SOCIAL BEHAVIOR IN MORMYRID FISH (MORMYRIFORMES, PISCES): SHORT AND LONG-TERM CHANGES ASSOCIATED WITH REPEATED INTERACTIONS

Abstract

1. The behavior of groups consisting of four weakly electric fish each (Gnathonemus petersii) was observed for 10 min periods on daily and weekly schedules. 2. During a testing day (short-term) locomotor activity decreased while group cohesion among fish increased. Over a period of six weeks (long-term) the same trends were observed for the mean values of these two behaviors (Fig. 2). Fig. 2. Changes in locomotor activity (I) and group cohesion (II) over time. a- Repeated observations during the same day (o, first day; =, following session 15 days later). b- Means (and S.D.) of the first four observations of the day through 7 days over a 6-week observation period. c- Weekly observation of the first 600 s period of trial. Figure options Download full-size image Download as PowerPoint slide 3. In 16 naive control groups, during 10 min of interaction, exploratory backwards swimming remained at a constant level, single file swimming of three and four fish was predominant during the first third, while agonistic interactions between two fish were prevalent during the remainder of the observation period (Fig. 3). Fig. 3. Mean frequencies of four behaviors: Backwards Swimming Individual (BWI), Body Attacks Displays (BAD), Antiparallel Displays (APD), Single File Swimming (SFS) over the 600 s periods of observation in 16 groups of naive fish. Upper graphs: mean frequencies per 10 s bins (dashed line indicates random distribution). Lower graphs: mean frequencies of 3 consecutive 200 s periods (* indicates significant differences at 0.05 level from random distribution). Figure options Download full-size image Download as PowerPoint slide 4. In the experimental group (same four fish), over a period of 16 weeks, agonistic, probing and exploratory behaviors decreased while ritualized lateral and parallel displays increased (Fig. 4). Fig. 4. Development of various behaviors in a group of fish. Sixteen weekly sessions with the same four fish were followed (arrow) by five sessions in which one fish was replaced by a naive. In a final session (=) the original group was restored. Mean frequencies are expressed as percentages of mean occurrences (in brackets with S.D.) calculated over 16 control groups. a- individual behaviors: Backwards Swimming (BWI); Chin Probing (CHP); Body Attacks Displays (BAD). b- reciprocal displays: Lateral Display (LAD); Parallel Display (PAD); Circling (CIR). c- reciprocal displays: Antiparallel Display (APD) and social behaviors: Occurrence of Group of 3 fish (AG3); Single File Swimming (SFS). (* indicates significant differences, at 0.05 level, from previous value or from control for the first column). Figure options Download full-size image Download as PowerPoint slide 5. Dishabituation trials (one experimental fish was replaced by a naive fish on week 17) yielded reversals in the above listed behaviors, thus demonstrating the existence of an individual or group recognition mechanism in G. petersii (Fig. 4).