34 W. A. SANDS 



diverging from very near its base. One difficulty in accepting this position has been 

 thought to be its lack of a soldier caste. However, 'soldierlessness' is an adaptive 

 character that could develop several times over; considering the rarity of soldiers 

 in a number of groups this is perhaps not a serious obstacle. 



Noirot and Noirot-Timothee noted two exceptions to the above scheme. Eburni- 

 termes, a monotypic genus described by Noirot (1966), was included by him in the 

 Amitermitinae mainly on the basis of the similarity of the worker mandibles to 

 Eurytermes and the soldier to those of the same group of genera. He pointed out 

 the resemblance of the gut to that of the Apicotermitinae, particularly in the arma- 

 ture of the enteric valve and the short first segment of the proctodeum. The 

 malpighian tubules are again attached some way up the midgut. There is no mixed 

 segment. This also applies to the other 'genus' mentioned by these authors, namely 

 those species of african ' Anoplotermes' known to them. In the latter group the 

 enteric valve was stated to be unarmed. 



One further genus has been described in the subfamily that does not fit the two 

 main groups as defined above. Deligne & Pasteels (1969) give details of the intestine 

 of Labidotermes. There is no mixed segment in this genus and the malphigian 

 tubules are attached to the midgut. The enteric valve armature has only a single 

 large tooth at the anterior end of each cushion. The worker mandibles are similar 

 to those of the Anoplotermes-Speculitermes group. 



In a preceding section the layout of the gut of the afiican soldierless termites 

 is described and an account given of its variations. Further details will be found 

 in the introductory passages to genera and in species descriptions. The salient 

 features are illustrated for all species. Within the group are found all stages in 

 the development of the mesenteric overlap with the proctodeum from a simple 

 transverse junction like those of Ebarnitermes and Labidotermes, to elongated mixed 

 segments with dilated ends. There are also all stages in the development of enteric 

 valve armature from nothing at all to elaborate spines, hooks and sclerotizations 

 everted through the valve opening. A similar though less extensive variety of 

 forms has been seen in the Neotropical species of Anoplotermes. Thus the excep- 

 tions mentioned by Noirot and Noirot-Timothee to the two divisions of the Amiter- 

 mitinae are not separable by any sharp division from the first type in which the 

 malpighian tubules are attached to the wall of the midgut. In addition in all these 

 forms the gizzard is feebly developed. The other important character is only 

 readily appreciated by examination of the layout of the gut in situ. Once uncoiled, 

 the relative positions of its parts are displaced. In the first type of Amitermitinae 

 where a mixed segment is found the extension of the mesenteron is always around 

 the inner curve of the coil of the gut on the same side as the insertion of the mal- 

 pighian tubules. 



In the second group of Amitermitinae the extended part of the mesenteron 

 forming the mixed segment is always around the outside of the loop of the gut on the 

 opposite side to the malpighian tubules where these are attached at the anterior 

 end of the proctodeum. A number of possible exceptions have been examined, 

 such as Prohamitermes . Here the mesenteron is effectively external, arising between 

 the two outer malpighian tubules. In a more advanced Eremotermes species, 



