( 245 ) 



bnt not Calamocichla iiewtoni from Madagascar, which has the tail not about eijiial 

 in length to the wing, bat always much longer, the length of the tail exceeding 

 that of the wing by about 8 — 13 mm., in adult specimens. The tail of G. newtoni 

 is more graduated, so that the bird is intermediate in structural characters between 

 Calamocichla and Nesillds Obcrh. {= Klli.^id Hartl.), while the streaks on the 

 throat, being always present, give the bird a resemblance to Bradijpterm 

 braclii/pterus. I therefore think that the bird cannot be included in the genus 

 Calamocichla, and accordingly create for it the genus 



Hemiellisia gen. no v. 



Type : Calamoherpe newtoni Hartl. 



Characters : Similar in external character to Calamocichla, but the tail fiir 

 more graduated, decidedly longer than the wing, though the difference in length 

 is not so great as in Nesillas. 



The other species remain united under the generic term Calamocichla, of 

 which the type is Calamodijta brecipennis Keulemaus. 



The Genus Calamocichla may shortly be described as follows : Similar in all 

 characters to Acrocephalus, but the first primary well developed and not rudimentary. 

 Wing rounded, especially in brcvipennis, more pointed in gracilirostris. Fourth 

 or filth i|uill longest, or both equal in lengtli. First (luill about half the length 

 of the second, being a little shorter in gracilirostris and ziiliiensis. 



Distribution : Tropical Africa, from Capetown to Northern Abyssinia * in the 

 East, and to Fernando Po in the West, and Cape Verde Islands. 



I do not pretend that the results arrived at in this revision are in every way 

 satisfactory. I cannot even say whether two of the species of Calamocichla occur 

 side by side in one and the same geographical province, the two species which 

 appear to occur together in East Africa, viz. parca and leptorhyncha, being the 

 most difiicult to distinguish, as the female of the former is extremely similar to 

 the male of the latter. 



I found that the length of the hindclaw, as compared with that of the wing, 

 is a good character by which the different species can be distinguished. I have 

 omitted to give the length of the culmen, which is difficult to take in these birds, 

 and is seldom measured in the same way by different ornithologists. 



The length of the hindclaw is taken in a straight line from the basal ventral 

 edge to the tip. 



Calamocichla ansorgei ansorgei Hart. 



Hartert, Bidl. B. 0. C. vol. xvi. (190G) p. b->. 



cJ. Upper-surface olive, somewhat lighter and more brownish on the rump and 

 upper tail-coverts. Under-surface pale olive-grey, upper throat, middle of abdomen, 

 and under tail-coverts whitish. 



Wing 82 ; tail 82 ; tarsus 30J ; hindclaw lU mm. 



Ilab. North Angola. 



Only one specimen known, collected by Dr. W. J. Ansorge at Dmiue de 

 Braganza, 



* The exact locality of the only specimen of le/ilorhi/ncJia collected by Wilke in Northern Abyssinia 

 (Berl. Mus.) is not known. The specimen geographically nearest is the one from Lake Zwai (Brit 

 Mus.) which is referred by me to C. ^avva. 



