THE BIRDS OE AUSTRALIA. 
noticeable. The head is bluish-green, not greenish-brown ; the back is also 
bluish-green, and the primaries edged with indigo. Tail blue. 
Mud Island, near Brisbane, Queensland.” 
The range is most peculiar and complex, as this recent recognition of a 
form as far south as Brisbane on the east coast was balanced by the discovery 
of another at the Monte Bello Islands on the west coast ; while Carter records 
it from the North-West Cape. It was known from the Aru Islands, and 
recently Ogilvie-Grant recorded it from south-west New Guinea, writing : “As 
might be expected, New Guinea examples of H, sordidus belong to the 
larger race of this species, which ranges from the Aru Islands to the 
Monte Bello Islands, Melville Island, North Australia and Queensland. 
It was previously unknown from New Guinea. In the birds from the 
Mimika River, the wing measurements are respectively: — (a) ^ 110; (6) 
d 107 ; (c) $ 104 mm. Dr. Hartert records a specimen from Dobbo, Aru 
Islands, with a wing measurement of 116 mm.; while in the British 
Museum Collection the largest examples are from the Hope Islands, off 
Cape Tribulation, East Queensland, a male and female collected there 
by Macgillivray measuring respectively 122 and 115 mm. In some specimens, 
however, the wing measurement is much smaller, and two from the 
Aru Islands measure 103 mm. The average measurement of sixteen 
specimens is 106 mm. The small race, which has been named H. s. 
colonus Hartert inhabits the Louxsiade Archipelago. In seven specimens the 
wing measurement varies from 86 to 93 mm., the average being 91*5 mm.” 
The small size and very different coloration entitle Hartert’s colonus 
to specific rank, while the variation between the specimens from Australia, 
save colcloughi as noted, is slight as regards colour and size. For 
the present, I am not recognising any Australian subspecies, except 
colcloughi, but the blue coloration of this form is very remarkable and 
makes the bird look like a distinct species. When better series are obtained 
it may be necessary to reinstate the subspecies here rejected, as has been 
the case in the congener S. sancta. I was in the same predicament 
with good series of the latter, but extended series indicated the sub- 
specific features. In an old genus hke the present one small differences 
constitute subspecific forms, but long series are necessary to recognise these, 
as the birds show their plumage changes slowly. 
I therefore admit : 
Sauropatis sordida sordida (Gould). Northern Australia, from 
Cooktown on the east to the North-West Cape on the west. 
Sauropatis sordida colcloughi (Mathews). South Queensland. 
196 
