2 
THOMPSON YATES LABORATORIES REPORT 
The identification of any particular bacillus may be either a very easy or a very difficult 
matter. The detection of tubercle bacilli in sputum can be pretty safely entrusted to any 
conscientious medical student who has done his course of practical pathology : the determination 
of typhoid bacilli in the excreta of a patient may require all the resources of an experienced 
bacteriologist. During the last two years, however, the development and proper understanding 
of previous isolated or misinterpreted observations has given us a new and generally trustworthy 
method of recognizing many species of micro-organisms. And more than this, by a simple 
reversal of the procedure it has been possible to supply a new means and aid to clinical diagnosis— - 
another addition to practical medicine derived from experimental science. 
So far back as 1889, Charrin and Roger (3) demonstrated that the growth of B. 
pyocyaneus (and of other bacteria) differed when made in the serum of animals immunized 
against B. pyocyaneus and when made in the serum of normal animals. The differences 
consisted in (i) serum remaining more, transparent, (ii) formation of granular agglomerations, 
and (iii) microscopically, the formation of chains of altered bacteria, in the former mode of 
cultivation. 
In 1 89 1, Metschnikoff, whilst investigating Vibrio Metschnikovi^ saw the same pheno- 
mena, and noted them as facts requiring further investigation (10). But he made no progress in 
this direction, being discouraged by failure to obtain the reactions with the bacillus of hog 
cholera. 
In 1893, IssAEFF made similar observations with Pneumococci, and later in conjunction 
with IvANOFF, with Vibrio Ivanovi. They attributed the effects obtained to the bactericidal and 
growth-hindering substances in the serum (8). 
In 1895, Washbourn saw similar appearances, and remarked that the mode of growth 
in the serum seemed to give a good indication of its protective power (13). 
All these observers appear to have worked with undiluted serum only, and that too acting 
for a considerable time. Although they examined the deposit with the microscope, they do not 
appear to have investigated, microscopically or otherwise, either the immediate effect of the 
serum on the bacilli, or the effect of diluting the serum. 
In 1894, R, Pfeiffer published what is now known as Pfeiffer's reaction (11), but it was 
BoRDET, who, in 1895, trying to simplify Pfeiffer's reaction, first described any observations in 
whicli the above-mentioned neglected aspects of the question were considered. He saw the 
agglomeration and loss of movement, but failed to attribute any specific importance to the 
reaction [2a). 
It was therefore left to Durham and Gruber, who had already been more than six 
months at work on the subject when Bordet's investigations were published, to show how 
valuable and characteristic this reaction could be made, and to establish its general applicability 
(4^, 617). 
In March, 1896, I applied the reaction to the diagnosis or enteric fever in man, but 
scarcity of material and other delays prevented any paper on the subject appearing before 
September (7^). Meanwhile, Widal, profiting by a hint thrown out by Gruber at the 
