4?3 CONCLUSIONS 



least remarkable feature of it is the great change in function 

 which has taken place from an organ connected with the 

 ciliary method of feeding to a ductless gland regulating the 

 metabolism of the body. This case is a good illustration of 

 the fact that function is no criterion whatever in questions of 

 homology, and that the sole condition which organs must 

 fulfil to be homologous is to be descended from one and the 

 same representative in a common ancestor. 



A fact which the vertebrates illustrate well, is that the 

 numerical correspondence of segments which give rise to 

 particular structures is not a necessary criterion for homology. 

 This is well shown by a consideration of the pectoral and 

 pelvic limbs. The fore limb is formed from trunk-segments 

 2, 3, 4, and 5 in the newt (Salamandra), whereas in the lizard it 

 arises from segments 6, 7, 8, and 9. Similarly the hind limb 

 arises from segments 16, 17, and 18 in the newt, but segments 

 26 to 31 in the lizard. Countless similar examples are afforded 

 by other vertebrates, and it is to be noticed that the limbs 

 not only vary in their position, but also in the number of 

 segments which have contributed to their formation. Yet 

 wherever they may be and however many or few segments 

 they may contain, fore limbs are homologous throughout the 

 vertebrates, and so are hind limbs. During evolution trans- 

 position has occurred ; new adjacent segments have taken to 

 contributing to the formation of the limb, and at the opposite 

 end segments which hitherto contributed may cease to do so. 

 In this way the limbs may become transposed over the trunk 

 of the animal much as a tune can be transposed over the keys. 

 But it is the same tune and the same limb. 



Another case is that of the position of the occipital arch at 

 the back of the skull. The neurocranium of Scyllium occupies 

 7 segments while that of a form as closely related to it as 

 Squalus occupies 9. Although they are situated in different 

 segments, there can be no doubt that the occipital arches of 

 these two animals are descended from the occipital arch of a 

 common ancestor, and are therefore homologous. 



A very interesting example of the same kind is furnished by 

 the number of gill-slits in various Selachii. Heptanchus has 



