6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I3I 



roaches are regularly present in the first two thoracic intersegments 

 but absent in the third, where their failure to appear is no doubt re- 

 lated to the near obliteration of the ventral region of the first abdomi- 

 nal segment, reduction of which is a very general and probably very 

 early feature of pterygote reorganization. However, the transverse 

 muscles are represented in the abdomen by the muscular attachments 

 of the ventral diaphragm on the anterior lateral angles of the second 

 abdominal sternum. A striking development of these abdominal 

 fibers is seen in the series Pycnosceliis, Diploptera, Nauphoeta, 

 Leucophaea, in which last they attain the status of a powerful trans- 

 verse muscle, SijA-SiiA (figs, i, 8, ii, 12: 34). The function of such 

 a muscle is unknown. Curiously, Blaberus, Cryptocercus, and Macro- 

 panesthia (figs. 5, 10, 17), which share other peculiarities of the four 

 genera just mentioned, do not show any tendency toward hypertrophy 

 of the transverse muscle of the ventral diaphragm, and in this respect 

 are more like the other species included in this study. The lateral 

 suspensions of the ventral diaphragm are not evident in the abdomi- 

 nal intersegments posterior to the first; and in general the ventral 

 diaphragm of cockroaches is much less extensive than that of some 

 other insects, e.g., phasmids and the acridid Orthoptera. 



Some authors have listed as transverse muscles structures such as 

 the ligament fus-fus, whose af^nities are, however, with the spina- 

 furcal muscles. 



b. Spinajiircal muscles. — Cockroaches all have the muscle isps-jui 

 and "3sps"-fus, the latter represented, as a result of loss of the third 

 spina, by fibrous ligaments that often appear as a single transverse 

 band, jus-fus. A corresponding 2sps-fu2 does not occur in any blattid 

 I have examined, and I believe Maki's record of this muscle (1938, 

 fig. 6, No. 41) in P. australasiae must rest on an error of transcrip- 

 tion, since all cockroaches have another, larger muscle, 2sps-fiii, that 

 is omitted from his figure and description. 



Miall and Denny's (1886) statement that the muscle 2sps-fui is 

 inserted on the base of the first leg in B. orientalis is misleading, for 

 the connections in B. orientalis (fig. 2: 16) are identical with those 

 of other cockroaches ; but the description reflects Miall and Denny's 

 awareness of a structural difference between the prothoracic sternal 

 arm and those of other segments, a distinction that seems to have 

 escaped comment by most others who have investigated the muscula- 

 ture of cockroaches. (See section 3, below.) 



The muscles isps-fug and 2sps-jiis are also universally present in 

 blattids as is their possible serial homolog, "3sps"-SnA, which is here 



