PROTOZOA AND OTHER ANIMALS 925 



faunules become known the tendency probably will be relatively to re- 

 duce this number. Many species are known from several or many hosts. 

 Trichonympha agilh probably occurs in all species of Reticulitermes, but 

 has not been found in other termites. Staurojoenina is widespread in 

 Kalotermes sensu lato, and there are few if any differences between 

 species of different hosts. Of twenty species of the genus Devescovina, 

 only nine have been found in but one host each. On the other hand, 

 there are species with many hosts widely separated geographically. D. 

 glabra has been identified in eighteen termites from Africa, Madagascar, 

 Java, and Sumatra; D. lemniscata has seventeen hosts in Central and 

 South America, the West Indies, Australia, the Pacific Islands, Africa, 

 Madagascar, Java, and India. A unique, elaborately organized deves- 

 covinid, when first found in a Ceylon termite, was thought to be a 

 strictly host-specific form; but it has since been found also in a 

 termite from Australia. The small, simply organized Tricercomitus, 

 which occurs in most if not all species of Kalotermes sensu lato, appar- 

 ently is one species, T . divergens, in all those in which it has been 

 studied. Another species exists in Zootermopsis. 



Many species of termite flagellates in all groups have a present host 

 distribution which indicates greater stability in characteristics than ex- 

 isted in the same period of time in the insects. Speciation has occurred 

 in the hosts without having taken place in certain of the symbionts. That 

 there are other termite flagellates which have evolved into different spe- 

 cies in single hosts is probable; but we cannot designate any one as cer- 

 tainly rigidly host-specific. Although there are many one-host forms, the 

 situation is such that finding any one of them in a termite, even in 

 another part of the world, would not be astonishing. Even although there 

 is only a single extant host, it would in no instance be unlikely that 

 formerly existing species not directly ancestral served as hosts of the 

 flagellate. 



But whether a flagellate species occurs in one or in several host species 

 of termites is far from being the question of greatest interest in host 

 distribution. It has, in fact, little significance for general considerations. 

 More important is the fact that there is limitation of certain flagellate 

 types to certain groups of termites. That is true mainly among poly- 

 mastigotes. There are also some very widely distributed flagellate types, 

 but that only adds to the significance of the instances of strict limitation. 



