A New Species Of Mabuya (Sauria : Scincidae) From The British Virgin Islands

16 Pages·2000·5.2 MB·English

by  G C Mayer

PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON
113(4):871-886. 2000.
A new species of Mabuya (Sauria: Scincidae) from the
British Virgin Islands
Gregory C. Mayer and James Lazell
(GCM) Department of Biological Sciences, University ofWisconsin-Parkside, Kenosha,
WI 53141, U.S.A.;
(XL) The Conservation Agency, 6 Swinburne Street, Jamestown, RI 02835, U.S.A.
—
Abstract. Mabuya macleani, new species, is a pallid, drab, almost pattern-
less skink which is abundant on Carrot Rock, British Virgin Islands. Mabuya
sloanii shows character divergence in pattern from the new species concordant
with geographic approach to within 400 m. On Puerto Rico, M. sloanii has a
broad middorsal bronzy area on the anterior dorsum. On most of the smaller
islands of the Puerto Rico Bank, as well as the Mona and Desecheo Banks to
the west, this middorsal area is much narrowed by the presence of well-de-
veloped dark dorsolateral stripes stretching from the head to behind the fore-
limbs. These two pattern types, which apparently intergrade in the vicinity of
northeastemmost Puerto Rico, are recognized as M. s. nitida Garman and M.
s. sloanii Daudin, respectively. The specific name "mabouya" Lacepede, for-
merly applied to Antillean skinks, is shown to be unavailable. The presence of
M. macleani and another endemic lizard (Anolis emestwilliamsi) on such a
small (1.3 ha), poorly isolated, and young (<3000 y) island as Carrot Rock
may be a striking case of rapid divergence of insular populations.
"The coloration is highly interesting . . . several imen was distinctive, we at first referred to
insularformsmaybedistinguishablewhenadequate it as Mabuya mabouya sloanii (Mayer &
series become available." Karl P. Schmidt (1928) Lazell 1988), the common skink ofthe Vir-
The scincid lizards ofthe genus Mabuya gin Islands (MacLean 1982, Lazell 1983),
are nearly tropicopolitan in distribution. We pending collection of further specimens.
have collected or examined them in num- Over the next several years JL returned to
bers from the Antilles, South America, Carrot Rock occasionally, and found an
tropical Asia, and Africa. Throughout this area where these peculiar pallid lizards
vast range, most species are brown with were abundant, and succeeded in capturing
near-black stripes extending the length of five more individuals (of dozens seen).
the body. A striking exception is on Carrot These specimens, as well as comparison
Rock, a very small (1.3 ha), steep-sided is- with Mabuya from throughout the islands
land offthe southeast end ofPeter Island in of the Puerto Rican Bank, have abundantly
the British Virgin Islands (Fig. 1). confirmed the distinctiveness of this popu-
On 13 July 1985, while one ofus (GCM) lation, and also brought into sharp focus a
climbed the biggest tree on the island in most intriguing evolutionary phenomenon.
search ofanoles, the other (JL) toiled in the We here describe this distinctive Mabuya
dust in the little gully on the windward side as:
of Carrot Rock in which the tree grew,
looking for Sphaerodactylus geckos. A far Mabuya macleani, ne—w species
larger, drab, pale lizard was turned out, and Mabuya mabouya sloanii. Mayer & La-
ran up his sleeve. Although this initial spec- zell, 1988:23 (in part).
872 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON
'j'^rii&^i
Fig. 1. Carrot Rock, British Virgin Islands, viewed from the north, 17 July 1988. Windward is to the left,
leeward to the right. The highest elevation is ca. 25 m. From a Kodachrome transparency by GCM.
—
Holotype. Museum ofComparative Zo- is distinguished from the geographically
ology (MCZ) 170884, collected on Carrot nearest populations of M. sloanii by the
Rock, south of Peter Island, British Virgin much reduced dorsolateral dark stripe, the
Islands, 18°19'45"N, 64°34'18'W, by J. La- continuous stripe beginning behind the
zell, 13 Jul 1—985 (Fig. 2D). head in M. macleani (on the head in M.
Paratypes. All from the type locality: sloanii), and extending only to the level of
MCZ 182270-72, 17 Jul 1988; MCZ the forelimbs (behind the forelimbs in M.
176728 and University of Michigan Muse- sloanii). —
um of Zoology (UMMZ) 197261, both 26 Description of the type. Rostral wider
Oct 1991. than high, bordered dorsoposteriorly by the
—
Diagnosis. A pallid tan to brownish- nasals and paired supranasals, which are in
gray New World Mabuya (Dunn 1936, narrow contact. The frontonasal is broader
Greer 1970) with one or two pairs of en- than long and in contact with the frontal.
larged nuchals (theircombined widths more The paired prefrontals are separated medi-
than 75 percent of the width of the parie- ally by the contact of the frontonasal with
tals); two frontoparietals; parietal overlap- the frontal. The frontal is about three quar-
&
ping upper anterior temporal (Greer ters as long as its distance from the poste-
Nussbaum 2000); midbody scales in 32-34 rior parietal edge. There are four supraocu-
rows; 16-18 subdigital lamellae under lars, the second the largest. There are three
fourth toe of pes; limbs moderately long; supraciliaries, the first by far the longest.
dark dorsal markings fragmented or absent The two frontoparietals are in contact with
on head and separated from dark dorsolat- the second, third, and fourth supraoculars,
eral stripes on nape; nape stripes reduced, bordered posterolaterally by the parietals,
separated by all or most oftwo dorsal scales and posteriorly by the interparietal, in
and not extending more than 21 dorsal which the parietal foramen is posteriorly
scales posterior to parietals; lateral dark situated. The large, paired parietals are in
stripes poorly developed. Mabuya macleani contact posterior to the interparietal, that on
VOLUME 113, NUMBER4 873
T T
~c~ "5"
Fig. 2. Dorsal patterns of skinks from the Puerto Rican Bank. A-C, geographically approaching Carrot
Rock: A) MCZ 6050, lectotype ofMabuya sloanii nitida, San Juan, Puerto Rico, ca 164 km WNW of Carrot
Rock. B) MCZ 36624, intergrade, M. s. nitida X sloanii, Cayo Icacos, ca. 102 km WNW ofCarrot Rock. C)
MCZ 182273, M. s. sloanii, Stoney Bay, PeterIsland, British Virgin Islands, 400 m NofCarrotRock. D) MCZ
170884, type of Mabuya macleani. Carrot Rock, British Virgin Islands. Bars below lizards indicate 1 cm in
each case.
the right extending further posterioriy. The eyelid about as wide as the ear opening.
parietals overiap the upper anterior tempo- The temporals are larger than the trunk
ral. There is one pair of transversely en- scales. There are no auricular denticles.
larged nuchals, but the second and third nu- There are two pairs of chin shields in con-
chals are enlarged on the left side. tact posterior to the mental.
The nasal is subrectangular in side view, Scales of body and limbs imbricating,
with the large nostril posteriorly located, subcycloid, regularly arranged in rows.
followed by the postnasal and two loreals; Thirty-four longitudinal rows at midbody,
the anterior one on the left is much larger 57 transverse rows dorsally from parietals
than the posterior, but the two are subequal to anterior edge of hind limb, 61 ventrally
on the right. The anterior loreal is in contact from mental to vent. The vent is bordered
with the prefrontal, but the posterior loreal anteriorly by eight subequal scales. Scales
is separated from the latter scale by a pre- of soles and palms tubercular, transition
supraciliary. The sixth supralabial on the from imbricate scales of limb to tubercular
left, and the fifth on the right, are about scales abrupt. Thirteen lamellae under
twice as long as the others, the enlarged su- fourth toe of manus, 17 under fourth toe of
pralabial on each side forming a long su- pes. Adpressed limbs do not meet.
bocular. There is a clear disk in the lower In life, the type was pallid beige-gray
874 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON
—
Table 1. Some characteristics ofMabuya macleani. "Scales" are numberofrows aroundtrunkatmidbody.
"Stripe" is the length ofthe dorsolateral dark stripe in dorsal scales posterior to the parietals (left/right).
Nuchal
SVL Scales Supra-nasals Stripe Pairs Adpressedlimbs
MCZ 170884 69.5 34 Contact 15/16 1 Fail to meet
MCZ 176728 71.0 32 Contact 18/18 1 Toes barely touch
MCZ 182270 80.5 32 Contact 17/18 2 Fail to meet
MCZ 182271 76.0 32 Contact 18/15 1 Fail to meet
MCZ 182272 63.0 32 Separated 17/18 2 Meet
UMMZ
197261 44.5 32 Fused 21/21 1 Toes overlap
with a faint trace of a lateral stripe extend- supraocular with the parietal, also on the
ing to just above the axilla. There are two left side. Contact of the supranasals is var-
darkdots on the frontonasal and alittle dark iable, being separated in one specimen,
gray flecking on the supraoculars. Theplain touching in four, and even fused medially
lead-gray dark dorsolateral stripes end 15 into a single scale in another. The prefron-
(left) or 16 (right) dorsal scales posterior to tals are never in contact medially. In lateral
the parietals.— view, the head squamation of M. macleani
Variation. Some characteristics ofMa- is essentially similar to that of M. sloanii
buya macleani are given in Table 1, and shown by Schmidt (1928:122).
measurements in Table 2. The holotype is The following summary of meristic var-
fairly typical of the type series in its squa- iation gives for each character the range,
mation, and the paratypes do not present a followed by the mean and sample size in
great deal of variability. Dorsal and ventral parentheses. For some bilateral characters
head scales of MCZ 182270 are shown in the sample size has been reckoned by the
Fig. 3. This specimen shows two unusual number of sides rather than specimens, and
conditions: the presence of an intercalary this is noted after the sample size. Supra-
scale separating the first supraocular from oculars: 4 (4.0, 12 sides); supraciliaries: 2-
contact with the frontal on the left side of 4 (3.0, 6 sides); supralabial subtending the
the head, and thepartial fusion ofthe fourth eye (subocular): 5-6 (5.3, 11 sides); mid-
—
Table 2. Measurements (mm) ofholotype and three paratypes ofMabuya macleani.
MCZ 170884 MCZ 182270 MCZ 182271 MCZ 182272
Holotype Paratype Paratype Paratype
Snout-vent length 69.5 80.5 76.0 63.0
Tail length 52+ 67+ 56+ 75.5
Axilla-groin length 38.0 45.0 40.5 34.5
Snout length 5.4 5.7 5.7 5.0
Snout width 5.5 5.5 5.4 4.9
Head length 11.9 13.0 12.8 11.4
Head width 8.2 10.2 10.2 8.0
Upper arm length 5.5 6.5 6.0 5.0
Lower arm length 4.9 5.5 6.5 4.7
Palm length 1.9 2.8 2.6 2.9
Fourth finger length 4.5 4.5 4.5 4.0
Upper leg length 8.0 8.0 8.0 7.0
Lower leg length 6.6 7.5 7.3 6.6
Sole length 3.4 3.8 4.3 3.1
Fourth toe length 6.5 7.0 7.0 7.5
+ Tail broken or regenerated.
VOLUME 113, NUMBER 4 875
have paired blotches on the frontonasal, but
they are scarcely evident on the largest
specimen (MCZ 182270), in which, were it
not for their presence in the other speci-
mens, they might be taken to be merely two
of several scattered dark mottlings on the
head, rather than reduced pattern elements.
In all specimens the dorsolateral stripe is
short and starts on the neck, beginning 3-5
(4.6, 12 sides) scales behind the parietals,
and extending to the 15th to 21st (17.7, 12
sides) transverse dorsal scale row behind
the parietals; the posterior end is thus at
about the level of the forelimbs. There is
also size related variation in the intensity of
the striping, the dorsolateral dark and light
stripes being relatively darker and lighter,
respectively, in the second smallest speci-
men (MCZ 182272) than in the larger ones.
The most distinctive pattern variation is in
(UMMZ
the smallest specimen 197261),
which, in addition to the frontonasal blotch-
es and dorsolateral dark stripe (which is
longest in this specimen, extending from
the 5th to 21st scale row), also has two
short stripes on the head, extending from
the second supraoculars to the parietals.
There appears to be allometric growth of
UMMZ
the limbs. The smallest individual,
mm
197261, 44.5 SVL, has proportionately
MCZ
by far the longest limbs, and 187220,
mm
80.5 SVL, the largest, has the shortest.
The other four are intermediate. The type,
Fig. 3. Aspects of head squamation in MCZ at 69.5 nmi SVL, has slightly shorter limbs
182770, Mabuya macleani paratype. The first pair of MCZ mm
proportionately than 176728, 71
enlarged nuchals are marked "n." Bar is 1 mm. SVL, but the overall impression is that limb
length fails to keep pace with body growth.
body scale rows: 32-34 (32.3, 6); trans- In studies of other New World Mabuya,
verse dorsal rows: 54-58 (56.0, 4); trans- based on larger series than available of M.
verse ventral rows: 58-66 (61.2, 6); fourth macleani, a similar decline in relative limb
toe of manus lamellae: 13 (13.0, 4); fourth length has been demonstrated: Rebou9as-
toe of pes lamellae: 16-18 (17.2, 5). Spieker (1974) found this to be the case for
All Mabuya macleani have a much re- most samples in her study of M. agilis, M.
duced pattern of striping compared to the caissara, and M. macrorhyncha; Rebougas-
near-black striping typical ofAntillean Ma- Spieker & Vanzolini (1990) found it in M.
buya (Fig. 2), but the extent of reduction carvalhoi; and Avila-Pires (1995) found it
varies, and seems to be size related, with in M. bistriata (=ficta sensu Rebou^as-
smaller specimens having less reduced Spieker) and M. nigropunctata {=bistriata
markings. All specimens of M. macleani sensu Rebou9as-Spieker).
876 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OFWASfflNGTON
The sexes are similar. Based on broad (MCZ 6050, MCZ 6052 [2 specimens]),
head, thick tail base, and enlargementofthe Ensenada (American Museum of Natural
medial pair of scales bordering the vent, History [AMNH] 6462), and Bayamon
MCZ 182271 wasjudged to be a male, and (AMNH 14007, UMMZ 73828) agree with
dissection confirmed that. Based on narrow Schmidt and Stejneger's descriptions, as did
head, abruptly tapering tail, and subequal the Puerto Rican specimens available to
MCZ
scales bordering the vent, 187220, the Grant (1931).
largest available specimen, was judged to A seventh specimen, from easternmost
be a female, and—dissection confirmed that. Puerto Rico (UMMZ 73829, Cape San
Comparisons. Mabuya macleani is Juan), shows some approach toward the
morphologically and geographically closest sloanii pattern. A specimen from Icacos
to M. sloanii. The latter species is wide- (MCZ 36624), an islandjust east of Puerto
spread, but nowhere abundant, in the Pas- Rico, is intermediate between nitida and
sage and Virgin Islands, and is also found sloanii of the smaller islands to the east
on Mona and Desecheo (Heatwole et al. (Fig. 2B). Based on these specimens and
1981, Lazell 1983, 1991, 1995; Mayer the literature, we recognize M. s. nitida as
1989). This skink is deep copper to choc- a valid taxon, and confirm Schmidt's re-
olate brown with prominent lateral stripes striction of type locality by selecting MCZ
of near-black and very bold jet black dor- 6050 as the lectotype. A similar pattern of
solateral stripes, beginning on the head, and geographic variation, with a Puerto Rican
extending continuously down the nape onto mainlandform and aPassageA'^irginIslands
the anteriortrunk, well past 20 dorsal scales form showing signs of intergradation in
posterior to the parietals (Figs. 2C, 4). The easternmost Puerto Rico and the adjoining
dark dorsolateral stripes are black and sep- cays is found in Anolis cristatellus (Hea-
arated by silvery-white on the median por- twole 1976). It is interesting that s. sloanii
tions of the two middorsal scale rows. On occupies the islands to the east and west of
Puerto Rico, skinks also have near-black the Puerto Rican main, with s. nitida oc-
lateral stripes extending onto the trunk, but cupying the main inbetween; asimilarperi-
the dark dorsolateral stripes are reduced to Puerto Rican distribution is shownbyHem-
heavy blackish blotching on the head, frag- idactylus mabouia (Kluge 1969) and Epi-
&
menting and dwindling to speckles on the crates monensis (Schwartz Henderson
nape (Fig. 2A). This pattern form was 1991), with related species on the Puerto
named M. nitida by Garman (1887) on the Rican main {H. brookii and E. inornatus).
basis of one specimen from Hispaniola and The overall picture is of Puerto Rico
three from Puerto Rico. Stejneger (1904) Bank Mabuya sloanii showing greaterchar-
described this form accurately, mistakenly acter divergence from M. macleani as the
under the name M. sloanii, based on one of latter's geographic distribution is ap-
two specimens numbered MCZ 6052 from proached (Figs. 2, 5). Mabuya s. nitida has
San Juan, Puerto Rico. He also noted the less dark anterodorsal pigment than doesM.
existence of the other pattern type (the ac- s. sloanii; however, M. s. sloanii occurs to
tual sloanii) under the name semitaeniata within 400 m ofM. macleani, but the latter
Wiegmann (1837). Schmidt (1928) also has the least dark pigment of all.
noted the two patterns, and restricted Gar- The Caicos Islands, although geological-
man's nitida to Puerto Rico. Grant (1931) ly part of the Bahamas, share a number of
recognized the distinction in pattern be- herpetofaunal elements with the Puerto
tween the Puerto Rican and small island Rico Bank {Anolis, Typhlops-Thomas
forms as well, using the same erroneous 1999). The skinks of these islands appear
names as did Stejneger Six Puerto Rican distinct from Puerto Rico Bank ones based
specimens available to us, from San Juan on the specimens we have seen. We have
VOLUME 113, NUMBER 4 877
Fig. 4. Daudin's (1802)illustrationofhisScincussloanii,thewidespreadformoftheVirginIslands,showing
pattern in side view.
.
MCZ
color notes in life for 182881 from tend dorsolaterally onto the costal region.
Long Cay, Caicos Bank: dark bronze- about one-third of the distance from axilla
brown dorsally, with two very bold cream- to groin; these light stripes are bordered
white stripes that begin at the snout and ex- ventrally by a near-black field, six scale
Fig. 5. The eastern portion of the Puerto Rico Bank showing localities from which specimens have been
examined. 1, San Juan, Puerto Rico, type locality forMabuya sloanii nitida. 2, Cayo Icacos, represented by an
apparent M. s. sloanii X nitida intergrade. 3, Isla Culebra, 4, Tortola, 5, Anegada, 6, Peter Island, all nominate
M. s. sloanii. 7, Carrot Rock, type locality for Mabuya macleani. Bar indicates 10 km. The dotted line is the
edge ofthe Puerto Rican Bank, ca. 100 m below sea level.
878 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON
rows wide at the forelimb insertion, and buya macleani differs from M. sloanii
flecked with white speckles. The hght dor- somewhat in having smaller scales, reflect-
solateral stripes are bordered dorsally by ed in higher midbody row counts: 32-34,
near-black stripes one scale wide at the as opposed to 30-32 in other Puerto Rico
forelimb insertion and separated by four BankMabuya (n = 20). Itwouldtakelarger
scales across the dorsum at that level. The sample sizes to even demonstrate statistical
posterior body is heavily speckled with significance. There may be selection pres-
black dots aligned in longitudinal rows. A sure for smaller scales in M. macleani. Car-
preserved series from the same island rot Rock's other endemic lizard, Anolis er-
(AMNH 80125-30) agrees well in pattern nestwilliamsi, has very small scales and is
with the live specimen described above. absolutely distinct from its closest relative
The striping pattern most resembles that of in this character (Lazell 1983). Interpreting
the Icacos intergrade (Fig. 2B), except that the adaptive significance ofscale size inUz-
the middorsal brown area is about twice as ards is, however, fraught with difficulties
wide, thus being more like nitida, but dif- and apparent contradictions (Lazell 1994,
fering from the latter in having a broader Dmi'el et al. —1997).
dorsolateral dark stripe, and in that the Ica- Etymology. The species is named in
cos specimen lacks the heavy, alignedbody honor of our late friend and colleague Dr.
spotting of the Caicos specimens. Such William P. MacLean, III, of the University
spotting is variable in M. s. sloanii, and of the Virgin Islands, who contributed so
UMMZ
may be fairly heavy (Grant 1931; much to knowledge of the Virgin Island
80585 [Buck Island off St. Thomas]). herpetofauna (MacLean 1982), and who
White speckling in the lateral dark stripe, aided and assisted our work, and that of
which is also seen in specimens from West many others, on numerous occasions (La-
Caicos (UMMZ 117392-4) and Six Hill zell & Mayer 1992). He was one ofthe first,
(UMMZ
Cays 117394-6), is not found in and still few, professional biologists ever to
Puerto Rican region specimens. Further have set foot upon Carrot Rock, and rec-
study, beyond the scope of the present ognize its biotic uniqueness.
work, may reveal that the Caicos Mabuya
should be recognized as a valid taxon. Discussion
Dunn's (1936) description of Mabuya
—
pergravis of the western Caribbean, "strip- Ecology. Carrot Rock has undergone
ing very indistinct; pale with dark dots major ecological changes since it was first
above," may sound superficially similar to visited by JL in 1980 (Lazell 1983). Then,
M. macleani, but the two forms are amply most of the windward (eastern) and north-
distinct. Mabuya pergravis is much more em portion ofthe top ofthe island was cov-
slender, and has fewer midbody scale rows ered with a sprawling growth of sea grape,
(28-30-Dunn 1936, Dunn & Saxe 1950). Coccoloba uvifera (Polygonaceae), which
Striping is not indistinct in a single speci- had to be either climbed over or crawled
men from San Andres, and in those from under. There were three thickets of sea
Providencia the dark dots are numerous grape on the edges of the scarp tall enough
(unlike macleani, in which there are few or to stand in the shade of: one on the leeward
no dark dots dorsally). coast, one on the northern windward coast,
In meristic and measurable characters, all and the biggest in the gully-locally called
New WorldMabuya, and most from the rest "ghut"-where the first Mabuya macleani
of the world, are slightly, and often only was collected.
&
modally, differentiated; as Greer Nuss- Severe drought characterized the climate
baum (2000) noted, "Few unequivocal of the Virgin Islands during the eighties. It
characters of scalation are available." Ma- seemed that more precipitation fell in the
VOLUME 113, NUMBER 4 879
—
form of dust said to—have blown all the the seagrapes in the ghutthicket were dead.
way from the Sahara than as water. JL's The large candelabra cactus, Pilosocereus
field notes of 13 July 1985 record: royenii, thathad crownedthe top oftheislet
(and housed the largest, uncatchableAnolis)
"The Rock is in terrible shape! Theseagrapelooks was "rotting pulp and stark skeleton."
90% dead; places I could crawl under before are Rainfall in the region began to increase
nowjust scattered dry sticks. Trying to dig out leaf in 1996. On 24 October 1996 a group ofus
litter was a nightmare ofdust." checked the Rock briefly. We did not at-
Despite the drought, Anolis remained tempt to collect specimens, but we quickly
conamon, the firstMabuya was secured, and located three Anolis emestwilliamsi and
Sphaerodactylus macrolepis (MCZ two Mabuya macleani, one of the latter
170890) was also collected-all in the one perched on a vine ca. 3 cm above the
&
remaining sea grape thicket in the ghut-in ground (Schwartz Henderson [1991] note
1985 (Mayer & Lazell 1988). This brought climbing in Mabuya mabouya). Again,
Carrot Rock into compliance with the "rule from JL's field notes: "Seagrapes are regen-
of three" for Caribbean islands, as predict- erating well; the place generally looks
ed (Lazell 1983). On 17 July 1988 condi- much better than last year."
tions were no better, but more M. macleani A brief vegetation survey of Carrot
were seen and collected than ever before. Rock, by Dr. Fred Kraus on 26 October
Over most of the boulder-jumbled surface 1991, included, in addition to seagrape and
of Carrot Rock, skinks have the advantage candelabra cactus, Mammillaria nivosa,
over would-be captors. In one small area Melocactus intortus, and Opuntia dillenii
near the top, however, there are few rocks, (all Cactaceae), andthe vines Capparisflex-
little vegetation, and a soil substrate. Here uosa (Capparidaceae), Stigmophyllon peri-
a group of us simultaneously sighted eight plocifolium (Malpighiaceae), and Cavana-
skinks in a 10 x 20 m (200 m^) plot (and lia maritima (Leguminosae). There are
caught three of the eight). Excluding the "various graminoids including the rare
bare rock faces and wave-washed talus of silky foxtail grass Pappophorum pappifer-
the edges, we estimate the top ofthe island um^ —
habitable for skinks at ca. 1.3 ha. Using our Differentiation on small islands. The
crude estimate, the total population of Ma- distinctiveness of populations inhabiting
buya macleani might be something like 520 small islands, and the apparently rapid evo-
individuals. In any case, a density of 12 in lutionary rates involved in achieving this
200 m^ (400 per ha) far exceeds that of differentiation, have long been known to
skinks or other comparable ground lizards students of the zoology of archipelagos
(e.g., Ameiva) anywhere in the Antilles. (Mayr 1963, Lazell 1972). Mabuya ma-
On 27 October 1994, three Mabuya ma- cleani appears to be an example ofthis phe-
cleani were seen in about two hours on the nomenon. Two aspects of Carrot Rock's
Rock. JL's field notes mention "the incred- geographic situation, in particular, argue for
ible drought," and the appearance ofthe is- rapid evolution (Fig. 6). First is its short
land as "dead gray still" and "really distance, approximately 400 m, from Peter
bleak." Hurricanes Luis and Marilyn struck Island. Given this short distance, and the
the Virgin Islands in September 1995. Low- predominant direction of the currents from
lying areas Uke Carrot Rockwereinundated the northeast, there seems a considerable
with sea water, butthere was relativelylittle probability of waifdispersal ofskinks from
mitigating rainfall. On this day the Rock adjacent parts ofPeterIsland (where typical
was visited with a group of 6 people, but M. sloanii does occur: MCZ 182273) or
in two hours ashore we saw but two skinks other islands to windward. Divergence of
and not a singleAnolis. Approximately half the Carrot Rock population, especially by
880 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OFWASHINGTON
genetic drift, would have to proceed at a species on the island, Sphaerodactylus ma-
high rate to offset the genetically homoge- crolepis is not distinct). As Mayr (1963)
nizing effects of immigration events. noted, production of evolutionary novelties
The second aspect is the short time dur- or new species in small, isolated popula-
ing which Carrot Rock has existed as a sep- tions is a rare event; ifthe initiating genetic
arate island. Lowered sea levels during the events are stochastic in nature, it is highly
last glacial period united all of the islands unlikely that, of all the many islands on the
of the Puerto Rico Bank into a single large Bank inhabited by Mabuya (Fig. 5) and A.
&
island (Heatwole MacKenzie 1967). The cristatellus (MacLean 1982, Mayer 1989),
age ofseparation oftwo islands on thebank these events should occur in both taxa on
can be inferred from the maximum depth of the same island. Weinferthatthereis some-
the water now separating them, and the thing about Carrot Rock itselfwhich is con-
time course of the Holocene sea level rise. ducive to divergence, rather than that there
We cannot say with certainty what the max- have been two independent occurrences of
imum depth of the channel between Carrot arare stochastic event. We can only suggest
Rock and Peter Island is, because the chan- that it is the unusual environmental condi-
nel is so shallow and strewn with rocks that tions of the island (see above under Ecol-
only the smallest of boats attempt to pass ogy) that are the conmion factor in diver-
through the strait, so that accurate sound- gence of the two species, but could only
ings are notpossible; it cannotbemorethan speculate about the exact conditions influ-
a few meters. Based on the time course of encing one or the other species. Although
Caribbean sea level rise (Fairbanks 1989), it is often argued that evolution proceeds
m
a depth of 5 would correspond to a sep- faster in smaller populations, adaptive di-
aration of 3000 years; as the actual depth is vergence in fact is faster and greaterinlarg-
&
almost certainly less than this, this is an up- er populations (Weber Diggins 1990,
per limit on the time of isolation of Carrot Coyne et al. 1997), making the situation of
Rock. two endemic lizards on Carrot Rock even
The causes of rapid divergence in island more remarkable.
populations have long been a matterofcon- An alternative explanation for the ende-
tendon (Williamson 1981, Berry 1986, micity exhibited by the Carrot Rock sau-
Grant 1998), with some arguing for the im- rofauna is that they are relicts, stranded
portance of stochastic factors (e.g., Mayr there by post-glacial sea level rise. While
1954), while others have stressed the adap- we cannot definitively rule out this possi-
tive nature of island differentiation (e.g.. bility, the close proximity and very recent
Grant 1968, Malhotra & Thorpe 2000). separation of Carrot Rock from the main
Carrot Rock's proximity to Peter Island ar- body of the Virgins, and its small size,
gues for a non-stochastic cause, since even make it an unlikely refuge for species
a low rate ofmigration is sufficientto coun- which have elsewhere gone extinct. Three
teract divergence due to founder effect or distributional patterns in the Virgin Islands
drift (Crow and Kimura 1970); gene flow herpetofauna suggest a relictual distribu-
is much less effective in counteracting se- tion, but none match that ofthe Carrot Rock
lection (Lande 1980). There is another, in endemics. Sphaerodactylus townsendi, an
our; opinion much stronger, argument for otherwise Puerto Rican species, occurs in
non-stochastic causes: the occurrence on the Virgins only on Frenchcap Cay (Hea-
Carrot Rock of another endemic lizard, An- twole et al. 1981), but this cay, separated
olis ernestwilliamsi, which, like Mabuya by depths of 22 m, was the earliest of the
macleani, has a close relative, A. cristatel- Virgins to be isolated, about 8000 yr ago
lus, widespread on other islands on the when Puerto Rican forms apparently ranged
Puerto Rican Bank (Lazell 1983; the third further east on the then exposed bank. Sev-