438 T. H. MOEGAN. 



wings removed. Corresponding to the 5th metamere of the abdo- 

 men we find a compound metamere^ so that one half-metamere 

 of the left side corresponds to two half-metameres of the right 

 side. A figure of the abdomen from below is drawn in 

 fig. 89, B, and the abdomen seen from the right side is seen in 

 fig. 89, c. On account of the interpolated half-metamere the 

 abdomen is bent towards the left. The interpolated segment 

 reaches exactly to the median line above and below. The 

 abdomen ends normally, having all of the accessory structures 

 of the male perfect in all details. 



Fig. 89, c, shows that each of the half-metameres on the 

 right side of the compound metamere bears stigmata, so that 

 there is one more of these on the right than on the left side of 

 the body, i. e. eight on the right (one more than normal) and 

 seven on the left. 



If we assume each of the right halves of the compound 

 metamere equivalent to a true half-metamere of the series, 

 then each of the half-metameres following the compound 

 metamere is joined to a half-metamere that is not its normal 

 half. As a consequence, one new additional half-metamere 

 will have to be formed at the end of the series on the right 

 side to make a perfect ending to the abdomen. In so highly 

 modified a form as the locust this view seems very improbable. 

 An alternative view would be to suppose that a half-metamere 

 has been intercalated between the 4th and 5th, or between 

 the 5th and Gth^ and that the intercalated half has united 

 across the middle line to the 5th in common with the normal 

 half of the 5th. We might think of this intercalation as due 

 to a half of one metamere dividing into two, or we may 

 suppose that when the ventral raesoblast broke up 

 into a series ofblocks it formed on one side one more 

 block than on the other (one more than the normal). 

 If the last view be true — and it seems more probable than any 

 of the other suggestions — we see at once the futility of trying 

 to explain the conditions on an assumption of predestined 

 riiiht and left half-metameres. When we recall that the whole 

 ventral plate of the insect becomes segmented at the same 



