STRUCTURE OF FOSSIL BONE. 67 



morant, is said to have no predilection even for swimming ; 

 as, however, the force by which it descends, evinced by the 

 cloud of spray it scatters around, must have the primary 

 effect of carrying the bird to a great depth, unless resisted 

 by muscular power, aided by the buoyancy of its body, its 

 legs become powerful instruments in bringing the gannet to 

 the surface so soon as it has secured (as it never fails to do) 

 its prey. The bill of the bird is strong, without clefts for the . 

 nostrils ; the furcula strong and bowed out, and the portion 

 especially that meets the resistance of the sea is bent down 

 at about a right angle with the coracoid, and it is there ex- 

 cessively strong, with a powerful articulation and a pointed 

 process, extending for half an inch on the inside — taking 

 the furcula, coracoids, and breast-bone together^ they form 

 an admirable piece of machinery, adapted to the habits of the 

 bird. The keel of the sternum is small and thrown forwards, 

 so that the outer edge and the fiu-cula constitute an extended 

 arch, very different to the depressed character of the furcula 

 in the shag ; the keel of the breast-bone is straight, without 

 any curve, as in the shag. The socket also, where the 

 coracoid fits into the sternum, has a very strong and expanded 

 articulating siu-face, and there is an impression both in the 

 sternum and the coracoid for an oblong ligament, which 

 braces the coracoid strongly to the sternum ; on its outer 

 edge it is not truncated as the shag's, but is shaped like the 

 point of a bill-hook. I will now, however, describe the 

 microscopic structure of some of the principal bones. The 

 humerus, radius, and ulna partake of the characters I have 

 observed in all birds whose wings are long and pointed. In 

 the humerus vertical section (fig. 17), the Haversian tubes 

 do not reticidate, but run nearly parallel, and ultimately 

 converging to a point, from which extends another tube which 

 converges in a similar manner. The lacunar are long, and, for 

 the most part, narrow ; in the transverse section the 

 Haversian tubes appear as round dots, the lacunae as small, 

 irregular specks, and the eanaUculi beautifully reticulate. 

 The ulna compares very nearly with the humerus ; in the 

 radius the Haversian tubes are more numerous, more parallel, 

 with fewer connections ; the lacunse also are rather longer. 

 There is nothing particular to notice in the scapulae, the 

 Haversian tubes obser\dng a pretty similar arrangement ; but 

 the fiu'cula, and especially the coracoids, are worthy of par- 

 ticular notice. 



In the furcula (fig. 18) the Haversian tubes are not so 

 numerous as in the humerus, have no connections, and the 

 lacunae are very fusiform. In the coracoid (fig. 19) a 



