_Correlation of Growth._--By this term I mean that the whole organisation is so connected, that when one part varies, other {168} parts vary; but which of two correlated variations ought to be looked at as the cause and which as the effect, or whether both result from some common cause, we can seldom or never tell. The point of interest for us is that, when fanciers, by the continued selection of slight variations, have largely modified one part, they often unintentionally produce other modifications. For instance, the beak is readily acted on by selection, and, with its increased or diminished length, the tongue increases or diminishes, but not in due proportion; for, in a barb and short-faced tumbler, both of which have very short beaks, the tongue, taking the rock-pigeon as the standard of comparison, was proportionally not shortened enough, whilst in two carriers and in a runt the tongue, proportionally with the beak, was not lengthened enough. Thus, in a first-rate English carrier, in which the beak from the tip to the feathered base was exactly thrice as long as in a first-rate short-faced tumbler, the tongue was only a little more than twice as long. But the tongue varies in length independently of the beak: thus, in a carrier with a beak 1.2 inch in length, the tongue was .67 in length; whilst in a runt which equalled the carrier in length of body and in stretch of wings from tip to tip, the beak was .92 whilst the tongue was .73 of an inch in length, so that the tongue was actually longer than in the carrier with its long beak. The tongue of the runt was also very broad at the root. Of two runts, one had its beak longer by .23 of an inch, whilst its tongue was shorter by .14 than in the other.