Suppl 5. The beta intercepts taken from our Bayesian multinomial logistic regression model. They illustrate the between season and year variation in moult timing.