SMALL DIVING PETREL. 
I think that exsul can only be a subspecies and that it does not occur 
in New Zealand. 
Two skins from Stephen’s Island are bigger than others from New Zealand. 
Wing : 123 female, and 135 male ; tail 42 and 45 ; culmen 17 ; tarsus 27.5 and 
27 ; middle toe and claw 32.5 and 34.5. These are the typical form. 
Five from other parts of New Zealand vary but little inter se. Wing 110-115 
(113); tail 39 ; culmen 14.5 to 16 (15) ; tarsus 23 to 25 (23.6) ; middle toe and 
claw 28.5 to 31 (30.3). 
Buller always maintained that Stephen’s Island birds differed from other 
New Zealand birds. 
The above shows the difference between the larger urinatrix and the sub¬ 
species belcheri ; it is only one of measurements. 
The form that occurs in Australia is not the large bird from Stephen’s 
Island with a wing of 135 mm. in the male, but the smaller bird Avith a wing 
varying between 110 and 125, or even 128. This bird does not appear to be 
different from the form from Chatham Islands ; the names, therefore, for the 
forms, will be :— 
PELECANOIDES URINATRIX URINATRIX (Gmblin). 
Queen Charlotte’s Sound, breeding on Stephen’s Island. 
PELECANOIDES URINATRIX BELCHERI Mathews. 
Australia (East); NeAv Zealand (other than Queen Charlotte’s Sound) ; 
Chatham Island : Eyes broAAm ; bill dark brownish-black ; anterior half of tongue 
blackish, remainder and palate pale fleshy-white; legs and feet light blue ; Avebs 
greyish-black ; claAA r s blackish-broAAm. 
Of this form chathamensis is a synonym. 
Dr. R. C. Murphy has decided that there are no subspecies of some of the 
dark Petrels, iioav that he has been able to see and handle a large series. 
With a blanket description of urinatrix and working on the same lines 
as Murphy, the measurements are the only thing on Avhich A\ r e can rely in this 
bird. It is impossible to differentiate belcheri from chathamensis on measure¬ 
ments ; and the other characters seem to be too uncertain and variable at the 
present stage of our knoAvledge. 
In this bird Loomis accepted a blanket description and wing measurements 
varying from 110 to 135, the average being 117.4. This is one inch in 4f inches. 
Surely this cannot be right ? 
In measuring large series of birds are they all used ? I ask this because 
Avhen I Avas Avorking on the Crows of Australia I had over 150 skins, over tAVO- 
thirds of these could not be used for one scientific reason or another. Some 
103 
