Summary
Naïve kin groups and mixed‐family groups of bluegill Lepomis macrochirus larvae were exposed to a novel predator cue. The larvae responded by increasing shoal cohesiveness in kin groups but not in mixed‐family groups; moreover, larvae sired by males of the ‘cuckolder’ life history tended to have an enhanced ability to respond to direct cues of kinship v. larvae sired by males of the ‘parental’ life history, which instead appeared to respond to cues of life history rather than relatedness per se. The increased shoal cohesion among related individuals probably confers a survival benefit and indicates that the antipredatory shoaling response is innate in L. macrochirus.
Methodology
In June 2006 and 2008, swimmers equipped with snorkelling gear conducted daily surveys of L. macrochirus breeding activity in Lake Opinicon. When spawning was discovered, mature parentals, cuckolders and females were collected opportunistically using snorkelling gear and dip nets and transported by boat to the aquarium facilities at the Queen's University Biological Station, which sits on the lake's shore. These fish were used to create families using in vitro fertilization techniques as described in Neff & Lister (2007). Briefly, sperm was collected in 2 ml syringes from cuckolders and parentals by applying pressure to the gonad region of the abdomen. Eggs were collected from gravid females by applying gentle pressure to her abdomen. Eggs were then fertilized with sperm from either a cuckolder or a parental and then reared in 500 ml glass jars filled with lake water and equipped with a small airstone. Each female and each male was used only once to form a family. Fifty per cent water changes were conducted daily until larval swim‐up (5–8 days post‐hatch), which signals the onset of exogenous feeding.
Antipredator response trials were conducted in brown translucent tanks measuring 409 mm × 282 mm × 150 m m filled to a depth of 8· 0 cm with water from Lake Opinicon. Each tank was visually divided using a horizontal grid of 28 equally sized rectangles (arranged as 4 × 7) positioned beneath the tank. On the first day of exogenous feeding, ‘pure’ and ‘mixed’ groups were formed by transferring larvae of known pedigree to trial tanks using a small plastic pipette. Pure groups consisted of 10 full‐siblings sired by either a parental or a cuckolder. Three types of mixed groups of 10 larvae were created: a group of the ‘mixed parental’ type consisted of five full‐siblings sired by a parental and five full‐siblings sired by a second parental; a group of the ‘mixed cuckolder’ type consisted of five full‐siblings sired by a cuckolder and five full‐siblings sired by a second cuckolder; and a group of the ‘mixed life history’ type consisted of five full‐siblings sired by a parental and five full‐siblings sired by a cuckolder. Within all mixed groups, the two sets of full‐siblings were themselves unrelated (i.e . the sets had different fathers and mothers). Families used in the trials were used once to form pure groups and a maximum of twice to form mixed groups. The larvae were then allowed to acclimate for c . 24 h. One hour before a trial began, a male pumpkinseed Lepomis gibbosus (L.), a known predator of L. macrochirus larvae: Gross & MacMillan, 1981; Neff, 2003) was placed in a tank filled with 10 l of lake water. This water conditioned by the L. gibbosus served as a predator cue in the trials. A trial began by recording the grid co‐ordinates of each larva. Then, for ‘control’ treatments, 100 ml of unconditioned lake water was added to the centre of the tank and for ‘predator’ treatments, 100 ml of predator‐conditioned water was added to the centre of the tank. Pilot trials were conducted to determine that a volume of 100 ml was small enough that the introduction of the cue would not disturb the fish in the tank, but was large enough to induce a response (Hain & Neff, 2006). The grid co‐ordinates of each larva were read by a naïve observer to a stenographer at 20 and 80 s after the addition of the cue. The times when the location of the fish was recorded was chosen to give one initial measure after sufficient time for the larvae to respond to the odour cue, and another measure 1 min later to give an indication of how stable these groups were over time. The locations of the most active larvae were recorded first, followed by the location of less active or stationary larvae, and this process typically took no more than a few seconds per time step. All groups were subjected to both treatments on consecutive days with approximately half of the trials (n = 31) starting with the control treatment and the other half (n = 30) starting with the predator treatment. Fifty per cent water changes were performed between trials using water from Lake Opinicon.