When we reflect upon the problem of song and consider the numerous and diverse forms in which it is manifested, we are apt to draw a comparison between the sounds we hear and those produced by musical instruments, and hence to conclude that each bird is gifted with a special instrument in virtue of which it produces its characteristic melody. But there is a very remarkable phenomenon connected with the singing of birds which shows that this is really not the case--I mean the phenomenon of imitation. There are plenty of good imitators amongst our native species, and the power of imitation is not the exclusive property of those which have reached a high degree of vocal development, nor, for the matter of that, of song-birds at all. Even the Jay, than which few birds have a more raucous voice, that "hoots" like the Wood-Owl, or copies the sounds produced by the tail feathers of the Snipe, will occasionally imitate the most melodious strains of some other species; and the Red-backed Shrike, whose sexual call is principally a few harsh notes rapidly repeated, bursts at times into perfect imitations of the song of the Swallow, Linnet, or Chaffinch. Nevertheless it is amongst such typical songsters as the Warblers that we find the greatest volume of imitation, and no limit seems to be placed upon their capacity. The Marsh-Warbler can utter the call of the Green Woodpecker, or sing as the Nightingale does, with as much facility as it sings its own song; and the Blackcap is well-nigh as proficient in copying the cries and melodies of surrounding species--and so, if it were necessary, we might proceed to add to the list. These examples demonstrate that different songs are not represented by a corresponding number of different physiological contrivances; for if the difference were really attributable to some structural peculiarity, then the range of sounds embraced in the call-notes and the sexual call of any given species, must be the measure of the capacity of its instrument; and no matter how great its power of imitation may be, it follows that it will only be capable of copying those sounds which fall within that range. There is plenty of evidence to show that the power of imitation is almost unlimited, at all events that it is not confined within such narrow limits as are here demanded. Hence it seems clear that the diversity of song is not to be sought in structure, but in some innate capacity to play one tune in preference to another; and if this be so, and if out of the same instrument, which has been primarily evolved to further the biological end of intercommunication, all manner of diverse sounds can be made to proceed, the problem of the origin of song is to that extent simplified.