sion from the results indicated in Table I may be that, considered from a purely bactericidal aspect, 3 per cent sodium hydroxide is the best of the materials tested.

TESTS IN LABORATORY DISHWASHING MACHINE

Experiment 1. Forty-five clean Petri dishes were soiled by filling them with ordinary bottled milk. The milk was poured off and the dishes, in the open and uncovered, were allowed to dry for 24 hours at room temperature. At the end of this time, fifteen of the dishes were washed with water alone in accordance with the technic (8) described previously, using a 30-second preliminary spray, a 3minute wash, and two 30-second rinses. Fifteen more of the soiled dishes were washed in a similar manner, using in the wash T.S.P. in a concentration approximating 0.23 per cent. The last fifteen dishes were washed in the same manner, using the same concentration of C. The water temperature was between 55° and 60° C. At the end of the final rinse, one dish was taken from each set, filled with nutrient agar, covered, and incubated for 48 hours at 37° C.; then the colonies were counted. The remaining fourteen dishes in each set were resoiled as described and rewashed the following day, when another dish from each set was removed for incubation. This procedure was continued for 15 days in exactly the same way. The average hardness of Pittsburgh city water for this 15-day period was 115 p.p.m. calcium carbonate. The results are presented in Table II.

TABLE II.-PLATE COUNTS ON DISHES REPEATEDLY SOILED WITH MILK AND WASHED

Experiment 2. Four Petri dishes were soiled in the same manner by filling with milk, emptying, and drying uncovered at room temperature for 24 hours. This soiling was repeated for 15 days without any washing during the whole period. By the fifteenth day each of the dishes was covered with a hard, dry crust. On the sixteenth day they were washed in the way described, with water, T. S. P., and C, respectively. Each plate was carefully placed in exactly the same position in the machine, so that the water forces during each of the washings would be identical. Several large pieces of dried milk residue remained on the water-washed dish. The dish washed with T.S.P. held a few pieces, about 1 sq. cm. in area; the dish washed in C was free from any soil whatever. The dishes were covered with nutrient agar and incubated for 45 hours, and the colonies were counted (Table III). The hardness of the water was 11S p. p. m. calcium carbonate.

TABLE III.-PLATE COUNTS ON DISHES ACCUMULATIVELY SOILED WITH MILK AND WASHED

TABLE IV.--PLATE COUNTS ON DISHES SOILED WITH B. PRODIGIOSUS-INOCULATED MILK AND WASHED

TABLE V.-PLATE COUNTS ON DISHES SOILED WITH MOUTH WASHINGS AND INOCULATED MILK AND THEN WASHED

On the first two dishes the presence of pieces of unremoved soil made an accurate count impossible; therefore these data are to be regarded as minima. The dish washed with C was perfectly clean and hence the figure is accurate as given. Experiment 3. Sterile whole milk was inoculated with B. prodigiosus. This organism ordinarily withstands a temperature of 60° C. and is easily differentiated from others because of its red color. Sterile Petri dishes were soiled with 1.5 cc. of the inoculated milk and allowed to dry for 24 hours. A count of the bacteria present in the milk, made at the time of soiling by the dilution-plate method, showed that each plate was inoculated with over 54 million bacteria. Naturally the bacteria multiplied rapidly in the incubator, which brought the inoculum on each plate into the billions.

One plate at a time was washed in the manner described, and each plate was carefully placed in the same position in the machine. A total of five plates was washed with each of three materials. These plates were then covered with nutrient agar and incubated for 48 hours, when plate counts were made. The hardness of the water used was 120 p.p.m. calcium carbonate. Average results are given in Table IV.

Such counts would of course include any organism in the rinse water; but counts on sterile dishes run through the same procedure with water alone failed to develop colonies, showing the absence of organisms in the water.

Experiment 4. Petri dishes were soiled with milk inoculated with washings obtained by rinsing the mouth with sterile salt solution; 1.5 cc. of the inoculated incubated milk was added to each dish, which was allowed to dry for 24 hours. The dishes were then washed as described. The average hardness of the water during this experiment was 124 p.p.m. calcium carbonate. After washing, the dishes were filled with nutrient agar and incubated for 48 hours, after which time the colonies were counted. The results are given in Table V.

It is interesting to note that on the fifth day the plates were rinsed in sterile distilled water. On the sixth day the plates were not rinsed at all.

Experiment 5. Forty Petri dishes were soiled with milk that had been previously inoculated with B. subtilis. The dishes were dried in an incubator at 37.5° C. for 24 hours. Ten of them were washed in 3 per cent sodium hydroxide in the manner previously described: another set of ten was washed in 0.25 per cent C; a third set was washed once in 3 per cent sodium hydroxide, followed by another washing in 0.25 per cent C; the remaining ten dishes were washed in 3 per cent sodium hydroxide. The average hardness of the water during this experiment was 85 p.p.m. calcium carbonate. The plates were covered with nutrient agar and incubated for 48 hours, when the colonies were counted. The results shown in Table VI were obtained.

Tables II to V indicate results which are not surprising. These findings show that the metaphosphate cleaner, C. yields dishes that are cleaner biologically than those washed with either water or T.S.P. under similar conditions of soiling and procedure.

Figure 1 presents graphically the relation between the results of the various experiments described. The dishes washed in plain hot water had on them considerable numbers of bacteria. Those washed with T.S.P. had fewer colonies, and those with C had still less, as might be expected from the results given in Table 1 and from a knowledge of the cleaning properties of this material. If the numbers of colonies counted on plates washed in all four of these experiments with each of the three materials are averaged, we find that the average number of colonies on the water-washed dishes was 185, on the T.S.P.-washed dishes, 63, and on the C-washed dishes, 9. If we arbitrarily assign to the water a value of 0 per cent bateria removal and destruction, we find that the T.S.P. removed or

killed 66 per cent of the bacteria and that C removed 95 per cent. When viewed in conjunction with each individual experiment, this averaging is indicative, at least in a broad sense, of the difference in results obtained with these materials.

Table VI. Plate counts on dishes soiled with B. subtilis-inoculated milk and washed

The experiments on dishwashing thus far described show merely that the use of C leads to more sanitary dishes, but they do not provide an explanation of this improvement. C may give better dishes, either because it effects physically cleaner dishes than T.S.P., as has been pointed out, or because it is a better germicide than T.S.P., which may be gathered from the data in Table I. Experiment 5, which was a comparison between 3 per cent sodium hydroxide solution and 0.25 per cent C solution, indicates that in actual washing operations the cleaning ability of a material is just as important as (if not more important than) its germicidal power. Three per cent sodium hydroxide solution has for some time been regarded as an efficient germicidal cleaner and is proving very useful in the field of beverage bottle washing. The results in Table I indicate, too, that this material is a better germicide than any of the others tested. Yet in experiment 5, where Petri dishes were soiled with milk inoculated with B. subtilis, the comparatively low concentration of C that were used to give results far superior to those of ?— with 3 per cent caustic solutions. Owing to the fact that the actual bactericidal efficiency of sodium hydroxide is higher than that of C, it must follow that the dishes washed with 3 per cent caustic were not so clean as these washed with C. The result, therefore, is attributed to the fact that sodium hydroxide is inferior to C as a detergent, rather than to the fact that it leaves a residue, because sodium hydroxide in water of the character of Pittsburgh water would lead less to film formation than the other commonly used alkalies. This

reasoning would not necessarily apply to a water containing bicarbonate hardness. At present, work is being continued for the purpose of determining its continuous large-scale operations, such as exist, for example, in dairies, the commonly used 3 per cent alkali solutions may be replaced with more dilute solutions of C with consequent improvement in results.

TESTS IN RESTAURANTS

These experiments were performed in two large restaurants, each of which was quipped with a modern mechanical dish-washing machines. Both machines were

typical of those usually met in first class restaurants, and the operating conditions of the two establishments were believed to be average and representative. Thorough examinations of the machines indicated that they were in good operating condition before the test period was started. No changes in procedure or operation were suggested by the investigator, because, for the purpose of this work, the results of ordinary practice were desired.

At the time the work was begun in resturant A, there was being used a common dishwashing material composed almost entirely of trisodium phosphate. The machine used in resturant A was a single-tank machine of the automatic conveyor type with direct-connected rinse. During wash and rinse the dishes were sprayed both from above and below. The average temperature of the wash water was 60° C., that of the rinse water was 82° C., and the average time of exposure to the hot rinse was one minute. Ten to fifteen dishes were collected daily and at random as they emerged from the dishwashing machine; they were carefully placed in previously sterilized paper bags and brought back to the laboratory; they were removed from the bags and on them were poured 35 cc. of nutrient broth. The dishes were immediately covered with previously sterilized pie tins and placed in an incubator for 2 hours. At the end of this time 1 cc. of broth from each dish was plated on agar, and after an incubation.period of 48 hours the colonies were counted.

When two weeks had elapsed, C was substituted for the trisodium phosphate cleaner and the same technic continued. No other changes were made. In so far as could be determined, all conditions remained the same as they had been when T.S.P. was used. C-washed dishes were collected for 2 weeks and treated as just described.

The machine used in restaurant B was a double-tank machine of the manual push-through type. The average washing temperature was 54° C., that of the rinse was 1.5 to 2 minutes. In this restaurant, which was using C when the work was started, dishes were collected for 2 weeks, after which time a T.S.P. cleaner was employed for 2 weeks. The results are as follows:

These experiments show that, under the particular conditions obtaining in each place, dishes washed in C were bacteriologically cleaner than those washed with T.S.P. The results in each of the two restaurants are quantitatively different, and this difference may be ascribed to differences in the washing procedure at each place. No change was made in the actual technic of washing other than change in cleaning material. In restaurant B the work was done carefully, the dishes were racked as they should be, and the washing was neither careless nor hurried. In restaurant A less care on the part of the washman yielded less satisfactory results. But in both restaurants, in spite of rather marked differences in the conditions existing, brighter, cleaner dishes were secured with C than were obtained with the other materials.

The superior sanitary quality of the dishes washed with the sodium metaphosphate preparation does not proceed from any specific germicidal action but must be attributed to its effects in facilitating both the washing action of the alkaline detergent and the rinsing. Thus the dishes are not only freed from soil, but they do not accumulate any organic or inorganic film that results from precipitation during cleansing.

Sanitary regulations governing machine dishwashing vary in different localities. The sanitational value of hot alkaline detergent solutions is well recognized. It is also recognized that, in machine washing, dishes may be suitably disinfected by exposing them to hot rinse water for a sufficient period. As an alternative, when an adequate supply of hot water is not available, the use of an auxiliary germicide, usually a compound of chlorine (2), is specified. In hand dishwashing, in which low temperatures and low alkalinities prevail, the use of an auxiliary germicide is well justified, especially in public eating places. Wherever absolute sterilization is demanded, it may be more readily obtained by the use of germi

cidal solutions when sodium metaphosphate is used in the washing process because of the nonexistence of protective films. Thus more effective and economical use of the germicidal solution is made possible.

Film formation in dishwashing is a source of great concern to health and restaurant officials because such films form an ideal environment for bacterial growth. It has been demonstrated that the utilization of the film-preventive function of sodium metaphosphate represents a distinct improvement in dishwashing sanitation. Moreover, the dishes washed present an unusually attractive appearance, owing to their extreme physical cleanliness. Thus the practical removal of dishwashing film represents a contribution to both sanitational and esthetic progress.

LITERATURE CITED

(1) Buchanan, J. H., and Levine, Max, American Bottlers of Carbonated Beverages. Educational Bull. 1 (rev.) 17 (1929).

(2) Cumming, J. C., and Yongue, N. E., Am. J. Pub. Health, 26, 237 (1936). (3) Devereux, E. D., and Mallman, W. L., Ibid., 26, 185 (1936).

(4) Gilmore, B. H., Oil & Soap, 12 No. 2, 29 (1935).

(5) Hall, R. E., U.S. Patent Reissue, 19,719 (Oct. 8, 1935); U.S. Patent 2,635,652 (March 31, 1936).

(6) Purdy, A. C., Marine News, 21 S3 (1935).

(7) Reddish, G. F., Am. J. Pub. Health, 17, 320 (1927).

(8) Schwartz, C., and Gilmore, B. H., IND. ENG. CHEM., 26, 933 (1934).

(9) Smith, G. W., Am. Dyestuff Rep., 23 161 (1934).

Received October 19, 1936. Contribution from the Calgonizing Fellowship at the Mellon Institute of Industrial Research.

[Reprinted from Journal of Dairy Science, July, 1940, Vol. XXIII, No. 7, pages 621-627] THE VALUE OF SODIUM METAPHOSPHATE IN DETERGENT MIXTURES IN THE CLEANING OF MILKING MACHINES 1

(By W. L. Mallmann and C. S. Bryan, Section of Bacteriology, Michigan Agricultural Experiment Station, and L. H. Bergman, Lansing Department of Health)

The process of cleaning the milking machine has claimed the attention of the dairy sanitarians ever since the advent of this device. Although numerous publications have appeared from time to time recommending various methods of cleaning and sterilizing, and although these methods can be applied with considerable success, the fact remains that milking machines are still serious foci of contamination in market milk. The mechanical structure of the machine, even though marked simplication have been affected, makes the cleaning and disinfecting procedure difficult when compared to the equipment necessary for hand milking. This does not mean that the equipment is really difficult to clean, but it does mean that the method of care is different and peculiar to this equipment. The milking machine is such a valuable adjunct to the dairy farm that the more or less insanitary conditions of these machines that prevails should be rectified by better methods of cleaning so that their use will not be discredited.

Not only may the dirty milking machines be an important source of bacterial contamination, but they may also be a reservoir of heat-resistant bacteria, which are responsible for the high count occasionally obtained in properly pasteurized milk. The role of milking machines as a source of heat-resistant bacteria is not unknown to the dairy sanitarian. Many times the dairy plants are accused of insanitary conditions or carelessness when most of the trouble lies with the producers who have been careless in the handling of their milking machines.

EXPERIMENTAL

In the spring of 1938, a large dairy in Michigan was having considerable trouble with high counts in its pasteurized milk. On several occasions samples were collected on three different routes from procedures using milking machines. These machines had received little attention from the milk inspector so they represented conditions which may prevail under lax inspection. The producer samples were tested by the standard plating procedure. Each sample was then pasteurized and another plating made to determine the reduction affected by this heating.

1 This study was made possible through a grant from Calgon, Inc., of Pittsburgh, Pennsylvania. Journal Article No. 363 n.s. from the Michigan Agricultural Experiment Station.