90 
Fishery Bulletin 99(1 ) 
P<0.0001) and 1998 (Wilk’s lambda=0.21, PcO.OOOl). In 
1997, the first canonical correlation (Table 6) indicated 
that the relative abundance of larval fishes with high posi- 
tive loading on factor 1 (E. cimbrius, H. plattesoides, and 
U. subbifurcata ) exhibited a positive association with tem- 
perature. These taxa also exhibited a weaker negative as- 
sociation with plankton volume and salinity. Fish taxa that 
loaded positively on factor 2 were negatively associated 
with temperature and positively associated with plankton 
volume and salinity. These species included Ammodytes 
sp., H. americanus, M. aenaeus, and P gunnellus. Fish taxa 
that loaded positively on factor 3 (L. atlanticus and P. amer- 
icanus) were positively associated with temperature and 
were present in every month sampled. In 1998, the first 
canonical correlation (Table 6) indicated that the relative 
abundance of larval fishes with high positive loading on 
factor 1 (C. maculatus, L. coheni, and P. gunnellus) and 
factor 2 (Ammodytes sp., C. harengus, and M. aenaeus) all 
exhibited a negative association with temperature and a 
positive association with plankton volume. There was no 
correlation between factor 3 scores and environmental data 
for the first canonical correlation (Table 6). 
Discussion 
Species composition 
Ichthyoplankton typical of the Gulf of Maine and the 
Northwest Atlantic was observed in Penobscot Bay: lipa- 
rids, sculpins ( Myoxocephalus spp. ), Ammodytes sp., P. gun- 
nellus, U. subbifurcata , and P. americanus. The dominant 
species are similar to those previously reported for areas of 
the Maine coast, including the Sheepscot estuary (Graham 
and Boyar, 1965; Hauser, 1973; Shaw, 1981), the central 
Maine coast and estuaries (Chenoweth, 1973), the Dam- 
ariscotta River estuary (Lee, 1975; Laroche, 1980; 1982; 
Townsend, 1981; 1984), and Sullivan Harbor (Townsend, 
1984). Of the twenty-two species collected in the Damar- 
iscotta estuary and 21 species collected in Sullivan Harbor 
(Townsend, 1984), 17 and 18 of these species, respectively, 
occurred in Penobscot Bay. The dominant species in these 
two systems was P. gunnellus along with the three species 
of Myoxocephalus and a stichaeid, Lumpenus lampraetifor- 
mis (Townsend, 1984). Twenty-two kinds of boreal larvae 
with centers of abundance north of the mid-Atlantic coast 
were found in the Damariscotta and Sheepscot estuaries 
and along the central coast of Maine (Chenoweth, 1973), 
and 18 of these taxa were also found in Penobscot Bay. 
Pholis gunnellus, Liparis sp., C. maculatus, L. lumpraeti- 
formis, and Cottidae accounted for 91% of the total catch; 
the number of larvae declined sharply in spring with the 
end of the larval stage of these dominant species before 
reaching a low point in July and August (Chenoweth, 
1973). The ichthyoplankton of the St. Lawrence estuary, 
principally osmerids, gadids, cottids, cyclopterids and pleu- 
ronectids, consisted of 25 species (Able, 1978) and 15 of 
these also occurred in Penobscot Bay. 
Significance of demersal eggs 
The ichthyoplankton community of Penobscot Bay is dom- 
inated by larvae that hatch from demersal eggs within the 
estuary, whereas larvae that hatched from pelagic eggs 
are rare. This finding was first noted for the Mystic River 
estuary (Pearcy and Richards, 1962) and subsequently for 
the Sheepscot and Damariscotta estuaries (Chenoweth, 
1973). Able (1978) observed that the St. Lawrence estuary 
was almost exclusively inhabited by larvae that hatched 
from demersal eggs, primarily cottids, stichaeids, P. ameri- 
canus, and C. harengus and that the larvae from pelagic 
eggs were merely strays from more offshore waters. In all 
of these areas, larvae that hatched from demersal eggs 
dominated the catch within the estuary, whereas larvae 
that hatched from pelagic eggs were more common at the 
mouth of the estuary or adjacent ocean. 
Table 6 
Canonical correlation coefficients for relative larval fish abundance factor scores and environmental variables in 1997 and 1998. 
Underlined values significant at P = 0.05, adjusted for multiple comparisons such that P < 0.004 for a significant correlation. 
1997 
1998 
Variables 
Standardized 
coefficients 
Correlation with 
canonical variables 
Standardized 
coefficients 
Correlation with 
canonical variables 
Larval fish relative abundance 
Factor 1 
0.63 
0.66 
-0.66 
-0.65 
Factor 2 
-0.58 
-0.61 
-0.74 
-0.74 
Factor 3 
0.46 
0.49 
-0.16 
-0.09 
Environment 
Depth 
-0.06 
-0.19 
0.24 
0.36 
Plankton 
-0.02 
-0.37 
-0.28 
-0.82 
Temperature 
1.00 
1.00 
0.66 
0.94 
Salinity 
0.07 
-0.21 
0.19 
0.37 
