containing an excess amount of fat were, -wiienever possible, 

 avoided.' Those containing agglutinins for the experimental 

 cells, although thorou^^hly annoying, can be used without in- 

 validating the results. '-Vhile a few hours difference in the 

 age of the sera results in no demonstrable difference in their 

 opsonic reactions, all sera for an experimental series were 

 collected as nearly as possible at the same time. 



The phagocytic mixtures were prepared according to the 

 Wright technique, and incubated at 37° for twentj'- minutes. 

 The smears and stains were made in the manner described by 

 Gross (18). All smears showing gross differences in the num- 

 ber or distribution of the leucocytes were discarded and the 

 preparations repeated. The organisms in fifty polymorpho- 

 nuclear leucocytes were counted and these cells were always 

 enumerated from corresponding areas on the control and experi- 

 mental slides. Cells containing an excessively large number 

 of bacteria show a tendency to collect in portions of the smear 

 which can be predicted, and the most confusing and contradic- 

 tory deviations present themselves if the selection of corres- 

 ponding areas for enumeration is not rigorously observed. All 

 preparations exhibiting marked variations from the normal or 

 expected were repeated throughout. Whenever the phagocytes 

 revealed unusual inequalities in the number of ingested organ- 

 isms, parallel preparations were made and the average enumera- 

 tion ta^cen as the true count. Polymorphonuclear leucocytes 

 alone were considered, and no attempt was made at enumeration 

 in cells containing m.ore than twenty-five organisms. In all 



(8) 



