316 VALUE OF COLOR CHARACTERS IN BIRDS. 



exclusively African and so also are the red-breasted species of Merops. In other 

 words no species with red in the plumage occur outside of Africa. Six of the 

 African species of M elittophagus form a well-marked group, with green backs, 

 yellow throats, and a distinct black or blue breast band. (PL XXVII, fig. 5.) 

 In Java and the Indo-Malay region occur respectively M. leschenaulti and 

 M. swinhoiiy which resemble the African group in the plumage of the lower 

 surface, but differ above in having the anterior portion (head to interscapulum) 

 chestnut color, and the rump blue. (PL XXVII, fig. 4.) No M elittophagus 

 occurs in Sumatra or Borneo but in these islands as well as in parts of the Malay 

 peninsula and north to southern China, occurs a Merops with precisely the same 

 distribution of chestnut above but the blue of the rump extending to the tail. 

 Below, however, the throat is blue and there is no breast band. (PL XXVII, fig. 2.) 

 A closely allied species inhabits the Philippines to the east while the European 

 Merops apiaster, whose distribution touches the ranges of these species on the 

 west, is the only other chestnut-backed species of Meropidce. 



It seems to me that this close correspondence of certain types of coloration to 

 certain geographic areas is significant, and I should much rather regard the elon- 

 gation of the central tail feathers as a tendency likely to crop out at any time and 

 place than to consider that the two types of tail structure developed first and that 

 several types of color combinations were later evolved in each group in almost 

 exact parallelism. We might argue that the environment had something to do 

 with the question, resulting in red-breasted or red-throated forms of each genus 

 in Africa and chestnut-backed forms in the Indo-Malay and European countries, 

 but unfortunately for such a theory other species inhabit these regions which 

 have totally different color patterns. Furthermore in West Africa we find side 

 by side several of the most aberrant species of the family so far as color is con- 

 cerned, differing widely from one another both in color pattern and colors. 



In arguing for phylogenetic significance in coloration I do not mean to say 

 that we can divide a family like the Meropidce into definite " color genera" nor 

 can we trace exact lines of descent for all the species. Naturally many con- 

 necting links have become quite extinct so that species like some of those from 

 West Africa have scarcely a single color character in common with any other 

 species of the family. In other groups we find numerous family color characters 

 in all sorts of combination, now one of them being suppressed, now another. 



It does, however, seem to me that color patterns can be used to advantage 

 in testing the real value of so-called structural characters. Where a different 

 color pattern is combined with a difference in structure of the bill as in Cardinahs 

 and Pyrrhuloxia, and especially where, as in this case, the distribution of t e 

 species of each genus is continuous, the evidence that we have a natural, pny °~ 

 genetic division of species into two groups is obvious. On the other hand wnere 

 we find the same color characters occurring in two genera which have bee 

 separated upon a trivial and variable character such as the elongation of 

 central rectrices, the inference is that the genera are artificial and not pny 0- 





