256 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



Table 7. — Mtdtiple regression of estimated population in 

 year i (F) on escapement inyeari—B (X{), escapement in 

 year i—4 (Xi), an d escapement in year i—1 (X3), with 

 population expected from regression (Y) and deviation from 

 expected population (Y— Y), in thousands of fish 



F=48.10655+0.52543.Yi+0.3129S.Yi+0.78338Ar3 

 'YXvXiXi=0.O3, P~0.25 

 'YXrX,Xi=0.2Z,3, P~0.4S 

 ' YXy Xi Xj= 0.818, P<0.01 



ANALYSIS OF VARIANCE 



Source: Degrees of freedom Sum of squares Mean square 



Total 12 1,040,752 



Mean 1 963,333 



Regression 3 64,490 21,497 



Error 8 12,929 1,616 



21 497 



J?==rrnr= 13.303, p<o.oi 



1, 616 

 JJ* = 0.833; R = 0.912, P<0.01 



ENVIRONMENTAL CHANGES 



The deviations from regression in table 7 reflect 

 the effects of other factors on the total population 

 of shad. Water temperatures, stream flow, and 

 dissolved oxygen data have been carefully studied, 

 but none of these variables exhibits changes or 

 trends that would account for the changes in total 

 population or the deviations from the predicted 

 populations. However, until representative 

 samples of scales are taken from the commercial 

 fishery for a series of j T ears and the year-class sizes 

 are estimated, it is not possible to determine fully 

 the effects of these variables during the spawning 

 season on the size of the year class. 



EXTRANEOUS MORTALITY 



One factor which may affect the deviations from 

 regression is the mortality which occurs among 

 the adult shad between fishing seasons. This 

 mortality could be from natural causes during or 

 following spawning in the river, natural causes at 

 sea, or from fishing outside the river. The 1951 

 tagging experiment and similar studies by Hollis 

 (unpublished manuscript), and Westman and 

 Bevelander (correspondence with Dr. Westman), 

 show that Connecticut River shad are taken by 



fishermen off the coast of Maine, in Sandy Hook, 

 Raritan and Lower New York Bays, off Long 

 Island, and along the New Jersey coast. 



The mortality that occurs among the adult 

 shad between fishing seasons will be referred to as 

 extraneous mortality in this paper. If scale 

 samples representative of the commercial catch 

 with respect to gear and sex were available for a 

 number of years, the extraneous mortality rates 

 could be easily estimated. Since most of the 

 scale samples have been collected from the highly 

 selective gill-net fishery in the past, it is necessary 

 to use scale data from the sport fishery at Enfield 

 Dam. 



The Connecticut State Board of Fisheries and 

 Game has collected scale samples at the Enfield 

 Dam sport fishery each year since 1944. These 

 samples represent from 10 to 33 percent of the 

 anglers' catches and are taken in the same manner 

 each year. Before these scale samples can be 

 used for mortality determinations, the assump- 

 tion must be made that the fish caught at Enfield 

 Dam are representative of the fish escaping the 

 commercial fishery. An examination of the 

 lengths and ages of fish taken by anglers at the 

 dam reveals proportionately more small 3-year- 

 old males in the sport-fishery samples than in the 

 commercial-fishery samples. The apparent dis- 

 proportionate number of small males may be due 

 to the fact that some of these fish are too small 

 to be caught by commercial gear, or may result 

 from a lack of samples from the haul-seine 

 fishery which generally takes more small male 

 shad than does the gfll-net fishery. Only when 

 more representative samples from the commercial 

 catch are obtained can the discrepancy be ex- 

 plained. In the meantime, I do not believe one 

 can safely assume that the small males in the 

 sport-fishery samples are representative of the 

 fish escaping the commercial fishery each year. 



Even though the relative abundance of all 

 other fish is assumed to be the same in the Enfield 

 Dam samples as in the group of fish escaping the 

 fishery, the effects of disproportionate numbers of 

 3-year-old males in the Enfield samples can cause 

 serious errors in mortality-rate determinations. 

 For this reason, in estimating mortality rates I 

 have selected 2 successive years when the 3-year- 

 old males were in the same proportionate abun- 

 dance in the sport-fishery samples. 



