FISHERY BULLETIN: VOL. 74, NO. 1 



determination (Steel and Torrie 1960) was used to 

 estimate the amount of variation in y associated 

 with variation in x. Residuals were used to 

 examine the data for differences due to categories 

 of classification such as year of collection. Size 

 ranges are given within which regressions were 

 linear. 



When first referred to, locations are followed in 

 parentheses by their approximate distances in 

 kilometers upstream from Marcus Hook, Pa., 

 which is situated about 90 km downstream from 

 the fall line at Trenton, N.J. and near the transi- 

 tion between brackish and fresh water. 



RESULTS AND DISCUSSION 



Total and Fork Length Conversion 



The relationship between total length (TL) and 

 fork length (FL) for 590 young and adult fish was 

 linear, and 99.96*7^ of the variation in one mea- 

 surement was explained by variation in the other. 

 Regression equations were FL = 1.28 + 0.88 TL 

 and TL = 1.00 + 1.13 FL. Extreme deviations 

 from regression were about ±7.6 mm for adults 

 and less for young. The slope of the regression of 

 fork length on total length coincides with La 

 Pointe's (1958) factor of 0.894 to convert total 

 length to fork length. 



Total Weight- Length Relationships 

 Prior to Spawning 



The relationships between total weight (TW) 

 and length determined for fish captured at Lam- 

 bertville were TW = 1,106.77 + 8.09 (FL - 427.98) 

 for 268 males and TW = 1,737.26 + 11.54 (FL - 

 476.71) for 244 females. About 81^f (males) and 

 78% (females) of the variation in total weight 

 was associated with variation in length. Valid 

 ranges for linear interpolation were about 330- 

 520 mm for males and 410-550 mm for females. 



The observed arithmetic mean weights with 

 95% confidence limits were 1,107 ± 36 g for males 

 and 1,737 ± 45 g for females. The smallest males 

 were 272 and 680 g and the smallest female was 

 1,089 g. The heaviest male and female fish were 

 1,905 and 2,585 g, respectively. 



Somatic Weight-Length Relationships 

 Prior to Spawning 



The relationships between log somatic weight 

 152 



(SW) and length determined for 85 males and 130 

 females captured at Lambertville in 1964 and 

 1965 were logio SW - 3.0047 + 0.0036 (FL - 

 428.20) for males and logio SW = 3.1807 + 0.0029 

 (FL - 480.73) for females. About 91% (males) and 

 81% (females) of the variation in log somatic 

 weight was associated with variation in length. 

 Valid ranges for linear interpolation were about 

 360-500 mm for males and 410-540 mm for 

 females. Mean somatic weights with 95% 

 confidence limits were 1,011 ± 56 g for males and 

 1,516 ± 38 g for females. 



Gonad Weight- Length Relationships 

 Prior to Spawning 



The relationships between log total gonad 

 weight (TGW) and length determined for 267 

 males and 244 females captured at Lambertville 

 were logio TGW = 1.8633 + 0.0033 (FL - 428.43) 

 for males and logio TGW = 2.3892 + 0.0024 (FL 

 - 476.93) for females. Valid ranges for linear in- 

 terpolation were about 330-520 mm for males and 

 410-550 mm for females. About 45% (males) and 

 26% (females) of the variation in log total gonad 

 weight was associated with length variation. 

 Much variation in gonad weight, especially for 

 females, is not explained by the regression equa- 

 tions. Much gonad development occurs during the 

 spawning run (Chittenden 1969), and residual 

 plots suggested that gonad weights were heavier 

 in 1963 than in 1964. These factors account for 

 some unexplained variation in gonad weight. 



Mean total gonad weights with 95% confidence 

 limits were 73 ± 7 g for males and 245 ± 22 g for 

 females. 



Duration of the Freshwater Residence 



Most fish begin to return seaward by about late 

 June. I observed hundreds of adults near Han- 

 cock, N.Y. (403) until 17 June 1964, but very few 

 were present on 14 July. Most fish had died or 

 migrated seawards during the interim period. De- 

 laware River shad runs begin in early April at 

 Lambertville and the peak occurs about 1 May, 

 depending upon the degree of pollution near 

 Philadelphia (34) (Chittenden 1969). This 

 suggests most fish probably spend a maximum of 

 2 mo in fi^esh water before returning seaward, in 

 agreement wdth Bean's (1892, 1903) observations. 



Many fish remain near the spawning grounds 

 well into summer. The Tri-State Surveys cap- 



