A Study of the B. coli Group. 
333 
that auto -agglutination was rarely seen among these bacilli. The tubes were 
placed in the incubator for one and a half hours, and at the expiry of tiiat 
period, at room temperature for half an hour; they were again replaced in the 
incubator for two hours, when readings were taken of the results. 
Ultimately they were allowed to stand at room temperature till next day, 
when further readings, if necessary, were made. It was usually 
noted that the variations in temperature produced in this way set ap con- 
vection currents in the fluid and this hastened agglutination. The degree of 
agglutination was determined by the amount of sediment in the various 
tubes as compared with the control or by the clarity or turbidity of 
the supernatant fluid as compared with the fluid in the control tube. 
Complete agglutination is signified in the tables by + + + +, and 
lesser degrees by + + + , + + , and + . 
Results Observed with Antisera to A Types 1, 2, 3 and 4. — Marked 
specificity for the individual strain on the part of these immune 
sera was observed. Tables VII, VIII, IX and X show that the only strains 
agglutinated to any extent by the correspo7iding antiserum were the particu- 
lar strains used for immunisation. Thus, the antiserum to strain 1 type 1 
agglutinated strain 1 in dilutions up to 1 : 50,000 ; 14 other strains of the 
same type were tested with the antiserum but none showed any agglutina- 
tion by dilutions higher than 1 : 500, and 8 were not even agglutinated by a 
dilution of 1 : 100. A strain of type 3 also exhibited little reaction with the 
type 1 antiserum (Table VII). Similar results were obtained with antisera to 
types 2, 3 and 4. 
While this restricted specificity was found to be the general rule with 
antisera to these common types of B. coli an exception has been noted : an 
antiserum to a type 2 (B. coli communis) was found to agglutinate a 
particular strain of type 1 (B. MacGonhey No. 71) in a four times higher 
dilution than in the case of the homologous strain (Table XI), and this type 1 
strain was not found to be specially susceptible to other B. coli agglutinins 
(Tables IX and X), i. e. it was not itself susceptible to other agglutinating 
sera nor did it show any tendency to auto-agglutination. Moreover, this 
agglutinin had no effect on a number of other strains of type 1 (Table 
XI). 
To ascertain whether this peculiarity was a function of the immune 
animal, another antiserum to the same type 2 strain was obtained and the 
serum behaved in practically the same manner, indicating that this property 
of " paragglutination " was dependent on the particular strain (Table XI). 
It is noteworthy that this instance of paragglutination is characterised by 
the more powerful effect of the paragglutinin than the primary agglutinin. 
As is well known, the agglutinin has marked affinities for the homologous 
bacillus, and is absorbed or used up by the organisms during the process of 
agglutination. Moreover, a bacillary emulsion is capable of absorbing much 
