1. The surface taint butter examined often contained large numbers of bacteria, as determined by the plate method, and with some of the samples the counts were very high; with a few samples the counts were comparatively low. The counts were higher on the surface portion of a sample than on the interior portion in all but 1 of 20 comparisons.

2. The surface taint butter examined usually contained large numbers of yeasts, with the surface portion of a sample commonly containing a larger number than the interior portion; the butter showed some mold counts that were high and a few that were surprisingly high, the counts on the surface portion generally being higher than on the interior portion.

3. The general types of bacteria found in surface taint butter by picking colonies into litmus milk from beef infusion agar plates were essentially the same as those found in any lot of butter containing considerable numbers of organisms.

4. Surface taint could not be produced in butter by inoculating a normal product, either salted or unsalted, with surface taint butter but could be developed by inoculating the defective butter into pasteurized cream and churning the cream; from 2 to 4 days were required for the surface taint to develop at 15.6°C. (60°F.) and from 7 to 10 days at 5°C. (41°F.).

5. Surface taint sometimes developed in commercial butter held at temperatures very favorable for bacterial growth; in most cases the butter that developed surface taint was unsalted or had a low salt content.

6. Excessive numbers of organisms were found when either experimental or commercial surface taint butter was examined microscopically.

7. The organisms which predominated on beef infusion agar plates poured with surface taint butter did not produce surface taint when inoculated into pasteurized cream and the cream churned.

8. An organism capable of producing surface taint when inoculated into pasteurized cream and the cream churned was first isolated from a sample of Canadian butter by plating on beef infusion agar and picking colonies into litmus milk; at the time of the appearance of surface taint in experimental butter the numbers of organisms per milliliter, as determined by the plate method, were comparatively small. The organism was believed to be an undescribed species and was tentatively designated Achromobacter putrefaciens.

9. By the use of an enrichment method consisting of inoculating butter into litmus milk, holding this at 5°C. (41°F.) and then plating on beef infusion agar and picking colonies into litmus milk, A. putrefaciens was secured from five additional samples of surface taint butter, three from Canada and two from the United States. From a considerable number of surface taint samples A. putrefaciens could not be secured.

10. Organisms, other than A. putrefaciens, which would produce surface taint were isolated from a number of samples of surface taint butter. These organisms were always found in small numbers.

11. Organisms which would produce surface taint were isolated from a total of 17 samples of commercial surface taint butter, 6 from Canada and 11 from the United States.

12. Organisms capable of producing surface taint could not be isolated from a considerable number of samples of surface taint butter, although with some of the samples the defect could be carried through a series of experimental churnings by using defective butter to inoculate the cream.

13. The organisms which would produce surface taint were greatly restrained by the use of medium salt percentages or butter culture in the making of butter. A. putrefaciens failed to grow in skimmilk acidified with lactic acid to 0.30 or 0.31 percent but did develop when the milk was acidified to 0.27, 0.28 or 0.29 percent.

14. In the trials carried out with Ps. fluorescens by inoculating pasteurized cream and churning the cream, rancidity regularly developed.

15. There appeared to be rather distinct variations in the samples of butter sent to the laboratory as examples of surface taint.