452 PKOCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 



spite of Mr. Seebolim's .statement, be hopeless ; but, luckily, they may 

 be distinguished by very recoguizable and distinct characters; the dif- 

 ferent construction of the wing-, in particular, rendering their separation 

 easy. 



A further examination of the birds included in the genus Geocichla 

 shows that in several species the sexes are alike, while in others they 

 are very differently colored ; but it seems that Mr. Seebohm attaches 

 no importance to this as a character of generic value. We cannot but 

 indorse this view, being much surprised, however, to lind that he makes 

 this difference in coloration between the two sexes the chief, not to say 

 the only, distinctive mark of the genera Turdus and Merula; in some 

 instances carried out to the utmost, while on the other hand several 

 species are included in Merula which have the sexes colored alike, and 

 other species showing not unimportant differences between male and 

 female are placed within the genus Turdus. In the one case the char- 

 acter is the only valuable one; in the other, again, it has no value at all! 



Having adopted the singular theory that structural characters did 

 not indicate natural relationship, while pattern of coloration was suffi- 

 cient for the purpose, the author has given us a right to expect as the 

 result of his investigations a more natural arrangement than any pre- 

 ceding it. Unfortunately, howcAer, it must be said that he has not 

 succeeded therein, for his own theory is so often and so violently ignored 

 that most of his genera are quite void of definite limits. 



It is hardly likely that any«one, be he ever so deeply enamored with 

 the coloration theory, will consider it as according with natural affini- 

 ties to arrange T. 7ummis^ u-ardii,pinicola, and sihiricus together in one, 

 and T. maranonicus, dryas, and inlaris in another subgeneric group, 

 when, at the same time, such birds as T. inJaris and torquatus were sep- 

 arated geiierically. And as the natural relationshii> in these cases has 

 been violated, so also have they in many others. 



It being thus evident that the new mode of defining the genera does 

 not lead to a more natural system than the rejected structural charac- 

 ters, it is to be doubly deplored that the generic groups resulting from 

 its application are so indefinite and their limits so unstable, that IVlr. 

 Seebohm (p. 14) needs to appeal to "the instinct of the ornitholog- 

 ical student," when he has not been "able to define the character of 

 each genus." This instinct may in most cases be sufficient to "tell a 

 Chat from a Redstart," but certainly it will be of no use when he shall 

 separate a Turdus Seeb. from a Merula of the same author. The exam- 

 l»le of Mr. Seebohm himself proves that this instinct is often misleading. 



The (coloration and the pattern of coloration may, in many cases, be of 

 very great value as indicating the relationship, but used as a distinctive 

 mark for defining genera in the manner of Mr. Seebohm, who often 

 only takes in consideration the colors of the male, it seems to me to 

 have no scientific value at all. 



It is an objection against the theory of coloration that in many genera 



