We chose not to use a factorial approach for the experimental design or analysis, as this would have required assigning arbitrarily temperatures as levels within the experimental factor and using poor replication within each level due to limited numbers of replicate animals.