Carpenter — Relationshijjs behceen Classes of Arthropoda. 321 



demonstration of the Arthropoclous affinities of the Peripatidte has led 

 many students to look upon those wonn-like creatures as indicating 

 the probable ancestors of Millipedes, Centipedes, and Insects, and to 

 believe in the derivation of those classes from an Annelidan stock 

 quite independently of the Crustacea. Thus the opinion seems to have 

 been slowly gaining ground that the Arthropoda can no longer be con- 

 sidered as a natural group of the Animal Kingdom (Hutton, &c., '97). 



The most extreme view of the multiple origin of the Arthropoda 

 is that put forward by Bernard ('96), who would derive each of the 

 great classes independently from an Annelid ancestry. Most recent 

 writers, however, consider that the present-day Arthropods have 

 developed along two main lines of descent. Kingsley ('94), for 

 example, recognising, with Lankester ('81), the Arachnidan affinities 

 of Limulus, refers the Crustacea and Arachnida to one great group, the 

 Insects, Centipedes, and Millipedes to another. But other zoologists 

 consider the manner of breathing to be the all-important character in 

 deciding the affinities of the Arthropod classes. Lang ('91), for 

 example, divides the Arthropoda into a Branchiate and a Tracheate 

 series, regarding Limulus and the Eurypterida as closely allied to the 

 Crustacea, and believing that the Arachnida were derived fi-om the 

 Insectan (Tracheate) stock by the fusion of the head with the thorax 

 and the disappearance of the feelers. 



Supporters of either of these two views agree in supposing a wide 

 divergence between Crustaceans and Insects ; they differ as to whether 

 the Arachnida should be associated with the former or with the latter 

 group. The special question of the affinities of the Arachnida will be 

 discussed later. The conflicting views of the various authors men- 

 tioned have been briefly sketched as an introduction to the argument 

 of this essay, which will endeavour to show that the various classes 

 of the Arthropoda are indeed truly related to each other, and that 

 ancestors with distinctly arthropodan characters must be predicated 

 for all of them. As has been recently pointed out by Lankester ('97), 

 the structural features in which all Arthopods agree — even if the hard, 

 segmented exoskeleton and the jointed limbs be left out of account — 

 are striking and remarkable. The heart with paired openings ; the 

 "pericardium" and the secondarily -formed body-cavity made up of 

 greatly enlarged blood-channels ; the reduced ccelom ; the variable 

 number of pairs of mesodermal excretory tubes ; the uniformly striated 

 muscle-fibres,^ and the complete absence of ciliated epithelium^ — all 



' Except among the Malacopoda. 



2 J92 



