Monitoring aggression
Fish used in the present experiment were acclimated in five 1 m3 outdoor holding tanks connected to a biological filter in a recirculation system. Water was aerated using submerged air diffusers connected to a regenerative blower. Nile tilapia were size sorted by hand, and 34 fish (46.0 ± 14.97 g; mean ± SD) were stocked in each tank. Each tank was assigned one of the five dietary treatments and fish were offered feed at 3% body mass (BM) divided into two daily portions, at 8:00 and 17:00 hours, for two weeks. At the end of the two weeks, a subset of the fish in each tank were relocated from the holding tanks to 52-L glass tanks connected to a biological filter, and a sand filter in a recirculation system, to monitor aggression. Five fish of similar sizes were randomly collected from each of the five holding tanks, group-weighed and stocked in the glass tanks. Fish were left to acclimate in the glass tanks for 24 hours, and offered their respective feeds at 3% BM, twice daily, at 8:00 and 17:00 hours. Fish were then monitored for signs of aggression using an adaptation of the Falter (1983) ethogram. Aggressive behavior such as nipping, mouth fighting/ attack, lateral fighting/ attack, chasing and flight, and cornering were quantified twice daily for 10 minutes (after morning feeding and before second feeding) on days 2, 3 and 5 after stocking. All observations were recorded by two people and their observations compared. Number of animals per tank hindered the quantification of aggressive behavior, therefore frequency of aggressive behavior was not considered but rather the presence or absence of aggressive behavior. Data collected was used to calculate an aggression index whereby the presence of a criterion elicits a score of “1”, and the absence a score of “0”. As such, each tank was given a score (on a scale from 0 to 5) and average scores of three monitoring days were recorded. Observers also monitored position of fish in the tank and coloration of fish. The monitoring procedure was replicated five times using new fish from the holding tanks (Table 2). Fish stocked in previous experimental weeks continued to be observed throughout the duration of the experiment but only data collected on days 2, 3 and 5 were used to calculate the aggression index.