SUBORDER PROCELLARHFORMES. 15 



is present and the wing is aquincubital. The pterylosis is discontinuous, as is 

 usual, and not specialised, the downy nestlings having very close down. 



On account of their great powers of flight not many fossil remains of Petrels 

 have yet been recorded. Such a record as Diomedea anglica Lydekker from the 

 London Clay can only be regarded with suspicion, as that author did not discriminate 

 between the various forms of Diomedeoidea, and the fragment is rather indefinite. 



SUBORDER PROCELLARHFORMES. 



The suborder is coequal with the order, being divided into two superfamilies, 

 the Procellarioidea and the Diomedeoidea, externally so different that no confusion 

 is possible, while internally there is almost as much distinction. Alexander has 

 recently given some notes on the Australian species advocating the admission of sight 

 records by good observers and the limits of Australian Seas as reaching half-way to 

 the next land . He has also lumped some generic forms in his essay. We have since 

 had the opportunity of discussing some of the points with him personally, and he 

 admitted that the matter was not so easy as he had previously concluded, and that 

 sight records of all but the most easily recognisable forms might still be doubted. As 

 to the lumping of genera, Garrodia proves to be more closely related to Fregetta than 

 to Oceanites, while Fregettornis is the most distinct of the series. Again Halobcena was 

 included in Prion because he was not familiar with that form, and consequently he 

 might correct that item. 



SUPERFAMILY PROCELLARIOIDEA. 



We divide this superfamily into three families : the first, of small birds with 

 nostrils opening into a single tube, and legs with tarsal covering either reticulated, 

 scutellate or booted ; the second, from medium to very large birds with the nostrils 

 separated by a septum, and tarsal covering always reticulate ; the third, small birds 

 with nostrils separated by a septum but the openings horizontal, and legs with 

 reticulate covering and no hind-toe ; in the two previous a hind-toe is present and 

 the nasal openings are never horizontal. The technical name of the first family 

 is in dispute, but their vernacular equivalent of Storm-Petrels is beyond all con- 

 troversy. It is remarkable that such a compact little group superficially should 

 have been disintegrated by the researches of anatomists, though had any ornitho- 

 logist dared to suggest such a division there would have been more reason. At the 

 present time some ornithologists, who have never studied the group carefully, dis- 

 allow any separation in the order, which is mere nonsense. If the anatomists were 

 correct in the valuation of the characters they deal with even more subdivision than 

 we allow would be made by them. As a matter of fact there seems to be reason in 

 still further subdividing the group. The Storm-Petrels are found more or less in 

 all seas from the Arctic to Antarctic but are more rare in the tropics. Recently an 

 Antarctic breeding form has been credited with a northward migration almost to the 

 Arctic, though otherwise the species are not given to much wandering. In connection 

 with the second subfamily, the Procellariidse, which includes the Fulmars and Shear- 

 waters, erratic movements, which can scarcely be termed migrations as that term 

 is commonly used, are common but certainly not at all known. The third family, 

 the Pelecanoididse, lives only round the Sub-antarctic Circle and the south-west coast 

 of South America, and may not be phylogenetically related at all closely to the 

 preceding families, as both superficially and internally it is very different. 



FAMILY THALASSIDROMID^S. 



Storm-Petrels are the smallest forms of Procellariae, and all have the tubes 

 opening on the top of the bill with superficially a single orifice, the tails square or 



