282 
Fishery Bulletin 114(3) 
Figure 4 
Length-frequency plot for the 3 myctophid species captured from sur- 
face waters in the eastern tropical Pacific during 2006. Sizes reflect 
our unique specimen collection and are not representative of the en- 
tire population because only surface-migrants were targeted. 
and 89°W, and nearshore off Ecuador (Fig. 7). Net 
samples revealed copepods were the most numerically 
abundant zooplankton (80.5%, compared with 63.9 and 
75.1% in the offshore and nearshore regions) in this 
region (Table 3). This intermediate region was charac- 
terized by moderate values of MLD, SSS, and SCHL 
(Fig. 3), and it appears to represent a transition zone 
between the offshore and nearshore regions. 
Terminal node 5 The myctophids sampled at 8 near- 
shore stations near the Panama Bight, east of 87°W 
and north of 3°N, (Fig. 7) consumed primarily euphau- 
siids (mean MNi. 45.0%). SSS in this region was low 
(<32.86 [Fig. 3]), the MLD was shallow (mean MLD: 
22.9 m), and SCHL concentrations (0.22 mg/m^) were 
greater than those at stations identified within the 
other regions. 
Interspecific patterns 
Collectively, the predominant prey of these myctophids 
came from 4 groups, copepods (MiV'j=37.7%), ostracods 
(34.9%), euphausiids (13.7%), and amphipods (9.1%), 
which accounted for more than 95% of the diet, by 
number. The remaining 5% comprised mollusks (ptero- 
pods and heteropods, 2.6%), larval fishes, decapods, fish 
eggs, one squid paralarva, and one terres- 
trial insect (Table 2). 
Interspecific dietary differences were 
apparent and might have been more de- 
finitive if the prey were identified at a 
lower taxonomic level. Previous research 
has, for example, indicated that these 
species are selective feeders (Van Noord 
et al., 2013b). We further assessed these 
previous findings by including a broad 
suite of predictor variables and found 
that “myctophid species” ranked relatively 
low in explaining diet patterns across the 
ETR Myctophum nitidulum fed on cope- 
pods (42.7%) and ostracods (41.5% [Table 
2]). Symbolophorus reversus fed primarily 
on copepods (32.5%), euphausiids (29.6%), 
and ostracods (24.4%). Gonichthys tenui- 
culus took prey from only 4 groups, pri- 
marily ostracods (34.6%) and amphipods 
(27.3% [Table 2]). 
Distribution patterns differed some- 
what among the 3 myctophids. Figure 
2 displays spatial trends in abundance; 
greater numbers of S. reversus and G. 
tenuiculus occurred in the nearshore and 
intermediate areas, respectively. The indi- 
viduals in this study, however, represent 
subsamples of the captured myctophids, 
and no quantitative distribution analysis 
was possible. However, representatives 
of each species were captured across the 
entire sampling region, resulting in ad- 
equate distributional overlap, but the tree 
analysis did not indicate that myctophid species are 
an important variable in characterizing the diet of the 
fishes in this study. 
Discussion 
We used a classification tree to examine the influence 
of spatial, biological, and oceanographic predictors on 
diet and found that feeding by the collection of surface- 
migrating myctophids in this study was controlled by 
prey distribution and resource-driven processes, such 
as mixed-layer depth, productivity, and sea surface sa- 
linity, whereas the influence of dietary resource par- 
titioning was a minor controlling factor. These myc- 
topMds shared a similar diet, consisting primarily 
of copepods, ostracods, euphausiids, and amphipods. 
Diet of all 3 species changed geographically, and with 
oceanographic conditions and zooplankton prey compo- 
sition. Myctophids consumed ostracods offshore where 
the mixed layer depth was deep and ostracods were 
more abundant in the prey community, euphausiids 
nearshore where the MLD was shallow, and copepods 
at intermediate stations between those stations where 
they were most abundant. Understanding myctophid 
feeding behavior can provide insight into how these 
