NO. I INSECT THORAX SNODGRASS 39 



ventro-lateral plates of the abdomen, beneath the projecting edges 

 of the terga, which in the adults are fused with the lateral edges of 

 the sterna and reflected dor sally. Since the homology of the so-called 

 " pleural " sclerites in the abdomen of the Coleoptera is not known, 

 the morphological position of the spiracles in this order cannot be 

 exactly stated, but the nature of the variations in the location of the 

 spiracles suggests that there has occurred, in some cases, dorsal and 

 ventral migrations of the spiracles theiuselves from a primitive site 

 in the pleural membranes. 



The spiracles of the thorax in adult insects usually occur at vary- 

 ing levels on the sides, hetzveen the chitinous pleura. They appear, 

 therefore, to be intersegmental in position. In embryonic and larval 

 stages, however, the thoracic spiracles are usually well within the 

 limits of the segments. In the typical Chilopoda (Pleurostigma) the 

 spiracles are in the pleural membranes and generally in the posterior 

 parts of the segments. In the Protura they lie in small plates in the 

 lateral margins of the mesothoracic and metathoracic terga (fig. 8, 

 Sp.). The tracheal branches from each abdominal spiracle in insects 

 are distributed mostly within the segment of the spiracle, and the 

 muscles of the closing apparatus of an abdominal spiracle arise from 

 the tergum and the sternum of the same segment. There can be little 

 doubt, therefore, that all spiracles are segmental organs, as claimed 

 by Lehman (1925) in his recent review of the insect tracheal system, 

 in which also he gives reasons for believing that the primitive tracheal 

 invaginations of insects may represent the ectodermal parts of the 

 ducts of the nephridial organs (nephromixia) of Peripatus. The 

 apparent intersegmental positions of the thoracic spiracles are evi- 

 dently the result of secondary displacements, which in many cases 

 can be traced during development, though some writers (Comstock, 

 1924, Keilin, 1924, de Gryse, 1926) have argued that all spiracles 

 are intersegmental in origin, and that their segmental positions are 

 owing to secondary migrations. A spiracle could not be truly inter- 

 segmental in a soft-bodied insect without interference with its func- 

 tion, and the idea of any organ or set of organs other than folds for 

 muscle attachments being intersegmental is at variance with our con- 

 ception of the nature of metamerism. 



The typical number of spiracles in insects is ten pairs, two pairs 

 being thoracic and eight abdominal. There is much reason from 

 developmental studies for regarding the usual first pair of spiracles as 

 belonging to the mesothorax, and the second to the mesothorax, and 

 for believing that their intersegmental positions, or the occasional 

 prothoracic position of the first are due to forward migrations. Yet, 



