i8 9 5. THE STRUCTURE OF ARCHMOPTERYX. 119 



femur it is impossible to measure the exact length in the Berlin 

 specimen. The numbers of vertebrae in the tails differ in the two 

 specimens, so that it will not be safe to take the ratio in length of the 

 two tails as a guide. There is no other bone which can be iden- 

 tified and measured with certainty in both specimens, so we may 

 adopt 8 : 9 as the relative sizes of the Berlin and the London specimens 

 respectively. 



But in thickness a different relation holds. In corresponding 

 bones of two similar animals we find that the ratio of thickness to 

 length is always greater in the larger animal. And this is true here : 

 all bones of the London specimen are stouter and more massive than 

 those of the Berlin specimen. Now, in the London specimen two 

 conspicuous bones (badly shown at cv. in Fig. 2) were identified by 

 Owen as the " third" and " fourth " metacarpals. They measure 39 

 and 33-5 mm. respectively in length. In thickness they are much 

 greater than any hand-bone of the Berlin specimen. Others have 

 regarded these bones as the second and third metacarpals. Suppose 

 this were the case, then we get these ratios between the London and 

 Berlin specimens. 39 : 27-5 = more than 11:8 and 33-5 : 26 which is 

 10-3 : 8. Further, the bones are utterly unlike their supposed equiva- 

 lents in the Berlin slab. They are far stouter, and the longer of the 

 two is exceedingly broad at the base, and therein is well-fitted to resist 

 torsional stress or twist at the joint. In their proximal halves, instead 

 of being slender and almost circular in section, they are stout and have 

 ridges which, when the two were fitted together, would have prevented 

 their movement one on the other. Whatever they are, they are 

 utterly unlike any bones visible in the Berlin specimen. Their position 

 with reference to the feathers of the wing, in spite of the dislocation 

 of other bones, is just that of the large metacarpals in an ordinary 

 bird's wing ; and the fact that these feathers are still in their normal 

 position in this wing (the left) justifies the belief that when the animal 

 finally settled down previous to fossilisation those feathers were still 

 bound to those metacarpals by ligament. 



This is proof no. 1 that those two bones are the metacarpals of 

 the digits IV and V. 



(2.) The second proof is a more formidable one. Some hundreds 

 of experiments extending over hundreds and even thousands of years 

 have shown the effect of " selection " upon dogs, horses, sheep, pigs, 

 pigeons, poultry, vines, roses, plums, apples, pears, strawberries, 

 gooseberries, blackberries, pansies, daisies, dahlias, chrysanthemums, 

 etc., etc., and the result is the same in all cases. Selection occurs in 

 Nature (Naudin, Darwin, Wallace), and its effect is the same as in 

 the case of artificial selection (Naudin, Darwin, Wallace, Bates, and 

 others). I do not think it necessary to repeat the proof of this state- 

 ment here : the proof is far too long, too well known, and too 

 widely accepted for me to need to say more about it. If anybody 

 should challenge me to do so, I will give a proof of it in a future 



