1. Solubility of a germicide in the medium (water or air) in which it is operating appears often to be the most important physical characteristic governing degree of activity on bacteria (partition law).

2. Evidence of the importance of solubility has only been derived from a study of the action of phenols and glycols, a high bactericidal activity never being registered unless solubility was low.

3. Other types of germicide were exceptional in that high solubility coincided with high bactericidal activity: iodine and maleic anhydride in the air; mercuric chloride and many organic acids in the test-tube.

4. Mixtures of germicide and bacteria as used in the test-tube were sprayed into the air, the bactericidal effect being usually increased or decreased according as to whether it was originally low or high in the test-tube.

5. The vapour pressures of most of the germicides used were deduced from the Clausius and Clapeyron formula, the boiling-points being determined by the bubbling method.

6. When the amount of a phenol used in the air is approximately proportional to the vapour pressure, the bactericidal effects of the different samples are more nearly parallel when the degree of saturation of the air is low than when it is high.

7. No phenol with a vapour pressure lower than 0·0021- mg. Hg, and tested at full saturation, sterilized the air during the first minute of the experiment, while all samples with, at least, 10 times this vapour pressure did so when at 12·5 % saturation.

8. Among the phenols tested, especially those having low vapour pressures, the concentration (w/w) in air is very much less than the concentration in water in the respective experimental conditions, where but few survivors is the criterion of efficacy. Owing to a number of different factors, the rapid decline in the initial concentration is more likely to be met with under air than under test-tube conditions.