All nestlings within a brood were hand-fed to satiation once daily from day 4. Nestlings were stimulated to beg by tapping the nesting box and fed until begging was no longer elicited. Sugar water was enhanced by the addition of the carotenoids lutein and zeaxanthin (available in liquid form, OroGLO®, Kemin Industries) at a final concentration of 100 µg mL−1. Lutein and zeaxanthin are the main carotenoids circulated by adult Hihi, and the ratio of these carotenoids in OroGLO® is very similar to that circulating in their blood plasma (see Ewen et al. 2006b for details). The carotenoid-enhanced food was no different to that used in previous successful manipulations of carotenoid availability in adult Hihi (Ewen et al. 2006c). The food received by hand-fed nestlings was identical to that which adults received and fed to their nestlings. In each treatment, adults were provided with additional supplementary food, 10 m from the active nesting box. Food supplements were changed every 2 days, and presented in hummingbird-type feeders familiar to the birds because of their permanent use on the island (Taylor & Castro 2000). The use of feeders after establishment was monitored, with all pairs finding and taking the food within 30 min. Furthermore, previous research of the same population found no difference in the use of sugar and carotenoid-enhanced supplementary food (Ewen et al. 2006c). Supplementary food was provided to both adults and nestlings for 7 days, ending when nestlings were 10 days of age.
Nests were allocated to feeding treatments sequentially in order of hatching date, with the exception of a small number of nests that occurred in close proximity to each other. These had to be assigned to the same adult feeding treatment. In order to address the possibility that groups varied prior to treatment, small blood samples (c. 150 µL, using brachial venipuncture) were collected from a random sample of males (N = 24) and females (N = 10) prior to supplementary feeding (at the end of egg laying). Blood was immediately centrifuged to separate plasma which was then stored at –20 °C until subsequent carotenoid analyses were performed (see below). There were no significant differences in parental circulating plasma carotenoid levels (mean µg mL−1 ± SE, control (N = 13): 8·68 ± 1·08, hand-fed (N = 11): 9·45 ± 1·59, Parentally-fed (N = 10): 9·33 ± 1·64, F(2,31) = 0·14, P = 0·87), parental experience (mean combined parental ages in years ± SE, control (N = 11): 3·91 ± 0·71, hand-fed (N = 10): 5·9 ± 1·03, parentally-fed (N = 10): 4·89 ± 1·06, F(2,27) = 1·86, P = 0·18), or timing of breeding (mean days from first nest found ± SE, control (N = 11): 40·55 ± 3·95, hand-fed (N = 10): 39·3 ± 5·72, Parentally-fed (N = 10): 33·56 ± 5·89, F(2,27) = 0·78, P = 0·47) between the experimental groups. Neither sex, nor its interaction with experimental group was significant for any variable. In addition, there were no significant differences in mean brood mass (g ± SE, Control: 8·57 ± 0·44, hand-fed: 8·88 ± 0·6, parentally-fed: 9·07 ± 0·6, F(2,23) = 0·35, P = 0·71) at the start of the experimental treatment.