Scharf et al.: Diet analysis of piscivorous fishes 
583 
unique among the fishes examined here (Fig. 2, E 
and F). The shape of the anterior process differs be- 
tween the two genera of Clupeidae. In alewife, the 
anterior process is sickle-shaped, whereas in Atlan- 
tic herring, the anterior process is more symmetri- 
cal, resembling a mushroom. Additionally, the tip of 
the ventral process is distinctly sharper in the 
cleithrum of alewife than in the cleithrum of Atlan- 
tic herring. Atlantic mackerel cleithra are gently 
curved and relatively thin and one posterior process 
is located in the extreme dorsal region of the bone 
(Fig. 2G). The dorsal process of the Atlantic mack- 
erel cleithrum is relatively small and rounded, 
whereas the ventral process is broad and does not 
end sharply as in the cleithra of the gadids. The 
cleithra of butterfish are only slightly curved along 
the dorsoventral axis and contain a broad, rounded 
posterior process. Butterfish cleithra can be distin- 
guished among the families examined here by the 
presence of a distinctly sharp, thin dorsal process 
(Fig. 2H). The cleithra of sand lance have a general 
Table 4 
Independent variables included in stepwise multiple re- 
gression models estimating original fish length. Abbrevia- 
tions for independent variables are those given in Tables 2 
and 3. * indicates that forward and backward models were 
not identical. 
Variables 
Variables 
included 
included 
in forward 
in backward 
Species 
stepwise model 
stepwise model 
Alewife 
CP, PF, OP CL 
CP, PF, OP, CL 
Blueback herring 
PP CL 
PF, CL 
Atlantic herring 
BD, PP OP, DN 
BD, PF, OP, DN 
Atlantic mackerel* 
CL, DN 
CL 
Butterfish* 
BD, CP, PF, CL 
ED, CP, PF, CL 
Sand Lance 
CP, DN 
CP, DN 
Red hake 
BD, ED, CL 
BD, ED, CL 
Silver hake* 
OP, CL, DN 
CL 
Haddock 
ED, CP, PF, CL 
ED, CP, PF, CL 
Atlantic cod 
BD, PF, CL, DN 
BD, PF, CL, DN 
Table 5 
Least squares regression equations relating measurements (in millimeters) of body depth (BD), eye diameter (ED), caudal pe- 
duncle depth (CP), pectoral fin length (PF), opercle length (OP), cleithrum length (CL), and dentary length (DN) to total weight 
(W) in grams for ten prey species in the Northwest Atlantic. s b = standard error of the regression coefficient; r 2 = coefficient of 
determination; %PE = mean percent prediction error; n = number of fish measured; Range = size range in total length. 
Species 
Equation 
s b 
r 2 
%PE 
n 
Range 
Alewife 
W = 0.73 xlO- 3 BD 2 940 
0.041 
0.97 
10.74 
137 
73-282 
W= 2.62x10 - 3 £D 4 305 
0.133 
0.89 
24.00 
137 
73-282 
W= 10.18 xlO“ 3 CP 3 194 
0.037 
0.98 
8.87 
137 
73-282 
W= 1.25 xlO-W 5 229 
0.053 
0.98 
11.31 
66 
73-282 
W = 2.95 xlO- 3 OP 3 331 
0.052 
0.98 
11.73 
66 
73-282 
W = 0.81 xlO _3 CL 3 383 
0.051 
0.99 
10.65 
66 
73-282 
W= 1.05x10 -3ZW 3 - 739 
0.055 
0.99 
10.66 
66 
73-282 
Blueback herring 
W = 0.64 xl0~ 3 BD 3 066 
0.103 
0.96 
6.83 
38 
83-134 
W = 6.68 xlO _3 EZ) 4 049 
0.340 
0.80 
15.73 
38 
83-134 
W = 13.80 xlO- 3 CP 3 061 
0.154 
0.92 
9.48 
38 
83-134 
W = 3.94 xlO-W 2 903 
0.158 
0.90 
9.88 
38 
83-134 
W = 10.40 xlO - 3 OP 2 851 
0.160 
0.90 
9.84 
38 
83-134 
W= 2.38 xl0“ 3 CL 3 027 
0.139 
0.93 
8.15 
38 
83-134 
W = 3.00 xlO -3DN 3 383 
0.201 
0.89 
9.96 
38 
83-134 
Atlantic herring 
W = 2.96 xlO" 3 #/) 2 734 
0.033 
0.99 
8.64 
84 
97-295 
W= 7.14xlO- 3 £D 4012 
0.092 
0.96 
16.58 
84 
97-295 
W = 7.96 xlO“ 3 CP 3 438 
0.060 
0.98 
11.81 
84 
97-295 
W = 3.76 xlO - 3 P£2 935 
0.065 
0.97 
12.06 
84 
97-295 
W = 10.53 xlO - 3 OP 3 030 
0.063 
0.97 
10.01 
71 
97-295 
W = 2.20 xl0 _3 CL 3 193 
0.069 
0.97 
10.01 
71 
97-295 
W = 3.52 xl0- 3 Z)AF 210 
0.084 
0.96 
12.87 
71 
97-295 
Atlantic mackerel 
W = 2.79 xl0 _3 BZ) 2 884 
0.066 
0.97 
11.73 
56 
158-330 
W = 2.20 xlO- 3 £Z) 4 - 762 
0.377 
0.75 
39.17 
56 
158-330 
W = 247.52 xl0 _3 CP 3 373 
0.061 
0.98 
9.12 
56 
158-330 
W = 1.49 xlO-W 3 399 
0.078 
0.97 
12.07 
56 
158-330 
W = 3.90 xlO - 3 OP 3 318 
0.049 
0.99 
6.77 
44 
158-330 
continued 
