12 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I3I 



position of true Us but also carry the dorsoventral muscles typical of 

 these structures. 



Furthermore, careful dissection of most cockroaches discloses a 

 second series of straplike ligaments, also of muscular derivation, that 

 run from the postcoxal membranes (i.e., from the intersegments) 

 to the f ureal arms of the respective preceding segments. These liga- 

 ments, here designated lils-jui, etc., are inserted on the furcal arms 

 near the seat of the furcophragmatal muscle (see figs. 2, 10: jj. 



It appears, then, that in the thoracic intersegments of cockroaches 

 the former Us may now be represented by as many as three seemingly 

 distinct sites: (i) the following episternum ; (2) the original Us; 

 and (3) the origin of the postcoxal ligament. How this separation 

 came about is by no means obvious. 



As already stated, the often straplike but still fibrous transverse 

 ligaments isps-iUs, etc., are evidently derived from former muscles, 

 and are even now represented in whole or in part by muscles in some 

 species. However, again in the light of relationships found in other 

 primitive groups (e.g., Dytisciis larva (Speyer, 1922), Corydalus 

 larva, etc.), these transverse muscles seem to have served also (after 

 loss of their contractile nature ?) as suspensory ligaments for a por- 

 tion of the longitudinal ventral intersegmental musculature. Vestiges 

 of this or an analogous arrangement are still present in the first 

 thoracic intersegment of some cockroaches. 



In Cryptocercus, which in this respect is the most primitive blattid 

 I have seen, both the ligament isps-iUs and the muscle isps-epso are 

 present and are, laterally, quite distinct (fig. 13 •.4, 5). However, the 

 mesal portion of isps-iUs, incidentally still composed of contractile 

 tissue, is so confluent with the adjacent fibers of isps-eps^ that a 

 separation of the two muscles in this region is hardly possible. Thus, 

 the anterior portion of the muscle isps-epss could be described as 

 "ligamcnt-epS2." Similarly, it is difficult to specify the origin of the 

 cruciate profurcal muscle epS2JUiX (8), for part of its fibers arise on 

 epS2 while the more ventral ones, not visible in figure 13, originate on 

 the ligament isps-iUs, from which they run with the others to the in- 

 sertion on the contralateral furcal apodeme fiii. 



Variations of these relationships are exemplified in a number of 

 other genera, viz, Pcriplaneta, Neostylopyga, Eurycotis, Blatta, and 

 BlatteUa. In none of these is the peripheral attachment of the liga- 

 ment isps-iils preserved ; but the central portion of the ligament is 

 present and extends anterolaterally from the spina as a noncontractile 

 septum on which fibers from epS2 are attached and from which origi- 



