314 L. R. CLEVELAND. 



mastix (Fig. 4) does not. It may be dependent on the other 

 protozoa or on the termites for its nourishment. It is difficult 

 to say just when Streblomastix begins to die since this genus is 

 much smaller than either of the other three and in the normal 

 or not-starved termite is greatly obscured by the countless 

 thousands of larger individuals, thus making it difficult to deter- 

 mine accurately the normal number present ; but the number has 

 perhaps diminished some by the end of the tenth day and by 

 the fifteenth day a great many have died, though not all; some 

 of them, in fact, live almost as long as the termites three to 

 four weeks. 



During starvation most termites are active and appear normal 

 for about fifteen days. However, as soon as many of the protozoa 

 have died, after six days, say, it is not necessary to make a micro- 

 scopical examination of the intestinal contents to determine 

 what has happened, for there is now much more fluid than 

 formerly present and it looks very muddy the difference is 

 quite characteristic and cannot be mistaken. 



Why do termites, when not given food (wood), lose their proto- 

 zoa? Do the protozoa die of actual starvation, and most of 

 them much more quickly than their host? Three experiments 

 were carried out which perhaps throw some light on this question. 

 When termites are fed cellulose instead of wood for several 

 months before being starved, and then are cellulose-starved, they 

 lose their protozoa more slowly. For instance, it takes them at 

 least one to two days longer to lose Trichonympha. The 

 writer ('250) has shown elsewhere that this termite (Termopsis) 

 can live for more than a year and perhaps indefinitely in a 

 perfectly normal manner on a diet of pure cellulose and may it 

 not be true that when it is fed nothing but cellulose (wood is 

 about 50 per cent.) for some time before being starved that it 

 really has more food in its intestine for the protozoa when 

 starvation is begun and for this reason the protozoa are able to 

 live longer? If so, this indicates that when Termopsis is wood- 

 starved that its intestinal protozoa, particularly Trichonympha, 

 Leidyopsis, and Trichomonas, die of actual starvation long before 

 their host. We should expect the protozoa to die first since 

 they digest the wood for themselves and their host. And when 

 they die, they perhaps give themselves as food to their host, 



