SYMBIOSIS BETWEEN TERMITE AND FLAGELLATES. 321 



oxygen as their host, and oxygen is really not any more toxic 

 for them than for their host. They die, then, while their host 

 is uninjured, because the oxygen percentage of their environment 

 increases many times more than that of their host. 



In larger animals with a different system of respiration, the 

 oxygen percentage of host and parasite may increase corre- 

 spondingly, in which case it would be impossible to kill the para- 

 sites of, say, a vertebrate by confining it in an oxygen atmosphere 

 without at the same time killing the vertebrate, unless, of course, 

 : s r en is actually more toxic for the parasites. At any rate, 

 other parasites may be killed by the use of oxygen if we can 

 develop a method of getting it to them. 



When termites are starved, their largest protozoa die first, 

 but when tlu \ .ire oxygenated, their next to the largest one dies 

 ln>t. What is the cause of this? The larger and more active 

 ones may require more food than the smaller ones, and for this 

 rra-on starve more quickly, if starvation is the actual cause of 

 death. Or it may be that the smaller ones are partly nourished 

 by their host or by their larger protozoan neighbors. In the 

 case t ygenation, the smaller ones may die first because of the 

 higher ratio of surface-volume exposure which they have. 



One interesting problem which this study brings out is, what 

 maintain-^ the fairly definite ratio between the four genera of 

 I >iot it/oar 1 The host may produce a reaction product for each 

 i;enn> \\liieh inhibits its multiplication beyond a certain point. 

 Ilout M i, tin.-- does not seem very likely, for when Tfichonympha 

 i- ienio\rd, /., i.jyopsis takes its place, and when Leidyopsis is 

 removed, Triiliotuonas partly takes its place. The protozoa 

 may inhibit the reproduction of each other be\<>ml a certain 

 point. And another possibility is the question of struggle for 

 food which mu>t ;<<> on where such a large number of protozoa 

 are present. 



SUMMARY AND CONCLUSIONS. 



The termite which was used in these experiments belong to 

 the genus Ternwpsis. Very probably two species, T. nrcadcnsis 

 Hagen and T. angusticollis Hagen, have been used. These two 

 species are so nearly alike that they can be distinguished at 

 present only by a study of their winged adults, which were not 



