THE SOLDIERLESS TERMITES OF AFRICA 35 



E. indicates, it appears on the inside of a gut loop, but it is the one further forward, 

 the proctodeum having elongated to a greater extent and pushed the very short 

 mesenteron back. The mesenteric extension remains on the same side of the gut, 

 and in E. nanus the mixed segment is of typical Amitermes form. In Synhamitermes 

 the proctodeal overlap appears to have elongated more than that of the mesenteron, 

 and has twisted across the latter. They remain morphologically on the same side 

 as in less developed forms. These elongated mixed segments suggest that this 

 structure may result more from a need to extend the proctodeum than the mesenteron. 



Thus within the Amitermitinae as hitherto recognized there are two morphologic- 

 ally quite distinct lines. A parallel development of the mixed segment has taken 

 place on opposite sides of the gut, and there is a radically different attachment of 

 the malpighian tubules. It is therefore relevant to enquire whether the two lines 

 resemble other subfamilies of Termitidae more than each other. In the Termitinae 

 the pattern of the gut is essentially the same as in the second type of Amitermitinae. 

 Minor modifications are found but they appear to be relatively small divergences 

 from the same basic stock. The similarity of the most primitive forms of the 

 first group of Amitermitinae to the Apicotermitinae has already been indicated, 

 and it has been pointed out that numerous intermediates between them and the more 

 specialized forms with a long mixed segment exist. The subfamily Apicotermitinae 

 as conceived by Grasse - & Noirot (1954) has not achieved general recognition up 

 to the present time. This is understandable in view of the small number of genera 

 involved, and the fact that the significance of the gut characters had not been fully 

 worked out. Sufficient evidence has now accumulated to provide a new basis 

 for subfamily classification within the Termitidae. These divisions will reflect 

 more clearly the phylogeny of the group and at the same time incorporate the existing 

 knowledge of other characters more satisfactorily. It is proposed to divide the 

 heterogeneous group hitherto named the Amitermitinae between the Apicotermitinae 

 and the Termitinae, as outlined above. The amalgamation of the second group of 

 Amitermitinae with the Termitinae results in the former name becoming a junior 

 synonym since this group includes the type-genus Amitermes. This action was 

 suggested by Noirot (personal communication) ; having considered the alternative 

 of retaining the five subfamilies, and examined many representatives of all of them 

 I now believe he is right, and here give formal expression to his idea. It may be 

 noted that Ahmad (1950) admitted to doubts of the subfamily status of Amiter- 

 mitinae. He based the only distinction between them and the Termitinae on the 

 length of the apical teeth of the mandibles. This does not bear comparison with 

 much greater differences within other subfamilies, nor does it hold when further 

 species are examined within the groups as he defined them. 



The Nasutitermitinae will be retained as a valid subfamily, although their intestinal 

 morphology is clearly derived from the same stock as the Termitinae. The Macro- 

 termitinae with their short intestine like the more primitive families, and four 

 evenly spaced malpighian tubules at the junction of mesenteron and proctodeum, 

 are distinct from the other three subfamilies. Thus the Termitidae will, with 

 the removal of the Amitermitinae, be divided into four subfamilies as shown in 

 Text-fig. 24. The subfamily Apicotermitinae as now constituted will contain 36 



