Bradbury et al. : Daily and seasonal activity patterns of female Tautogolabrus adspersus 
651 
not foraging. The female cunners we studied exhib- 
ited similar behavior, seeking shelter beneath rocks 
or in crevices during the day. 
Some workers have assumed that temperate 
wrasses use cover to avoid predation ( Olla et al., 1979; 
Hobson et al., 1981), although threat of predation 
has not been well documented as a factor. Whoriskey 
(1983) interpreted the diurnal use of shelter by cun- 
ners in Massachusetts as predator avoidance. Al- 
though he may have been correct, our field observa- 
tions in Newfoundland do not support this hypoth- 
esis. Predation on adult cunners in Conception Bay 
is very low as judged by over 400 hours of diving ob- 
servations during which no predation, or attempted 
predation, on adults was observed. 
Females may seek shelter to avoid conspecifics 
with courting and chasing behaviors, especially ter- 
ritorial males. This explanation, however, is inad- 
equate in elucidating why males exhibit the same 
behavior or why the behavior occurs so frequently 
outside the spawning period. 
A reduction in energy expenditure may be associ- 
ated with such behavior because cunners probably 
require less energy to maintain a position in a shel- 
ter than in the water column, even when water move- 
ment from currents and waves is minimal, and be- 
cause the length of inactivity bouts increased with 
increased water turbulence. However, such an ex- 
planation is weak unless it can be shown that con- 
tinued foraging would result in a net loss of energy. 
In many diurnal fishes, including cunners, the 
onset and cessation of daily activity coincides closely 
with the rising and setting of the sun (e.g. Hobson, 
1972, 1973; Hawkins et al., 1974; Olla et al., 1974, 
1975; Clark and Green, 1990). As expected, females 
exhibited a marked seasonal decrease in the dura- 
tion of their diurnal activity (from 16.5 h in June- 
July to 11.0 h in October-November) as day length 
decreased. Light intensity at sunrise and sunset was 
affected seasonally by the surrounding topography 
at the study site (e.g. in the fall the sun “set” behind 
a range of hills rather than at sea level), and this 
topography may have accounted for some of the sea- 
sonal change in the onset and cessation of diurnal 
activity. However, the considerable variation in the 
onset and cessation of daily activity among cunners 
suggests that this variation is not simply a response 
to a threshold light intensity. 
Contrary to the expectation that female cunners 
would maximize foraging opportunities prior to en- 
tering winter torpor, they spent a larger percentage 
of the diurnal period inactive as the length of the 
photoperiod decreased. Why cunners should signifi- 
cantly reduce their foraging activity, at a time when 
food is still available and they could acquire more 
energy for somatic growth and winter torpor, is not 
clear. Although water temperature is decreasing dur- 
ing this period, our analyses show that above ~5°C, 
temperature has little direct effect in determining 
the ratio of activity to inactivity. Mean daily water 
temperatures during June-July and October-No- 
vember were approximately the same (8.2°C and 
9.2°C, respectively )(Fig. 1), yet there were large dif- 
ferences in the amount of time cunners spent in shel- 
ter. During June and July, females were outside their 
shelter for about 12.5 hours of the day, whereas from 
October to November cunners were active, on average, 
only 3 hours of the 11-h sunrise to sunset period. 
Fall or winter decreases in the activity or feeding 
behavior (or both) of fishes in the absence of changes 
in water temperature are common, although the 
mechanisms underlying these decreases are not un- 
derstood. Smith et al. (1993) for example found that 
in Atlantic salmon ( Salmo salar), seasonal reductions 
in swimming activity and feeding were more closely 
related to day length and changes in day length than 
to other environmental variables, including water 
temperature. This also seems to be true for cunners. 
Presumably their temporal pattern of activity is 
adaptive and important to their success at northern 
latitudes. Cunners survive six months or more of 
torpor that can begin at a time not predictable by 
exogenous cues in the marine environment. At our 
study site, the date at which winter torpor com- 
mences (i.e. when seawater temperature remains 
below ~5°C) can vary year to year by at least four 
weeks. Perhaps an endogenous mechanism sensitive 
to changes in day length, or to some other environ- 
mental cue, regulates the physiological processes 
associated with successful winter torpor. Although 
such a mechanism may exist in cunners, the identi- 
fication of endogenous rhythms in fishes is difficult 
(Boujard and Leatherland, 1992). 
Our findings concerning seasonal changes in the 
activity patterns of cunners have implications for 
estimating the size of their populations. For example, 
population estimates based on visual surveys by 
divers should take into account that, depending on 
when the survey is conducted, a significant and vari- 
able proportion of the population will be out of sight, 
under cover. Significant errors in estimates of popu- 
lation size are likely, and errors will not be consis- 
tent for different times of the year. This caution may 
apply to other species with similar behavior patterns. 
Acknowledgments 
We thank Robert Dunbrack for many helpful discus- 
sions and his review of the manuscript. John Gibson 
