\received
DD MMMM YYYY
\acceptedDD MMMM YYYYStatistical analyses
All statistical analyses were conducted in R v.4.2.2 (R Core
Team, 2022). To test for influences on somatic growth, the effects of
diet, habitat (plus interaction), and sex on fork length for all trout
(n = 180) were modelled with a linear model.
To test factors influencing the cognitive performance of observer trout,
a linear model modelled the effects of diet, habitat (plus interaction),
and sex on the difference in time for each triad between the observer
trials and naïve trials (i.e. observer trial time minus naïve
trial time) until the dominant/subordinate relationship was clearly
established. Triads that were unable clearly to establish
dominance/subordination in either the naïve or observer trial were
omitted from this and subsequent models. To ensure that triads were
appropriately size-matched, linear mixed effects models using triad as a
random effect confirmed that whether individuals would become dominant
or subordinate was not affected by small differences in size after
approximate size matching (naïve trail:F 1,104 = 0.013, p = 0.911; observer trial:F 1,104 = 0.06, p = 0.809).
To test factors influencing encephalisation, the effects of diet,
habitat (plus interaction) and sex on relative whole-brain mass for all
trout whose brains were preserved in formalin were modelled with a
linear model. To consider brain morphology, MANOVA tested the effects of
diet, habitat (plus interaction), and sex on relative volumes of brain
regions. To test the effects of encephalisation and brain morphology on
cognitive performance, a linear regression modelled time differences
between naïve and observer trials on relative brain mass and relative
volumes of brain regions in addition to habitat and sex of those
observer trout whose brains had been extracted intact (n = 52);
diet was omitted as it was colinear with, and causally linked to,
encephalisation.
To examine differences between treatment groups in how individual fatty
acids were routed to specific tissue/lipid types (i.e. brain and
muscle tissue, polar and neutral lipids), mean percentages of total
lipids composed of individual fatty acids across tissue/lipid types were
tested with one-way ANOVAs followed by Tukey’s HSD post hoc . To
determine whether fatty acid contents of various tissue/lipid types were
influenced by treatment, the effects diet, habitat (plus interaction)
and sex on the percentage of each fatty acid in the n-3 and n-6
bioconversion pathways were tested with MANOVA. Effects on each fatty
acid percentage were then tested with ANOVA for individual tissue/lipid
types. Differences between treatment groups in the depletion of
Δδ 2H in fatty acids across tissue/lipid types
were evaluated by modelling the effects of diet, habitat (plus
interaction) and sex on the Δδ 2H of individual
fatty acids and tested with MANOVA. For specific tissue/lipid types,
effects on Δδ 2H were tested with ANOVA.