104 TROPICAL WILD LI IK IN BHITLSH (iUIANA 



peter or parrakeet or red-breasted l)lackl)ird was an unthink- 

 able phenomenon, although at least in the case of the first 

 two, the birds remain paired for life. 



There is another type of sociability, that between dif- 

 ferent sj^ecies, and in the jungle this was one of the most 

 common phenomena and has been noticed by almost every- 

 one who has spent any time there. The hosts of species of 

 small and medium sized birds drifted together when their 

 nesting season was past and roamed the jungle in small 

 bands. It was remarkable how many different kinds were 

 to be found in each little gathering. Sometimes when such 

 a flock worked toward and across a glade it was possible to 

 make an approximately complete census. I have comited 

 twenty-eight birds in a flock of this kind, including twenty- 

 three distinct species. The association reminded me strongly 

 of birds migrating at night. There was the same steady 

 drift in one direction and the same constant intercourse by 

 means of soft chirps and twitters, woodhewer calling to fly- 

 catcher, and manakin to antbird. 



Reviewing the whole host of Bartica birds, w^e are im- 

 pressed with the tremendous extremes, not even approached 

 by the avifauna of a corresponding temperate region. In 

 size, for example, our specimens ranged from the pygmy 

 amethyst hummingbird less than three inches from beak to 

 tail, which probed the corollas for tiniest of insect food, to 

 the great harpy eagle over a yard in length, with talons ca- 

 pable of striking down any sloth or monkey of the jungle. 



A word as to color, more to indicate our line of investi- 

 gation in this direction than to provide any satisfactory the- 

 sis for the solution of this tremendously interesting factor of 

 life. Working with the same methods with which I judged 

 of protective coloration among the pheasants, I was able to 

 divide the Bartica birds into graded groups with a fair de- 

 gree of assurance. My estimate of protective coloration was 

 based on tlie action of the birds themselves at the approach of 

 danger. A bird Mhich flew at once, either to a point of van- 



