Genus— P 0 E P H IL A. 
PoEPHiLA Gould, Birds Austr., pt. vi,, March 1st, 
1842. Type (by subsequent designation) 
id., Proc. Zool. Soc. (Lond.) 1842, p. 18, Nov. Amadina acxiticauda Gould. 
Undeb the genus Erythura I have given a sketch of the variations structurally 
of the Green Finches, and here analyse the Poephila group as arranged in the 
Catalogue Birds British Museum. 
Sharpe separated the group thus : 
“ Tail produced and pointed, the centre feathers elongated, so that 
the distance between the outer and centre feathers is greater than 
the length of the tarsus. 
“ TaU graduated, the two centre feathers longest but not 
produced to a tluead-hke point Uroloncha. 
“ TaU not so strongly graduated, but the centre feathers 
produced to a thread-like point Poephila.’’’’ 
Then as members of the genus Poephila he classed the species acuticauda, 
cincta, personata, leucotis, gouldice and mirabilis. The first four were ahke 
in coloration, the last two (only one) different, and the taU of personata agi’eed 
\vith Sharpe’s diagnosis of his Uroloncha, not of his Poephila. 
I make leucotis and personata to be subspecies oiUy, and with acuticauda 
and cincta these constitute a colom-gi-oup quite distinct from gouldice, and 
if the latter is at all closely aUied it has varied immensely. 
The variation in coloration in the forms acuticauda, cincta and personata 
is almost neghgible, and the bUl formation is very simUar; but AvhUe 
acuticauda and persanata have striking yeUow to red bUls, that of cincta is 
smaller and black. 
The tail of cincta is square with the two central feathers scarcely exceeding 
the others, but the tips attenuate. The taU of cLcuticauda is wedge-shaped, 
aU feathers pointed and central two mth very long, attenuate tips; the next 
two are longer and more attenuated than the central feathers of cincta. The 
taU of personata is wedge-shaped, the central ones thimier, but not hair-hlie. 
While the coloration shows these to be closely aUied forms, had the 
coloration been distinct the structm’al differences would have long ago been 
utihsed as valid for generic separation. In the present work these are recognised 
231 
