PROTOZOA AND OTHER ANIMALS 965 



it contains, or devour leaves; in fact, almost all types of vegetable mat- 

 ter are utilized by certain members of the group. Hodotermitidae forage 

 for grass and herbs, even eat straw from unbaked bricks; some on the 

 Karroo collect twigs. Kalotermitidae and Rhinotermitidae can live on 

 paper; even, as stated above, cotton cellulose and a lignin-cellulose com- 

 plex (see Cleveland, 1924, 1925b) . The wood-boring roach Cryptocercus 

 punctulatus eats the wood of fallen timber, well-decayed or sound. 



Cleveland (1923) pointed out that, in the groups of termites that 

 use a uniform diet of wood, all species examined had rich faunas of 

 Protozoa; and this has been confirmed by studies by the writer of more 

 than a hundred additional species. In Termitidae, with varied food 

 habits, such faunas are absent, though there are some Protozoa in many 

 species (Kirby, 1937). Cleveland et al. (1934) remarked that the 

 correlation of wood feeding and intestinal flagellates is not so close as 

 he at first supposed, since there are some Termitidae that eat wood and 

 have no (xylophagous) flagellates. 



We know very little of the actual nutrient substances among the 

 varied materials taken in by termites as a group. Matter that has passed 

 through the digestive tract is used extensively by higher termites in 

 building mounds, fungus gardens, and carton nests. Cohen (1933) and 

 Holdaway (1933) analyzed mound material of Eiitermes exitwsus, 

 which contains no Protozoa and feeds on wood. They found cellulose 

 to be much reduced, though some passes out undigested, whereas lig- 

 nin is unaffected. These results agree with those reported by Oshima 

 (1919) after analyses of wood and nest material of Coptotermes formo- 

 sanus, which does contain xylophagous flagellates. Oshima concluded 

 that the principal food of that termite is cellulose and that there is no 

 decrease of lignin. In termites of still another group, Zootertnopsis, 

 Hungate (1936, 1938) found essentially the same thing by analyses of 

 uneaten wood and pellets. 



Tissue-produced cellulase is absent from Cryptoceycus punctulatus 

 (Trager, 1932), Kalotermes jidvicollis (Montalenti, 1932), and Zoo- 

 termopsis angustkollh (Hungate, 1938). Probably none is present in 

 any Kalotermitidae or Rhinotermitidae, though, in the light of the 

 situation with wood-boring beetles, one should not generalize from 

 limited data. Termitidae have not been investigated for cellulase. Man- 

 sour and Mansour-Bek (1934a) suggested that some termites may be 



