Tidal and spontaneous activity patterns in fiddler crabs

Abstract

1. We described locomotor activity patterns in fiddler crabs by the frequency distributions of the duration of single activity bouts and of single rests. No significant interrelations between consecutive activity bouts and rests were detectable. The analysis of the frequency distributions led to the conclusion that the patterning of activity behaviour in constant conditions can be described by stochastic processes (Lehmann, Neumann, Kaiser, 1974). 2. Using the very parameters of the analysis we simulated activity patterns with two models. In model 1 the duration of each activity bout or rest is determined at its onset by one stochastic decision using the observed frequency distributions (Fig. 1). In model 2 the transition from one state (activity or resting) to the alternative state is controlled by a series of consecutive stochastic decisions at regular intervals (Fig. 2). 3. The simulated activity patterns produced by the two models using the data of freerunning crabs are similar to the original patterns. They are rather random and give only for short sequences the impression of a weak rhythmicity (Figs. 5, 9, 11). 4. Simulation with the data of the only one crab which displayed a spontaneous circadian rhythm yielded rhythmic patterns despite the fact that no serial correlations and no phase control were fed into the simulation (Fig. 14). 5. Simulation with the data of crabs which were entrained by artificial tides also yielded distinctly rhythmic patterns (Figs. 17, 18). 6. The results of 4. and 5. demonstrate that the preference of a certain duration of the rests is sufficient to organize an activity pattern rhythmically without additional control by an endogenous oscillator. 7. Entrainment of the activity by a Zeitgeber regime was simulated with an extended model 2: The supposed external cycle influences the transition probabilities from activity to resting and vice versa. According to the strength of the external influence the simulation produces more or less strictly synchronized patterns (Fig. 21). 8. The variation in the results of the stochastic simulations demonstrates the effects of chance. Simulations may be used to scan the range of random variations. 9. The involvement and the properties of an endogenous oscillator can not be deduced with certainty from an observed rhythmic pattern only, since the properties of behavioural mechanisms may be sufficient to produce a fairly clear periodicity.