REVIEWS — A MONOGRAPH OF THE TROCHILIDAE. 49 



specific characters. He may probably intend — we tope be does — to 

 give at the conclusion of his work a careful digest of his views and a 

 synopsis of the order, with the requisite characters of genera and 

 species. "Whilst waiting for this it is hardly possible for us fairly to 

 discuss the goodness of generic groups^ of which we have to collect 

 the distinctions for ourselves, or search them out, scattered through 

 various works. 



The most obvious character for dividing the Humming birds is 

 perhaps found in the bill, which is straight, slightly curved, sickle- 

 shaped, recurved, and in a few instances furnished with recurved teeth 

 along one portion of both mandibles. Then we have the various 

 forms of the tail, the crests, tufts, and other appendages, the 

 position of the gem feathers, to which we believe Mr. Loddiges 

 attached much importance, and the pretty downy boots or muflFs on 

 the feet of many species, besides size, general distribution of colour- 

 ing, and peculiar habits or instincts. From all these sources we may 

 expect combinations of good and sufiELcient characters. What is 

 needed is to form the numerous species into natural groups, as many 

 as may be found necessary to express our observations on their 

 resemblance and differences, but taking care that these shall be of 

 real importance, minor distinctions only constituting sections of 

 genera and having no claim to burden science with additional names. 

 "When we think we clearly perceive which species must stand 

 together, we then observe carefully whatever is common to them all, 

 and select from what is thus collected concise characters, sufficient to 

 exclude all other species. It would be a manifestly wrong course to 

 contrive characters from abstract principles in relation to the number 

 of genera that must be found in a family, or the points that must be 

 deemed important, since these will differ in different families and are in 

 each case to be learned from observation. Every apparent distinction 

 must not be assumed to be a good generic mark. To combine well 

 is more useful as well as more difficult than to divide. The tendency 

 to make much of small distinctions and to elevate minor sections into 

 genera is the bane of Natural Science, and it is not always understood 

 how much easier a task it is than to find out the true boundaries of 

 natural groups. "We have fallen into this train of thought in 

 reflecting on the eighty-eight genera of Trochilidae already given in 

 Mr. Gould's sixteen numbers, containing two hundred and forty spe- 

 cies, giving two and a fraction species for each genus, and, as many of 



VOL. IV. E 



