SYMBIOSIS BETWEEN TERMITE AND FLAGELLATES. 313 



I.. \PERIMENTS AND OBSERVATIONS. 

 a. Starvation. 



Ten experiments, employing altogether approximately five 

 hundred termites, were carried out as follows: The termites 

 were removed from wood and carefully freed of all wood particles 

 which they attempted to cling to. Then they were placed in 

 large Petri dishes which were kept in moist chambers. In this 

 way the amount of moisture which previous work had shown 

 most desirable u.i- constantly supplied. All tin- way from one 

 to fifty individuals were placed in a Petri dish during starvation, 

 with tin- -ame rc-nlts in all instances. 



When starved in this manner, Termopsis begins to lose its 

 large, dominant and principal wood-ingesting protozoon, Tri- 

 ihonymphd ( I i^. i), by the end of the third day, and by the end 

 of the fourth day perhaps half the individuals of this genus 

 in. illy present are dead, although few, if any, of the other 

 genera of proto/oa have died. By the end of the fifth day, some 

 termite.- h.ive lost all their Trichonympha, while others still 

 retain a few slowly moving, apparently weak individuals. Hy 

 the end of the sixth day, no Trichonympha can be found in any 

 in mite-, and perhaps half the individuals of the next l.tr. 

 l>ri>to/n.iii, Leidyopsis (I'ig. 2), have died. And in a few termites 

 l>erh. ij i- .ill or ne.irly all of Leidyopsis may be dead, but this is 

 exceptional. Also, by this time a few individuals of the next 

 Miialle-t genus, Trichomonas (Figs. 5, 6), ma\ have died, but 

 not man\ . It -tarvation is continued through the seventh day, 

 -me termite- completely lose their infection of Leidyopsis, while 

 others harbor a few individuals until near the end of the eighth 

 da\ . After eight days of starvation, then, the two large protozoa, 

 Trichonympha and I.<id\opsis, have all disappeared entirely. If 

 the Marvation i- continued, Trichomonas now begins to die 

 rapidly and by the end of the tenth day nearly all the individuals 

 of tlii- i^enus are dead, though perhaps one to two per cent, of 

 the total number originally present live sometimes four or five 

 (.lays longer. Trichomonas struggles along and dies more slowly 

 than Trichonympha and Leidyopsis. All of these protozoa feed 

 on the wood particles which their host has eaten, but Streblo- 



