Carpenter — Relationships between Classes of Arthropoda. 335 



remain free from each other and from the head. The abdomen of 

 Nebalia sho-^s a remarkable likeness both to the abdomen of the 

 higher Malacostraca and to that of the Phyllopoda. It agrees with 

 the latter and differs from the former in the reduction of the limbs 

 on the hinder segments and in bearing a terminal fui'ca; while it 

 approaches the Malacostracan abdomen in the limited number of its 

 segments, eight being present, the last two of which are limbless. In 

 the Malacostraca there are seven abdominal segments, the sixth bear- 

 ing a strong pair of appendages. (See Table, pp. 354-5.) 



Having seen that the structure of the cephalic and thoracic limbs 

 in Nebalia leads us to regard it as more primitive than Apus, we are 

 prepai'ed to compare the abdominal region in the two animals, and 

 to admit that the numerous abdominal segments in the latter may well 

 have arisen by the multiplication of a primitively moderate number. 

 On the other hand, if the Malacostraca have developed, as is almost 

 universally believed, from Leptostracan ancestors, it is easy to con- 

 ceive that one abdominal segment has been lost, in connexion probably 

 with the strong development of the Malacostracan ui'opods. If these 

 limbs belong to the true sixth abdominal segment, then there may be 

 two fused segments in the telson ; or, as is perhaps more probable, 

 when the formation of the Insect abdomen is recalled, the uropods 

 may be in reality the limbs of the seventh abdominal segment which 

 has become united with the sixth. Future researches on the embry- 

 ology of the Malacostraca will doubtless clear up this point. 



If we carry our investigation still further back, and speculate as 

 to the nature of the ancestors of the Leptostraca, we naturally com- 

 pare them with that other ancient group — the Trilobita. Beecher's 

 restoration ('00) of the appendages of the Trilobites, as suggested by 

 the study of Triarthus, is now well known. The head bears a pair 

 of simple feelers and four pairs of biramous limbs, not differing from 

 the succeeding limbs of the trunk. The hinder trunk -limbs are 

 specially modified as swimmerets by the flattening of the endopodital 

 segments, and not, as in Apus, by the foliation of the protopodite. 

 There can be little doubt that the feelers of the Trilobite are Crustacean 

 antennules ; and that the other four head-appendages represent the 

 antennae, mandibles, and two pairs of maxillae. Now the distinct and 

 conspicuous palps of the mandibles and maxillae in Leptostraca caiiy 

 us some way towards the very primitive condition of the appendages 

 of the Trilobites, as do also the comparatively simple biramous 

 thoracic limbs of Paranebalia (Sars, '87). There need be no hesitation, 

 therefore, in deriving the Leptostraca from an ancestral form in which 



R.I. A. PllOC, VOL. XXIV., SEC. B, '2 £ 



