RELATION OF BURSARIA TO FOOD 9 



The uniformity in size of the yolk grains is of course of para- 

 mount importance in many of the experiments and for some of 

 the conch] sions which will be drawn from them. In order that 

 the degree of uniformity might be tested and indicated quanti- 

 tatively, a large number of measurements of grains of the pre- 

 pared yolk suspension were made at different times by means 

 of a stage micrometer. The following shows a typical result 

 of one set of these measurements: 105 grains were measured and 

 the numbers divided at random into three sets of 35 each and 

 the average of the diameters taken. These gave respectively 

 0.0890, 0.0906, and 0.0837 mm. The range of variation of the 

 diameter was from 0.060 to 0.130 mm. The distribution of the 

 variations are shown by the following figures: 



Diameter of grain mm 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.13 



Frequency 12 19 16 14 20 13 6 5 



By thus being able to obtain a very constant average diameter 

 of a comparatively small number (30 to 50 grains) the errors 

 introduced by the individual variation in size, which in the above 

 example is about as ^ is to 1, is largely eliminated. In order to 

 remove the objection to experimental results based on the vol- 

 ume of granules of varying size, a large number of individuals 

 (20 to 100, depending upon the purpose of the experiment) were 

 used in each experiment and the number of grains eaten was 

 counted; furthermore the experiments were always repeated 

 whenever there could be any doubt as to the validity or signifi- 

 cance of the results. Hence, as will be shown later, the lack of 

 strict individual uniformity of the unit volume is corrected (a) 

 by the fact 'that the average size of the yolk grain is practically 

 constant, (b) by using a large number of individuals in each 

 experiment, and (c) by repeating the experiment. 



Thus having the form, weight and volume of the units of food 

 eaten made practically constant, we may vary one of their prop- 

 erties — as for example, their chemical nature — by letting them 

 adsorb different kinds of toxic and non-toxic substances which 

 are diffusible or non-diffusible in the native medium, tap or dis- 

 tilled water. We may therefore test the responses to variations 



