Published online by Cambridge University Press: 15 May 2009
The present investigation deals with the routine bacteriological examination of 525 samples of water, 265 of which contained lactose fermenting bacilli; 1441 strains were isolated and various differential tests were used in the classification of these organisms.
All the strains were tested in dextrose phosphate culture for the methyl red and the Voges and Proskauer reactions, and although most of them gave normal results (M.R. + V.P. 0 or M.R. 0 V.P. +), there were 25 strains, 1·73 per cent., which gave either double positive or double negative reactions. Similar cases have already been discussed in an earlier portion of this paper, and they leave no doubt that there are types which may vary from the normal as far as the methyl red and the Voges and Proskauer reactions are concerned. At the same time these abnormal types are not sufficiently numerous to make any real practical difficulties in the use of these tests for routine purposes.
The Koser test, considered by many bacteriologists to be of great differential value, was found to give imperfect correlation with the other reactions.
1395 strains were isolated which could be classified in the B. coli group as a result of the production of acid and gas in lactose, the clotting of milk, and the non-liquefaction of gelatine, but when the methyl red and the Voges and Proskauer results were considered it was found that nearly 11 per cent. of these strains were of the B. lactis aerogenes type, and were therefore of no value as indicators of faecal pollution.
The indol reactions were somewhat variable. Among 1244 strains of the B. coli type (M.R. + V.P. 0), more than 10 per cent. failed to produce indol; while among 151 strains of the B. lactis aerogenes type (M.R. 0 V.P. +), nearly 51 per cent. were indol producers. Preliminary extraction with ether was carried out with 448 of these strains before Böhme's reagent was added. With this technique it was found that among 400 methyl red positive types there were 87 strains (22 per cent.) which produced no indol, and among 48 methyl red negative strains there were 12 (25 per cent.) which gave the reaction. Chen and Rettger (1920) were unable to establish any definite correlation between the indol test and the other reactions.
When the sample results were studied, it was found that although 262 waters contained organisms which were included in the B. coli group on the basis of the lactose, milk and gelatine tests, nearly 6 per cent. of these samples contained only organisms which gave a methyl red negative, Voges and Proskauer positive reaction, and were probably associated with pollution from soil washings rather than contamination from faecal sources. This means that the application of the methyl red and Voges and Proskauer reaction does make a considerable practical difference in the interpretation of results in the bacteriological examination of water.
More detailed examination as regards the fermentation of the various carbohydrates is useful for purposes of classification, but is of no significance, in the light of our present knowledge, in assessing the probability of excretal pollution.
It is suggested that, in considering the results of a bacteriological examination of water, organisms should be regarded as significant which are of typical morphology and staining reaction, produce acid and gas in lactose, fail to liquefy gelatine, clot milk, and give a positive methyl red and a negative Voges and Proskauer reaction. The production of indol in a medium containing peptone may be added as an extra test if desired, but the available evidence does not suggest that failure to produce indol is a sufficient basis for the exclusion of an otherwise typical bacillus from the B. coli group.
My thanks are due to Dr G. D. Dawson, in conjunction with whom the present investigation was started; and also to Prof. W. W. C. Topley whose interest throughout the course of the research was made manifest by his helpful advice and criticism.