426 fVilUmu Morton Wheeler 



ment and remain unaffected by the surrounding processes of 

 growth and differentiation till the imaginal stage is attained. In 

 holometabolic insects the secondary sexual characters are, of 

 course, segregated in the imaginal discs, or histoblasts, and even 

 in hemimetabolic and ametabohc insects there must be a similar 

 isolation of the cell-materials which will produce the somatic 

 sexual peculiarities of the adult. 



The opinion here advocated, namely, that in insects the pri- 

 mary and secondary characters are very loosely correlated dur- 

 ing ontogenetic development or in a very different manner from 

 what they are in vertebrates or even in the Crustacea, receives 

 indirect support from two interesting classes of facts. One of these 

 classes comprises the anomalies known as gynandromorphs,which, 

 though alwa)s rare, are neverthelese much more frequently found 

 among insects than am.ong any other animals. These anomalies 

 consist in combinations of male and female somatic characters 

 in the same individual, usually in such a manner that the two lat- 

 eral halves or the anterior and posterior portions of the body are 

 of different sexes. In the former combination the reproductive 

 organs may be hermaphroditic and correspond with the sex of the 

 halves of the body in which they lie, but this is not always the 

 case, and in anteroposterior, or frontal, or in mosaic, or decus- 

 sating gynandromorphs, which exhibit an irregular mingling of the 

 the sexual characters, the gonads may nevertheless be unisexual. 

 Herbst ('01) and Driesch ('07) have emphasized the obvious 

 inference that these various arrangements of the male and female 

 characters cannot owe their origin to internal secretions, or 

 hormones, and indeed all those who have speculated on the ori- 

 gin of these anomalies are unanimous in holding that they must 

 arise either from peculiarities in the structure of the egg or from 

 irregularities in its fertilitation or early cleavage stages at the 

 very latest. Among recent speculations on the origin of gynan- 

 dromorphism those of Boveri ('02) and Morgan ('05, '09) may be 

 mentioned. Boveri believes that the gynandromorph arises from 

 an egg which has segmented prematurely, so that the male pro- 

 nucleus unites with one of the cleavage nuclei. Morgan is of the 

 opinion "that the results may be due to two (or more) spermato- 



