On the changes of colour in the feathers of birds, independent of moulting