812 
Fishery Bulletin 96(4), 1998 
Not all times of day were sampled equally; there- 
fore, abundance of harbor porpoise in relation to time 
of day was compared by using number of porpoise 
observed per minute to standardize the data. Mean 
number of harbor porpoise observed per minute for 
each hour of daylight was compared with Kruskal- 
Wallis nonparametric analysis of variance and 
Kolmogorov-Smirnov goodness-of-fit analyses (Zar, 
1984). Nonparametric statistics were used for data 
with non-normal distributions or unequal variances. 
Fixed boat surveys 
To determine temporal changes in harbor porpoise 
distribution between 1991 and 1992, six 8-km 
transect lines (hereafter called fixed transects; Fig. 
2), placed in areas of harbor porpoise occurrence (pre- 
liminary harbor porpoise surveys and information 
from Orca Hotline, The Whale Museum, Friday Har- 
bor, Washington), were surveyed regularly from 27 
July to 26 August 1991 and from 24 July to 28 Au- 
gust 1992. In 1991, there was only one observer per 
survey; therefore, only one half the transect (bow out 
to 90° port or starboard) was completed during each 
survey. To be consistent in 1992, one observer sur- 
veyed from bow to 90° port while the other surveyed 
from bow to 90° starboard so that one half of each 
transect could randomly be compared to 1991 
transects. 
Harbor porpoise were counted from the same 7.3-m 
vessel as in random surveys during a Beaufort sea 
state of 0 or 1. Each fixed transect survey was con- 
ducted at an average speed of 11 km/h and completed 
in 40 to 45 minutes. This vessel speed was chosen in 
1991, and to be consistent, 1992 fixed transect sur- 
veys were conducted at the same speed (instead of 9 
km/h as in random boat surveys). Harbor porpoise 
locations were calculated as in random boat survey 
methods. 
Mean number of sightings of harbor porpoise per 
survey between 1991 and 1992 was compared by us- 
ing a /-test. Because both sides of the vessel were 
observed during a single survey in 1992, each side 
could not be considered an independent sample. 
Therefore, one side of the vessel was randomly cho- 
sen from each survey in 1992 to compare with 1991. 
Power tests (Cohen, 1988) were performed when re- 
sults were not significant. 
Results 
Random boat surveys 
There were 301 sightings of 526 harbor porpoise (Fig. 
3) during random boat surveys. Of these, 20 sightings 
(39 porpoise) were possible resightings (i.e. observer 
believed the porpoise had already been seen during 
that survey, given the location 
and direction of travel of por- 
poise), therefore, these pos- 
sible resightings were not used 
in analyses. An average of 4.4 
harbor porpoise sightings were 
recorded per hour (8.1 harbor 
porpoise per hour), with group 
sizes of 1 to 8 (mean=1.87, 
SE=0.06, n= 278) individuals. 
Thirteen cow and calf pairs 
were observed between June 
and September. Harbor por- 
poise were sighted during 75% 
of surveys at a mean perpen- 
dicular distance of 237 m 
(SE=13.89, n=250, range: 0 to 
1060 m) from the trackline. 
The half-normal (hermite) 
model, truncated at 750 m, 
best fitted the frequency dis- 
tribution of perpendicular dis- 
tance of harbor porpoise 
sighted from the trackline (Fig. 
4). Using harbor porpoise 
sightings (n=250) for all sec- 
Figure 2 
Harbor porpoise locations along fixed boat transects ( 1, 2, 3, 5, 6, 7) in 1991 and 1992 off the 
northern San Juan Islands, Washington. The “x” denotes locations of harbor porpoise sighted 
in 1991; the denotes locations in 1992. 
