sensus, at least among authors of the mostwidely cited plate tectonic models, that thisarc originated in the Pacific (Malfait andDinkelmann, 1972; Burke et al., 1984; Le-clere and Stephan, 1985; Pindell and Barrett,1990; Pindell, 1994), although other inter-pretations also exist (fig. 5; Turner, 1972;

Donnelly, 1989a; Iturralde-Vinent, 1994a,1997b). However, whatever its starting po-sition might have been, there is little evi-dence to support the view that the continentswere physically united by the arc. In princi-ple, connection might have occurred duringmajor uplift events recorded in associationwith various unconformities (between theNeocomian and early Albian, between theConiacian and Santonian, and in the earlyCampanian). However, information is lack-ing concerning whether these erosional sur-faces were continuous or synchronous alongthe trend of the arc (Iturralde-Vinent 1994a,1996b, 1997b). During the late Campanianand early Maastrichtian (ca. 70–80 Ma) sub-stantial subaerial exposure existed along theCretaceous arc and adjacent continental mar-gins, as indicated by evidence of deforma-tion, angular uncomformities, hiatuses, deep-seated erosion, mountain building, and ter-restrial sedimentation (including conglomer-ate and paleosol development) (Khudoleyand Meyerhoff, 1971; Mattson, 1984; Push-charovski et al., 1989; Maurrasse, 1990;Lewis et al., 1990; Iturralde-Vinent, 1994a,c, 1995, 1996b, 1997b; Beccaluva et al.,


Fig. 5. Plate tectonic model of the Caribbean region for Jurassic through Eocene (after Iturralde-Vinent, 1997b, slightly modified). Maps are designed to display tectonic information; they are notpaleogeographically accurate.


1996). At this time magmatic activity ter-minated along large segments of the volcanicarc, including its western and eastern extrem-ities (i.e., modern central and western Cubain the north and west, and the Netherlandsand Venezuelan Antilles and CaribbeanMountains in the south and east). Cessationof the magmatic phase is coincidental on awider scale with the Subhercinian orogeny(Schwan, 1980; Leonov and Khain, 1987),an important global tectonic event, and sev-eral drops in eustatic sea level (Haq et al.,1987). It is reasonable to infer that if anyland connection existed between North andSouth America in the last part of the Creta-ceous, this connection most probably wouldhave occurred during the late Campanian andearly Maastrichtian. However, if in fact con-tact occurred it would have been brief, be-cause transgressive late Maastrichtian marinesediments are recorded in the Cretaceousvolcanic arc as well as in North and SouthAmerica (table 1; appendix 1).

There does not appear to be any reliableevidence of permanent islands or island–con-tinent connections in the Caribbean regionduring the early part of the Paleogene, al-though we cannot reject this possibility al-together (see Biogeographical Hypothesesand Caribbean Paleogeography). An uncon-formity dated to the K/T boundary is knownto exist in some parts of the Caribbean seafloor, but uninterrupted sedimentation evi-dently continued elsewhere in the CaribbeanBasin, even within the area of the inactiveCretaceous arc. This unconformity couldhave been caused by submarine erosion rath-er than by cessation of deposition if, for ex-ample, it was induced by tsunamis triggeredby the K/T impactor landing in Chicxulub(Pszczolkowski, 1986; Maurrasse and Sen,1991) and/or in the Yucatan Basin (Iturralde-Vinent, 1992).

After a period of relative quiescence, vol-canic activity began again in the Caribbeanregion in the Paleocene (Iturralde-Vinent,1994a, 1997b), as indicated by the occur-rence of a set of magmatic, sedimentary, andmetamorphic rocks that widely outcrop in theregion (eastern Cuba, Jamaica, Hispaniola,Puerto Rico, Virgin Islands, Cayman Ridge,Nicaragua Rise, and Aves Ridge/Lesser An-tilles). This resumption of activity ushered in

the formation of the Paleogene volcanic arc.However, magmatism was short-lived in thisnew arc, effectively ending by the MiddleEocene. Within the Paleogene arc, marinesediments uninterruptedly filled some basins,indicating that land contacts between arccomponents and the continents were not (orwere no longer) in existence (Lewis andStraczek, 1955; Bresznyanszky and Iturralde-Vinent, 1985). Uninterrupted sedimentationfrom Paleocene through Middle Eocene is re-corded in basins located on the Caribbeanseafloor (Edgar et al., 1973; Sigurdsson et al.,1997), as well as in foldbelt areas, along thetrend of the inactive segments of the Creta-ceous arc (e.g., deep marine sediments out-cropping through much of La Habana andMatanzas provinces in west central Cuba)(Bronnimann and Rigassi, 1963; Pszczol-kowski, 1987; Bresznyanszky and Iturralde-Vinent, 1985; Bralower and Iturralde-Vinent,1997). Furthermore, a major transgressionoccurring between the late Early and MiddleEocene produced extensive deposits of shal-low- and deep-water marine carbonate rockson previously positive areas throughout theCaribbean (Lewis and Straczek, 1955; Bron-nimann and Rigassi, 1963; Iturralde-Vinent,1982, 1994a; Holcombe et al., 1990; Lewiset al., 1990; Maurasse, 1990; Edgar et al.,1973; Sigurdsson et al., 1997). The occur-rence of this transgression militates againstthere having been any considerable exposureof land during the early Middle Eocene.

Another global tectonic event, the Illyrianphase of tectogenesis (Leonov and Khain,1987), produced a profound modification ofthe Caribbean tectonic regime between Mid-dle and Late Eocene (fig. 5). Extensive upliftand deformation occurred not only in the Ca-ribbean Basin per se but also in the surround-ing continental margins and oceanic domains(Khudoley and Meyerhoff, 1971; Gonzalezde Juana et al., 1980; Mattson, 1984; Lewiset al., 1990; Iturralde-Vinent 1981, 1994a,1994b, 1994c, 1996a; Maurrasse, 1990).With reorganization of the geodynamic re-gime, the relative motion of the CaribbeanPlate shifted eastward, and relative motionbetween the North and South Americanplates decreased markedly (Pindell, 1994). Inconcert with these developments, several de-formed belts were consolidated and accreted


against continental areas, and magmatic ac-tivity shifted to new locations (CentralAmerica and Lesser Antilles). The north-western edge of the Caribbean Plate (NWGreater Antilles Belt) collided with the Yu-catan and Bahamas margins, while the south-eastern edge (Aruba/Tobago Belt) interactedwith the South American margin. Also, be-ginning in the latest Eocene, new tectonic el-ements were defined within the Caribbean re-gion. Transverse faulting divided the plateinto several microplates and block-terranes,and their subsequent displacement disruptedthe original structure (see appendix 2).

In summary, existing data indicate thatsubaerial entities were formed along the Cre-taceous and Paleogene volcanic arcs andnearby continental margins from time to timefrom the Jurassic into the Eocene. However,there is no evidence that any of these entitieslasted for long periods; indeed, none seemsto have survived as emergent land into thesubsequent interval (latest Eocene to MiddleMiocene). Nevertheless, if the Cretaceous arcever connected North and South America,this most likely occurred during the lateCampanian to early Maastrichtian (ca. 70–80Ma ago), just after extinction of the Creta-ceous volcanic arc.

LATEST EOCENE TO MIDDLEMIOCENE PALEOGEOGRAPHY

In this section, three ‘‘snapshot’’ intervalsare discussed in detail: Eocene–Oligocenetransition (35–33 Ma), Late Oligocene (27–25 Ma), and early Middle Miocene (16–14Ma). Basic data are presented in appendix 1(see also figs. 6–8 and tables 2–4). These in-tervals were chosen in order to contrast pe-riods of maximum and minimum land de-velopment. The Eocene–Oligocene transitionwas a time of general uplift; therefore, theamount of subaerial land in the Caribbeanshould have been at a maximum. The LateOligocene was a time of high sea level, andtherefore of mimimum exposure (and, prob-ably, interconnectedness) of emergent areasin the early Cenozoic. In the early MiddleMiocene, further isolation of land areas tookplace as a consequence of active tectonic dis-ruption of the northern and southern Carib-bean Plate boundaries. In the case of theGreater Antilles, this resulted in the subdi-

vision and separation of block-terranes pre-viously acting as continuous landmasses.This subdivision may have been significantbiogeographically if it caused island–islandvicariance (as opposed to continent–islandvicariance; see Biogeographical Hypothesesand Caribbean Paleogeography).

EOCENE–OLIGOCENE TRANSITION (35–33 Ma)

The transition between the end of the Eo-cene and the beginning of the Oligocene(zones P16 to P18 of Berggren et al., 1995)coincides with the Pyrenean phase of tecto-genesis (Schwan, 1980; Leonov and Khain,1987), the effects of which are well repre-sented in the Caribbean region (MacPhee andIturralde-Vinent, 1995). In this phase, gen-eral tectonic uplift coincided with a majoreustatic sea level drop at ca. 35 Ma (Milleret al., 1996). As a result, subaerial exposurewithin the Caribbean basin was probablymore extensive then than at any other timein the Cenozoic, including the late Quater-nary. The map in figure 6 reflects this fact(see also table 2). However, it is important tocompare this map with other ‘‘Oligocene’’reconstructions which represent this periodas one of overall minimum land exposure(e.g., Gonzalez de Juana et al., 1980; Gallo-way et al., 1991; Macellari, 1995). The ex-planation for the difference in treatment liesin the fact that most Oligocene reconstruc-tions depict the paleogeography of the mid-to later Oligocene, by which time the Pyre-nean orogenic phase had terminated (see dis-cussion of next map).

Evidence for Pyrenean uplift can be seenin stratigraphic sections as well as submarinedredge samples, drill cores, and seismic linesrecovered from many parts of the Caribbeanand surrounding continental borderlands (ta-ble 2; appendix 1). Stratigraphic sectionsconsistently lack marine sediments of latestEocene–Early Oligocene age, presenting in-stead hiatuses, red beds, and other kinds ofterrestrial deposits (table 2; appendix 1). Inmany sedimentary basins located near formertopographic highs, latest Eocene/Early Oli-gocene sediments carry abundant land-de-rived debris, chiefly very coarse conglomer-ates (matrix or clast-supported) and sand-stones. This type of sedimentary unit is socommon that it may usefully be dubbed the


‘‘Eocene–Oligocene transition conglomerateevent.’’ In places far removed from terrestrialsediment sources, such as basins in the Ca-ribbean sea floor, deep-marine depositionwas condensed (i.e., pelagic sedimentationoccurred at a low rate). Low sedimentation

rate is in agreement with the combined ef-fects of lowered sea level and restricted sea-water circulation (Haq et al., 1987).

The Aves Ridge deserves special mentionbecause it has been proposed as the site of apotential landspan between the Greater An-


Fig

.6.

Pal

eoge

ogra

phy

ofC

arib

bean

regi

onas

reco

nstr

ucte

dfo

rla

test

Eoc

ene/

Ear

lyO

ligo

cene

(35–

33M

a).

The

reco

nstr

ucti

onis

desi

gned

tode

pict

cond

itio

nsob

tain

ing

inth

eC

arib

bean

regi

ondu

ring

the

peri

odof

max

imum

Cen

ozoi

cla

ndex

posu

re,

occa

sion

edby

very

low

eust

atic

seal

evel

and

wid

espr

ead

regi

onal

upli

ft(P

yren

ean

orog

enes

is).

Alt

houg

hsu

baer

ial

cond

itio

nsex

iste

dal

ong

atle

ast

part

ofth

eA

ves

Rid

gedu

ring

this

peri

od,

depi

cted

land

span

shor

elin

esan

dco

ntac

tpo

int

wit

hS

outh

Am

eric

ash

ould

bere

gard

edas

conj

ectu

ral,

asis

the

illu

stra

ted

conn

ecti

onbe

twee

nth

eS

outh

ern

His

pani

olan

and

Blu

eM

ount

ains

Blo

cks

(for

expl

anat

ion,

see

text

and

appe

ndix

1).


tilles and northern South America (Borhidi,1985; MacPhee and Iturralde-Vinent, 1994,1995). This structure—presently almost com-pletely submerged—was originally continu-ous with the Greater Antilles Ridge and isconsidered to have constituted a single entityin latest Eocene/Early Oligocene time. Weargue that exposure of the ridgecrest created,for a short time ca. 33–35 Ma, a series oflarge, closely spaced islands or possibly acontinuous peninsula stretching from north-ern South America to the Puerto Rico/VirginIslands Block (see GAARlandia Landspanand Island–Island Vicariance).

Among the points in the paleogeographi-cal reconstruction that require further refine-ment and explanation are the close position-ing of southwestern Hispaniola and the BlueMountains Block, as well as inferred per-manent exposure of parts of Jamaica as earlyas 33–35 Ma (see appendix 1). Also note-worthy is the fact that western Cuba wouldhave been separated by deep-water environ-ments (Havana–Matanzas Channel) fromcentral and eastern Cuba at this time.

Other features characteristic of the presentCaribbean sea floor that did not exist in Eo-cene–Oligocene times (and are therefore notdepicted) include the Cayman Trench andAnegada Trough, among others (see appen-dix 1) (Calais et al., 1989, 1992; see discus-sion by MacPhee and Iturralde-Vinent, 1994,1995).

LATE OLIGOCENE (27–25 Ma)

The Late Oligocene (zones P21b–P22 ofBerggren et al., 1995) was a time of exten-sive marine invasions, probably due to acombination of tectonic subsidence and highsea level stands (table 3). Marine sedimentsof this age are common in North and SouthAmerica and the Greater Antilles (fig. 7). In-undation of terrestrial environments began asearly as zone P19 (Berggren et al., 1995) andcontinued into zone P22. Evidence of LateOligocene transgression is seen in strati-graphic sections, dredge samples, drill cores,and seismic lines recovered from many partsof the Caribbean region and the FloridaBlock. Stratigraphic sections consistently re-cord mid-Late Oligocene marine depositsoverlying older rocks. In the more distal sed-

imentary basins of the Caribbean sea floor,the rate of sediment accumulation increasesat this time and deeper water environmentsupwardly dominate Oligocene sections. Theamount of land-derived debris in such sec-tions is substantially reduced, with fine-grainand biogenic deposits dominating. Neverthe-less, heights-of-land remained persistentlysubaerial along the axis of GAARlandia, asshown by the existence of rocks formed innonmarine environments, presence of land-derived sediments and plant remains in prox-imate marine basins, and depositional hiatus-es in sections. These results suggest that, inan area such as the Caribbean region, inwhich vertical motions have been marked,the amount of terrestrial exposure or inun-dation cannot be simply read out from sea-level curves (e.g., Haq et al., 1987).

EARLY MIDDLE MIOCENE (16–14 Ma)

The early Middle Miocene paleogeogra-phy (zones M5–M7 of Berggren et al., 1995)of the Caribbean region shows the effects ofdisruption of the deformed foldbelt boundingthe Caribbean Plate (fig. 1; table 4). The pro-cess of Neogene disruption, the tectoniccharacteristics of which have been extensive-ly investigated (Mann et al., 1990; Pindelland Barrett, 1990; Pindell, 1994; Macellari,1995), is recorded east of Cuba along theplate’s northern boundary, and within theNetherlands and Venezuelan Antilles to Trin-idad and Tobago along its southern bound-ary. Localized extension occurred along bothboundaries as grabens, pull-apart basins, andtrenches began to form. This, combined withcontinuing marine transgression, served tofurther isolate fault-bounded block-terranesfrom one another along plate boundaries. Ex-amples of extensional features formed or ac-tivated at this time in the central CaribbeanBasin include the Cayman Trench betweenCuba and Hispaniola, the Anegada Troughbetween the northern Virgin Islands and St.Croix/Aves Ridge, the Puerto Rico Trench(Mann and Burke, 1984; Mann et al., 1990),and the Tortuga Basin (fig. 1). Significantly,extension also took place along the axis ofthe Nicaragua Rise (Droxler et al., 1989)

Figure 8 shows Puerto Rico and Hispan-iola as being connected by a neck of landinto the Miocene. The existence of this con-


Fig

.7.

Pal

eoge

ogra

phy

ofC

arib

bean

regi

onas

reco

nstr

ucte

dfo

rL

ate

Oli

goce

ne(2

7–25

Ma)

.D

urin

gth

ispe

riod

,ge

nera

lsu

bsid

ence

and

high

erse

alev

els

grea

tly

dim

inis

hed

land

area

wit

hin

the

Car

ibbe

anS

ea,

sund

erin

gth

ela

ndsp

anco

nnec

tion

betw

een

the

Gre

ater

Ant

ille

san

dno

rthw

este

rnS

outh

Am

eric

a(f

orex

plan

atio

n,se

ete

xtan

dap

pend

ix1)

.


nection is indicated, although not proven, byevidence that (1) the formation of the MonaPassage was a neotectonic event (Larue etal., 1990; Jany et al., 1990; Masson andScanlon, 1991), and (2) the first late Ceno-zoic marine sediments to onlap eastern His-paniola and western Puerto Rico were Pleis-tocene limestones, implying the existence ofa barrier of some kind. Additional data fromthe floor of the Mona Passage would help toclarify the history of this connection.

Comparison of figures 7 and 8 reveals that

the amount of land in the central part of theCaribbean Basin was approximately the samein the Late Oligocene and the early MiddleMiocene. However, by the Middle Miocene,block-terranes fringing the Caribbean Platewere already widely separated; many ofthese would never be reunited again, evenduring the glaciations (low-stand events) ofthe late Quaternary. Late in the Middle Mio-cene, western Cuba finally achieved drylandcontact with central Cuba after the disap-pearance of the Havana–Matanzas Channel.

BIOGEOGRAPHICAL HYPOTHESES AND CARIBBEANPALEOGEOGRAPHY

This section is not intended as a review ofthe recent systematically oriented literatureon Caribbean biogeography. Such an under-taking, if properly comprehensive, would bea major undertaking unto itself and thereforerequires separate treatment (MacPhee andIturralde-Vinent, in prep.). Our narrower pur-pose in this paper is to examine specificproblems in light of the new paleogeograph-ical reconstructions developed in precedingsections, with an emphasis on ‘‘how they didit’’ as opposed to ‘‘who did it.’’ We pursuethis by considering, in turn, three quite dif-ferent models proposed by Rosen (1975,1985), Hedges and co-workers (1992, 1994;Hedges, 1996a, 1996b), and MacPhee andIturralde-Vinent (1994, 1995).

CONTINENT–ISLAND VICARIANCE:MODEL OF ROSEN

Rosen’s (1975, 1985) continent–island vi-cariance model was the earliest attempt tocreate a cladistically oriented biogeographyof the Caribbean region with an emphasis onvertebrates.4 Rosen’s (1975) principal inno-

4 As a matter of historical record, the first intuitivemodel of Antillean vicariance was proposed by J. Issacdel Corral (1940) in order to explain the origin of theCuban mammal fauna by Wegenerian ‘‘continentaldrift.’’ According to this author, before the Late Miocenethe Greater Antilles were attached to northern SouthAmerica (Colombia and Venezuela), whence they re-ceived their mammalian fauna. Later on, the islandsdrifted northward to their current positions. Vandel(1973) used Issac del Corral’s (1940) model to accountfor the relationship of troglobytic faunas in Cuba andnorthern South America.

vation was to wed modern ideas concerningthe relationships of whole faunas and areas(see Croizat, 1964; Nelson and Platnick,1981; Humphries, 1992; Morrone and Crisci,1995) to the emerging theory of plate tecton-ics, for one of the world’s most complicatedbiological and geological regions. Vicariancetheory has been critically explored with spe-cial reference to Antillean faunas in severalrecent works (e.g., Guyer and Savage, 1987,1992; Kluge, 1988, 1989; Page and Lydeard,1994; Roughgarden, 1995). In the main,however, cladistic biogeographers have notconcerned themselves with updating or re-vising Rosen’s (1975: 453-454) paleogeo-graphical inferences, despite his exhortationsthat they do so.

In Rosen’s model, the ‘‘Antillean archi-pelago’’ or ‘‘proto-Antilles’’—here under-stood as the Cretaceous and Paleocene–Eo-cene volcanic arcs, although he did not makethis distinction—were assumed to have orig-inated as a series of closely spaced islandson the leading edge of the Caribbean crustalplate in roughly the position occupied bypresent-day Central America. As these is-lands were tectonically transported eastward,they interacted with adjacent continentalmargins in such a manner that they were ableto receive the greater part of their biota inessentially one event (Rosen, 1975: fig. 8).As Perfit and Williams (1989) pointed out,although Rosen (1975) described this com-mon-cause event as vicariant in nature, hewas ultimately noncommittal as to how im-migration actually occurred. (Rosen’s maps


Fig

.8.

Pal

eoge

ogra

phy

ofC

arib

bean

regi

onas

reco

nstr

ucte

dfo

rea

rly

Mid

dle

Mio

cene

(16–

14M

a).

Neo

gene

tect

onic

mov

emen

tsan

dde

form

atio

nsal

ong

the

mar

gins

ofth

eC

arib

bean

Pla

tesu

bdiv

ided

form

erst

ruct

ural

ridg

es(G

reat

erA

ntil

les

and

Ave

sR

idge

s),

crea

ting

isol

ated

tect

onic

bloc

ksan

dte

rran

esse

para

ted

byde

ep-w

ater

gaps

(e.g

.,C

aym

anT

renc

h,M

ona

Pas

sage

,A

nega

daT

roug

h).

Sou

ther

nC

entr

alA

mer

ica

and

Sou

thA

mer

ica

not

inco

ntac

tac

ross

Pan

aman

ian

regi

on(D

uque

-Car

o,19

90;

Kol

arsk

yet

al.,

1995

a,19

95b;

for

expl

anat

ion,

see

text

and

appe

ndic

es1

and

2).


imply that some short-distance, overwatertransport would have been required, which ofcourse muddies a primary distinction be-tween vicariance and dispersal.) Subsequentplate motions, Rosen argued, carried the is-lands and their resident faunas to their cur-rent locations. During this postvicariant in-terval, amounting to all of the Cenozoic, thefaunas of the islands were further shaped byextinction, local radiation, and, in a smallnumber of cases, by later overwater dispers-als (Rosen, 1975: figs. 9–12). He ended thissection of his paper (pp. 453–454) with a listof nine ‘‘general remarks’’ on his ‘‘combinedvicariant and geophysical model,’’ remarkingthat he incorporated ‘‘many provisional ideasthat urgently need independent testing.’’

In assessing Rosen’s (1975) views in lightof modern tectonic theory, it should be ap-preciated that the tectonic model that Rosenutilized—basically that of Malfait and Din-kelmann (1972)—is not very different in itsconceptual framework from current modelsof the evolution of the Caribbean Plate (e.g.,Pindell, 1994). Instead, the essential problemlies with Rosen’s paleogeographical scenar-io: he simply assumed an identity relation-ship between geological units and geograph-ical entities in his discussion of the originand early history of the ‘‘proto-Antilles,’’ asbecomes obvious when his maps are exam-ined (on this point see also Rosen [1985] andcomments by Perfit and Williams [1989] andHedges [1996a]). In short, he viewed the pa-leoislands which existed in the position ofCentral America in the Cretaceous as some-how the ‘‘same’’ as the ones in existence to-day, as the ‘‘transposed, original archipelago,the Antilles’’ (Rosen, 1975: 432; see alsoRosen, 1985: 652).

This view of paleogeographical continuityacross 80 Ma or more of earth history is fun-damentally flawed, because it does not takeinto consideration the effects of tectonic pro-cesses on Antillean biogeography betweenthe late Campanian and Recent (see Iturral-de-Vinent, 1982). As noted above, the fold-belt which constitutes the geological base-ment of the present-day Greater Antilles wascreated by complex phases of amalgamationand deformation of rock units comprising theCretaceous and Paleogene arcs, together withassociated oceanic crust and continental mar-

gin terranes (Iturralde-Vinent et al., 1996a).As we have emphasized, any islands then ex-isting along the axis of the developing fold-belt would have been intensely modified bythrusting, folding, subsidence, and burial be-neath thick tectonic nappes. Therefore, thegeography of the ‘‘proto-Antilles’’ (i.e., theCretaceous and Paleogene volcanic arcs)probably has little or nothing to do with thegeography of the existing islands (Iturralde-Vinent, 1982).

This point was also missed by Guyer andSavage (1987, 1992), who in some respectswent even further than Rosen did in assum-ing the permanency of islands. Guyer andSavage (1987: 526) proposed that, with theexception of a very few recent dispersals, theancestors of the anole faunas of the large is-lands were emplaced all at once, and diver-sified there on ‘‘the remnants of the oncemore-or-less continuous land connection (theproto-Greater Antillean block) that originallylay between North and South America duringlate Cretaceous to early Tertiary.’’

Although Rosen (1975) thought that fossilevidence would prove to be of great impor-tance for testing vicariance, at that time theAntillean paleontological record was of littlevalue for testing hypotheses of vicariance be-cause it was almost exclusively Quaternary(MacPhee and Wyss, 1990). In the yearssince Rosen wrote, the vertebrate fossil re-cord has improved marginally, but not to theextent that it can provide a critical test. Asnoted by MacPhee and Wyss (1990), a strongtest of vicariance would require the discov-ery of (1) numerous representatives of con-tinental lineages (2) of the same or similarage that are (3) not simply early members ofclades represented in the Antillean Quater-nary.

As to (1) and (2), the discovery that a rhin-ocerotoid perissodactyl (Hyrachyus sp.) livedon an Early Eocene landmass now incorpo-rated into present-day western Jamaica(Domning et al., 1997) is certainly importantbecause it establishes that Rosen’s mecha-nism (continent–island vicariance) may in-deed occur. However, to date this discoveryremains the only fossil-based evidence forRosen-style vicariance in the Antillean ver-


tebrate record.5 However fascinating its dis-covery may be on other grounds, one Eocenerhinocerotoid from Jamaica is of limited ex-planatory significance. If this species livedon a terrane that was part of Central Americaprior to the latter’s separation and amalgam-ation with other terranes represented in mod-ern Jamaica—as is indeed possible, given itssubstantial age—then it qualifies as a vicar-iantly emplaced taxon in the very sense thatRosen (1975) intended. There may be moresuch taxa; the only way to find out is to pros-pect in the correct contexts on other islands.

Other mammalian fossil discoveries madein recent years include a ?megalonychidsloth femur from an Early Oligocene contextin southwestern Puerto Rico (MacPhee andIturralde-Vinent, 1995), a possible insecti-vore in Early Miocene Dominican amber(MacPhee and Grimaldi, 1996; see also Itur-ralde-Vinent and MacPhee, 1996), and vari-ous remains attributable to a platyrrhine pri-mate, a capromyid rodent, and another me-galonychid from Early Miocene Domo deZaza, central Cuba (MacPhee and Iturralde-Vinent, 1994, 1995). Although these discov-eries significantly extend the insular recordsof several higher-level mammalian taxa tothe early Neogene/late Paleogene, all of themlie within clades that survived into the An-tillean Quaternary and are therefore not un-ambiguously representative of a ‘‘continen-tal’’ faunal aliquot. Reptile fossils have beenfound in Tertiary contexts on the islands, butfew have been adequately published.

PASSIVE OVER-WATER DISPERSAL:MODEL OF HEDGES AND

CO-WORKERS

In several recent papers, Blair Hedges andhis colleagues (Hedges et al., 1992, 1994;Hedges 1996a, 1996b) have argued that anal-ysis of immunological distances among a va-riety of reptiles and amphibians reveals thatoverwater dispersal has been by far the most

5 Additional, albeit nonvertebrate, fossil evidence forvicariance is provided by the diverse ant fauna recoveredfrom Early Miocene Dominican amber. This fauna hasa more continental character than the extant Hispaniolanfauna and includes forms never known to occur on oce-anic islands (see commentary by Mayer and Lazell,1988: 1477).

important mechanism of faunal formation inthe West Indies. A centerpiece of their ar-gument is that ‘‘all groups examined hadlower estimates of divergence than would bepredicted by proto-Antillean vicariance, sug-gesting an origin by over-water dispersal inmid- to late Cenozoic’’ (Hedges, 1996a: 97).Although hypotheses other than Rosen-style,mid-Cretaceous vicariance are briefly refer-enced (e.g., that of MacPhee and Iturralde-Vinent, 1994, 1995), discussion is otherwiseessentially bipolar: if classic continent–islandvicariance can be rejected, it seems, over-water dispersal must be correct. With respectto historical arguments no such certainty isever possible, and this is decidedly the casewith Antillean vertebrate colonizations. Thelater papers in the series by Hedges and co-workers introduce minor updates of the al-bumin data and their interpretation. Graphicrepresentations of times of origin also differin various, sometimes subtle ways (cf. Hed-ges et al., 1992: fig. 1; Hedges et al., 1994:fig. 2; Hedges, 1996a: fig. 2).

Page and Lydeard (1994) have compre-hensively discussed a number of systematicand interpretative issues raised by Hedges etal. (1992); Hedges et al. (1994) should beconsulted for replies to their criticisms. Muchof the discussion in these papers is beyondthe scope of the present investigation andwill not be summarized here. However, otherpoints are clearly pertinent. These may begrouped under three headings: (1) prelimi-nary issues, (2) sources of error in estimatingtimes of lineage origins, and (3) passivetransport and Cenozoic surface–current pat-terns.

PRELIMINARY ISSUES

Hedges’ basic database is impressive, butpartitioning it in different ways brings outsome of the underlying uncertainties and am-biguities in his presentation of the informa-tion. Some of these are detailed in the fol-lowing paragraphs, but others would requirea larger investigation than we have the com-petence to pursue.

(1) Number of lineages analyzed. Hedges(1996a: 113) concluded that the ‘‘major find-ing of this analysis is that all but one or twoof the 77 independent lineages of amphibians


and reptiles . . . [sampled for] the West Indiesapparently originated by dispersal in the Ce-nozoic.’’ In fact, age-of-origin estimates arespecifically made for only 72 lineages, not77; we use the corrected figure in computingproportions in subsequent paragraphs. Line-ages lacking such estimates are Hyla heilpri-ni, Phyllodactylus wirshingi, Mabuya lineo-lata, the Leptotyphlops bilineata group, andGeochelone sp. (described as ‘‘unknown’’ inthe text although his table 3 presents an es-timate of ‘‘0–2?’’ Ma). Lineages that cannotbe dated as to origin cannot contribute to theargument that their emplacement must havebeen essentially random with respect totime.6

(2) Mixture of morphological and immu-nodiffusion data. Although Hedges et al.(1992) presented ID evidence for divergencetime for a number of taxa, for many other—more than 40 (56%) in Hedges’ (1996a) mostrecent analysis—ID data are not providedand morphological divergence is used insteadas a proxy measure. Although these latterdata are separately analyzed, their valueseems to us to be incidental to demonstratingthe validity of the main argument (i.e., thatRosen-style vicariance is falsified by the lackof temporal patterning in divergence times asestimated by the ID data). Since there is nolinear clock that can be applied to rates ofmorphological divergence, estimated timesof lineage origins can only be expressed inthe broadest terms. Thus Gymnophthalmuspleei, an endemic Lesser Antillean teiid, isviewed on morphological grounds as a closerelative of northern South American G. li-neatus. Time of origin (by dispersal) is listedas 0–45 Ma, the lower limit being based ex-clusively on an estimate of the ‘‘geologic or-igin of the Lesser Antilles in the Eocene

6 Gregory C. Mayer (personal commun.) has calledour attention to three errors in enumerated entries inHedges’ (1996a) table 3: (1) 11, Crocodylus intermedi-us, known from only one or two vagrant individuals,cannot be considered to be established in the West In-dies; (2) 26, Iguana iguana does not occur on the Cay-man Islands; and (3) 41, Mabuya bistriata is presumablya lapsus for Mabuya mabuya; M. bistriata is a Brazilianspecies that is unlikely to be conspecific with anythingin the West Indies. Dropping these taxa from the list willalso affect the count, but we have not made this correc-tion here.

(Pindell and Barrett, 1990)’’ (Hedges, 1996a:104). An analogous induction is made for theteiid Cnemidophorus vanzoi of the Maria Is-lands (St. Lucia). Origin bandwidths of thismagnitude—essentially equivalent to all ofthe post-Paleocene Cenozoic—could be con-sistent with practically any biogeographicalhypothesis, not just dispersal.

(3) Taxa are not discriminated in terms ofinterpretative significance. Even if it wereaccepted that the leading cause of faunal for-mation in the West Indies was overwater dis-persal, this phrase as usually understood cov-ers several quite different mechanisms. Themajority of lizards, for example, would pre-sumably not be capable of dispersing bylong-distance swimming (i.e., self-poweredoverwater dispersal, unassisted by rafting,palm ‘‘boot’’ transport, or other classicallyinvoked means). By contrast, species of croc-odilians (Crocodylus) and chelonians (Geo-chelone, some pelomedusids) that are able totolerate saltwater conditions could have at-tained their known distributions under theirown power, whether their swimming was di-rected or not. Their propagules could havedispersed by rafting as well, but includingtaxa that are inherently ambiguous as toprobable method of dispersal adds nothing toresolution of the debate. The pertinence ofthis point becomes obvious when viewed inlight of the problems with Hedges’ (1996a)evaluation of surface-current flow in the Ca-ribbean Sea (see below).

(4) Overrepresentation and ambiguoussignificance of nonendemics. The bulk ofHedges’ (1996a) taxon list (37/72 5 51%)consists of lineages that are defined as ‘‘non-endemics’’ (species having populations onthe mainland as well as one or more WestIndian islands). Hedges assumed that the ex-istence of allopatrically distributed popula-tions of single species is good evidence thatsome island colonizations occurred so re-cently that source populations have not dif-ferentiated from colonizing ones. From a di-versity standpoint, the number of recent (0–2 Ma) lineages is certainly overstated relativeto their importance. However, from a mech-anism standpoint, if each colonization werean event separate from every other, the rateof dispersal must have been higher in the re-cent past (assuming that extinctions can be


ignored; see below). Even so, the mechanismof emplacement need not have been alwaysthe same. For example, although continent–island vicariance can be excluded from con-sideration, for very recent introductions de-termining whether range extensions were dueto natural dispersal or to anthropogenic trans-port cannot be settled merely by observingthat the latter is assumed to be less likely.Hedges (1996a: 99) noted that lineages ‘‘thatclearly are the result of human introductionare mentioned but not treated in the analy-ses.’’ However, taxa that are ambiguous inthis regard, such as Gonatodes albogularis,Hemidactylus haitianus (5 H. brookii haiti-anus of other authors), and H. mabouia, arenevertheless included by implication in var-ious statements to the effect that ‘‘nearlyall,’’ ‘‘99%,’’ and ‘‘virtually all’’ lineagesoriginated by natural overwater dispersal.The degree of human involvement in pro-ducing some herp distributions will probablynever be known except by determining fromfossil evidence that the taxon existed in theWest Indies well before the earliest presumeddate of human arrival (ca. 7000 yrbp; Burneyet al., 1994). Currently, however, the pale-ontological record is of little assistance inthis regard as few of the herp taxa listed byHedges as having a Quaternary origin haveindependent, associated radiometric datesfrom sites in the West Indies (cf. Morgan andWoods, 1986; MacPhee et al., 1989). Al-though it can be argued, trivially, that in-stances of anthropogenic dispersal still countas dispersals, in a paper designed to test sce-narios of faunal formation it is surely prob-lematic to overrepresent such taxa in the da-tabase.

(5) Low number of nonendemic lineagesin the Greater Antilles. According to the datapresented by Hedges, the great majority ofnonendemic herp lineages live on islandsother than the Greater Antilles. Setting asidecrocodiles and tortoises, only 6 of 33 Qua-ternary nonendemic lineages listed by Hed-ges (1996a) have distributions that include acontinental mainland and one or more of theGreater Antilles (Gonatodes albogularis,Hemidactylus haitianus, H. mabouia, Iguanaiguana, Mabuya ‘‘bistriata,’’ and Nerodiaclarki). Of these, at least three are of ques-tionable pertinence given the ambiguity of

their method of transport to the Greater An-tilles (see above). However, in terms of his-torical biogeography, the real issue is thecomparative lack of nonendemic populationsin the Greater Antilles in contrast to the Less-er Antilles and islands on the fringe of theCaribbean Sea. How can this pattern be ex-plained?

Not knowing whether the six Greater An-tillean herp lineages noted by Hedges wererepresentative of many more that he couldhave cited, we examined all distributions list-ed by Schwartz and Henderson (1991) intheir exhaustive catalog. In addition to thosealready discussed, and resolving cases of tax-onomic doubtfulness in Hedges’ favor, thereare only eight other apparent nonendemics(all saurians) whose distribution includes amainland and one or more of the Greater An-tilles (five species of Anolis and three ofSphaerodactylus). We cannot comment onwhether any of these may qualify as addi-tional ‘‘independent lineages’’; judging fromHedges’ comments under individual genera,he seems uncertain as well. However, itseems reasonable to conclude that the num-ber of ‘‘Quaternary dispersals’’ to the Great-er Antilles counted according to Hedges’methodology amounts to only a fraction ofthose that can be enumerated for the otherislands, even though the Greater Antillescomprise 90% of the land area in the WestIndies. It may be that the herpetofaunas ofsmall islands, with their low endemism, in-dicate that dispersal is possible; but the fau-nas of the larger islands, with their very highrates of endemism, show that colonization ishard. But the cays of Cuba do not seem tohave gained unusual numbers of nonendem-ics by comparison to the main island, despitethe fact that some of the cays are also closeto a continental margin (e.g., Archipielago deSabana-Camaguey; fig. 1).

(6) Unknown shaping influence of extinc-tion. Finally, if, as Hedges’ interpretation ofhis data implies, it were possible for manycontinental taxa to emit successful propa-gules during the past 2 Ma, then it shouldhave been possible throughout the Cenozoic.That is, it should have been possible if, atany and all times during the last 65 Ma, thelikelihood of passive dispersal was equipo-tential. Hedges (1996a: 116) addressed this


point indirectly, noting only that, for theWest Indies in general, the ‘‘large number of[Quaternary] lineages almost certainly is anartifact due to fewer extinctions expected forrecent lineages.’’ Once again, we suggest,fossil evidence will turn out to be criticallyimportant for determining whether massivefaunal turnovers or taxon cycles occurred onthese islands.

SOURCES OF ERROR IN ESTIMATING

TIMES OF LINEAGE ORIGINS

Hedges et al. (1992, 1994; see also Hedg-es, 1996a, 1996b) allowed that their estima-tions of divergence times for investigated lin-eages could be subject to three sources oferror. The first two are concerned with theinterpretation of the immunological datathemselves; they concern the effect of dif-ferences in reciprocal estimations of IDs be-tween taxa, and interclade variablity in ratesof albumin evolution. Hedges and co-work-ers discounted the first on the ground thaterrors (if any) would not lead to consistent(unidirectional) underestimation of IDsacross many groups. In discounting the sec-ond, they acknowledged that rate variabilityexists, but at such a low level (they cite 10–15% as a reptile maximum) that it wouldhave no effect on the kinds of interpretationthey pursued in their paper. Their certaintythat these potential sources of error affectedtheir conclusions only marginally is open tochallenge (Page and Lydeard, 1994), but weare particularly interested in the underpin-nings of their third source of potential error.Hedges et al. (l992: 1910-1911) acknowl-edged that

. . . the species used here may not be representativeof the most recent divergence event between the lin-eages examined (i.e., a member of the mainland taxonclosest to the island taxon examined inadvertentlywas not used); this type of error always will result inan overestimation of the time of lineage divergencefor the taxa from different land masses. If this sys-tematic error could be corrected, some distances re-ported here could only be lower, further suggestingdispersal as the primary mechanism for vertebratecolonization of the West Indies.

We take this passage to mean that it is notcrucial that actual sister taxa (hereafter, ‘‘ex-act sisters’’) be used, so long as (by impli-cation) the taxa compared are indeterminate-

ly near relatives. To restate their main points,and assuming the valency of their molecularclock model, we are told that (1) ‘‘systematicerror’’ of this kind must always be unidirec-tional (i.e., will always result in overestima-tion of time since divergence), and (2) anycorrections resulting from the discovery andutilization of exact sisters would only shortendivergence times. We explore the implica-tions of these points by means of simple phy-logenetic scenarios superimposed on equallysimple biogeographical ones (fig. 9A, B).

In figure 9A, imaginary taxa A–F areknown to be phylogenetically related in themanner indicated in the small cladogram in-set on the lower right. The groups of mono-phyletic terminal taxa are holophyletic; thereare no unknown (undiscovered) taxa abovethe lowermost node. Taxa A, C, and F areextant (solid stems). Taxa B, D, and E areextinct (outline stems). The main cladogramis superimposed on two landmasses (‘‘Main-land,’’ ‘‘Island’’) separated by an expanse ofwater (suggested by double-headed arrow).The same conventions are followed in figure9B, except that the distribution of taxa on thelandmasses is different.

In figure 9A, one major clade, of which A,B, and C are the terminal taxa, was alwaysresident on Mainland. (The sense of time istoward top of page, as indicated by verticaltime line on right.) At a specific point intime, the other major clade, of which D, E,and F are the terminal taxa, came to be iso-lated on Island. The diagram is purposelyambiguous as to whether the basal split oc-curred though dispersal or vicariance. Allthat is assumed is that a barrier separatedpopulations of the last common ancestor ofall terminal taxa, interrupting gene flow. As-sume now that albumins of A and F (taxa inboxes outlined in heavy black) are comparedwithin the same analytical framework as theone followed by Hedges et al. (1992) andyield an ID of 100, for the sake of this ex-ample. In this case, the geological age of thebasal split between the branches ending in Aand F, as estimated by ID (open circle ontime line), is for certain the same as the ap-pearance of the stipulated water barrier be-tween populations of the last common an-cestor (closed circle). Such a situation, inwhich ID time, cladogeny, and the event of


Fig. 9. Splitting lineages vs. splitting lands: Limitations on interpretation (see text).

interest all intersect on the same graph(dashed line on inset cladogram), may be de-scribed as one of congruence.

In figure 9B, all initial factors are thesame, except that extinct taxa D and E wereMainland residents. Only the branch repre-sented by terminal taxon F was isolated onIsland. If the extinct taxa were not known, itwould be assumed that taxa A and C consti-tute the sister group of F. Measured IDwould still be 100 (between A and F or Cand F), but it would not date the appearance

of the barrier between F and its closest exactsister, the clade terminating in taxa D and E(hence ‘‘?1D’’). The method is blind to latersplits (including the one that installed the an-cestor of F on Island) because the criticaltaxa are extinct and cannot therefore be sam-pled. The result is one of incongruence.

In answer to criticisms by Page and Ly-deard (1994), Hedges et al. (1994: 48) statedthat errors in taxonomic sampling do notmatter, because overestimates are always uni-directional, and furthermore ‘‘whether the


IDs coincide or not would not affect thatconclusion.’’ However, it actually does mat-ter, because filling a matrix with overesti-mates can obscure whatever pattern—includ-ing any concentration of splits—that may ex-ist within the phylogeny (or sets of comparedphylogenies). In their original figure illus-trating times of origin (Hedges et al., 1992:fig. 1), for example, the ‘‘West Indies–Main-land’’ row is filled with a flock of ID mea-surements distributed across the range 55–32Ma. Some cluster fairly tightly around 28–24 Ma, while others group at approximately18 and 11 Ma. In a later version of this figure(e.g., Hedges et al., 1994: fig. 2), which in-cludes errors of the estimates, no pattern isevident, and the lack thereof (i.e., apparentrandomness) is taken to be evidence for over-water dispersal of relevant taxa/lineages.However, this conclusion is also unjustified,because they do not know (as they acknowl-edge) whether they are judging distances be-tween exact sisters, i.e., whether congruenceobtains. In the context of the present paper,it would be especially important to know ifthere was a concentration of splits during theearlier part of the Oligocene, when we thinkpaleogeographical conditions were especiallyfavorable for overland colonization (seeGAARlandia Landspan and Island–IslandVicariance). And indeed there could be, ifthe pre-28 Ma splits represent overestimatesof just the sort Hedges et al. (1992) deemunimportant.

We take the point that validly identifiedpost-Oligocene splits cannot be explained byeither continent–island vicariance or thelandspan hypothesis (see below), althoughsome intra-Caribbean splits could be relatedto island–island vicariance events in the Ear-ly and Middle Miocene. Some post-Oligo-cene lineage origins were probably occa-sioned by overwater dispersals, although fortheir identification using Hedges’ methodol-ogy, much would depend on whether all oftheir molecular clocks ticked true all thetime. Nevertheless, the timing of splits esti-mated from albumin divergences cannot betaken as proxies for dating actual coloniza-tion events unless one knows the true branch-ing sequence of a phylogeny (cf. Hedges etal., 1994: fig. 1).

PASSIVE TRANSPORT AND CENOZOIC

SURFACE-CURRENT PATTERNS

SURFACE-CURRENT PATTERNS AND FLOTSAM

DISPERSAL: Hedges (1996b: 186) contendedthat ‘‘an overwhelming majority (99%)’’ ofall ‘‘independent lineages’’ of Antillean ver-tebrates originated from dispersants arrivingas passively transported ‘‘flotsam,’’and that,furthermore, this ‘‘dispersal pattern can beexplained by the nearly unidirectional currentflow [in the Caribbean Sea] from the south-east to the northwest, bringing flotsam fromthe mouths of South American rivers (e.g.,Amazon, Orinoco) to the islands of the WestIndies.’’ Within the Caribbean itself, taxicsimilarity among nonvolant faunas of differ-ent islands is to be explained by subsequentinterisland dispersals occurring in the samemanner, i.e., passive dispersal by surface cur-rents.

It is critical to note that Hedges’ (1996a,1996b) argument implicitly requires that themodern surface-current pattern, in which av-erage flow is indeed ‘‘almost unidirectionallyfrom southeast to northwest,’’ obtained forthe whole period in which the Antillean fau-na was being formed (i.e., virtually the entireCenozoic acording to Hedges’ own analysis).The only causative mechanism for the ‘‘al-most unidirectional’’ flow that Hedges(1996a, 1996b) refers to, however, is the Cor-iolis effect, i.e., the tendency for the trajec-tories of moving objects on the earth’s sur-face to be deflected either to the right (north-ern hemisphere) or the left (southern hemi-sphere). He gives little attention to theinfluence of varying paleogeographical con-figurations of the Caribbean region on cur-rent flow. As we show in this section, presentevidence supports the conclusion that the ex-isting pattern of surface currents within theCaribbean Sea is characteristic only of thelast 4 million years. Before that, other pat-terns predominated. Some of these, we argue,are incompatible with the history of faunalemplacement in the Caribbean region as en-visaged by Hedges (1996a, 1996b).

SURFACE-CURRENT PATTERNS AND PALEO-GEOGRAPHY: Figure 10 compares reconstruct-ed patterns of marine surface currents for theLate Eocene–Oligocene (35–32 Ma), LateOligocene/Middle Miocene (30–14 Ma),


Fig. 10. Paleoceanography of Caribbean region for selected intervals between latest Eocene andRecent, modified from Mullins et al. (1987), Duque-Caro (1990), and Droxler et al. (in press). Thepurpose of these reconstructions is to depict possible effects of GAARlandia and other features onsurface-current patterns during the latter part of the Cenozoic (MacPhee and Iturralde-Vinent, 1995;Iturralde-Vinent et al., 1996a). Since the latest Eocene the course of westward-flowing currents in themid-Atlantic and Caribbean area has been greatly affected by the appearance and disappearance ofvarious land barriers. Thus, the Circumtropical Current, which originally passed into the western Pacific,was temporarily disrupted in the latest Eocene and Early Oligocene by the emergence of GAARlandia,and was permanently disrupted by the completion of the Isthmus of Panama in the Late Pliocene. Thiscurrent has played a leading role in mediating climate in the Caribbean region during the Cenozoic.Gulf coast of North America not intended to be paleogeographically accurate.


Late Miocene (6–5 Ma), and Pliocene–Ho-locene (4–0 Ma) (see also figs. 6–8). Thesereconstructions, slightly modified from thosedeveloped by Mullins et al. (1987), Duque-Caro (1990), and Droxler et al. (1998), at-tempt to consider the effects of GAARlandiaand other recently described paleogeograph-ical features on surface currents (MacPheeand Iturralde-Vinent, 1995; Iturralde-Vinentet al., 1996a).

Latest Eocene to Early Oligocene (35–32Ma): During this interval, the surface-currentpattern in the Caribbean region would havediffered radically from patterns that obtainedbefore or after this time. There are two pa-leogeographical reasons for this: the exis-tence of the Panamanian Seaway, connectingthe Caribbean Sea with the Pacific Ocean;and the conformation of GAARlandia, thenat its subaerial maximum. The effect of thesetwo features would have been to enhancecommunication with the Pacific while im-peding it with the equatorial Atlantic (Don-nelly, 1989b). Connection with the Atlanticwould have continued at higher latitudes, viathe Yucatan and Havana–Matanzas Chan-nels. However, throughput from the Carib-bean Sea to the Atlantic via the Strait of Flor-ida seems to have been rather low (Iturralde-Vinent et al., 1996a). Because of the config-uration of land masses, the Gulf Streamwould have been fed mostly by the north-westerly current flowing parallel to Bahamas.In South America, the early Magdalena andOrinoco Rivers would have flowed directlyinto the the Caribbean Sea (Hoorn et al.,1995).

Under such conditions of at least partialisolation, there was probably a very low rateof water circulation across different portionsof the Caribbean Sea. Sub-basins such as theGulf of Mexico may have behaved as large,semi-independent embayments, in which cy-clonic circulation patterns dominated, withseasonal changes in direction of flow. Theseinterpretations are supported by two majorlines of evidence. Seismostratigraphic studies(Mullins et al., 1987; Denny, 1992; Hine,1997; Denny et al., 1994) carried out in thesouthern part of the Strait of Florida indicatethat the West Florida ramp was in an aggra-dational regime from the Late Eocene to theLate Oligocene, with little winnowing and no

evidence of erosional surfaces. Also, the rateof pelagic sedimentation in this period wasextremely low. (See description of Holes 998and 999, Leg 165, by Sigurdsson et al.[1997]; Oligocene sediments are absent insome DSDP cores [Edgar et al., 1973].)

Late Oligocene to Middle Miocene (30-14Ma): Water circulation patterns were sub-stantially altered by paleogeographicalchanges during mid-Oligocene/Middle Mio-cene highstands. The seaways to the Pacificand northern Atlantic persisted, and the Ori-noco and Magdalena still drained into theCaribbean Sea (Hoorn et al., 1995). At thebeginning of this period, communication be-tween the Pacific and equatorial Atlanticwould have been greatly improved followingthe subsidence and inundation of much of theAves Ridge, as indicated by the existence ofshared foraminiferal assemblages on the Pa-cific and Caribbean sides of the Panamanianseaway (Duque-Caro, 1990). Warm water de-rived from the equatorial Atlantic wouldhave been more widely distributed by theCircumtropical Current (Mullins et al., 1987;Droxler et al., 1998), now no longer pre-vented by southern GAARlandia from fullycommunicating with the Caribbean Sea. In-fusion of warm Atlantic water may have in-fluenced regional climate, as it did elsewere(Tsuchi, 1993). Within the Caribbean Sea it-self, a minor current headed northwest topush warm water through the Havana–Ma-tanzas Channel and Yucatan Channel, there-by contributing to the early Loop surfacecurrent in the Gulf of Mexico (Iturralde-Vi-nent et al., 1996a). However, flow in the ear-ly Loop was not as strong as that of the mod-ern system (Mullins et al., 1987; Hine, 1997).These new water masses had a pronouncedsedimentological effect in the southern partof the Strait of Florida: the West Floridaramp began to prograde from the east, as il-lustrated by west-dipping clinoforms.

Later, in the Middle Miocene, rejuvenationof tectonism modified the pattern of watercirculation by redistributing the flow carriedby the Circumtropical Current (Iturralde-Vi-nent et al., 1996a). In the Late Miocene, up-lift of the Andes drastically altered the fluvialpattern of northern South America, eventu-ally causing the Orinoco to empty into theAtlantic (Hoorn et al., 1995). Foraminiferal


assemblages in sediments of late Middle toLate Miocene age in the Atrato Basin on thePacific side of Colombia exhibit an over-whelmingly ‘‘Californian’’ aspect (Duque-Caro, 1990), suggesting that at this time therewas little throughput from the Caribbean sideof the Panamanian seaway. Nevertheless, im-pairment of circulation across the seawaywas evidently short-lived, because Late Mio-cene to Pliocene rocks in the Atrato basincontain assemblages having a mixed Pacific/Caribbean aspect (Duque-Caro, 1990).

Late Miocene (6–5 Ma): In the Late Mio-cene (fig. 10), water circulation within theCaribbean Sea may have been diverted to thenorthwest as a result of (1) reduction in thewidth of the Panamanian seaway as the isth-mus advanced southwards; (2) extension andsubsidence of the Nicaragua Rise, which hadpreviously impeded flow into the YucatanChannel (Mullins et al., 1987; Droxler,1995); and (3) closure of the Havana–Matan-zas Channel, thereby forcing all northwest-bound flow into the Yucatan Channel (Itur-ralde-Vinent et al., 1996a). The increasedvolume of water diverted into the Loop Cur-rent modified depositional conditions in theStrait of Florida. Erosion took place on thePourtales Terrace (Gomberg, 1974) and theMiami Terrace (Mullins and Newmann,1979), and there was a 50% decrease in therate of sediment accumulation in the south-ern Strait of Florida (Austin et al., 1988). Themodern Loop Current/Gulf Stream circula-tion was initiated and carbonate depositionchanged drastically, becoming a pelagicslope-front-fill system (Mullins et al., 1987).

Pliocene to Holocene (4–0 Ma): The lastmap in this series illustrates the modern pat-tern of water circulation (Atlas Nacional deCuba, 1970; Emery and Uchupi, 1972). Thispattern was initiated around the Miocene–Pliocene boundary, following the completeclosure of the Panamanian waterway and thetermination of the Circumtropical Current.Its successor, the Caribbean Current, is feddirectly by the Atlantic equatorial current;surface currents are now mostly directed to-ward the northwest. Like other changes dis-cussed in this section, this final series of cur-rent reorganizations must have had profoundeffects on terrestrial climates as well (Frakes,1979).

SURFACE-CURRENT PATTERNS AND PROXY

DATA: A current’s average vector is not theonly influence operating on material passive-ly carried along the sea surface, as experi-mental data from free-floating objects illus-trate. Molinari et al. (1979) and Kinder(1983) employed satellite telemetry to track23 free-drifting buoys released from stationsin the Lesser Antilles (fig. 11A). After wan-dering widely, often for several months,bouys ended up not only along or near thecoasts of Jamaica, Cuba, Hispaniola, PuertoRico, and several other Caribbean islands,but also those of Central America, Yucatan,and the Gulf of Mexico. The point here isnot that the free-floating buoys achieved acertain distribution, but instead that the av-erage gross direction of surface currents pro-vided only the most general guide to theprobable movement of any given buoy. Fur-ther, the path followed by each buoy wassubstantially affected by local eddies, deepcyclonic water circulation, storms, and otherevents (see Molinari et al., 1979; Kinder,1983; Kinder et al., 1985; Sou et al., 1996).Significantly, such forces generally acted toincrease rather than decrease trip length, animportant consideration in evaluating thelikelihood of successful overwater dispersalfor animals that are not physiologicallyadapted for temporally lengthy sea journeys(such as land mammals).

Brucks’ (1971) drift bottle study yieldedbroadly similar results. Brucks released bot-tles at various locations off the Windward Is-lands (southern Lesser Antilles), Gulf ofHonduras, and southwestern Caribbean(northern coast of Panama). In the WindwardIslands release area, bottle recovery recordsindicated that surface currents are cyclonic,but with a definite northward trend (insteadof a due westward trend as stipulated in pre-vious literature). Bottles came ashore in thenorthern Windward Islands and a wide vari-ety of other locations (fig. 11B, table 5) dis-tributed throughout the Caribbean Sea andGulf of Mexico. Of interest is the fact thatthe greatest number of reports (38% of total)were from Central America; only 15% werefrom the Greater Antilles (as much as Yu-catan alone). Calculated minimum speeds ofbottles varied from 0.1 to 2.0 knots, withfaster rates of movement in summer and mid-


Fig. 11. Surface-current patterns in the Caribbean Sea and nearby waters, as revealed by bouy anddrift bottle experiments. A. Tracks of bouys released by Molinari et al. (1979) between October 1975and June 1976 (redrawn from original). Arrows have been added to pinpoint last recorded positions ofselected bouys. Results of this experiment demonstrate that ‘‘average’’ marine-current patterns, as rou-tinely depicted on oceanographic maps, fail to capture substantial variance in surface-water movementin the Caribbean Sea. B. Surface-current pattern derived from drift bottle experiments conducted byBrucks (1971, modified from original). Bottles were released from sites in the eastern Caribbean. Themajority of returned bottles transported to points west of the Lesser Antilles were deposited along theeast coast of Central America, Gulf of Mexico, and Florida. Relatively few bottles were returned fromthe Greater Antilles. These results strongly imply that, given existing surface-current patterns, flotsamemitted from the Orinoco and Amazon Rivers is much more likely to end up in southeastern NorthAmerica or Central America than in the Greater Antilles.


winter than in spring and fall. These resultsare in good agreement with work on the dy-namic sea-surface topography of the Carib-bean Current, which subdivides into a faster-moving southern current and a slower north-ern one (Duncan et al., 1982).

The relatively slow speed at which pas-sively transported objects move under aver-age conditions in the West Indies is furtherillustrated by recent work on passively dis-persed pelagic larvae (e.g., Roberts, 1997).These investigations indicate that distancesactually or potentially travelled by such lar-vae during one- and two-month ‘‘transportenvelopes’’ are relatively small. For exam-ple, the two-month envelope for larvae dis-persing in the northeastern Caribbean spansonly the distance from Anguilla to easternPuerto Rico. A route crossing the entire Ca-ribbean Sea from southeast to northwest, asenvisaged in Hedges’ model, would presum-ably take much longer. Except under unusualcircumstances, therefore, it would appear thatmodern surface currents in the Caribbean Seado not flow rapidly enough to ensure the sur-vival of terrestrial amniotes dispersing byrafting (with the possible exception of somereptiles).

This limitation might be overcome by hur-

ricane transport, and there is at least one re-cent case in which this is the only feasibleexplanation for a dispersal event. A numberof individuals of Iguana iguana, a specieswhich does not occur in Anguilla, were con-clusively identified on that island in themonth following the passage of HurricaneLuis across the northern Lesser Antilles inAugust, 1995 (Censky et al., 1998). It isprobable but not demonstrated that the dis-persants came from Guadeloupe or one of theother islands south of Anguilla that supportIguana iguana. Although northward trans-port of propagules by hurricanes might ex-plain some aspects of Lesser Antillean sau-rian biogeography, modal hurricane tracksmay well have differed in previous epochs.Before the completion of the Panamanianisthmus, for example, storm tracks may havebeen deflected relatively southward by thewarm Circumtropical Current flowing intothe Pacific. Whether this occurred or not hasnot been properly modelled, but any south-ward deflection of high-energy storms wouldhave reduced the chances of successful dis-persal to the Greater Antilles.

Tracking studies provide another body ofrelevant proxy data. Mean sediment concen-trations in sea water can be tracked and mea-sured using satellite imagery. In one suchstudy (Richardson, 1996), it was found thatwater discharged by the Amazon does notalways move northwestward into the LesserAntilles, but is sometimes forced to floweast–southeast. This finding is in agreementwith the fact that the North Equatorial Coun-tercurrent carries much of the Amazon’s out-flow eastward into the central Atlantic (Rich-ardson, 1996). Interestingly, sediment track-ing studies have also shown that outflowfrom the Orinoco is normally directed to thenorthwest, where it can be detected as farnorth as Puerto Rico. By contrast, outflowsfrom the Magdalena and Lake Maracaiboquickly become disorganized after enteringthe Caribbean Sea: pigment can be traced foronly a short distance before becoming dilut-ed (Richardson, 1996: fig. 21-12).

OTHER CONSTRAINTS: Hedges’ (1996a) IDdata indicate that two surviving herp lineages(Eleutherodactylus and Cricosaura typica)could have been established on landmassesin the Caribbean Sea as early as the latest


Cretaceous or early Paleogene (see alsoHedges et al., 1992, 1994). In his view, em-placement at this early date could have oc-curred through either continent–island vicar-iance or overwater dispersal. However, dur-ing the latest Eocene and Early Oligocene,circulation within the Caribbean Sea waslimited, and, as discussed above, in any casethe patterns then existing would not haveuniformly favored the movement of watermasses toward the northwest (figs. 6, 10).Further, given the paleoposition of themouths of the Magdalena and Orinoco, flot-sam issuing from these rivers would havebeen much more likely to end up on the is-lands that then comprised much of southernCentral America and the Nicaraguan Risethan, say, the Cuban/Hispaniolan end ofGAARlandia.

The same constraints would have contin-ued into the subsequent period (30–14 Ma)that we have modelled. In our view, it is like-ly that any natural rafts coming out of largenorthwestern South American rivers wouldhave been sent into the Pacific (figs. 8, 10;see also Frakes, 1979; Kennett, 1985; Mul-lins et al., 1987; Droxler et al., 1998). At thattime part of the present Amazon Basin wasoccupied by a large marine embayment(Hoorn et al., 1995). Drainage off the Guy-ana Shield would have been the only signif-icant source of flotsam directed into the At-lantic; however, from this source any pas-sively transported material would have beenas likely to drift toward Africa as the WestIndies (cf. Richardson, 1996). Similar con-ditions would have existed during the sub-sequent 10 Ma. However, it is necessary toobserve that increasing amounts of waterwould have been directed toward the north-west as Central America grew southward inthe Late Miocene. From this it follows thatHedges’ (1996a) emplacement scenariowould increase in likelihood the closer onecomes to the present. The trouble is that, ifemplaced by overwater dispersal, the speci-fied ‘‘overwhelming majority’’ of indepen-dent insular lineages would have to haveoriginated 4 million years ago or less. Hedg-es’ (1996a) table 3 may be said to underlinethis very possibility, since no fewer than 55investigated origins (76% of total) are datedto, or overlap with, the period 4–0 Ma,

which may explain the acknowledged recen-cy of so many relatively undifferentiatedSouth American herps in the southern LesserAntilles. Yet if these figures were truly rep-resentative of the vertebrate fauna as awhole, then it would be necessary to concludethat either (1) the islands had a taxically mi-nuscule vertebrate fauna prior to the Plio-Pleistocene, or (2) there was a virtually com-plete faunal turnover in the latter part of theNeogene. At least for mammals, the first pointis flatly contradicted by fossil finds made inthe Greater Antilles in the last decade: everymajor taxon of land mammals known fromthe Antillean Quaternary now has a minimumorigin date of Early Miocene or earlier(MacPhee and Iturralde-Vinent, 1994, 1995;MacPhee and Grimaldi, 1996). The possibilityof high faunal turnover, however, is not di-rectly tested by this evidence as no ‘‘unex-pected’’ taxa other than Hyrachyus in Jamaicahave been recovered (see MacPhee and Wyss,1990). For lizards the data are also unhelpful,inasmuch as recent dispersals (as evaluated byHedges) have added very little to the faunallyrich Greater Antilles, even though there mayhave been many independent events (G. May-er, personal commun.).

The proxy data for the effect of surface-current flow as a dispersal agent suggest that,on the basis of existing current patterns, theOrinoco is more likely to release flotsam thatfinds its way into the central Caribbean Seathan is the Amazon. Whether this pattern wasalways the rule in earlier times is moot, butit is surely relevant that the Amazon has beenan Atlantic coastal river only since the LateMiocene (Hoorn et al., 1995). If rivers actedas potential sources of animal-bearing flot-sam in the manner contemplated by Hedgesfor geologically long periods of time, then itis the basins of northwestern South America,rather than the Amazon itself, that should beunder consideration.

A final occasion of lack of fit betweenHedges’ model and the paleogeographical re-constructions offered here concerns the avail-ability of lands in the Caribbean Sea at pur-ported times of land vertebrate colonization.The data of Hedges and co-workers suggestthat 6 to 11 herp lineages originated earlierthan the Eocene–Oligocene transition. Al-though the number of such ‘‘early’’ lineages


is obviously small and therefore does notbear much analysis, the fact that any lineagesmay go back to the earlier Paleogene con-flicts with our view that the islands that nowexist in the Caribbean are post-Middle Eo-cene in origin (Iturralde-Vinent, 1982;MacPhee and Iturralde-Vinent, 1994, 1995).Hedges (1996b: 166) resolved this conflictby stating that ‘‘the recent suggestion thatthere were no permanently subaerial land-masses in the Greater Antilles prior to [theend of the Middle Eocene] is speculative; itcan neither be refuted nor supported withcurrent evidence.’’ We interpret the nature ofthe current evidence differently, but, notwith-standing that, we pose the following conun-drum. If evanescent islands existed in the Ca-ribbean Sea before the Middle Eocene—andit is highly probable that they did—then anyresident faunas must have either perished atthe time of inundation or subsidence, or theyfound a means to transfer to newly risenlandmasses elsewhere. Hedges (1996a) men-tioned the former possibility in relation to thesterilizing effects of giant tsunamis generatedby the K/T bolide impact, but speculated thatthere were, nevertheless, some places wherefaunal elements could have persisted into theCenozoic. As we have repeatedly asked inthis paper, if such places existed, where werethey, and what is their paleogeographical his-tory?

GAARLANDIA LANDSPAN ANDISLAND–ISLAND VICARIANCE:

MODEL OF MACPHEE ANDITURRALDE-VINENT

LANDSPANS, VICARIANCE, AND

DIVERSITY SCENARIOS

A ‘‘landspan’’ is here defined as a subaer-ial connection (whether continuous or punc-tuated by short water gaps) between a con-tinent and an off-shelf island (or island arc).Although few continuous landspans exist atpresent (e.g., Kamchatka), during Pleistoceneglaciations several significant ones were cre-ated during phases of lowered sea level.Among these were the Greater Palawan land-span connecting Bulabac-Palawan-Calami-ans with Borneo (and therefore with south-eastern Asia; Heaney and Regalado, 1998)

and the Southern Ryukyu landspan extendingfrom Taiwan at least as far as Okinawa (andtherefore joining these islands with continen-tal eastern Asia; Ota, 1998). This usagestands in contrast with that favored for‘‘landbridge,’’ here restricted to mean landlinkages between continental regions (e.g.,DeGeer, Panamanian, Beringian landbrid-ges). It should also be noted that on-shelf (or‘‘continental’’) islands are simply extensionsof continents that may function as submergedmargins, islands, or peninsulas (or often asall three over time), and they are alwaysbroadly affected by the same tectonic andisostatic regimes as their mainlands. Off-shelf islands are not by any useful definitionextensions of continents, even if they incor-porate continental fragments (e.g., Madagas-car). Off-shelf islands may be affected bymajor tectonic events that also affect nearbycontinents, but they may otherwise evolvequite independently.

We introduce this terminological refine-ment in order to differentiate between twoquite dissimilar contexts in which faunal dis-tributions can be influenced by the appear-ance of a land connection. In general, trans-fers from one faunal area to the other acrossa newly developed land connection will becontrolled by a host of factors (i.e., filters)that affect the likelihood that any given spe-cies will complete the journey successfully.However, the outstanding feature of land-bridges is that they connect areas having con-tinental-scale faunal diversity. At least in the-ory, the entire diversity of each area is avail-able for interchange (although in practice theactual number of transfers in either directionis usually much less than the theoretical max-imum). That flow actually occurs bidirection-ally is amply demonstrated by relevant pa-leontological records (cf. Webb, 1976), evenif it occurs predominantly in one direction.Landspans are markedly different becauseone terminus lacks continental-scale diversi-ty (and, indeed, may initially have no faunaat all). Although in principle any number ofcontinental faunal elements might cross anewly created landspan to colonize an islandor island chain, long-term survival after ini-tial colonization will be highly correlatedwith the availability of appropriate habitat inwhat are, after all, absolutely small places. In


Fig. 12. Schematic paleogeographical scenario for Caribbean region, middle Mesozoic to late Ceno-zoic, emphasizing permanency of terrestrial enviroments and fluctuations in land–land connections. Gen-erally speaking, light shading implies temporary lands or connections between landmasses. ‘‘Evanescentislands’’ are purely conjectural, and their number and position are not be taken literally. Late Mesozoicand Miocene landbridges between the Americas are not confirmed. Black arrows signify direction ofpropagule dispersal over land connections, as established by fossil evidence. Thus, dispersal was bidirec-tional in case of Plio-Pleistocene Panamanian landbridge (Great American Biotic Interchange), whichpermitted exchange between two large continental faunas. By contrast, in case of GAARlandia and Chortis/western Jamaica landspans, propagule movement (at least for vertebrates) was neccesarily unidirectional,as only the continental terminus possessed a fauna when the landspan was formed. Timing of majortectonic events also indicated. Chronometric scale purposely nonlinear. Interconnections existing at anygiven time can be easily visualized by placing a ruler horizontally across page.

contrast, continent-sized areas are muchmore likely to support a great diversity ofhabitat types, increasing the chances of suc-cess of a substantial variety of immigrants.

The two-part landspan/vicariance model ofMacPhee and Iturralde-Vinent (1994, 1995,this paper) attempts to infer mechanisms offaunal formation in the Greater Antilles fromdetailed paleogeographical reconstructions,fossil evidence, and species/area relation-ships. Central to the hypothesis is the argu-

ment, sustained at length in this paper, thatthe Cenozoic paleogeography of the Carib-bean region strongly favored emplacementover land (as opposed to over water) onlyonce in the past 65 Ma. (Details of Cenozicconnections among lands is schematicallypresented in figure 12.)

The first component of the model seeks toexplain how land mammals might havereached the northern Greater Antilles fromnorthwestern South America by dispersing


across a short-lived landspan during a re-stricted period in the mid-Cenozoic (ca. 33–35 Ma according to our current estimate).Specifically, it is hypothesized that, duringthe Eocene-Oligocene transition, the AvesRidge became subaerial for a short interval,possibly as short as one or two million years.At that time the islands on the northern partof the Greater Antillean Ridge (central andeastern Cuba, north-central Hispaniola,Puerto Rico, Virgin Islands) were in a close-packed array; they either constituted a single,large island, or a series of islands separatedby very narrow water gaps. By connectingproximally with the eastern end of the Great-er Antilles Ridge and distally with north-western South American microcontinent, thesubaerial Aves Ridge completed the GAAR-landia landspan. As in the case of the north-ern connection, the linkage between GAAR-landia and northwestern South Americawould have existed for only a short time (i.e.,between major occurrences of postmagmaticarc uplift in latest Eocene and general sub-sidence in Oligocene).

The second component seeks to explainhow certain distributions of faunal elementsmight have been produced via island–islandvicariance, due to the subdivision of the is-lands themselves. Mid-Oligocene and Mio-cene marine transgressions and neotectonics(Mann and Burke, 1984; Mann et al., 1990)substantially affected the disposition and pa-leogeography of the Greater Antilles Ridge.Several pull-apart basins and related featuresopened or expanded along the northern Ca-ribbean plate boundary (Cayman Trough,Mona Canyon, Sombrero Basin/AnegadaPassage), creating deep-water channels andbasins between tectonic units. Paleogeo-graphically, the end result was the gradualsubdivision of the subaerial parts of theGreater Antilles Ridge. For example, easternCuba and northern Hispaniola, physicallyconnected during the Early Oligocene, weresundered by the expansion of the WindwardPassage later in that epoch (Iturralde-Vinentand MacPhee, 1996); by contrast, the con-nection between central Hispaniola andPuerto Rico probably lasted until late in theMiocene (figs. 6–8, 12). These subdivisionswould have divided the ranges of terrestrialfaunal elements previously emplaced by dis-

persal. The ‘‘orderly,’’ multi-island distribu-tion of lower level monophyletic units ofland mammals (particularly sloths and insec-tivores) possibly supports this inference ofisland–island vicariance and ought to be test-able cladistically (MacPhee and Iturralde, inprep.; White and MacPhee, in prep.).

This model has major implications for An-tillean vertebrate paleontology. As noted ear-lier, Hedges’ (1996a, 1996b) ID data appearto show that a large proportion of the WestIndian herpetofauna originated extremely re-cently (although much of its diversity is dueto Tertiary colonizations, as his data alsoshow). By contrast, Rosen’s (1975) andMacPhee and Iturralde-Vinent’s (1994, 1995)models require that the Antillean fauna wasformed much earlier (Late Cretaceous orLate Eocene/Early Oligocene, respectively).If the overwater dispersal model is a gener-ally accurate narrative of faunal formation inthe West Indies, it follows that the faunamust have formed in an episodic, accretion-ary manner. Accordingly, one would not ex-pect to find a faunal assemblage that wasmarkedly different or systematically more di-verse than the modern one at any point in thepast (low diversity scenario). On the otherhand, if either Rosen-style vicariance or thelandspan hypothesis were correct, it wouldbe expected that diversity would be muchgreater at the crucial times during which thefauna was being emplaced (high diversityscenario). Figure 13 explores how features ofa good paleontological record might be ableto help distinguish between high-diversityand low-diversity scenarios.

In the figure, taxa a–p (left column) arethe initiators or founding species of a seriesof different imaginary clades. T1, T2, andT3 are three different time transects (for thesake of this example, T1 represents Late Eo-cene; T2, Early Oligocene, T3, Pleistocene/Holocene boundary). The cartouches, repre-senting the separate clades, are each por-trayed as having a ‘‘lifespan’’ in relation tothe time axis.

This example permits models of supposedinitial faunal diversity to be compared (seebox in upper right for shading conventions).The light gray shading represents possibleexpectations under the high initial diversitymodel (taxa 5 16). Dark gray represents ex-


Fig. 13. Speculating on the faunal history of the Greater Antilles: Significance of fossils, ‘‘ghosts,’’and incomplete evidence (see text).

pectations under the low initial diversitymodel (taxa 5 4). Gradient-shaded dots rep-resent fossil discoveries; single dots repre-sent isolated finds, numerous dots manyfinds. To make comparisons realistic, taxasurviving beyond T3 are shown as havingvery good late Quaternary fossil records (m,n, o, and p; n and o now extinct).

Scenario 1 (high initial diversity, few orno dispersals to augment original composi-tion): At about T1 a fauna of high diversity(taxa a–p, light grey cartouches) is emplacedessentially simultaneously through operation

of some common cause (e.g., vicariance,landspan). (Prior to T1, it is assumed thatthere was no fauna at all.) Later on, fauna isreduced by extinctions. By time T2, taxa a–c are already extinct; at various times there-after, taxa d–l also disappear. Taxa m–p sur-vive into the late Quaternary as the much-diminished remnants of a once-diverse fauna.

Scenario 2 (low initial diversity, faunagradually augmented by new dispersals stag-gered over a long period): In this example,taxa a–l were never part of the fauna; theentire complement at any time derives from


taxa represented by dark gray cartouches(m–p), whose initiators arrive at differenttimes (asterisks). Faunal diversity (by clade)was never more extensive than it was in theQuaternary.

Scenarios 1 and 2 cannot be distinguishedmerely by inspection of the neontologicalfauna. However, they can be critically eval-uated with fossil evidence. Recovery of sub-stantial numbers of ‘‘unexpected’’ fossils(i.e., evidence of clades not represented inthe Quaternary) would tend to favor the highinitial diversity model. Failure to find suchfossils would imply that the known, latestCenozoic level of diversity is all there everwas, and that the low initial diversity sce-nario is correct.

As noted above, recent paleomammalogi-cal discoveries in the northern Greater An-tilles (Cuba, Hispaniola, and Puerto Rico)have significantly extended the records ofseveral clades (MacPhee and Iturralde-Vi-nent, 1994, 1995; MacPhee and Grimaldi,1996). Nevertheless, all Tertiary taxa recov-ered to date from these islands appear to beclosely related to clades known from theQuaternary, which favors the low initial di-versity model. At present any firm conclu-sion would be premature, as there are as yetno fossil vertebrate sites that date to the cru-cial period in the Late Eocene when perma-nent land environments were first establishedin the northern Greater Antilles. Since thepaleogeographical history of Jamaica hasbeen markedly different from that of the oth-er Greater Antilles, it would be equally pre-mature to read anything into the remarkablediscovery of Hyrachyus on that island(Domning et al., 1997). There may be moresuch taxa, in Jamaica and elsewhere; the onlyway to find out is through intensive pros-pecting.

Specific colonization timetables are al-ways subject to refutation paleontologically,because fossils provide minimum dates ofoccupation. Thus discovery of an ‘‘early’’fossil attributable to taxon m outside the tem-poral limits of its cartouche (dark gray shad-ing) is evidence that the colonization timeoriginally inferred for m is incorrect. Fossilsdating first appearances are also of great sig-nificance if they represent numerous clades(whether expected or unexpected) and occur

within a narrow time slice (e.g., first-appear-ance fossils of taxa m–p). Multiple first ap-pearances at a specific time could be evi-dence of the operation of a common cause.(For specific examples of the use of thesecriteria in analyzing the Antillean mamma-lian colonization record, see MacPhee andIturralde-Vinent, 1995).

DISCUSSION

Most of the topics that require discussionin relation to the geology and paleogeogra-phy of the landspan hypothesis have beendealt with at length in earlier sections of thiswork. Our purpose here is to summarizewhat we consider to be the oustanding prob-lems with this hypothesis, as a guide to fur-ther work.

(1) As noted in earlier sections, we cannotyet offer detailed geological and paleogeo-graphical reconstructions of the Caribbeanarea during the Mesozoic/early Paleogene,and therefore we are unable to settle whetherlandspans/landbridges existed in this regionbefore the mid-Cenozoic (but see fig. 12 andpoint 3 below).

(2) Although geographically a member ofthe Greater Antilles, Jamaica has had a tectonichistory quite different from that of the otherislands in the group. The only tectonic unitcurrently incorporated into Jamaica that mighthave had some relationship to evolvingGAARlandia is the Blue Mountains Block. If,as some evidence indicates, the Blue Moun-tains Block lay relatively close to the northernGreater Antilles during the Cenozoic, it mayhave received immigrants directly fromGAARlandia, either over water or over a landconnection with southern Hispaniola (fig. 12).Collision of Western Jamaica with the BlueMountains Block—if such an event actuallytook place—would not have occurred earlierthan the Miocene.

(3) In principle, some propagules couldhave reached Caribbean landmasses from themainlands either before or after the landspanperiod, if colonizations occurred by over–water dispersal or continent–island vicari-ance. However, in our view, earlier coloni-zations (if they occurred) were ultimatelydoomed to failure (no permanent landmassesbefore Late Eocene), and later colonizations


would have become increasingly difficult(owing to the disappearance of landspans andtectonic dismemberment of GAARlandia).Nevertheless, two herp lineages identified byHedges (1996a, 1996b) may be examples ofearly emplacement, although there is no fos-sil evidence relating to these taxa. JamaicanHyrachyus does not constitute a counterex-ample: perched on its Viking funeral ship, itremains a Central American/North Americantaxon, despite its allochthonous presence inthe Greater Antilles (Domning et al., 1997).

Very late originations, as apparently oc-curred in the Lesser Antilles among manyherp groups (Hedges, 1996a, 1996b) and per-haps some mammals (e.g., sigmodontine ra-diation of northern Lesser Antilles and Ja-maica) cannot be explained by the GreaterAntilles landspan model. They are the resultof the operation of some other mechanism,evidently natural dispersal (or, in some cases,possibly human transport).

(4) Although island–island vicariance mayprovide a neat solution for the distributions ofseveral tightly related Quaternary taxa, wecannot account for all cases. For example, thepresence of apparently endemic species ofcapromyid rodents and Nesophontes in theQuaternary of the Cayman Islands (Morgan,1994) cannot currently be explained as a con-sequence of island–island vicariance, inas-much as the Cayman Islands are probablyvery recent geographical entities that are un-likely to have had any land connection witheither Cuba or Jamaica during the Eocene–Oligocene transition (figs. 6–8). A very lateland connection between the Cayman Islandsand Cuba might have occurred during the up-lift of the Sierra Maestra during the Plio-Pleis-tocene, as these mountains are located in thesame trend and geological unit as the Cay-mans (Perfit and Heezen, 1978; Case et al.,1984; Sigurdsson et al., 1997). However, thispossibility would require a substantial amountof subsidence during the last 5 Ma, as seafloor depths between Cayman Brac and CaboCruz (western end of Sierra Maestra) are inexcess of 1000 m (G. Morgan, personal com-mun.). It may also be noted that colonizationsby lineages that were evidently not proxi-mately South American (e.g., solenodontids,with proximate ancestry in North America orpossibly the Old World), not well constrained

by this or any other model, will have to beelucidated by additional geological, paleogeo-graphical, and paleontological research.

(5) Detailed paleogeographical reconstruc-tions of the now-submerged Aves Ridge arenot currently possible. However, as noted inappendix 1, there is evidence that portions ofthis ridge must have been subaerially elevat-ed at one time (e.g., regional pattern of post-magmatic phase uplift in the late Paleogene,thin Cenozoic sedimentary cover, presence ofOligocene and Early Miocene land-derivedconglomerates on ridge). The critical issue iswhether the uplifted, emergent Aves Ridgecould have formed a corridor of some sortbetween the Greater Antilles and northwest-ern South America. Wells drilled along theridge’s structural highs might provide the latePaleogene record still needed to meaningful-ly constrain the corridor hypothesis, althoughit cannot be predicted whether such data willbe detailed enough to determine if GAAR-landia constituted a single subaerial entity atany given time. Transitory water gaps, forexample, may have intervened for short pe-riods of time between land masses situatedon the Aves Ridge or between segments ofthe Greater Antilles. Importantly, however,none of them would have had the durationor depth of the Havana–Matanzas Channelbetween western and central Cuba. Alterna-tively, faunal elements might have dispersedin a pulsed manner within the confines of asingle event, if uplift and subsidence affecteddifferent parts of GAARlandia at differenttimes. Episodic movements of this kind, ifthey occurred, might have increased thechance of extinction of faunal elements un-able to transfer to the next landmass on thechain, further diminishing overall diversity.

(6) The degree to which the NWSA mi-crocontinent was physically separated fromthe rest of the continent by marine barriershas not been fully clarified. From the Eo-cene–Oligocene transition, the Middle Mio-cene, and especially during high sea-levelstands in the Late Miocene, Pliocene, andQuaternary a large marine embayment in thelocation of the present-day Orinoco River ba-sin isolated the microcontinent from theGuyana highlands to the east (figs. 2–4; Nu-tall, 1990; Webb, 1995; Rasanen et al., 1995;Cooper et al., 1995; Kay and Madden, 1997).


Prior to the rise of the Andes in southernColombia and Ecuador, the microcontinentmay or may not have been bounded by an-other water gap or lowland to the southwestat the location of the Guayaquil Portal(Domning, 1982; Hoorn et al., 1995; Webb,1995). Thus, to a greater or lesser degree, thenorthwestern and eastern parts of the conti-nent were at least partially isolated from oneanother for long periods of time, until com-paratively recently. These points raise an in-teresting issue. If the northwestern microcon-tinent was substantially isolated from the restof South America during the late Paleogeneto early Neogene by marine and orogenicbarriers, then only that fraction of the SouthAmerican biota occupying the northwesterncorner of the continent would have been ina position to cross over the evanescent land-span into GAARlandia. Unfortunately, weknow very little about the faunal composi-tion of northwestern South America at theend of the Paleogene (Marshall, 1985; Kayand Madden, 1997). Nevertheless, the factthat the Antillean fauna has apparently al-ways lacked representatives of many South

American groups that ‘‘should’’ have madethe journey (Simpson, 1956) may havesomething to do with the relative paleogeo-graphical isolation and faunal diversity of thesource area as well as the nature and durationof the landspan.

(7) The Late Miocene occurrence of twosloth taxa in North America and the equallyearly presence of procyonids (Cyonasua) andnow a ?cuvieroniine gomphothere in SouthAmerica (Webb, 1985; Frailey et al., 1996)may indicate that these continents experi-enced limited biotic interchange prior to theformation of the (last) Panamanian bridge inthe late Neogene. The mechanism that per-mitted this limited interchange—if that iswhat it was—is obscure; a possible drylandconnection is signalled by the 12.9–11.8 Mahiatus in the Atrato Basin during the lateMiddle Miocene (fig. 12; Duque-Caro,1990), but its existence is unconfirmed. Wedoubt that the sloths swam the distance, al-though we note that at least one extinct phyl-lophage (Thalassocnus) is thought to havebeen highly aquatic (cf. de Muizon andMcDonald, 1995).

CONCLUSIONS

(1) Number of intercontinental landbrid-ges. The last time that western Laurasia(North America) and western Gondwana(South America) were physically connectedas continental areas was during the MiddleJurassic, ca. 170 Ma. Terrestrial connectionsbetween these continental areas since thencan only have occurred via landbriges. In theCretaceous, three major uplift events, record-ed as regional unconformities, may have pro-duced intercontinental landbridges involvingthe Cretaceous Greater Antillean island arc.The late Campanian/early Maastrichtian up-lift event is the one most likely to have re-sulted in a landbridge, as it would have beencoeval with uplift of the dying Cretaceousarc. However, the evidence is too limited forany certainty on this point.

Whether the Cretaceous island arc was in-volved in the formation of a late Mesozoiclandbridge between North and South Amer-ica carries no necessary biogeographical im-plications. As we have stated, it is not un-

likely that there were islands in the Carib-bean Sea from the time of its opening in theJurassic onward. On the other hand, it is veryunlikely that any of the early islands contin-uously remained as such (i.e., as subaerialgeographical entities) into later times, due torepeated transgressions, subsidence, and, notincidentally, the K/T bolide impact and as-sociated mega-tsunamis (cf. Hedges et al.,1992).

Since the close of the Mesozoic, any land-bridge between North and South Americawould have to have involved Central Amer-ica. The existing bridge (Panamanian isth-mus) was completed only in the Plio-Pleis-tocene. Evidence for a precursor bridge latein the Middle Miocene is ambiguous at thistime.

(2) Role of GAARlandia landspan. Thereis evidence that northwestern South Americawas briefly connected during the Eocene–Ol-igocene transition with large landmassesemergent on the Greater Antilles Ridge and


Aves Ridge. The massive uplift event thatapparently permitted these connections wasspent by 32 Ma; a general subsidence fol-lowed, ending the landspan phase. Thereaf-ter, Caribbean neotectonism resulted in thesubdivision of remaining land areas, possiblycausing multiple instances of true vicarianceamong vertebrate species then present.

(3) Modes of faunal formation in theGreater Antilles. Currently, there are threemain models of faunal formation in the WestIndies of interest to vertebrate biogeogra-phers: strict dispersal, strict continent–islandvicariance, and one that combines dispersaland vicariance in a two-phase process. Thispaper reviews recent contributions to theoryrelevant to each of these major modalities.Continent–island vicariance in the classicsense of Rosen (1975, 1985) appears to beexcludable for any period since the mid-Ju-rassic; even if vicariance occurred at thattime, its relevance for understanding the or-igin of modern Antillean faunas is minimal.Hedges and co-workers (Hedges et al., 1992,1994; Hedges, 1996a, 1996b) have stronglyespoused overwater dispersal as the majorand perhaps only method of vertebrate faunalformation in the Caribbean region. Notwith-standing their well-argued case, surface-cur-

rent dispersal of passively transported prop-agules seems to us to be an ineffective ex-planation of observed patterns of faunal dis-tribution in the Greater Antilles. Even thougha general tendency exists for Caribbean sur-face currents to flow northward with respectto the South American coastline, experimen-tal evidence indicates that it is highly unpre-dictable where passively floating objectscaught in these currents will be deposited.Prior to the Pliocene, regional paleogeogra-phy was such that current-flow patterns frommajor rivers should have delivered mostSouth American waifs to the Central Amer-ican coast, not to the Greater or Lesser An-tilles. Since at least three (capromyid rodents,pitheciine primates, and megalonychidsloths) and possibly four (solenodontid in-sectivores) lineages of Antillean mammalswere already on one or more of the GreaterAntilles by the Early Miocene (MacPhee andIturralde-Vinent, 1995), Hedges’ inference asto the primacy of overwater dispersal appearsto be at odds with the facts.

The landspan model is consistent withmost aspects of Antillean land-mammal bio-geography as now known (MacPhee andIturralde-Vinent, 1995); whether it is consis-tent with the biogeography of other groupsremains to be seen.

REFERENCES

Albear, J., and M. A. Iturralde-Vinent1985. Estratigrafıa de las provincias de La

Habana. In M. A. Iturralde-Vinent(ed.), Contribucion a la geologıa de lasprovincias de La Habana y Ciudad deLa Habana, pp. 12–54. La Habana: Ed-itorial Cientıfico-Tecnica.

Algar, S., and J. Erikson1995. Correlation of the Jurassic through Ol-

igocene stratigraphic unit of Trinidadand northeastern Venezuela. Int. Geol.Rev. 37: 313–334.

Alvarado, G. E.1988. Centroamerica y las Antillas: Puente,

barrera y filtro biologico entre Norte ySudamerica (Cretacico al Presente):GEOISTMO, vol. 2(1): 9–25.

Anderson, T., and V. Schmidt1983. The evolution of Middle America and

the Gulf of Mexico-Caribbean regionduring Mesozoic time. Geol. Soc. Am.Bull. 94: 941–966.

Areces-Mallea, A.1990. Piazopteris branneri (White) Lorch,

helecho del Jurasico Inferior-Medio deCuba. Rev. Soc. Mex. Paleontol. 3(1):25-40.

Atlas Nacional de Cuba1970. Atlas Nacional de Cuba. La Habana

and Moscow: Academies of Sciencesof Cuba and USSR.

Austin, J. A., W. Schlager, A. A. Palmer, and ODPLeg 101 Scientific Party

1988. Proceedings of the Ocean Drilling Pro-gram, Initial Reports (Pt. A), Vol. 101.College Station, TX: Ocean DrillingProgram.

Balkwill, H. R., G. Rodrigue, F. I. Paredes, and J.P. Almeida

1995. Northern part of Oriente basin, Ecua-dor: reflection seismic expression ofstructures. In A. J. Tankard, R. SuarezSoruco, and H. J. Welsink (eds.), Petro-


leum basins of South America. Am.Assoc. Pet. Geol. Mem. 62: 559–571

Bartok, P. E.1993. Prebreakup geology of the Gulf of

Mexico-Caribbean: its relations to Tri-assic and Jurassic rift systems of theregion. Tectonics 12: 441–459.

Bartok, P. E., O. Renz, and G. E. G. Westermann1985. The Siquisique ophiolites, northern

Lara State, Venezuela: a discussion oftheir Middle Jurassic ammonites andtectonic implications. Geol. Soc. Am.Bull. 96: 1050–1055.

Beccaluva, L., M. Coltorti, G. Giunta, M. Iturral-de-Vinent, E. Navarro, F. Siena, and F.Urbani

1996. Cross sections through the ophioliticunits of the southern and northern mar-gins of the Caribbean plate in Venezue-la (Northern Cordilleras) and CentralCuba. Ofioliti 21(2): 85–103

Beets, D. J., W. V. Maresh, G. T. Klaver, A. Mot-tana, R. Bocchio, F. F. Beunk, and H. P.Monen

1984. Magmatic rock series and high-pres-sure metamorphism as constraints onthe tectonic history of the southern Ca-ribbean. In W. E. Bonini, R. B. Hard-grave, and R. Shagam (eds.), The Ca-ribbean–South American plate bound-ary and regional tectonics. Geol. Soc.Am. Mem. 162: 95–129.

Bellizia, A., and G. Dengo1990. The Caribbean mountain system, north-

ern South America; a summary. In G.Dengo and J. E. Case (eds.), The ge-ology of North America, vol. H, TheCaribbean region: 167–177. Boulder,CO: Geological Society of America.

Berggren, W. A., D. V. Kent, C. C. Swisher, andM.-P. Aubry

1995. A revised Cenozoic geochronology andchronostratigraphy. Soc. Econ. Paleon-tol. Mineral. Spec. Publ. 54: 129–212.

Bird, D. E.1991. An integrated geophysical interpreta-

tion of the Grenada basin. Unpub. M.S.thesis, Dep. Geosciences, Univ. Hous-ton, TX.

Bird, D. E., S. A. Hall, and J. F. Casey1993. Interpretation of magnetic anomalies

over the Grenada basin. Tectonics 12:1267–1279.

Bock, W. D.1972. The use of foraminifera as indicators of

subsidence in the Caribbean. In Trans-actions of the 6th Caribbean Geological

Conference, Caracas: 439–440. Cara-cas: Cromotik.

Bonini, W. E., R. B. Hardgrave, and R. Shagam(eds.)

1984. The Caribbean-South American plateboundary and regional tectonics. Geol.Soc. Am. Mem. 162.

Borhidi, A.1985. Phytogeographic survey of Cuba. 1.

The phytogeographic characteristicsand evolution of the flora of Cuba. ActaBot. Hungarica 31: 3–34.

Bouysse, P., P. Andreief, M. Richard, J. C. Baub-ron, A. Mascle, R. C. Maury, and D.Westercamp

1985. Aves swell and northern Lesser Antil-les ridge: rock-dredging results fromArcade 3 cruise. In A. Mascle (ed.),Caribbean geodynamics: 65–76. Paris:Edit. Technip.

Bowin, C.1975. The geology of Hispaniola. In A. E. M.

Nairn and F. G. Stehli (eds.), The oceanbasins and margins, vol. 3: Gulf ofMexico and Caribbean Sea: 501–552.New York: Plenum Press.

Bralower, T., and M. A. Iturralde-Vinent.1997. Micropaleontological dating of the col-

lision between the North Americanplate and the Greater Antilles Arc inwestern Cuba. Palaios 12: 133–150.

Bresznyanszky, K., and M. A. Iturralde-Vinent1978. Paleogeografıa del Paleogeno de Cuba

oriental. Geol. Mijnb. 57: 123–133.1985. Paleogeografıa del Paleogeno de las

provincias de La Habana. In M. A. Itur-ralde-Vinent (ed.), Contribucion a lageologıa de las provincias de La Ha-bana y Ciudad de La Habana: 100–115.La Habana: Editorial Cientıfico-Tecni-ca.

Bronnimann, P., and D. Rigassi1963. Contribution to the geology and pale-

ontology of the area of the city of LaHabana, Cuba, and its surroundings.Eclog. Geol. Helvetiae 56: 193–480.

Brouwer, S. B., and P. A. Brouwer1982. Geologıa de la region ambarıfera ori-

ental de la Republica Dominicana. InW. Snow et al. (eds.), Trans. 9th Carib-bean Geol. Conf., Santo Domingo, Do-minican Republic 1: 305–322.

Brucks, J. T.1971. Currents of the Caribbean and adjacent

regions as deduced from drift-bottlestudies. Bull. Mar. Sci. 21: 455–465.


Buffler, R. T., W. Alvarez, G. Suarez, A. Camargo,A. Ocampo, and K. Pope

1995. Chicxulub crater, Tertiary carbonatemargins, and the cenote ring. Ann.Meet. Geol. Soc. Am., abstracts withprogram, p. A-348.

Buffler, R. T. and J. M. Hurst1995. Disintegration of the Jurassic-Lower

Cretaceous megabank, Cuba-Florida-Bahamas. 1st SEPM Congr. Sedimen-tary Geology, program and abstracts, p.35. St. Pete Beach, FL.

Bunce, E. T., J. D. Philips, R. L. Chase, and C.O. Bowin

1970. The Lesser Antilles arc and the easternmargin of the Caribbean Sea. In A. E.Maxwell, (ed.), The sea 4: 359–385.New York: Wiley-Interscience.

Burke, K., C. Cooper, J. Dewey, P. Mann, and J.Pindell

1984. Caribbean tectonics and relative platemotions. In W. E. Bonini, R. B. Hard-grave, and R. Shagam (eds.), The Ca-ribbean–South American plate bound-ary and regional tectonics. Geol. Soc.Am. Mem. 162: 31–64.

Burney, D. A., L. P. Burney, and R. D. E.MacPhee

1994. Holocene charcoal stratigraphy fromLaguna Tortuguero, Puerto Rico, andthe timing of human arrival on the is-land. J. Archaeol. Sci. 21: 273–281.

Butterlin, J.1960. Geologie generale et regionale de la

Republique d’Haiti. Universite de Par-is, Trav. Mem. Inst. Hautes Etud. Am.Latine, vol. 6: 194 pp.

Butterlin, J. and F. Bonet1966. Les formations cenozoıques de la partie

mexicaine de la presqu’ıle de Yucatan.Trans. 3rd Caribbean Geol. Conf.,Kingston, Jamaica: 75–90.

Calais, E., B. M. de Lepinay, V. Renard, and M.Tardy

1989. Geometrie et regime tectonique le longd’une limite de plaques en coulissage:la frontiere nord-Caraıbes de Cuba aHispaniola, Grandes Antilles. C. R.Acad. Sci., Paris 308: 131–135.

Calais, E., B. M. de Lepinay, P. Saint-Marc, J.Butterlin, and A. Schaaf

1992. La limite de plaques decrochante nordcaraıbe en Hispaniola: evolution paleo-geographique et structurale cenozoıque.Soc. geol. France Bull. 163: 309–324.

Case, J. E.1975. Geophysical studies in the Caribbean

Sea. In A. E. M. Nairn and F. G. Stehli

(eds.), The ocean basins and margins,vol. 3: Gulf of Mexico and CaribbeanSea: 107–180. New York: PlenumPress.

Case, J. E., T. L. Holcombe, and R. G. Martin1984. Map of the geologic provinces in the

Caribbean region. In W. E. Bonini, R.B. Hardgrave, and R. Shagam (eds.),The Caribbean–South American plateboundary and regional tectonics. Geol.Soc. Am. Mem. 162: 1–30.

Case, J. E., R. Shagam, and R. F. Giegenganck1990. Geology of the northern Andes; an

overview. In G. Dengo and J. E. Case(eds.), The geology of North America,vol. H, The Caribbean region: 177–200. Boulder, CO: Geological Societyof America.

Censky, E. J., K. Hodge, and J. Dudley1998. Over-water dispersal of lizards due to

hurricanes. Nature 395: 556.Cooper, M. A., F. T. Allison, R. Alvarez, M. Coral,

R. H. Graham, A. B. Hayward, S., J.Howe, J. Martınez, R. Naar, R. Penas,A. J. Pulham, and A. Taborda

1995. Basin development and tectonic historyof the Llanos basin, Eastern Cordillera,and Middle Magdalena Valley, Colom-bia. Am. Assoc. Pet. Geol. Bull. 79:1421–1443.

Craw, R. C., and P. Weston1984. Panbiogeography: a progressive re-

search program. Syst. Zool. 33: 1–13.Croizat, L.

1964. Space, time, form; the biological syn-thesis. Caracas (published by the au-thor).

Darlington, P. J.1938. The origin of the fauna of the Greater

Antilles, with discussion of dispersal ofanimals over water and through the air.Q. Rev. Biol. 13: 274–300.

De La Torre, C.1910. Osamentas fosiles encontradas en las

casimbas de la Sierra de Jatibonico.Comprobacion de la naturaleza conti-nental de Cuba a principios de la eraCuaternaria. Rev. Fac. Cienc. LetrasUniv. La Habana 10: 78–88.

De Zoeten, R.1988. Structure and stratigraphy of the Cen-

tral Cordillera Septentrional, Domini-can Republic. Unpubl. Master’s thesis,Univ. Texas at Austin.

De Zoeten, R., and P. Mann1991. Structural geology and Cenozoic tec-

tonic history of the Central CordilleraSeptentrional, Dominican Republic. In


P. Mann, G. Draper, and J. F. Lewis(eds.), Geologic and tectonic develop-ment of the North America–Caribbeanplate boundary in Hispaniola. Geol.Soc. Am. Spec. Pap. 262: 265–281.

Dengo, G., and J. E. Case (eds.)1990. The geology of North America, vol. H,

The Caribbean region. Boulder, CO:Geological Society of America.

Denny, W. A.1992. Seismic stratigraphy and geologic his-

tory of Mid-Cretaceous through Ceno-zoic rocks, southern Straits of Florida.Unpubl. Master’s thesis, Univ. Texas atAustin.

Denny, W. A., R. T. Buffler, and J. A. Austin1994. Seismic stratigraphy and geologic his-

tory of Mid-Cretaceous through Ceno-zoic rocks, southern Straits of Florida.Am. Assoc. Pet. Geol. Bull.78: 461–487.

Dolan, J., P. Mann, R. de Zoeten, C. Heubeck, J.Shiroma, and S. Monechi

1991. Sedimentology, stratigraphy, and tec-tonic synthesis of Eocene-Miocene sed-imentary basins, Hispaniola and PuertoRico. In P. Mann, G. Draper, and J. F.Lewis (eds.), Geologic and tectonic de-velopment of the North America–Ca-ribbean plate boundary in Hispaniola.Geol. Soc. Am. Spec. Pap. 262: 217–241.

Domning, D. P.1982. Evolution of manatees: a speculative

history. J. Paleontol. 56: 599–619.Domning, D. P., R. J. Emry, R. W. Portell, S. K.

Donovan, and K. S. Schindler1997. Oldest West Indian land mammal: rhin-

ocerotoid ungulate from the Eocene ofJamaica. J. Vertebr. Paleontol. 17: 638–641.

Donnelly, T. W.1985. Mesozoic and Cenozoic plate evolution

of the Caribbean region. In F. G. Stehliand S. D. Webb (eds.), The great Amer-ican biotic interchange: 89–121. NewYork: Plenum Press.

1989a. Geologic history of the Caribbean andCentral America. In A. W. Bally and A.R. Palmer (eds.), The geology of NorthAmerica, vol. A, An overview: 299–321. Boulder, CO: Geological Societyof America.

1989b. History of marine barriers and terres-trial connections: Caribbean paleogeo-graphic inference from pelagic sedi-ment analysis. In C. A. Woods (ed.),Biogeography of the West Indies: 103–

118. Gainesville, FL: Sandhill CranePress.

Donnelly, T. W., G. S. Horne, R. C. Finch, and E.Lopez-Ramos

1990. Northern Central America; the Mayaand Chortis Blocks. In G. Dengo and J.E. Case (eds.), The geology of NorthAmerica, vol. H, The Caribbean region:37–76. Boulder, CO: Geological Soci-ety of America.

Donovan, S. K.1994. Trinidad. In S. K. Donovan and T. A.

Jackson (eds.), Caribbean geology, anintroduction: 209–228. Kingston, Ja-maica: The University of the West In-dies Publishers’ Association.

Donovan, S. K. and T. A. Jackson (eds.)1994. Caribbean geology, an introduction.

Kingston, Jamaica: The University ofthe West Indies Publishers Association.

Draper, G.1989. Comparison of the Cretaceous-Paleo-

cene age rocks of northern Hispaniolaand southeastern Cuba: implication forthe tectonic evolution of the GreaterAntilles. 1st Congr. Cubano Geol., re-sumenes y programa, p. 211.

Draper, G., G. Gutierrez, and J. F. Lewis1997. Thrust emplacement of the Hispaniola

peridotite belt: orogenic expression ofthe mid-Cretaceous Caribbean arc po-larity reversal? Geology 24: 1143–1146.

Droxler, A. W.1995. Control of the last 35 M.Y. Caribbean

tectonic evolution on interoceanic wa-ter exchange and heat/salt latitudinaltransfer. Ann. Meet. Geol. Soc. Am.,abstracts with program, p. A-153.

Droxler, A. W., A. A. Staples, E. Rosencrantz, R.T. Buffler, and P. A. Baker

1989. Neogene tectonic disintegration of acarbonate Eocene-Early Miocene me-gabank along the northern NicaraguaRise, Caribbean Sea. 12th CaribbeanGeol. Conf., St. Croix, U. S. Virgin Is-lands, abstracts with program, p. 41.

Droxler, A. W., K. Burke, A.D. Cunningham, A.C. Hine, E. Rosencrantz, D. S. Duncan,P. Hallock, and E. Robinson

1998. Caribbean constraints on circulationbetween Atlantic and Pacific oceansover the past 40 million years. In T.Crowley and K. Burke (eds.), Tectonicboundary conditions for climate recon-struction: 160–191. Oxford: OxfordUniv. Press.


Duncan, R. A., and R. B. Hargraves1984. Plate tectonic evolution of the Carib-

bean region in the mantle referenceframe. In W. E. Bonini, R. B. Hard-grave, and R. Shagam (eds.), The Ca-ribbean–South American plate bound-ary and regional tectonics. Geol. Soc.Am. Mem. 162: 81–93.

Duncan, C. P., S. G. Schladow, and W. G. Wil-liams

1982. Surface currents near the Greater andLesser Antilles. Int. Hydrogr. Rev. 59:67–78

Duque-Caro, H.1990. Major Neogene events in Panamic

South America. In R. Tsuchi (ed.), Pa-cific Neogene events, their timing, na-ture and interrelationships: 101–114.Tokyo: Univ. Tokyo Press.

Eberle, W., W. Hirdes, R. Muff, and M. Pelaez1982. The geology of the Cordillera Septen-

trional (Dominican Republic). In W.Snow et al. (eds.), Trans. 9th CaribbeanGeol. Conf., Santo Domingo, Domini-can Republic, 1: 619–632.

Edgar, N. T., J. B. Saunders, H. M. Bolli, R. E.Boyce, W. S. Broecker, T. W. Donnelly,J. M. Gieskes, W. W. Hay, R. M. Ho-rowitz, F. Maurrasse, H. P. Nieto, W.Prell, I. P. Silva, W. R. Riedel, N.Schneidermann, and L. S. Waterman

1973. Initial Reports of the Deep Sea DrillingProject, vol. 15. Washington, D.C.:U.S. Gov. Printing Office.

Emery, K. O., and E. Uchupi1972. Western North Atlantic Ocean: Topog-

raphy, rocks, structure, water, life, andsediments. Am. Ass. Pet. Geol. Mem.17: 1–532.

Erikson, J. P., and J. L. Pindell1993. Analysis of subsidence in northeastern

Venezuela as a discriminator of tecton-ic models for northern South America.Geology 21: 945–948.

Escalante, G.1990. The geology of southern Central Amer-

ica and western Colombia. In G. Dengoand J. E. Case (eds.), The Geology ofNorth America, vol. H, The Caribbeanregion: 201–230. Boulder, CO: Geolog-ical Society of America.

Eva, A. N., and N. A. McFarlane1985. Tertiary to early Quaternary carbonate

facies relationships in Jamaica. InTrans. 4th Latin American Geol. Conf.,Port-of-Spain, Trinidad, 1979, 1: 210–219.

Fernandez de Castro, M.1884. Prueba paleontologica de que la isla de

Cuba ha estado unida al continenteamericano y breve idea de su constitu-cion geologica. An. Acad. Cienc. Ha-bana 21: 146–165.

Fox, P. J., W. F. Ruddiman, W. B. F. Ryan, and B.C. Heezen

1970. The geology of the Caribbean crust, I:Beata Ridge. Tectonophysics 10: 495–513.

Fox, P. J., E. Schreiber, and B. C. Heezen1971. The geology of the Caribbean crust:

Tertiary sediments, granitic and basicrocks from the Aves Ridge. Tectono-physics 12: 89–109.

Frailey, C. D., K. E. Campbell, and L. Romero-Pittman

1996. Early proboscideans in the Amazon Ba-sin of South America. J. Vertebr. Pa-leontol. 16 (Suppl.): 34A.

Frakes, L. A.1979. Climates throughout geological times.

New York: Elsevier.Frost, S. H., J. L. Harbour, D. K. Beach, M. J.

Realini, and P. M. Harris1983. Oligocene reef trace developments,

southwestern Puerto Rico. Sedimenta9: 1–144.

Galloway, W. E., D. G. Bebout, W. L. Fisher, J.B. Dunlap, R. Cabrera-Castro, J. E.Lugo Rivera, and T. M. Scott

1991. Cenozoic. In A. Salvador (ed.), The ge-ology of North America, vol. J, TheGulf of Mexico basin: 245–324. Boul-der, CO: Geological Society of Ameri-ca.

Garcıa, E., and F. Harms1988. Informe del mapa geologico de la Re-

publica Dominicana, escala 1:100 000:Hoja San Juan. Santo Domingo: Direc-cion General de Minerıa.

Gayet, M., J. C. Rage, T. Sampere, and P. Y. Gag-nier

1992. Modalite des exchanges de vertebrescontinentaux entre l’Amerique du Nordet l’Amerique du Sud au Cretace su-perior et au Paleocene. Soc. geol.France Bull. 163: 781–791.

Geddes, A. J.1994. Jamaica–Geology, 1:250 000 map.

Kingston, Jamaica: Mines and GeologyDivision, Ministry of Mining and Nat-ural Resources.

Gingerich, P. D.1985. South American mammals in the Paleo-

cene of North America. In F. G. Stehliand S. D. Webb (eds.), The great Amer-


ican biotic interchange: 123–137. NewYork: Plenum Press.

Gomberg, D.1974. Geology of the Pourtales Terrace. Flor-

ida Sci. 37, suppl. 1: 1–15.Gonzalez de Juana, C., J. M. Iturralde de Arozena,

and X. Picard Cadillat1980. Geologıa de Venezuela y de sus cuen-

cas petrolıferas, 2 vols. Caracas: Edi-ciones Foninves.

Graham, A.1990. Late Tertiary microfossil flora from the

Republic of Haiti. Am. J. Bot. 77: 911–926.

Guyer, D., and J. Savage1987. Cladistic relationships among anoles

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

1992. Anole systematics revisited. Ibid. 41:89–110.

Hamilton, W. B.,1988. Plate tectonics and island arcs. Geol.

Soc. Amer. Bull. 100: 1503–1527.Haq, B. U., J. Hardenbol, and P. R. Vail

1987. Chronology of fluctuating sea levelssince the Triassic. Science 235: 1156–1167.

Hay, W. W., and N. Wold1996. A simpler plate-tectonic history for the

Caribbean. Zentbl. Geol. Palaontol. Teil1 (7/8): 917–934.

Heaney, L. R., and J. C. Regalado1998. Vanishing treasures of the Philippine

rain forest. Chicago: The Field Muse-um.

Hedges, S. B.1996a. The origin of West Indian amphibians

and reptiles. In R. Powell and R. W.Henderson (eds.), Contributions toWest Indian herpetology: a tribute toAlbert Schwartz: 95–128. Ithaca, NY:Society for the Study of Amphibiansand Reptiles.

1996b. Historical biogeography of West Indianvertebrates. Annu. Rev. Ecol. Syst. 27:163–196.

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

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

1994. Reply: Toward a biogeography of theCaribbean. Cladistics 10: 43–55.

Henderson, I. M.1991. Biogeography without area? Aust. Syst.

Bot. 4: 59–71.Heubeck, C., and P. Mann

1991. Structural geology and Cenozoic tec-

tonic history of the Southeastern ter-mination of the Cordillera Central, Do-minican Republic. In P. Mann, G. Drap-er, and J. F. Lewis (eds.), Geologic andtectonic development of the NorthAmerica–Caribbean plate boundary inHispaniola. Geol. Soc. Am. Spec. Pap.262: 315–336.

Heubeck, C., P. Mann, J. Dolan, and S. Monechi1991. Diachronous uplift and recycling of

sedimentary basins during Cenozoictectonic transpression, northeastern Ca-ribbean plate margin. Sediment. Geol.70: 1–32.

Hine, A. C.1997. Structural, stratigraphic, paleoceano-

graphic development of the margins ofthe Florida Platform. In A. F. Randazzoand D. S. Jones (eds.), The geology ofFlorida: 169–194. Gainesville, FL:Univ. Florida Press.

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

tion of Caribbean ridges and rises withspecial references to paleogeography.In A. Azzaroli (ed.), Biogeographicalaspects of insularity, Atti ConvegniLincei 85: 611–626. Rome: AccademiaNazionale dei Lincei.

Holcombe, T. L., J. W. Ladd, G. Westbrook, N. T.Algar, and C. L. Bowland

1990. Caribbean marine geology; ridges andbasins of the plate interior. In G. Dengoand J. E. Case (eds.), The geology ofNorth America, vol. H, The Caribbeanregion: 231–260. Boulder, CO: Geolog-ical Society of America.

Hoorn, C., J. Guerrero, G. A. Sarmientos, and M.A. Lorente

1995. Andean tectonics as a cause for chang-ing drainage patterns in Miocene north-ern South America. Geology 23: 237–240.

Humphries, C. J.1992. Cladistic biogeography. In P. L. Forey

(ed.), Cladistics: A practical course insystematics: 137–159. Oxford: Claren-don Press.

Hunter, V. F.1978. Notes on the Tertiary stratigraphy of

Margarita Island. Geol. Minjb. 57:189–192.

Issac del Corral, J.1940. El geosinclinal cubano. Rev. Soc. Cu-

bana Ing. 34(4): 485–623.Iturralde-Vinent, M. A.

1969. Principal characteristics of Cuban Neo-


gene stratigraphy. Amer. Assoc. Pet.Geol. Bull. 53: 1938–1955.

1972. Principal characteristics of Oligoceneand Lower Miocene stratigraphy ofCuba. Ibid. 56: 2369–2379.

1975. Problems in the application of twomodern hypothesis to Cuba and the Ca-ribbean region. Ibid. 59: 838–855.

1976–77. Estratigrafıa de la zona Calabazas-Achotal, Mayarı Arriba, Oriente. Rev.Minerıa en Cuba 5: 9–23 (pt. 1), 6: 32–40 (pt. 2).

1978. Los movimientos tectonicos de la etapade desarrollo plataformico de Cuba.Geol. Mijnb. 57: 205–212.

1981. Nuevo modelo interpretativo de la ev-olucion geologica de Cuba. Rev. Cienc.Tierra Espacio 3: 51–90.

1982. Aspectos geologicos de la biogeografıade Cuba. Ibid. 5: 85–100.

1988a. Naturaleza geologica de Cuba. La Ha-bana: Editorial Cientıfico-Tecnica.

1988b. Consideraciones generales sobre elmagmatismo de margen continental deCuba. Rev. Tecnologica 18(4):17–24.

1991. Deslizamientos y descensos del terrenoen el flanco meridional de la SierraMaestra, Cuba sudoriental. In Morfo-tectonica de Cuba Oriental: 24–27. LaHabana: Inst. Geografıa, Academia deCiencias de Cuba.

1992. A short note on the Cuban late Maas-trichtian megaturbidite (an impact de-rived deposit?). Earth Planet. Sci. Lett.109: 225–228.

1994a. Cuban geology: a new plate tectonicsynthesis. J. Pet. Geol. 17: 39–70.

1994b. Interrelationships of the terranes inwestern and central Cuba: comment.Tectonophysics 234: 345–348.

1994c. Meetings Reports: tectonostratographiccorrelation of the NW Caribbean: Do-minican Republic. J. Pet. Geol. 17:243–245.

1995. Meetings Reports: Caribbean ophiolitesand volcanic arcs: Jamaica, October1994: J. Pet. Geol. 18: 234–235.

1996a. Introduction to Cuban geology andgeophysics. In M. A. Iturralde-Vinent(ed.), Cuban ophiolites and volcanicarcs. UNESCO International Geologi-cal Correlation Program Project 364,Special Contribution 1: 3–35. Miami,FL.

1996b. Cuba: El arco de islas volcanicas delCretacico. Ibid. 179–189.

1996c. Estratigrafıa del arco volcanico enCuba. Ibid. 190–227.

1996d. Cuba: El archipielago volcanico Paleo-ceno-Eoceno. Ibid. 231–246.

1996e. Magmatismo de margen continental deCuba. Ibid. 121–130.

1997a. Meeting Reports. 1. Stratigraphy andcorrelation of Cretaceous volcanic arcrocks, Dominican Republic (IGCP364): July 1997. J. Pet. Geol. 20: 489–491

1997b. Introduccion a la geologıa de Cuba. InG. Furrazola Bermudez and K. N.Cambra (eds.), Estudios sobre geologıade Cuba: 35–68. La Habana: CentroNacional de Informacion Geologica.

Iturralde-Vinent, M. A., and R. D. E. MacPhee1996. Age and paleogeographical origin of

Dominican amber. Science 273: 1850–1852.

Iturralde-Vinent, M. A., and M. Norell1996. Synopsis of Late Jurassic marine rep-

tiles from Cuba. Am. Mus. Novitates3164: 17 pp.

Iturralde-Vinent, M. A., G. Hubbell, and R. Rojas1996a. Catalog of Cuban fossil Elasmobran-

chii (Paleocene-Pliocene) and paleo-geographic implications of their Lowerto Middle Miocene occurrence. J. Geol.Soc. Jamaica 31: 7–22.

Iturralde-Vinent, M. A., G. Millan, L. Korpas, E.Nagy, and J. Pajon

1996b. Geological interpretation of the CubanK-Ar database. In M. A. Iturralde-Vi-nent (ed.), Cuban ophiolites and vol-canic arcs. UNESCO InternationalGeological Correlation Program Project364, Special Contribution 1: 48–69.Miami, FL.

Jackson, T., and S. Donovan1994. Tobago. In S. K. Donovan and T. A.

Jackson (eds.), Caribbean geology, anintroduction: 193–208. Kingston, Ja-maica: The University of the West In-dies Publishers’ Association.

Jackson, T. A., and E. Robinson1994. The Netherlands and Venezuelan Antil-

les. In S. K. Donovan and T. A. Jackson(eds.), Caribbean geology, an introduc-tion: 249–264. Kingston, Jamaica: TheUniversity of the West Indies Publish-ers Association.

Jany, I., K. M. Scanlon, and A. Mauffret1990. Geological interpretation of combined

seabeam Gloria and seismic data fromAnegada Passage (Virgin Islands,North Caribbean). Mar. Geophys. Res.12: 173–196.

Jones, B.1994. The Cayman Islands. In S. K. Donovan


and T. A. Jackson (eds.), Caribbean ge-ology, an introduction: 87–110. King-ston, Jamaica: The University of theWest Indies Publishers Association.

Kay, R. F., and R. H. Madden1997. Paleogeography and paleoecology. In

R. F. Kay, R. H. Madden, R. L. Cifelli,and J. J. Flynn (eds.), Vertebrate pale-ontology in the Neotropics: the Mio-cene fauna of La Venta, Colombia:520–550. Washington, D. C.: Smith-sonian Institution Press.

Kennett, J. P.1985. The Miocene oceans: paleoceanogra-

phy and biogeography. Geol. Soc. Am.Mem. 163: 1–340.

Kesler, S. E., N. Russell, C. Reyes, L. Santos, A.Rodrıguez, and L. Fondeur

1991a. Geology of the Maimon Formation,Dominican Republic. In P. Mann, G.Draper, and J. F. Lewis (eds.), Geologicand tectonic development of the NorthAmerica–Caribbean plate boundary inHispaniola. Geol. Soc. Am. Spec. Pap.262: 173–186.

Kesler, S. E., N. Russell, J. Polanco, K. McCurdy,and G. L. Cumming

1991b. Geology and geochemistry of the EarlyCretaceous Los Ranchos Formation,Central Dominican Republic. In P.Mann, G. Draper, and J. F. Lewis (eds.),Geologic and tectonic development ofthe North America–Caribbean plateboundary in Hispaniola. Geol. Soc.Am. Spec. Pap. 262: 187–202.

Khudoley, C.1967. Principal features of Cuban geology.

Am. Assoc. Pet. Geol. Bull. 51: 668–677.

Khudoley, K., and A. A. Meyerhoff1971. Paleogeography and geological history

of Greater Antilles. Geol. Soc. Am.Mem. 129: 1–199.

Kinder, T. H.1983. Shallow currents in the Caribbean Sea

and the Gulf of Mexico as observedwith satellite-tracked drifters. Bull.Mar. Sci. 33: 239–246.

Kinder, T. H., G. W. Heburn, and A. W. Green1985. Some aspects of Caribbean circulation.

Mar. Geol. 68: 25–52.Klitgord, K. D., and H. Schouten

1986. Plate kinematics of the central Atlantic.In P. R. Vogt and B. E. Tucholke (eds.),The geology of North America, vol. M,The western North Atlantic region:351–378. Boulder, CO: Geological So-ciety of America.

Kluge, A.1988. Parsimony in vicariance biogeography:

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

1989. A concern for evidence and a phylo-genetic hypothesis of relationshipsamong Epicrates (Boidae, Serpentes).Ibid. 38: 7–25.

Kolarsky, R. A., P. Mann., and S. Monechi1995a. Stratigraphic development of south-

western Panama as determined from in-tegration of marine seismic data andonshore geology. In P. Mann (ed.),Geologic and tectonic development ofthe Caribbean plate boundary in south-ern Central America. Geol. Soc. Am.Spec. Pap. 295: 159–200.

Kolarsky, R. A., P. Mann., and W. Montero1995b. Island arc response to shallow subduc-

tion of the Cocos ridge, Costa Rica. InP. Mann (ed.), Geologic and tectonicdevelopment of the Caribbean plateboundary in southern Central America.Geol. Soc. Am. Spec. Pap. 295: 235–262.

Ladd, J., T. Shih, and C. Tsai1981. Cenozoic tectonics of Central Hispan-

iola and adjacent Caribbean Sea. Am.Assoc. Pet. Geol. Bull. 65: 466–489.

Larue, D. K.1994. Puerto Rico and Virgin Islands. In S.

K. Donovan and T. A. Jackson (eds.),Caribbean geology, an introduction:151–165. Kingston, Jamaica: The Uni-versity of the West Indies PublishersAssociation.

Larue, D. K., and H. F. Ryan1990. Extensional tectonics in Mona Passage,

Puerto Rico and Hispaniola. In D. K.Larue and G. Draper (eds.), Trans. 12thCaribbean Geol. Conf., St. Croix, U.S.Virgin Is.: 223–230. Miami, FL: MiamiGeological Society.

Larue, D. K., J. Joyce, and H. Ryan1990. Neotectonics of the Puerto Rico

Trench: extensional tectonism andforearc subsidence. In D. K. Larue andG. Draper (eds.), Trans. 12th CaribbeanGeol. Conf., St. Croix, U.S. Virgin Is.:231–247. Miami, FL: Miami Geologi-cal Society.

Lebron, M. C., and M. R. Perfit1993. Stratigraphic and petrochemical data

support subduction polarity reversal ofthe Cretaceous Caribbean island arc. J.Geol. 101: 389–396.


Leclere, A., and J. F. Stephan1985. Evolution geodynamique des points

chauds. In A. Mascle (ed.), Caribbeangeodynamics: 21–34. Paris: Ed. Tech-nip.

Leonov, Yu. G., and V. E. Khain1987. Problems of a systematic approach to

the study and correlation of tectonicphases. In Yu. G. Leonov and V. E.Khain (eds.), Global correlation of tec-tonic movements: 1–12. New York:Wiley.

Lewis, G. E., and J. A. Straczek1955. Geology of south-central Oriente Prov-

ince, Cuba. U. S. Geol. Surv. Bull. 975-D: 171–336.

Lewis, J. F., G. Draper, C. Bourdon, C. Bowin, P.Mattson, F. Maurrasse, F. Nagle, and G.Pardo

1990. Geology and tectonic evolution ofnorthern Caribbean margin. In G. Den-go and J. E. Case (eds.), The Geologyof North America, vol. H, The Carib-bean region: 77–140. Boulder, CO:Geological Society of America.

Lewis, J. F., A. Amarante, G. Bloise, J. G.Jimenez, and H. D. Domınguez

1991. Lithology and stratigraphy of UpperCretaceous volcanic and vulcaniclasticrocks of the Tireo Group, DominicanRepublic and correlation with the Mas-sif du Nord in Haiti. In P. Mann, G.Draper, and J. F. Lewis (eds.), Geologicand tectonic development of the NorthAmerica–Caribbean plate boundary inHispaniola. Geol. Soc. Am. Spec. Pap.262: 143–164.

Lopez-Ramos, E.1975. Geological summary of the Yucatan

Peninsula. In A. E. M. Nairn and F. G.Stehli (eds.), The ocean basins andmargins, vol. 3, The Gulf of Mexicoand Caribbean: 257–282. New York:Plenum Press.

Lugo, J., and P. Mann1995. Jurassic-Eocene tectonic evolution of

Maracaibo basin, Venezuela. In A. J. R.Tankard, R. Suarez Soruco, and H. J.Welsink (eds.), Petroleum basins ofSouth America. Am. Assoc. Pet. Geol.Mem. 62: 699–725.

Macellari, C. E.1995. Cenozoic sedimentation and tectonics

of the southwestern Caribbean pull-apart basin, Venezuela and Colombia.In A. J. R. Tankard, R. Suarez Soruco,and H. J. Welsink (eds.), Petroleum ba-

sins of South America. Am. Assoc. Pet.Geol. Mem. 62: 757–780.

MacLaughlin, P. P., I. P. Gill, and W. A. van derBold

1995. Biostratigraphy, paleoenvironmentsand stratigraphic evolution of the Neo-gene of St. Croix, U.S. Virgin Islands.Micropaleontology 41: 293–320.

MacPhee, R. D. E., and D. A. Grimaldi1996. Mammal bones in Dominican amber.

Nature 380: 489–490.MacPhee, R. D. E., and M. A. Iturralde-Vinent

1994. First Tertiary land mammal from Great-er Antilles: an Early Miocene sloth(Xenarthra, Megalonychidae) fromCuba. Am. Mus. Novitates 3094: 13pp.

1995. Origin of the Greater Antillean landmammal fauna, 1: New Tertiary fossilsfrom Cuba and Puerto Rico. Ibid. 3141:31 pp.

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

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

MacPhee, R. D. E., D. C. Ford, and D. A. Mc-Farlane

1989. Pre-Wisconsinan mammals from Ja-maica and models of late Quaternaryextinction in the Greater Antilles. Quat.Res. 31: 94–106.

Malfait, B., and M. Dinkelmann1972. Circum-Caribbean tectonics and igne-

ous activity and the evolution of theCaribbean plate. Geol. Soc. Am. Bull.83: 251–272.

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

Geophys. Space Phys. 22: 309–362.Mann, P., and R. A. Kolarsky

1995. East Panama deformed belt: structure,age, and neotectonic significance. In P.Mann (ed.), Geologic and tectonic de-velopment of the Caribbean plateboundary in southern Central America.Geol. Soc. Am. Spec. Pap. 295: 111–130.

Mann, P., C. Schubert, and K. Burke1990. Review of Caribbean neotectonics. In

G. Dengo and J. E. Case (eds.), Thegeology of North America, vol. H, TheCaribbean region: 307–338. Boulder,CO: Geological Society of America.

Mann, P., G. Draper, and J. F. Lewis1991. An overview of the geologic and tec-

tonic development of Hispaniola. In P.Mann, G. Draper, and J. F. Lewis (eds.),Geologic and tectonic development of


the North America–Caribbean plateboundary in Hispaniola. Geol. Soc.Am. Spec. Pap. 262: 1–51.

Mann, P., F. W. Taylor, R. L. Edwards, and T-L.Ku

1995. Actively evolving microplate formationby oblique collision and sideways mo-tion along strike-slip faults: an examplefrom the northeastern Caribbean platemargin. Tectonophysics 246: 1–69.

Marshall, L. G.1985. Geochronology and land-mammal

biochronology of the TransamericanFaunal Interchange. In F. G. Stehli andS. D. Webb (eds.), The great Americanbiotic interchange, pp. 49–85. NewYork: Plenum Press.

Marton, G., and R. T. Buffler1993. Application of simple-shear model to

the evolution of passive continentalmargins of the Gulf of Mexico Basin.Geology 21: 495–498.

Mascle, A., M. Cazes, and P. Le Quellec1985. Structure des marges et bassins caraı-

bes: une revue. In A. Mascle (ed.), Ca-ribbean geodynamics: 1–20. Paris: Ed.Technip.

Masson, D. G., and K. M. Scanlon1991. The neotectonic setting of Puerto Rico.

Geol. Soc. Am. Bull. 103: 144–154.Matthew, W. D.

1918. Affinities and origins of the Antilleanmammals. Geol. Soc. Am. Bull. 29:657–666.

Mattson, R.1984. Caribbean structural breaks and plate

movements. In W. E. Bonini, R. B.Hardgrave, and R. Shagam (eds.), TheCaribbean–South American plateboundary and regional tectonics. Geol.Soc. Am. Mem. 162: 131–152.

Maurrasse, F.1982. Survey of the geology of Haiti; guide

to field excursions in Haiti. FloridaGeol. Soc. Spec. Pub.: 1–103.

1990. Stratigraphic correlation of the Circum-Caribbean region. In G. Dengo and J.E. Case (eds.), The geology of NorthAmerica, vol. H, The Caribbean region:pls. 4, 5A, and 5B. Boulder, CO: Geo-logical Society of America.

Maurrasse, F., and G. Sen1991. Impacts, tsunamis, and the Haitian Cre-

taceous-Tertiary boundary layer. Sci-ence: 252: 1690–1693

Maury, R. C., G. K. Westbrook, P. E. Baker, P.Bouysse, and D. Westcamp

1990. Geology of Lesser Antilles. In G. Den-

go and J. E. Case (eds.), The geologyof North America, vol. H, The Carib-bean region: 141–166. Boulder, CO:Geological Society of America.

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

239: 1477.Maze, W.

1984. Jurassic La Quinta Formation in the Si-erra de Perija, northwestern Venezuela.Geology and tectonic environment ofred beds and volcanic rocks. In W. E.Bonini, R. B. Hardgrave, and R. Sha-gam (eds.), The Caribbean–SouthAmerican plate boundary and regionaltectonics. Geol. Soc. Am. Mem. 162:263–282.

McFarlane, E., and L. S. Menes1991. Lower Cretaceous. In A. Salvador

(ed.), The geology of North America,vol. J, The Gulf of Mexico basin: 181–204. Boulder, CO: Geological Societyof America.

McKenna, M. C.1973. Sweepstakes, filters, corridors, Noah’s

arks, and beached Viking funeral shipsin paleogeography. In D. H. Tarling andS. K. Runcorn (eds.), Implications ofcontinental drift to the earth sciences,1: 293–306. New York: AcademicPress.

Meyerhoff, A. A., and C. W. Hatten1968. Diapiric structure in Central Cuba. Am.

Assoc. Pet. Geol. Mem. 8: 315–357.1974. Bahamas salient of North America: tec-

tonic framework, stratigraphy and pe-troleum potential. Am. Assoc. Pet.Geol. Bull. 58: 1201–1239.

Meyerhoff, H. A.1933. The geology of Puerto Rico. Univ.

Puerto Rico Monograph Series B, no.1: 1–360.

Michalzik, D.1987. Sedimentacion y sucesion de facies en

un margen continental pasivo del Trias-ico al Cretacico temprano del norestede la Sierra Madre Oriental, Mexico.Univ. Nac. Auton. Linares, Fac. Cienc.Tierra Actas 2: 27–31.

Miller, K. G., G. S. Mountain, Leg 150 ShipboardParty, and members of New JerseyCoastal Plain Drilling Project

1996. Drilling and dating New Jersey Oligo-cene-Miocene sequences: ice volume,global sea level, Exxon record. Science271: 1092–1095.


Mills, B.1994. Venezuela surface geology, scale 1:

1 000 000. Caracas: CXY International.Molinari, R. L., D. K. Atwood, C. Duckett, M.

Spillane, and I. Brooks1979. Surface currents in the Caribbean Sea

as deduced from satellite tracked drift-ing buoys. Proc. Gulf Caribb. Fish.Inst. 1979: 106–113.

Monroe, W. H.1980. Geology of the middle Tertiary forma-

tions of Puerto Rico. U. S. Geol. Surv.Prof. Pap. 954: 1–93.

Montadert, L., J. Sigal., B. Biju-Duval, A. Mascle,and A. Eva

1985. Precisions sur la stratigraphie des seriescretacees et la tectonique neogene de laJamaıque. In A. Mascle (ed.), Carib-bean geodynamics: 419–426. Paris: Ed.Technip.

Montgomery, H., E. A. Pessagno, and J. L. Pin-dell

1994. A 195 Ma terrane in a 165 Ma Sea:Pacific origin of the Caribbean plate.GSA Today 4(1): 1–7

Moran-Zenteno, D. J., P. Corona-Chavez, and G.Tolson

1996. Uplift and subduction erosion in south-western Mexico since the Oligocene;pluton geometry constraints. EarthPlanet. Sci. Lett. 141: 51–65.

Morgan, G. S.1994. Late Quaternary fossil vertebrates from

the Cayman Islands. In M. A. Bruntand J. E. Davies (eds), The Cayman Is-lands: natural history and biogeogra-phy: 465–508. Boston: Kluwer Aca-demic.

Morgan, G., and C. A. Woods1986. Extinction and the zoogeography of

West Indian land mammals. Biol. J.Linn. Soc., 28: 167–203.

Morrone, J. J., and J. V. Crisci1995. Historical biogeography: introduction

to methods. Annu. Rev. Ecol. Syst. 26:373–401.

Mueller, R. D., J.-Y. Royer, and L. A. Lawver1993. Revised plate motions relative to the

hotspots from combined Atlantic andIndian Ocean hotspot tracks. Geology21: 275–278.

Muizon, C. de, and H. G. McDonald1995. An aquatic sloth from the Pliocene of

Peru. Nature 375: 224-227.Mullins, H. T., and A. C. Newmann

1979. Geology of the Miami Terrace and itspaleoceanographic implications. Mar.Geol. 30: 205–232.

Mullins, H. T., Gardulski, A. F., S. W. Wise, andJ. Applegate

1987. Middle Miocene oceanographic eventin the eastern Gulf of Mexico: impli-cations for seismic stratigraphic succes-sion and Loop Current/Gulf Stream cir-culation. Geol. Soc. Am. Bull. 98: 702–713.

Nagle, F.1972. Rocks from the sediments of escarp-

ments on the Aves ridge. In Trans. 6thCaribbean Geol. Conf., Caracas: 409–423. Caracas: Cromotik.

Nagy, E., K. Brezsnyanszky, A. Brito, D. Coutın,F. Formell, G. Franco, P. Gyarmati, G.Radocz, and P. Jakus

1983. Contribucion a la geologıa de CubaOriental. La Habana: Editorial Cientı-fico

Nelson, G., and N. I. Platnick1981. Systematics and biogeography: cladis-

tics and vicariance. New York: Colum-bia Univ. Press.

Nemec, M. C.,1980. A two phase model for the tectonic

evolution of the Caribbean. In W. Snowet al. (eds.), Trans. 9th Caribbean Geol.Conf., Santo Domingo, Dominican Re-public, 1: pp. 24–34.

Nunn, P. D.1994. Oceanic islands. Oxford: Blackwell

Nutall, C. P.1990. A review of the Tertiary non-marine

molluscan faunas of the Pebasian andother inland basins of northwesternSouth America. Bull. Brit. Mus. (Nat.Hist.), Geol. 45:165–371.

Ota, H.1998. Geographic patterns of endemism and

speciation in amphibians and reptiles ofthe Ryukyu archipelago, Japan, withspecial reference to their paleogeo-graphical implications. Res. Popul.Ecol. 40: 189-204.

Page, R. D. M., and C. Lydeard1994. Toward a cladistic biogeography of the

Caribbean. Cladistics 10: 21–41.Pardo, G.

1975. Geology of Cuba. In A. E. M. Nairnand F. G. Stehli (eds.), The ocean ba-sins and margins, vol. 3, The Gulf ofMexico and Caribbean: 553–616. NewYork: Plenum Press.

Parnaud, F., Y. Gou, J.-C. Pascual, M. A. Capello,I. Truskowski, and H. Passalacqua

1995. Stratigraphic synthesis of northern Ven-ezuela. In A. J. R. Tankard, R. SuarezSoruco, and H. J. Welsink (eds.), Petro-


leum basins of South America. Am.Assoc. Pet. Geol. Mem. 62: 681–698.

Perfit, M. R., and B. C. Heezen1978. The geology and evolution of the Cay-

man Trench. Geol. Soc. Am. Bull. 89:1155–1174.

Perfit, M. R., and E. E. Williams1989. Geological constraints and biological

retrodictions in the evolution of the Ca-ribbean sea and its surroundings. In C.A. Woods (ed.), Biogeography of theWest Indies: 47–102. Gainesville, FL:Sandhill Crane Press.

Pessagno, E.1963. Planktonic foraminifera from the Juana

Dıaz Formation, Puerto Rico. Micro-paleontology 9: 53–60.

Pindell, J.1994. Evolution of the Gulf of Mexico and

the Caribbean. In S. K. Donovan andT. A. Jackson (eds.), Caribbean geolo-gy, an introduction: 13–40. Kingston,Jamaica: The University of the West In-dies Publishers Association.

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

region: a plate tectonic perspective. InG. Dengo and J. E. Case (eds.), Thegeology of North America, vol. H, TheCaribbean region: 405–432. Boulder,CO: Geological Society of America.

Pinet, B., D. Lajat, P. Le Quellec, and P. Bouysse1985. Structure of Aves Ridge and Grenada

Basin from multichannel seismic data.In A. Mascle (ed.), Caribbean geodyn-amics: 53–64. Paris: Ed. Technip.

Pszczolkowski, A.1978. Geosynclinal sequences of the Cordil-

lera de Guaniguanico in western Cuba;their lithostratigraphy, facies develop-ment and paleogeography. Acta Geol.Polonica 28: 1–96.

1986. Megacapas del Maestrichtiano de Cubaoccidental y central. Bull. Polish Acad.Sci., Earth Sci. 34: 81–87.

1987. Paleogeography and paleotectonic evo-lution of Cuba and adjoining areas dur-ing the Jurassic-Early Cretaceous. Ann.Soc. Geol. Poloniae 57: 127–142.

Pszczolkowski, A., and R. Flores1986. Fases tectonicas del Cretacico y del Pa-

leogeno en Cuba occidental y central.Bull. Polish Acad. Sci., Earth Sci. 34:95–111.

Pushcharovsky, Yu., A. Mossakovsky, G. Nekra-sov, S. Sokolov, and M. A. Iturralde-Vinent

1989. Tectonics of the Republic of Cuba. Ex-

planatory note to the tectonic map ofCuba scale 1:500 000 [in Russian].Moscow: Geological Institute, USSRAcademy of Sciences, Academia Nau-ka.

Randazzo, A. F., and D. S. Jones1997. The geology of Florida. Gainesville,

FL: Univ Florida Press.Rasanen, M. E., A. M. Linna, J. C. R. Santos, and

F. R. Negri1995. Late Miocene tidal deposits in the Am-

azonian foreland basin. Science 269:386–389.

Ravenne, C., P. Le Quellec, and P. Valery1985. Depots carbonates profonds des Baha-

mas. In A. Mascle (ed.), Caribbeangeodynamics: 255–271. Paris: Ed.Technip.

Richardson, P. L.1996. Tracking ocean eddies. In R. Gordon

Pirie (ed.), Oceanography, contempo-rary readings in ocean science, 3d ed.:88–104. Oxford Univ. Press.

Roberts, C. M.1997. Connectivity and management of Ca-

ribbean coral reefs. Science 278: 1454–1457.

Robinson, E.1965. Tertiary rocks of the Yallahs area. J.

Geol. Soc. Jamaica 7: 18–27.1994. Jamaica. In S. K. Donovan and T. A.

Jackson (eds.), Caribbean geology, anintroduction: 111–127. Kingston, Ja-maica: The University of the West In-dies Publishers Association.

Rojas, R., M. A. Iturralde-Vinent, and P. W. Skel-ton

1995. Stratigraphy, composition and age ofrudist-bearing deposits. Revista Mexi-cana de Ciencias Geologicas. 12: 272–291

Rosen, D. E.1975. A vicariance model of Caribbean bio-

geography. Syst. Zool. 24: 431–464.1985. Geological hierarchies and biogeo-

graphical congruence in the Caribbean.Ann. Missouri Bot. Garden 72: 636–659.

Rosencrantz, E.1990. Structure and tectonics of the Yucatan

basin, Caribbean sea, as determinedfrom seismic reflection studies: Tecton-ics 9: 1037–1059.

1995. Opening of the Cayman trough andevolution of the northern Caribbeanplate boundary. Ann. Meet. Geol. Soc.Am., abstracts with program, p. A-153.

1996. Basement structure and tectonics in the


Yucatan basin. In M. A. Iturralde-Vi-nent (ed.), Cuban ophiolites and vol-canic arcs. UNESCO InternationalGeological Correlation Program Project364, Special Contribution 1: 36–47.Miami, FL.

Ross, M. I., and C. R. Scotese1988. A hierarchical tectonic model of the

Gulf of Mexico and Caribbean region.Tectonophysics 155: 139–168.

Ross, M. I.1992. PaleoGeographic Information System/

Mac. Earth in Motion Technologies,Users Manual, 1–32.

Roughgarden, J.1995. Anolis lizards of the Caribbean. Oxford

Univ. Press.Rull, V. and C. Schubert

1989. Evolucion de la hipotesis sobre el ori-gen del Caribe. Interciencia 14(2): 74–85.

Russell, N., and S. E. Kesler1991. Geology of the maar-diatreme complex

hosting precious metal mineralizationat Pueblo Viejo, Dominican Republic.In P. Mann, G. Draper, and J. F. Lewis(eds.), Geologic and tectonic develop-ment of the North America–Caribbeanplate boundary in Hispaniola. Geol.Soc. Am. Spec. Pap. 262: 203–216.

Salvador, A.1987. Late Triassic-Jurassic paleogeography

and origin of the Gulf of Mexico Basin.Am. Assoc. Pet. Geol. Bull. 71: 419–451.

1991. Origin and development of the Gulf ofMexico Basin. In A. Salvador (ed.),The geology of North America, vol. J,The Gulf of Mexico basin: 389–444.Boulder, CO: Geological Society ofAmerica.

Saunders, J. B., P. Jung, and B. Biju-Duval1986. Neogene paleontology in the northern

Dominican Republic, 1. Field surveys,lithology, environment, and age. Am.Paleontol. Bull. 89: 1–79.

Schlager, W., R. T. Buffler, D. Angstadt, and R.L. Phair

1984. Geologic history of the southeasternGulf of Mexico. In Initial Reports ofthe Deep Sea Drilling Project 77: 715–738. Washington D.C.: U.S. Gov. Print-ing Office.

Schwan, W.1980. Geodynamic peaks in alpinotype orog-

enies and changes in ocean floorspreading during Late Jurassic–Late

Tertiary time. Am. Assoc. Pet. Geol.Bull. 64: 359–373.

Schwartz, A., and R. W. Henderson1991. Amphibians and reptiles of the West In-

dies: descriptions, distributions, andnatural history. Gainesville, FL: Univ.Florida Press.

Shaw, P. R., and S. C. Cande1990. High-resolution inversion for South At-

lantic plate kinematics using joint al-timeter and magnetic anomaly data. J.Geophys. Res. 95: 2625–2644.

Sigurdsson, H., M. Leckie, G. Acton, and ODPLeg 165 Scientific Party

1997. Proceedings of the Ocean Drilling Pro-gram, Initial Reports, Leg 165. CollegeStation, TX: Ocean Drilling Program.

Simpson, G. G.1956. Zoogeography of West Indian land

mammals. Am. Mus. Novitates 1759:28 pp.

Smith, A. G., D. G. Smith, and B. M. Funnell1994. Atlas of Mesozoic and Cenozoic coast-

lines. Cambridge: Cambridge Univ.Press.

Sohl, N. F., E. Martınez, P. Salmeron-Urena, andF. Soto-Jaramillo

1991. Upper Cretaceous. In A. Salvador (ed.),The geology of North America, vol. J,The Gulf of Mexico basin: 205–244.Boulder, CO: Geological Society ofAmerica.

Sou, T., G. Halloway, and M. Eby1996. Effect of topographic stress on Carib-

bean Sea circulation. J. Geophys. Res.101C7: 16,449–16,453.

Stephan, J.-F., B. Mercier de Lepinay, E. Calais,M. Tardy, C. Beck, J.-C. Olivet, J.-M.Vila, P. Bouysse, A. Mauffret, J. Bour-gois, J.M. Thery, J. Tournon, R. Blan-chet, and J. Decourt

1990. Paleodynamic maps of the Caribbean:14 steps from Lias to present. Soc.geol. France Bull. (8) 6: 915–919.

Stockhert, B., W. V. Maresch, M. Brix, C. Kaiser,A. Toetz, R. Kluge, and G. Kruckhans-Lueder

1995. Crustal history of Margarita Islands(Venezuela) in detail: constraint on theCaribbean plate-tectonic scenario. Ge-ology 23: 787–790.

Stoddart, D. R., J. D. Taylor, F. R. Fosberg, andG. E. Farrow

1971. Geomorphology of Aldabra atoll. Phi-los. Trans. R. Soc. London 260B: 31–65.

Tankard, A. J., R. Suarez and H. J. Welsink (eds.)1995. Petroleum basins of South America.


Am. Assoc. Pet. Geol. Mem. 62: 1–792.

Toloczyki, M., and I. Ramırez1991. Mapa geologico de la Republica Dom-

inicana, escala 1:250 000, comp. by Di-reccion de Minerıa, D.R., and Bundes-anstalt fur Geowissenschaften undRohstoffe, Germany.

Tsuchi, R.1993. Neogene evolution of Pacific Ocean

gateways with reference to the event-biostratigraphy of Southwest Japan. InF. Hehuwat, E. P. Utomo, and A. Dhar-ma (eds.), Proceedings of an Interna-tional Workshop on the Neogene Evo-lution of Pacific Ocean Gateways(UNESCO IGCP–355), Bandarlam-pung, Indonesia, 3-5 August, 1993:2–9.

Turner, M. D.1972. Geology of the San Sebastian area of

northwestern Puerto Rico and the tec-tonic development of Puerto Rico andsurrounding areas. Unpubl. Ph.D. diss.,Univ. Kansas, Lawrence, KS.

Vandel, A.1973. Les isopodes terrestres et cavernicoles

de l’Ile de Cuba. In T. Orghidan et al.(eds.), Resultats des expeditions bio-speleologiques cubano-roumaines aCuba, vol. 1: 153–190. Bucarest: Edi-torial Academiei Republicii SocialisteRomania.

Van den Bold, W.1981. Distribution of Ostracoda in the Neo-

gene of Central Haiti. Am. Paleontol-ogy Bull. 79:1–136.

Viniegra, O. F.1981. Great carbonate bank of Yucatan,

southern Mexico. J. Pet. Geol. 3: 247–278.

Webb, S. D.1976. Mammalian faunal dynamics of the

great American interchange. Paleobi-ology 2: 220–234.

1985. Late Cenozoic mammal dispersals be-tween the Americas. In F. G. Stehli andS. D. Webb (eds.), The great Americanbiotic interchange: 357–386. NewYork: Plenum Press.

1995. Biological implications of the MiddleMiocene Amazon seaway. Science 269:361–362.

Weidie, A. E., W. C. Ward, and R. H. Marshall1979. Geology of Yucatan Platform. In Field

trip guidebook: geology of Cancun,Quintana Roo, Mexico: 50–68. Mid-land, TX: West Texas Geological So-ciety.

Westercamp, D., P. Andreieff, P. Bouysse, A. Mas-cle, and J. C. Baubron

1985. The Grenadines, southern Lesser Antil-les: I. Stratigraphy and vulcano-struc-tural evolution. In A. Mascle (ed.), Ca-ribbean geodynamics: 109–118. Paris:Ed. Technip.

Woods, C. A.1989. The biogeography of West Indian ro-

dents. In C. A. Woods (ed.), Biogeog-raphy of the West Indies: 741–797.Gainesville, FL: Sandhill Crane Press.

APPENDIX 1: RECONSTRUCTING CARIBBEAN PALEOGEOGRAPHY:AN ANALYTICAL GUIDE

This appendix presents additional maps, tables,stratigraphic columns, and other information rel-evant to interpreting the geological history of theCaribbean region and the Late Tertiary paleogeo-graphical reconstructions presented elsewhere inthis paper. Information in the appendix is linkedto tables 1–4, which analytically summarize rel-evant geological data and literature. Each tablecovers a particular temporal interval, organized bygeological units (first column). These units, de-fined in terms of their current location on the ge-ode (second column), are the entities traced on thepaleogeographic maps. The third and fourth col-umns synthesize the evidence available for spe-cific environmental indicators, as preserved inrock-stratigraphic records.

Age of each entity was verified according to

stratigraphic position and fossil context (whereapplicable) in light of the new chronostratigraphicframework provided by Berggren et al. (1995).Use of this new scale necessitated modification ofpublished time-stratigraphic positions of severalformations. Paleoenvironmental interpretations ofeach unit were assessed using data and interpre-tations in the literature as well as the results ofour own fieldwork.

Yucatan PeninsulaThe Yucatan Peninsula (Maya Block or MB;

fig. 14) has been part of the North American platesince the late Cretaceous (Marton and Buffler,1993). Tectonic activity has been largely concen-trated in the southwestern part of MB, far fromthe areas immediately relevant to this study. Nev-ertheless, there is considerable information on this


unit and its Caribbean borderland (Butterlin andBonet, 1966; Case, 1975; Lopez-Ramos, 1975;Weidie et al., 1979; Viniegra, 1981; Mascle et al.,1985; see also the paleogeographical maps com-piled in papers edited by Salvador, 1991).

Stratigraphic data indicate that MB was prob-ably uplifted from Jurassic through Barremiantime (early Cretaceous). The block was coveredby shallow seas from late Aptian through LateEocene, although small islands probably existedon it from time to time (Salvador, 1991; Mc-Farlane and Menes, 1991). Brief emergence ofMB may have occurred around 88–90 Ma, as ahiatus has been reported within the Turonian (Lo-pez Ramos, 1975; Weidie et al., 1979; Viniegra,1981). This correlates well with a Turonian hiatusrepresented in rock sequences in the northern Ca-ribbean and Gulf of Mexico (Meyerhoff and Hat-ten, 1968; Schlager et al., 1984; Pszczolkowskiand Flores, 1986; Iturralde-Vinent, 1994a).

Seismic refraction studies of the Chicxulub bo-lide crater on the NW corner of the Yucatan Pen-insula indicate that islands produced by impactdebris might have had a brief existence there dur-ing the early Paleocene, before they vanished andwere covered by younger marine carbonate de-posits (Buffler et al., 1995). Unfortunately, thereis not enough information available on thesestructures to offer a plausible reconstruction oftheir position or duration.

Several geologic units in western and centralCuba (Guaniguanico, Pinos, and Escambray ter-ranes) are allochthonous (Iturralde-Vinent, 1994a,1994b). These terranes were detached from theiroriginal location along the Yucatan borderland(Pszczolkowski, 1987; Rosencrantz, 1990) be-tween the Late Paleocene and the Middle Eocene,during the formation of the Greater Antilles Fold-belt (fig. 5; Bralower and Iturralde-Vinent, 1997).However, the tectonic processes involved (intensefolding, thrust faulting, metamorphism) took placeat significant depth, and it is quite improbable thatany of these terranes include units from MB thatwere actually emergent at the time of their incor-poration into Cuba.

According to Butterlin and Bonet (1966) andGalloway et al. (1991), parts of MB have beenpermanently subaerial since Late Eocene. How-ever, latest Eocene to Early Miocene deposits arerare and have been found only in the northern andnorthwestern parts of the Yucatan Peninsula. Inthe Oligocene and Miocene, the northeastern sec-tion of the peninsula was covered by water, as wasthe westernmost extremity of Cuba and Isla de laJuventud during the Miocene (Iturralde-Vinent,1969). Furthermore, data from offshore seismiclines clearly indicate that the Yucatan Channel isan ancient feature that was in existence long be-

fore the Late Eocene (probably since the Maas-trichtian; Case, 1975; Mascle et al., 1985; Rosen-crantz, 1990, 1996). Therefore, for most of theCenozoic, if not longer, the Cuban terranes (wher-ever they were then stationed) and MB have beenseparated by a significant water barrier (see alsodiscussion in MacPhee and Iturralde-Vinent,1995).

Northern Central America, Nicaragua Rise, andWestern Jamaica

Geologically, northern Central America (com-prising Nicaragua, Honduras, Guatemala, andsouthern Mexico) consists of a single tectonic ter-rane, the Chortis Block (CB; fig. 15). In most cur-rent plate tectonic models, CB is assumed to haveoriginated off Mexico on the Pacific margin ofNorth America, and to have been rotated into itspresent position late in the Cenozoic (Malfait andDinkelmann, 1972; Donnelly et al., 1990). TheMonagua–Polochic deformed system, the hingezone between CB and terranes farther north inMexico, was particularly active from Middle Eo-cene through Middle Miocene (Pindell, 1994;Moran-Zenteno et al., 1996). Emplacement of CBis generally correlated with the evolution of theCayman Trench system (Pindell, 1994; Rosen-crantz, 1995; Moran-Zenteno et al., 1996). Ap-preciable parts of CB have been uplifted for mostof the Cenozoic (Maurrasse, 1990; Donnelly etal., 1990).

Geographically, the Nicaragua Rise (NR) is theprolongation of CB into the Caribbean Sea. Geo-logically, however, these units are quite different(Holcombe et al., 1990; Donnelly et al., 1990).Stratigraphic data for NR are spotty, but isolatedwells, dredge hauls, and seismic stratigraphy aresufficient to create a broad-brush picture. Mostimportantly, these data confirm that NR and theWestern Jamaican Block (WJ) have shared a con-siderable amount of geological history.

Assuming that the Cretaceous basement of WJcan be correlated with basement rocks of NR (alsoknown to be Cretaceous [Holcombe et al., 1990]),it appears that this terrane (i.e., Western Jamaicaand Nicaragua Rise together) was the site of vol-canic arc activity in the last part of the Mesozoic,under mostly submarine conditions. Local hiatus-es in the Albian and Turonian suggest bouts oftemporary uplift of the terrane in the Late Creta-ceous (Case, 1975; Mascle et al., 1985; Perfit andHeezen, 1978; Holcombe et al., 1990; Maurrasse,1990). Although different marine environmentsdominate the Paleocene and Eocene section (Hol-combe et al., 1990; Robinson, 1994), occurrenceof at least one terrestrial vertebrate in Early Eo-cene rocks of WJ indicates the local occurrenceof land (Domning et al., 1997). Perhaps, therefore,


Fig. 14. Yucatan Peninsula (Maya Block): top left, Simplified pre-Miocene terrestrial geology; topright, late Tertiary stratigraphic columns (compiled from many sources; see appendix 1); bottom, cross-section along offshore seismic line (courtesy Institute for Geophysics, University of Texas at Austin).Cenozoic land connections between Maya and Western Cuban Blocks are ruled out by the shape of thesection underlying the Yucatan channel and the presence of sedimentary rocks of Mesozoic and Cenozoicage.


CB, NR, and WJ were terrestrially connected toNorth America for a short time in the Early Eo-cene. (Eastern Jamaica, comprising the BlueMountains Block, has had a different and morecomplex history; see below.) NR was nearly atambient sealevel during the latest Eocene, andpositive areas were partly or completely upliftedduring most of the Oligocene. Miocene to Recentmarine rocks are found throughout, suggestingthat the NR became once again a submarine fea-ture by the Late Miocene.

Two types of late Tertiary sedimentary environ-ments can be defined for WJ (Eva and McFarlane,1985; Robinson, 1994). One environment is in-dicated by shallow marine carbonate rocks, re-sembling those of NR. The other environment isindisputably deep water, suggesting that Jamaicawas located at the eastern edge of the Rise. Thepossible existence of a short-lived hiatus at thebase of the shallow-water Oligocene section inWJ (Eva and McFarlane, 1985; Robinson, 1994)correlates well with Oligocene uplift postulatedfor NR. Final uplift of WJ, to create most of theisland as it is known today, probably occurredduring the important tectonic deformation thattook place during the Middle Miocene (Eva andMcFarlane, 1985; Mann et al., 1990, 1995; Rob-inson, 1994).

Southern Central AmericaAccording to data compiled by Escalante

(1990) and Kolarsky et al. (1995a, 1995b), South-ern Central America (SCA) is underlain by Me-sozoic oceanic crustal rocks and late Campanian–Eocene oceanic crust and volcanic arc suites. TheLate Eocene to Recent section is summarized infigure 16 (after Escalante, 1990; see also Kolarskyet al., 1995a, 1995b). The Eocene–Oligoceneboundary interval (35–33 Ma) is marked by localuplift and deposition of angular, poorly sortedconglomerates, derived from a local source. Aselsewere in the Caribbean region, the Late Oli-gocene of SCA is transgressive and features ma-rine rocks indicative of shallow- and deep-waterenvironments. An important event occurred latein the Middle Miocene, as elsewhere in the Ca-ribbean region, when strong volcanic activity andgeneral uplift were initiated (see Dengo and Case,1990; Duque-Caro, 1990). According to Duque-Caro (1990), late in the Middle Miocene, SCAmay have been sufficiently uplifted to serve as aconnector between Central and South America.The existence of late Neogene deformation andthrust faulting recorded at the hinge zone betweenPanama and South America (Mann and Kolarsky,1995) tends to support this inference, although itis far from demonstrated.

Northwestern South AmericaThe geology of this region has been investiga-

ted by many authors (fig. 17); key references in-clude Gonzalez de Juana et al. (1980) Bonini etal. (1984), Duque-Caro (1990), Dengo and Case(1990), and Tankard et al. (1995). The regionalgeological picture is too complicated for usefulsummary here, but some general points need tobe briefly canvassed for the purposes of this paper.Detailed paleogeographic reconstruction of north-western South America (NWSA Microcontinent)prior to the latest Eocene is beyond the scope ofthis report (but see Tankard et al., 1995).

During most of the late Tertiary, the NWSAMicrocontinent has moved predominantly east-ward with respect to the motion of the rest ofSouth America (Gonzalez de Juana et al., 1980;Case et al., 1990; Bartok, 1993; Pindell, 1994;Balkwill et al., 1995). Also, this region has beenstrongly affected by vertical movements since thelatest Eocene. Marine sediments corresponding tothe P17–18 zones (35–33 Ma) of Berggren et al.(1995) are not known on the microcontinent prop-er, indicating that most of the area was upliftedduring that time (tables 1, 2). However, mid-Oli-gocene (P19 zone) and younger marine depositshave been found in several basins (Gonzalez deJuana et al., 1980; Duque-Caro, 1990; Cooper etal., 1995), indicating that uplift was followed fair-ly rapidly by subsidence or higher sea stands (orboth). This must have been accompanied by con-siderable modification of the pattern of riverdrainage (Hoorn et al., 1995). The epicontinentalseaway that apparently converted northwesternSouth America into a large island was in existencefrom the latest Eocene (fig. 17; Cooper et al.,1995; Kay and Madden, 1997).

Aruba/Tobago BeltThe structure and history of the Aruba/Tobago

Belt (ATB) has to be considered separately (fig.18). The basement of these islands consists ofMesozoic oceanic crust and Cretaceous volcanicarc units. Equivalent rocks are also found allo-chthonously in the Caribbean Mountains, abovecontinental margin sediments, all partially meta-morphosed (Gonzalez de Juana et al., 1980; Beetset al., 1984; Mascle et al., 1985; Jackson andDonovan, 1994). The deformation of ATB tookplace when the Caribbean Plate interacted withthe South American continental margin (Beets etal., 1984; Erikson and Pindell, 1993; Macellari,1995). Between 35 and 33 Ma this interactionproduced general uplift of the Belt (see above), aswell as progressive subsidence of local basinsalong the continental margin (Macellari, 1995; Er-ikson and Pindell 1993; Stockhert et al., 1995). InTrinidad, sedimentation in deep-water conditions


Fig. 15. Chortis/Nicaraguan Rise and Western Jamaican Blocks: top, Location map, drilling sites,and important geological features; middle, late Tertiary stratigraphic data for Nicaraguan Rise andWestern Jamaica (dredge samples from wall of Cayman Trench); bottom, simplified cross-section, Ni-caraguan Rise to Colombian Basin (compiled from many sources; see appendix 1). Stratigraphic datasuggest that Nicaraguan Rise was at or near sealevel in latest Eocene, and remained positive into Earlyand Middle Oligocene. Off the Rise, Late Oligocene subsidence is indicated by deep-water sedimentsin ODP 999 hole.


Fig. 16. Southern Central America and northern South America (Atrato Basin area): right, Locationmap, major features, and late Tertiary rocks; left, stratigraphic columns (compiled from Escalante, 1990;Duque-Caro, 1990; Kolarsky et al., 1995a, 1995b).

began in the Late Oligocene and persisted untillate in the Miocene (Algar and Erikson, 1995).Similarly, in the Orinoco River basin, marine sed-iments of Eocene to Recent age occur as outcrop-ping or subsurface deposits (Gonzalez de Juana etal., 1980; Cooper et al., 1995). Stratigraphic se-quences in ATB and several basins in the NWSAMicrocontinent present a record of the same crit-ical events in late Tertiary geological history thathave already been described for other parts of theCaribbean region, such as the 35–33 Ma hiatusand subsequent Oligocene transgression (Gonza-lez de Juana et al., 1980). In some areas, the hi-atus seems to have involved long-lasting subaerialexposure, with shorter intervals of marine inun-dation in the Late Oligocene and the late EarlyMiocene. For example, this interpretation seemsto apply to the exceptional hiatus between the

Late Eocene and Late Miocene seen in sectionsfrom Aruba and Margarita. This area is provision-ally regarded as the zone in which the inferredland connection between the Aves Ridge and con-tinent was formed during the Eocene–Oligocenetransition.

Greater AntillesThe Greater Antilles have been the subject of

detailed geological research for most of this cen-tury. Key modern sources include Khudoley andMeyerhoff (1971), Dengo and Case (1990), Mannet al. (1991), Donovan and Jackson (1994), andIturralde-Vinent (1996a). The unusual complexi-ties of this region militate against easy paleogeo-graphical reconstruction, and it is recognized thatmuch more field and laboratory work will have tobe undertaken before paleomapping projects will


Fig

.17

.N

orth

wes

tern

Sou

thA

mer

ica:

Sch

emat

icla

teT

erti

ary

geol

ogic

alm

ap,

show

ing

loca

tion

ofO

rino

coan

dea

ster

nV

enez

uela

nba

sins

sepa

rati

ngN

WS

AM

icro

cont

inen

tfr

omG

uyan

ash

ield

.N

ote

that

rock

sco

nfor

min

gin

age

toE

ocen

e–O

ligo

cene

boun

dary

are

abse

ntth

roug

hout

,w

hile

mar

ine

depo

sits

ofea

rlie

stO

ligo

cene

age

are

rest

rict

edto

Ori

noco

and

east

ern

Ven

ezue

lan

basi

ns(c

ompi

led

from

man

yso

urce

s;se

eap

pend

ix1)

.


Fig. 18. Late Tertiary stratigraphic columnsfor selected basins on NWSA Microcontinent(northwestern South America) and Aruba/TobagoBelt (compiled from many sources; appendix 1).Batuco Fm (Aruba–Bonaire) was originally datedas Eocene–Early Oligocene, but age markers areambiguous and suggest Late Oligocene to EarlyMiocene age (Lepidocyclina, Heterostegina, Par-arotalia, Antiguastrea cellulosa [sensu Gonzalezde Juana et al., 1980]). Lithology as in figure 20A.

achieve the needed level of detail and authorita-tiveness. However, thanks to recent investigationssome headway has been possible (see Paleoge-ography of the Caribbean Region: Evidence andAnalysis).

The basement of the Greater Antilles Foldbelt(GAF) consists of old continental-margin suites,Mesozoic oceanic crustal units, Cretaceous/Paleo-gene volcanic arcs, and latest Cretaceous to Re-cent sedimentary basins—all partly metamor-

phosed, deformed, and jumbled together under thepervasive effects of tectonic forces (Fig. 5; Dengoand Case, 1990; Iturralde-Vinent, 1994a). Widelyvarying opinions concerning the plate tectonicevolution of these elements can be found in theliterature (cf. Pindell, 1994; Iturralde-Vinent,1994a; Hay and Wold, 1996). Most aspects of thetectonic history of GAF prior to the latest Eocenelie outside the special concerns of this paper (butsee fig. 5 and main text for background).

From the literature we have compiled strati-graphic data (figs. 19, 20) pertinent to interpretingthe post-Eocene paleogeography of the GreaterAntilles (Cuba: Bronnimann and Rigassi, 1963;Nagy et al., 1983; Albear and Iturralde-Vinent,1985; Iturralde-Vinent, 1969, 1972, 1996a,MacPhee and Iturralde-Vinent 1994, 1995; His-paniola: Butterlin, 1960; Van den Bold, 1981;Eberle et al., 1982; Maurrasse, 1982, 1990; Saun-ders et al., 1986; Garcıa and Harms, 1988; Mannet al., 1991; Heubeck et al., 1991; Toloczyki andRamırez, 1991; Iturralde-Vinent and MacPhee,1996; Blue Mountains Block: Robinson, 1965,1994; Eva and McFarlane, 1985; Maurrasse,1990; Geddes, 1994; Puerto Rico: Meyerhoff,1933; Pessagno, 1963; Monroe, 1980; Frost et al.,1983; MacPhee and Iturralde-Vinent, 1995; Vir-gin Islands: Larue, 1994; MacLaughlin et al.,1995).

Eastward shift of tectonic activity has been amarked trend in the evolution of the Greater An-tilles since the end of the Eocene. Thus, in west-ern and central Cuba, vertical movements, withlimited sinistral strike-slip faulting, has been therule since latest Eocene (Iturralde-Vinent, 1978).By contrast, east of the Guacanayabo–Nipe faultin eastern Cuba, sinistral faulting and transpres-sional tectonics have been dominant. This has re-sulted in strong deformation and subdivision ofGAF into a series of block-terranes correlatedwith the opening of trenches, grabens, and pull-apart basins (Ladd et al., 1981; Larue et al., 1990;Larue and Ryan, 1990; Jany et al., 1990; MacPheeand Iturralde-Vinent 1995; Mann et al., 1990; Ca-lais et al., 1992). These disrupted block-terraneshave to be returned to their original, latest Eoceneposition in order to reconstruct Greater Antilleanpaleogeography (figs. 3, 6).

Figure 3 is a palinspastic reconstruction of acritical area in the northern Greater Antilles thatencompasses present-day eastern Cuba, northernHispaniola, and Puerto Rico. Offsets along majorstrike-slip faults have been calculated for somefaults (Iturralde-Vinent, 1981; De Zoeten andMann, 1991); these figures are used where avail-able. Our reconstruction is based strictly on con-cordances between identical or strongly correlat-


able geological units (particularly those of LateEocene age or older), as follows:

(1) ?Neocomian–Campanian Cretaceous vol-canic-arc complexes: These rocks outcrop fromwest-central Cuba across Hispaniola into PuertoRico and the Virgin Islands.

(2) Maastrichtian massive conglomerates, dom-inated by ophiolite pebbles and overlain by Pa-leocene/Early Eocene white tuffaceous rocks:This suite of rocks only outcrops east of Holguınin eastern Cuba and in a small area in northwest-ern Hispaniola. This suite is of unique importancefor correlating terranes (Iturralde-Vinent, 1994b).

(3) Ophiolite trend: Outcropping ophiolites inCuba follow the same trend as those in Hispan-iola, especially when their paleoposition is recon-structed palinspastically.

(4) Distinctive metamorphic rock units: Thisseries consists of four distinctive rock suites—marble and schists of the Bahamas margin com-plex, amphibolites (metaophiolites), serpentiniteswith blocks of eclogite, and metamorphosed Cre-taceous volcanic arc rocks. This combination ofmetamorphics outcrops in easternmost Cuba andnorthwestern Hispaniola (Puerto Plata–Samana).

(5) Paleogene volcanic arc rocks: Rocks de-rived from the Paleogene volcanic arc outcrop ineastern Cuba as well as the northern peninsula ofHaiti, central Hispaniola, and Puerto Rico.

(6) Latest Eocene/Oligocene sedimentaryrocks: Units of this age in the Guantanamo Basincorrelate precisely with those found in the the Ci-bao-Altamira Basin of Hispaniola (Calais et al.1992; Iturralde-Vinent and MacPhee, 1996).

The paleoposition of Puerto Rico/Hispaniola isnot so well constrained as that of Cuba/Hispan-iola, although it is known that the Cretaceous vol-canic arc complex outcropping in eastern Hispan-iola also forms a large portion of the basement ofPuerto Rico. The most important correlatableunits are the Duarte complex of Hispaniola andthe Bermeja complex of Puerto Rico, which lie inthe same trend. Also, outcrops of Paleogene rockson the eastern side of Hispaniola lie in the sametrend as their equivalents in Puerto Rico.

The close match between the main structuralfabric and compositional elements of easternCuba, Hispaniola, and Puerto Rico/Virgin Islandsas illustrated in figure 3 is valid only for the in-terval between the latest Eocene and the mid-Ol-igocene (35–30 Ma). Before the latest Eocene, ov-erthrusting and extensive superposition of terranestook place in this area (Meyerhoff and Hatten,1968; Pardo, 1975), indicating that a different pal-eogeographic organization prevailed at that time(see main text). Subsequent to the Late Oligocene,movement along several sinistral faults has

strongly disrupted the assemblage of block-ter-ranes (cf. figs. 1–3).

Blue Mountains BlockA possible land connection between the south-

ern peninsula of Hispaniola (SH) and the BlueMountains Block (BM) of eastern Jamaica is il-lustrated for the Eocene–Oligocene transition (fig.6). The evidence for this connection is circum-stantial at present, and therefore its inclusion inpaleogeographical reconstructions of the GreaterAntilles requires some elaboration.

It is generally accepted that Jamaica originatedas a single crustal unit in the Cretaceous arc (Pin-dell, 1994). However, as previously noted, thereis evidence that Jamaica is structurally and litho-logically divisible into two major terranes, con-sisting of a large western block (Clarendon andHanover Blocks of Lewis et al., 1990) and asmaller Blue Mountains Block. These two ter-ranes differ radically in crustal composition, de-gree of metamorphism, and stratigraphy (includ-ing the span of temporally correlated units), as isevident from several recent papers and mappingprojects (Geddes, 1994; Montadert et al., 1985;Lewis et al., 1990; Robinson, 1994). It is true that,after the Middle Eocene, resemblances betweencoeval formations in different parts of Jamaicagreatly increase (e.g., Bonnie Gate Fm; Robinson,1965, 1994). However, lithology by itself has lim-ited correlation value in this case, as composition-ally similar formations of late Tertiary age out-crop widely in the Greater Antilles.

According to Pindell’s (1994) model of the or-igin of Jamaica, the island’s basement as a wholewas originally part of the Cretaceous volcanicarc located on the leading edge of the CaribbeanPlate. As the plate moved east during the lateCretaceous, the basement rocks of Jamaica re-mained attached to northern Central America. Itis assumed by necessity that these rocks werecarried to their present-day position when theNicaragua Rise (which originated in the Pacific)was inserted into the Caribbean (Pindell, 1994:fig. 2.6). If this were so, one would expect tofind strong similarities between the ophioliticand Cretaceous-arc suites of western Cuba andJamaica, since they were located side by side inthe original arc (sensu Pindell, 1994). Yet thereis virtually no similarity between relevant geo-logical sections (cf. Iturralde-Vinent, 1996a,1996b, 1996c and Robinson, 1994). By contrast,there are evident resemblances in the ophioliticand metavolcanic sequences of the Eastern Cu-ban Block and BM, suggesting that these ter-ranes belong to the same geological province(Iturralde-Vinent, 1995). Most importantly, thegeological composition of the Mesozoic rocks of


Fig

.19

.G

reat

erA

ntil

les:

Sim

plifi

edge

olog

ical

map

ssh

owin

gm

ajor

terr

anes

and

faul

tsac

tive

duri

ngla

teT

erti

ary.

Num

bers

refe

rto

geo-

grap

hica

llo

cati

onof

late

Ter

tiar

yst

rati

grap

hic

colu

mns

depi

cted

infi

gure

20A

,B

.


Fig. 20. Greater Antilles: Late Tertiary stratigraphic columns for selected basins: A, Blue Mountains,Cuba (this page); B, Hispaniola, Puerto Rico/Virgin Islands (compiled from many sources; see appendix1; opposite page).

SH and BM are also remarkably similar. (Cf. de-scriptions of southern Hispaniola by Butterlin[1960] and Maurrasse [1982] with descriptionsof Blue Mountains by Robinson [1994] andMontadert et al. [1985].)

These observations can be made concordant ifit is accepted that BM originated as part of thenorthern Greater Antilles, while WJ evolved fromthe leading edge of the Nicaragua Rise (sensu Pin-dell, 1994). In this interpretation, these terranesmaintained a separate existence until the MiddleMiocene, when they were conjoined during tec-tonic deformations recorded in the island (Mon-tadert et al. 1985). We acknowledge that this dual-origin hypothesis represents a break with the or-thodox view, and that further substantiation is re-quired (see also Stephan et al. 1990: pls. 8–10).

With regard to the possible presence of land onthe Blue Mountains unit during the early Tertiary,

one important piece of the puzzle is the identityof the 33–35 Ma hiatus (corresponding to zonesP17–18) occuring in BM and SH. In SH, thedeep-water G. ampliapertura to G. kugleri zonesof the Jeremie Fm unconformably overlie olderrocks (fig. 20B). Rocks deposited in the time-slicecorresponding to zones P17–18 have not been re-ported; this may indicate that SH was uplifted atthat time, as were many other terranes in the Ca-ribbean (including blocks having similar oceaniccrustal structure, e.g., Beata Ridge).

Likewise, well-dated sediments of 35–33 Maage have not been recognized in the Blue Moun-tains either. Eva and McFarlane (1985) identifiedthe Early Oligocene in Jamaica on the basis of anassemblage including Dictyoconus cookei, Ar-chaias angulatus, and several small species of Pe-neroplis, but these taxa are ambiguous indicatorsof this age, because A. angulatus and Peneroplis


Fig. 20. Continued.

are extant taxa, while D. cookei is not restrictedto the Oligocene (see Bronnimann and Rigassi,1963; Albear and Iturralde-Vinent, 1985). Simi-larly, Robinson (1965, 1994) identified the hemi-pelagic Bonny Gate Fm from localities surround-ing the central Blue Mountains as being Eoceneto Late Oligocene in age. However, he does notlist fossils consistent with a P17–18 zone alloca-tion (sensu Berggren et al., 1995). In any case,the use of benthic forams as index fossils to datenarrow time intervals is problematic, especially in

contexts like the Bonny Gate Fm in which allo-chthonous biodetritus is frequently encountered(Eva and McFarlane, 1985; Robinson, 1965,1994).

The possibility that uplift occurred within BMand SH 35–33 Ma is underscored by the ubiqui-tous presence of a hiatus of that precise age inexisting positive structures almost everywhere inthe Caribbean, including the submarine BeataRidge. Absence of this hiatus in eastern Jamaicawould be completely anomalous.


The nonpalinspastic maps of Eva and Mc-Farlane (1985) indicate the occurrence of land inthe Blue Mountains during the Paleocene and Eo-cene, and, much later, after the Middle Miocene.Although this scenario, involving repeated emer-gence and submergence of the Blue Mountains,cannot be directly challenged on the basis of thefew facts available, we argue that it is more par-simonious to hypothesize that the Blue Mountainshave been permanently subaerial since the latestEocene. This view is consistent with (1) the ab-sence of Eocene and younger rocks from the coreof BM; (2) the lithological character of the BonnyGate Fm, in which the presence of coarse clasticdebris (olistostromes, sandstones, silts) in the bas-al Lloyd member (Maurasse, 1990) indicates theexistence of land during the late Paleogene; and(3) the occurrence of shallow-water biodetritusthroughout the Bonny Gate section (Eva andMcFarlane, 1985), suggesting shelf or coastal en-vironments prevailing at the time of deposition.Better evidence of the emergence of the BlueMountains and the nature of their connection withHispaniola is sorely needed.

Aves Ridge, Lesser Antilles, and Grenada BasinIn the last three decades, a large amount of geo-

logical and geophysical data has been collectedconcerning the Aves Ridge (AR), Lesser Antilles(LA), and Grenada Basin (GB). For an overviewand additional references, the reader is referred topapers by Fox et al. (1971), Bouysse et al. (1985),Pinet et al. (1985), Holcombe et al. (1990), andMaury et al. (1990). Here we concentrate onphysical paleogeography.

Cretaceous and Paleogene volcanic and pluton-ic rocks of island arc affinities occur in AR (Bun-ce et al., 1970; Fox et al., 1971; Nagle, 1972;Bouysse et al., 1985; Westercamp et al., 1985;Holcombe et al., 1990), as do Mesozoic and Eo-cene volcanic rocks in LA (fig. 15). This basiccompositional similarity suggests that, from Cre-taceous through Eocene time, AR and LA were asingle entity: the AR–LA Volcanic Arc (Pinet etal., 1985; Bouysse et al., 1985). This arc was pre-sumably linked geologically to the Aruba/TobagoBelt in the south and the eastern Greater Antillesin the north, because all of these landmasses pos-sess a similar Cretaceous volcanic arc-ophiolitebasement.

If AR and LA once comprised a single arc, itcan be concluded that, at some time in the past,the GB that now separates these two entities didnot exist. However, the age of this basin has notbeen well constrained. Inconclusive seismic evi-dence suggests that GB is filled by sedimentaryrocks of Paleocene(?) to Recent age (Pinet et al.,1985; Bouysse et al., 1985; Bird, 1991), while

dredge hauls from the basin’s margins consist ofmostly Eocene and younger sedimentary and vol-caniclastic rocks (fig. 21).

According to Pindell (1994), GB opened be-tween the Paleocene and Late Eocene, but wepostulate a somewhat younger date (Late Eoceneor younger) for the following reasons. If GB isinterpreted as a back-arc basin, the disjunction ofthe AR–LA arc into two independent geologicalunits (Aves Ridge remnant arc and Lesser Antillesactive arc) would have probably been caused bya local change in the subduction regime (e.g., al-teration of angle of dip of lower slab, or migrationof position of subduction zone). We hypothesizethat this event was correlated with Late Eocenecessation of volcanic activity in AR (and a con-committantly great increase in activity in LA) andincreased thickness of Oligocene and youngersediments in GB (see seismic sections in Nemec[1980] and Pinet et al. [1985]).

Figure 21 depicts islands and other features ofthe eastern Caribbean as they appear today, to-gether with simplified late Tertiary stratigraphiccolumns for AR, GB, and LA. The thickness ofTertiary sediments (Pinet et al., 1985) indicatesthat AR and LA have been positive for most ofthe Cenozoic. Positive topography is also indicat-ed by the occurrence of shallow-water limestonesof Eocene to Lower Miocene age dredged fromridge walls of these features. In addition, EarlyOligocene slope deposits have been recoveredfrom cores from Saba Well (Nemec, 1980); theircompositional character suggests that they werederived from some nearby area (presumably AR)that had been block-faulted and significantly up-lifted (Pinet et al. 1985).

Calculations by Holcombe and Edgar (1990)establish that, given certain assumptions, beforethe Miocene most of the topographic highs on ARmay have been subaerial, or nearly so, and wouldhave formed a string of emergent lands along anorth–south axis. This inference is corroboratedby the frequent occurrence of conglomerates insamples dredged from the Ridge (Nagle 1972;Bouysse et al., 1985). These poorly sorted con-glomerates contain rounded pebbles and cobblesof andesite, up to 10 cm long and showing altered(weathered?) cortices in a calcareous–tuffaceousmatrix. The matrix has yielded large benthic fo-rams and algae, which Bock (1972) identified asOligocene, Early Miocene, or older (Nagle, 1972).

The existence of these conglomerates stronglyimplies the existence of subaerial conditions justbefore and at the time that they were beingformed. Observations on the effects of weatheringon granular igneous and sedimentary rocks in theGreater Antilles (so-called ‘‘big boulder bed’’ ofBronnimann and Rigassi [1963]; M. A. Iturralde-


Fig. 21. Aves Ridge, Grenada Basin, and Lesser Antilles: left, Present-day topography and thicknessof Tertiary sediments; right, stratigraphic columns (compiled from many sources; see appendix 1). AvesRidge and Grenada Basin columns based on combined seismic interpretation and dredge-haul samples.Relative thinness of Paleocene–Oligocene sediments contrasts markedly with thick Miocene–Recentdeposits (Holcombe et al., 1990), and is consistent with interpetation that Aves Ridge has been positivefor much of the Cenozoic. Existence of terrestrially derived conglomerates establishes that portions ofthe ridge were subaerial in Oligocene.


Fig. 22. Cayman Rise, Ridge, and Islands:Late Tertiary stratigraphic columns (compiledfrom Perfit and Heezen, 1978; Jones, 1994; Si-gurdsson et al., 1997). ODP hole 998 drilled onCayman Rise shows very low sedimentation ratesoccurred during Eocene–Oligocene transition andlate Middle/early Late Miocene. Low rates cor-relate well with hiatuses and uplift events in Cay-man Islands and elsewhere in Caribbean region.

Vinent [unpubl. data]) indicate that, under tropicalconditions, surface weathering of the cortices ofgranular rocks causes material to be concentrical-ly lost, eventually resulting in the formation ofrounded cobbles. Most rivers in the Greater An-tilles are far too short to produce roundedness byabrasion alone, which suggests that conglomeratescontaining large, rounded cobbles may be diag-nostic of highland conditions prevailing at thetime of deposition. Incorporation of rounded cob-bles into a calcareous matrix is obviously second-ary, the result of transport to marine environ-ments. In our view, AR conglomerates are readilycorrelatable with the Eocene–Oligocene conglom-erate event recorded elsewere in the Greater An-tilles and South America (MacPhee and Iturralde-Vinent, 1995).

These facts suggest that AR was a topographichigh (Donnelly, 1989b) from the Eocene to theLower Miocene, and was actually emergent forsome indefinite (but probably short) period withinthe latest Eocene/Early Oligocene, after the ter-mination of arc magmatism and related uplift inthe Greater Antilles and Aves Ridge. AR rapidlysubsided thereafter, and was already deeply sub-merged by the Middle Miocene, as indicated byseismic profiles and the occurrence of MiddleMiocene and younger deep-water sediments inwells (DSDP 30, hole 148; Edgar et al., 1973) anddredge hauls (Nagle 1972; Bock, 1972; Bouysseet al., 1985).

Structurally, the southern portion of AR–LA ispart of the deformed and partly obducted volcanicarc complex that also forms the basement of theCaribbean Mountains and islands along the Aru-ba/Tobago Belt (Gonzalez de Juana et al., 1980;Jackson and Robinson, 1994). That the NWSAMicrocontinent (Gonzlez de Juana et al., 1980;Balkwill et al., 1995; Parnaud et al., 1995) sharesa significant portion of its Cenozoic geologicalhistory with the southern AR–LA is also indicatedby the fact that they were extensively and coter-minously uplifted around the Eocene–Oligoceneboundary (35–33 Ma). The possibility that theywere physically connected by emergent land dur-ing this time is strongly suggested by the absenceof marine sediments of this age in northwesternSouth America, as well as in the Aruba/TobagoBelt and AR–LA arc.

The stratigraphic record of LA (Maury et al.,1990) shows a history of activity in marine aswell as subaerial contexts since the Eocene. Thepresence of late Tertiary marine sedimentaryrocks intercalated within volcanic sequences isevidence that the islands have not been perma-nently uplifted since the Eocene, but have had acomplex history involving emergence, subsi-dence, and migration of topographic highs. In


their present form the majority of them are cer-tainly young (Pliocene to Recent).

Beata RidgeThe geological history of Beata Ridge (BR) has

been recently reviewed by Fox et al. (1970), Case(1975), Mascle et al. (1985), Holcombe et al.(1990), and Maurrasse (1990). According to Case(1975) and Holcombe et al. (1990), BR was anundifferentiated part of the Caribbean oceaniccrust until the onset of orogenic movements in theLate Cretaceous. Organizationally, BR consists ofa set of tectonic blocks (fig. 4, cross section, el-evations considerably exaggerated) that have beenpositive since the end of the Maastrichtian(?). BRwas markedly affected by uplift (up to 1000 m)in both Middle Eocene/Oligocene and MiddleMiocene time, but there is no direct evidence thatit was actually emergent at any stage. In particu-lar, the lack of conglomerates in wells and dredgehauls speaks against the existence of high, dryareas on BR (cf. discussion of Aves Ridge). Hi-atuses do occur in the BR section, but none ofthem shows an unequivocal signature of emer-gence. For example, the Middle Miocene hiatusobserved in cores may be due to an interruptionin deposition or to the effects of submarine ero-sion on BR (Mullins et al., 1987; Iturralde-Vinentet al., 1996a). Beata Island, the only emergent partof the ridge, probably became subaerial in theQuaternary. Therefore, Beata Ridge is of no sig-nificance for interpreting the history of land con-nections in the Caribbean region, as it was evi-dently never a structural or paleogeographical linkbetween a mainland and the Greater Antilles(Heubeck and Mann, 1991) (figs. 6–8; tables 1–4).

Nevertheless, evidence that BR was subjected

to latest Eocene/Early Oligocene uplift is impor-tant, because it establishes that areas composed ofthick oceanic crust were affected in the samemanner as geological units composed of continen-tal or island-arc crust, underlining the all-embrac-ing character of this event (MacPhee and Iturral-de-Vinent, 1995).

Cayman Islands and Cayman RidgeBecause of the small size of outcrops on the

Cayman Islands and difficulties in recovering suf-ficient dredge samples from the walls of the Cay-man Trench (CT), information is limited on theseunits (Perfit and Heezen, 1978). The most recentreview of the geology of the Cayman Islands isby Jones (1994); additional topics of interest arecovered by Case (1975), Holcombe et al. (1990),and Rosencrantz (1990, 1995).

Stratigraphic columns (fig. 22) indicate that be-tween 35 and 33 Ma the Cayman Ridge was cov-ered by shallow water. Evidence of deeper waterenvironments occurs only in Early Miocene (andlater) rocks dredged from the walls of CT, indi-cating that this is when the trench system beganto open (Perfit and Heezen, 1978). On the islandsthemselves, shallow-water limestones of Oligo-cene–Miocene and Middle Miocene age havebeen documented, as has a hiatus within the EarlyMiocene. These data underline the very recentcharacter of the islands on the Cayman Ridge.However, they do not necessarily preclude thepossibility of a recent land connection betweenthe Cayman Islands and eastern Cuba, as both arelocated along the same structural trend, i.e., theCayman Ridge (but see point [4] in discussionsection under GAARlandia Landspan and Island–Island Vicariance).

APPENDIX 2: A PLATE TECTONIC MODEL OF THE CARIBBEANFROM LATEST EOCENE TO MIDDLE MIOCENE

Models of the plate tectonic evolution of theCaribbean region tend to agree on the major is-sues (Malfait and Dinkelmann, 1972; Ross andScotese, 1988; Pindell and Barrett, 1990; Pindell,1994), but many details remain uncertain. Mostdiscrepancies among models concern tectonic de-velopments prior to the latest Eocene, althoughcontroversy attends several aspects of plate move-ment in the crucial interval between the end ofthe Eocene and the Middle Miocene. In order topresent new data and interpretations bearing onlate Tertiary tectonics, we constructed a tectonicmodel for the interval 35–14 Ma (see figs. 23–26,table 6). Basic assumptions underlying our recon-struction are presented below and in the captionof figure 23.

1. Crustal plates are deformableInteractions between plates commonly result in

profound deformations of crustal blocks and ter-ranes, not only along plate margins but also with-in intraplate domains. Typical deformations in-clude crustal shortening and superimposition ofunits as a consequence of folding and thrust fault-ing (see figs. 2 and 5) as well as the partial orcomplete destruction of microplates, blocks, andterranes at subduction zones. These processes op-erate at all scales, resulting in modification of thesize and configuration of individual blocks as wellas entire plates. From this it follows that tectonicmodels that purport to be realistic must take someaccount of these processes; if not, results will beinterpretatively problematic. The recent tectonicmodel for the Caribbean published by Hay and


Fig

.23

.C

arib

bean

tect

onic

fram

ewor

k:C

urre

ntpo

siti

ons

ofge

olog

ical

unit

sac

tive

from

late

stE

ocen

eto

Lat

eM

ioce

ne.

Uni

tsha

vebe

enpa

lins

past

ical

lyre

stor

edto

Lat

eE

ocen

e( ;

35M

a)pa

leop

osit

ions

usin

gth

efo

llow

ing

four

cons

trai

nts:

(1)

Cay

man

Tre

nch

syst

emm

ovem

ents

sequ

enti

ally

abso

rbed

byG

uaca

naya

bo–N

ipe

and

Ori

ente

sini

stra

lfa

ults

and

loca

lun

dert

hrus

ting

wit

hin

His

pani

ola

(Man

net

al.,

1995

);(2

)Ja

mai

caac

tsas

two

dist

inct

terr

anes

,W

este

rnJa

mai

ca(o

rigi

nall

yas

soci

ated

wit

hN

icar

agua

nR

ise)

and

Blu

eM

ount

ains

(par

tof

Car

ibbe

ancr

ust,

but

stru

ctur

ally

rela

ted

toE

aste

rnC

uban

and

Sou

ther

nH

ispa

niol

anB

lock

s);

(3)

sout

hern

Cen

tral

Am

eric

aor

igin

ates

sout

hwes

tof

Cho

rtis

Blo

ck,

acco

unti

ngfo

rsi

nist

ral

mov

emen

tsas

soci

ated

wit

hal

kali

nevu

lcan

ism

alon

gH

ess

esca

rpm

ent;

and

(4)

NW

SA

Mic

ropl

ate

acts

aspa

rtof

Car

ibbe

anpl

ate

for

mos

tof

late

Ter

tiar

y,du

eto

acti

vity

alon

gM

erid

a–B

ocon

ode

xtra

lfa

ult.


Wold (1996) may be cited as an example of thelatter. In their model, tectonic blocks and terranesmove, but they do not deform, even after theelapse of many millions of years (Hay and Wold,1996: figs. 2–7). The resulting lack of realism isevident in the evolution of Hispaniola (our Centraland Northern Hispaniolan Blocks, fig. 23). His-paniola is depicted by these authors as sufferingno deformations or alterations from the late Me-sozoic onward; to fit within the space availableper time slice, it has to be sequentially movedfrom a position within the Pacific realm (150–130Ma) to the margin of the Chortis Block (100 Ma),thence to the margin of the Maya Block (67.5Ma), thence south of western Cuba (58.5 Ma),thence south of eastern Cuba (49.5 Ma), finallyending up east of Cuba at 24.7 Ma (Hay andWold, 1996: figs. 2–7). Further contortions are in-troduced by unconstrained rotation of the terranealong its major axis from N–S at 130 Ma to ENE–WSW at 49.5 Ma. These proposed lateral dis-placements and rotations find no support in thegeological composition or structure of central andnorthern Hispaniola (in addition to main text, seefigs. 3 and 5 and appendix 1; Draper, 1989; Mannet al., 1991).

2. Global phases of orogeny have had importantimplications for Caribbean plate evolution

Many different bouts of orogeny, from regionalto global, have affected the Caribbean region andsurrounding areas during the past 170 Ma. Themost significant of these took place in the lateAptian (120–110 Ma), late Campanian–earlyMaastrichtian (75–70 Ma), and Middle Eocene(45–40 Ma) and had worldwide effects. In the Ca-ribbean, these effects included (1) modification ofthe rates of plate movement, (2) rotation of majorstress axes, (3) modification of the orientation andextension of volcanic arcs, (4) alteration of arcmagmatic geochemistry, and (5) consolidation offoldbelts (fig. 5; Schwan, 1980; Mattson, 1984;Pszczolkowski and Flores, 1986; Iturralde-Vinent,1994c; Iturralde-Vinent et al., 1996b; Bralowerand Iturralde-Vinent, 1997). The orogeny whichoccurred in the Middle Eocene is especially note-worthy. Correlated with this orogeny were (1) re-duction in the relative motion of the North andSouth American plates (Pindell, 1994: fig. 2.3),(2) reorientation of the Caribbean Plate stress fieldfrom mainly NE–SW to dominantly E–W, and (3)formation of numerous microplates, blocks, andterranes along plate margins (Case et al., 1984).Many of the critical geological units discussed inthis paper (e.g., various Cuban blocks) wereformed after this event, as were many major fault-bounded structures such as the Windward andMona grabens, Cayman Trench, and the Provi-

dencia and Muertos Troughs. In the Greater An-tilles, the Middle Eocene orogeny was associatedwith cessation of magmatic activity and uplift ofvolcanic structures formed in the Paleocenethough early Middle Eocene. As magmatism end-ed in the Greater Antilles, it also terminated onthe Cayman and Aves Ridges, thereafter shiftingpermanently to the Lesser Antilles arc (cf. figs. 5and 6). This orogeny additionally led to deacti-vation of the Yucatan Basin spreading center andthe shifting of ocean crust production to the Cay-man center (Rosencrantz, 1990). Due to subse-quent movements, crustal segments that wereoriginally part of the foldbelt created by this eventare now distributed in a broad circum-Caribbeanswath, involving southern Mexico, Greater Antil-les, Aves Ridge, Aruba/Tobago Belt, CaribbeanMountains, Columbian/Venezuelan Andes, andCentral America (fig. 23).

3. Island arc magmatic activity on the Caribbeanplate occurred in discrete stages and was not acontinuous process

The most important magmatic events in the his-tory of the Caribbean area were (1) continentalmargin magmatism (170–110 Ma) in associationwith the break up of Pangaea (Maze, 1984; Bar-tok, 1993; Iturralde-Vinent, 1994a), (2) oceanicmagmatism (170–110 Ma) related to the forma-tion of the proto-Caribbean oceanic crust betweenNorth and South America (Pindell, 1994), (3)eruption of alkaline volcanoes related to intraplatetectonic activity along major faults (Dengo andCase, 1990), and (4) the evolution of the volcanicarcs.

With respect to the last of these phenomena, arcmagmatic activity on the Caribbean plate, severaldiscrete stages are evident: (1) ?Neocomian toAptian (120–110 Ma), (2) Albian to Coniacian–Santonian (100–87 Ma), (3) Santonian to ?earlyMaastrichtian (87–70 Ma), (4) mid-Paleocene toearly Middle Eocene (60–55 Ma), and (5) latestEocene to Recent (37–0 Ma). Each of these re-gionally discrete magmatic stages exhibited a spe-cific geological signature marked by structural un-conformities due to tectonic deformation and up-lift, hiatus formation related to erosion and non-deposition, and deposition of coarse clastic andcarbonate sedimentary rocks (see Paleogeographyof the Caribbean Region: Evidence and Analysis,Early Middle Jurassic to Late Eocene Paleoge-ography). This conception of periodic arc mag-matism, punctuated by nonvolcanic intervals,contradicts the widely held view originally for-mulated by Malfait and Dinkelmann (1972), whoenvisaged a single ‘‘Great Arc’’ continuously de-veloping on the leading edge of the Caribbeanplate from the Jurassic onward (see also Burke et


Fig

.24

.P

late

tect

onic

reco

nstr

ucti

on,

Car

ibbe

anR

egio

n,E

ocen

e–O

ligo

cene

tran

siti

on(3

5–33

Ma)

.F

igur

es24

–26

wer

ege

nera

ted

usin

gth

epr

ogra

mP

LA

TE

S(I

nsti

tute

for

Geo

phys

ics,

Uni

vers

ity

ofT

exas

atA

usti

n),

inw

hich

cons

eque

nces

ofsp

ecifi

cdi

spla

cem

ents

can

bein

vest

igat

edov

era

set

inte

rval

(her

e,35

–14

Ma)

rela

tive

toa

fixe

dm

aste

rre

fere

nce

unit

(her

e,N

orth

Am

eric

anpl

ate;

see

tabl

e6)

.P

rese

nt-d

ayco

astl

ines

are

used

for

orie

ntat

ion

but

dono

tre

pres

ent

pale

ogeo

grap

hica

lre

alit

y.S

ome

stru

ctur

alel

emen

tssu

bdiv

ided

into

smal

ler

unit

sto

pres

erve

tect

onic

accu

racy

.C

onta

ctbe

twee

nN

orth

Am

eric

anpl

ate

(NO

AM

)an

dC

arib

bean

plat

e(C

AR

IB)

occu

rsal

ong

Mot

agua

/Sw

an/N

ipe–

Gua

cana

yabo

/nor

ther

nH

ispa

niol

antr

ansf

orm

faul

ts,

and

betw

een

Car

ibbe

anpl

ate

and

Sou

thA

mer

ican

plat

e(S

OA

M)

alon

gM

erid

a-B

ocon

otr

ansf

orm

faul

ttr

end.


Fig

.25

.P

late

tect

onic

reco

nstr

ucti

on,

Car

ibbe

anre

gion

,L

ate

Oli

goce

ne(2

7–25

Ma)

.F

orab

brev

iati

ons

and

othe

rde

tail

s,se

efi

gure

24.

Inth

isst

age,

Cay

man

Rid

gean

dE

aste

rnC

uban

Blo

ckbe

cam

eat

tach

edto

NO

AM

,an

dco

ntac

tw

ith

CA

RIB

inea

ster

nC

uba

jum

ped

toth

eO

rien

tetr

ansf

orm

faul

t.C

AR

IB–S

OA

Mco

ntac

tco

ntin

ued

toli

eal

ong

the

tren

dof

Mer

ida-

Boc

ono

tran

sfor

mfa

ults

,bu

tO

ca–P

ilar

faul

ttr

end

beca

me

acti

vein

asso

ciat

ion

wit

hal

kali

nevo

lcan

icac

tivi

ty.

Act

ive

plat

eco

nver

genc

eco

ntin

ued

inL

esse

rA

ntil

les

and

Pac

ific

Oce

anm

argi

nof

Cen

tral

Am

eric

a.H

ess

Esc

arpm

ent

faul

ttr

end

was

acti

vein

asso

ciat

ion

wit

hal

kali

nevo

lcan

oes.

Str

ong

defo

rmat

ion

and

accr

etio

nto

okpl

ace

alon

gpl

ate

mar

gins

inea

ster

nC

arib

bean

.


Fig

.26

.P

late

tect

onic

reco

nstr

ucti

on,

Car

ibbe

anR

egio

n,M

iddl

eM

ioce

ne(1

6–14

Ma)

.F

orab

brev

iati

ons,

see

figu

re24

.In

this

reco

nstr

ucti

on,

Nor

ther

nH

ispa

niol

aB

lock

isno

wat

tach

edto

NO

AM

,an

dco

ntac

tw

ith

CA

RIB

has

jum

ped

toS

epte

ntri

onal

faul

tin

His

pani

ola.

CA

RIB

–SO

AM

cont

act

islo

cate

das

befo

re,

but

Oca

–Pil

arfa

ult

tren

dw

ashi

ghly

acti

vein

this

peri

od.

Act

ive

plat

eco

nver

genc

eco

ntin

ued

alon

gL

esse

rA

ntil

les

and

Pac

ific

Oce

anm

argi

nof

Cen

tral

Am

eric

a,w

ith

exte

nsio

nof

volc

anic

acti

vity

into

Pan

ama

area

.P

edro

and

Hes

sE

scar

pmen

tfa

ults

tren

dsw

ere

acti

ve,

prod

ucin

gex

tens

ion

alon

gtr

end

ofC

hort

is/N

icar

agua

nR

ise

(for

site

sof

Oli

go-M

ioce

nevo

lcan

icac

tivi

ty,

see

fig.

15).

At

the

sam

eti

me,

stro

ngde

form

atio

nan

dse

dim

ent

accr

etio

nal

ong

unde

rthr

usti

ngfr

onts

cont

inue

dat

plat

ebo

rder

inea

ster

nC

arib

bean

.


al., 1984; Pindell, 1994). Recent supporters of thecontinuous-development model include Mann etal. (1995: fig. 36A–C), who argued that the con-vergent front of the Caribbean Plate was activewith subduction deepening to the south fromMaastrichtian until Middle Eocene times. How-ever, no magmatic activity subsequent to the lateCampanian is recorded in western and centralCuba (Iturralde-Vinent, 1994a), Aruba/TobagoBelt (Hunter, 1978; Jackson and Robinson, 1994),or the Caribbean Mountains (Bonini et al., 1984;Macellari, 1995). Additionally, volcanic arc rocksof mid-Paleocene to early Middle Eocene age ineastern Cuba are structurally unconformable topre-Maastrichtian Cretaceous arc rocks (fig. 5).The subduction zone of this arc was located to thesouth and deepened to the north, rather than viceversa (fig. 5; Iturralde-Vinent, 1994a, 1996d; Si-gurdsson et al., 1997).

4. Stress fields have rotated eastward within theCaribbean region

Stress-field rotation during the formation andevolution of the Caribbean was first proposed byIturralde-Vinent (1975). This phenomenon is ev-ident in the present-day N–S orientation of theconvergence front (island-arc subduction zone) ofthe Lesser Antillean and Central American arcs,and in the extension of arc magmatism southwardin Central America during the last 25 Ma. It isalso evident in the location of post-Eocene trans-form faults and associated deformations along thenorthern and southern margins of the CaribbeanPlate, and in the sequential shifting of plate

boundaries along major faults (in the north, fromNipe–Guacanayabo to Oriente to Septentrional; inthe south, from the Merida/Bocono suture towardthe Oca–Pilar fault; fig. 23–26 and appendix 1).

Migration of volcanic activity and the otherphenomena noted above have been interpreted asa consequence of the oblique collision and result-ing ‘‘escape to the east’’ (or ‘‘escape to theocean’’) of the Caribbean plate as its leading edgeprogressively collided with the Bahamas platform(e.g., Mann et al., 1995). However, Bralower andIturralde-Vinent (1997) have rejected this inter-pretation as it concerns Cuba, on the ground thatthe Cuba–Bahamas collision is conventionallydated to Early Eocene but arc extinction actuallyoccurred much earlier (15 Ma previously, in theLate Cretaceous; see also Iturralde-Vinent, 1994a,1994c). Earlier extinction of the Cretaceous arc isalso seen in the Caribbean Mountains (Bonini etal., 1984; Macellari, 1995; Beccaluva et al., 1996)and the Aruba/Tobago belt (Jackson and Robin-son, 1994), indicating that the Cuban case is notanomalous (fig. 5).

The mechanism of stress-field rotation is notunderstood. Speculatively, it might be assumedthat the phenomenon is driven by the same pro-cess that also affects the movement of tectonicplates. From this perspective, tectonic events re-corded in the lithosphere may be thought of as aconsequence of interactions between individualplates as they accommodate reorientations (rota-tions) of the deep-seated source (mantle-core) ofthe stress field.

paleogeography of the caribbean region: implications...

Documents