Geographically widespread species commonly show genetic differences among different areas of their distribution. Differences presumably arise from some degree of isolation by ecological barriers, or mere distance, with natural selection favoring different genotypes in different areas. Isolation is assumed to be prerequisite for selection to produce statistically demonstrable genetic differentiation. In contrast, genetic uniformity is taken to indicate free gene flow among localities, or the existence of some form of balancing selection (Prakash et al., 1969). The existence of genetic differences among relatively restricted areas of prolific marine organisms with planktonic dispersal stages (Frydenberg et al., 1965; M0ller, 1968, 1969; Koehn et al., 1973) has led us to question this kind of reasoning. In the cod, for instance, passive dispersal of the planktonic eggs and larvae by ocean currents should last for weeks or months, notably in colder parts of its range. Ocean current velocities are such that dispersal for hundreds of kilometers should be common. However, expected genetic uniformity over hundreds or even thousands of kilometers is not supported by observation. There are consistent differences in the hemoglobin allele frequencies along the Norwegian coast (Frydenberg et al., 1965) as well as between inshore and offshore samples (M0ller, 1969). There are additionally, significant differences with respect to depth of the samples in both hemoglobin and blood type allele frequencies (M0ller, 1968). Microgeographically, similar observations can be made, such as the significant differences at the tetrazolium oxidase locus in the marine pelecypod Modiolus demissus between high and low areas of the tidal zone (Koehn et al., 1973). We here suggest that isolation is not prerequisite to the above described genetic differentiation, but that every locality derives its adult stock from the same zygote pool. Given selection of sufficient intensity, striking genetic differences can be produced between the sedentary adult stocks that are widely dispersed as young and belong to the same panmictic population. Also, recent literature that isoenzyme variation is responsive to local environmental conditions (Koehn and Rasmussen, 1967; O'Gower and Nicol, 1968; Johnson et al., 1969; Koehn, 1969; Prakash et al., 1969; Richmond, 1970; Marshall and Allard, 1970; Koehn et al., 1971; Smith and Koehn, 1971; Koehn and Mitton, 1972; Merritt, 1972) suggests that genetic isolation is not prerequisite for differing genetic composition throughout the range of the species. While most marine organisms with pelagic larvae breed in various geographic areas, and therefore do not offer a simple situation in which the relative importance of selection and isolation can be estimated, in the American eel, Anguilla rostrata, all divergence between localities must necessarily arise in the absence of isolation. The thorough investigations of Schmidt (1922, 1925) indicate that all spawning by the American eel takes place in a circumscribed region northeast of the West Indies. O'cean currents disperse the leptocephalus larvae to American shores from northern South America to the Arctic. The leptocephali in coastal waters metamorphose into elvers that move into estuarine or fresh waters, where they grow to adult size. These resi-