Lazzari: Dynamics of larval fish abundance in Penobscot Bay, Maine 
89 
Table 5 
Product-moment correlations between principal component scores and the angularly transformed relative abundances of larval 
fishes. Underlined values significant at P < 0.05, adjusted for multiple comparisons such that P < 0.004 for a significant correlation. 
Rotated factor 1 explained 36% of the total variance, factor 2 explained 20%, and factor 3 explained 15%’ in 1997, and rotated factor 
1 explained 35% of the total variance, factor 2 explained 28%, and factor 3 explained 11% in 1998. 
Taxon 
1997 
1998 
Factor 1 
Factor 2 
Factor 3 
Factor 1 
Factor 2 
Factor 3 
Clupea harengus 
0.05 
0.51 
-0.04 
Mallotus villosus 
0.02 
-0.16 
0.08 
Enchelyopus cimbrius 
0.44 
-0.14 
0.11 
Lumpenus lampraetiformis 
-0.07 
0.34 
0.02 
Ulvaria subbifurcata 
0.97 
-0.22 
0.10 
-0.11 
-0.11 
0.05 
Pholis gunnellus 
-0.22 
0.46 
-0.41 
0.67 
0.40 
0.58 
Cryptacanthodes maculatus 
-0.14 
0.23 
-0.39 
0.63 
-0.01 
0.11 
Ammodytes sp. 
-0.17 
0.98 
0.05 
0.06 
0.92 
0.31 
Hemitripterus americanus 
-0.15 
0.37 
-0.27 
0.13 
0.02 
0.73 
Myoxocephalus aenaeus 
-0.12 
0.31 
-0.13 
0.41 
0.58 
0.35 
M. octodecimspinosus 
-0.14 
0.07 
-0.41 
0.80 
0.58 
-0.07 
M. seor-pius 
0.56 
0.66 
0.17 
Liparis atlanticus 
0.39 
-0.13 
0.55 
-0.42 
0.16 
-0.05 
L. coheni 
0.67 
0.10 
0.40 
Pleuronectes americanus 
0.06 
-0.23 
0.89 
-0.20 
-0.09 
-0.13 
Hippoglossoides platessoides 
0.48 
-0.07 
0.27 
□ 
Usu 
□ 
Pgu 
□ 
Pel 
a 
Pam 
□ 
Msp 
□ 
Lsp 
□ 
Cma 
■ 
Asp 
— i 
Mar Apr 
1998 
Figure 3 
Percent density (number/100 m 3 ) by month of the dominant larval 
fishes collected with 1.0-m plankton net in Penobscot Bay, Maine, 
from April to June 1997 and from March to April, 1998. Asp -Ammo- 
dytes sp., Cma = Cryptaeanthod.es maculatus. Lsp = Liparis spp., 
Msp = Myoxocephalus spp., Pam = Pleuronectes americanus. Pel = 
pelagic spp., Pgu = Pholis gunnellus, Usu = Ulvaria subbiffurcata. 
brius, H. plattesoides, U. subbifurcata ) were uni- 
formly positively correlated with factor 1. Species 
that were most abundant earlier in the year (Am- 
modytes sp., H. americanus, M. aenaeus, P. gunnel- 
lus) were positively correlated with factor 2. Values 
for several species (C. maculatus, M. octodecimspi- 
nosus, and P. americanus) were not significantly 
correlated to either factor 1 or 2, but were cor- 
related to factor 3. In 1998, MANOVA and post 
hoc comparisons of mean factor scores indicated 
a significant temporal effect on factors 1 and 2 
(Wilk’s lambda=0.341, P<0.0001). Values for taxa 
that were more abundant in March ( Ammodytes 
sp., C. maculatus, C. harengus, L. coheni, M. aenae- 
us, M. octodecimspinosus, M. scorpius ) were uni- 
formly positively correlated with factor 1 or factor 
2, or with both factors. Values for H. americanus 
were not significantly correlated to either factor 1 
or 2, but were highly correlated to factor 3 as their 
relative abundances increased in April. 
Assemblage composition was similar among 
sampling locations as indicated by MANOVA of 
mean factor scores in 1997 and 1998. In 1997, 
MANOVA (Wilk’s lambda=0.985, P>0.10) indicated 
that means for lower bay, midbay and upper bay 
sites were not significantly different from each oth- 
er for any factor. Similarly, MANOVA (Wilk’s lamb- 
da=0.831, P>0.08) indicated that means for lower estuary 
and bay sites were not significantly different for any factor 
in 1998. 
CCA revealed significant overall association between fac- 
tor scores from the PCA of the relative fish abundances 
and environmental variables in 1997 (Wilk’s lambda=0.17, 
