218 



Fishery Bulletin 105(2) 



Table 2 



Frequency distributions of segmented dorsal- and anal-fin rays in the Caranx hippos species complex. 



Dorsal-fin rays 



Anal-fin rays 



Species 



19 20 21 22 23 24 n 



16 17 



18 19 



93 



10 



135 21.9 



63 19.6 



161 20.0 



100 19.8 



17 



54 9 



103 58 



69 31 



104 



14 



135 18.0 



63 16.1 



161 16.4 



100 16.3 



Dorsal + anal rays 



Pectoral-fin rays 



Species 



35 36 37 38 39 40 



41 



42 43 n 



18 



19 20 



21 



C. fischeri 



C. hippos (E. Atlantic) 

 C. hippos (W. Atlantic) 

 C. caninus 



15 



18 



83 



13 



135 39.8 



63 35.7 



161 36.3 



100 36.1 



1 12 



1 18 



3 68 



15 



57 

 21 

 50 

 38 



12 

 1 

 2 



(1959) included excellent illustrations of young C. hip- 

 pos and C. latus. As both he and Laroche et al. (2006) 

 discussed, small juveniles of these two species can not 

 be distinguished solely by pigmentation. Thus, as might 

 be expected, juveniles of C. fischeri and hippos also ap- 

 parently cannot be distinguished by color pattern. 



Comparisons and relationships The unique pigmen- 

 tation of the pectoral fin in adults, pattern of breast 

 squamation (a relatively small number of C hippos and 

 C. caninus are atypical in having the naked area of the 

 breast continue without interruption to the pectoral- 

 fin base), and the relatively large symphyseal dentary 

 canines, which are shared by all members of the hippos 

 complex, indicate their common ancestry. Of the three 

 extant species, C. fischeri is readily distinguished by 

 typically having more dorsal- and anal-fin rays (Table 2), 

 and in specimens >20 cm FL the anterior dorsal- and 

 anal-fin rays are relatively longer, and the body is deeper. 

 The pattern of bones that exhibit hyperostosis is mark- 

 edly different from that of the other species (Table 1), 

 in neither of which is the posttemporal bone hyperossi- 

 fied. The anal-fin lobe is white anteriorly in adults of C. 

 fischeri, in contrast to the uniformly lemon yellow lobe of 

 C. hippos. Adults of C. hippos also differ in having the 

 underside of the caudal peduncle bright yellow. 



The presence of hyperostosis in a particular bone is 

 presumed to be a derived condition (and conversely, the 

 absence of hyperostosis is uninformative). On the basis 

 of shared character states 5-6 (Table 1), C. hippos and 

 C. caninus are considered to be sister species, and the 

 geologically recent (-3.1 mya) rise of the Panamanian 

 Isthmus was the likely vicariant event leading to the 

 isolation and subsequent speciation of C. caninus. The 

 common ancestor of C. fischeri and C. hippos-caninus 

 presumably originated in the proto-Atlantic Ocean; 

 and the sympatric occurrence of both C. fischeri and C. 



hippos in the eastern Atlantic is likely indicative of an 

 earlier phylogenetic origin. 



Distribution African coast from Mauritania south at 

 least to Mo?amedes, southern Angola (Franca, 1954), 

 and at least historically it was present in the Mediter- 

 ranean Sea (Fig. 2). The collection of an adult C. fischeri 

 from Ascension Island indicates at least the occasional 

 vagrant occurrence at insular localities. Unconfirmed 

 historical reports of C. hippos from both Ascension 

 (Clark, 1915) and St. Helena (Edwards, 1990) are likely 

 based on misidentifications, possibly of C. fischeri. 



Tortonese (1952) discussed historical Mediterranean 

 specimens dating from the 1890s in the Giglioli Col- 

 lection and Geneva Museum and he identified these 

 specimens as Caranx hippos. Our efforts to locate these 

 or recent Mediterranean specimens of C. hippos have 

 been unsuccessful (see Tortonese, 1973, for status of 

 historical fish collections in Italy). Data that Tortonese 

 (1952) provided for two of his five specimens, as well as 

 an accompanying photograph of one of them, confirm 

 their identification as C. fischeri. We assume that all 

 five specimens were conspecific, and all Tortonese's 

 Mediterranean distributional records are plotted in 

 Figure 2. Papaconstantinou (1988) and Bilecenoglu et 

 al. (2002) cited a few additional unconfirmed literature 

 records of C. hippos from the Mediterranean, which we 

 presume were also based on misidentifications of C. 

 fischeri; these records are not shown on the distribu- 

 tion map (Fig. 21. See discussion of probable erroneous 

 recent photographic record of C. hippos from the Medi- 

 terranean in the following species account. 



This species is often found in brackish water, some- 

 times ascending rivers. The paratype series includes 

 collections, mostly of juveniles, from three different 

 river drainages. In their account of C. hippos, Norman 

 and Irvine (1947) quoted a secondary source as report- 



