1. This paper reports the results obtained by the cultivation of the bacteria of ten batches each of ten mussels, 100 mussels in all, in lactose bile broth.

2. From these mussels 371 cultures of bacteria were obtained. 30·7 per cent, of these were not Gram-negative bacilli, 10·0 per cent. were Gram-negative bacilli which did not ferment lactose and glucose, 59·3 per cent. were coliform bacilli.

3. By reading the results in lactose bile broth after 24 hours' incubation, and by regarding as positive those tubes which show strong acid and good gas production the most reliable results were obtained. 93·0 per cent. of all typical, and 83·8 per cent. of all typical and atypical B. coli found gave this reaction. 37·6 per cent. of the organisms giving this reaction were typical B. coli, 61·1 per cent. of them were typical or atypical B. coli, and 97·2 per cent. of them were coliform bacilli.

4. The writer's method is described in detail. Its essentials are the rejection of shell fluid, the standardisation of the volume of each mussel at 25 c.c. by the addition of saline, the inoculation of three volumes of mussel mince into lactose bile broth—0·5, 0·1 and 0·02 cc., each in duplicate, and the regarding as positive each of these amounts only when acid and gas are produced in 24 hours in both tubes.

5. It is suggested that a batch of mussels showing more than seven positives in the tubes inoculated with 0·5 cc.. of mince, or more than three in the tubes inoculated with 0·1 c.c. or more than one in the tubes inoculated with 0·02 c.c. should be considered as contaminated to an undesirable extent.

6. Three types of solid media—Salle's, MacConkey's and Burke-Gaffney's— were tested. None were considered suitable for the bacteriological examination of mussels.

7. Reasons are given for considering any solid medium inferior to fluid medium for the bacteriological examination of mussels.