SOME USES OF STATISTICAL ANALYSIS IN CLASSIFYING RACES OF 



AMERICAN SHAD (Alosa sapidissima) 



By Donald R. Hill, Fishery Research Biologist, Bureau of Commercial Fisheries 



The commercial catch of American shad (Alosa 

 sapidissima) has declined since the beginning 

 of the twentieth century. In 1950 a study of 

 this species was undertaken by the Fish and 

 Wildlife Service acting as the primary research 

 agency of the Atlantic States Marine Fisheries 

 Commission. The objectives of the investiga- 

 tion were to determine the causes for the decline 

 in shad abundance, determine conditions favoring 

 recovery, and provide basic information so that 

 the fishery can be managed to obtain optimum 

 yields. 



Most American shad landed on the Atlantic 

 coast are captured in rivers; however, pound 

 nets fished off the coasts of New York and New 

 Jersey take large numbers of them each spring. 

 The racial origin of these fish must be known 

 for the intelligent management of the species. 

 In this study, use is made of meristic counts 

 for shad to separate two races or populations. 



Meristic data were collected under the super- 

 vision of Louella E. Cable of the U. S. Fish and 

 Wildlife Service. Grateful acknowledgement is 

 made to G. B. Talbot. Chief. Middle Atlantic 

 Fishery Investigations, for supplying these data 

 and reviewing the manuscript, to C. H. Walburg 

 in preparing the manuscript for publication, to 

 T. M. Widrig and D. D. Worlund for suggestions 

 concerning the estimation of relative abundance; 

 and to T. A. Bancroft of the Statistical Laboratory, 

 and K. D. Carlander of the Zoology Department, 

 Iowa State College, for numerous suggestions. 



STATEMENT OF THE PROBLEM 



Studies of the shad populations of the Hudson 

 (Talbot 1954) and Connecticut (Fredin 1954) 

 Rivers have shown the effect of fishing effort 

 on catches made in subsequent years. As a 

 result of analysis of catch-and-effort statistics 



N'ote.— Approved for publication December 4, 1957. Fishery Bulletin 147. 



and tagging experiments, the size of the runs in 

 previous years was determined for each river. 

 Table 1 shows the size of the catches for the two 

 rivers for 1938-51 and the calculated fishing 

 rates for each of these years. These catches 

 include only the fish taken in the rivers and not 

 fish caught in the ocean. 



Table 1. — Total shad catches and estimated fishing rates 

 for the Hudson and Connecticut Rivers from 1938 to 1951 



1 Catch for Connecticut River estimated by a factor of 3 pounds per fish. 



Through an analysis of data on scales of 6-year- 

 old shad from the Connecticut River, Fredin 

 (1954) found for shad an extraneous mortality of 

 about 40 percent occurring outside the river fish- 

 ery. This was nearly as great as the fishing mor- 

 tality in the river. He suggested that the pound 

 nets in the New York Bay area and along the 

 New Jersey coast could be the cause of some of 

 this mortality. The number of pound nets in 

 operation increased from 144 in 1946 to 180 in 

 1950. For these same years, the estimated popu- 

 lations (estimated by a regression analysis of 

 escapements in previous years) were higher than 

 the populations calculated from the catch-and- 

 effort data. Fredin stated that the increase in 

 pound-net effort may account for these deviations. 



269 



