16. P Jung, Palaeontol. Schweiz. Abh. 111, 1 (1989) Ing any other interval.

any other interval. The subgenera character- extended through the late Pliocene and early
17. All species in (16) plus those from our new istic of different oceans were: Sincola (Sinaxila), Pleistocene (30).
Caribbean and Pacific isthmian collections (3) exclusively Caribbean; S~ncola(Sincola), restrict- 32. H J. Dowsett and R Z. Poore, Quat. Sci Rev. 10,
18 Fossil collections are from the formatlons de- ed to Caribbean since the late Miocene, Strom- 189 (1991); H. J. Dowsett et a/., Science 258,
scribed in (3) Age Intervals nearly colncide with bina (Recun/lna) and Cotonopsis (Cotonopsis), 1133 (1992)
the time scales of W. A. Berggren, D. V Kent, J J. restr~ctedto the eastern Paclfic except for a single 33. D. H Geary, T. A. Brieske, B. E. Bemis, Palaios 7,
Flynn, J. A. Van Couvering, Geol. Soc Am. Bull. Pllocene species in each 77 (1992); D. H. Geary et a/., Paleontol. Soc
96, 1407 (1985) and W B. Harland et a1 [A 24 Seven specimens of Strombina (Strombina ?) aff Spec. Publ. 6 (1992), p 108.
Geologic Time Scale, 1989 (Cambridge Unlv. pumiliofrom a slngle locality in the eastern Pacific 34 F M. Bayer, G. L. Voss, C. R.Robins, Bioenviron-
Press, Cambr~dge, 1990)l. However, all our RIO Corredor section of the Armuelles Formation mental and Radiological Safety Studies Atlant~c-
Pllocene sections overlap during the interval be- are most s~milarto Strombina (Strombina) pumilio Pacific Interoceanic Canal. Repori on the Marine
tween 3.5 and 3 6 million years ago (3). There- from the late Pliocene to modern Venezuela and a Fauna and Benthic Shelf-Slope Communities of
fore, faunas from sections ranging upward from slngle modern speclmen from the Gulf of Darien. the Isthmian Region (University of Miami, Mlami,
thls interval were counted as late Pliocene and Five specimens of Cotonopsis (Cotonopsis) cf. FL, 1970).
those ranging downward from the same interval mendozana from four localities scattered through 35. A. A Olsson and T. L. McGinty, Bull. Am. Paleon-
were counted as early Pl~ocene.L~kewise,age the late Pliocene Caribbean Escudo de Veraguas to1 39, 1 (1958), G. E Radwin, Trans. San Diego
assignments for the Lomas del Mar section over- Formation are most simllar to Cotonopsis (Coto- Soc Nat. Hist. 15, 229 (1969); M Arrunategui and
lap with those for the uppermost type M o ~ nFor- nopsis) mendozana, known only from the modern J. B. C Jackson, unpublished data.
matlon by 100,000 years, but the former IS unam- eastern Pacific. 36. We thank H. Dowsett and L. Bybeli for datlng the
biguously younger, on the basis of superposition 25 A. Coates, J. Obando, H. Gonzalez, Paleontol lsthmlan sections, Y. Ventocilla and K. Muller for
T h ~ srelatlon is the basis for the separation of the Soc. Spec Publ 6 (1992), p. 62. preparing the samples; J. Glbson-Smith and W.
Lomas del Mar collect~onsfrom those of the type 26 L. S Collins, /bid., p. 67 Gibson-Smith for helping to compile the faunal lists
Moin and uppermost Escudo de Veraguas Forma- 27. N. Knowlton, L. A Weigt, L. A. Solorzano, D. K. and identify taxa to genus or subgenus, H. Fortu-
tions. Intervals, approximately 200,000 years before Mills, E Bermingham, Science 260, 1629 (1993). nato for helping to analyze the data and preparing
the Pliocene-Pleistocene boundary, Lomas del Mar 28 S M. Stanley and L. D. Campbell, Nature293, 457 the flgures, L. Anderson, M. Arosemena, T Collins,
above the top of the type Moln Formatlon: late (1981); S M. Stanley, Palaios 1, 17 (1986). A Guevara, E. Lombardo, J. Obando, R. Pan-
Pliocene, upper Rio Banano Formatlon type section 29 J. G~bson-Smithand W Gibson-Smith, Geos 24, chaud, A Velarde, and many others for provldlng
at La Bomba, upper Escudo de Veraguas Forma- 11 (1979). additional assistance; and E Bermingham, A.
tion, Moln Formatlon type section, lower Rio Banano 30 A F. Budd, T. A Stemann, K. G. Johnson, Pale- Budd, A. Cheetham, E. Leigh, N. Knowlton, and P
Formation at Quitarla, lower Escudo de Veraguas onto/. Soc. Spec. Publ. 6 (1992), p 43. Morr~sfor thelr useful advlce. Supported by the
Formation, early Pliocene, Cayo Agua Formatlon, 31. Ages of the Important sectlons in Flor~da,Vene- National Geographic Society, the National Science
upper Shark Hole Point Formation, late Miocene, zuela, and the isthmus are not finally resolved (3, Foundat~on, Kugler Fund of Basel Naturhisto-
lower Shark Hole Point Formation, Nancy Point For- 16), D. S. Jones etal., J Geol 99, 637 (1991), W. risches Museum, Smlthsoniah Fellowship Pro-
matlon, Rio Sandbox section of Uscari Formation, D Allmon, Palaios, In press] but are consistent grams, George Washington University, and Smith-
mlddle and upper Gatun Formatlon The envlron- wlth the hypothesis that massive molluscan turn- sonian Troplcal Research lnstltute
ment of deposition was less than 200 m for all over throughout the region occurred between 2.0
formatlons except the Nancy Point and Uscari For- and 1.5 million years ago. Turnover of reef corals 15 January 1993, accepted 26 April 1993
mations [L S. Collins, J Paleontob, in press], whlch
contribute little to total late Mlocene dlverslty. Exclu-
sion of these deeper water faunas would only in-
crease the disparity In numbers of subgenera be-
tween late Miocene and younger faunas in Fig. 1.
Taxa were identifled to genus or subgenus by P.
Diversity of Atlantic Coastal Plain Mollusks
Jung, so that names are consistent throughout.
Numbers of taxa are total unique genera and sub-
Since the Pliocene
genera. Except for strombinids, identlficatlon to spe-
cies IS impractical pend~ngmonograph~crevis~ons Warren D. Allmon,' Gary Rosenberg, Roger W. Portell,
because most specles are undescribed.
19 H. L. Sanders, Am. Nat. 102, 243 (1968); J. C. Kevin S. Schindler
T~pper,Paleob~ology4, 423 (1979), A. E. Magur-
ran, Ecological Diversity and its Measurement About 70 percent of tropical western Atlantic mollusk species have become extinct since
(Croom Helm, London, 1988). the Pliocene, which has led to perceptions of a corresponding decline in diversity. However,
20. By interpolation from Flg. 1 to the nearest 100
specimens. 2500 specimens for the Plio-Ple~s-
a compilation of gastropod species from Plio-Pleistocene faunas of the United States
tocene, 5900 for the late Pliocene, 5500 for the Atlantic coastal plain and from Recent western Atlantic faunas indicates that regional
early Pllocene, and 15,600 for the late Miocene. diversity has not changed since the Pliocene. Gastropod diversity in the Pliocene Pinecrest
Sampling is inadequate to estlmate total diversity Beds in Florida approximates that seen today on either coast of Florida. Gastropod diversity
during any intervai, as shown by the persistently
steep slopes of all the curves. is not demonstrably different in the Recent tropical western Atlantic than in the Recent
21. The only comparative transisthmian study (34) tropical eastern Pacific. High extinction rates must have been balanced by high origination
recorded 331 mollusk species from 58 eastern rates.
Pacific stations versus 277 species from 105
Caribbean stations, collected with the same
equipment. These counts suggest that diversity in
the eastern Pacific is twice as large as that in the
southwestern Caribbean ( 1l ) , on the assumption
of similar abundance of specimens (no data are
T h e late Cenozoic history of mollusk di- western Atlantic mollusk species are
given). However, mollusks are generally much versity in the tropical and subtropical west- thought to have become extinct since the
less abundant in the Caribbean than in the east- ern Atlantic has traditionally been viewed late Pliocene, whereas only about 30% of
ern Pacific (9, 34), so even this difference in
diversity may be due to sampling bias.
as one of a decrease in species richness by a eastern Pacific species became extinct over
22. W. D. Allmon, R. Portell, G. Rosenberg, and K. series of extinction episodes, from a Mi- the same interval (2, 3). We have compiled
Schindler [Paleontol. Soc. Spec Publ. 6 (1992), ocene-Pliocene peak to the depauperate lists, in part based on our own new collec-
p. 4: W. D. Allmon, G. Rosenberg, R. Portell, K. fauna of today ( 1 4 ) . In what has been tions, of gastropod species from the Recent
Schindler, Science 260, 1626 (1993)l made a
new compilation of described shelled gastro- termed a regional mass extinction, 70% of and Plio-Pleistocene of the western Atlan-
pods from Florida to northeastern Brazil. Their tic. We have used these data to test wheth-,.
list contains 2643 species, compared with 2475 er extinction has led to a substantial de-
species from the same latitudinal range In the W. D. Allmon, Paleontological Research Institution,
eastern Paclflc. Ithaca, NY 14850. crease in diversity in the region, both in
G. Rosenberg, Malacology Department, Academy of absolute terms and relative to the Recent
23 The species In common were Strombina (Strom-
Natural Sciences of Philadelph~a,Philadelphia, PA
bina) colinensis, Strombina (Strombina) lessepsi- eastern Pacific fauna.
19103
ana, Sincola (Sincola) chiriquiensis, and Sincola
(Dorslna) pigea in the late Miocene, Strombina
R. Portell and K. Schlndler, Florida Museum of Natural The fossil database comprises' lists of
(Strombina) lessepsiana and Sincola (Sincola) Histow, Universitv of Florida, Gainesvilie, FL 3261 1 species from three Plko-Pleistocene faunas
chiriquiensis in the early Pliocene: and none dur- *To whom corres~ondenceshould be addressed (Pinecrest, Caloosahatchee, and Bermont)

SCIENCE VOL. 260 11 JUNE 1993

 at two localities in southern Florida (5-7). Similarly, numbers for the living fauna of differ substantially from that reported in the
We concentrated on the Pliocene Pinecrest eastern Florida are approximately 4% great- overlying Caloosahatchee Formation (Table
fauna because it is the oldest and is there- er than those for the Pinecrest fauna. Thus, 2), which has a documented fauna of 386
fore critical to understanding patterns of although the Pinecrest fauna is about 70% species of gastropods >5 mm in maximum
diversity in the region (8). Data for mollus- extinct, it is not ~ossibleto demonstrate size. This number is about 16% fewer than
can diversity in the Pliocene of the Atlantic that ~ e c e n tdiversity in Florida has de- the 460 Pinecrest species, a difference com-
coastal plain north of Florida were taken clined as a result of Pliocene or subsequent p arable to the 12% difference todav between
from previously p~~blished sources (9-1 1 ) extinctions. counts for the east and west coasts of Florida
(Table I). The Recent database was com- The molluscan fauna of the Pinecrest has (Table 3). Documented diversitv is consid-
piled from the published literat~lre(12); it been estimated to contain as many as 1200 drably ldwer in the Bermont than in the
includes all described shelled marine species species. This number exceeds that of the ~lnderlvineCaloosahatchee. The Bermont
of gastropods known to occur between Cape Recent fauna of Florida by as much as three beds, however, were not recognized as con-
Hatteras, North Carolina, and Rio de Jan- times (2, 15, 16) and approximates Recent taining a distinct fauna until the 1960s (8)
eiro (35"N to 2 3 5 , including Bermuda). diversity of the entire Caribbean region (2). and so have been undersampled relative to
Our results give a tally of 460 species of Estimates such as these have created the the longer known Pinecrest and Caloosa-
gastropods >5 mm in maximum adult di- im~ressionof a tremendous diversity for the hatchee faunas. Not surprisingly, recent
mension in the Pliocene Pinecrest Beds fossil beds of Florida, while underestimating st~~diesindicate that the Bermont contains a
(Table 1). Of these species, 173 (38%) are Recent diversity there, which is r 1147 spe- considerable number of undescribed or Lln-
known to occur in the overlying Caloosa- cies. Although further work may reveal that reported species [for example, (18)l.
hatchee Formation (Table 2). Of the 500 the Pinecrest contains considerably more Com~arisonof diversities over time is
total gastropod species reported from the species than we have been able to docu- difficult because the Recent fauna has been
Pinecrest, only 150 survive today, a number ment, species are still being discovered in sampled better geographically than have
that reflects an extinction rate of 70%. the Recent fauna of Florida. throueh u
both the fossil faunas, whereas the latter are
About 95% of the extant Pinecrest species range extensions and new species (1 7). much better sampled temporally. The
are known to occur in depths <30 m in the Therefore, we expect that additions to the count of 333 species that'we report for the
Recent fauna (12). In addition, independent fossil fauna will be approximately balanced Pinecrest at Sarasota (Table 1) is an over-
paleoenvironmental evidence suggests that by additions to the Recent. estimate of local diversity in that it includes
the Pinecrest Beds were deposited in depths Diversity in the Pinecrest also does not many discrete beds, each of which may
5 3 0 m (1 3). Therefore, we used 30 m as the
cutoff point in making comparisons between
Table 1. Diversity of gastropods in Pllo-Pleistocene stratigraphic unlts of the Atlantic coastal plain.
Recent and Pinecrest diversity.
To obtain an estimate of diversity repre-
Locality Total Specles Data
senting all size ranges, we added to the 460 species >5 m m source
species > 5 mm a proportion of species with
maximum observed size 5 5 mm that corre- Bermont Belle Glade 261 225 7
sponds to the proportion present in the Bermont Leisey 95 '86 30*
Recent Florida shallow-water fauna (22%) Bermont Total 304 259 7, 30*
James City Total 99 99 10
(Table 3). We thus obtained a total of Waccamaw Total 21 3 200 9
about 590 species of gastropods. About 21 1 Caloosahatchee Sarasota? 42 40 *
bivalve species have been reported in the Caloosahatchee Total 495 386 9
Pinecrest after intensive recent studv, ( 3,),.
~
Chowan River Total 113 113 10
giving a total for bivalves and gastropods of Bear Bluff Total 22 22 10
801 (14). Duplin North Carol~na 21 1 195 9
Duplin South Carolina 21 6 198 9
Pinecrest bivalve diversity can be used Jackson Bluff Total 170 144 9
to estimate expected gastropod diversity. Yorktown North Carolina 117 101 9
There are 320 species of bivalves known to Yorktown Virginia 293 275 9
occur in water depths <30 m in Florida Pinecrest Sarasota$ 333 333 31*
[compiled from Florida references in (12)l. Pinecrest Total 500 460 9, 31*
Thus, the gastropod:bivalve ratio is 2.58: 1 *Florida Museum of Natural History collections. tunit 1 , APAC Quarry (13). SUn~ts10 to 2, APAC Quarry
for the Recent Florida fauna (827:320). (13).
Applying this ratio to the 211 Pinecrest
bivalves, we expect a total of 544 Pinecrest Table 2. Molluscan origination and extinction for selected Plio-Pleistocene units of the Atlantic
gastropods. We consider our ,estimate of coastal plain. See (8)for stratigraphic relations. The calculat~onsare based only on species >5 m m
590 Pinecrest gastropods reliable because it (Table 1 ) .YDJP, putatlve synoptic time slhce of late Pliocene age comprised of the Yorktown, Duplin,
is within 8% of the number expected on the Jackson Bluff, and Pinecrest units; CB, Chowan River and Bear Bluff formations;JWC, James City,
basis of diversity of bivalves. Waccamaw, and Caloosahatchee formations. The gastropod data are from Table 1; bivalve data,
Our estimates of diversitv of Pinecrest indicated by the abbreviation b, are from (3).
gastropods are comparable to' our estimates
for the Recent shallow water gastropod Number of species
Apparent Apparent
fauna of Florida (12) (Fig. I). The count of Unit extinction orlglnation
Older Younger Shared
460 fossil species > 5 mm is about 7% unit unit (%) ?/.)
higher than the 429 species in this size
range now living in western Florida; the P~necrest-Caloosahatchee 460 386 173 62.4 55.2
estimate of 590 Pinecrest species of all sizes Pinecrest-Caloosahatchee (b) 21 1 150 110 47.9 26.7
YDJP-CB 600 100 53 91 .1 47.0
is only 8% greater than the 545 living CB-JWC 100 462 75 25.0 83.8
species documented in western Florida.

SCIENCE VOL. 260 11 JUNE 1993 1627

reflect an average of diversity over time Table 3. Specles diversity of Recent shelled mus, the onset of Northern Hemisphere
(13). The diversity of a single horizon may gastropods in Florida. The slze and depth rang- glaciation (28), or both. These mechanisms
es are those covered by analysls of Pinecrest could also be used to explain a simultaneous
be considerably lower than 300 species. and Caloosahatchee fossll assemblages (<30
Two st~~dies of restricted local shallow mol- m depth and >5 mm In maxlmum adult dimen- increase in molluscan originations and ex-
luscan faunas in Florida reported 181 and sion). The counts for western (W.) and eastern tinctions in the late Pliocene.
155 species of gastropods, respectively (19). (E.) Florlda do not include tallies for the Florida
Thus. local diversitv, mav, not have changed - Keys. REFERENCES AND NOTES
significantly in Florida since the Pliocene. - -

1. A A. Olsson, Mollusks of the Tropical Eastern Pacif-

Confirmation of this hv~othesis
A, would re- Specles ic, Parliculariy from the Southern Half of the
quire quantitative sampling in Recent sites Total Specles <30 m and Panamic-Pacific Faunal Province. Panamic-Pac~fic
specles <30 m 2 5 mm Pelecypoda (Paleontolog~calResearch Institution,
and fossil horizons that represent compara-
(%) Ithaca, NY, 1961);W. P. Woodring, Proc. Am Philos.
ble environments. Soc. 110, 425 (1966); G J. Vermelj, Biogeography
There is thus no evidence that mollus- Florlda 1171 827 645 (78 0) and Adaptation (Haward Univ. Press, Cambridge,
can diversity has changed appreciably in W Florlda 710 545 429 (78 7) MA, 1978); Science 253. 1099 (1991)
E Flor~da 778 617 490 (79 4) 2 S. M. Stanley and L. D. Campbell, Nature293, 457
Florida from the late Pliocene to the Re- (1981).
cent: the Pinecrest, Caloosahatchee, and 3. S. M. Stanley, Palaios I,17 (1986).
Recent faunas do not differ substantially in 4. A. M. Keen, Sea Shells of Tropical West America
diversitv in the western Atlantic. The Re- (Stanford Univ. Press, Stanford, CA, ed. 2, 1971).
number of species. The late Neogene re-
5. To construct these lists, we identified every gas-
gional mass extinction recognized in the cent fauna of the tropical western Atlantic tropod 2 5 mm In marlmum dimension from stratl-
western Atlantic (2,, . 3).
, , and best reoresent- has traditionally been viewed as less diverse graphically collected bulk samples, found in the
ed in strata exposed in southern Florida (3), than that of the tropical eastern Pacific Pinecrest Beds and Caloosahatchee Formation
exposed in two quarries in Sarasota County and In
must therefore have been a phenomenon of (14, having declined from approximate the Bermont Formation exposed in a quarry in
not only high extinction but also high parity before the uplift of the Central Hillsborough County. Specimens <5 mm were
origination. During the time of this extinc- American Isthmus to a depauperate condi- excluded because of the difficulty of sortlng and
Identitying mollusks of this size range. The Pine-
tion, origination rates exceeding 50% oc- tion at present. However, our data show crest Beds [units 10 to 2 of Petuch (IE)], were
curred not only in Florida but also farther that the western Atlantic has almost 7% sampled at APAC Quarry and Quallty Aggregates
north along the Atlantic coastal plain (Ta- more species of shelled gastropods than in Quarry, northern Sarasota County, Florida. The
Caloosahatchee Formation [unit 1 of Petuch (16)]
ble 2). Further research is needed to parse the comparable latitudinal range of the was sampled at APAC Quarry [see (13) for site
these originations into components of spe- tropical eastern Pacific (3 1°N to 6"s) (23) description]. The Bermont Formation was sam-
ciation, immigration, and lineage transfor- (2643 versus 2475); the Gulf of Mexico and pled at Leisey Shell Plt, H~llsborough County,
Caribbean alone (30°N to 8"N) contain Florida [see S. D. Webb et a/., Quat. Res. 32, 96
mation. Nevertheless, the regional mass (1989) for site description]. Taxa that we could
extinction in Florida may be an example of 2362 species (1 2). The tropical fauna of the identify only to genus level (some probably unde-
the often observed linkage between specia- western Atlantic is therefore not demon- scribed) were included in the tallies. Specimens
tion and extinction (20) and may be similar strablv less diverse than that of the eastern from these collections are housed in the Florida
Museum of Natural Hlstory (FLMNH).
in this respect to other regional episodes of pacific. If it has suffered more extinction 6. We supplemented data f r ~ m the bulk samples by
extinction in the geological record (2 1). than has the eastern Pacific, then it has also reference to earlier species lists for the Pinecrest
Similar patterns have been documented for seen more origination. [(IE); E. J. Petuch, J Coastal Res. 2, 391 (1986);
ibid. 3, 189 (1987); W H. Dall Paleontol. Res
gastropods in the late Neogene record of Western Atlantic faunal history farther Center Spec. Publ. 1 (1991); J. J. Tripp, Ann.
the southern Caribbean, where there was north may have been different. In North Carnegie Mus. 57, 259 (1988); E H. Vokes, Tu-
also substantial origination with no decline Carolina and Virginia, diversity has de- lane Stud. Geol Paleontol. 23, 1 (1990); ibid. 25,
1 (1992); (g)] and for the Caloosahatchee (9).
in diversity during this time (22). clined since the Pliocene. Virginia has 7. We supplemented bulk sample data from the
Even at the level of faunal province, we about 100 Recent shallow water gastropod Bermont wlth that of S. E Hoerle [Tulane Stud.
see no evidence for an overall decline in species >5 mm in maximum size, and Geol. Paleontol. 8, 56 (1970)l We updated Hoe-
North Carolina has about 230 such species rle's species count on the basis of nomenclatural
changes since 1970 and our comparison with
(12), whereas the Pliocene Yorktown For- newer collections at FLMNH.
mation of Virginia has 275 species in this 8. Stratigraphic terminology in the Florida Plio-Pleis-
size range (Table I). The reason for this tocene is unclear. Most unlts have been defined
faunally rather than lithostratigraphically and thus
difference may be different climatic histo- properly should not be referred to as formations.
ries north and south of Cape Hatteras (24). See W. G. Lyons, Bull. Fla. Mus. Nat Hist. 35, 131
Virginia and North Carolina north of Cape (1991) and T M. Scott and W D. Allmon, Eds.,
"The Plio-Pleistocene stratigraphy and paleontol-
Hatteras have temperate Recent faunas, ogy of southern Florida," FI. Geol Sun/ Spec
and climatic cooling may have caused the Pub. 36 ( I 992), p 1. Some aspects of correlation
loss of diversity since the Pliocene (2, 3, between Plio-Pleistocene units !n Florida and
those of the Atlantlc coastal plain are disputed.
25). South of Cape Hatteras, however, We have used the relations presented by J. E.
0 large degrees of Neogene extinction cannot Hazel [Smithsonian Contnb Paleobiol 53. 81
1 10 100
Depth (m)
1000 -
be attributed to climatic cooling alone. (1983)l and in (13, 25)
9. L D Campbell, S Campbell, D Colquhoun, J.
Our results suggest that any environ-
Ernlssee S.C. Geol Notes 19, 51 (1975) We
Fig. 1. Cumulative number of species by depth mental change associated with Plio-Pleis- used data from thls publication as given, assurn;
for Recent Floridian gastropods >5 mm in tocene faunal change produced not only ing synonymies to be correct
maximum size. Each species is counted from high rates of extinction but also high rates 10. L W. Ward and B W. Blackwelder, Smithson.
its m~nimumknown depth in the western Atlan- Contrib Paleobiol 61, 113 (1987) (James Clty
of origination (26). Other mechanisms may Formation), L W. Ward and N L Glllnsky Va
tic. The circle representing total P~necrestdl-
versity (Table 1) is placed at 30 m, on the basis
have been present in this system, such as Mus. Nat. H~st.Mem , In press (Chowan Rlver and
changes in sea level (25, 27) or, more Bear Bluff formations)
of independent paleoenvironmental data (73).
probably, changes in productivity caused by 11 To make these data comparable to our sample data,
Symbols: filled squares, all Florida, open we determined a marlmum adult slze of each spe-
squares, East Florida; and diamonds, West variations in oceanic circulation associated cies from the collection at FLMNH or from the
Florida. with closure of the Central American Isth- l~teratureComparisons among the faunas were car-

SCIENCE VOL. 260 11 JUNE 1993

 r ~ e dout on the bas~sof only those specles known to 21. K. W. Flessa et a / , In Patterns and Processes in cifically and unambiguously described as
reach sizes >5 mm in any linear dlmens~on the Histow of Life, D M. Raup and D. Jablonski,
Eds (Spr~nger-Verlag,Berlin, 1986) p p 232-257.
each other's closest relatives on the basis of
12 G. Rosenberg, Am Malacol. Bull, in press. More
than 1000 publications on western Atlantic mol- 22. J B. C. Jackson, P. Jung, A G Coates, L. S. morphological criteria (1 0). Collections
lusks were consulted in the compilat~onof this Coll~ns.Science 260, 1624 (1993) along the two coasts and adjacent islands of
data base 23. Eastern Pac~ficdata from (4) as rev~sedby C. central Panama at depths less than 5 m
13 W D Allmon, Palaios 8, 183 (1993) Skoglund, Festivus (suppl.) 24, 1 (1992).
14 We estlmate that the total shelled molluscan fauna 24. T M Cron~n,Quat Sci. Rev 10, 175 (1991).
revealed unrecognized sibling species in ad-
of the P~necrestis about 1050 species. We ob- 25. L W Ward, R. H. Bailey, J. G Carter, in The dition to these pairs (I I). In total, we
tained this number by tak~ngthe estimated 801 Geology of the Carolinas, J. W Horton and V. A examined 17 taxa (Table I ) : two unambig-
P~necrestgastropods and bivalves and add~ng Zullo, Eds. (Un~v,of Tennessee Press, Knoxv~lle,
20% for specles yet to be discovered and 5% for uous pairs (P4-C4, P5-C5), three triplets
1991). pp. 274-289
two less diverse classes, the ch~tonsand scapho- 26. S. M. Stanley, Paleobiology 12, 89 (1986): W. D
(P3-C3, P3-C3'; P6-C6, P6'-C6; P7-C7,
pods (29). This number is the same as a widely Allmon Oxford Suw. Evol. Biol. 8, 219 (1992) P7'-C7), and one quartet (PI-Cl, PI-C2,
c ~ t e dest~matefor the P~necrest( 9 ) ,wh~chl~sted
300 species, including 170 gastropods, and gave
27 B W. Blackwelder, Palaeogeogr Palaeoclimatol P2-Cl. P2-C2). We used shared anatomi-
Palaeoecol 34, 87 (1981), T. M Cronin, Geol. cal and color pattern character states (12)
a total of about 1000 species.
Soc. Am Bull 92, 812 (1981); H. J. Dowsett and
15 A A Olsson, in Miami Geological Society Fieldtrip
T M. Cron~n,Geology 18, 435 (1990). to posit relations within the triplets and
Guidebook, R D. Perk~ns,comp~ler(Mlam~,FL
28. W D. Allmon, Palaeogeogr. Palaeoclimatol. quartet. The result was seven morphologi-
1968), p 75
Palaeoecol 92,41 (1992), and references therein cally defined transisthmian sister species
16 E J Petuch, Proc Acad. Nat. Sci. Philadelphia
134, 12 (1982). 29. D. Nicol Nautilus 91. 4 (1977). pairs (bold in Tables 1 and 2).
17. In 1977, the Florida fauna was tallled at 1085 30 R. W Portell, K. S. Sch~ndler,G M. Morgan, in
"The Plio-Pleistocene strat~graphyand paleontol- For each taxon, we characterized allo-
gastropods and b~valves (29); here we have
tallled 1147 (827 gastropods and 320 b~valves), ogy of southern Florida," T. M. Scott and W. D zymes by using conventional starch gel elec-
an Increase of almost 6%. Allmon, Eds., Fla. Geol. Sun/ Spec Publ 36 trophoresis (13) and sequenced a segment of
18. E. J Petuch, Neogene Histow of Tropical Ameri- (1992), pp. 181-1 94
31. R W. Portell, K. S Sch~ndler,D S. Jones, Tulane
the mtDNA cytochrome oxidase I (COI)
can Mollusks (Coastal Education and Research
Foundation, Charlottesville, VA, 1988). Stud. Geol Paleontol., in press. gene (14). Aggressive behavior was used as
19. J K. Reed and P. M. Mikkelsen, Bull. Mar Sci. 40, 32. We thank A Collazos K. Ketcher, and S Schellen- an estimate of behavioral comDonents of
99 (1987). W. G Lyons, Fla. Mar Res. lnst Publ. berg for assistance in data analysis and T. Cronin, J. reproductive compatibility (1.5) because
,
47 (1989) Jackson, D. Jones, and L. Ward for comments on
20. S. M Stanley, In Causes of Evolution, a Paleontolog- earl~erdrafts. Th~spaper 1s Un~versityof Flor~da snapping shrimp attack heterospecific indi-
icalPerspective, R M. Ross and W D Allmon, Eds Contribut~onto Paleontology no 41 5. viduals and all conspecifics except potential
(Univ. of Chlcago Press, Chicago, 1990), pp. 10% mates (16). We calculated genetic diver-
127. 19 January 1993: accepted 23 April 1993
gence between transisthmian pairs using
Nei's D for allozymes and Kimura's corrected
percent sequence divergence for mtDNA
(17). We estimated divergence in behavioral
Divergence in Proteins, Mitochondria1 DNA, and compatibility by standardizing measures of
Reproductive Compatibility Across the tolerance and intolerance for transisthmian
pairs against values observed in intraoce-
Isthmus of Panama anic, conspecific control matings (1 5).
These three measures of divergence con-
Nancy Knowlton,* Lee A. Weigt, Luis Anibal Solorzano, sistently support assignments of transisth-
mian sister species pairs on the basis of
DeEtta K. Mills,? Eldredge Bermingham morphology and color pattern. Within the
It is widely believed that gene flow connected many shallow water populations of the
Caribbean and eastern Pacific until the Panamaseaway closed 3.0 to 3.5 million years ago.
Measurements of biochemical and reproductive divergence for seven closely related,
transisthmian pairs of snapping shrimps (Alpheus) indicate, however, that isolation was
staggered rather than simultaneous. The four least divergent pairs provide the best es-
timate for rates of molecular divergence and speciation. Ecological, genetic, and geological
data suggest that gene flow was disrupted for the remaining three pairs by environmental
change several million years before the land barrier was complete.

Geographic isolation is thought to permit has prompted study of transisthmian sister

divergence and speciation by disruption of taxa to test the accuracy of molecular clocks
gene flow (1). Pairs of marine sister taxa and to estimate the timing of other evolu-
separated by the Isthmus of Panama are tionary events (3, 4, 8). It has been difficult
ideal for studying these processes (2-5) to interpret inconsistencies and estimate
because isolation of the Caribbean and the possible phylogenetic differences in diver-
eastern Pacific is well dated and relatively gence rates (9), however, because of the
recent (6, 7). This geological framework limited number of taxa and characters stud-
ied. To address these problems, we investi-
Fig. 1. Single most parsimon~ousphylogenetic
Smithsonian Tropical Research Institute, Apartado
gated divergence in allozymes, mitochon-
tree constructed on the basis of mtDNA se-
2072, Balboa, Republic of Panama. drial DNA (mtDNA), and reproductive quences with PAUP (78). Trans~tionswere giv-
*To whom correspondence should be addressed compatibility for seven shallow water trans- en one-quarter the weight of transversions
Alternate mailing address Smithsonian Trop~calRe- isthmian pairs of sister taxa in the snapping (based on the fourfold greater-abundance of
search Inst~tute,Unit 0948, APO AA 34002-0948, shrimp genus Alpheus. transitions than transvers~onsIn our data), and
United States.
?Present address Route 2 Box 538, Joshua, TX We used the taxonomic literature to trees were rooted by the P7-P7'-C7 clade.
76058 identify transisthmian pairs that were spe- Taxon codes are as in Table 1

SCIENCE VOL. 260 11 JUNE 1993