THE SOLDIERLESS TERMITES OF AFRICA 7 



nomic value, but the width of the narrowest part between the wing processes of 

 both nota is important. The presence or absence of a dark sclerotized suture at 

 this point, the internal attachment of the oblique lateral dorsal muscles, is a further 

 character. The 'medial frons spot' of Fuller (1925) lies immediately in front of the 

 fontanelle and is the attachment of the medial muscle of the labium. It is some- 

 times raised or depressed, and will be referred to simply as the medial spot. A 

 pair of crescent-shaped or semicircular smooth marks between the ocelli and the 

 posterior margin of the postclypeus were named 'antennal organs' by Holmgren 

 (1909). Whatever their morphological origins, this name seems confusing and they 

 will be referred to here as frontal marks. The antennae always have 15 articles 

 in imagos and 13 in workers. 



A considerable range of measurements of both imago and worker castes has been 

 undertaken. These will be referred to again in the sections on multivariate methods. 

 Although the analyses of measurements were done first, they are described after the 

 multivariate similarity study based on coded characters. This is done because 

 the latter follows on more naturally from the description of the non-measured 

 characters. In listing measurements in the species descriptions, the limits of the 

 range are rounded off to two decimal places. Where more than two specimens 

 were measured a mean value was calculated and when measured series numbered 

 five or more specimens the standard deviation is also given. These calculated 

 values are given to three places of decimals. Where the rounded figures for the 

 range coincide they are omitted and only the mean given. A single figure in the 

 range column indicates invariance in the character for the specimens measured, 

 in which case no mean is given. 



In the worker caste a new set of characters based on features of internal anatomy 

 is used. The existence of a histologically mixed segment in the digestive tract at the 

 junction of the mid-gut and the hind-gut in the worker castes of the Termitidae 

 was first demonstrated by Sutherland (1934). This author was also led by the 

 appearance of its spiny armature to assume that the enteric valve functioned as a 

 second 'gizzard'. The enteric valve is found in all termites where the first part of 

 the hind gut enters or connects with the large pouch of the second, posterior part, 

 sometimes referred to as the 'large intestine'. Grasse and Noirot (1954) studied 

 in more detail the anatomy of the mixed segment and the enteric valve of a number 

 of Termitidae, and pointed out the value of these characters in the systematics of 

 the group. The differences in intestinal anatomy led them to propose separating 

 certain genera in a new subfamily, the Apicotermitinae. Noirot and Kovoor 

 (1958) carried the investigation of the gut in the Termitinae further and discussed 

 its phylogenetic implications. Recently Noirot (1966) has included descriptions 

 of the gut of the worker in describing two new Amitermitinae. Noirot and Noirot- 

 Timothee (1969) summarized the available information on the intestine of all 

 termite families and Kovoor (1969) has carried out a more detailed study of the 

 Nasutitermitinae that confirms the great importance of these characters in inter- 

 preting the phylogeny. 



The detailed anatomy of the gut has not previously been investigated in the 

 soldierless termites, but it forms an indispensable part of this revision. There is 



