370 BULLETIN 15 3, UNITED STATES NATIONAL MUSEUM 



mens from Somaliland. It is true that the measurements average 

 very slightly larger in the northern part of the range, but the differ- 

 ence is so slight, the overlapping so extensive, and the total bulk of 

 material available for study so meagre that it Avould be wholly un- 

 justifiable to split the species into two subspecies. The measure- 

 ments of the type specimen, as given by Oustalet ^^ are larger than of 

 any specimen I have seen or than any reported by Erlanger or 

 Zedlitz. The type measurements are : Wing 80, tail 78, and culmen 

 29 millimeters. The last probably refers to the exposed culmen, not 

 the total length, as the other figures indicate a large specimen. The 

 colored figure (pi. i) in Oustalet's paper is not very accurate. The 

 rufous tawny of the underparts extends too far up the breast (in fact 

 even the chin is represented as tawny) ; the same color is far more 

 extensive and noticeable on the back in the figure than in any bird 

 examined, the suj)erciliaries are represented as being wider, more 

 prominent, and bluer than they actually are, and the middle rectrix 

 is figured as having a brilliant blue median portion laterally bordered 

 with green, while in no specimen is this true. The rectrices vary 

 somewhat in their relative blueness and greenness, but never ap- 

 proach the condition shown in the plate. The same is true for the 

 alula and outer secondaries, which are far too blue in the figure. 



The female taken at Hor on June 28 and the male from Endoto 

 Mountains, July 20, were molting the tail feathers when collected. 

 The material is too slight to enable me to work out the order of the 

 molt, but it seems to be irregular — which means that it is worth study- 

 ing as all irregularities turn out, on proper investigation, to be modi- 

 fied regularities, which often suggest clues as to the factors involved 

 in regulating ecclysis. 



This bee eater appears to be quite common in most of its range. 

 Lonnberg ^® found it to be rather common in the thornbush country 

 north of the Guaso Nj'iro, but never saw it in flocks, single birds 

 being the rule. Likewise, Erlanger ^° recorded it as very common in 

 the steppes and thorn veldt of southern Somaliland. Considering 

 its high numerical status within its distributional area, the limita- 

 tions of that area appear all the more interesting. Its southern 

 limit is abrupt ; there is no gradual diminution in numbers resulting 

 finally in complete absence of the species as one goes southward from 

 Ethiopia to the Guaso Nyiro. Lonnberg ^^ definitely states that as 

 soon as he had passed over to the northern side of the Guaso Nyiro 

 he met with this little bee eater. Similarly, in the northern part of 

 its range, Elliot *" writes that in Somaliland it was not met with 

 before reaching the plateau south of the Goolis range. 



" Miss. R6voil Qomalis, Ois., p. 5, 1882. 



ssKunjil. Sv. Vet. Akad. Handler., vol. 47, 1911, p. 74. 



3»Journ. f. Ornith., 1905, p. 455. 



"Field Colum. Mus. publ. 17, Oruith. series, vol. 1, No. 2, 1807, p. 55. 



