Time series problem

This is my first posting.

I have collected a time series dataset.
Individually recognisable fish were observed at three times of day. I wish to test whether their behaviours (e.g. maximum distance from shelter) differ with time of day. Since I have measured the behavioiur of the same fish at each of the three time periods there is an independence problem. It seems like a repeated measures problem (well that's how I initially conceived it), but there is no 'treatment' group variable, so a standard repeated measures anova cannot be performed. It seems as though it should be simple. If I had 2 time periods it'd be a simple paired t-test.
Any suggestions?