H. MUNEO FOX 567 



collecting there in ever increasing numbers (Fig. 2) , This collection 

 is not a surface energy phenomenon which might be caused by any 

 solid object in the suspension, for if a larva which has just been 

 killed by being dropped into hot water is used in place of the living 

 larva, there is no aggregation of flagellates at its surface. The Bodo 

 are simply attracted towards the surface of the living larva because 

 in this neighborhood the respiratory activity of the insect has lowered 

 the concentration of oxygen dissolved in the water. The reason for 

 the collection is the same as that for the spontaneous aggregation 

 previously described.^ The flagellates move into a region of lower 

 oxygen concentration caused, in the case of spontaneous aggregation, 

 by their own respiration, in this experiment by the respiration of the 

 larva. Here, of course, the aggregation of flagellates takes place 

 much more rapidly than in spontaneous aggregation for the larva 

 consumes relatively much more oxygen than the flagellates. If spon- 

 taneous aggregation occurred rapidly, it would interfere with the 

 collection of the flagellates on the respiratory surfaces of the larva. 

 The danger of spontaneous aggregation is, however, averted by a 

 preliminary filtering of the Bodo suspension through bolting-silk. 

 This not only removes fragments of debris from the culture jar but 

 the exposure of the water to the air during filtration allows it to 

 become saturated with oxygen. This greatly retards subsequent spon- 

 taneous aggregation.^ 



More flagellates collect on some parts of the insect's surface 

 than on others. These are the most actively respiring surfaces. 

 When the aggregation has reached certain dimensions, a clear space 

 free from flagellates appears between it and the surface of the larva 

 (Fig. 3). This is due to the same cause as the appearance of an 

 area free from flagellates in the center of a spontaneous aggregation; 

 the dissolved oxygen has been reduced to a certain low concentration 

 at which the flagellates leave the region. The aggregation of flagel- 

 lates on the surface of the larva now gradually becomes a band sepa- 

 rated from the surface by a clear space, this happening first at those 

 surfaces which are most actively respiring. Eventually the flagel- 

 lates move out everywhere from the surface of the larva and come all 

 to lie in a band encircling it (Fig. 4) . In the case of an animal, such 

 as a Simulium pupa, which respires only in one part, namely through 



