Parental physiology and behaviour
For the parental individuals, the
multi-model selection approach identified two models within 2 ΔAIC,
one for which cortisol levels were
only affected by body weight (estimate: -3.39; t-value: 4.22, df=1,
p=0.001) and a second model (ΔAICc=1.87) which included both body weight
(estimate=-3.03; t-value: 2.86, df=1, p<0.001) and environment
(enriched or poor) (t-value: 7.38, df=1, p=0.03) as significant factors
affecting cortisol levels, with individuals from enriched environments
having lower cortisol levels (Fig. 1a-b; Table S3). Basal metabolic rate
decreased with body weight (estimate: -2.40; t-value:-2.35, df=1,
p=0.01) but was not affected by the parental environment (Table S3b). A
linear regression analysis between cortisol levels and basal metabolic
rate showed strong correlation between them (adjusted R² = 0.41;
F-value=16.73, df=1, p<0.001) (Fig. 1c).
Parental activity significantly decreased with body weight (estimate:
-2.88; z-value: 31.58, df=1, p=0.01), was lower in individuals reared in
poor environments (estimate: -0.25; z-value: 34.98, df=1,
p<0.001) and was influenced by with the interaction between
environment and body weight (estimate: 2.45; z-value: 38.76, df=1,
p=0.02) (Fig. 1d; Table S3). The number of inspections of the novel
object was significantly explained by body weight (estimate: -9.50;
z-value: -2.21, df=1, p=0.02; Fig. 1e; Table S3). The number of contacts
with novel object was significantly affected by environment, with
individuals from poor environments having higher number of contacts
(estimate:0.92; z-value: 2.98, df=1, p=0.001) (Fig. 1f; Table S3e). No
other equivalent model was found for activity, number of inspections and
number of contacts. Two of the individuals, one form the enriched and
one from the poor environment, were identified as outliers
(P<0.01) for the number of contacts (7 and 14 respectively)
but re-running the analyses without these two individuals did not change
the significance or direction of the difference in number of contacts
between groups. The same individual from the poor environment was
identified as an outlier for the inspections (10) and after its removal
from the analysis, neither body weight (estimate: -2.82; z-value: -0.51,
df=1; p=0.89) or the environment (estimate: 0.05; z-value: 0.12; df=1;
p=0.60) significantly influenced the number of inspections. Parental
latency to leave acclimatisation zone variation was not affected by body
weight or rearing environment
(estimate: 0.12; z-value: 1.08,
df=1, p=0.10).