data from tagging experiments were also avail- 

 able . The tagging data require considerable 

 modification of conclusions based on meristic 

 or morphometric data alone (as Royce has dis- 

 cussed in some detail) . 



Ahlstrom concludes from his review 

 that, while the indirect meristic and morpho- 

 metric studies generally show heterogeneity 

 within a population, only direct methods will 

 demonstrate the actual amount of intermingling. 

 Carefully designed tagging experiments are in- 

 dicated as the solution to this dilemma. 



Turning now to some recent studies, 

 Cope has had the opportunity of studying cut- 

 throat trout in Yellowstone Lake. This sub- 

 species has long been naturally isolated (and no 

 introductions have been made) in a situation of- 

 fering a diversity of ecological habitats and, 

 presumably, a like amount of opportunity for 

 micro -evolution . 



This trout ordinarily inhabits the lake 

 and only ascends tributary streams to spawn. 

 The trout spawning in five of these tributaries 

 were examined. There was a marked tendency 

 for adult fish to return to the same stream in 

 which they had previously spawned. (Young 

 downstream migrants were not marked for hom - 

 ing studies). In addition, between -stream dif- 

 ferences were observed in the time of the 

 spawning migration, the size and age composi- 

 tion, and the size and number of eggs produced. 



Cope has accepted these differences as 

 evidences of the existence of at least five dis- 

 tinct subpopulations of cutthroat trout in Yellow- 

 stone Lake. Since there is no experimental 

 evidence, it might equally well be argued that 

 the only characteristic which is heritable is that 

 of homing to the natal stream (whichever it may 

 be) and that all the other differences are pheno- 

 rypic differences imposed by the natal stream . 

 Even if this perhaps extreme view should prove 

 to be true, nevertheless as long as the fish do 

 show persistent homing there is an effective 

 method of isolation operating. If the fish of dif- 

 ferent streams have not yet genetically diverged, 

 it may be expected that they will eventually do so 

 under the selection of the differing environments 

 inhabited during at least part of their lives. The 



nature of this problem and of the natural habitat 

 of the fish suggests interesting experiments that 

 could be done in situ. 



Raney has continued his studies of the 

 striped bass, considering here the possible ex- 

 istence of subpopulations in Chesapeake Bay and 

 its tributaries. He concludes that there are 

 three subpopulations, one in the James River, 

 one in the York and RappahannocK Rivers and 

 one in the bay and tributaries north of the Rap- 

 pahannock. He also believes the James 

 subpopulation to be differentiated at a higher 

 level than the others. 



His study is based on meristic characters, 

 including dorsal spines and dorsal and anal rays. 

 The recognition of subpopulations on this kind of 

 evidence requires that assumptions be made 

 about the genetic nature of the observed differ- 

 ences. Raney has made these assumptions, yet 

 he clearly recognizes that they are assumptions 

 and has gone to some length to examine their 

 validity, or at least to remove some of the ex- 

 trinsic variables. To this end he has examined 

 only fish of the year which were collected at or 

 near their place of origin, he has considered 

 characters by year-classes and has followed 

 them through several year -classes, he has con- 

 sidered characters over wide geographic areas 

 and he has drawn on supplementary sources of 

 information such as the results of tagging experi- 

 ments. Definitive and final answers to this 

 problem must, as Raney suggests, await experi- 

 mental evidence . 



In reviewing the subpopulation problem 

 in the Pacific sardine, I have attempted to gather 

 all available evidence in an effort to clearly de- 

 fine the problem and to suggest methods of 

 attacking it. The single most important body of 

 information in this respect is the data on the 

 space -time distribution of spawning and on the 

 absolute magnitude of each space -time group (of 

 which there are at least four). 



The crucial problem is, I believe, to de- 

 termine the amount of interchange between these 

 space -time groups. This obviously imposes the 

 requirement of sampling on the spawning grounds . 

 The more direct the attack on this problem the 

 more definitive will be the results. A tagging 



