Before observing two conspecifics inspect a predator, the observer’s
preference for these two fish was determined in an apparatus similar to that of
Dugatkin (1992b) (Fig. 1). This initial preference test allowed us to examine the
possibility that observers were choosing to associate with others based on intrinsic
preferences rather than what they observed during the inspection bout. The
observer was placed in a 9.5 l aquarium and the two remaining fish (inspectors in
the subsequent inspection trial) were each randomly placed in one of the smaller
compartments on each end. After 15 min of acclimatization time, removable
opaque partitions between the observer and each side fish were lifted to allow
visual contact for 5 min. The quarter of the observer’s tank closest to each side
fish was designated as a preference area. The total number of seconds the
observer spent in each preference area was recorded on two stopwatches from
behind a black curtain. Preference was designated if the observer: (1) spent more
than half of the preference period in the preference areas and (2) spent >60% of
this time in one of the preference areas. Immediately following the preference test,
the guppies were transferred to the breeders for 48 h of isolation to minimize the
influence of their prior prior choice on subsequent choices in the post-inspection
preference test.