476 
Fishery Bulletin 99(3) 
the anterolateral spines, convexity of the carapace, sur- 
face granulation patterns, and varied morphometric mea- 
surements (Taissoun 1969, 1972). However, neither these 
characters nor additional ones proposed by other authors 
(Williams, 1974; Rodriguez, 1980) have proven to be of 
consistent use in separating the species. Carapacial color 
and granulation vary widely, and the outer frontal teeth 
of many specimens cannot be classified as a definitive tri- 
angular (C. maracaiboensis ) or obtuse (C. bocourti) shape. 
These concerns have led some authors to report only pos- 
sible co-occurrence of C. maracaiboensis in their samples 
of C. bocourti , as in records from Puerto Rico (Buchanan 
and Stoner, 1988). 
Recently, molecular markers have proven particularly 
valuable in decapod crustacean systematics. In addition to 
reconstructing phylogenetic relationships, molecular se- 
quences are of value in recognizing questionable species 
separations (e.g. Geller et ah, 1997; Sarver et al., 1998; 
Schneider-Broussard et ah, 1998; Schubart et ah, 1998; 
Schubart et ah, in press). In our study, we used sequences 
of a mitochondrial gene (16S rRNA) to examine genetic 
differentiation between the morphologically similar spe- 
cies C. bocourti and C. maracaiboensis and to compare 
these sequences to those of other swimming crabs of the 
genus Callinectes in a phylogenetic analysis. 
Materials and methods 
Swimming crabs were caught by hand in shallows of the 
Golfete de Cuare (68°37'W, 10°06'N, State of Falcon, Vene- 
zuela) in November-December 1998 and March 1999, and 
in the shallows of El Mojan, Lago de Maracaibo (71°39'W, 
10°58'N, State of Zulia, Venezuela) in October 1999. Crabs 
were transported to the Institute Venezolano de Investig- 
aciones Cientificas, Caracas, in liquid nitrogen and then 
stored in a -70°C freezer. Two walking legs and one swim- 
ming leg were separated from each specimen, preserved 
individually in 85% ethanol, and shipped to the University 
of Louisiana, Lafayette (U.S.A.), for molecular analyses. 
Specimens were assigned to Callinectes bocourti or 
C. maracaiboensis according to morphological characters 
provided by Taissoun (1969, 1972), Williams (1974, 1978), 
Fischer (1978), and Rodriguez (1980). Eight specimens of 
C. bocoui'ti and six of C. maracaiboensis from Venezuela 
were used for the molecular analysis. In addition, a short 
fragment (-300 base pairs) of 16S mtDNA was obtained 
with an internal primer for one formalin-preserved speci- 
men of C. bocoui'ti from Colombia. These voucher speci- 
mens, lacking limbs (which had been removed for analy- 
sis), were cataloged in collections of the Laboratorio de 
Ecologia y Genetica de Poblaciones, Institute Venezolano 
de Investigaciones Cientificas (IVIC), and in the Univer- 
sity of Louisiana Zoological Collection (Table 1). For out- 
group comparisons, specimens of C. sapidus, C. oniatus 
Ordway, 1863, C. danae Smith, 1869, and Portunus ord- 
wayi (Stimpson, 1860) were sequenced (Table 1). 
Genomic DNA was isolated from the muscle tissue of one 
walking leg with a phenol-chloroform extraction (Kocher 
et al., 1989). Selective amplification of a 585-basepair (bp) 
product (547 bp excluding primer regions) from the mito- 
chondrial 16S rRNA gene was carried out by a polymerase- 
chain-reaction (PCR) (35-40 cycles; 1 min at 94°C, 1 min at 
55°C, 2 min at 72°C (denaturing, annealing, and extension 
temperatures, respectively ) with primers 16Sar (5'-CGCCT- 
GTTTATCAAAAACAT-3'), 16Sbr ( 5'-C CGGTCTGAACTC A- 
GATCACGT-3'), 1472 ( 5 '- AGATAGAAAC C AAC CTGG-3 ' ) , 
and 16L15 ( 5'-GACGATAAGACCCTATAAAGCTT-3') (for 
references to primers see Schubart et al., 2000). 16L15 is 
an internal primer and, in combination with 16Sbr, was 
used for partial amplification of the formalin-preserved 
specimen. PCR products were purified with Microcon-100® 
filters (Millipore Corp., Bedford, MA) and sequenced with 
the ABI BigDye® Terminator Mix (PE Biosystems, Welles- 
ley, MA) in an ABI Prism 310 Genetic Analyzer® (Applied 
Biosystems, Foster City, CA). Sequences were manually 
aligned with the multisequence-editing program ESEE 
(Cabot and Beckenbach, 1989). New sequences were ac- 
cessioned to the European Molecular Biology Laboratory 
(EMBL) genomic library (see Table 1). Sequences avail- 
able online (by electronic database accession numbers that 
follow) were obtained for “ Callinectes sapidus ” (U75267), 
C. ornatus (U75268), C. similis (U75269), and C. sapidus 
(AJ130813). 
Sequence divergence was analyzed by using Kimura 
2-parameter distances, UPGMA cluster analysis, and 
neighbor-joining (NJ) distance analysis (Saitou and Nei, 
1987) with the program MEGA (Kumar et al. 1993). Sta- 
tistical significance of groups within inferred trees was 
evaluated by the interior branch method (Rzhetsky and 
Nei, 1992). As a second phylogenetic method, a maximum 
parsimony (MP) analysis was carried out with a heuristic 
search and random sequence addition (tree-bisection and 
reconnection as branch-swapping option) and by omission 
of gaps with the program PAUP (Swofford, 1993). Signifi- 
cance levels were evaluated with the same software and 
the bootstrap method with 2000 replicates. 
Results 
Sequencing of 15 specimens of Callinectes bocourti and 
C. maracaiboensis from Golfete de Cuare and El Mojan 
(both Venezuela) and the Rio Sinu estuary (Colombia) 
revealed the existence of seven different haplotypes. Two 
haplotypes (ht-1 and ht-5) clearly predominated (together 
64.3%) and were found in both species (ht-1 in two Cal- 
linectes bocourti and three C. maracaiboensis from Golfete 
de Cuare, ht-5 in three C. bocoui'ti and one C. maracai- 
boensis from El Mojan). The other five haplotypes (ht-2, 
ht-3, ht-4, ht-6, ht-7) each differed from ht-1 or ht-5 in 
not more than two positions and were found in only single 
individuals (Fig. 1, Table 2). Consequently there is not 
a single diagnostic molecular character in our sequenced 
unit of 16S mtDNA that would discriminate between the 
two species of swimming crabs. On the other hand, hap- 
lotype distributions differed between the two sampled 
populations; ht-1 is found in only the Golfete de Cuare 
(-550 km east of Lake Maracaibo) and ht-5 is restricted 
to El Mojan (within Lake Maracaibo). Comparison of the 
