Annals of the Transvaal Museum. 
251 
also have been though Marais, who collected these specimens, was 
always very careful in his sexing. It remains, therefore, to be seen 
whether this difference in size is constant. Grant doubts Shortridge’s and 
Davies’s records of “ viridis ” from Pondoland, stating that he would 
not be surprised to find that the specimens would turn out to be I. e. 
marwitzi. Shortridge’s specimens I have not seen, but Davies’s from 
Flagstaff are certainly referable to “ viridis .” A single specimen from Port 
St. Johns District, taken by Swinny, is much smaller than even juvenile 
specimens from Grahamstown ; it is sexed as a and I have no doubt 
from a knowledge of the conscientious work of this naturalist, and the 
general smallness of birds from Pondoland, that this is correct. A small 
race may therefore be found near the coast in Pondoland. 
|| \i ]At p. 285, Grant makes the following statement : — ££ Irrisor erythro- 
rhynchus marwitzi Reichw., Orn. Monatsb., 1906, p. 171 : Makalama in 
the Wemberesteppe, German East Africa. 
££ Of this, Irrisor erythrorhynchus brevirostris Gunn, and Rob. (Annals 
of the Transvaal Museum , Yol. Ill, p. 113 : Villa Pereira, Boror, Portuguese 
East Africa ; see also Journ. S.A. Orn. Union , Yol. VIII, 1912, p. 26 
becomes a synonym, as birds from Portuguese East Africa agree perfectly 
with specimens from the north and south, and the bill is no£ shorter.” 
p |i jThe range he gives as: ££ Natal, Swaziland, Transvaal, Portuguese 
East Africa, Matabeleland, Mashonaland, Nyasaland, Northern Rhodesia, 
Belgian Congo (Dikulwe Valley), German East Africa (Mombasa to Rift 
Valley), Uganda up to the Turkwel River, island of Zanzibar.” 
At p. 281, he records a £ specimen from Amala River, 200 miles north 
from the type locality of ££ marwitzi ,” and a $ from Turkwel River, still 
another 200 miles further north ; all three localities are on different river 
systems, an important detail where these birds are concerned, owing to 
their frequenting heavily wooded places such as are commonly distributed 
along river banks. I mention this because the measurements he records 
of these two specimens show that the A has a long wing and a short bill 
and the $ a long bill and a short wing, and unless we take account of such 
differences we lose sight of the main object of trinomial classification. It 
is this want of appreciation of details which concern geographical 
distribution which has caused Grant to reject the name of ££ brevirostris .” 
If it is possible to separate Meyer’s parrot only by the study of series of 
specimens, why should not the same trouble be taken in regard to these 
Hoopoes, which present more clearly defined characters on which they can 
be separated ? I am not aware on what grounds Grant states that the bill 
of ££ brevirostris ” is not shorter ; but the series in the Trans\aal Museum 
Collection certainly shows this to be wrong. In the table of measurements 
given hereafter of the specimens in the collection before me, it will be 
seen that both the wing and tail, besides the bill, are shorter in specimens 
from the Zambesi valley than in those from the Transvaal, whfe they 
approximate to those of the eastern parts of the Cape Province. This 
I noted before and therefore compared the race with ££ viridis ,” stating that 
the bill was stouter and straighter. Specimens from the Transvaal appear 
to come nearer to I. e. angolensis than to I. e. marwitzi , to judge by Grant’s 
short diagnoses. The prismatic colours of these birds are not easily 
