An endeavour to show tbat the tracbese arose troxa sëtipàrous sacs. 523 



Summary. 



The deduction of tracheae from setiparous glands is thus found 

 to harmonise many of the anomalies presented by the tracheal systems 

 of the Arthropoda. We explain the diffuse arrangement of the tracheae 

 in Peripatus by deducing them from the bristle glands which were 

 scattered over the surface of the body, as is still the case in many 

 living Chaetopods and in most other Tracheata. 



The regular metameric arrangement of the tracheae of the Hexa- 

 poda is explained by referring them back to the acicular glands of 

 vanished dorsal parapodia. This derivation finds some support in 

 the claims of Bütschli and Carrière that the tracheae are homo- 

 logous with the larval spinning glands and salivary glands of the 

 Insecta. The row of lateral segmentally arranged stigmata found 

 in the embryo is no longer quite complete in the adult owing to later 

 specialisation of parts of the body. 



The conditions in the Myriapoda are also to some extent explained. 

 The lateral row of stigmata may be considered as the acicular glands 

 of vanished dorsal parapodia ; the stink glands occurring just above 

 them in some families may be groups of parapodial setiparous glands. 

 The dorsal "lungs" of Scutigera may also be groups of bristle glands 

 joined in the middle line, perhaps homologous with the stink glands 

 of the Iulidœ. In the Myriapoda we also find setiparous glands on 

 the coxae of the legs, as in Peripatus. 



In the Arachnida, the row of derivatives from the dorsal acicular 

 glands has disappeared, at least, they are not developed in the lateral 

 line as in the Hexapoda. Instead of a lateral, we have a ventral 

 row of structures which may be deduced from the acicular glands 

 of either the dorsal or the ventral parapodia ; a point which can only 

 be determined when we know more about the origin of the Arach- 

 uidan limb. These structures show apparently almost exactly the 

 same modifications as those of the lateral rows of the Hexapoda 

 and Myriapoda. We have salivary or poison glands , tracheae 

 (lungs or tubes) and spinning glands. This ventral row in the 

 Arachnida has, however, certain modifications (the coxal glands) 

 which apparently have nothing to correspond with them in the 

 Hexapoda. The Crustacea help us to homologise these also with aci- 

 cular glands, as I shall endeavour to show in a subsequent publication. 



In the foregoing argument we have assumed little which may 

 not be readily granted , that is , apart from the question of the 



