STRUCTURE AND HABITS OP ARCHJEOPTERYX. 275 



And again, the digits I, II, III of Archcsopteryx (which the large 

 size and perfectly ossified bones show to be an adult, as also do the 

 well-developed feathers) are, by virtue of their very great slenderness 

 and narrowness at the joints, incapable of resisting a great torsional 

 stress. Unless those feathers exert a great torsional stress on the 

 bones supporting them they are useless. I have shown they were 

 not useless. Therefore they exerted a great torsional stress, and 

 therefore they were supported by bones not yet seen in the Berlin 

 specimen, although those of the left wing are seen in the London 

 specimen. It follows, therefore, that the first three digits were used 

 for climbing, and that one or more others were present to support the 

 feathers. 



(3.) The third proof is incomplete. It shows only that the digits 

 I, II, and III did not support the feathers, and that, therefore, some- 

 thing else must have existed to do so. Its simplicity is unsurpassed. 

 It will appeal even to those who ignore both the principles of mechanics 

 and the actiou of Natural Selection. The figure 10 is placed on the 

 surface of the right wing in Plate XV. In this region the dorsal surface 

 of the wing is convex. A rule or " straight-edge " placed on the 

 wing across this point, parallel to the ulna and resting upon the first 

 and second digits, touches the wing along the whole of its length from 

 number 10 backwards. In front of this the feather-surface curves 

 downwards, so as to be perhaps 2 mm. below the edge of the rule 

 near the digits. The loiver surface of the metacarpal and of the first 

 and second phalanges of the second digit lies fully 1 mm. above that 

 feather-clad surface. The bones of the third* digit are closely pressed 

 down upon, but not sunk below, that surface. Therefore those digits 

 did not lie in but upon the feathered wing when that animal finally 

 sank dead upon the mud in which it has been preserved. Therefore, 

 further, other bones (or bone) were present to support those feathers. 

 No argument from embryology will shake that conclusion. 



The pelvis is seen in the London specimen only, and in this 

 specimen nothing is to be learnt from the left innominate, while even 

 the right one is imperfect. This innominate appears to have been 

 about 50 mm. long. The acetabulum is perforate. I believe there is 

 no anti-trochanter, though in absence of the specimen I would not 

 make the statement definite. It is characteristically avian pelvis so 



