700 



Fishery Bulletin 94(4). 1996 



heterozygosities ranging from 0.083 in O. atlanticum 

 to 0.181 in N. sp.: the four commercial species (A. 

 niger,A. verrucosus, N. rhomboidalis, and P. macu- 

 latus) had values ranging from 0.105 to 0.127. De- 

 spite the lowest mean sample size, N. sp. showed the 

 highest proportion of polymorphic loci (46.2%) and 

 the highest average heterozygosity. 



Oreosoma atlanticum was the most divergent of 

 the oreosomatids (Table 5). Its average genetic iden- 

 tity (Nei, 1978) (0 indicates complete dissimilarity 

 and 1 complete similarity) with the other species was 

 0.371 (range 0.313 to 0.426 for 26 loci). The two most 

 similar species were N. rhomboidalis and N. helgae, 

 with a high genetic identity of 0.973. The third 

 Neocyttus species, N. sp., had a relatively lower iden- 

 tity with the other two Neocyttus species: 0.903 with 

 N. rhomboidalis and 0.884 with N. helgae. The two 

 Allocyttus species had a genetic identity of only 0.695. 



The three outgroup species were very divergent 

 from both the oreosomatids and each other (Table 

 5). The acanthurid N. tuberosus diverged most from 

 the oreosomatids, with an average genetic identity 

 of 0.112 (range 0.085 to 0.180, from 23 loci). Cyttus 

 australis had a mean identity with the oreosomatids 

 of 0. 17 1 (range 0. 108 to 0. 199, 26 loci ) and B. splendens 

 a mean identity of 0.164 (range 0.115 to 0.222, 22 loci). 



In the acanthurid N. tuberosus, 16 of the 23 

 scorable loci were diagnostic (no shared alleles with 

 any oreosomatid); in the zeid C. australis, 15 of 26 

 loci; and in the berycid B. splendens, 15 of 22 loci 

 were diagnostic (Table 3). In a comparison of 

 oreosomatids with one another, O. atlanticum had 

 eleven diagnostic loci, P. maculatus, three, A. 

 verrucosus, two, whereas the other four species re- 

 vealed that only the muscle protein patterns were 

 diagnostic (Table 3; Fig. 1 ). However, when the seven 

 oreosomatids were compared pair-wise, each pair 

 (except N. rhomboidalis with either N. sp. or N. 



helgae) had at least one and up to fourteen diagnos- 

 tic allozyme loci, other than the general protein dif- 

 ference (Table 6). With the addition of between 4 and 

 13 loci showing significant allele frequency differen- 

 tiation (P<0.05, with Bonferroni adjustment for mul- 

 tiple tests) (Table 6), even the closely related N. 

 rhomboidalis and N. helgae were found to differ at 

 five loci (FH*, GPI-A*, GPI-B*, G3PDH-2*, and 

 sSOD*) and N. rhomboidalis and N. sp. at eight loci 

 (sAAT-2*, CK-A*, ESTD*, GPI-B*, LDH-1*, MP1*, 

 PGDH*, and PGM-1*). The locus PGDH* was diag- 

 nostic between N. sp. and N. helgae, with significant 

 allele frequency differences at a further seven loci 

 (CK-A*, ESTD*, G3PDH-2*, LDH-1*, MPI*, PGM- 

 1*, and sSOD*). Oreosoma atlanticum differed from 



AN 

 (226) 



W 

 (178) 



NA 

 (13) 



Ml 

 (23) 



NR 



(4221 



OA 



(25) 



I'M 

 (248) 



Figure 1 



Banding pattern observed for each oreosomatid species 

 with a general protein Coomassie blue stain [PROT*). 

 Dashed line indicates sample origin. Boldest band repre- 

 sents most common CKA* product in each species. Spe- 

 cies abbreviations are defined in Table 1. Numbers in pa- 

 rentheses are number of individuals scored. 



