INSECT SPINASTERNAL MUSCULATURE — CIIADWICK 1 37 



the living representatives. A corollary of this interpretation is that 

 individual muscles that we have listed are in most cases homologous 

 throughout the orders and segments in which each occurs. Nowhere 

 have we found evidence for the creation in phylogeny of wholly new 

 spinasternal muscles; rather we have observed that the ruling tend- 

 ency has been one of increasing restriction of the ancestral pattern as 

 the more recently differentiated orders evolved. In those instances 

 where there is a change in the complement of spinasternal muscles 

 during ontogeny, one detects the same trend toward reduction of the 

 spinasternal musculature that is so clearly manifest in phylogeny. 



The spinasternal musculature also provides reasonably good indi- 

 cations that all three thoracic segments were at one time nearly iden- 

 tical in respect to it. Such relationships are not seen anywhere at 

 present, for the proximity of the prothorax to the head and of the 

 metathorax to the legless abdomen have dictated that the several 

 thoracic segments should diverge in structure. 



If we ignore the unusual spinatergal and spinatrochanteral muscles, 

 whose distribution has not been sufficiently studied to allow their inclu- 

 sion here, we may suppose that the typical body segment, for which 

 we shall use the mesothorax as a model, originally had the following 

 spinasternal muscles : 



1. 2SpS-2ils 3. 2SpS-Sils 6. 2SpS-CXi 



2. 2sps-jui 4. 2sps-iils 7. 2sps-3sps 



5. 2SpS-CX» 



From these seven basic types we can, with a fair degree of probability, 

 derive all the known spinasternal muscles that have been considered 

 above. The proposed derivations are summarized in the following 

 paragraphs, which carry the same numbering as the corresponding 

 muscles in the list just given. 



I. The transverse muscle, typified by 2sps-2ils, is thought early 

 to have split ofif a second band with attachment on the anterior margin 

 of the following subcoxa; i.e., the present 2sps-cpS3. However, a 

 derived muscle of this nature seems to have been confined to the first 

 two thoracic intersegments. There is no indication that a correspond- 

 ing postmetathoracic band ever existed ; and if one developed in the 

 cervical region, it was soon obliterated. In the pterothorax, modern 

 insects may show either, neither, or occasionally both bands. 



The muscle 2sps-2ils and its serial homologues, isps-rils and ssps- 

 Sils, have frequently been replaced by a ligament, apparently as a re- 

 sult of increasing sclerotization and consolidation of the thoracic seg- 

 ments. Loss of the spinasternal attachment, or disappearance of the 



