THE ORIGIN OF GYNANDROMORPHS. 33 



two plants — the epidermis of one species and a core of the other s{x?cie8. 

 The mosaic shown in Cytisus adami, a hybrid resulting from fi;raftinK 

 Cytisus purpureus and Laburnum vulgare, seems also to be due to a 

 similar sort of combination. In animals mosaics have been produced 

 in hydra by King by grafting pieces of a deep-green race on a light 

 one, and by Whitney by destroying the green pigment of one indi- 

 vidual and grafting pieces of it onto a normal green hydra. In tad- 

 poles combinations of different species caused by grafting have been 

 made by Born, Harrison, Morgan, and others. A result strictly 

 comparable to the periclinal chimaeras of plants has been reached by 

 grafting a piece of the tail of one species on to the amputated stump 

 of another species. As the new tail grows the skin of the stock is 

 carried out over the core derived from the graft, and as a result an organ 

 is formed with an outer layer of one and a core of another species. 



The mosaic seeds of corn that are striped with red and white have 

 been shown by Emerson to arise through a mutation in the gene for 

 striping. The ''half-and-half" mosaic grains that have been recorded 

 by Correns (1899), Weber (1900), East and Hayes (1911), Emerson 

 (1915), and Collins (1919) have been variously accounted for — re- 

 calling the different interpretations that have been advocated for 

 gynandromorphs in animals. Emerson (1915) reviews these theories 

 and advances the explanation of somatic mutation. It seems not 

 improbable that elimination will account for those mosaics in which 

 the triploid endosperm nucleus is involved. 



CLASSIFICATION AND DESCRIPTION OF GYNANDROMORPHS 



OF DROSOPHILA. 



The main group includes the gynandromorphs that are adequately 

 explained by chromosomal elimination. It is subdivided according to 

 the type of gynandromorph into : (1) those approximately bihiteral, (2) 

 those mainly female, (3) those mainly male, (4) those in which the 

 type is largely "fore and aft," and (5) those in which the mother was 

 known to have been an XXYfemale, but in which simple elimination 

 is sufficient to account for the results. Another group (6) includes 

 those in which the distribution of parts is irregular. These types are 

 only approximations and by no means mutually exclusive; it is often 

 somewhat difficult to decide to which type a specimen belongs. 



The highly interesting group of special cases (7) is undi\ided, 

 though it calls for three or four different genetic explanations, based, 

 however, on special modes of distribution of the sex chromosomes. 

 In the Appendix are included those cases in which our records are 

 incomplete as to parentage or in which the specimen has been lost, 

 so that the description is sketchy. This grou]) contains many of the 

 very early gynandromorphs. To this subdivision is added a brief 

 review of previously pubhshed gynandromorphs in Drosophila. 



