92 BULLETIN 50, UNITED STATES NATIONAL MUSEUM. 



2. Throat streaked, head blackish (at least in adult males). 



P. migratorius (North America). 



P. confinis (Lower California). 



P. magellanicus (southern South America). 



P.falklandicus (Falkland Islands). 



3. Throat streaked, head not black; chest grayish; abdomen, or sides and 



flanks, fulvous. 

 P. flavirostris (Mexico). 

 P. rufiventris (Brazil, etc.) 

 P. olivaceus (Africa). 



4. Throat streaked, head not black; a white patch on lower throat; no fulvous 



on under parts. 

 P. jainaicensis (Jamaica). 



P. tristis and subspecies (Mexico and Central America). 

 P. phxopygus and subspecies (South America). 



5. Throat streaked, head not black; no white patch on lower throat; color 



plain grayish or brownish, paler below. 

 P. grayi (Mexico and Central America). 

 P. obsoletus (Central America). 

 P. plebejus (Central America). 



P. leucomelas, P. albivcntris, and allies (South America). 

 P. nigrirostris (Lesser Antilles). 



P. gymnophthalmus (Lesser Antilles and northern South America). 

 P.fumigatus (northern South America). 

 d. Throat not streaked, the general coloration uniform sooty brown or dusky. 

 P. nigrescens (Central America). 

 P. gigas and allies (Andean district). 

 P.fuscatra, P. cacozela, and allies (Andean district). 



That these color-diflferences are of only secondary value in classifi- 

 cajtion, so far as this group is concerned, is shown by the fact that 

 very often the species most similar in color show the greatest struc- 

 tural differences, for example: P. jamaicensis compared with P. tristis, 

 P. iiihseopygus, and allies; P. nigrescens compared with P. gigas and 

 related forms; and P. gymno'pMhalmus compared with P. alhiventris, 

 P. leucomelas, etc. 



Of all the American species none approach very closely in certain 

 characters, especially the exceedingly compressed bill, long gonys 

 (nearly twice as long as mandibular rami), and arched mesorhinium, 

 to P. merula, and I am not at all certain that they are really congeneric. 

 Probably a more thorough study of the group than I am at present 

 able to give it would develop characters justifying the separation of 

 all the American forms, together with at least some of the African 

 ones, from P. merula and most of the other Old World forms, either 

 as a single generic group or several. P. fumigatus almost certainly 

 should be separated on account of the much greater adhesion of the 

 anterior toes, no other thrush that I have been able to examine hav- 

 ing practically the entire length of the basal phalanx of the middle 

 toe united to the outer toe. 



