208 



Fishery Bulletin 90(1). 1992 



Kornfield et al. 1982, King 1984, Ryman et al. 1984). 

 The conclusion that herring spawning groups are not 

 discrete genetically -distinct stocks is further supported 

 by the results of a recent study by Kornfield and 

 Bogdanowicz (1987). They investigated the genetic 

 relationships of ripe female herring from three loca- 

 tions, including Jeffries' Ledge and some of the 1983 

 Trinity Ledge samples analyzed in this study, by 

 restriction endonuclease analysis of mitochondrial 

 DNA (mtDNA). In other species, this technique has 

 revealed genetic differentiation not uncovered by tradi- 

 tional enzyme electrophoresis (A vise et al. 1986). Korn- 

 field and Bogdanowicz (1987) found that these spawn- 

 ing groups were not completely distinguished by the 

 composite mtDNA digestion patterns generated, and 

 no consistent geographic patterns were found for the 

 unique composites. Therefore, they concluded that this 

 approach also provided no evidence for the existence 

 of genetically distinct stocks in the Gulf of Maine. 



The significant departures from C-H-W equilibrium 

 found in this and previous studies (Grant 1981, Ryman 

 et al. 1984) may be considered contradictory to the 

 hypothesis of the existence of a genetically homogenous 

 herring population. However, these departures seem 

 to be a feature of pelagic fish stocks (Smith et al. 1989). 

 These disequilibria have been variously attributed to 

 chance due to the low frequency of occurrence (Grant 

 1981, Ryman et al. 1984) and assortative mating (Smith 

 et al. 1989). The significant departure in the present 

 data has derived from an excess number of hetero- 

 zygotes of one particular allelic combination in the JL 

 1984 data. One significant departure in 16 tests is 

 slightly higher than would be expected by chance alone 

 at the 5% probability level. An excess of heterozygotes 

 can result from negative assortative mating; however, 

 the data are not sufficient to support that hypothesis. 

 Importantly, the C-H-W equilibrium applies to all gen- 

 erations in a population, thus significant departures 

 may occur if sampling does not measure all generations 

 in the same proportion in which they occur in the 

 population. Based on SL, few immature and old fish 

 were included, so this sample bias may have con- 

 tributed to the significance level. Thus, the departure 

 from C-H-W equilibrium is probably due to chance and 

 perhaps some sampling bias. However, the distribution 

 of alleles across generations within a population may 

 warrant further investigation as disequilibrium, though 

 explicable, is a feature of herring populations and some 

 age-based selection may be occurring. 



Although the genetic evidence argues for a single 

 population of herring, significant phenotypic differ- 

 ences between spawning groups have been demon- 

 strated (Parrish and Saville 1965, Burd 1969, Anthony 

 1972, Cote et al. 1980, Ryman et al. 1984). Morpho- 

 metric and meristic characters, which have a complex 



underlying genetic structure, are believed to be great- 

 ly influenced by environmental parameters (Sinder- 

 mann 1979, Ryman et al. 1984). Thus phenotypic dif- 

 ferences may not reflect genetic differentiation, and 

 small but detectable genetic differences may not sig- 

 nificantly alter phenotypic characters. Differences in 

 biochemical genetic and phenotypic variation can best 

 be demonstrated when genetic and phenotypic analyses 

 are performed on the same specimens. In their study, 

 Ryman et al. (1984) screened 17 loci from herring 

 caught in 17 locations ranging from the Gulf of Bothnia 

 to the northeast Atlantic off Norway's western coast, 

 and found significant allelic heterogeneity at only 4 loci. 

 They concluded that the results resembled those of 

 samples drawn from a single breeding population, as 

 both the genetic diversity index and genetic distances 

 were very small. They chose numbers of vertebrae and 

 keeled scales as morphological characters. Morpho- 

 logical distances were used to construct a dendogram 

 which differentiated herring in central Baltic fall 

 spawning groups from a spring spawning Baltic group 

 and the other fall spawning groups. Thus these meristic 

 characters differed to some extent despite genetic 

 similarities. Morphologic variation was partitioned by 

 nested ANOVA with localities nested within larger geo- 

 graphic areas, and genetic variation was partitioned 

 by genetic diversity analysis. They found over 99% of 

 the gene diversity within a locality, compared with 50% 

 of the phenotypic variation. Most important, <1% 

 of the gene diversity was explained by between-geo- 

 graphic-group differences, while these differences 

 explained 40% of the phenotypic variation. 



The partitioning of variance in our samples was 

 similar in many respects to that of Ryman et al. (1984). 

 Over 99% of the genetic variance in our samples also 

 occurred within a locality within a year, compared with 

 approximately 50% of the morphometric variance com- 

 ponent. However, the percent of the morphometric 

 variance component explained by differences between 

 spawning groups was similar for both the genetic and 

 morphometric components (0.1%), in contrast to the 

 large between-group morphometric variation found by 

 Ryman et al. (1984). Results from both these studies 

 demonstrate that most genetic diversity lies within a 

 single locality at one point in time, further supporting 

 the hypothesis that herring form a single panmictic 

 population. Thus the current situation seems to be that 

 despite the existence of discrete, defined spawning 

 groups and apparent high homing fidelity, enough gene 

 flow exists between spawning groups to prevent North- 

 west Atlantic herring from evolving into genetically 

 distinct stocks. Alternatively, herring may have begun 

 this process in recent geographic time, so that genetic 

 differences have not had time to evolve. This lack 

 of genetic differentiation also means that observed 



