VOl 'lfi^ XI1 ] Murphy, Birds of Trinidad Islet. 343 



" Rio Plata " to be the type locality given by Gmelin, Syst. Nat. I, 1789, 

 p. 565. In this citation an error has evidently been made, for Gmelin, 

 nowhere mentioning Rio Plata, simply follows Latham, Synopsis, III, 

 1785, p. 400, where it is stated that the species " Inhabits Turtle and 

 Christmas Islands." Pending further light on the subject, M. alba must 

 therefore continue to stand as a doubtful synonym of JE. neglecta. 



We first saw JEstrelata arminjoniana on April 4, 1913, in latitude 25° S., 

 longitude 30° 40' W., nearly four degrees south of Trinidad. This species 

 became increasingly common on the fifth, sixth, and seventh of the month, 

 as we approached its headquarters. JE. trinitatis was seen only on April 8, 

 in the immediate vicinity of the island, and neither bird was noted again, 

 although I kept a sharp lookout for several days after we had proceeded 

 on our northward journey. 



The validity of the species trinitatis has long been questioned, Salvin 

 (Cat. B. Brit. Mus.) and others considering it merely a dark form of 

 arminjoniana. It has lately been noticed by both Wilson and Nicoll that 

 trinitatis breeds at a higher altitude on Trinidad than arminjoniana, and this 

 fact, together with the apparently constant color differences of plumage 

 and feet, is the present warrant for granting specific distinction to the two 

 birds. The evolution of several well-differentiated representatives of one 

 circumscribed section of a genus, whether they be true species or merely 

 color phases, is an interesting and rather common phenomenon in the 

 genus JEstrelata (viz. jamaicensis and neglecta), and indeed among Tubinares 

 in general. As regards the relationships of the petrels inhabiting the small 

 oceanic island of Trinidad, it is not improbable that the parti-colored forms 

 (arminjoniana, " wilsoni," chionophara) are of relatively recent origin, and 

 that this small group of birds is still specifically unstable. Such a hypo- 

 thesis can at least be made to fit the facts, although a final decision must be 

 reserved until a large series of specimens, representing every stage of growth, 

 can be studied. Assuming the uniformly colored trinitatis to be nearest the 

 parental stem, I find that I can formulate the following progressive arrange- 

 ment, partly on the basis of my own specimens, partly on published in- 

 formation: 



a. Downy young of all the species, so far as known, dark gray. 



b. trinitatis, immature. Bill black; tarsi and feet black; plumage uni- 

 formly blackish-brown; concealed portions of the feathers light gray with 

 dark shafts. 



c. trinitatis, adult. The same, except that the concealed portions of the 

 feathers are pure white, including the shafts. 



d. arminjoniana, immature (dark "wilsoni" phase). Bill black; 

 " tarsi and basal half of the toes very dark brown " (Nicoll), distal half of 

 toes black; dorsal plumage like that of trinitatis; breast dark gray. 



e. arminjoniana, older than the last but not fully mature — or possibly 

 the mature bird in fresh plumage? (light " wilsoni " phase). Bill black; 

 tarsi, and basal half of web and two inner toes, flesh color; distal half of 

 foot black; breast more or less dark, sometimes showing only on the ex- 



