82 
Fishery Bulletin 99(1) 
Figure 1 
Map of stations sampled biweekly with a 1.0-m plankton net in Penobscot Bay, 
Maine, from 4 April to 25 June 1997 and from 18 March to 30 April 1998. II = 
Iselboro Island; NH = North Haven Island; and VH = Vinalhaven Island. 
In 1998, four one- or two-day cruises (98I-98IV) were 
conducted biweekly from 18 March through 30 April 1998 
and data collection was the same as in 1997 except that 
ten ichthyoplankton stations were sampled in the lower 
bay only. Eight of these stations were sampled in 1997 
including five midbay stations (B1-B5) and three lower 
bay stations (02-04). Two additional lower bay stations, 
01 and 05, were added in 1998. Larvae from both years 
were preserved in 5% formalin for later identification to 
the lowest taxon possible by the Atlantic Reference Center 
of the Huntsman Marine Biological Laboratory in St. An- 
drews, New Brunswick, and for quantitative determina- 
tion of larval fish densities (number of larvae per filtered 
100 m 3 ). Fish larvae were measured for standard length, 
or in some cases notochord length, to the nearest mm. 
Plankton volume standardized by volume filtered was de- 
termined for each tow by displacement of the unidentified 
plankton. 
Data analysis 
Larval fish abundance and environmental data were 
analyzed by using three multivariate techniques: princi- 
pal components analysis (PCA), multivariate analysis of 
variance (MAN OVA), and canonical correlation analysis 
(CCA). PCAs of the variance-covariance matrices derived 
from both environmental and larval abundance data were 
performed to reduce intercorrelated variables to a smaller 
number of uncorrelated variables. This procedure provided 
a concise description and comparison of complex spatial 
