almost at random, and the majority are destroyed before the next spring. This is proved by the fact that the birds which return in spring are as a rule not more numerous than those which came the {27}preceding spring, whereas those which went away in autumn were two or three times as numerous. Those young birds that do get back, however, have learnt by experience, and the next year they take care to go with the old ones. The most striking fact in favour of the "instinct" of migration is the "agitation," or excitement, of confined birds at the time when their wild companions are migrating. It seems probable, however, that this is what may be called a social excitement, due to the anxious cries of the migrating birds; a view supported by the fact stated by Marcel de Serres, that the black swan of Australia, when domesticated in Europe, sometimes joins wild swans in their northward migration. We must remember too that migration at the proper time is in many cases absolutely essential to the existence of the species; and it is therefore not improbable that some strong social emotion should have been gradually developed in the race, by the circumstance that all who for want of such emotion did not join their fellows inevitably perished. The mode by which a passage originally overland has been converted into one over the sea offers no insuperable difficulties, as has already been pointed out. The long flights of some birds without apparently stopping on the way is thought to be inexplicable, as well as their finding their nesting-place of the previous year from a distance of many hundreds or even a thousand miles. But the observant powers of animals are very great; and birds flying high in the air may be guided by the physical features of the country spread out beneath them in a way that would be impracticable to purely terrestrial animals.