words, the song is always specific, and this is the most noticeable, as it is the most remarkable, characteristic.
There is still, however, another quality to which I would draw attention—that of loudness. The sounds produced are on the whole alike penetrative, and the individuals of any given district, even though the climate by affecting their vocal muscles may have modified the character of the song, are at no disadvantage in this respect; neither are the females on the same account the less likely to hear the undeveloped voice of the immature male.
We have then the following considerations: firstly, there is the widespread and remarkable phenomenon of imitation, from which we can infer that the diversity of song is not due to structural differences but must be sought in some innate capacity to play one tune in preference to another; secondly, not all the individuals of the same species play a similar tune—we find that there is in certain directions a noticeable variation which nevertheless does not seem to affect the question of success or failure in the attainment of reproduction; in the third place, in contrast with this variation, we can observe a striking uniformity in two important particulars, namely in the specific character and penetrative power of the song—qualities which we know are essential for the purposes of "recognition" and "warning"; and finally, from the general course of our investi-
