54 W. A. SANDS 



grouped at some distance from the rest. Various arrangements were tried in efforts 

 to obtain well-defined genera. These were based in turn on differences in the posi- 

 tion and form of the worker enteric valve seating, in the proportions of imago and 

 worker mandibles, and in the widths of imaginal thoracic nota. However, the 

 results of cluster and vector analyses of the similarity matrix placed the species 

 in a single rather large loose cluster, near to the genera Adaiphrotermes, Acholotermes 

 and Anenteotermes (Text-fig. 56). When the stage of preparing keys to genera was 

 reached it quickly became clear that the groupings previously tried would not key 

 out at all. For practical purposes the grouping indicated by overall similarity 

 was better than any of the others tried earlier, and was therefore adopted. It should 

 be noted that consequent on the heavy weighting imposed on the enteric valve 

 characters by the coding used, the genus is essentially defined by its lack of armature 

 in this feature. 



Species-Groupings 



Although the earlier attempts at clustering proved to be at too low a level, some 

 of the groups so formed remain recognizable as species-groups within the genus 

 Astalotermes. Having considered the possibility that they could be treated as 

 subgenera, I now think it is better to deal with them informally. They are of 

 value in indicating more clearly the affinities of Astalotermes with other genera, 

 but the groups are not so well-defined that any useful purpose would be served by 

 naming them. As usual, there are some species that appear to be related to more 

 than one group and do not readily fall into any of them. The arrangement of 

 the species descriptions following the keys conforms to these general groupings, as 

 indicated below. 



Group I. Smaller species, W, (Imago) 0-69-0-86; apical teeth of mandibles 

 longer, La/Li (Imago) 0-50-0-79, (Worker) 0-56-0-86; imago, meso- and metanota 

 rather consistently narrower at constriction, M/W, 0-20-0-28; worker, enteric valve 

 seating lateral or dorsolateral in unopened abdomen, sessile to strongly necked, 

 weakly to distinctly two-lobed. Species included: A. amicus sp. n., A. benignns sp. 

 n., A. comis sp. n., A. concilians (Silvestri), A. eumenus sp. n., A. amicus and 

 A . eumenus have longer apical teeth than the other species. These, and their strongly 

 necked enteric valves may indicate a leaning towards Acholotermes. At the other 

 end of the group, A. comis is clearly related to Acidnotermes. 



Group II. Larger species, W, (Imago) 0-98-1-23; apical teeth of mandibles a 

 little shorter in some species, (Imago) 0-55-0-66 (Worker) 0-62-0-79; imago, meso- 

 and metanota variably wider at constriction, M/W, 0-26-0-32; worker, enteric valve 

 seating ventrolateral to lateral in unopened abdomen, long-necked, distinctly two- 

 or three-lobed. Species included: A. aganus sp. n., A. hapalus sp. n., A. ignavus 

 sp. n., A. mitis sp. n. and A. murcus sp. n. A. ignavus and A. murcus are 

 more closely related to each other than to the rest, with their extremely long head- 

 setae. The other three species may represent a link with genera such as Alysco- 

 termes and perhaps Astratotermes. It is also likely that Adynatotermes and Agano- 

 termes are derived from forms like these. 



