THE BIRDS OF AUSTRALIA. 
Saunders’s ideas of immature birds of this group seem puzzling, as can 
be best understood by reference to a footnote in the Cat. Birds Brit. Mus.^ 
Vol. XXV., p. 70, 1896. Saunders placed in the synonymy of Sterna dougallii 
the reference “ ? Larus polo-candor Sparrman, II., fasc. III., No. 83, 1788,” and 
starred it for a footnote, which reads : “A very young bird, little more than 
a fledgling ; the white edges to the primaries indicate that it belongs to this 
species or to S. melanauchen.^^ 
I do not think any modern student who refers to Sparrman’ s figure and 
description, will look anywhere near either S. dougallii or S. melanauchen to 
find Sparrman’s bird. 
Though there is not much variation in colour in this species, and very 
little difference in wing-measurements, the Australian breeding birds are 
undoubtedly smaller in the wings, and have, moreover, longer bills. A long 
series from the Andamans agree fairly with Sumatran birds, and a few from 
the Celebes agree in that the bill never measures more than 35 mm., while 
the wing goes over 225 mm. 
Australian breeding birds have biUs over 37 — mostly 39, while the wing 
never reaches 220 mm. 
At present I can only recognise 
Sterna swnatrana sumatrana Raffles ; Andamans, Sumatra to Celebes. 
S. melanauchen Temminck is a synonym. 
Sterna smnatrana hempi Mathews ; North-east Australia. 
Specimens from Foochow may represent another race, while birds from 
Fiji, Pelew Islands, and Phoenix Island do not seem easily referable to the 
North-east Australian form. 
In the synonymy of this species Saunders {Proc. Zool. Soc. (Lond.) 1876, 
p. 661) included “ Gygis, sp. ? et Gygis decorata Hartlaub, Ibis, 1864, p. 232 : 
Godeffroy’s Cat. I. (1864), p. 5.” In the first place Gygis sp. occurs in a list 
of birds from the Feejee Islands, where also a description is given which 
undoubtedly refers to this bird. If this description is named Gygis decorata 
in the second place, to which quotation I have not access, Hartlaub’s name 
win be available for the Fijian form when series are at hand to determine 
the subspecific characters. 
372 
