EVOLUTION OF INTERSPECIES INTEGRATION AND ECOSYSTEM 



717 



metabolic products. In addition, the oxygen 

 needs of the termite maintain a low oxygen 

 tension which (6) permits the protozoa to 

 survive and at the same time insures that their 

 metabolism shall be anaerobic and result in 

 products useful to the termite. The unique 

 conditions which the protozoa require and 

 which the termite supplies may explain in part 

 the restricted distribution of the protozoa." 



Perhaps the best evidence for the de- 

 pendence of the protozoa on their roach 

 and termite hosts is the fact that all the 

 species of flagellates of the orders Polymas- 

 tigina and Hypermastigina so far recorded 

 from these insects are wholly confined to 

 the termites or the single species of wood- 

 eating roach. Of the eight families of hy- 

 permastigotes, six are found in both Cryp- 

 tocercus punctulatus and in termites, al- 

 though, with one exception (Trichony- 

 mpha) , the genera are exclusively found 

 either in the roach or the termites. Of the 

 polymastigotes, one family (Pyrsonymphi- 

 dae) has one genus in the roach and two 

 genera exclusively in termites. Another 

 family (Trichomonadidae) has one genus 

 in the roach, in termites and in other in- 

 sects, while eight genera of the subfamily 

 Devescovininae are confined to termites. 



Many of the protozoa are species-specif- 

 ic or are confined to closely related groups 

 of host species. In some instances closely 

 related species of the same protozoan 

 genus occur in a single host species and 

 appear to have originated in this narrow 

 ecological niche (p. 628). These proto- 

 zoans seem to be incapable of an inde- 

 pendent existence, and contrary to the 

 potentialities of most free-living flagellates, 

 nearly all are incapable of forming cysts 

 that might enable them to survive period- 

 ically unfavorable conditions. The proto- 

 zoan phylogeny seems to be largely cor- 

 related with the phylogeny of their hosts— 

 so much so that both Cleveland and Kirby 

 think that the mutualism goes back to the 

 ancestors of the roaches and the termites. 



Martynov (1937) presents evidence that 

 the modern order of roaches (Blattodea or 

 Blattaria), with their parchment-like fore- 

 wings, could not be the ancestors of ter- 

 mites (Isoptera) with their membranous 

 wings, even though primitive roaches are 

 well represented in Pennsylvanian rocks 

 (Carpenter, 1930). Martynov thinks that 

 the common ancestor belonged to an ex- 



tinct hypothetical order of Devonian or 

 Carboniferous age. 



All the species of the four primitive fam- 

 ilies of termites (Mastotermitidae, Kalo- 

 termitidae, Hodotermitidae, and Rhinoter- 

 mitidae) have intestinal flagellates, and 

 there is fair circumstantial evidence that 

 these families were well established in the 

 Mesozoic age. It may therefore be assumed 

 that this mutualistic relationship between 

 the insects and their intestinal flagellates 

 has been in existence and evolving for 150 

 to 250 miUion years. In 1948, 528 species 

 of termites were known to harbor these 

 flagellates, and from them about 250 spe- 

 cies of flagellates have been described. 

 Numerous additional species of termites 

 and protozoans await description. 



It is also fairly obvious that the family 

 and social life of the roaches and termites 

 evolved in part as a necessary adjustment 

 for the transmission of the protozoa, and 

 that the difi^erence in colonizing behavior 

 between the roaches and the termites is 

 also associated with difi^erences in the mode 

 of protozoan infection. Cleveland, Hall, 

 Sanders, and Collier (1934, p. 209) say: 



"Once the protozoan irrfection is acquired 

 in Cryptocercus, it is never lost until death; 

 and any individual after acquiring it is capable 

 of living by itself during the rest of its life. 

 This, however, is in direct contrast to termites 

 ( even lower ones where the reproductive in- 

 dividuals do not lose the ability to feed on 

 wood) since the protozoa are lost at each 

 moult, and colony life is essential in order 

 that reinfection from non-moulting individuals 

 may take place. On the other hand, two 

 sexually mature first form reproductive adult 

 termites may leave the colony and start a new 

 one; but this is impossible in Cryptocercus. 

 for sexually mature adults very probably do 

 not moult and, unless they do, they could not 

 infect their young with protozoa. Hence, in 

 order to start a new colony, it is necessar)' 

 for nymphs to accompany adults, so that when 

 the nymphs moult they can transmit protozoa 

 to the young which the adults in the meantime 

 have produced." 



It should be noted in this connection that 

 all termite species have winged colonizing 

 reproductives, while Cryptocercus is wing- 

 less in both sexes. 



It seems that in this case at least we are 

 forced to the conclusion that the mutualis- 

 tic relationship has resulted in the evolu- 

 tion of something closely approaching an 



