REPORT ON THE AMPHIPODA. XV 



present state of knowledge appear to have the largest number of important affinities. 

 But v/hat affinities should be considered important for classification it is by no means 

 easy to determine. Animals genealogically very wide apart may have adopted similar 

 modes of life, and in so doing have become modified on parallel lines, while on the other 

 hand, in species nearly related by descent, great divergence of character may have resulted 

 from difference of habits, such as the assumption of a parasitic life by one branch of 

 a family, when the other branches have remained independent. In classifying the 

 Gammarina authors have usually placed the Orchestidse first. In the order of evolution they 

 might rather be placed last. Among these alone of the Amphipoda has a capacity for 

 terrestrial existence been acquired ; some of them are gradually adding the faculty of 

 walking upon dry land to the ordinary movements of sliddering and leaping ; all of them 

 have lost the mandibular palp. Delage, founding his view upon the circulatory 

 apparatus,' suggests that the Corophidse are the ancestors in common of the other 

 Gammarina and the Caprellidse. But Corophium volutator (Pallas), the subject of 

 Delage's investigation, is far removed from a typical Amphipod. Though it has not 

 the variety of movement found among the Orchestidge, yet, by having a Ijody 

 flattened instead of laterally compressed, it is perfectly capable of walking. It cannot 

 perhaps, strictly speaking, be said to walk upon dry land, Init it walks freely over moist 

 mud in the open air. Of the three pairs of lateral orifices to the heart, so generally found 

 among the Amphipoda, Delage has observed that the first two pairs are wanting in 

 Coro2)hium vohitator, and that they are small and inactive in the Caprellidge. But it 

 may safely be said that if the Gammarina and Caprellidse were descended in separate 

 lines from the Corophium, the degraded and inert Caprellidae would never have acquired 

 the two additional pairs of orifices for which they have, it seems, no urgent need, and which 

 their supposed ancestor of a higher type and more active habits is able to dispense with. 

 Thus, while the character of the heart makes it very improbable that Coropldum should 

 have been an ancestor of the Caprellidse, its shape and habits make it quite as unlikely 

 that it should have been an ancestor of the Gammarina, so few of which have any 

 activity out of water, and so many of which, the Orchestidse included, have the body 

 laterally compressed. 



On the supposition of a common origin of all the Amphipoda, it is obvious that 

 families will have been gradually separated by the successive acquisition of distinctive 

 characters. The supposition itself is based upon the fact that some characters are 

 common to many families, since that fact is explained most simply on the principle of 

 inheritance from a common ancestor. In the search, then, for ancestral characters, we 

 must look away from what is rare and exceptional to what is commonplace and un- 

 attractive. When any single character is investigated in all the known species, some 

 form will often be found of marked simplicity and completeness, round which the rest 



' See p. 526. 



