EXPEDITION TO FUNK ISLAND. 519 



uiuth, pair of ribs, and one had an extra rib on one side only. In one 

 case the ninth pair of ribs was completely fused with the sacrum, show- 

 ing not the slightest trace of former articulation, although the eighth 

 pair was free. Twenty-three sacra also, but not those just noted, were 

 composed of fifteen vertebne, owing to the inclusion of the seventh 

 dorsal by anchylosis. Of course the number of caudals included in the 

 sacral mass may be one less or one more, this depending largely on the 

 age of the individual; but the normal -'sacrum" is composed of five 

 presacrals, three true sacrals, and six uro-sacrals. 



The Great Auk has the characteristic alcine sternum, and this is 

 subject to the same variations found in other members of the family, 

 the xiphoidal extremity being entire, perforate or notched. 



These differences are largely due to age, notches becoming trans- 

 formed into foramina, and foramina being obliterated by the extension 

 of ossification into the membrane filling the vacuities. The same thing 

 may be seen to some extent in the Loons, and the character of the 

 xiphoid extremity of the sternum has by no means the same value among 

 the generalized water birds that it has among the more specialized 

 Passeres. 



The extra ribs furnish another case in point, for although variation 

 in the number of ribs is liable to occur among mammals, and is not in- 

 frequent in passerine birds, such variation would naturally be oftener 

 found in more generalized forms. 



The greater the number of similar skeletal parts and the more gen- 

 eralized th.e form the greater would seem the tendency to variation. 

 Thus among the Urodele Batrachia the number of presacral vertebras 

 is not specifically constant, and so frequent are variations in the post- 

 sacral region that total number of vertebrae is almost of no importance 

 even as a specific character. 



Fishes might seem to offer an exception to this for in the families Ser- 

 ratiidce, Sparidce and Caraiigidw the number of abdominal and caudal ver- 

 tebras is not only remarkably constant for the species, but even for the 

 family. But these fishes have a comparatively small number of verte- 

 bral segments (generally ten to fourteen), while in other members 

 of the same superfamily, groups in which the number of vertebrae is 

 greater, the amount of variation is greater. And the specialized Aca- 

 thopterygians mentioned stand in relation to other fishes much as the 

 Passeres do to other birds.* Occasionally eight pairs of ribs were con- 

 nected with the costal margin of the sternum, the normal number thus 

 attached being seven. 



It would be interesting to know if the sterna with eight pairs of cos- 

 tal facets belonged to the birds, with nine pairs of ribs, but this ques- 

 tion can not be answered, although from one or two similar instances 

 that have come under my notice among other birds it is quite probable 

 that this was the case. 



* For the above facts pertaining to fishes I am indebted to Dr. Theo, Gill. 



