SERUM DIAGNOSIS 171 



be well beyond the range of error are, as a rule, 

 accompanied by other changes which constitute 

 a more reliable means of diagnosis. The varia- 

 tion in the counts is much greater than most ex- 

 aminers believe. North has collected a series 

 of tests, carried out in nearly all the important 

 laboratories in the Eastern United States that 

 are working upon opsonins. The results re- 

 corded prove absolutely that, while an av- 

 erage counting error of only about ten per cent 

 is present, there may be an exceptional error of 

 at least 100 per cent ; and one of at least 20 per 

 cent may be expected once in about every ten 

 determinations, even by experienced workers. 

 Most value was placed by Wright upon the 

 opsonic index as a guide to the size and fre- 

 quency of doses of bacterial vaccines in the 

 treatment of disease. As a routine measure, 

 however, the opsonic index has fallen into dis- 

 use, and vaccine therapy is being largely guided 

 by the clinical evidences of reaction and the con- 

 dition of the patient. 



