575 
Enhancing diet analyses of piscivorous 
fishes in the Northwest Atlantic 
through identification and 
reconstruction of original prey 
sizes from ingested remains 
Frederick S. Scharf 
Department of Forestry and Wildlife Management 
University of Massachusetts, Amherst, Massachusetts 0 1 003 
Present address: Texas Parks and Wildlife Department, Coastal Fisheries Division 
Seabrook Marine Laboratory 
RO. Box 8, Seabrook, Texas 77586 
E-mail address: fred.scharf@tpwd.state.tx, us 
Richard M. Yetter 
Department of Forestry and Wildlife Management 
University of Massachusetts, Amherst, Massachusetts 0 1 003 
Adam R Summers 
Organlsmic and Evolutionary Biology 
University of Massachusetts, Amherst, Massachusetts 0 1 003 
Francis Juanes 
Department of Forestry and Wildlife Management 
and 
Graduate Program in Organismic and Evolutionary Biology 
University of Massachusetts, Amherst, Massachusetts 0 1 003 
Abstract.— Biological interactions 
among species can play a dominant role 
in structuring marine fish communi- 
ties. Specifically, predation may repre- 
sent a significant source of mortality for 
larval and juvenile fishes. Analysis of 
predator diet requires accurate infor- 
mation on the identity as well as the 
sizes of prey consumed. In examina- 
tions of stomach contents of piscivorous 
fishes, the condition of recovered prey 
items varies substantially not only in 
the large range of digestive states en- 
countered but also in the occurrence of 
partially consumed fishes. To estimate 
the original sizes of well-digested and 
partially consumed prey fishes we con- 
structed a series of predictive equations 
relating total length, fork length, and 
weight of fish to seven morphometric 
measurements including dorsoventral 
body depth, eye diameter, caudal pe- 
duncle depth, pectoral-fin length, 
opercle length, cleithrum length, and 
dentary length for ten common prey 
fishes in the Northwest Atlantic. All 
relationships were highly significant, 
with coefficients of determination typi- 
cally exceeding 0.90 and mean percent 
prediction errors less than 10%, indi- 
cating that reliable original size esti- 
mates are obtainable from incomplete 
fish remains. To aid in field-based iden- 
tification of prey fishes, we extracted 
and examined opercles, cleithra, and 
dentaries from each fish. Careful ex- 
amination of bones revealed prominent 
diagnostic characteristics with clear 
differences among family taxa, demon- 
strating their potential utility as iden- 
tification tools. Used collectively, the 
predictive equations and the diagnos- 
tic features of the bones should allow 
for inclusion in diet analyses of prey 
items previously designated as uniden- 
tifiable or unmeasurable, and thus in- 
crease the amount of dietary information 
obtainable from stomach contents analy- 
ses of Northwest Atlantic piscivores. 
Manuscript accepted 22 October 1997. 
Fishery Bulletin 96:575-588 (1998). 
Biological interactions among spe- 
cies can significantly affect the dy- 
namics of marine fish populations 
(Overholtz et al., 1991; Rothschild, 
1991). Specifically, predation by pi- 
scivorous fishes is recognized as an 
important mechanism in structur- 
ing fish communities (Hackney, 
1979; Lyons and Magnuson, 1987; 
Tonn et al., 1992), particularly in 
the role it plays to potentially regu- 
late natural mortality rates and re- 
cruitment of larval and juvenile 
fishes (Sissenwine, 1984; Houde, 
1987; Bailey and Houde, 1989). In 
order to estimate consumption rates 
of key predators, details are needed 
on the types and sizes of prey fishes 
consumed across spatial and tem- 
poral scales. However, because of 
the substantial variability in the 
condition of recovered prey items, 
the collection of stomach contents 
data in the field is often limited to 
only those items that are readily 
identified and measured, resulting 
in the loss of potentially important 
diet information. 
Identification of piscine prey in- 
gested by fish, avian, and mamma- 
lian predators has frequently in- 
volved the use of diagnostic bones 
including vertebrae (Pikhu and 
Pikhu, 1970; Feltham and Marquiss, 
1989; Carss and Elston, 1996), pha- 
ryngeal arches (McIntyre and Ward, 
1986; Raven, 1986), opercular bones 
(Newsome, 1977), the axial skeleton 
