THE success attained by chemists in preparing purified or concentrated extracts of tissue containing enzymes and other biological products has stimulated similar attempts upon the tissues of animals infected with filterable viruses.1 2 8 The virus of vaccinia has received considerable attention in this respect, due in part perhaps to the economy and facility of laboratory manipulations and to the accessibility of applying practically, some of the observations made. Thus, in the last few years, many investigators have reported on adsorption,4 5 purification,6' 7, 8 and filtration 9, 10, lo experiments with vaccinia virus. We wish to report a number of observations which after further investigation should lead toward a highly purified and readily filterable vaccinia virus. In order to give a clear understanding of the reasons for our steps it is necessary to describe the actual sequence of our experiments. An attempt was made to determine the duration of survival of vaccinia virus in milk. The Levaditi neurostrain, maintained by testicular passage in rabbits, was employed. Two series of flasks were set up as follows: