272 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



Table 2. — Summary of areas from which meristic data were 

 collected on adult American shad 



[The figures represent the number of shad in each sample. Data were not 

 complete for all meristic characters in all samples] 



This paper will not include a complete analysis of 

 the data available. It is hoped that this prelim- 

 inary analysis will show that a complete analysis 

 would be warranted, and that further research 

 along these lines would be fruitful. No analysis 

 of the information on the juveniles has been 

 attempted, hence there is still much to be learned 

 by combining this with the data on the adults. 



REPRESENTATIVENESS OF THE SAMPLES 



The need for separation of races of fish is 

 apparent. However, before races can be sepa- 

 rated, it should be established that they are 

 present. Tagging experiments have been offered 

 as the best evidence supporting the theory of a 

 separate race of shad in each Atlantic coast river. 

 In numerous tagging programs carried out by the 

 Fish and Wildlife Service, few tagged shad have 

 been recaptured in rivers other than the one in 

 which they were tagged. No shad tagged on the 

 spawning ground of one river system has ever 

 been recaptured on the spawning ground of 

 another river system. The operation of the 

 homing instinct may not be 100 percent for shad, 

 but examination of tag returns indicates that the 

 percentage is very high. 



If there is a race of shad in each river, that is, 

 a group of fish and their offspring which return to 

 the same spawning area year after year, the fish 

 within a river should be more like one another 

 than to the shad from other rivers. This could 

 be expected because of environmental differences 

 or genetic isolation. Conversely, if the spawning 

 ground of each fish is determined by a completely 

 random process, a single homogeneous shad popu- 

 lation would be expected. Therefore, if consist- 

 ent differences between shad in the several rivers 



can be found for some measurable characters, this 

 can be used as further evidence to support the 

 race theory. 



Before proceeding with an attempt to verify 

 this theory, one assumption should be investi- 

 gated. If the available data are to be used to 

 establish differences between rivers, it is essential 

 that the samples be representative of the various 

 populations. It is impossible to assume that 

 these are random samples, because the shad fisher- 

 men know that they can control the size of fish in 

 their catches by changing the mesh size of their 

 nets. They know that if they fish a 5%-inch 

 stretched -mesh gill net, they will catch large-roe 

 shad, and if they use a 4%- to 5-inch stretched -mesh 

 net, they will catch proportionately more small 

 shad of both sexes. This selectivity occurs witli 

 drift, anchor, and stake gill nets. Pound nets 

 and haul seines may be much less selective, but 

 if the gill nets are catching large fish and per- 

 mitting smaller ones to escape, the population 

 being sampled by the haul seines is not the total 

 population of the river, but the total population 

 minus the fish removed by gill nets. The result, 

 of course, would be an excess of smaller fish in the 

 haul seine samples. 



Since it is known that some fishing gear tends 

 to select fish by length (and all correlated meas- 

 urements such as depth and thickness), it cannot 

 be assumed that the samples are random. How- 

 ever, this selectivity may take place only in the 

 size of the fish in the samples and not in some of 

 the other characteristics. If the number of rays 

 in the pectoral fin is being investigated, the 

 samples may be representative of this character 

 even if there is selectivity of size. This would be 

 true of any character which is not correlated with 

 length. Therefore, the characters were tested 

 for a correlation with length and if none was 

 found, the samples were considered representative. 



In examining the catches of shad in the St. 

 Johns River of Florida and the Connecticut 

 River, it is apparent that the shad in the Connec- 

 ticut are larger. Some of this can probably be 

 explained by the difference in the age distributions 

 of the two populations. Can these age distribu- 

 tions be used in separating races of shad? The 

 author feels that they cannot be used, since they 

 will fluctuate from year to year with changes in 

 fishing effort and catches. Age and length are 

 correlated, and this is another reason for excluding 



