96 EMBRYOLOGY OF INSECTS AND MYRIAPODS 



succeeding one, thus forming the two longitudinal trunks. Lateral 

 branches then arise near the junction of the longitudinal trunks with the 

 main spiracular trunk, to form the transverse commissures. Finer 

 branches from the transverse trunks then pass to the alimentary canal 

 and other viscera. In the more primitive insects the longitudinal trunks 

 may be lacking, the tracheae being arranged in segmental clusters the 

 tubules of which may be branched or unbranched. Though most chilo- 

 pods and the majority of insects have branched tracheae, most diplopods 

 have tracheae consisting of segmental clusters of unbranched tubules. 



The apertures at the surface represent the spiracles. The time of the 

 appearance as well as the position and number of spiracular openings 

 varies with the group to which the animal belongs. Tracheal pits are 

 lacking on the head in most cases. Nelson (1915) states that in the 

 embryo of the honeybee tracheal invaginations are found on the anterior 

 part of the second maxillary segment above the bases of the labial 

 rudiments. These give rise to four diverticula which produce the ante- 

 rior ends of the main tracheal trunks, including the anterior tracheal 

 commissure or loop, and also produce the trachea supplying the head. 

 Later the orifice to the exterior closes, leaving no trace of its existence. 

 In some Collembola cervical spiracles are present which may be persisting 

 examples of the second maxillary spiracles, although tracheae are wholly 

 lacking in most members of this group. The presence of tracheal 

 invaginations in the second maxillary segment is an obstacle to the 

 \aew held by certain writers that the tentorial and some other invagina- 

 tions of the head are homologous to the tracheae. 



Prothoracic spiracles may be present in embryos in some cases, but 

 they usually disappear before hatching. Postembryonic stages of all 

 insects except Diplura do not possess more than two pairs of thoracic 

 spiracles. In the embryo of Lepidoptera as well as in some other orders, 

 the spiracles are formed on the meso- and metathorax, but in the later 

 embryos the first pair moves forward onto the prothorax whereas the 

 second pair becomes reduced and nonfunctional. A similar condition 

 is found in some Diptera. Strindberg (19136) found that tracheal 

 invaginations in Isoptera are formed in the first two thoracic segments 

 but not in the third. In some species of Diplura there are three or four 

 pairs of spiracles on the thorax. 



Eight pairs of spiracles usually develop on the abdomen; but Cholod- 

 kowsky (1891) claims to have found a ninth pair in Blattella, and Hey- 

 mons (1897) found in the embryo of Lepisma the rudiment of a tenth. 

 Nine pairs of embryonic spiracles have been reported in Hydrophilus 

 and Donacia. In Chironomus (str. sens) the tracheal system does not 

 develop until the late larval stages and then only in the thorax. On 

 the other hand, in the subfamily Orthocladiinae of the same family, the 



