586 
Fishery Bulletin 96(3), 1 998 
Discussion 
Each of the morphometric measurements, including 
those from diagnostic bones, were significantly re- 
lated to total length, fork length, and weight. Mea- 
surements taken from diagnostic bones, especially 
cleithrum length, appear to be highly reliable pre- 
dictors of original size for prey fish species in the 
Northwest Atlantic. Internal hard parts of fishes have 
historically proven accurate for estimating original 
sizes of prey fishes recovered from the stomachs of 
several freshwater piscivores (Knight et al., 1984; 
Trippel and Beamish, 1987; Hansel et al., 1988). Our 
results suggest that examination of diagnostic bones 
should also benefit diet analyses of marine piscivores. 
The ability to estimate original length and weight 
of common prey fishes from a suite of morphometric 
measurements should result in a considerable in- 
crease in the amount of size-specific diet informa- 
tion for several predatory fishes in the Northwest 
Atlantic. Accurate information on the sizes of prey 
consumed across temporal and spatial scales is criti- 
cal in order to calculate predator consumption rates 
and to determine predator impact on prey popula- 
tions. Moreover, many piscivorous fishes feed selec- 
tively on specific sizes of prey (Juanes, 1994), sug- 
gesting that predatory impact may be greatest on a 
small range of prey sizes. Therefore, knowledge of 
prey sizes consumed is necessary in order to deter- 
mine the extent of size-selective feeding patterns in 
Northwest Atlantic piscivores and their role in struc- 
turing marine fish populations. Finally, several au- 
thors have demonstrated the importance of body size 
in determining the outcome of predator-prey inter- 
actions among fishes (Werner and Gilliam, 1984, 
Miller et al., 1988, Stein et al., 1988). More complete 
information on predator-prey size relationships in the 
Northwest Atlantic may improve our understanding 
of the effects of body size on prey vulnerabilities to 
predation during various life history stages. 
Clearly, the estimation of original prey sizes from 
diagnostic bone dimensions is not as susceptible to 
measurement error as estimates from external mor- 
phological features, where measurements are usu- 
ally associated with soft tissues. Moreover, external 
morphology can often be altered during digestive 
processes, which may cause some or all external fea- 
tures to yield unreliable measurements. However, for 
recently consumed prey fishes, external morphologi- 
cal measurements may be highly reliable estimators 
of original prey size and represent a suitable alter- 
native to diagnostic bones. For example, the linear 
relations of both eye diameter and caudal peduncle 
depth with total length were found to be reliable for 
reconstructing original prey lengths of six prey spe- 
cies consumed by juvenile bluefish (Scharf et al., 
1997). Similarly, Serafy et al. (1996) found the lin- 
ear relation between eye diameter and total length 
of red drum to be particularly useful in estimating 
the size of angled fish. In addition to cleithrum 
length, maximum body depth and caudal peduncle 
depth were consistent predictor variables for esti- 
mating original weight of prey fishes examined in 
this paper. Differences in fish condition factor may 
not be closely correlated with differences in other 
morphological measurements. Therefore, indices of 
fish girth, such as body depth and caudal peduncle 
depth, should be reliable indicators of condition 
factor and should be most closely related to body 
weight. Further, the reconstruction of original prey 
sizes from external morphological measurements has 
the advantage of being comparatively less labor in- 
tensive than dissection and removal of internal hard 
parts. 
The reconstruction of original prey sizes from di- 
agnostic bones or external morphological measure- 
ments does, however, have some important limita- 
tions. The potential effects of preservatives on bone 
dimensions need to be considered if stomach contents 
are stored in a chemical preservative. The fish used 
in this study were not exposed to preservatives but 
were frozen and thawed before measured. The use of 
boiling water to aid in the removal of soft tissue may 
cause shrinkage and deformation of extracted bones 
if sufficient time elapses between the boiling process 
and the completion of bone measurements. However, 
bone measurements for our analyses were completed 
immediately after dissection and removal. In addi- 
tion to the potential effects of preservation, previous 
studies indicate that estimation of prey sizes from 
partial remains will likely underestimate the pro- 
portion of small prey fish in the diet, because bones 
and external features of larger prey fish should be 
more resistant to digestion and be recovered from 
piscivore stomachs more frequently (Trippel and 
Beamish, 1987; Hansel et al., 1988). Lastly, allomet- 
ric relationships may not remain constant with 
changes in fish size. Therefore, regression equations 
should not be applied to fish outside the size range 
used for their generation. 
The results of our study indicate that, for the fami- 
lies of fishes examined, the opercle, cleithrum, and 
dentary represent diagnostic tools that can poten- 
tially aid in the identification of prey fishes recov- 
ered from the stomachs of Northwest Atlantic 
piscivores. The diagnostic features unique to each of 
the three bones appear to perform equally well in 
distinguishing family taxa from one another. How- 
ever, within the families Gadidae and Clupeidae, 
cleithra and dentaries are better suited to illustrate 
