SEA-FISHERIES LABORATORY. 443 



react positively to at least three out of the four tests 

 employed" (Houston, 1904, p. 236). Thus even the 

 meagre array of proof originally suggested by this 

 bacteriologist appears to be too much in actual public 

 health practice. 



This will be seen by considering Houston's 

 " Quintuple Preferential Bacillus Coli Test,"* a title 

 which at once suggests some really well-planned method 

 of analysis. Let us suppose we use Houston's four tests 

 to identify a bacillus — if all are positive the identifica- 

 tion is made. But it may be that one or more of the 

 tests fail; each of them is then given a preferential 

 value, thus fermentation of glucose = 2, fermentation 

 of lactose = 1, production of indole = \, production of 

 fluorescence = \, and a negative result with cane- 

 sugar = £. The total value of the five tests is therefore 

 3|, and the more nearly the value of the tests approaches 

 to 3^ " the stronger would be the evidence derived from 

 its presence in favour of recent pollution by matter of 

 excremental origin." 



At first sight it sounds quite reasonable, but on 

 looking into the matter it is not difficult to detect the 

 confusion of thought involved in the suggestion. The 

 covert suggestion is that of a comparison with the results 

 of a chemical analysis : a solution of prussic acid is 

 always prussic acid, whether it contains 5 per cent., or 

 0"5 per cent., or 0"00005 per cent. Let us regard the 

 characters of a microbe then as all of the same order, so 

 to speak, and give each a percentage value; the more 

 of these characters the greater the percentage. Thus we 



* I have been unable to procure a copy of the publication in which 

 this test was described (Minutes of the Metropolitan Water Board for the 

 year 1907). 



I therefore describe it as used by Buchan , Journal of Hygiene, Vol. 

 X, p. 476, 1910. 



