578 Messrs. G. M. Mathews and T. Iredale on some 



It should be observed tliat the figure of Procellaria tethys 

 Bp. in the Monograph shows a square tail ; this is wrong, 

 and though the specimen from which the plate was prepared 

 is in poor condition and there is a good excuse for the 

 artist so drawing it, Salvin wrote, from examination of the 

 same skin, that the tail was forked. 



The tail in '' Oceanodroma" castro is not '''perceptibly" 

 or, as written in the ' Catalogue,' "distinctly" forked. At 

 the best it would be described as ''emarginate," while some 

 specimens show little emargiuation even. Further, the 

 tarsus is distinctly longer than the middle toe and claw, so 

 that if the definitions in the ' Monograph ' or ' Catalogue ' 

 were of any value it ivould be a typical Procellaria = Hydro- 

 hates = Thalassidroma. 



Nevertheless we conclude that phylogenetically it is a 

 Cymochorea. 



When Cones separated Cymochorea and restricted Oceano- 

 droma to f areata he included with that species hornhyi Gray. 

 This bird was unknown to him save by description, and he 

 followed Bonaparte, though remarking upon the different 

 coloration. The type-specimen is still unique and is one of 

 the most puzzling Petrels we have seen. It differs absolutely 

 in coloration from Oceanodroma or the dusky Petrels of the 

 Hydrubatee-Halucyptena-Cymochorea-^rovi-^. It recalls to us 

 a similar anomalous form from the ^o\xi\\, Pelagodroma marina 

 Latham. In this case structural difi'crences are co-existent 

 Avith the strange coloration, and consequently the genus 

 Pelagodroma is recognised by all workers. The species 

 hornbyi has just as distinctive coloration, and as it does not 

 correlate at all with any of the other members of the 

 family, we propose for it the generic name 



Bannermania, gen. nov. 

 We would point out that the unique specimen differs in 

 wing formula, having the second primary noticeably the 

 longest, the first equal to the third and much longer than 

 the fourth. The value of this from one specimen alone we 

 cannot exactly gauge : tlie skin has been unmounted, and 



