• Rogers, Chandra M. C.
  • Neff, Bryan D.
  • Knapp, Rosemary


Male parental care in vertebrates often involves both defensive and nurturing behaviors. Whether androgens differentially mediate these two types of behaviors, or a trade-off exists between them, has been studied by behavioral endocrinologists for years but predominantly in species with biparental care. In such systems, potential detrimental effects of elevated androgens on parental care behaviors are often compensated for by changes in behavior of the unmanipulated parent. In contrast, for species where only one parent provides care, manipulation of androgen levels may more clearly determine if there are differential effects of androgens on these two types of behaviors and whether the proposed trade-off between defensive and nurturing behavior exists. Here, we manipulated androgen levels in two ways in bluegill sunfish (Lepomis macrochirus), a species where males provide sole parental care for the young. At the onset of the care period, males were implanted with 11-ketotestosterone, a major teleost androgen, the androgen receptor antagonist flutamide, or a blank implant. A separate control group experienced no manipulation. Males were then observed over several days and tested for their aggressiveness towards an experimentally-presented brood predator and for nurturing behavior (fanning of the eggs, removal of dead or fungal-infected eggs). Males implanted with 11-ketotestosterone displayed 64% more aggressive behaviors and 71% fewer nurturing behaviors than control groups. In contrast, males implanted with flutamide displayed 7% fewer aggressive behaviors and 126% more nurturing behaviors than control males. Taken together, these results show that aggression and nurturing behaviors are mediated by androgens and suggest that there is a trade-off between the two behaviors during parental care in this species.


Treatment assignment and implants

From June 4 to June 19, 2010, swimmers monitored the shores of Lake Opinicon for formation of bluegill colonies and mating activity. The day after spawning, 98 males were captured from their nests using dip nets and were brought to a nearby boat, where initial blood samples (~ 300 μL) were taken from the caudal vein. Numbered tiles were placed on the shore side of each nest for fish identification and nest covers were placed on nests to protect the eggs from brood predators while parental males were being handled. Blood collection time, measured from the time the fish was caught until the needle was removed from the caudal vein, averaged 67 s (range: 35–174 s). Fish were then anesthetized using clove oil, weighed (to the nearest gram) and measured for total length (to the nearest mm). Fulton's condition factor was calculated using the equation (mass/length3) × 105, which estimates an individual's energetic state (Neff and Cargnelli, 2004). An egg score was also assigned to each parental male's nest following Claussen (1991) for a score of 1 (1–4900 eggs), score 2 (4600–29,000 eggs), score 3 (27,000–53,000 eggs), score 4 (49,000–87,000 eggs) or score 5 (82,000–113,000 eggs). The fish were then implanted in the abdominal cavity with either a sham implant filled with castor oil (S; n = 17), a flutamide implant (F; n = 17), one 11-ketotestosterone implant (KT1; n = 20) or two 11-ketotestosterone implants (KT2; n = 23). A fifth group consisted of males that did not undergo surgery or receive an implant, which served as our controls (C; n = 21). Fish were collected randomly from the colony and treatment was assigned by rotation through the five treatments. Sham and KT implants were made with silastic tubing (Konigsburg Instruments, Pasadena, CA, United States) and were 7 mm in length (1.47 mm i.d., 1.96 mm o.d.) with 1 mm silicone sealant on the ends. The KT implants were made using crystalline KT (Steraloids, Newport, RI, United States) dissolved in ethanol and subsequently mixed into castor oil (concentration ~ 8 mg KT/mL oil). Flutamide implants were made by packing the interior of the tubing with crystalline flutamide (Sigma Aldrich, Oakville, ON, Canada). These implants were 8 mm in total length (1.47 mm i.d. and 1.96 mm o.d.) and were sealed with 1 mm silicone sealant on the ends. KT and flutamide doses were calculated based on doses used in previous studies on fish and birds while also calculating for average parental male bluegill body mass (Alonso-Alvarez et al., 2007, Kindler et al., 1991, Ros et al., 2004, Yamaguchi et al., 2004, Yamaguchi et al., 2005).

Males that underwent implant surgery were administered with 50 μL of antibiotic solution (oxytetracycline; Sigma-Aldrich, Oakville, ON, Canada) into the incision to prevent infection and the wound was sealed with New Skin (Prestige Brand Holdings, Inc., Irvington, NY United States). Fish were placed in a bucket filled with lake water and allowed to recover for 5 min before they were placed back on their nests, where they typically resumed care within a few minutes.

Behavioral observations

On the first and second days after implantation, males were observed by one of six swimmers for aggressive and nurturing behaviors between 0900 and 1700 EST. The presence of observers does not tend to affect the behavior of bluegill (e.g., Coleman et al., 1985, Gross, 1982). All swimmers were blind to the treatment group of each fish. Aggressive behaviors were assessed by placing a ca. 16-cm pumpkinseed sunfish (Lepomis gibbosus) on the edge of the parental male's nest. The pumpkinseed was held in a plastic bag attached to a 1.5-m pole. On each of the two days, the number of bites, opercular flares and lateral displays directed towards a brood predator were recorded over two 30-second periods, with a 30-second break away from the nest in between presentations (see Neff and Knapp, 2009). These numbers were later summed as a measure of total aggressive behavior (i.e. based on the full 2 min).