I think the burden of proof is on the "orange cats have statistically-significant behavioural specificity" side