Proceedings of the Biological Society of Washington

Home , Rhizophora mangle

MAT (J BIOLOGICAL SOCIETY OF 200 j } ) PROCEEDINGS OF THE WASHINGTON /J 114(1): 1-33. 2001. ^'^'^^^ three-toed sloth (Mammalia: Xenarthra) from Panama, with a review of the genus Bradypus

Robert P. Anderson and Charles O. Handley, Jr.

(RPA) Natural History Museum & Biodiversity Research Center, and Department of Ecology & Evolutionary Biology, 1345 Jayhawk Blvd., University of Kansas, Lawrence, Kansas 66045-7561, U.S.A.;

(COH, Jr. —deceased) Division of Mammals, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560, U.S.A.

Abstract. —Morphological and morphometric analyses of three-toed sloths {Bradypus) from the islands of Bocas del Toro reveal rapid differentiation of several populations during the Holocene. These islands, lying off the Caribbean coast of western Panama, were separated from the adjacent mainland by rising sea levels during the past 10,000 years. The sequence of island formation and the approximate ages of the islands are known. In at least four independent events, sloths on five of the islands evolved smaller size following insularization. Sloths on the younger islands remain conspecific with mainland populations of Bradypus variegatus. On Isla Escudo de Veraguas—the oldest and most remote island of the archipelago—however, the three-toed sloth has dif- ferentiated to the species level, and we here describe it as Bradypus pygmaeus. We provide updated diagnoses and distributions for the species of Bradypus, including a key to the genus. Resumen.—Se realizaron analisis morfologicos y morfometricos de los perezosos de tres dedos {Bradypus) de las islas de Bocas del Toro, que mostraron una diferenciacion rapida de varias de las poblaciones durante el Holoceno. Estas islas, que se ubican en la costa caribena del occidente de Panama, se separaron de tierra firme debido a aumentos en los niveles del mar durante los ultimos 10.000 afios. Se conoce la secuencia de formacion de las islas y sus edades aproximadas. Los perezosos de cinco de las islas evolucionaron hacia un tamaho corporal menor en por lo menos cuatro eventos independientes, siguiendo el proceso de insularizacion. Consideramos que tanto los perezosos de tierra firme como los de las islas jovenes son representantes de la especie Bradypus variegatus; sin embargo en la Isla Escudo de Veraguas, la mas vieja y mas remota del archipielago, el perezoso de tres dedos ha logrado el nivel de especie y lo describimos aca como una especie nueva, Bradypus pygmaeus. Presentamos caracteres diagnosticos y distribuciones para las especies de Bra- dypus, incluyendo una clave de las especies del genero.

Together with armadillos and anteaters, ranges, one species of Choloepus and one sloths make up the Neotropical order Xe- species of Bradypus occur together in the narthra (Gardner 1993, or magnorder Xe- same habitat, exhibiting biotic sympatry ( = narthra sensu McKenna & Bell 1997). Two syntopy; Sunquist & Montgomery 1973,

distantly related genera of sloths, Choloe- Wetzel 1985, Taube et al. 1999). The two pus (two-toed sloths) and Bradypus (three- genera are easily distinguished by the num- toed sloths) are extant. Over much of their ber of clawed digits on their forelimbs (two PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

for Choloepus; three in Bradypus), by the Bradypus is about half that of most mam- blunter muzzle of Bradypus, and by denti- mals (Britton & Kline 1939); its muscle tion—stronger and more complex in Chol- mass to surface area ratio may not be oepus, while simple and peg-like in Bra- enough to create sufficient heat to maintain dypus (Naples 1982, Wetzel 1985). Here we a constant body temperature. Clearly, the consider only the three-toed sloths, Brady- low level of energy expenditure by three- pus. toed sloths for both movement and ther- The three nominal species of Bradypus moregulation directly relates to their diet of can be distinguished both externally and leaves. cranially (Wetzel & Avila-Pires 1980, The natural history of Bradypus indicates Wetzel 1985). The maned sloth (Bradypus a low potential for dispersal. Sloths avoid torquatus) of southeastern Brazil has a dis- predation largely by avoiding detection, tinctive plume or mane of long, jet-black moving very slowly in trees (Brattstrom

hair from its nape to the middle of its back, 1966). Their small home ranges average 1.6 and its skull is characterized by inflated ha (Montgomery & Sunquist 1975, see also pterygoid sinuses (illustrated by Wetzel Chiarello 1998a). Furthermore, their outer 1985:10). Both the pale-throated sloth, Bra- fur harbors an alga, which grows in grooves dypus tridactylus (Guianas, eastcentral Ven- in the surface of the hair (Alston 1879:183, ezuela, and northcentral Brazil), and the Aiello 1985), giving the pelage a green tint brown-throated sloth, B. variegatus (Hon- and providing camouflage. Sloths move duras to Argentina), lack the mane and in- even more slowly on the ground than in flated pterygoids. Adult males of these two trees, traveling on average 0.4 km per hour species also have a large orange patch (Britton & Kline 1939). Surprisingly, they (speculum) on the dorsum. They may be are known to swim well in rivers (Beebe distinguished from each other by the bright 1926:7-9, Carvalho 1960), but we have golden-yellow throat and face in B. tridac- found no reference to their swimming in

tylus, whereas the throat is brownish, at salt water. Perhaps they have a behavioral least at the base of the hairs, in B. varie- aversion to salt water or to rough water and gatus. Most B. variegatus also possess a fa- wave action. Their relatively large size, re- cial stripe not present in B. tridactylus. A stricted diet, and low dispersal potential single pair of large foramina in the antero- make sloths a model system for investigat- dorsal nasopharynx in B. tridactylus are ing the evolution of body size in large in- lacking in B. variegatus (illustrated in sular mammals. Wetzel 1985:10). Emmons & Peer (1997) The islands of Bocas del Toro. —The provided external color illustrations of these province of Bocas del Toro is located on sloths. the Caribbean coast of northwestern Pana-

Three-toed sloths are arboreal folivores. ma adjacent to Costa Rica (Fig. 1). Just off They eat leaves of a variety of trees, in- the coast lies a group of continental islands cluding, but by no means limited to, Cec- that were formed during the Holocene as a ropia spp., which is a common early suc- result of postglacial events, including rising cessional tree in Neotropical rainforests sea level and continental submergence due (Carvalho 1960, Montgomery & Sunquist to meltwater loading and redistribution of 1975, Chiarello 1998b). Concomitant with the Earth's magma. Rising sea levels iso- their energy-poor diet, they have low met- lated hilltops and ridges, first as peninsulae, abolic rates and are not fully homeothermic and then eventually completely sepaiated (Britton & Atkinson 1938). Interestingly, them as islands. The islands of Bocas del temperature regulation is more effective in Toro have low elevations and occupy a pregnant females (Morrison 1945). The per- Tropical Moist Forest life zone, bosque hii- cent of body weight made up of muscle in medo tropical (OEA 1959). They vary in VOLUME 114, NUMBER 1 age, size, distance from the mainland, and (1989). Bartlett & Barghoom (1973) used depth of surrounding water. Mangrove the pollen of Rhizophora mangle in deep- swamps (primarily red mangroves, Rhizo- sea cores from the Gatun Basin in Panama phora mangle) fringe parts of the shoreline to produce a similar curve. Rhizophora of Bocas del Toro and the coasts of some mangle is an obligate saltwater species and of the islands, possibly acting as added bar- represents the major component of Neo- riers to the dispersal of some species. tropical coastal mangrove swamps. Handley Combining ocean floor topography with and Vam then created a composite curve studies of pollen and coral cores from the with years-before-present and depth-below- western Caribbean allowed Handley and M. current-sea-level as axes. Using this curve,

Vam (in litt.) to determine the sequence of they roughly dated each of the bathymetric island formation and to estimate the dates maps and thus estimated island ages from of separation events for the various islands dates of disappearance of land bridges be- of Bocas del Toro. We present their general tween islands and the mainland. Even if conclusions as an introduction to this island their absolute dates err in one direction or system and to interpret the evolution of the other, relative dates of island formation three-toed sloths in Bocas del Toro. Assum- will be correct to the extent that sea-floor ing that the present-day submarine topog- contours in this region have remained con- raphy of Bocas del Toro is not very differ- stant through the Holocene. ent from that of the terrestrial topography Isla Escudo de Veraguas (= Isla Escudo) 10,000 years ago (before flooding), then the occupies a position well outside the Laguna depths of water at which land bridges to de Chiriqui (Fig. la), and was the first of various present-day islands disappeared the islands to be separated from the main- should be apparent from current sea-floor land of Bocas del Toro (ca. 8900 years maps. Thus, given estimates of sea level at B.P.). It fragmented from the eastern shore various time intervals in the past, it is pos- of the Peninsula Valiente and is not directly sible to estimate the approximate date of related to the other islands (Fig. lb). To the isolation of each island. Handley and Vam northwest of the Peninsula Valiente and Isla obtained sea-floor data from maps of Bocas Escudo, the islands of the Laguna de Chi- del Toro produced by the U.S. Army Map riqui are much younger (Fig. la). They Service. Using geographic information sys- fragmented sequentially from the Peninsula tems (GIS) software, they digitized data Tierra Oscura, which was once a long, J- points from isobars below present sea level. shaped peninsula jutting out from the south- With a program produced by the Morpho- western shore of the Laguna de Chiriqui. metries Laboratory of the National Museum The peninsula was formed by the opening of Natural History, they converted the data of the Boca del Drago Pass at the western points to a database transferable to Surfer 4 end of the Laguna (Fig. lb). These islands and then connected them to produce maps are related to each other, but not to Isla Es- of bathymetric contours of Bocas del Toro cudo. The outermost, facing the ocean, are for various depths below present sea level. about 5000 years old: Isla Colon, which

To estimate sea levels over the past was the first to split off of the Peninsula 10,000 years, Handley and Vam utilized Tierra Oscura (ca. 5200 years B.P.), and Isla three models based on coral and pollen Bastimentos, which separated from the pen- cores taken in the western Caribbean. Ra- insula along with what now is Cayo Nancy dioisotope dating of Acropora palmata, a (ca. 4700 years B.P.). Cayo Agua became coral restricted to the upper 5 m of water, isolated from the adjacent mainland (now formed the bases of the curves of depth- part of Isla Popa) about 3400 years B.P below-current-sea level vs. time produced Cayo Nancy recently split from Isla Basti- by Lighty et al. (1982) and Fairbanks mentos proper (<1000 years B.P), and PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

10 20

kilometers

Fig. 1. Maps of Bocas del Toro showing major islands and place names on the mainland (A, upper) and reconstruction of the sequence of island formation (B, lower). In B, the dashed line approximates sea level at

10 m below present. Major events in the formation of the islands (Handley & Varn, in litt.) are as follows: 1) Isla Escudo separated from the southern coast of the province, ca. 8900 years B.P.; 2) the Boca del Drago pass opened, creating a ^-shaped peninsula in the Laguna de Chiriqui, ca. 6300 years B.R; 3) Isla Colon became isolated from the peninsula, ca. 5200 years B.P.; 4) the superisland Isla Bastimentos-Cayo Nancy became separated from what remained of the peninsula, ca. 4700 years B.R; and 5) Cayo Agua was formed, ca. 3400 years B.P. More recently (in the past 1000 years), Isla Cristobal and Isla Popa separated from the mainland, and Cayo Nancy split from Isla Bastimentos. ——

VOLUME 1 14, NUMBER 1 mangroves fringe the shallow channel be- differentiation from their mainland rela- tween it and Isla Bastimentos. Isla Popa and tives. For example, a fruit-eating bat, Arti- Isla Cristobal each separated from the main- beus incomitatus, underwent rapid differ- land in the past 1000 years. They are iso- entiation on Isla Escudo (Kalko & Handley lated from the mainland only by narrow, 1994). shallow channels through mangroves. Is- Collectors found three-toed sloths (Bra- land area and distance from the mainland dypus) on all of the major islands, as well follow, for each of the major islands: Cayo as at the mainland sites. The Bradypus on Agua— 14.5 km^, 6.6 km; Cayo Nancy several of the islands were notably small, 6.8 km^, 9.5 km; Isla Bastimentos—51.5 and some lived in the red mangroves rather km2, 6.3 km; Isla Colon—59.0 km^, 1.5 km; than in upland forest trees as elsewhere. On Isla Cristobal—36.8 km^, 0.3 km; Isla Es- Isla Escudo, Bradypus was found only in cudo—4.3 km^, 17.6 km; Isla Popa—53.0 mangroves. Except for one purchased at Ti- km^, 1.8 km. Further discussion of the his- erra Oscura from a local boy who claimed tory of Bocas del Toro and surrounding re- to have caught it in a mangrove, no three- gions can be found in Jackson et al. (1996). toed sloths were found in mangroves on the Biological interest in the islands of Bocas mainland of Bocas del Toro. The 1993 ex- del Toro emerged recently (summarized in pedition searched in vain for sloths in ex- Handley 1959, Olson 1993, Kalko & Han- tensive areas of mangroves near Nuri. This dley 1994). Early collecting took place ecological separation, coupled with the ob- from 1958 to 1967 and intensified from served size differences, spawned the current 1987 to 1993, when scientists from the study. Smithsonian Institution sampled the biota on all of the islands and at several sites on Materials and Methods the adjacent mainland. The major mainland collecting sites of sloths were Almirante Museum specimens. —We examined a to- and Tierra Oscura in the west, and Nuri and tal of 531 specimens of the genus Bradypus the Peninsula Valiente in the east (Fig. la). in 13 natural history collections (see Spec- Smaller collections of Bradypus were made imens examined) identified as follows: at Sibube and Changuinola in western Bo- American Museum of Natural History, New cas del Toro. We assume that the fauna of York (AMNH); British Museum (Natural the coastal plain of Bocas del Toro was rel- History), London (BM, now Natural His- atively uniform as the islands sequentially tory Museum of London); Field Museum, became isolated from the mainland (Han- Chicago (FMNH); Instituto de Ciencias Na- dley 1959, Olson 1993, Kalko & Handley turales, Universidad Nacional de Colombia, 1994). Colinvaux (1997) has shown that de- Bogota (ICN); Instituto del Desarrollo de spite significant climatological fluctuations, Recursos Naturales Renovables, INDER- the vegetation of lowland tropical forests in ENA, Bogota (IND-M; specimens now part Panama remained intact during glacial of the collection of the Instituto Alexander times, lending support to this premise. Sub- von Humboldt, Villa de Leiva); Michigan sequent to their isolation, numerous spe- State University Museum, East Lansing cies—including bats, rodents, cats, and (MSU); Museo del Instituto La Salle, Bo- weasels present on the nearby mainland gota (MLS); Museum of Comparative Zo- have been extirpated from some or all of ology, Harvard University, Cambridge the islands. Conversely, a few apparently (MCZ); United States National Museum of relict species no longer found on the adja- Natural History, Washington, DC (USNM; cent mainland are present on the islands. * denotes specimens returned to Panama- Many of the species that have survived on INRENARE); Universidad del Cauca, Po- the islands exhibit marked morphological payan (UC); Universidad del Valle, Call —

PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

(UV), University of Kansas Natural History sent, based on the total length of specimens Museum, Lawrence (KU); and University in that population verified as adult by their of Michigan Museum of Zoology, Ann Ar- cranial characters. bor (UMMZ). Information provided by any We found seven pelage characters that source other than the collector is placed in varied among populations. Overall facial [brackets]. Numbers in (parentheses) after color is either tan (off-white to pale brown) the country name indicate the total number or yellow (golden). Orange around the eyes of specimens examined for that country. is present to various degrees, or absent. In the Specimens examined sections, lat- Brow-color categories are: dark brow (a ter- itude and longitude are given after the place minal band of 2-3 cm of dark brown or name to which the coordinates belong. Co- black hair), some dark hair on brow (a nar- ordinates provided by the collector appear row terminal band of only about 1 cm of in parentheses. We provide latitude and lon- dark brown hair), or dark hair not present gitude in brackets for localities that appear (in which case the brow is generally pale in the following standard references: Bra- brown, with no dark terminal band). A dis- zil—Paynter & Traylor (1991); BoUvia tinct boundary line on the brow between the Paynter et al. (1975); Colombia—Paynter dark hair of the brow and the longer, paler except exact coordi- (1997), where more hair of the crown is visible in some speci- nates are given Hershkovitz in by (1947) or mens, but in others the color of the brow original sources cited in Anderson (1999), hair blends in with that of the crown. Some as noted; Costa Rica McPherson — (1985); individuals have a stripe down the midsag- Ecuador Paynter (1993); French Guiana, — ittal plane of the back. Dorsal underfur al- Guyana, and Suriname Stephens Tray- — & ways appears blotchy, with pale and dark lor (1985); Nicaragua—USBGN (1956), patches, but the surface coloration varies. except where more exact coordinates ap- Overall dorsal appearance is blotchy with pear in Genoways (1973); Panama Fair- — brown and beige patches when the outer fur child & Handley (1966); Peru Stephens & — color corresponds to the color of the un- Traylor (1983); and Venezuela—Paynter derfur in that particular area of the dorsum. (1982); additional coordinates for localities In other animals, the surface color is uni- in several countries were taken from Hersh- formly pale, regardless of the color of the kovitz (1977), and are so noted. underfur. The color of the underfur can be Pelage analyses. —We analyzed geo- observed by pushing aside the outer fur and graphic variation in pelage characters in the examining the shorter underfur. In some Bradypus from Bocas del Toro and in B. specimens, the fur of the crown and the variegatus from other regions, principally sides of the head is extremely long, over- from Nicaragua, central Panama, and Co- hanging the forehead and sides of the face lombia. Pelage analyses were based on and creating the aspect of a hood. In other specimens in the ICN, IND-M, KU, MLS, specimens, the fur of the crown and sides USNM, UC, and UV collections. Because of the head is shorter and not noticeably different methods of field preparation or overhanging the face. tanning can affect the color of fur, we did Cranial analyses. For cranial analyses, not consider subtle color differences. We — we included only sloths that had reached focused on striking differences in color and adult size (Age classes 2 and 3, as defined color pattern. We eliminated juvenal and below). Because no explicit standards exist immature animals from the pelage analyses for aging Bradypus skulls (but see Naples by including only individuals whose skulls 1982:6-7), developed the following sys- indicated them as adult (see Cranial analy- we tem of age categories. ses) if skull was available, or that had clearly reached adult size if only a skin was pre- Age class 0, newborn and juvenile: all su- VOLUME 1 14, NUMBER 1

tures open; size small; anterior skull el- del Toro (Fig. 2). To examine geographic ements small and pooriy developed; mas- variation in size in B. variegatus from other seter-temporal fossa smooth; postmastoid parts of its range, we measured only great- fossa not indicated; frontal sinuses little, est length of skull (GLS) on series of B. if at all, swollen; lambdoidal crest not variegatus of Age classes 2 and 3, using formed. either dial or digital calipers to the nearest

Age class 1, immature: all sutures open; 0.1 mm. We chose GLS because prelimi- size intermediate; anterior skull elements nary morphometric analyses indicated that

nearing adult proportions; masseter-tem- it was most highly correlated with general poral fossa usually smooth; postmastoid size in Bradypus. Specimens from the fossa may be indicated; frontal sinuses American Museum of Natural History were somewhat swollen; lambdoidal crest pre- not included in the quantitative analyses, sent (immatures may retain some juvenal but we report them as additional confirmed characters). distributional records. Cranial nomenclature Age class 2, young adult: all sutures open; follows Naples (1982). We here define and size large; anterior skull elements at adult illustrate (Fig. 2) our measurements for proportions. Young adults must also have Bradypus. some of the following: masseter-temporal Greatest length of skull (GLS): Distance fossa rugose to the touch; postmastoid between the anteriormost point of the na- fossa prominent; frontal sinuses swollen; sals and a line connecting the posterior- lambdoidal crest sharp-edged. most surfaces of the occipital condyles. Age class 3, full adult: some or all sutures Anterior zygomatic breadth (AZB): Great- closed; size large; anterior skull elements est breadth across the jugal (anterior) zy- fully developed. Fully adult sloths have gomata. most or all of the following: masseter- Posterior zygomatic breadth (PZB): Great- temporal fossa visibly rugose; postmas- est breadth across the squamosal (poste- toid fossa prominent; frontal sinuses rior) zygomata. swollen; lambdoidal crest sharp-edged. Postorbital breadth (POB): Least breadth Individuals reach adult size by Age class across the constriction of the frontals, 2 in Bradypus. Skulls with all cranial su- posterior to the postorbital processes. tures closed clearly represent adults, but Squamosal process length (SPL): Distance closure of even one suture signifies full- between anteriormost point of the squa- adult status. The nasal (intemasal) and in- mosal process of temporal (posterior zy- terparietal sutures are usually the last to gomata), and the notch formed by the close. Closure of the basioccipital-basisphe- junction of the posterior border of the noid suture represents a good indicator of bulla and the mastoid process. maturity in bats and rodents, but this suture Maxillary toothrow length (MTRL): Great- closes late in Bradypus after adult size is est alveolar length from the anteriormost attained. Its lack of closure should not be edge of the anterior chisel-shaped tooth used alone to judge adulthood. to the posteriormost edge of the last mo- Measurements. —We recorded external lariform tooth in a maxillary toothrow. measurements—total length (TOTAL); tail Postpalatal length (PPL): Distance between length (TAIL); hind foot length (HF); and the anteriormost margin of the mesopter- ear length (EAR)—in mm and body mass ygoid fossa and the anteriormost margin in kg from museum labels. Using dial cal- of the foramen magnum. ipers to the nearest 0.1 mm, we took 14 Palatal breadth (PB): Greatest alveolar measurements on all adult-sized skulls of breadth across the lateral margins of the Bradypus (Age classes 2 and 3) from Bocas first molariform teeth. PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

ABAR

1 cm

DJPL

Fig. 2. Dorsal, ventral, and lateral views of a cranium and lateral view of a mandible of Bradypus variegatus, illustrating method of taking cranial measurements. Abbreviations and measurements are defined in the text.

Braincase depth (BD): Greatest distance be- nearest point on the dorsal border of the tween the medioventral surface of the ba- jugal process. sioccipital and the dorsalmost point of the Greatest external auditory meatus diameter braincase. (EAM): greatest internal diameter of the Antorbital bar breadth (ABAR): Least external auditory meatus. breadth across the flattened antorbital bar Squamosal process breadth (SPB): Breadth (jugal process of zygomata) anterior to its of the squamosal process, taken 5 mm division into the ascending and descend- posterior to the anterior tip of the process. ing jugal processes, taken in ventral view. Ascending mandibular ramus breadth Descending jugal process length (DJPL): (ARB): Least distance between the anter- Distance between the ventralmost point iormost point of the angular notch of the of the descending jugal process and the mandible, between the condylar and an- VOLUME 1 14, NUMBER 1

gular processes, and the nearest point on Peninsula Valiente, and Tierra Oscura—we the anterior margin of the ascending ra- tested for a difference in mean PC I scores mus below the coronoid process. between sexes, again controlling for locality in a general linear model. Finally, we Statistics. calculated descriptive —We tested for differences between all pairs of statistics and performed analyses of vari- localities on PC I, with the same protocols ance using 11.12 (ANOVAs) MINITAB as in the comparisons of scores on the first software for personal computers (MINI- two canonical axes. TAB 1996) and used SAS 6.12 for UNIX We used measurements of greatest length (SAS 1990) to examine the data using mul- of skull (GLS) to compare the small sloths tivariate statistics. We used a Type-I error from Bocas del Toro with Bradypus varie- = rate of a 0.05 for all tests. The probabil- gatus from mainland localities outside Bo- ity levels that we report should be consid- cas del Toro, as preliminary analyses indi- ered approximate, however, because our cated that GLS correlated highly with PC I, sample sizes were too small to test ade- and thus represented a good measure of quately for departures multivariate from overall size. Series from Bonanza, El Re- normality. creo, and Tepeyac (Nicaragua); Bajo Cali- We conducted a multiple-group discrim- ma-Rio Raposo and Tumaco (Colombia); inant function analysis (DFA) on the Bra- and Mojui dos Campos (Brazil) were our del dypus of Bocas Toro using log,o-trans- six mainland localities outside of Bocas del formed measurements. Collection locality Toro, as few specimens were available from denoted group length membership. Ear and other sites. We conducted an ANOVA on body mass were excluded from the DFA GLS measurements of the six mainland because few individuals carried those mea- samples outside Bocas del Toro and our surements. We tested for multivariate dif- samples from the five islands with small ferences among localities by F-statistics for sloths in Bocas del Toro (Cayo Agua, Cayo Mahalanobis distances between pairs of Nancy, Isla Bastimentos, Isla Colon, and group centroids using Holm's (1979) mod- Isla Escudo), using a Tukey's test with a ification of the Bonferroni correction for family-wide error rate of a = 0.05. multiple comparisons. We then conducted multiple unplanned comparisons between Results all pairs of localities for scores on the first two canonical axes, using Tukey's test with Pelage. —Externally, the Bradypus of a family-wide error rate of a = 0.05. Bocas del Toro resemble specimens from Additionally, we performed a principal central Panama more closely than they do components analysis on the same data ma- those of Nicaragua or South America (Table

trix without regard to collection locality. 1). All Central American specimens have The first principal component (PC I) of the tan faces, whereas faces of most South covariance matrix of log,o-transformed American sloths have a yellowish cast. In measurements was used as the best measure addition, while the sloths of Bocas del Toro, of overall sloth size in further analyses. For central Panama, and many localities in Co- the two populations with several individuals lombia have at least some orange coloring of both Age classes 2 and 3 (Cayo Nancy around their eyes, Nicaraguan specimens and Isla Escudo), we tested for a difference lack this trait. Most individuals from Bocas in mean PC I scores between age classes del Toro have either some (~1 cm) or much while controlling for locality, using a gen- (2—3 cm) black or dark brown brow hair. eral linear model. Likewise, for those lo- Specimens from central Panama and ex- calities with multiple individuals of each treme NW Colombia match those from Bo- sex—Cayo Agua, Cayo Nancy, Isla Popa, cas del Toro in this trait, but Nicaraguan 10 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

and most other South American sloths lack (N); mixed m follows: % O < Et^ >; ^ variegatus. 3 +j a dark brown brow. A distinct brow line is no > Z z Z >^ PQ Z or are < found in specimens from central Panama, as ^ extreme NW Colombia, and in some local- (S), samples adypus viations ^5 ities in Bocas del Toro, but not in speci- 00 JH >^ z z >H pa z Dme ^-S < 5 mens from Nicaragua or elsewhere in Cen- i^ " i: a U < ^ .XI . r. dJ tral America or western Colombia. Thus, bz-s t^ & 5 5 the Bradypus from Bocas del Toro group cS'S H >- >^ >- >^ D Z ^ with other Central American and western 2 ^S yj ca .. ^ m »

Colombian three-toed sloths, with closer af- the bro (Y) Oh O u 5 finity to those from central Panama and ex- WtoS H >^ >H >- JH pa >H out on -es 3 E treme NW Colombia to the east rather than ^ M ^ 'y 3 < with sloths from Nicaragua to the north- da o s hood throu from z >H H C/5 z >^ west. "^ §z s (N); d d 2 c _ C o reg (B) arra > § 5 '« characters, often linking island populations H t/3 z > pa z (sit), cted chy are XJ c with nearby populations on the mainland u ^ O M >o < sel bli u (Table 1; Fig. 1). For example, the only .2 sligh uping £9 er or 6 populations in Bocas del Toro with individ- H >^ c^ Z >^ D Z J= ^ ^ o c« O >> o >. D feb uals lacking a dorsal stripe are found at Ti- ( 1—1 «

yes erra Oscura and on the adjacent islands of from 2 a. Irt ographic U C3 55 uniform H >^ z z >- oa z Cristobal, Popa, and Cayo Nancy. For dor- as ace, sal appearance, there is a west-to-east cline u c ?^ O u < from uniform to blotchy. The populations w o .1^ O 5 as pearan( slash. H >-' C/5 z z D z with the highest frequencies of individuals ange ea o o with a distinct brow line are two proximate o J^ x; localities in the west (Isla Colon and the DS ^ u C £2 o U H >^ 00 z z D Z adjacent mainland at Almirante) and two in 5 -a <" -" 03 O .. the east (Isla Escudo and Nuri, the nearest >^ E >-- site on the mainland). Finally, while sloths cu ^ CQ H >- 00 z JH D Z i=. . o n< from most localities in Bocas del Toro have T)o mon from or ^2 ( c3 ti patches, central lo- orange eye those of four u g z (Y) com 5 5 tan H >- 00 z z P z calities (Almirante, Isla Cristobal, Isla adypus yes or, lost Popa, and Peninsula Valiente) have little if c U 2 Br col m any orange. Overall, o pelage of specimens ripe, ;e the u or Si H >^ >" JH DZ '^^ a from Isla Escudo closely matches pelage (^1 ^ s^ « U wit w ers sal 2 characters found at Nuri. However, char- 's: o CO o o S pa acters variable in the Nuri population ap- sll 1 < ^^ 5 2 ' - c H >- W2 z JH DZ c ti; --^ pear to be fixed on Isla Escudo. Sloths from o 7 DO ^^ (11 ca I- Isla Escudo were unique in possessing long a ca 3 1 no y >- are H z Z z pa z elag( tes C z hair on the crown and sides of the head, p r o sta giving the impression of a hood. ?(Y)o ta iations o Crete oth u Quantitative measurements.—Both uni- b o Dis^ ye; o c

VOLUME 1 14, NUMBER 1 11

o 6 00 00 vo 3 VJ •a CT; (N vq vo OS 00 vq r^ 0\ O -''^ ^"* lO"* tN — « (J\ 6 •a p fi m o tN ri ^ -^ ^i-: _^r-^ _^[~^ __;^ T3 'c .c +1 +1 +1 +1 C o d L. T3 tN ^ tN m (A O c; t ca « 1) t/3 C3 o T3^ "m a 'i-l O 00 ro r-; OS tN [^ 00 00 so OS tN so 60 T3 c iri '-' -; _; (N —< "^ d K tN 0(3 Qs d d ^^ ^ , < in • c^ C/3 • +1 +1 "T !2. +1 +1 - +1 +1 I +1 I +1 t S 7^ t S • +' O a - — so J 00 1 SO OS c < E "1 o6 n (N ^ so o 00 OS 00 + 1 ^ d ^ Ci. .. o p C C/l T3 ca C a n 00 p « o — tN tN SO tN (N ^' tN tN tN tN so -^ o C3 _ ^ 01 GO 00 so 00 ca - - . X + 1 I S+l + 1 I S+i +1 I +1 +1 I S +1 I s in in vn t^ ON C u ^• -* o 'n od o ^ so o u 13 d > O so tN SO en t~~ so OS t-~ d OS c^ o 00 r~ 00 2 'So E 13 'fi D CJ c« > V) X o T3 3 o r^ O OS 00 m in (N in g 1 w-> '"i CO ^ tN ^ O m (N in c o O o '3 o so in in c I-l (N 1-3. (2) 3-4. 2-4. T in "^ f-)" P sn o o 7P° ± (10) ± T +1

•a 3 C3 tS in 00 in O so o O SO OS ^ OS in SO as o ro OC) d 1 p

1 't ^ tN 1 Iso^^r-Jin^^dso^^^"^ '=^2 2 O ^^ O ,_^ o ^ ,^ 2 m^^— (3) (5) ± 8-14 1^ 10-12 11-18 ^- 1 -. + 1 -^ + 1 +1 + 1 +1 +1 +1 +1 (J U _L ± £+1 7 s T S 7 ^ 7 § 7 o J^ i O •* O r<-i O O tN ^H in ^H o 6 inin— in— ro— o-^ o — m o r-H i! ^E 3 •:= o E -g c so tN so o ^ in o 00 (N ^ a 2 OS r^ 00 O ^ t^ ro =2 o 00 (N 00 so o ^ SO ^ o ^ "* r<-i 6J) T3

r- t<^ 1 (N 1 +1 +1 +1 +1 +1 1 m 7 7 + 1 1 +1 1 P +1 1 + ^ 1 s + 1 1 ^ 1 1 § 1 E iS so 00 so in 5 o ^ I^ o ^ tN CM p tN tN 1-H r-^ .—1 ^^ . ^ 00 in ^ 00 — d 2 (N tN

^^ o > Vi (D .B 2 w •a SO o so O O * 'S u c OS SO [^ in 2 E ca o o * => C^ _ OCJ OO o o „ tN en in *

I I - I o N cS ;+!'?§ ^ +1 ^+1 S +1 I +1 in OS c^ tN tN c "^ ^^ o ^^ ca tN 00 m so m r»-i so t~^ ^ in ^ so ^^ >. u tN so 00 ca xt ob in in o E a. =y 1 ^^ T3 la c ca o E U so (A >< <2 u o o '^ O 00 in rn tT sd (N m -* -H SO in in tN •* (N to _CJ c in so so so 0) •? m so

I I I - + 1 + 1 +1 +1 I +1 +1 in J^ w IZ) 13 c 00 tN in so in tn "^ "^ i u ca in (N —; —; M N o OS 00 0) in 'cfl > E ca Oj ca Q. C "H. > •a S £ ca M c/3 3 Q 2 -a E o c CO ca D u H .£ 3 CJ (N C4-. 13 E DO _U a 3 < o o Cfl o IS CA E ca ca ca '3 '3 pa U U H fc o ca o E •a u U H 'Z. 12 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

B in Ov o (^ O VD NO 00 NO 5 _ (^ m <^. oo f^i 00 (N.0 —: 00

is u "^ (0° f^ "n in ''^ f:° I V ^ I "V +1 +1 5i vi w2 O) in NO in NO iri in in NO in -o 3 in in in in in

1) tN

•a 15 d O o X: 1 1 1 1 + + + oi + ., + 1 c oi ON r<-i ON ro 00 NO en < X) ^ m

in (N 00 00

00 00 00 NO r~- en JZ o d 00 ^ d d ON ^^ d ON "rt ^^ ^ S r- ^D I +1 + 1 l: 'a +1 3. +1 I +1 7 S +1 7 C:- C3 NO a. X C ^ iri vd NO _,

00 ON rj m (N p On r- r-. r^ NO q en ON -^ O NO (N On oi d rn —^ en in o i3 ^ ^^ d <3 CNl (o D- c +1 I I s +1 T !{3 +1 'f + 1 t 7 s+i 7 ~~ o ON 00 ON a- ON ON "00 >n r<-) NO r<-i 00 en en O ro en

00 O On in On (N ro — -sf 00 >. c d * d in d ^ _ \o d NO d NO O -2 \o (N . I I I + 1 +1 + 1 + 1 i o O r- CNl :s ^ ^ o o

in NO 00 ON t^ — 00 ^o in W) q O 'jr. c d (N d 06 ^ NO d _; d „ o ON o ~" CM

o ^ NO rn in NO o B £ d in NO NO NO NO S (N (N rj (N ^ o S 2 +1 ^ +1 I I ^ +1 + 1 I +1 I I NO 00 O ON 6 X) a. rn P r^ q r^ CNl in CNl in CNl in cN

"ra C3 L» O CI. -J < O u a, U 1 1 :' '

VOLUME 114, NUMBER 1 13

1—^ 5 O .&&H O (N o o O 00 ^ O rn O NO O ON o t~~ o d d o d d dS O o ""! "1 ": C " d d "1 d d +1 I ^ + o O o c S I I +1 I :4i? + 1 I 00 >o * 00 - 00 O in m On ^S fN VO — (N On NO ^ o

"rt On on NO ON a -^ ON ^ r<-) 00 o r~- ON NO '^ NO c d Tf o d d — o ^ _J ^ O |U I IT) ?^ +1 :+' ', I s +1 T ^ +,.1 T, S +1 T S + 1 S +1 I OM o f^ lO 00 o r- 00 VO NO NO NO in NO ^ NO — rj • ^ r-i r

B1 I E I

x: bO (N fN o (N o ON in NO ro r- ON sa^ -* a 3 =« ON ON NO VO >o (N 00 in r- ro 00 NO u-i (N tN 00 >n 00 o ^ o 3 U d -* ^ d t~^ ^ d NO 00 ^ d ^ d ^ d ON ^ d o6 ^ d 3 xi b c ^ X) r? 1 s:^ "^ +1 1 C: +1 S +1 + 1 7 £+1 T £ +1 7 !2 +1 'a +1 S+l 7 +1 S +1 7 n in NO NO in in r

^ ^ 'S CS cS ON o ro NO On Cfl 4) r^ "^^ ^ o ^ O ^ ON On o <^ ON i^ <^ in o ,-1 *. ON O t< O O O (^^ ;;. ^ ^ S^ ^ ^ ' « +1 • +1 +1 +1 ^ +1 ^ +1 +1 -- +1 • a .A ck C in cJn ri. C^ ^ 6 ^ ^' §^8 f<^ in r<-i xf o ^ ^ in NO NO O ON in r~; in 00 in >n NO in NO NO NO NO in in in t>n S Vha

~~o d >, (U •1— ^ e >o +3 3 c u 13 3 OX) ^o to o O (/! O. c < "o 'u C3 JO o o CO hJ pq 03 o U u a. H ja S iS"« u 'Ifl < 1 1~~

~~14 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON~~

5 60 to c 'P C/3 3 '$ '« ta " d Z E C/D "ca >i CO < O >o U CO u 'S CO ca B c c Z ca II 'a U 3 14, ca C4-1 > ^ ON tN U -a -; d u C/5 e c ca 03 c« u 3 2 60 ca < ca 'c3 O ^ -J (3n 60 CO d C/3 fc B 'c/3 0) "u d ^ u -; o '>< ca > 60 ca U _u H C •a 3 "ca '$ E 1) T3 'I c 'S 'S >, ca 60 CJ Dh

'crt 6 ca jj U 43 'S * "ft 60 E ca c^ S U CO « ,— c C/) 1) a; B c x: u to to u C3 c/} 0, c ca T3 , c OJ XI "ca C/1 > C T3 CO 60 ^ a> D — u u u u "ft H <' •T3 * >— * KH * 1—T * hH * ^ ca E C u -- ^ ON 0\ o _ "a ca _ca CO o S XI CO U ^ UU ^ U 0- "5 C u-iUr-^CXiy-iDnvdOn^Q-inU u i" Ul '3 ca 60 e .2 C/3 '> C .D 2 J^= , M-i OJ « ca E U ft ^ "ca ^ aj r >^ q c ca C C3 E V} c w >. C/3 ca u *- < CJ 45 u u U U U ^ cC E "ca CN B CJ C/5 CO G. 'S C c ,. On"OOn' 60 ca 3 CO CO (N U >n U eLivdCL,r-^CL.^Cuu-iU c CO NOJ

ca e "ca _0J C/3 E u "ft CO X & ca E >*; c/5 ca u E c CO CA > ca 6 c OJ m u "3 U UUU U cc ' > ^ ^ r * u • ' '60 t 'E E "* ' VO ft CO (N _ "a. (A (N ^ t^ c/) ( ) ^ U0N(J-H(J-.(JVPU (U E (-n U x; ca T3 E co' a en )„ "co ro a H 'C V C/5 x; Sa c *-* E 60 ca 'e C4-H H ca OJ < O X (N X2 UU H VOLUME 1 14, NUMBER 1 15

~~Colon, and Isla Escudo—in comparisons the sexes and the sex-locality interaction with those from the younger islands (Isla also were nonsignificant {F^^^ ,23 = 1.32, P~~

Popa and Isla Cristobal) and sites on the = 0.262; F,e..,oca„ty 4,23 = 1 -04, P = 0.408). adjacent mainland (Tables 2 and 3). The These tests suggest that our pooling of the discriminant function analysis showed dif- sexes and Age classes 2 and 3 was justified. ferences in both size and shape. In the DFA, Scores on PC I, however, were significantly 32 of 55 comparisons between pairs of different among localities in a one-way group centroids were significantly different ANOVA (F,o,46 = 21.79, P < 0.001). Pat- and followed a clear pattern (Table 3, Fig. terns of localities differing significantly on

3). The sloths on Isla Escudo were mor- PC I paralleled the results for the first ca- phometrically distinct from all other popu- nonical axis (Table 3). PC II and PC III lations. Furthermore, sloths on the outer is- represent shape differences, but given the lands were generally distinct from those of multiple groups involved they probably do Isla Popa and Isla Cristobal as well as from not represent the most appropriate way to those from all mainland sites in Bocas del examine shape.

Toro. The first canonical axis accounted for In comparing the small sloths from Bo- most (63%) of the variation among groups, cas del Toro with six series of Bradypus and the first and second together encom- variegatus from its range outside Bocas del passed 76% of the variation (Table 4). The Toro, only the population on Isla Escudo is first canonical axis may be interpreted as significantly smaller than all mainland sam- general size, with smaller sloths having ples (Table 6) in the greatest length of skull. lower scores. Most measurements loaded Sloths from the other differentiating popu- strongly and positively on this axis; exter- lations in Bocas del Toro—Cayo Agua, nal auditory meatus diameter (EAM) had a Cayo Nancy, Isla Bastimentos, and Isla Co- negative loading, but its magnitude was lon—fall within the size variation in B. var- small enough to be negligible, indicating iegatus from Colombia and Brazil. that EAM diameter did not correlate with general size. The second canonical axis Discussion contrasted tail length and EAM diameter to squamosal process breadth and ascending The three-toed sloths of the outer islands mandibular ramus breadth (Table 4). Sloths of Bocas del Toro—Isla Colon, Isla Basti- with relatively long tails, large EAMs, nar- mentos, Cayo Nancy, Cayo Agua, and Isla row squamosal processes, and narrow rami Escudo—are significantly smaller than the scored high on this axis. Bradypus of the adjacent mainland of Bo- In the principal components analysis of cas del Toro, as evidenced by Canonical

overall variation in the sloths of Bocas del axis 1, Principal component I, and univar- Toro, the first component (PC I) accounted iate statistics. Furthermore, sloths on those for 51% of the variation among individuals, five islands themselves vary in mean size, without regard to locality (Table 5). Speci- with those from Isla Escudo being the mens from Isla Escudo again plotted far smallest (Tables 2 and 3). The samples from from all other specimens from Bocas del Isla Popa and Isla Cristobal, which are Toro (Fig. 4). The loadings on PC I indicate young islands close to shore, are not sig- that it represents overall size, with EAM nificantly different in size from sloths on loading so slightly as to be immaterial. No the mainland of Bocas del Toro (Table 3). difference between Age classes 2 and 3 or Sloths on Isla Escudo display differences age-locality interaction was detected for PC in cranial shape when compared with other I scores in a general linear model = populations of Bradypus from Bocas del (F^ê i 6 = 0.23, P = 0.650; F.êîocauty 1,6 = O.od, P Toro. The position of specimens from Isla 0.981). Similarly, the difference between Escudo on the second canonical axis (Fig. 16 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

~~> A All individuals 4 - V Csl A W A x A o m • (0 o °o • o oO o o "c • o o ^ On • • • c O O (§)« Q oCO u o~~

~~r I 1 -3~~

~~B Group centroids 4 - aE CN /^f^ (/} A 'x CD oN Id g - 'c o COo oCA TO* \ c CD Bo o -4 -~~

~~1 1 1 1 -6 -3~~

~~Canonical axis 1~~

Fig. 3. Plot of specimen scores (A, upper) and locality centroids (B, lower) on the first two canonical axes from a multiple-group discriminant function analysis of three-toed sloths (Bradypus) from Bocas del Tore, Panama, showing morphometric distinctiveness of sloths of Isla Escudo and moderate dwarfing on four other islands. In A, specimens from Isla Escudo are marked with solid triangles; specimens from the four islands of intermediate age (Cayo Agua, Cayo Nancy, Isla Colon, and Isla Bastimentos) are represented by open circles; and specimens from the mainland and from the two youngest islands (Isla Cristobal and Isla Popa) are denoted by solid circles. Abbreviations for locality centroids in B follow, with symbols following the same scheme as in A; A, Almirante; B, Isla Bastimentos; CA, Cayo Agua; CR, Isla Cristobal; CO, Isla Colon: E. Isla Escudo;

~~N, Cayo Nancy; NU, Nuri; P, Isla Popa; TO, Tierra Oscura; V, Peninsula Valiente.~~

3) is due primarily to their large external mainland populations that were once con- auditory meatus, narrow squamosal pro- tiguous (Table 1, Fig. 1). These analyses cess, and narrow ascending mandibular ra- show that the small sloths on the outer is- mus (Table 2, Table 4). For the other island lands share no discrete pelage characters. and mainland populations in Bocas del The pelage traits are independent of body Toro, variation in shape is minimal as com- size, which is notably predisposed to con- pared with differences in size. vergence (Roth 1992). The few cranial Although the sloths from the outer is- characters common to the small sloths on lands share small size, our examinations of various islands are all gracile traits associ- pelage show similarities between island and ated with size reduction and ontogenetic —

~~VOLUME 1 14, NUMBER 1 17~~

Table 4.—Loadings (correlation coefficients) of data, with an area cladogram based on the external and 14 cranial measurements (log,o- three sequence of island formation derived from transformed) on the first three canonical axes of a mul- sea level mapping (Brooks McLennan tiple-group discriminant function analysis of three-toed & sloths {Bradypus) from 1 1 localities in Bocas del Toro, 1991:197-198, Avise 1994). Panama. Eigenvalues and the corresponding cumula- Thus, we suggest that the evolution of tive percent of total dispersion explained are given for smaller body sizes occurred at least four each axis. For the analysis, specimens were grouped times in Bocas del Toro: independently on by collection locality. See Materials and Methods for Isla Escudo, Isla Colon, and variable abbreviations. The first canonical axis repre- Cayo Agua sents a measure of general size. each of which formed separately—and once on Cayo Nancy and Isla Bastimentos to-

C 1 C 2 C 3 gether (they became isolated from the TOTAL 0.804 0.026 0.081 mainland as a unit and only recently have TAIL 0.423 0.620 -0.136 become separated from each other). The HF 0.822 -0.158 0.252 sloths from the outer islands are not linked GLS 0.895 0.046 0.132 a corrunon ancestry as might be pre- AZB 0.856 -0.141 0.096 by PZB 0.797 -0.178 -0.004 sumed because of their small size, but rath- FOB 0.742 0.027 0.268 er they adapted separately as isolated pop- -0.324 -0.058 SPL 0.795 ulations while experiencing similar envi- MTRL 0.452 0.009 0.202 romnental changes following insularization, PPL 0.862 0.124 0.202 PB 0.716 0.027 0.398 in an instance of parallel evolution. The BD 0.885 -0.027 0.136 sloth of Isla Escudo has clearly reached the ABAR 0.760 -0.258 0.107 species level, but we consider that the pop- DJPL 0.236 0.160 -0.199 ulations on Cayo Agua, Cayo Nancy, Isla EAM -0.152 0.437 0.515 Bastimentos, Isla con- SPB 0.477 -0.467 -0.355 and Colon remain ARB 0.681 -0.481 0.070 specific with Bradypus variegatus. Future will the factors that Eigenvalue & 17.7800 3.6367 2.2703 work evaluate may (cumulative % of have led to these instances of dwarfism. total dispersion) (63.3%) (76.3%) (84.4%) We have documented the extremely small size of the three-toed sloth on Isla

Escudo, as well as its unique cranial and truncation (e.g., thin zygomatic arches, pelage characters relative to other known weakly developed temporal crests). The species of the genus. For these reasons, we geographic distribution of the pelage traits here formally describe this endemic insular apparently represents the relictual manifes- population as: tation of previously continuous geographic variation that was subdivided into isolated populations when the islands formed. Bradypus pygmaeus, new species Movement of sloths between islands or to Fig. 5 or from the mainland probably has been in- significant. We propose that those popula- Holotype.—\]SNM 579179, adult fe- tions independently underwent selection for male, skin and skull (Fig. 5), collected on smaller size when separated from the main- 27 Mar 1991 by Charles Handley and Pen- land, under a vicariant model consistent ny Nelson, from Panama: Bocas del Toro: with the patterns of island formation elu- Isla Escudo de Veraguas, West Point. Orig- cidated by Handley & Vam (see Introduc- inal number EPN 166. Also examined: nine tion). This hypothesis of strict vicariance is paratypes from Isla Escudo: USNM currently being tested by Anderson and L. 578413, 579171-579175, 579176 (returned Olson by comparing a population-level to Panama-INRENARE), 579177-579178. phylogeny produced from DNA sequence Etymology.—From the Latin pygmaeus, 18 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

~~A • C 0.1 - c0) • • % o o o ^ Q. ^ o •• A & * E o 0.0 - O o O 0»» rm "to Oo o • Q. o • O -0.1 - ^o^ c o o ^~~

~~. -~~

~~1 1 -0.2 0.0 0.2~~

~~Principal component~~

Fig. 4. Plot of specimen scores on the first two axes of a principal components analysis of three -toed sloths (Bradypus) from Bocas del Toro, Panama, showing dwarfing of sloths on Isla Escudo and four other islands. Specimens from Isla Escudo are marked with solid triangles; specimens from the four islands of intermediate age (Cayo Agua, Cayo Nancy, Isla Colon, and Isla Bastimentos) are represented by open circles; and specimens from the mainland and from the two youngest islands (Isla Cristobal and Isla Popa) are denoted by solid circles.

~~The first principal component represents general size.~~

meaning dwarf or pygmy. Suitable vernac- mm; tail, X = 49.7 mm, range 45-60 mm; ular name is pygmy three-toed sloth. hind foot, X = 102.4 mm, range 94-110 Distribution. —Known only from Isla Es- mm; mass, X — 2.9 kg, range 2.5-3.5 kg); cudo de Veraguas, Province of Bocas del dorsal coloration usually blotchy and al-

Toro, Republic of Panama, where it is ways with a midsagittal stripe; adult males found exclusively in red mangroves at sea with orange speculum, woolly around an- level (Fig. 6). terior margin; fur of crown long and shag- Diagnosis. —A species of Bradypus char- gy, hanging over short hair of face to give acterized by the following combination of a hooded appearance; brow very dark with characters: size small (Table 2); orange abrupt posterior termination (line on brow speculum on dorsum of adult males; nape present); face buff with orange wash around without black mane; face tan with distinc- dark eye stripe; throat grizzled gray-brown. tive dark band across forehead; long hair of Skull small (n = 6 adults: greatest length, forehead hanging over face, giving the im- X = 69.1 mm, range 67.5-72.2 mm; ante- pression of a hood; pterygoids not inflated; rior zygomatic breadth, X — 41.5 mm, no foramina in anterodorsal nasopharynx; range 38.3-45.7 mm; maxillary toothrow frontal sinuses swollen; stylomastoid fora- length, X = 23.3 mm, range 22.3-24.7 mm; men miniscule, external carotid foramen see also Table 2) and gracile (Fig. 5); pa- usually absent or minuscule; external audi- rietal ridges weak and usually convex tory meatus large; ventral surface of hyoid (hourglass-shaped when viewed dorsally); (stylohyal) smoothly concave; descending masseter-temporal fossa rather smooth; process of jugal long and thin; coronoid pterygoids thin, not inflated; no foramina process of mandible thin and strongly fal- present in anterodorsal nasopharynx; pre- cate. maxillae minute, barely if at all articulated Description. —Size small (n = 7 adults: with maxillary; zygomatic arch incomplete, total length, X = 505.4 mm, range 485-530 anterior and posterior roots slender; de- VOLUME 1 14, NUMBER 1 19

Table 5.—Loadings, eigenvalues, and cumulative scending process of jugal long and thin; explained for the first three axes percent of variance lambdoidal crest continuous across poste- of a principal components analysis undertaken on rior margin of occiput, lateral margins of three-toed sloths (Bradypus) from Bocas del Toro, crest straight lateral re- Panama, using the covariance matrix of log,o-trans- in view; occipital formed values of 14 cranial measurements and three gion barely projecting posterior to lamb- external measurements for each of 57 individuals. See doidal crest; hyoid (stylohyal) smoothly Materials and Methods for abbreviations. The first arched on ventral surface to point of artic- principal component represents general size, which is ulation with epihyal on anterior limb; pos- uncorrelated with external auditory meatus diameter (EAM). terior limb of hyoid usually wider than anterior limb; external auditory meatus large;

PC I PC II PC III stylomastoid foramen minute; external carotid usually closed or miniscule; TOTAL 0.786 0.302 0.103 foramen TAIL 0.331 0.788 0.456 ventral mandibular surface strongly con- HP 0.801 0.141 -0.057 cave; coronoid process of mandible thin GLS 0.878 0.221 0.075 and strongly falcate. AZB 0.905 0.049 -0.218 Tooth formula: (anterior chisel-shaped PZB 0.895 -0.010 -0.214 teeth 1/1, molariform teeth 4/3) X 2 = 18 FOB 0.682 0.129 0.044 SPL 0.874 -0.062 0.073 (terminology of Naples 1982). Upper ante- MTRL 0.610 0.164 0.214 rior chisel-shaped tooth tiny or absent; low- PPL 0.839 0.269 0.017 er anterior chisel-shaped tooth anteropos- PB 0.712 0.236 -0.042 teriorly compressed. BD 0.794 0.054 -0.187 Measurements the holotype. Total ABAR 0.866 -0.092 -0.419 of — DJPL 0.354 0.348 -0.251 length, 510 mm; tail length, 54 mm; hind EAM -0.044 0.481 -0.252 foot length, 94 mm; mass, 3.5 kg. Cranial SPB 0.689 -0.477 0.455 measurements (in mm): GLS, 68.8; AZB, -0.191 ARB 0.842 0.200 43.6; PZB, 40.7; POB, 20.5; SPL, 21.3; Eigenvalue & 0.0164 0.0046 0.0033 MTRL, 22.8; PPL, 33.3; PB, 16.3; BD, (cumulative % 24.7; ABAR, 3.2; DJPL, 16.2; EAM, 6.7; variance explained) (51.2%) (65.5%) (75.9%) SPB, 5.0; ARB, 12.8.

Table 6. —Results of ANOVA of greatest length of skull (GLS) measurements. All possible pairwise comparisons were made among samples of Bradypus from the five outer islands of Bocas del Toro, Panama and six samples of B. variegatus from mainland localities outside Bocas del Toro using Tukey's procedure with a family- wide error rate of a = 0.05. Results for comparisons of island populations vs. mainland populations outside Bocas del Toro are presented here. Significant comparisons are marked with an asterisk (*), whereas nonsignificant ones are denoted "«..s." Bonanza, El Recreo, and Tepeyak are localities in Nicaragua; Bajo Calima-Rio

~~Raposo and Tumaco lie along the southern Pacific coast of Colombia; and Mojui dos Campos is located in the~~

lower Amazon of Brazil. Note that sloths from Isla Escudo are significantly smaller than those of all mainland localities outside Bocas del Toro, whereas the Bradypus from the four other outer islands of Bocas del Toro fall within the range of variation of B. variegatus in South America. Descriptive statistics (mean ± 2 standard errors, minimum-maximum, and sample size) are given here in mm for mainland localities outside Bocas del Toro; see Table 2 for descriptive statistics of localities in Bocas del Toro.

Bajo Calima- Bonanza El Recreo Tepeyak Ri'o Raposo Tumaco Mojui dos Campos 78.9 ± 1.12 78.5 ± 1.15 83.2 ± 3.08 73.7 ± 2.18 76.9 ±4.18 72.8 ± 1.15 75.3-82.3 76.1-80.7 80.0-86.5 68.9-76.8 70.6-79.3 67.0-78.3 (12) (7) (4) (6) (4) (21)

~~Isla Colon * * * n.s. n.s. n.s.~~

~~Isla Bastimentos *** n.s. * n.s.~~

~~Cayo Nancy n.s. n.s. * n.s. n.s. n.s.~~

~~Cayo Agua * * * n.s. n.s. n.s. Isla Escudo * * * * * * 20 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON~~

Fig. 5. Dorsal, ventral, and lateral views of the crania and lateral view of the mandibles of the holotype of Bradypus pygmaeus (USNM 579179, right) from Isla Escudo and a specimen of B. variegatus from the adjacent mainland on the Peninsula Valiente (USNM 578423, left). Note the diminutive size and gracile qualities of B. pygmaeus and the open external carotid foramen (ECF) of USNM 578423, which is closed in USNM 579179. Also note the strongly falcate coronoid process on the mandible of B. pygmaeus and the large external auditory

~~meatus (EAM) characteristic of that species, despite its overall small size. —~~

~~VOLUME 114. NUMBER 1 21~~

~~85°~~

~~Bradypus variegatus o Bradypus pygmaeus -^~~

~~10°~~

~~Pacific Ocean~~

Fig. 6. Verified Central American distribution of Bradypus. Georeferenced collection localities reported here are plotted for B. variegatus (open circles) and B. pygmaeus (star). The distribution of Bradypus variegatus continues northward into Honduras (voucher specimens from the Rio Patuca region [see Specimens examined] and additional sight records reported in Marineros & Martinez-Gallegos 1998) and south into South America (see Fig. 7). Bradypus pygmaeus is endemic to Isla Escudo de Veraguas in western Panama. See Specimens examined sections for full provenience and museum catalogue numbers.

~~Comparisons.—Compared with popula- Discrete cranial characters separate Bra- tions of Bradypus variegatus on the adja- dypus pygmaeus from all other species of~~

cent mainland, B. pygmaeus averages ap- the genus. It lacks the distinctly inflated proximately 40% smaller in mass, 15% pterygoid sinuses and the two or three smaller in total length, and 12—16% smaller roughly circular foramina in each side of

in most cranial dimensions. It is smaller the anterodorsal nasopharynx of B. torqua- than any studied population of Bradypus tus (Wetzel & Avila-Pires 1980, Wetzel variegatus in Central or South America (Ta- 1985). It lacks the pair of oblong foramina bles 2, 3, and 6). Additionally, its external present in the medial roof of the anterodor-

auditory meatus is conspicuously large for sal nasopharynx of B. tridactylus (Wetzel a sloth of overall small size. The diameter 1985). From populations of the closely re- of the EAM decreases as the bulla ossifies; lated B. variegatus—including both main- the development of this region of the skull land sloths and the moderately dwarfed is especially truncated in B. pygmaeus (Fig. sloths of other islands in Bocas del Toro 5). the pygmy sloth differs in having (Fig. 5): 22 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON external carotid foramen, through which the de Veraguas (9°06'N, 81°33'W), West carotid artery normally passes, absent or Point, 9 (USNM 579171-579175, 579176*, minute in most specimens (the homologous 579177-579179). foramen is markedly larger in B. variegatus, suggesting a different pattern of cranial Bradypus torquatus Illiger, 1811 circulation); stylomastoid foramen at the Distribution. —Restricted to the Atlantic posterior external base of the auditory bulla forests of southeastern Brazil (Fig. 7; Wetz- tiny (the facial nerve exits this foramen in el & Avila-Pires 1980). B. variegatus, in which the foramen is Diagnosis. —Size large; no speculum on markedly larger and almost always visible dorsum (contra Eisenberg & Redford 1999: to the naked eye); external auditory meatus 94); nape with distinct black plume or large (usually smaller in B. variegatus and mane; facial and body pelage grizzled; hair all other three-toed sloths); ventral edge of of forehead short; pterygoids distinctly in- stylohyoid usually smoothly concave (an- flated; two or three small, circular foramina gular or undulating in B. variegatus); cor- present on each side of anterodorsal naso- onoid process of the mandible slender and pharynx. strongly falcate (usually thick and straight- Comparisons. —This species is easily edged or only moderately curved in B. var- separated from all other species of the ge- iegatus). No other insular population of nus by its black dorsal mane and inflated three-toed sloth in Bocas del Toro shows pterygoids. autapomorphic cranial characters. Comments. —Extremely rare in museum Externally, Bradypus pygmaeus may be collections. Endangered due to deforesta- separated from B. torquatus of both sexes tion in its restricted range (Emmons & Feer by the lack of a black dorsal mane origi- 1997). Considered the most basal species of nating at the nape and by the presence of Bradypus (Wetzel & Avila-Pires 1980) and short, tan facial pelage with a black stripe placed in its own subgenus, Scaeopus. lateral to the eye; and in adult males by Specimens examined. —Bradypus torqua- possessing a dorsal speculum. From B. tri- tus, total 4. Brazil (4). Bahia. Itabuna, near dactylus, the pygmy sloth is distinguished Ilheus [14°48'S, 39°16'W], 1 (USNM by its tan facial and gular pelage and dark 259473); Tres Bracos, Fazenda Piabanha stripe lateral to the eye. In contrast, B. tri- (13°32'S, 39°45'W), 37 km N, 34 km E Je- dactylus has brilliant golden hair on the quie, 2 (USNM uncatalogued, field num- brow, face, and throat. The pygmy sloth dif- bers MTB 1706-1707). State unknown: no fers externally from B. variegatus by long specific locality, 1 (MCZ 1024). hair projecting over the brow, creating the aspect of a hood (Table 1). This character Bradypus tridactylus Linnaeus, 1758 provoked Handley to refer to this sloth in the field as the "monk sloth." No other Distribution.—The Guianas and adjacent sloth of the Bocas islands is hooded, and regions of eastcentral Venezuela (Estado the "ruff" on the brow (hair projecting Bolivar) and northcentral Brazil, principally over the forehead) mentioned for other north of the Amazon (Fig. 7). Distribution Central American Bradypus by Goldman probably does not extend southwest of the (1920:57), Goodwin (1946:352), Hall Rio Negro or as far south of the Amazon (1981:279), and others is not so long and as indicated in Eisenberg & Redford distinctive. (1999), where it is replaced by Bradypus Specimens examined. —Bradypus pyg- variegatus. maeus, total 10. Panama (10). Bocas del Diagnosis. —Size average for genus; or- Toro: Isla Escudo de Veraguas (9°06'N, ange speculum present on dorsum of adult

~~81°33'W), 1 (USNM 578413); Isla Escudo males; nape without black mane; face and VOLUME 114, NUMBER 1 23~~

~~Bradypus variegatus o Bradypus tridactylus a Bradypus torquatus~~

~~Pacific Ocean~~

Fig. 7. Verified South American distribution of Bradypus. Georeferenced collection localities are mapped for B. variegatus (open circles), B. tridactylus (solid triangles), and B. torquatus (solid squares). The distribution of Bradypus variegatus continues northwest into Central America (see Fig. 6) and south to Argentina (voucher specimen from Jujuy province, see Specimens examined). Wetzel and Avila-Pires (1980) reported additional distributional records for B. torquatus, which ranges southward from the localities of that species confirmed and plotted here. See Specimens examined sections for full provenience and museum catalogue numbers.

forehead golden with no dark stripe at level in South America possess golden faces, of eyes (contra Emmons & Peer 1997:43); however, and a few Brazilian populations throat golden to the base of the hairs, or even have throats frosted with golden- predominantly golden with bases of hairs tipped fur (e.g., localities on the lower Rio smoky gray; hair of forehead short and stiff; Tapajos), but the base of the gular hairs is pterygoids not inflated; a single pair of ob- characteristically brown for most of the long foramina present in the anterodorsal length of the hair in those populations. The nasopharynx; frontal sinuses seldom swol- golden facial and gular hair of B. tridacty- len. lus is generally shorter and stiffer than in Comparisons. —Bradypus tridactylus B. variegatus. See Comments in B. varie- may be easily separated from B. torquatus gatus. No B. pygmaeus have golden throats by the lack of a black dorsal mane or in- or faces. The dorsum of B. tridactylus is flated pterygoid sinuses. It is most similar often speckled or blotchy, but this character

~~to B. variegatus, whose range it probably does not serve to distinguish it from B. var-~~

contacts in Venezuela and Brazil. It differs iegatus (contra Eisenberg 1989, Eisenberg from that species by possessing a pair of & Redford 1999), which often displays this oblong foramina in the anterodorsal naso- trait as well. pharynx, and by its golden throat. All B. Comments.—In the older literature, in- tridactylus have golden faces and throats, dividuals of Bradypus variegatus were of- either golden to the base of the hair or with ten incorrectly reported as B. tridactylus. a slight smoky gray color at the base of the Specimens examined. —Bradypus tridac- hairs. In contrast, most B. variegatus have tylus, total 50. Brazil (6). Amazonas: Ma-

~~tan faces. Many specimens of B. variegatus naus [3°08'S, 60°01'W], 1 (AMNH 143012); 24 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON~~

Rio Amazonas, bought in Manaus [3°08'S, Bradypus variegatus Schinz, 1825 60°01'W], 4 (FMNH 165441-165444); Rio Distribution. From eastern Honduras to Amazonas, Manaus [3°08'S, 60°01'W], Ha- — northern Argentina (Wetzel & Avila-Pires cienda Rio Negro, 1 (AMNH 78968). 1980, McCarthy et al. 1999). Widespread in French Guiana (1). Cayenne [4°56'N, forested areas at low-to-middle elevations 52°20'W], 1 (AMNH 77891). Guyana (20). of eastern Central America (Fig. 6); South Cuyuni-Mazaruni: Essequibo, Kartabo Point America west of the Andes to southern Ec- [6°23'N, 58°4rW], 12 (AMNH 42454, uador; east of the Andes in South America 42871-42872, 42888, 48369, 74134-74137, throughout Amazonian forests (but not in 142932, 142934, 142992); Kalacoon [= Ka- the Guianan region, where replaced by B. lakun, 6°24'N, 58°39'W], 2 (AMNH 48103, tridactylus); and in some areas of south- Kartabo [6°23'N, 58°41'W], 1 269846); eastern Brazil and northern Argentina (Fig. Kartabo River, 1 (AMNH 48180); (AMNH 7). Absent from the open llanos of Colom- 74131); Kyk-over-al [island in Mazaruni bia and Venezuela, the Brazilian cerrado, River facing Karatabu Point (= Kartabo and other savanna habitats—contrary to Point), AMS 1944], 1 (AMNH 48104). De- distributional maps provided by Emmons & merara-Mahaica: Dunoon [6°25'N, 58°18'W, Feer (1997) and Eisenberg & Redford Hershkovitz 1977], 1 (UMMZ 46410). Up- (1999). per Takutu-Upper Essequibo: Dadanawa Diagnosis. —Size variable but most pop- [2°50'N, 59°30'W], 20 mi E, 1 (USNM ulations averaging at least 72 mm in GLS; 362241); no specific locality [in former Ru- orange speculum present on dorsum of pununi], 1 (USNM 395070). Suriname (7). adult males; nape without black mane; face Brokopondo: Saramacca Rivier, Loksie Hat- tan or golden—if golden, hairs usually with tie [5°10'N, 55°28'W, Hershkovitz 1977], 1 dark brown bases; face usually with dark (FMNH 95443). Paramaribo: near Paramar- band lateral to eye; throat brown or occasionally brown frosted with golden; hair of ibo [5°50'N, 55°10'W], 1 (MCZ 19570); forehead variable in length but never hang- Paramaribo [5°50'N, 55°10'W], 900 ft, 2 (FMNH 93297, 95446); Paramaribo ing over face giving the appearance of a hood; pterygoids not inflated; no foramina [5°50'N, 55°10'W], brush land, 900 ft, 1 present in anterodorsal nasopharynx; fron- (FMNH 93296). Saramacca: La Poule, 2 tal sinuses often but not always well-swol- (FMNH 95444-95455). Venezuela (16). Bo- len; stylomastoid foramen large; external livar: Camarata Valley, 450 m, 1 (AMNH carotid foramen large; external auditory 135474); Ciudad Bolivar [8°08'N, meatus medium in size; ventral surface of 63°33'W], 3 (AMNH 16134-16136); El hyoid (stylohyal) distinctly bent or undu- Manaco (6°17'N, 61°19'W), 59 km SE El lating, not smoothly concave; descending Dorado, 150 m, 1 (USNM 374821); La process of jugal variable, but usually rela- Bomba [7°02'N, 61°33'W], 1 (AMNH tively short and robust; coronoid process of 30738); Los Patos (7°11'N, 62°22'W), 25 mandible thick. km SE El Manteco, 350 m, 2 (USNM Comparisons. —This species lacks the 374822, 387803); Maripa [7°26'N, black dorsal mane and inflated pterygoid si- 65°09'W], 2 (AMNH 21305-21306); Rio nuses characteristic of B. torquatus. Bra- Suapure [6°48'N, 67°01'W], 4 (AMNH dypus tridactylus has a pair of oblong fo- 16932-16934, 17560); Rio Supamo ramina in the anterodorsal nasopharynx (7°00'N, 62°15'W), 50 km SE El Manteco, lacking in B. variegatus (although in young

~~150 m, 1 (USNM 374818); Rios Mato of Age class 1, the roof of the nasopharynx~~

~~[7°09'N, 65°07'W] and Caura, 1 (AMNH is poorly ossified, complicating the identi- 30201). fication of newborn and juvenal individuals —~~

~~VOLUME 1 14, NUMBER 1 25~~

of B. variegatus and B. tridactylus). Also, only named insular form of the genus prior B. variegatus has a brown throat (rarely to this study) does not show dwarfing sim- frosted with golden), in contrast to B. tri- ilar to that of B. pygmaeus, and it is con- dactylus, which has a brilliant golden throat specific with B. variegatus. Bradypus gor- with hairs golden to the base of the hairs, gon is from Isla Gorgona, an island with an or with only a slight smoky gray tint to the area of 15.6 km^ (Aguirre-C. & Rangel-Ch. bases. See account of B. pygmaeus for com- 1990) approximately 30 km off the south- parisons with that closely related species. western coast of Colombia. Isla Gorgona is Comments. —Bradypus variegatus is the located on the continental shelf of South only species of the genus that displays no- America (von Prahl 1986) and has strong table geographic variation. In Central biological affinities with the lowlands of

America, it almost always has a tan face. western Colombia and Ecuador (Alberico Many specimens from South America—es- 1986, Rangel-Ch. 1990b). Although its geo- pecially from southwestern Colombia, logical history remains controversial

western Ecuador, and northcentral Brazil (Aguirre-C. & Rangel-Ch. 1990), it may possess strikingly golden faces, although represent the tip of a sunken volcanic peak the bases of the facial hairs are usually dark belonging to a fourth (coastal) Colombian brown. A few populations in northern Bra- Cordillera (Haffer 1970, Alberico 1986). zil (e.g., on the lower Rio Tapajos) also This coastal cordillera, or at least an arc of show a strong golden frosting on the throat. volcanic islands, was formed at the end of This species also varies widely in the the middle Eocene, including the Serrania

blotchiness of its dorsal coloration. Crani- del Baudo, Serrania de los Saltos, and Alto ally, specimens from west of the Andes de Nique, in western Colombia and extreme tend to have more elongated, strongly hour- eastern Panama (Haffer 1970, see also glass-shaped skulls, whereas many popula- Hershkovitz 1969, Coates & Obando 1996). tions east of the Andes possess proportion- In the late Pliocene, the Atrato-San Juan sea ately shorter, wider skulls. All populations corridor (= Bolivar Geosyncline) closed, agree with the diagnoses of Wetzel (1985) uniting these volcanic blocks with the main in lacking foramina in the anterodorsal na- body of South America (Alberico 1990, sopharynx. Coates & Obando 1996). Subsequently, gla- Some populations show moderate dwarf- cial cycles alternately raised and lowered ing in size. The series from Mojui dos Cam- sea level; during one or more of these epi- pos represents one of the few such South sodes, Isla Gorgona presumably became American populations. In Central America, isolated from the adjacent Chocoan low- populations of Bradypus variegatus on sev- lands. eral islands of the Laguna de Chiriqui in Thomas justified naming Bradypus gor- Bocas del Toro, Panama average smaller gon partly on the basis of size: "Size small, than most but not all populations of the spe- about as in tridactylus, infuscatus, and cies that we examined from the mainland ephippiger, the skull markedly smaller than (Tables 2 and 6). If similar series were in the Ecuadorean macrodon'' (Thomas available from throughout the species' 1926:309—310). Bradypus infuscatus, B. range, a detailed study of geographic vari- ephippiger, and B. macrodon are currently ation might show that these populations de- considered junior synonyms of B. variega- serve subspecific status. Given the present tus (see Gardner 1993). Three of the four state of understanding of geographic varia- taxa that Thomas compared with the sloth

tion within B. variegatus, however, it is pre- from Isla Gorgona occur only east of the mature to recognize subspecies of this Andes, making them poor comparisons. wide-ranging and highly variable species. Thomas (1917) restricted the fourth one, B. Bradypus gorgon Thomas, 1926 (the ephippiger, to NW Colombia, and Cabrera 26 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

(1957) further restricted it to the Rio Atrato (USNM 238668); Rio Mamore [10°23'S, region. Thus, a better comparison for the 65°23'W], 5 km S Guayaramerin, 1 (AMNH Bradypus from Isla Gorgona would have 209940); Rio Mamore [10°23'S, 65°23'W], 8 been with B. epipphiger rather than with B. km N Exaltacion, 1 (AMNH 211663). Co- macrodon, as the occurrence of B. ephip- chabamba: Todos Santos [16°48'S, 65°08'W], piger west of the Andes in Colombia places 1 (AMNH 38784). Santa Cruz: Buena Vista it in the biogeographic region from which [17°27'S, 63°40'W], 4 (AMNH 61792; the sloth on Isla Gorgona almost certainly FMNH 21393-21394, 21430); Buena Vista was derived. [17°27'S, 63°40'W], 450 m, 1 (FMNH Contrary to Thomas' conclusions, in the 51871); Rio Surutu [17°24'S, 63°51'W], 400 two specimens from Isla Gorgona for which m, 1 (AMNH 61791); Rio Yapacani we have measurements (both adults), the [16°00'S, 64°25'W], 1 (FMNH 51870); Santa greatest length of the skull averages near Cruz de la Sierra [17°48'S, 63°10'W], 1 that of adults from the closest available (AMNH 133435); 5 km E Rio PalometiUas, mainland populations in southwestern Co- 300 m, 1 (AMNH 261304). Brazil (130). lombia (Isla Gorgona: GLS, X = 16.1 , stan- Amazonas: Rio Amazonas, Santo Antonio de dard error of the mean (SEM) = 1.20, n = Amatary, 1 (AMNH 93103); Rio Amazonas, 2; Rio Raposo & Bajo Calima: GLS, X = south bank. Villa Bella Imperatriz [= Parin- 73.7, SEM = 1.09, « = 6; Tumaco: GLS, tins, 2°36'S, 56°44'W], 12 (AMNH 93104- X = 76.3, SEM = 2.85, n = 3). We agree 93115); Rio Madeira, Rosarinho [3°42'S, with Wetzel & Avila-Pires (1980) and 59°08'W], 4 (AMNH 92335, 92828-92829, Gardner (1993) in considering Bradypus 92845); Rio Madeira, Rosarinho, Santo An- gorgon to be conspecific with B. variega- tonio de Uayara, 3 (AMNH 92332-92334); tus. Rio Negro, Cacao Pereira [3°08'S, 60°05'W], Isla Gorgona is twice as far from the Igarape, 2 (AMNH 80447-80448); Rio Ne- mainland as any of the islands of Bocas del gro, launari [0°31'S, 64°50'W], 1 (AMNH Toro, falls within the size range of those 79396); Rio Negro, Manaus [3°08'S, islands, and surely has been isolated longer, 60°01'W], 1 (AMNH 91353); Rio SoUmoes, since it is separated from the mainland by Codajas [3°50'S, 62°05'W], 1 (FMNH water about 70 m deep (Alberico 1986), 50906); Sohmoes, 1 (AMNH 37155). Espi- whereas the greatest depth of water sepa- ritu Santo: Lagoa Juparana [19°20'S, rating Isla Escudo from the adjacent main- 40°04'W, Hershkovitz 1977], 4000 ft, 1 land is ca. 29 m (Kalko & Handley 1994). (AMNH 78844). Para: Altamira, 85 km SW,

This raises the question of why the three- east bank Rio Wri (3°50'S, 52°40'W), 1 toed sloth on Isla Gorgona has not under- (USNM 549523); Belem [1°27'S, 48°29'W1, gone a decrease in size similar to that of 2 (MCZ 31001; USNM 393816); Belem Bradypus pygmaeus on Isla Escudo. We [1°27'S, 48°29'W], Utinga, 2 (USNM speculate that dwarfism in B. pygmaeus 339631-339632); Curralinho, 2 (AMNH may be related to foraging in mangroves, 133438, 133457); CurraUnho, Ilha de Marajo which are absent from Isla Gorgona (Ran- [1°00'S, 49°30'W], 8 (AMNH 133406, gel-Ch. 1990a). Another scenario that 133415, 133419, 133421, 133426, 133432- should not be dismissed is the possibility 133433, 133455); Ilha de Marajo [1°00'S, that three-toed sloths could have been re- 49°30'W], 4 (FMNH 34401, 34712-34714); cently introduced to Isla Gorgona from the Patagonia, 12 mi, 2 (AMNH 75140-75141); mainland by humans (Alberico 1986). Rio Amazonas, Igarape Piaba [1°55'S, Specimens examined. —Bradypus variega- 55°33'W], 3 (MCZ 30993, 30995, 31002); tus, total 467. Argentina (1). Jujuy: no spe- Rio Majary, Recreio [1°42'S, 52°12'W], 1 cific locahty, 1 (FMNH 21672). BoUvia (13). (AMNH 95841); Rio Tapajos, Aramanay

~~Beni: Beni River [10°23'S, 65°24'W], 1 [2°45'S, 54°59'W, Hershkovitz 1977], 3 VOLUME 1 14, NUMBER 1 27~~

(AMNH 95101-95103); Rio Tapajos, Caxir- 77°41'W], 100 m, 1 (FMNH 90060). Cesar: icatuba [2°50'S, 55°08'W], 1 (AMNH Colonia Agricola de CaracoUcito [10°18'N, 95104); Rio Tapajos, east bank, Fordlandia 74°00'W, Hershkovitz 1947], 2 (USNM

~~[3°40'S, 55°30'W], 1 (FMNH 94551); Rio 281352-281353); Valledupar, Rio Cesar, El Tapajos, Igarape Amorin [2°26'S, 55°00'W, Orinoco [10°09'N, 73°26'W, Hershkovitz~~

~~Hershkovitz 1977], 1 (AMNH 95329); Rio 1947], 1 (USNM 281354). Choco: Andagoya~~

~~Tapajos, Igarape Bravo [2°26'S, 55°00'W, [5°06'N, 76°41'W], 1 (FMNH 86760); Jurado~~

~~Hershkovitz 1977], 2 (AMNH 95105- [7°07'N, 77°46'W], 100 m, 1 (UC 3909);~~

~~95106); Rio Tapajos, Inajatuba, 4, (AMNH Quibdd [5°42'N, 76°40'W], 1 (AMNH 95325-95328); Rio Tapajos, Santarem 42838); Rio Baudo, Rio Sando [5°03'N,~~

[2°26'S, 54°42'W], nearby, 1 (FMNH 76°57'W], 160 m, 2 (FMNH 90061, 90314); 21551); Rio Tapajos, Tauary [3°05'S, Riosucio, corregimiento de Cacaricas, Rio 55°06'W], 3 (MCZ 30996-30997, 31731); Perancho [7°40'N, 77°10'W], Parque Nacion-

~~Rio Tapajos, west bank, Sao Raimundo al Natural Los Katios, 1 (IND-M 3907); Un-~~

~~[3°27'S, 55°17'W], 1 (FMNH 92079); Rio guia [8°01'N, 77°07'W, Hershkovitz 1977],~~

~~Tocantins, Baiao [2°41'S, 49°41'W], 1 Golfo de Uraba, 4 (FMNH 69587-69590).~~

~~(AMNH 96255); Rio Tocantins, Cameta Cordoba: Arboletes, 1 (ICN 12978); Catival~~

[2°15'S, 49°30'W], 1 (MCZ 30998); Rio To- [8°17'N, 75°41'W], upper Rio San Jorge, 1 cantins, Ilha do Taiuna [2°15'S, 49°30'W], 14 (FMNH 68921); Rio San Jorge [9°07'N, (AMNH 96241-96251, 96256, 97315); Rio 74°44'W], 2 (AMNH 32699-32700); Upper Tocantins, Mocajuba [2°35'S, 49°30'W], 2 Rio Sinu [7°51'N, 76°17'W, Hershkovitz (AMNH 96253-96254); Santarem [2°26'S, 1977], 2 (FMNH 68919-68920). Cundina- 54°42'W], 5 (USNM 111636/49590, 111637/ marca: Salto del Tequendama [4°35'N,

~~49591, 239454-239455, skin number 74°18'W], 1 (IND-M 3906). Guaim'a: Caiio~~

~~49592); Santarem, Cuiaba, km 35, 1 (USNM Carbon, Puerto Inirida [3°52'N, 67°56'W], 1 461731); Santarem [2°26'S, 54°42'W], near, (ENfD-M 3964). La Guajira: Puerto EstreUa~~

~~2 (AMNH 40829-40830); Santarem, Mojui [12°21'N, 71°19'W], 1 (USNM 216665).~~

dos Campos (2°26'S, 54°42'W), 27 (USNM Meta: ViUavicencio [4°09'N, 73°37'W], 1 545911-545937); no specific locality, 7 (ICN 801). Naririo: Barbacoas [1°41'N, (FMNH 25315-25319, 34402; UMMZ 78°09'W], 1 (AMNH 34153); Tumaco

~~53929). Sao Paulo: Jaragua [23°27'S, [1°49'N, 78°46'W], 1 (IND-M 4112); Tu-~~

~~46°44'W] 1 (FMNH 94296). Colombia (83). maco, Inguapi del Guadual, Rio Mira, 2 (UV Amazonas: Leticia [4°09'S, 69°57'W], Rio 8131, 10920); Tumaco, 15 km E, Inguapi del~~

~~Amazonas, 1 (MLS 2213); no specific local- Guadual, 4 (UV 4657, 4658, 8132, 8133).~~

~~ity, 1 (E^-M 387). Antioquia: Dabeiba Norte de Santander: Catatumbo, Petrolea~~

~~[7°01'N, 76°16'W], Rio Sucio, 2000 ft, 2 [8°30'N, 72°35'W], 1 (MLS 576); Ciicuta (AMNH 37792-37793); MedeUin [6°15'N, [7°54'N, 72°31'W], "comprado en Cucuta,"~~

~~75°35'W], 1 (MCZ 5015); Turbo [8°06'N, 1 (MLS 578). Putumayo: Puerto Leguizamo~~

~~76°43'W], 1 (ICN 800); Zaragoza, 23 km S, [0°12'S, 74°46'W], Caiio Caucaya, Finca Ve- 22 km W, at Providencia [7°21'N, 75°03'W], lasquez, entrando por Limonconcho NW de~~

400 m, 1 (USNM 449524). BoKvar: San Juan Leguizamo, 1 (IND-M 590); Rio Mecaya Nepomuceno, 167 m [9°58'N, 75°04'W, [0°28'N, 75°20'W], 185 m, 2 (FMNH Hershkovitz 1977], 2 (FMNH 68916-68917). 70812-70813). Santander: Barrancabermeja,

~~Caqueta: no specific locahty, 1 (FMNH Peroles, Caiio Muerto [7°10'N, 73°55'W],~~

~~140254). Cauca: Isla Gorgona [2°59'N, 150-200 m, 1 (ICN 2952). Sucre: Chocho,~~

~~78°12'W], 2 (BM skin 24.12.6.17/skull Sincelejo [9°18'N, 75°24'W], 1 (IND-M 24.16.6.17 [holotype of Bradypus gorgon 4133); Coloso [9°30'N, 75°21'W], Las Cam-~~

Thomas, examined by E. KaUco and N. Sim- panas, 1 (FMNH 68918). Valle del Cauca: mons]; IND-M 2613); Rio Saija [2°52'N, Bajo Cahma [4°00'N, 76°56'W, Gonzalez-M. 28 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

& Albeiico 1993], Quebrada La Brea, parte 97876-97880); Hacienda Tepeyac [13°11'N, alta, ca. 250 m, 2 (UV 4078, 10919), Bue- 85°56'W, Genoways 1973], 3 (KU 104581; naventura, Bajo Calima [4°00'N, 76°56'W, USNM 337556-337557). Rio San Juan: Gonzalez-M. & Alberico 1993], 35 m, 3 (UV Greytown [= San Juan del Norte, 10°56'N, 4079-4081); Buenaventura, Rio Raposo 83°42'W], 2 (USNM 59010, 16352/23251); [3°41'N, 77°05'W, Alberico 1983], ca. sea La Esperanza, 2 (KU 108389; USNM level, 17 (UV 4843-4859); Calima, 300 m, 361231). Zelaya: Bonanza [13°57'N,

~~1 (MSU 2077); Rockefeller Research Station, 84°32'W], 16 (KU 99451-99458; USNM 5 mi up Rio Raposo from the Pacific, 20 mi. 338773-338780); Bonanza [13°57'N,~~

SE Buenaventura, 1 (UMMZ 115803); Sa- 84°32'W], 3.5 mi SW of, 780 ft, 1 (KU baletas, 500 m, 4 (MSU 2078-2081); Zaba- 96356); Bonanza [13°57'N, 84°32'W], 4 mi letas, 500 m, 3 (FMNH 86761-86762, NE of, 800 ft, 1 (KU 96357); El Recreo 86879). Costa Rica (8). Alajuela: JabiUo, San [12°09'N, 84°26'W], 71 km ENE Bluefields, Carlos [= Vijagual and La Vieja de San Car- 5 (KU 104368-104369; USNM 337713- los, 10°20'N, 84°30'W, see also Goodwin 337715); El Recreo [12°09'N, 84°26'W],

1946], 1 (AMNH 139833). Cartago: Angos- north side Rio Mico, 25 m, 2 (KU 115212- tura [9°53'N, 83°38'W], 1 (USNM 12871/ 115213); El Recreo [12°09'N, 84°26'W], 14104). Heredia: Rio Sarapiqui, Puerto Viejo south side Rio Mico, 25 m, 3 (KU 106317, [10°38'N, 84°01'W], 300 ft, 2 (UMMZ 111343-111344); Escondido River [12°09'N,

~~112319-112320). Limon: Pacuare [9°55'N, 83°46'W], 1 (USNM 51273). Panama (127;~~

~~83°34'W], 1 (USNM 12870/15961); Tala- specimens collected in the former "Canal manca, 2 (USNM 11381, 12103/14215). Zone" are now placed in Colon or Panama~~

Puntarenas: Pahnar [8°57'N, 83°28'W], 1 provinces, as Usted here] Bocas del Toro: Al- (AMNH 139313). Ecuador (20). El Oro: Por- mirante (9°18'N, 82°24'W), 9 (USNM tovelo [3°43'S, 79°39'W]; Cuatro Lomas, 1 315847-315852, 315855-315856, 399052); (AMNH 46552). Esmeraldas: Achote, 1 Cayo Agua (9°10'N, 82°02'W), 18 (USNM

(MSU 9339); Dogola, 1 (MSU 9338); Haci- 324249-324260, 578414*, 578415-578419); enda de Tinbre, near Quininde [= Rosa Zar- Cayo Nancy (9°19'N, 82°11'W), 7.3 km ESE ate, 0°20'N, 79°28'W], 1 (MSU 8675); Mon- Bocas del Toro (town), 6 (USNM 464853- tana de Cole, near Quininde [= Rosa Zarate, 464855, 464856*, 464857-464858); Chan-

0°20'N, 79°28'W], 1 (MSU 8676); Montaiias guinola (9°27'N, 82°31'W), 2 (USNM de Chancamita, 1 (MSU 8664). Los Rios: 315853-315854); Elena Farm, 1 (USNM Vinces [1°32'S, 79°45'W], 8 (AMNH 62877, 291179); Isla Bastimentos (9°19'N, 62879-62885); Vinces [1°32'S, 79°45'W], 82°08'W), 2 (USNM 324248, 324261); Isla Hacienda Pijigual, 2 (AMNH 62876, 62878). Bastimentos, Cedar Creek (9°19'N, Manabi: Rio de Oro [0°28'S, 79°36'W], 1 82°08'W), 6 (USNM 335267-335172); Isla (AMNH 34270). Napo: Rio Suno [0°42'S, Colon (9°24'N, 82°16'W), La Gruta, 5

~~77°08'W], below Loreto, 1 (FMNH 31119). (USNM 464849*, 464850-464852, 464859);~~

~~Pastaza: Puyo, east of (1°29'S, 77°57'W), Isla Popa, south shore (9°10'N, 82°08'W), 1~~

2000 ft, 1 (MSU 3724). Pichincha: Santo Do- km E Sumwood Channel, 7 (USNM mingo de los Colorados [0°15'S, 79°09'W], 579164*, 579165-579170); Isla San Cristo- bypass road, 1 (USNM 528706). Honduras bal, Bocatorito (9°15'N, 82°16'W), 7 (USNM (2). Gracias a Dios: Patuca River, 2 (USNM 449525-449530, 449531*); Nuri (8°56'N, 21010/36058, 21011/36059). Nicaragua (44). 81°48'W), 10 (USNM 575379-575381, Boaco: Chontales, 2 (AMNH 28477-28478). 575382*, 575383-575388); Pemnsula Val-

~~Jinotega/Nueva Segovia: Rio Coco, 1 iente, Punta Alegre (9°10'N, 81°54'W), 8 (AMNH 29441). Matagalpa: Finca Tepeyac (USNM 578412, 578420-578424, 578425*,~~

~~[13°11'N, 85°56'W, Genoways 1973], 10.5 578426); Sibube (9°36'N, 82°47'W), 1 km N, 9 km E Matagalpa, 960 m, 5 (KU (USNM 335466); Tierra Oscura (9°11'N, VOLUME 1 14, NUMBER 1 29~~

82°15'W), 3.5 km S Tiger Key, 8 (USNM Rio Samiria [4°42'S, 74°13'W], 4 (AMNH 449542*, 449543-449549). Colon: Brujo 188193-188196); Rio Ucayali, Sarayacu Point, 3 (UMMZ 64942, 64950; USNM [6°44'S, 75°06'W], 7 (AMNH 76402-76403, 256182); Frijoles [9°10'N, 79°49'W], 3 76408, 76423, 76495-76497); Yurimaguas,

~~(UMMZ 56659, 58926, 59971); Fort Davis Puerto Arturo [5°50'S, 76°03'W], 1 (FMNH [9°15'N, 79°56'W], 3 (USNM 296408- 20132). Ucayali: PucaUpa [8°23'S, 74°32'W],~~

296409, 298712); Gatun [9°15'N, 79°56'W], 200 m, 1 (AMNH 147462). Venezuela (8). 2 (AMNH 36816; USNM 170889); Gatun, Amazonas: Mount Duida, Esmeralda near, Rio Indio [9°15'N, 79°59'W], 2 (USNM [3°10'N, 65°33'W], left bank Rio Orinoco, 1 170950-170951); Lion Hill [9°13'N, (AMNH 76904); Rio Casiquiare, oriUa iz-

~~79°54'W], 1 (USNM 172729); Loma de quierda. El Merey [3°05'N, 65°55'W], 1~~

Leon [= Lion Hill, 9°13'N, 79°54'W], 1 (AMNH 78515); San Juan (5°18'N, (MCZ B8427), Monte Lirio, 1 (USNM 66°13'W), Rio Manapiare, 163 km ESE 256178); Rio Pequeni, Salamanca Hydro- Puerto Ayacucho, 155 m, 1 (USNM 406693). graphic Station [9°17'N, 79°36'W], 1 (MCZ Aragua: Rancho Grande [10°22'N, 67°41'W],

34334). Darien: Cerro Tacarcuna (8°10'N, 1 (AMNH 144824); CaraboboA^aracuy: 10 77°18'W), 4600 ft, 2 (USNM 338124- km NW Urama (10°32'N, 68°23'W), 25 m, 338125); Cituro [8°00'N, 77°36'W], 1 1 (USNM 374817). Miranda: San Andres (AMNH 38191); El Real [8°06'N, 77°45'W], (10°22'N, 66°50'W), 16 km SSE Caracas,

3 (AMNH 37619-37621); Marraganti 1 144 m, 2 (USNM 372832-372833). ZuUa: [8°08'N, 77°44'W], about 2 mi above Real El Rosario (9°09'N, 72°36'W), 42 km WNW de Santa Maria, on the Rio Tuyra, ca. sea Encontrados, 24 m, 1 (USNM 443760). level, 1 (USNM 179551 [holotype of Bra- dypus ignavus Goldman]); Mount Sapo Key to three-toed sloths (Bradypus) [7°58'N, 78°22'W], 1 (MCZ 19844); TapaUsa

[7°59'N, 77°26'W], 400 ft., 1 (AMNH 1. Distinctive black mane originating at 38102). Panama: Balboa [8°57'N, 79°35'W], nape and extending halfway down the back; pterygoids inflated; two or three 1 (USNM 296410); Barro Colorado Island [9°09'N, 79°51'W], 2 (FMNH 30738; foramina present on each side of the anterodorsal nasopharynx; dorsal specu- UMMZ 64943); Cerro Azul [= La Zumba- lum never present; no differentiation dora], (9°14'N, 79°21'W), 1 (USNM between color or length of facial and 306856); La Chorrera [8°52'N, 79°48'W], 2 dorsal pelage. Known only from the Fort (AMNH 31427; USNM 324956); Kobbe Atlantic forests of southeastern Brazil. [8°54'N, 79°36'W], 2 (USNM 296293, sub- 318366). San Bias: Mandinga (9°29'N, genus Scaeopus, Bradypus (S.) torquatus 79°05'W), 2 (USNM 305593-305594). Peru r. No black mane on nape; pterygoids not (31). Loreto: Alto Amazonas, Rio Morona, inflated; no foramina present on lateral boca Rio Amaya [4°39'S, 77°07'W], 200 m, walls of the anterodorsal nasopharynx; 1 (FMNH 88893); Boca Rio Curarray orange speculum present on the dorsum of adult males; facial hair shorter than [2°22'S, 74°05'W], 1 (AMNH 71822); Iqui-

dorsal hair and of a different color. . . . tos [3°46'S, 73°15'W], 6 (AMNH 98532- subgenus Bradypus, 2 98533, 98536, 98542, 98545-98546); Nauta, 2. Single pair of large oval foramina pre- Rio Samiria, Santa Elena [4°50'S, 74°13'W], sent on the dorsal roof of the antero- 130 m, 1 (FMNH 86896); Rio Amazonas, dorsal nasopharynx; throat with stiff Apayacu [3°19'S, 72°06'W], 1 (AMNH hairs golden-yellow to the base or with 74429); Rio Amazonas, Orosa [3°26'S, a slight smoky gray tint to the base, no 72°08'W], 2 (AMNH 73758-73759); Rio dark stripe extending laterally from eye; Amazonas, Puerto Indiana [3°28'S, face and forehead covered with golden- 73°03'W], 6 (AMNH 73571-73575, 73757); yellow hairs to the base; hair of fore-

~~/ 30 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON~~

head short. Known only from the ed access to specimens under their care: C. Guianas adjacent regions of Vene- and Norris and R. Voss (AMNH), P. Jenkins zuela and Brazil (BM), A. Cadena (ICN), H. Chirivi (IND- Bradypus (B.) tridactylus M), L. Heaney, B. Patterson, and W. Stan- T . No foramina in any part of the antero- ley (FMNH), T. Holmes and R. Timm dorsal nasopharynx; throat with soft hairs, typically brown or cream-colored, (KU), M. Rutzmozer (MCZ), Hermano R. but occasionally dark brown frosted Casallas and A. Rodriguez (MLS), L. with golden; dark stripe usually present Abraczinskas and B. Lundrigan (MSU), the

lateral to eye; face variable in color, late J. Negret (UC), P. Myers and P. Tucker tan or yellowish hair; long with either (UMMZ), and M. Alberico (UV). L. Gor- brownish hair of dorsum usually ex- don facilitated our research at the USNM. tending to forehead We thank our colleagues who provided B. variegatus-group, 3 helpful comments and discussion this 3. Size small; skull small and gracile; ex- on ternal carotid foramen usually closed or project—especially M. Carleton, A. Gard- minute; stylomastoid foramen minute; ner, D. Handley, R. Holt, E. Kalko, V. Na- external auditory meatus both absolute- ples, S. Olson, T Peterson, N. Slade, R. ly and relatively large; coronoid process Timm, D. Wilson, and an anonymous re- of mandible strikingly falcate; long hair viewer, who each read earlier drafts of this of forehead and shoulders forming an manuscript. Handley's work in Bocas del obvious hood around short facial hair. Toro was supported by grants from the Na- Known only from Isla Escudo de Ver- tional Museum of Natural History (Nelson aguas, Panama. .. Bradypus (B.) pygmaeus

y . Size variable but usually large; skull ro- Fund and Research Opportunities Fund); bust; external carotid foramen large; Soundprints of Norwalk, Connecticut; the stylomastoid foramen larger, usually National Science Foundation; and a gift visible to naked eye; external auditory from J. and P. Nelson of Hillsboro, Oregon. meatus both absolutely and relatively Transportation in Bocas del Toro was pro- smaller; posterior border of coronoid vided by the U.S. Army, the Smithsonian process of mandible straight or only moderately curved; hair of forehead not Tropical Research Institute (STRI), United especially long, never giving the ap- Brands (Changuinola Division), and the pearance of a hood. Wide-ranging in Asociacion Nacional para la Conservacion both Central and South America de la Naturaleza, Panama (ANCON). The (B.) variegatus Bradypus Instituto Nacional de Recursos Naturales Renovables of Panama (INRENARE) pro- Acknowledgments vided collecting and export permits. The staff of STRI expedited administrative and We express thanks to all who participated logistical details. Our collaboration on Bra- in the field work, especially F. Greenwell, dypus began during Anderson's tenure in J. Jacobs, E. Kalko, P. Nelson, B. Schad, the National Museum of Natural History's and M. Varn; and to the late L. Paget for Research Training Program. His subsequent her hospitality and friendship to COH in data analyses and museum work were sup- Bocas del Toro. R. Chapman and N. Slade ported by a Putnam Scholarship (Kansas provided statistical advice. M. Varn shared her research on island ages and formation State University); a Fulbright Fellowship sequence. E. Kalko and N. Simmons ex- (Colombia-USA); a Panorama Society amined types of Bradypus in the British Small Grant and the E. Raymond Hall Fund Museum. V. Naples helped us understand (University of Kansas Natural History Mu- cranial anatomy in Bradypus. The follow- seum); and a National Science Foundation ing curators and collection managers grant- Graduate Research Fellowship. — —

~~VOLUME 1 14, NUMBER 1 31~~

Literature Cited comparative biology. University of Chicago Press, Chicago, 434 pp. Cabrera, A. 1957 [1958]. Catalago de los mamiferos Aguirre-C. J., & O. Rangel-Ch. 1990. Introduccion. de America del Sur.—Revista del Museo Ar- Pp. 9-10 in J. Aguirre-C. and O. Rangel-Ch., gentino de Ciencias Naturales "Bernardino Ri- eds.. Biota y ecosistemas de Gorgona. Fondo vadavia," Ciencias Zoologicas, 4(1): 1-307. para la Proteccion del Medio Ambiente Carvalho, C. T 1960. Notes on the three-toed sloth, FEN—Colombia and Editorial Presencia, Bo- Bradypus tridactylus. —Mammalia 24:155-156. gota, 303 pp. Chiarello, A. G. 1998a. Activity budgets and ranging Aiello, A. 1985 [1986]. Sloth hair: unanswered ques- patterns of the Atlantic forest maned sloth Bra- tions. Pp. 213-218 in G. G. Montgomery, ed.. dypus torquatus (Xenarthra: Bradypodidae). The evolution and ecology of armadillos, sloths, Journal of Zoology 246:1—10. and vermilinguas. Smithsonian Institution Press,

. 1998b. Diet of the Atlantic forest maned sloth Washington, 451 pp. Bradypus torquatus (Xenarthra: Bradypodi- Alberico, M. 1983. Lista anotada de los mamiferos del dae). Journal of Zoology 246:11-19. Valle. —Cespedesia 45-46:51-71. — Coates, A. G., & J. A. Obando. 1996. The geologic . 1986. Biogeografia terrestre. Pp. 225-244 in evolution of the Central American isthmus. Pp. H. von Prahl and M. Alberico, eds., Isla de Gor- 21-56 J. gona. Talleres Graficos Banco Popular and in B. C. Jackson, A. E Budd, and A. G. Coates, eds.. Evolution and environment in Universidad del Valle, Bogota, 252 pp. tropical America. University of Chicago Press, . 1990. A new species of pocket gopher (Ro- Chicago, 425 dentia: Geomyidae) from South America and its pp. P. 1997. history of forests on the isth- biogeographic significance. Pp. 103-111 in G. Colinvaux, The the Ice to the present. Pp. 123- Peters and R. Hutterer, eds.. Vertebrates in the mus from Age 136 in A. G. Coates ed.. Central America: a tropics. Museum Alexander Koenig, Bonn, 424 natural and cultural history. Yale University pp. Connecticut, Alston, E. R. 1879-1882. Biologia Centrali-America- Press, New Haven, 277 pp. Eisenberg, J. F. 1989. of the Neotropics. Vol. na. Mammalia. Taylor and Francis, London, 220 Mammals 1, the northern Neotropics: Panama, Colombia, pp. Guyana, Suriname, French Guiana. AMS. 1944. British Guyana, 1:125,000, Sheet 22, Bar- Venezuela, University of Chicago Press, Chicago, 449 tica. Army Map Service, U.S. Army, Washing- pp.

, K. H. Redford. 1999. of the Neo- ton. & Mammals central Neotropics: Ecuador, Anderson, R. R 1999 [2000]. Preliminary review of tropics, Vol. 3, the University of the systematics and biogeography of the spiny Peru, Bolivia, Brazil. Chicago pocket mice {Heteromys) of Colombia.—Revis- Press, Chicago, 610 pp. L. H., Feer. 1997. Neotropical rainforest ta de la Academia Colombiana de Ciencias Ex- Emmons, & F mammals: a field guide, 2nd edition. University actas, Fisicas y Naturales 23(Suplemento espe- cial):613-630. of Chicago Press, Chicago, 307 pp. Fairbanks, R. G. 1989. 17,000-year glacio-eustatic Avise, J. C. 1994. Molecular markers, natural history A and evolution. Chapman and Hall, New York, sea level record: influence of glacial melting rates on the Younger Dryos event and deep- 511 pp. circulation. Nature 342:637—642. Bartlett, A. S., & E. S. Barghoorn. 1973. Phytogeo- ocean — graphic history of the Isthmus of Panama during Fairchild, G. B., & C. O. Handley, Jr. 1966. Gazetteer the past 12,000 years (a history of vegetation, of collecting localities in Panama. Pp. 9-22 in climate, and sea-level change). Pp. 203-299 in R. L. Wenzel and V. J. Tipton, eds., Ectopara- A. Graham, ed.. Vegetation and vegetational sites of Panama. Field Museum of Natural His- history of northern Latin America. Elsevier Sci- tory, Chicago, 861 pp. entific Publishing Co., New York, 393 pp. Gardner, A. L. 1993. Order Xenarthra. Pp. 63-68 in Beebe, W. 1926. The three-toed sloth Bradypus cucul- D. E. Wilson and D. M. Reeder, eds.. Mammal liger cuculliger Wagler.—Zoologica l:\-61. species of the world: a taxonomic and geo- Brattstrom, B. H. 1966. Sloth behavior.—Journal of graphic reference, 2nd edition. Smithsonian In- Mammalogy 47:348. stitution Press, Washington, 1207 pp. Britton, S. W, & W. E. Atkinson. 1938. Poikilotherm- Genoways, H. H. 1973. Systematics and evolutionary ism in the sloth.—Journal of Mammalogy 19: relationships of spiny pocket mice, genus Liom- 94-99. ys. —Special Publications of the Museum, Texas

, & R. E Kline. 1939. On deslothing the Tech University 5:1-368. sloth.—Science 90(2323): 16-17. Goldman, E. A. 1920. Mammals of Panama.—Smith- Brooks, D. R., & D. A. McLennan. 1991. Phylogeny, sonian Miscellaneous Collections 69(5): 1-309. ecology, and behavior: a research program in Gonzalez-M., A., & M. Alberico. 1993. Seleccion de —

~~32 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON~~

habitat en una comunidad de mamiferos en la Nacional de Ambiente y Desarrollo, Tegucigal- costa pacifica de Colombia.—Caldasia 17:313- pa, 374 pp. 324. McCarthy, T. J., D. L. Anderson, & G. A. Cruz-D. Goodwin, G. G. 1946. Mammals of Costa Rica.—Bul- 1999. Tree sloths (Mammalia: Xenarthra) in letin of the American Museum of Natural His- Nicaragua and Honduras, Central America. tory 87:275-473. The Southwestern Naturalist 44:410-414.

Haffer, J. 1970. Geologic-climatic history and zoogeo- McKenna, M. C, & S. K. Bell. 1997. Classification of graphic significance of the Uraba region in north- mammals above the species level. Columbia western Colombia.—Caldasia 10:603-636. University Press, New York, 631 pp. Hall, E. R. 1962. Collecting and preparing study spec- McPherson, A. B. 1985. A biogeographical analysis of imens of vertebrates.—Miscellaneous Publica- factors influencing the distribution of Costa Ri- tions, University of Kansas Museum of Natural can rodents.—Brenesia 23:97-273. History 30:1-46. MINITAB. 1996. MINITAB reference manual. Re-

~~. 1981. Mammals of North America, 2nd edi- lease 11 for Windows®. Minitab, Inc., State tion. John Wiley and Sons, New York, 1181 pp. College, Pennsylvania.~~

~~Handley, C. O., Jr. 1959. A review of the genus Ho- Montgomery, G. G., & M. E. Sunquist. 1975. Impact plomys (thick-spined rats), with description of a of sloths on Neotropical forest energy flow and~~

new form from Isla Escudo de Veraguas, Pan- nutrient cycling. Pp. 69-98 in F. B. Golley and ama.—Smithsonian Miscellaneous Collections E. Medina, eds.. Tropical ecological systems: 139(4): 1-10. trends in terrestrial and aquatic research.

~~Hershkovitz, P. 1947. Mammals of northern Colombia, Springer-Verlag, New York, 398 pp.~~

Preliminary report No. 1: squirrels (Sciuri- Morrison, P. R. 1945. Acquired homiothermism in the dae).—Proceedings of the United States Nation- pregnant sloth.—Journal of Mammalogy 26: al Museum 97(3208): 1-46. 272-275.

. 1969. The evolution of mammals on southern Naples, V. L. 1982. Cranial osteology and function in continents VI. The Recent mammals of the the tree sloths, Bradypus and Choloepus. — Neotropical region: a zoogeographic and eco- American Museum Novitates 2739:1-41. logical review.—The Quarterly Review of Bi- OEA. 1959. Mapa ecologico de Panama, escala 1: ology 44:1-70. 1,000,000. Instituto Interamericano de Ciencias

~~. 1977. Living New World monkeys (Platyrrhi- Agricolas de la Organizacion de Estados Amer-~~

ni), with an introduction to Primates, Vol. I. icanos (OEA), San Jose, Costa Rica. University of Chicago Press, Chicago, 1117 pp. Olson, S. L. 1993. Contributions to avian biogeogra- Holm, S. 1979. A simple sequentially rejective multi- phy from the archipelago and lowlands of Bo- ple test procedure.—Scandinavian Journal of cas del Toro, Panama.—The Auk 110: 100— Statistics, Theory and Applications 6:65-70. 108.

Illiger, K. 1811. Prodromus Systematis Mammalium et Paynter, R. A., Jr. 1982. Ornithological gazetteer of Avium. Salfeld, Berlin, 301 pp. Venezuela. Harvard University, Cambridge, Jackson, J. B. C, A. E Budd, & A. G. Coates, eds. Massachusetts, 245 pp.

1996. Evolution and environment in tropical . 1993. Ornithological gazetteer of Ecuador, America. University of Chicago Press, Chicago, 2nd edition. Harvard University, Cambridge, 425 pp. Massachusetts, 247 pp.

Kalko, E. K. V, & C. O. Handley, Jr. 1994. Evolution, . 1997. Ornithological gazetteer of Colombia, biogeography, and description of a new species 2nd edition. Harvard University, Cambridge, of fruit-eating bat, genus Artiheus Leach (1821) Massachusetts, 537 pp.

~~from Panama.—Zeitschrift fiir Saugetierkunde , & M. A. Traylor, Jr. 1991. Ornithological gaz- 59:257-273. etteer of Brazil. Harvard University, Cam-~~

~~Eighty, R. G., I. G. Maclntyre, & R. Stuckenrath. bridge, Massachusetts, 788 pp.~~

1982. Acropora palmata reef framework: a re- , , & B. Winter. 1975. Ornithological liable indicator of sea level in the western At- gazetteer of Bolivia. Harvard University. Cam- lantic for the past 10,000 years. —Coral Reefs bridge, Massachusetts, 80 pp. 1:125-130. Rangel-Ch.. O. 1990a. Tipos de vegetacion. Pp. 109-

Linnaeus, C. 1758. Systema naturae per regna tria na- 126 in J. AguiiTe-C. and O. Rangel-Ch., eds.. turae, secundum classis, ordines, genera, spe- Biota y ecosistemas de Gorgona. Fondo para la cies, cum characteribus, differentiis, synonymis, Proteccion del Medio Ambiente—FEN—Co-

~~locis, 10th edition. Vol. I. Laurentii Salvii, lombia and Editorial Presencia, Bogota, 303 pp.~~

~~Stockholm, 824 pp. . 1990b. Caracterizacion ecologica, sintesis fi-~~

~~Marineros, L., & F. Martinez-Gallegos. 1998. Gui'a de nal. Pp. 277-293 in J. Aguirre-C. and O. Ran- campo de los mamiferos de Honduras. Instituto gel-Ch., eds.. Biota y ecosistemas de Gorgona. —~~

~~VOLUME 1 14, NUMBER 1 33~~

Fondo para la Proteccion del Medio Ambien- lus) along the Sinnamary River, French Gui- te—FEN—Colombia and Editorial Presencia, ana.—Biotropica 31:686-691. Bogota, 303 pp. Thomas, O. 1917. Some notes on three-toed sloths. Roth, V. L. 1992. Inferences from allometry and fos- Annals and Magazine of Natural History, Series 8, 19:352-357. sils: dwarfing of elephants on islands. —Oxford . 1926. On mammals from Gorgona Island, Surveys in Evolutionary Biology 8:259-288. with the description of a new sloth.—Annals SAS. 1990. SAS/STAT user's guide. Version 6, 4th and Magazine of Natural History, Series 9, 17: edition. SAS Institute, Inc., Cary, North Caro- 309-311. lina. USBGN. 1956. Gazetteer No. 25, Nicaragua. United Schinz, H. R. 1825. Das Thierreich eingetheilt nach States Board on Geographic Names (USBGN) dem Bau der Thiere als Grundlage ihrer Natur- and Office of Geography, Department of the In- geschichte, Vol. IV. Stuttgart and Tubingen. terior, Washington, 49 pp. Stephens, L., & M. A. Traylor, Jr. 1983. Ornithological von Prahl, H. 1986. Notas sobre la geologfa, climato- gazetteer of Peru. Harvard University, Cam- logi'a, regimen de mareas y oceanograffa. Pp. bridge, Massachusetts, 271 pp. 21-27 in H. von Prahl and M. Alberico, eds., Isla de Gorgona. Talleres Graficos Banco Pop- , & . 1985. Ornithological gazetteer of the Guianas. Harvard University, Cambridge, ular and Universidad del Valle, Bogota, 252 pp. Wetzel, R. M. 1985 [1986]. The identification and dis- Massachusetts, 121 pp. tribution of Recent Xenarthra (= Edentata). Pp. Sunquist, M. E., & G. G. Montgomery. 1973. Activity 5—21 in G. G. Montgomery, ed.. The evolution patterns and rates of movement of two-toed and and ecology of armadillos, sloths, and vermilin- three-toed sloths (Choloepus hojfmanni and guas. Smithsonian Institution Press, Washing- Bradypus infuscatus). —Journal of Mammalogy ton, 451 pp. 54:946-954. , & F. D. Avila-Pires. 1980. Identification and Taube, E., J.-C. Vie, P Fournier, & C. Genty. 1999. distribution of the Recent sloths of Brazil Distribution of two sympatric species of sloths (Edentata). —Revista Brasileira de Biologia 40: (Choloepus didactylus and Bradypus tridacty- 831-836.