Can scalar timing explain variability in scanning patterns?

Why are scanning patterns so variable? Theory predicts that for prey species facing non-stalking predators scans to monitor predators should occur at fixed rather than unpredictable times. Yet, empirical distributions of inter-scan intervals (ISIs) are very variable. One hypothesis to explain variability is that animals initiate several of their scans in response to external disturbances that occur at random times. I propose, instead, that animals actually aim to initiate scans at fixed times, which are adjusted to perceived predation risk, but well-established cognitive processes on interval timing induce variability in ISIs. Signatures associated with scalar timing, a leading theory of interval timing in animals, include a linear increase in the standard deviation of ISIs as a function of mean ISI duration. The increase is expected to be proportional to mean ISI duration, which implies that the CV (SD*100/Mean) of ISIs is unrelated to mean ISI duration. Finally, the distribution of ISIs should be gamma-like with right skew. I tested these predictions in groups of domestic fowls (Gallus gallus domesticus) under controlled conditions and in groups of American flamingos (Phoenicopterus ruber) in the field. I found support for most but not all predictions in these two species. In particular, CV of ISIs increased with the mean, a deviation that I attribute to non-independent vigilance amongst group members. Cognitive processes associated with scanning patterns warrant further empirical testing.