Asian Giant honeybees (Apis dorsata) nest in single combs which are covered by several layers of bees forming a curtain. When moderately threatened the bees flick their abdomens, synchronously performed by a great number of individuals, which is said to repel potential predators effectively. Such colony responses can be evoked by visual stimuli like approaching wasps or birds, or even by wasp-like dummies (as we used). In selected video documents (in total: 5 hours observation) movements of bees were visualized using the Optimas^® image analysis tool. For that, movements were thresholded in 40 ms steps by using arithmetical, filtering and morphological operations combined with statistical processing.

Abdominal shaking is performed only by bees of the roofing layer of the curtain in all regions of the nest, it is initiated by a single wing stroke of 80-160 ms duration followed by an abdominal thrust of another 200-250 ms; still position is achieved again after further 200 ms. The abdominal shaking movements of thousands of bees are tightly coordinated, they form a cascadic colony response in temporal and spatial terms, emanating from one or more centers, radiating over a nest (e.g. with 80.000 individuals) within one second.

Such fast response patterns are initialized at particular sites of the nest, where groups of guard bees obviously respond more rapidly and powerfully than others. Visual and mechanoceptive sensation is here the obvious basis to let this colony process shoot up within 300 ms (revealing a maximum waving area of 130 cm²; the horizontal span of the small experimental nest was 80 cm).

While the wave affects more and more parts of the nest in the declining phase during the subsequent 600 ms, the total of "shaking" areas and dispersal speed decrease. Shaking activity is distributed saltatorically, whereas sub-centers are likely to amplify and distribute the process. The dummy’s direction is significantly (P<0.001) correlated to that of the shaking wave, but only 20 cm around the primary trigger center. Shaking is disseminated differently in different nest areas; in the mouth area near the rim, coupling is stronger (0.6) than beneath the mouth (<0.2). Shaking is spread fastest (<1 ms/cm) horizontally and close to the primary trigger center, and slower (>20 ms/cm) in the periphery. Such kind of directivity is probably caused by the posture of curtain bees which are hanging side to side with the abdomen down and the head up.