WAPLES and ROSENBLATT: LARVAL DRIFT IN SOUTHERN CALIFORNIA 



strength of the agreement (or disagreement) be- 

 tween these two sets of rankings: 



1 - (6 Z rff/[n(w2 - 1)]) 



(2) 



where n is the number of items ranked (in this case, 

 4 locahties or 6 pairs of localities) and d, is the dif- 

 ference in rankings of the i* locality (or pair of 

 localities). 



RESULTS 



The electrophoretic analysis provided information 

 regarding variation at 32-42 presumptive gene loci 

 in the 10 species. The genetic interpretation of band- 

 ing patterns was guided by comparisons of observed 

 and expected number of bands exhibited by pre- 

 sumed heterozygotes, by tissue specificity of iso- 

 zyme expression, and by quality and consistency of 

 resolution. A detailed discussion of results for each 

 enzyme can be found in Waples (1986). Except for 

 Semicossyphus pulcher (discussed below), no over- 

 all departures of heterozygote frequencies from 

 those expected under conditions of Hardy- Weinberg 

 equilibrium were found (Waples 1986, in press). 



Table 2 summarizes the allozyme data. Average 

 heterozygosities for the 10 species (mean H = 0.031; 

 range = 0.009-0.087) are somewhat lower than the 

 mean value of 0.055 reported for over 100 marine 

 fishes by Smith and Fujio (1982), but at least 5 loci 

 {Embiotoca jacksoni) and as many as 19 loci (Ly- 

 thrypnus dalli) were found to be polymorphic in each 

 species. Space does not permit reporting here the 

 allele frequencies for all of these variable loci; these 

 data appear in Waples (1986), or can be obtained 

 from the first author. 



Interpopulational genetic distance values (Table 

 2) were generally fairly small: for half of the species 



{Alloclinus holderi, Chromis punctipinnis, Girella 

 nigricans, Medialuna calif omiensis, Paralabrax 

 clathratus) all possible pairwise comparisons of 

 populations yielded D values <0.001. Even the 

 largest observed D value (0.029 for the Guadalupe- 

 Punta Eugenia comparison in E. jacksoni) is well 

 within the range of values typically found between 

 conspecific populations of fish species (Shaklee et 

 al. 1982; Thorpe 1983). Nevertheless, it is apparent 

 that populations of most of these shore fishes do not 

 behave as a single panmictic unit. For 8 of the 10 

 species, significantly nonzero single-locus Fgj 

 values indicate heterogeneity of allele frequencies 

 among populations (Table 2; see also Waples 1986). 

 Furthermore, the statistically significant tendency 

 for species that are better dispersers to have lower 

 mean D values (Waples in press) suggests that the 

 relatively small D values reported here for most 

 species contain valid information relating to popula- 

 tion structure. 



Our interest here is primarily to identify recur- 

 ring patterns (across species) of genetic similarity 

 between areas. One way to approach this topic is 

 to compute, for each species, a mean of all the pair- 

 wise D values involving each locality. In Table 3 

 these mean D values have been ranked within each 

 species, thus providing an indication of which 

 populations are most similar (or dissimilar) gene- 

 tically to the other populations as a whole. Two 

 species (A. holderi, L. dalli) that could be collected 

 from only three of the four areas have been deleted 

 from this analysis. 



The two southern populations, Guadalupe and 

 Punta Eugenia (total of rankings for each = 15), are 

 consistently more divergent than are La Jolla (24.5) 

 and San Nicolas (25.5). Substitution of these totals 

 and values for a (8 species) and b (4 localities) into 

 Equation (1) yields a x^ value of 7.54 with 3 df. This 



Table 2.— Summary of electrophoretic results. Number of loci surveyed (T), number polymor- 

 phic (P), and number with significantly nonzero Fgj values (F) are indicated. H = average 

 heterozygosity; L = La Jolla; C = Channel Islands; E = Punta Eugenia; G = Islade Guadalupe. 



'Mean of comparisons involving Cabo Thurloe and Islas de San Benitos. 



