8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I3I 



has been found in the immature mantid, Tcnodera sinensis Sauss. 

 (fig. 7:^i). 



The spinaspinal muscles are the only portion of the longitudinal 

 ventral musculature that has obviously retained its primary relation- 

 ships in blattids. Both attachments are still unmistakeably interseg- 

 mental. Like the other somatic muscles, the spinaspinal muscles are 

 paired bilaterally ; but the right and left bands of isps-2sps are often 

 so closely appressed in the midline that they seem like a single ele- 

 ment, and they have been so described by some authors. The pos- 

 terior insertions of 2sps-"^sps" are usually well separated on the 

 ligament fits-jus (e.g., fig. g:2^), and in some instances may even 

 seem to be on ftis at the site of attachment of the ligament. Care must 

 be taken, therefore, not to confuse them with the usually better de- 

 veloped spinafurcal muscles, 2sps-fii3, from which they are morpho- 

 logically distinct. 



The ligament fug-fus also serves, in several cockroaches, as a base 

 of attachment for paired muscular strands that course posteriorly to 

 join the mesh work of contractile fibers and membrane that consti- 

 tutes the ventral diaphragm (figs. 2-5, 8, 11, 18:50). These strands, 

 "3sps"-ve}itr. diaphr., may be serial homologs of the muscles isps- 

 2sps, 2sps-'3sps." I did not succeed in finding these delicate strands 

 in all species, but could not be sure, in the cases where they seemed 

 absent, that I had not destroyed them. 



In pterygote insects, there is no homolog of the spinaspinal muscles 

 anterior to isps; but Maki (1938) has recorded muscles that are 

 probably homologous in the prothorax of some Apterygota. 



e. Spinabdominal muscles. — The spinabdominal muscles of cock- 

 roaches include only 2sps-SiiA, "3sps"-Sji,i and " 3sps"-ventr. diaphr. 

 The last two have already been mentioned in this section, b and d, 

 above. They arise on the ligament jus-fi's, and not on the base of 

 fug as some have stated. The muscle 2sps-SnA is characteristic of 

 blattids, and is present in all of them I have seen, though it is weak 

 in Leiicophaea. Elsewhere, it has been recorded only from Gryl- 

 lohlatta (Walker, 1938, No, iiih). It is interesting as an example 

 of a muscle more than one segment in length, a type that is of infre- 

 quent occurrence in pterygote insects. In Macropancsthia, Peri- 

 planeta, Eurycotis, and Neostylopyga the abdominal insertion of some 

 or all the fibers is actually on Sjua- Apparently this modification 

 may occur readily because 2sps-SnA is ordinarily inserted along 

 the antecosta of Sha. dorsal to the usual longitudinal bands, Sua- 

 SiiiA, with which 2sps-SnA is more or less continuous. Dissolution of 

 the integumental attachment at sjja adds one segment of muscle 



