368 



Fishery Bulletin 91 12). 1993 



We interpret the significant heterogeneity in 

 haplotype frequencies between specimens from the east- 

 ern Gulf of Mexico and southern Brazil as genetic stock 

 differences between Spanish sardines of the western 

 North and South Atlantic, rather than as species dif- 

 ferences as, for example, between S. aurita and S. 

 brasiliensis. The one specimen of this study referable 

 to S. brasiliensis, with a gill-raker count of 154 at 

 150mmSL, came from southern Brazil, but it had one 

 of the mtDNA composite haplotypes (haplotype 2; Table 

 4) present among specimens from the eastern Gulf of 

 Mexico. All other specimens from both Gulf and Bra- 

 zilian waters possessed the lower gill-raker counts typi- 

 cal of S. aurita. 



This interpretation also agrees with results of the 

 electrophoretic study of Wilson & Alberdi ( 1991 ). After 

 comparing 350 specimens from the eastern Gulf of 

 Mexico with 41 specimens of Spanish sardine from off 



southern Brazil at 36 presumptive gene loci, they found 

 no genetic evidence of more than one species of Span- 

 ish sardine in those two areas. That is, there were no 

 fixed allelic differences that would suggest the pres- 

 ence of fully segregated gene pools, as in two species. 

 Neither in the present study nor in that of Wilson & 

 Alberdi (1991), which used the same Brazilian speci- 

 mens as here, was our nominal S. brasiliensis distin- 

 guishable from S. aurita by electrophoretic genotype 

 or mtDNA haplotype. 



We think it likely that S. brasiliensis is conspecific 

 with S. aurita. Whitehead (1970) removed S. 

 brasiliensis from the synonymy of S. aurita on the 

 observation that the syntypes of S. brasiliensis from 

 off Brazil had both high and low gill-raker counts at 

 similar lengths. That is, he concluded that the syntypes 

 with low gill-raker counts were S. aurita and that the 

 others with high gill-raker counts were S. brasiliensis 

 of Steindachner. Whitehead (1973) again proposed the 

 specificity of S. brasiliensis on this character based on 

 his review of published data on gill-raker counts of 

 western Atlantic Sardinella, but he did not report data 

 from any additional specimens. Whitehead (1985) pro- 

 posed that S. brasiliensis might be distinguished from 

 S. aurita based on the shape of the gill rakers forming 

 the branchial basket, i.e., flattened in S. aurita vs. 

 curled in S. brasiliensis, but he did not specifically 

 connect this character to catalogued specimens which 

 he examined. No mention was made of this character 

 in his prior discussion of the syntypes (Whitehead 

 1970). 



To our knowledge, no thorough quantitative study 

 of variation in gill-raker counts with SL among Brazil- 

 ian sardines has been published. Montero & Perez 



