Weinrich et a I.: A shift in distribution of humpback whales, Megaptera novaeangliae, in reponse to prey 
829 
were used to record prey density on Stellwagen Bank 
in the immediate area where whales had been ob- 
served. The echo-sounder was used for 83 days dur- 
ing 1990 (9 May to 20 October; 153 total hours), 98 
days during 1991 (9 May to 28 September; 221 total 
hours), and 69 days during 1992 (24 April to 24 Oc- 
tober; 60 total hours). Clear readings throughout the 
water column (i.e. with no interference present) were 
obtained for 69 hours in 1990, 166 hours in 1991, 
and 60 hours in 1992. An echo-sounder operating at 
this frequency is likely to detect the presence of fish 
but unlikely to detect plankton unless it is present 
in extreme densities (Dolphin 4 ). The echo-sounder 
was started as the boat slowed to begin whale obser- 
vations and turned off when the vessel left the ob- 
servation area to return to port. Because echo- 
sounder tracings were obscured by noise when the 
vessel was moving at cruising speed (e.g. moving from 
one group of whales to the next), tracings performed 
at cruising speed were eliminated from analysis. A 
timing mark was placed simultaneously on both the 
echo-sounder chart and the data sheets by the sec- 
ond observer at 10-min intervals. 
The echo-sounder chart was later sampled at 2- 
min intervals by interpolating between the 10-min 
marks. For each sampling point, prey presence was 
scored visually in 3.3 m ( 10 ft) vertical increments 
from the surface to the bottom, with a sliding score 
of zero (for no prey) to 10 (prey throughout that 3.3 
m interval). From these readings mean values for 
vertical bait density were calculated for each quar- 
ter of the water column and the total water column. 
Mean depth in which readings were taken was 38.4 m 
(SD=15.1 m). No echo-sounder data were recorded 
on Jeffreys Ledge. 
Although such data give an idea of the availabil- 
ity of prey in the immediate vicinity of whales, they 
do not reflect an area where whales were not present. 
Hence, there could have been very similar or differ- 
ent prey concentrations very nearby, without that 
information ever being recorded. However, since each 
year’s data set came from numerous days and con- 
tained data points from several different locations 
(albeit within a 3-4 mile radius) within each day’s 
observations, we feel they at least give a crude over- 
view to overall prey densities in the vicinity of whales. 
Data management and analysis 
Both daily whale sighting data and prey density data 
were stored in PC-based computer files and analyzed 
with commercially available statistical software 
4 Dolphin, W. F. 1994. Department of Biomechanical Engineer- 
ing, Boston University, Boston, MA02215. Personal commun. 
(SPSS/PC+, Kinnear and Gray, 1992). For daily sight- 
ing data, an Xbase program was written to isolate 
the sightings of each whale and to calculate statis- 
tics summarizing that individual’s within-year sight- 
ing history (including occurrence and occupancy — 
see below) in each part of the study area. These val- 
ues were then stored in a separate data file and ana- 
lyzed with the same statistical software. Temporal 
trends were analyzed with least-squares regression 
(Snedecor and Cochran, 1967) of individual data 
points with the year of observation as the indepen- 
dent variable, although only annual means are pre- 
sented in our tables for occurrence and occupancy 
scores. The slope of the regression line (B) and the 
probability value ( P ) from a test of the null hypoth- 
esis that the slope did not differ from zero are pre- 
sented for each test. Calves were eliminated from 
these analyses because we assumed that a calf is 
merely following the mother in her choice of habitat. 
Definitions 
“Occurrence” is defined as the number of days on 
which an individual whale was photographed in a 
single year. “Occupancy” is the number of days 
elapsed from the first to the last recorded sighting of 
an individual whale within a year. These definitions 
are consistent with those used by Clapham et al. 
(1992). 
Results 
Stellwagen Bank 
Total number of humpback whales identified per 
year The number of humpback whales identified 
in any single year on Stellwagen Bank ranged from 
258 ( 1990) to a low of 7 ( 1994), with a mean of 153.6 
(SD=88.4) ( Fig. 2). These values show a statistically 
significant declining trend (B=-32.82, P=0.033). 
When the total number of whales was broken into 
age class, differences in annual trends were appar- 
ent. Numbers of adult whales identified on Stell- 
wagen ranged from 173 (1990) to 3 (1994; mean= 
102.8, SD=60.4). These values also showed a statis- 
tically significant declining trend ( B=-0.84, P=0.018). 
Number of juveniles identified in each year varied 
from 85 (1990) to 4 (1994; mean=50.71, SD=29.2). 
These also showed a downward trend, although not 
statistically significant (B=-23.50, P=0.099). The 
ratio of identified adult whales to identified juveniles 
varied from 2.5:1 (in 1988) to 0.75:1 (in 1994). Num- 
bers of cow-calf pairs throughout the study period 
