INSECT SPINASTERNAL MUSCULATURE — CHADWICK 1 43 



sanura, so that it is reasonable to suppose that it must also have been 

 present relatively recently in the ancestry of the Orthoptera. But a 

 form that possessed this muscle in addition to those already found in 

 the present Orthoptera could serve as a potential direct ancestor for 

 both the Orthoptera and Psocoptera, for the various other orders 

 now theoretically derivable from them, namely, the Hymenoptera, 

 Embiodea, Trichoptera, Lepidoptera, Ephemeroptera, Thysanoptera, 

 Homoptera, Hemiptera, and Diptera ; and for the Plecoptera. 



Similar small changes would allow the inclusion of yet other groups. 

 The hypothetical composite that we have formed, with its 17 muscles, 

 contains only one element, 2sps-fu2, that is not found in the present 

 blattid pattern of 19 muscles. But, although this element is missing 

 in present-day cockroaches, the muscle 2sps-fu2 is known from 1 1 

 other pterygote orders and the Thysanura, and it is matched in the 

 blattids by the serial homologues isps-fui and ^sps-fu3. Cockroach an- 

 cestry must have had 2sps-fuo. Almost certainly, too, the ancestral 

 blattids must have possessed a distinct pps-cxz, which occurs in 

 several other primitive groups and is apparently still represented in 

 existing cockroaches as a portion of the thus transformed fitz-cx^ 

 (Badonnel, 1934; Maki, 1938; Chadwick, 1957). Now, the addition 

 of only these two muscles, ^sps-fuz and ^sps-cxs, to the present blat- 

 tid complement creates a composite of 21 muscles, from which could 

 be derived the existing patterns of the Isoptera, i\Iantodea, and Gryl- 

 loblattodea, in addition to those of all the groups already mentioned 

 as possible descendants of the Blattariae and of the combined Orthop- 

 tera and Psocoptera. 



The steps thus outlined dispose of all those more or less primitive 

 orders, possessing upward of nine spinasternal muscles, that are most 

 like what we shall designate for brevity as the orthopteroid stock ; and 

 the small number and logical nature of the changes proposed show 

 that these orders probably form a natural and relatively closely knit 

 group. 



Turning to the remaining orders, we see that in general those groups 

 with more than four spinasternal muscles share some tendency toward 

 retention of a greater proportion of muscles with attachments on the 

 Us, and of muscles of the third spina, while they lack spinaspinal 

 muscles. Apart from these similarities, which are by no means ex- 

 pressed universally among them, these orders are rather diverse in 

 spinasternal structure ; and there is thus good reason for thinking that 

 many of them are only distantly related to one another and to the 

 series of orders already considered. 



