Can sequential vigilance be predicted?

Sequential randomness is one of the three important assumptions for Pulliam's vigilance model (1973). Here we tested the sequential randomness in Black-necked cranes Grus nigricollis, to see if the vigilance sequence can be predicted. Not similar to other recent studies, we found that most vigilance sequences (44/46) passed runs randomness test, and the length of an inter-scan interval was usually unrelated to its previous scan duration. Our findings suggest high predation risk might favor a random pattern of vigilance.