will not exceed 50° C., and also, in several cases, by the inoculation with lactic acid bacteria.

As to the fermentative and cheese-making adaptabilities, milk may, without perceptible variations in its organoleptic characteris tics, be unsuited for cheese making because of modifications in its chemicozymatic constitution, in its relation to rennet and its restraining or germicidal action. These modifications which can not be corrected by pure cultures are a result of secretory troubles induced by an abnormal mammary bacterial flora, which can create cellular reaction even without noticeable inflammation in the udder. To prevent the occurrence of such results by avoiding stagnation of milk in the udder, the milking must be done in a complete, proper, and correct manner.

(2) Quality of pure cultures.-It is advisable, in principle, to use a mixture of simple lactic bacteria, indicated by Freudenreich (which do not attack the casein in acid reaction), with my lactoproteolytic ferments which render casein soluble even in acid medium. This is independent of their coccoid or bacillary form, for it is their function which counts.

(3) Methods of using pure cultures.-The milk, before being put into the vat, must be held at a very low temperature, possibly around 5° C., in order that the number of bacteria of all kinds, including both dangerous and favorable bacteria, be held as low as possible. It would always contain, nevertheless, a certain quantity of lactoproteolytic bacteria which I have shown exist normally in the mammary bacterial flora and contribute to the ripening of the cheese. The inoculation of selected cultures serves to make up for the lack of other types of ferments.

During the working it is always necessary in each operation to carefully watch the manufacture of the cheese as if it were a matter of preparing a culture of caseous microbes.

THE CONNECTION BETWEEN THE BACTERIAL CONTENT OF MILK FROM WHICH CHEESE IS MADE, AND THE RAPIDITY OF RIPENING OF THE CHEESE.

CHR. BARTHEL, Ph. D., and E. HAGLUND, departments of bacteriology and dairy husbandry, Central Agricultural Experiment Station, Experimentalfaltet, near Stockholm.

In two communications from this experiment station, one of us (Barthel) (1, 2), has shown that lactic acid bacteria belonging to the group Streptococcus lactis possess the power of decomposing paracasein to a marked degree, especially at a temperature of 14 to 20° C. i. e., at the usual cheese-ripening temperature. The fact that earlier investigations of the casein-splitting powers of such lactic acid bacteria usually gave negative results, is explained by the much higher temperatures at which they were carried out, usually at 35° C., and never under 20° C. Prior to our investigations referred to above, it was generally supposed that the action of the common lactococci in cheese ripening was chiefly of an indirect nature, i. e.. that by forming lactic acid they would cause the formation of the

curd, help the proteolytic action of the rennet, and hinder the development of putrefactive bacteria.

The proof that the common lactic acid streptococci possess a marked power of splitting casein at temperatures coinciding with the usual storage temperatures for hard cheeses, made it highly probable that the activities of these bacteria in cheese ripening were not confined to producing the above mentioned indirect effects, but that they really played a direct part in the decomposition of the paracasein during the ripening of the cheese.

This view is also supported by the fact that in most hard cheeses it is just these lactic acid bacteria which are absolutely the dominating ones in the bacterial flora, at any rate during the first months of ripening, while the practical dairyman considers that he can influence to some degree the course of the ripening by adding to the cheese milk greater or smaller amounts of starter, i. e.. sour or coagulated skim milk or buttermilk. From our investigations it appeared that different strains of bacteria belonging to the group in question showed quite important differences in their ability to decompose casein. This circumstance made practical cheesemaking trials highly desirable. By determining the rate of ripening, when lactic acid bacteria of different casein-splitting powers are added to the cheese milk, it should be possible to elucidate to some degree not only the general significance of these lactic acid bacteria in cheese ripening but also in what manner the proved difference in casein-splitting power of such strains comes to bear on the rate of ripening of the cheese.

This reasoning led us to carry out a number of preparatory cheesemaking trials, which will be briefly described here.

The milk for these trials was holder-Pasteurized at 63° C. for 30 minutes, then cooled on an ordinary milk cooler and transferred to the cheese vat. The milk was then brought to a convenient temperature for rennet coagulation (30° C.) and treated with lactic acid bacteria of known casein-splitting powers, after which it was at once coagulated with rennet. The lactic acid bacteria to be used in the trials were cultivated in sterilized skim milk. The same amount of culture, corresponding to 1.5 per cent of the amount of cheese milk, was used in all the trials, and three different strains of lactic acid bacteria, differing in their casein-splitting power, came into use in these cheesemakings. In all the trials whole-milk cheeses were made with gouda texture and salted in brine. After the cheeses had been kept for about three months the amounts of soluble nitrogen and amino nitrogen were determined (the latter by Van Slyke's method) and the nitrogen found was expressed in percentages of the total nitrogen of the cheese. As these trial cheese makings were of quite a preparatory nature, we considered that their number should be limited, and we therefore only carried out three cheese makings, one with each of the bacterial strains. These strains were characterized thus: No. 1 did not possess any appreciable power to decompose casein in milk treated with chalk, while No. 3 had a marked ability to do so. No. 2 was intermediate in this respect.

The results of these trials can be seen in Table 1.

It will be seen in the table that the time required for making the cheese, i. e., the sum of the times for curdling, heating, and stirring afterwards, is obviously dependent on the bacterial strain which was added to the cheese milk; thus, No. 1 required the longest and No. 3 the shortest time; i. e., 152 and 133 minutes, respectively. That these differences could not be due to the fat content of the cheese milk is shown by the next column, in which the fat contents of the fresh cheeses are given, expressed in percentages of the total solids. The water contents of the fresh cheeses, expressed in percentages on the cheese less fat, are also given in the table. These water figures show that when strain No. 1 was added to the cheese milk cheese with 61.6 per cent of water, calculated on the fat-free curd, was obtained in spite of the cheese being treated for 152 minutes in the vat; on the other hand, when No. 3 was used the water content of the cheese reached only 58.4 per cent, although the working time occupied only 133 minutes. Strain No. 2 takes an intermediate position in these respects.

The difference in working times given is, of course, the result of our endeavors to obtain cheeses with the same water content. When the cheese gave off whey slowly the working time was lengthened. In spite of this, the water content of the cheese treated in the vat for 155 minutes became the highest and that of the cheese treated for 133 minutes the lowest. The differences in the working times were thus not sufficient to counteract the differences in the tendencies of the cheeses to separate off their whey.

As this tendency to separate off the whey is very largely dependent on the rapidity with which lactic acid is formed inside the curd particles, it is easy to explain the differences in question as being due to the differences in character of the lactic acid bacteria added to the cheese milk.

In the last two columns of the table are given the amounts of soluble nitrogen and amino nitrogen, formed after keeping the cheese for about three months.

For all three cheese-making groups the percentages are practically the same. Thus, the investigations gave quite a negative result in this respect, no differences in effect due to the different caseinsplitting powers of the lactic acid bacteria added being noticeable.

We did not, however, consider that we should give up making further experiments on the effect of different strains of lactic acid bacteria on the rapidity of cheese ripening. The technique employed

in making the cheeses in the above trials was in several respects faulty, and, moreover, we were entirely without knowledge as to the real bacterial contents of the milk at the moment of curdling.

The technique of the cheese making should be perfected so that the water content should, if possible, be the same in all batches, and so that this water content should agree with that of similar cheese made under perfectly normal conditions. The bacterial content of the milk at the time of heating is influenced partly by the fresh infection which may take place after Pasteurizing and chiefly by the number of bacteria in the starter added to the cheese milk. The use of the same amounts of culture in all batches is no guaranty that the cheese milk will constantly contain the same number of bacteria, and only under the last-mentioned condition can the results be strictly comparable.

Since we had gained knowledge of the normal water content of cheese of the type which we aimed to produce in continued experiments (by exhaustive investigations, the description of which would carry us too far), we sought to form an opinion as to how far the bacterial content of the cheese milk at the moment of adding rennet could have any effect on the rapidity of the cheese ripening. The impression very common among practical dairymen that the addition of an increased amount of starter to the cheese milk shortens the time necessary for ripening the resulting cheese speaks in favor of the view that there is such an effect. That the bacterial content of the milk at the time of addition of rennet has the effect in question, in spite of the long time usually required for ripening the cheese, can be explained by the fact that the bacteria in milk live in a liquid while in the cheese they live in a solid medium. If the milk is rich in bacteria when curdling occurs, then the curd particles should contain an appreciably larger number of colonies of bacteria than when the milk is poor in bacteria when rennet is added. The solid nutrient medium to which the curd particles may be regarded as corresponding, thus becomes much more thickly strewn with bacterial colonies in the former case than in the latter. L. D. Bushnell's investigations (3) show that 95 to 98 per cent of the bacteria which the milk contains at the time of curdling are found again in the curd formed.

From a purely practical point of view, the conditions in question find expression in the different rapidities with which the acidity of the whey increases while the cheese is being treated in the vat. If an appreciable amount of the whey is removed before the heating and the subsequent stirring take place, the acidity of the whey increases particularly fast because the amount of curd is large in proportion to that of the whey left behind. If, on the other hand, the whole amount of the whey is kept in the vat while the cheese is being worked, then the acidity of the whey increases much more slowly. It is within the curd particles that most of the lactic acid formation takes place, a sign that these particles are much richer in bacteria than the surrounding whey. From G. J. Hucker's investigations (4) it even follows that the bacterial content of the milk when rennet is added really exercises a certain influence on the nature of the resulting cheese. Hucker worked with milk which showed vary

ing bacterial contents on arrival in the dairy. He had no further knowledge as to the nature of the kind of bacteria occurring in the milk, and this is surely the reason why he obtained such widely differing results. From these experiments, however, it appeared that the milk which contained the largest number of bacteria yielded a uniform cheese of good quality, while on the other hand, a low bacterial content in the milk had the effect of producing cheeses varying considerably in quality.

Before taking up new investigations on the influence of different bacterial strains on the ripening of cheese it was therefore desirable to investigate how far the bacterial content of the milk at the time of curdling exerted any appreciable effect on the rapidity of the ripening of the cheese; for should this be the case, then in making the next trials it would be necessary to take into account not only the properties of the bacterial strains added to the cheese milk but also the total bacterial content of the milk. In order to throw some light on this question a number of preparatory investigations were made in which milk containing varying numbers of lactic acid bacteria was made into cheese under conditions as similar as possible. These trials were made as follows:

The milk used came directly by motor from a well-kept farm not far from Stockholm. For each trial 75 liters of milk were taken and separated cold, after which the fat percentage was adjusted to 1.5. Of this milk with 1.5 per cent of fat 70 liters were transferred to the cheese vat, Pasteurized there at 63° C. for 30 minutes, and cooled to the coagulating temperature. Then sufficient hydrochloric acid was added to bring the acidity of the milk back to the same degree as before Pasteurization (the Pasteurization generally caused a lowering of the acidity by about 0.5 degree Soxhlet-Henkel). Finally 1 per cent of starter, or pure culture, was added to the milk, which was then left to ripen at 28° to 30° C.

As an approximate measure of the increase in the bacterial content during ripening the rise in acidity was used. If the rennet was added immediately after the admixture of the starter the rise in acidity was indicated as 0.0 S.-H. An increase of the acidity by 0.5 or 1.0 thus means that the acidity had increased by 0.5 or 1.0 S.-H. above that shown by the milk when the 1 per cent of starter had been added.

The curdling temperature was 30° C. and the heating temperature 34° to 36° C.; the same water percentage was always aimed at; i. e., 59 to 60 per cent calculated as percentage on the fat-free cheese. Rennet powder was used for curdling.

The bacterial counts were made microscopically on the addition of the starter and after the milk had ripened to the desired acidity.

The cheeses were stored under normal conditions for two months, after which samples for analysis were taken each month, while the cheeses were examined for texture and taste at the same time. The chemical analysis included the determination of total nitrogen, soluble nitrogen, amino nitrogen, fat content in solids, and water percentage calculated on the fat-free cheese. Table 2 gives a summary of the results obtained in these trials. Here the total nitrogen and the amino nitrogen are not included.