Breeding design and rearing condition

Three-spined sticklebacks were captured in the Rio Ulla (Spain) in February 2013 before sexual maturation and used for breeding. Like in other populations near the southern edge of the species’ range [38], the majority of fish in this population reproduce repeatedly throughout a single relatively long breeding season, after which they die. A total of 32 F1 families were produced by breeding16 sires and 16 dams. Each breeder mated twice with two different mates during April-May 2013 (for details, see [39]). Thus, each F1 fish had full-sibs and maternal and paternal half-sibs. The breeders were housed in individual tanks and paired with a single mate at a time. The fertilised clutches were collected from the nests within 3 h and incubated in a 100-l tank, following the standard egg husbandry protocol [40]. Each full-sib clutch was isolated in a hatching tank with a sponge filter prior to hatching; then hatchlings were reared there until age 40 days (mean ± SE number of fry per full-sib family: 57.8 ± 3.0, n = 32 families). Survival rate of the F1 families during the first 40 days was extremely high (mean survival rate: 0.971 ± 0.006).

At age 40 days, fry in each F1 full-sib family were allocated among two (n = 7 families) or four (n = 25 families) 8-l growth tanks (n = 114 tanks). Each growth tank initially housed 11 or 12 juvenile fish. The rest of the fish were housed separately and used in other studies. The growth tanks were connected to four closed water systems (30 tanks/system), in which water was continuously filtered for nitrification, aerated and temperature-controlled by the combined flow-through function. Juvenile fish were fed daily ad libitum on a progressive diet of newly hatched Artemia and a commercial pelleted diet (Gemma Micro, Skretting, Norway). We analysed the total carotenoid concentration in these food items. Carotenoids were repeatedly extracted using n-hexane, and then carotenoid concentration was determined in a spectrophotometer (Synergy HT, BioTek, Winooski, VT, USA) at 440 nm using a lutein curve as standard. Both food items contained high levels of carotenoids (wet Artemia larvae: 16.9 μg g⁻¹; dry food pellets: 103.9 μg g⁻¹). The programmed photoperiod in the tanks reflected the natural seasonal pattern in the region. This fish stock was reared also for an experiment to test reaction norms of life-history traits in response to winter temperature; a half of each F1 family replicates (i.e., growth thanks) were maintained at 14 °C and the other half experienced a gradual temperature change to 9 °C during winter [37]. This difference in winter conditions was taken into account in all statistical analyses.

At age 6 months juvenile fishes were permanently marked with colour elastomer tags (Northwest Marine Technologies, Shaw Island, WA, USA) under a low dose of benzocaine anaesthetic to track individual life-histories (n = 1038 individuals). Randomly selected samples from the fish stock were sacrificed with an overdose of benzocaine anaesthetic during the growth period to be used in the study of temperature manipulation (n = 221 in Sep-Nov 2013; n = 364 in Feb 2014). The juvenile survival rate was calculated in each full-sib family replicate as the proportion of individuals surviving to maturity (i.e., expression of red colouration in males and spawning in females) in the growth tank, excluding the sacrificed individuals.