HIGH OR LOW TEST MILK?

Hawera Star, Volume XLVII, 18 June 1928, Page 7

HIGH OR LOW TEST MILK?

Search for Best Cheese Cow

HAWERA LABORATORY’S REPORT

Of Profound Importance to Dairying

future course towards greater PROSPERITY INDICATED

Opinion regarding the relative values of low and high testing milk has up till the present been divided. On one side it has been laid down that the best butter-fat producing cow is the best cow for all dairying purposes; on the other hand it has been con tended with equal assurance that the low test cow produces more of the solids necessary to the manufacture of cheese, wmcn o these two contentions is correct?

It was with the clearly defined object of settling this question upon a scientific and practical basis that the Hawera Dairy Laboratory undertook nearly two years ago an experiment on a large scale. That experiment has been known locally as the “Tokaora Test” and its result has been awaited with the deepest interest by dairymen and breeders of stock. Mr P. O. Veale, researc chemist and bacteriologist, in charge of the laboratory has now completed the test and compiled his report and his findings will be published through the columns of ‘ ‘ The Hawera Star. The first instalment of this voluminous report is given in this issue and succeeding instalments will be published daily until the whole has been made public.

The report covers such a wide field of inquiry and furnishes such a large number of answers to hotly-debated questions, that it is difficult to summarise the findings in a few words, but if one statement can be selected as of paramount importance it is perhaps Mr. Veale’s arresting remark that many dairy farmers m Taranaki are losing large sums of money by their present adherence to high-testing cows for cheese-making purposes. . The features of the experiment which have led to this conclusion aje, of course, dealt with fully in the report, but it may be mentioned that striking evidence is adduced in support of the contention.

It has been found, for instance, that the cheese produced from the milk of the low testing herd returned more than 2d per pound butter-fat than the product of the high-testing herd. On the other hand the old contention that Jersey cheese would sell at a premium on the Home market, due to its higher butter-fat content, has been borne out, it having averaged 2d a cwt over the whole season in advance of the price secured for the low-testing product. However, this premium on the total sales, actually made little impression upon the increased production per lb of butter-fat of the low-testing milk, as is evidenced by the higher average of 2d per lb just mentioned.

AIM AND SCOPE OF RESEARCH. Tho aim of the experiment has •been to determine as a result of the year’s eheeseniaking with the •miflks of the Jersey, Ayrshire and Friesian breeds. What class of milk and what type of cheese is best suited to the export trade. To this end, the milks/ of the three breeds were made into cheese under ordinary factory conditions, in three experimental vats specially made for the purpose, and situated in am ordinary cheese factory. Full analyses by standard official methods were being made daily in the laboratory of all constituents of the milk, whey and cheese, and these tests were continued daily throughout the whole season. EatSh batch of cheese was weigher! separately, graded in New Zealand, regraded, weighed and valued in London, amd full reports oil its quality and •suitability for the trade were supplied by experts appointed in London.

curing and export, the experimented cheese, identical with the usual factory methods as practised in New Zealand.

From tlie text of the report given below are omitted the graphs and tabulated figures provided in the original report. /Mr. Veale's explanatory comment in illustration and proof 'of his statements are, however, so lucid that the reader will find no difficulty in following his line of reasoning. In the ■original text of the report facts supporting the /contentions are given with meticulous detail. In our slightly condensed version we have endeavoured to follow the main line of the statement, omitting the supporting data.

THE TEXT OF THE REPORT. Introduction And Statement of Problem. Milk is a substance which is well known to vary in chemical composition, and the existence of “rich” and “poor” milks is universally recognised. Of the various constituents of milk, butterfat has been the most widely accepted as a criterion of- richness; and varieties of milk are classed as ‘‘high testing” or “low testing” according to the percentages of butter-fat contained in them. Similarly, various breeds of dairy cattle have come to be designated “high testing” or “low testing” according as typical milk yielded by them contains approximately 5 per cent, of butterfat or falls to a content of about 3 per cent, of this substance. » Cheddar cheese made from milks of different compositions show variations in chemical analysis, and these variations are correlated with modifications in the process of manufacture, and with different characteristics in the resulting cheese. More important still from the economic point of view, high and low testing milks contain different percentages and different relative proportions of butterfat and casein, this resulting in different yields of cheese of diverse chemical analysis from equal bulks of milk, and again in different quantities of cheese from equal weights of butterfat derived from these milks. Hence has arisen the controversy on the relative values of high and low testing miIKS for eheesemaking. Those acquainted with the cheesemaking industry are aware that low testing milk such as that given by the Holstein-1 riesian or Ayrshire breeds yields more cheese per pound of butterfat than high testing milk such as that from the Jersey or Guernsey breeds. On the other hand the cheese from the Jersey milk is richer in butterfat than that from the Friesian milk and might be expected to sell at a premium over the product of the lower testing milk. Jersey, Ayrshire and Friesian Herds.

Among some of the items of fresh knowledge which have been obtained, may 'bo mentioned: Whether high or low testing milk is best suited to the manufacture of cheese UNDER NEW ZEALAND CONDITIONS (which are different from conditions in other countries); whether (to quote various partisan statements) cheese made from high testing milk is rich and mellow or whether it is greasy and uninviting; whether cheese made 'from low testing .milk is well balanced and rich in nutritious proteins, or whether it is hard, brittle and indigestible. It was l also sought to establish whether the homo trade has any special preferences in regard to cheese, and whether the ‘'"Trade” will really pay a 'higher price for checs'e having three or four per cent, more .butter-fat than that of the Dominion’s competitors. Further, it has been the aim of the experiment to establish whether New Zealand is wasting money, or on the other 'hand, creating a reputation for quality, by using so much high testing milk for cheesnuaking; if any variety of milk gives better manufacturing returns than another or is more conducive to unavoidable losses; if any variety of cheese is more prone to shrinkage' or to develop any merit or fault —all these facts were sought and the results will have the profoumdest bearing upon the future of New Zealand’s cheese industry.

Basis of Pay-out. Arising out of the great mass of data compiled during the experiment, some consideration has been given to the basis of payment for milk for cheesemaking, but 'to state that this has been the beginning and end of the investigation is to misrepresent the position entirely. At the same time the experiment has certainty created, a fund ol knowledge, specially relating to all phases of the manufacture of cheese under New Zealand. conditions without which no consideration of the basis of payment would be possible and Mr. Veale has availed himself of the knowledge so

The milk used in this experiment was supplied by the three herds as follows: High Testing.—Jersey herd belonging to 'Mt. H. E. liatten. how Testing.—Ayrshire herd belonging to Mr. F. Mills low Testing.— Friesian herd belonging to Mr. H. F. Johnson.

gained to evolve a basis for arriving at payment which has not hitherto been suggested and one 'which 'will provide a means of arriving at. an average when assessing the amount of a dairy factory’s payout that will be as nearly fair to suppliers of all classes of rnillc as possible. Ti. is a well-known fact, that, up To the present, figures relating to analyses of milk and its products, and all statistics relating to the manufacture of dairy produce arc quoted al'inost exclusively from American authoriti'Cs. New Zealand' factory managers have developed numerous practices peculiarly their own and this fact, coupled with the recognised differcuces' in the climates, and methods of feeding and treatment, of cattle have led Mr. Veale to believe that American results will not apply exactly here. Hence he has made all conditions of production, manufacture,

Each of the 'aimers mentioned were actual suppliers to the Tokaora branch of the Hawera Co-operative Dairy Factors' Company, Ltd., and their cows were fed and treated in accordance with iho usual practices of New Zealand dairy farmers. The three farms are situated within a couple of miles of each other, and are practically nextdror neighbours in the same locality. Thus throughout the whole season of the experiment the herds were exposed to similar variations of weather and temperature, and were likely to _be affected similarly by all meteorological factors which were capable of affecting the cows or their food supply. The report here gives a few details 1 o f the general conditions of each farm i and herd as set out in a communicn- ■ tion from Mr. A. J. Glasson. Fields il Instructor of the New Zealand Depart - , ment of Agriculture.

Mr. H. E. Johnson’s Friesian herd consisted of nearly 70 Holstein-Friesi-an>s, including a large percentage of purebreds. This herd, like Mr. Batten’s. had not wintered too well and came in to calve in only fair condition, but when seen on October 8, 1926, the pastures were showing a wealth of growth and the condition and general appearance of the herd had greatly improved.

Mr. Mills’ Ayrshire herd, numbering 43 to <3O, consisted entirely of purebred Ayrshires. Ensilage was fed out in the autumn and until the end of July, and a<ter that hay and mangolds till tho end of September. The cows had wintered well, and were showing nice condition.

Mr. Batten’s Jersey herd consisted of about 55 cows, mostly a good class Jersey, and were wintered on grass on Mr. Batten’s Mountain Road property; the cows were brought in to the home farm about a fortnight before cabtng, put on to fresh pasture, and given a daily ration of silage and green oats. The herd showed the effects of the absence of extra feeding during the winter months, and, generally speaking, came into profit in poor condition. On October 8, however, Mr. Glusson found a remarkable improvement in their appearance. Since no supervision was attempted of the treatment and feeding of the ton’s, the present investigation i.s not concerned with tarm economics, and has no bearing upon the yield of milk < r of ciieese per cow or per acre. Nevertheless, it will be found capable of indicating clearly the type of cow which, from the factory point of view, makes the best return for cheesemaking, leaving to the individual farmer to decide from his own individual experience whether differences in feeding costs, if any, will be sufficient to reduce appreciably', or to nullify altogether, the financial advantage gained by the use of such cows for supplying milk for cheesemaking.

Scope of Investigation. The actual scope of tho invostiga lion is as follows:

(A) To collect statistical matter and experimental figures in regard to the johowing portions of the nmuuiaotuie or cheese under New Zealand factory conditions:

1. Bai/y and seasonal fluctuations in the chemical analysis oi the milksof various breeds. 2. Actual yields of cheese per pounu of butterfat obtained from milks oi various .at. tests under New Zea

land conditions of eheeseniaking

3. rat and casein losses occurring in the whey from milks of varying richness. 4. Chemical analyses of diherent types of cheese, and the relationship between these and the opinions of graders and buyers as to qua.uy and texture, both in New Zealand prior to export, .and in London at time of sale. 5. The variations in quality and shrinkage between various types of cheese during curing in New Zealand and during the voyage Lorn New Zealand to London. (B) To secure answers, if possible, tro the following questions: 1. W hat is the chemical analysis of the type ox cheese which best suite the export trade for which New Zealand caters, and from what type of milk is it derived f 2. Can cheese of varying fat or moisture percentage be readily distinguished by grader or buyer? 3. Is the present New Zealand .standard of 50 per cent, of fat in the ary matter of cheese entirely satisfactory to our buyers, ancl entirely in our own interests? 4. Does cheese having substantially more than 50 per cent, of lat in the dry matter carry any premium over less fatty varieties? 5. Is this premium, if any, sufficient to compensate lor the lower yield of cheese per pound of butter-fat derived from high testing milks ? 6. If net, can any \ ariation from the straight butter-fat basis be devised for the payment for milk for cheesemaking, in order to compensate the lower testing supplier for the increased yield of cheese per pound of butterfat derived from hiis milk ? . 7. .Should the New Zealand cheesemaking. industry continue in its present tendency to use, year by year, continually richer and richer milk for cheesemaking

Outline of Experimental Work,

(A) IN THE FACTORY. For the actual cheesemaking, three special experimental vats, each of 2o() gallons capacity, were set .up in the Tokaora branch of the Hawera Dairy Company. The vats were identical in widtn and depth and in general construction with the ordinary 830 galled . ats, except that the smaller capacity was secured by a reduction in length. Each vat was fitted with its own steam et and rotary agitator.

The milks from the three breeds were weighed daily and run by chutes each into its own special vat. immediately the flow had ceased the milk was agitated lor a few moments and then sampled for analysis. ‘ Cheesemaking then proceeded in the three \ats in the usual way. Attempts were made to handle the vats daily in a. manner as nearly identical as possible, but in each case full details oi the slay’s work were kept in respect of each vat throughout the season. The whey running irom each vat was sampled for analysis at about the middle of the period of the “run-oft.” At the comlusn.n of each day’s work the cheese was hooped separately. Proper allowance was made from day to day for the exact weight ox any curd left over insufficient in quantity to make a “medium” cheese, and thus the weight of cheese made daily irom each quantity of milk was known with certainty. The cheese were numbered serially, and the different breeds were distinguished by letters which were continually ..hanged, generally at intervals ol a week. Throughout the whole investigation, the cheese were identified by their letters and numbers, arid tlioir real identity . was known to no one except the scientist in charge of the research. The batches of cheese remained overnight in the presses, and were removed the lollowing morning. After standing for 24 hours in the factory, they were transported to the company’s centra! curing room and weighed and sampled for analysis. Here they remained fur ; 4 days, being turned daily in accordance with the usual procedure. . After 14 days on the shelves the cheese were again weighed and crated lor export. Records were taken of the identity of the cheese in each, ciuto, and the crates were branded with lotto; s corresponding to the batch Jetteis of the cheese which they contained. The various batches of cheese, each corresponding approximately to one YYCcIv ’S make from Tcho milk of one paitirular breed, were graded by the officers of the New Zealand Dairy Division. and grading certificates furnished as in the usual way for export cheese throughout the season. Upon arrival in London the cheese were again weighed and graded .b\ officers of the New Zealand Dairy Division, and returns forwarded. The

cheese were also examined by a committee of English buyers and salesmen, who forwarded special reports upon each batch of cheese with regard to quality, flavour, texture, and general suitability or otherwise for the English trade. Each crate of cheese was given its own market value in shillings per hundredweight and a basis of comparison mrnislied for finest and first grade New Zealand and Canadian white and coloured cheese. After March 10, 1927, the work originally undertaken by the London committee was carried on alone by Mr. A. YV . Heyes, of Messrs H. A. Lane and Co., London, and the special thanks of the federation are due to this gentleman tor his painstaking care. (B) LABORATORY WORK.

Throughout the season the following programme of analytical work was followed in the laboratory :

1. Duplicate determinations of fat, casein, and total solids in the three milks used for eheeseniaking. 2. Duplicate determinations of fat, casein, and total solids in the whey running from each vat. 3. Duplicate determinations of fat, moisture, and protein matter in the green cheese (.two days old) made from each class of milk.

General Survey of Statistics. The investigation commenced on September 1, 1926, and finished on' April i.O, 1927. Be.ore and after these dates the quantity of milk furnished by at least one of the herds was deemed insufficient to handle in the va,ts for the purpose of making good quality cheese. During the season 472,394 lbs of milk were supplied for the special cheesemaking by the Friesian herd; 256,085 los by the Ayrshire, and 290,795 lbs oy the Jersey, making in all a grand total of 1,019,274 lbs of milk, handled during the investigation. It might be mentioned that, during the portion ox the season covered by the investigation, the Friesian herd actually produced 491,427 lbs of milk, but it was found impossible to use the whole of this, tor the reason that during most of October, November, and December the daily production exceeded the capacity (2500 lbs) of the vat provided. When this occurred, the milk was received in three equal portions in the factory weighing machine, and .at the same time a sample was taken for 'factory testing, an aliquot part ■ was removed by baling. Tlie weight of the remainder was noted, and the milk was then run into the cheese vat. This procedure was followed for each weighing of milk, anti thus the hulk was reduced to within the capacity of the vat without altering in any way the average composition of the whole hulk of milk secreted by the entire herd. The milk used in the investigation was manufactured intd a total of 1619 cheese, consisting partly of “Export” size (8911) approximately) and partly o. “Medium ’ (451 b) size. In all .09,6531 lbs of cheese (approximately 49 tons) were made during the season and exported in a total of 7.14 crates. The investigation covered practically the whole of the usual lactation period of New Zealand dairy cattle. It was conducted in all respects in accordance with the ordinary factory practices used in the manufacture and export of .New Zealand eheddar cheese, and, as will lie seen from the above statistics, upon a scale suliiciently large to furnish conclusions of value.

Variations in Analysis of Milks Throughout The Season.

A) FAT. The percentage of fat Jn the milk used for cheesemaking varied between the extreme limits o. 3.08 per cent, on November 7 to 6.63 per cent, on April 27, these figures being the Lwe.st and highest daily fat tests of the riesian and Jersey breeds respectively lor the whole season. The Friesian herd commenced' the season with an average test during the first 10-day period of 3.45; this fell to a minimum ox 3.25, and finally rose at the end of tlie season to a maximum of 4.18. Similar figures for the Ayrshire herd were 3.75 and 3.55, to a maximum ol 4.47; and for the Jersey, 4.53 and 4.33, to a maximum of 5.78.

Daily tests of the butterfat contents of the three milks showed much wider i nidations, the minimum .fat tests of the three being respectively 3.08 on Novemcer 7, 3.33 on November 8, and 0.90 on October 19, while the respective maxima were 4.53 on April 28, 4.75 on April 29, and 6.63 on April 27. There was thus an ample range of milks of v arying butterfat content used for cheesemaking. The butter-fat test of a herd is not, as is popularly supposed, a constant quantity from day to day, but, on the contrary, is a most variable quantity, dependent upon numerous factors, weather • and, climatic conditions obviously have a contributing effect, as the graphs which liaco been computed and drawn for this report furnish evidence of well-marked rises and Tails occurring simultaneously in the fat tests of the three herds, all of which were close neighbours. It is not within the pro- , ince of the present discussion to do more than point out the existence of these daily variations in the analysis of milk, but it is hoped to throw some light upon them in a subsequent communication.

Positive and negative variations were approximately regular and equal, the result being ‘that a graph of the average tests over a number of days such as the customary 10-day factory period showed only the slow and regular variation of tiie normal lactation curve, without the irregular outline characteristic of the daily record.

An important observation will he made when the fat records arc considered in relation to the legal limit of 3.25 per cent, .lor fat in, milk. It will be observed that four times in September, eight in October, six in November, and one in December, the Friesian milk fell below the legal limit of 3.25 per cent, of fat as required ior human consumption. This proves that, when analysis reveals, according to the health regulations, a deficiency in fat, some knowledge of the breed of the tows supplying the milk is necessary before an opinion of any value can be gicen as to whether adulteration with water is likely, and furthermore proves that when a" sample of milk from a large herd does fall below the legal limit for fat, addition of water is not thereby established in every case. For cheesemaking, of course, no legal limit for fat exists, because it has never been proved, nor indeed can it ever be proved, that normal milk of low fat tost, is at all harmful for this purpose.

Bearing of The Daily Variations in Fat Test Upon The Question of Cow Testing or Herd Testing.

The graphical record of the clailv tat tests rf the three breeds of milk shows that there is no such tiling as “the test” of a cow or of a herd, i.e., there is no scientific foundation for the popular belief that the buttervat test of a cow or of a herd is a fixed and immutable quantity, or even that it shows only minor variations from day to ’day. Variations in butterfat tests from dav to day, even in a herd of over 60 ftows, are slihwn to be quite consider-

able, and numerous instances will lie lountl in the present investigation or variations as great as 0.5 per cent, between one day and the next. This, ot (ourse, does not represent the maximum, as one example is recorded for the Jersey breed near the end ol April when the fat test rose from 5.4 per cent, to slightly over 6.6 per cent, and then fell to its original value on three successive days.

The question therefore arises as to whether cow-testing or herd-testing returns, based upon the fat test of a single day or. at the most, two days, are ready reliable. The hrst answer to this must undoubtedly be that returns based upon a two-days’ test are more likely to approximate to the production ot a period than those based upon a single day’s test only. Sharp rises and falls nearly always occurred in immediate succession, and thus the positive variation of one day would normally be followed by the negative counterbalancing variation of the next. When the herd is tested on one day only, there is the possibility, or rather the certainty, that it will be observed on a day upon which it will perform either slightly above or slightly below its true average butterfat test. Thus the results of a single period should not be expected to correspond closely with factory returns based upon true composite samples, although such a fortuitous corresponding of results is frequently obtained. However, over a whole season, with regular sampling at intervals, the law of probability will ensure that the number of times when positive variations are obtained will be aoout equal to that upon which corresponding negative variations occur. Thus, over a whole season, there is no reason why the return of butterfat per cow, based upon either one or two days’ observation, should not coincide, to within a very few pounds of butter.at, with the actual production, notwithstanding the existence of daily . ariations in test. (R) PERCENTAGE OF CASEIN.

This varied throughout the season between the utmost limits oi 2.23 per .eat. on September 19 and 3.40 per cent, on April 25, and showed a most interesting abnormality which has been correlated to some extent with certain marked weather conditions. In tho case of each breed, tlie casein content of the milk showed on the whole a progressive rise till towards the end of November. At this stage the Friesian and Ayrshire milks showed a value in the vicinity of 2.65 per cent., while the Jersey was sligtly above 2.90 per cent. Thereafter, save for a slight temporary recovery early in December, the casein contents of the milks steadily declined until moie than linkway through February, at which stage the low-testing milks gave under 2.5,) per cent, of casein, and the high testing about 2.8 j per cent. There is reason to believe tiiut this behaviour on the part or casein content of milks was connected with a long dry ol weather lasting from November to February, and was broken only by transient and insufficient showers. After the drought had definitely broken and lieav v rain had ialien, the casein content'of the milk rapidly recovered and reached, towards the end of the season, maxima of about 2.80 per cent, by the low testing milk, and nearly 3.4- per cent, by the high testing. A graph of the season’s results of the casein analysis of the three milks shows the same irregular daily variations as in the case or the fat percentages, and conforms in fact to the general 'rule for nil constituents ot milk The casein content of the milk is apparently more dependent upon weather conditions than the lat, as the examples of coincidence in sharp rises and falls are more numerous and more marked. These will be found in a number of cases to correspond with similar variations in the fat content, as for example, on September 4,5, 6, i, and again on October 28, 29, 30, 31. On s.me occasions a rise in the vat is accompanied by a rise in the casein, while on others the reverse is the case, thus producing variation's in the ratio of casein to fat quite as sharp as any ariations in the constituents themselves. These will be discussed later. ’(( i PERCENTAGE OF TOTAL SOLIDS. . .

These figures show a larger variation than any of the constituents taken separately, the .values fluctuating between a minimum of 11.8;) per cent, and a maximum of 16.05 per cent. Daily variations were very pronounced, many o: these corresponding with the rises' and falls already noted in the .ases of the fat and casein percentages. The difference between the Jersey milk and that of the other two breeds was most marked, and, from about January onwards, generally amounted to about a full two per cent., with a tendency to a slight increase as the season advanced. On the other hand, the Friesian and Ayrshire milks maintained an approximately constant difference throughout the entire lactation period. D) PERCENTAGE OF SOLIDS NOT FAT. These constituents are of especial interest by reason of the fact that theii value is' used as a means of arriving at the genuineness of the sample of milk. The legal limit for these constituents is B.A) per cent., and as it is an offence under the Sale of hood and Brims Act to sell milk with solids-not-,at “falling below this figure, it is generally assumed that milk falling below this standard is adulterated by the addition of water.

The close approximation of the values of the solids-not-fat through the entire reason confirms the tact already established that the solids-not-fat are less affected than any other constituent of milk by breed, climatic conditions, or period of lactation, and emphasising the soundness of the legal aspect of accepting the solids-not-fat a.s a criterion of the genuineness or itherwise of a sample of milk. There is no tendency for the graphs of the three breeds to lie in separate nelliefined regions, and only towards the end of the milking season did the fersey milk diverge at all noticeably from the two lower testing breeds. On ne or two occasions during the season * lie Jersey milk, which is normally a limb-testing milk, showed lower solids-not-fat than either of the other two, while on other occasions the Friesian, normally the lowest testing milk, showed the highest solids-not-fat.

The Legal Limit. An important fact comes to light when the values of' the solids-not-fat are considered in relation to the legal limit of 8.50 per cent., namely that each of the samples of milk fell below the legal limit on one or more occasions. " The Jersey milk touched the limit once and fell below the limit once, early in September. The Friesian and Ayrshire, being milks of lower solid content, fell below the legal limit more often the actual occurrences being as follow : Friesians : September 4 times, October once, December once, Februarv *yve times, March, four times; total, 15 times. Ayrshires : September eight times, October once, December once. January three times, February eight times, March four times, April five times; total, 30 times.

These figures prove conclusively that it is quite possible for the mixed milk of 50 or more cows to fall below the

legal limit of 8.50 per cent, for solids-, not-fat without deliberate additions of; water Laving been made. Under the I Sale of Food and Drugs Act it is an • offence to sell such milk for household: use, but nevertheless an. analysis revealing a deficiency in solids-not-fat does not necessarily prove adulteration by means of extraneous water.

It must be emphasised, of course, that the legal limits mentioned above apply only to milk intended for sale lor household purposes. No such limits exist in regulations concerning milk for cheesemaking, and the above departures from the standard • cannot therefore be construed as necessarily implying an unsuitability of the lowtesting milks for this purpose.

Yield of Cheese Per Pound Butter-Fat.

Since full daily records were made in respect to the milks and the cheese manufactured from them, each day’s cheesemaking gave a record of a yield of cheese per pound of butterfat from the milk of each breed. These daily results, being dependent upon the daily variations in the composition of the 'milk, as well as upon the moisture content of the cheese, and upon the losses occurring in manufacture, naturally showed a wide variation themselves from day to day. Few- deductions of value regarding the yield of cheese from milk of a given fat test can be made from a consideration of any one individual daily result, and accordingly tlie average yields per pound of butterfat have been calculated for each “period” of 10 or 11 clays, in accordance with the usual factory procedure.

Discussion of Different Yields. From tables which have computed from analysis of the milk yields of tiie herds it is at once obvious that the low testing breeds have secured a bigger yield of cheese per pound of butter-fat than the high testing, the Friesian breed as the lowest tester, being outstanding in this particular. Actually the lowest yield of the Friesian breed (2.652) for any 10 day period, was higher than the highest of the Jersey (2.621) for any similar period diking the whole season. These remarkable differences in yield of cheese pei- pound of butter-fat are caused bv difference's in the proportion cf casein to fat in the .milks of the various breeds'. The act of. coagulating milk with rennet followed by beating and agitation is practically that of separating the fat and casein from the remainder of the constituents ol* milk. . If this could be done separately, we should find that each pound of butter-fat from normal low testing miUr such as Friesian or Ayrshire was associated with a greater weight of casein than in the case of another pound of hutt.er-fat derived from high testing milk such as Jersey. Since casein is actively instrumental in retaining moisture without appreciable softening of body, the cheese made from low testing milk contains a higher proportion of protein matter, and an increased moisture capacity. Consequently each pound of butter-fat derived from low testing milk makes up into more cheese; and, the lower the test, the higher the yield per pound of hutter-fat, provided that the milk is normal.

It must- be observed, that different milks of the same fat test are not necessarily associated with the same amount of casein, and that therefore the same yields per pound of butter-fat cannot be expected from them, even with identical methods of'cheesemaking. Numbers of examples will tbc found of different milks' with practically identical fat tests, but with yields appreciably different, when the fat rises, the yield does hot necessarily fall, and viceversa. It is therefore obvious that the common practice of comparing seasonal statistics of different factories and of imputing inferior workmanship to a manager who ,with the same average fat test as a neighbour,' cannot produce as high a yield per pound of butter.ot. may be extremely unfair, and, in any case, is scientifically unsound. Differences in the breeds of the covvs sue owing the milk; variations in methods/of feeding, and in nature and quantities of feed; climatic conditions and numerous other factors' all tend to have their effect upon the ratio of casein to fat in milk, and to produce variations in the yield of cheese per pound of butter-fat from that mulk. Much additional research will be required to rib'-ever the exact why localities and factories noted either for very high or very low yields of cheese per pound of butter-fat exist, but the present investigation throws some light upon the pr.djbni by establishing that different milks of the same fat test do mot necesurily yield the same amount of cheese, an 1 that climatic conditions • and the period of the season have a profound bearing upon the amount of cheese obtainable from milk. Discussion of Seasonal Fluctuations.

By consulting the graphical record of tlie season’s yields of cheese per pound of butter-fat, one may observe the following facts: — 3. The Friesian and Ayrshire breeds are noticeably ahead of the Jersey. 3. There is a distinct tendency in all three graphs, to rise .to a maximum in the middle of November, 1926. 3. From November 1926, till February, 1927, there is a marked fall in the yield of cheese peT pound of butter-fat, and the season’s minimum is reached on or 'about the last period in January or the first period in February. In the case of the Jersey breed only, the yield for the - last period in April is very slightly loss than that reached during the first period of February. 4. After February, the yields all increased and showed an upward tendency for about a month before -evcuu-tnlly declining again as the lactation periods of the cattle drew to a close.

This remarkable fall in the yield of cheese from November to February followed by a sharp rise, is quite abnormal. ' It is explained, however, by the table, which shows am exactly.simiilar seasonal fluctuation in the ratio of casein to fat occurring in the three milks. Tire -season 1926-27 was- remarkable for a prolonged dry spell of weather, lasting approximately from November to February, with only a few slight showers occurring; and during that time, the yields of many Taranaki cheese factories fell away to abnormally low figures. There its season to believe that this long dry ispell was tire causative agent in lowering the proportion of casein relative to fat in the milk-s. Thus the ratio of casein. to fat. is shown to be am important controlling factor in the determination of yield of cheese per pound of,butter-fat . . . milks such a;s the Friesian or Ayrshire, containing a relatively high proportion of casein "to each part of fat will pro- [ duee cheese having a high moisturecarrying capacity, anl will therefore yield more cheese per pound of butte-r----f fat than ' Jersey milk. The increased

yield is, of course, partly made up by the actually greater weight of casein associated with each pound of 'butterfat contained in low testing milk, but the increased moisture-carrying capacity of the casein is the main and controlling factor.'

New Zealand Grading of Cheese Exported.

In accordance with the- usual factory procedure,'the cheese was graded by the New Zealand Government graders’ at the, port of Patea, prior .to export. Every cheese had stood at least 14 days upon' the shelves prior to being despatched, but some members of each shipment had stood varying periods up to 21 days, as it was necessary to- accumulate' at least a week’s make of cheese from cadi variety of milk in older to secure a batch sufficiently large to justify shipment. (A summary of the official grading figures .for each of .the cheese throughout the season-is given.)

Out of a total of 714 crates, 36 crates or 5.04 per cent were graded “second.” This is perhaps a. little in advance of the usual percentage of “second grade” chees made by unpasteurising factories, but considering, the fact that the experimental cheese was no doubt examined extremely carefully, indicates that, in general, a good line was manufactured It must be mentioned that as all the experimental cheese was. made from raw milk it could not be expected to reach the highest saondards in New Zealand grading. Broadly summarising the results of live .season, we observe that- the Jersey milk made up into cheese which has more consistently pleased the taste of the New. Zealand graders than did either of the other two. For the largo majority of. the season, the Jersey choose maintained practically 92 points ot more, whereas the low testing varieties fluctuated mostly between 904-and 91 points. The Ayrshire cheese showed lhe greatest variation in quality throughout the season, and, on the average was a little inferior in grading points to the Friesian. These general differences in quality between tlie Jersey, Friesian and Ayrshire cheese however, are connected more with bacterial quality of. milk than with their relative richness in but-ter-fat. Throughout the season, reductase tests were made at intervals upon the milks, and, as a general, rule, the Jersey milk was of the highest quality, closely followed by the Friesian and then at a lower grade, by the Ayrshire. The fact the a “bucket” type of milking machine was used upon the Ayrshire far in, whereas “rcleaser” plants were used upon tlie other two farms serves, no doubt, to explain the difference in quality, as “bucket” plants offer much creator facilities for contamination of the milk, and, require much more work Ho maintain sterile.

At the same time, the. differences in fat test between the three types of cheese have been reflected to some extent in the points awarded for “body and texture” and for “colour.’ ’ There were many occasions when the two types of cheese scored uniformly 28 or 284 points out of 30 for .body and texture, but during the majority of the spring and . summer months-, -the high testing cheese tended to show.an advant age of at least half a point in body and texture, thus indicating that, in the opinion of the New Zealand grading officers, Friesian and Ayrshire cheese requires something, presumably more fat, in -order ait attain the ideal body and texture. In addition, a deduction of half a point for “colour” was made from both low testing cheeses for the large majority of the season. This is presumably an adverse expression of opinion on the part of the graders -•’gainst the dead witen'ess of the low testing cheese. . As against this, /however, it must be noted'that, for the last two months of -the’season, and at other isolated times during the earlier portion, the Jersey cheese also suffered, a deduction of half a point for colour.- It therefore seems evident that, if those’ colour deductions have any bearing at all upon, the question of the fat content of the cheese, they must mean that neither the dead whiteness of the low testing cheese nor the yellow creaminess -of the high testing is considered ideal in a so-called *‘ white” cheese.

The influence of the bacterial quality of the’milk upon the body and texture, of tho cheese is universally admitted, and therefore because slight differences in this respect have been noted 'between, the two extreme types of cheese,? it

must not be assumed .that these differences are necessarily due to the variations in fat content.'ln support of this statement, Tables 1-X,: XI and XII will furnish numerous instances of the two extreme types of cheese at the same period -of the year scoring identically the same points for body and texture. The very great difference in analysis between these types of cheese will be pointed out 'later and the facts of this investigation clearly show that there is no necessary difference perceivable, by physical examination alone, between the body and texture of normal Friesian cheese low in fat and high in moisture and that of Jersey cheese high in fat and low in moisture.

Grader’s Remarks on Jersey Cheese. line important difference can 'be obseried between the two extreme types of cheese during the months of February, March, and April, i.e., as the dairying season drew to a close. Whereas the low testing (Friesian) cheese maintained its average of ‘‘first’’ grading points, the high testing (Jersey) cheese fell away distinctly, and finished up the season “second” grade. An examination of the detailed points awarded shows that, although there was some slight falling away in flavour, the failure to secure “first grade” points can be ascribed to a progressive loss of points for “body and textux-e,” for “colour,” and finally for “finish.” At this stage in the season the- fat test of the Jersey milk was so high that the body of the cheese was always pen and greasy, and finally even the iind's became greasy with free fat. Towards the end of the season remarks oil the grade notes such as “greasy 'tody” and “make irregular” became common in connection with ..Jersey cheese, and the fall from “first grade” can he ascribed to. an objectionable greasiness caused by abnormally high I rat test of the Jersey milk. Similar remarks do not apply* to the Ayrshire cheese during the last three months of the season, as an examination of the ■ 'etailed grading points reveals that the failure to obtain “first grade” Was t 6 he ascribed mainly to lass of points for 'favour, coupled, no doubt, with an incidental effect upon 'body and texture. This was shown to be due to the bacteriological state of the milk, and, as already stated, is not a fault inherent to Ayrshire milk as a class, but merely due to the handicap imposed upon the bacterial quality of the Ayrshire milk by a “bucket” type of milking plant, coupled with the existence of cowshed facilities of not quite afi high a standard as those on the other two farms.

During the months of March aud April the fat test of the Jersey milk was frequently in the vicinity of 0 per cent, and actually reached 6.62 per cent.’on April 27, with an average per period of ten days, varying irom 5.2 to 5.8. During the same two months the Friesian and Ayrshire rarely went above 4.5, while the highest period test reached by the Friesian breed was only 4.19, during the last ten days of April. These figures indicate that 4 to 4.5 per cent, of butter-fat represents the limit beyond which it is unsafe to go if one would make a consistently good cheese of uniform body. Certainly, good cheese can be made, and often is made, from milk of higher fat test, but very frequently greasine'ss of body and rind results, and, at the same time, irregularities of make are bound to occur when the. normal daily variations of test raise the percentage of butterfat much above 5 per cent.

Analysis of Experimental Cheese

MOISTURE CONTENT.

The difference between the cheese made from the low testing and high testing milk was very marked in the matter of its moisture .content. The two low testing milks, Friesian and Ayrshire, were, with only a few exceptions, practically identical in their moisture contents throughout the season, and were substantially in advance of the Jersey. A graphical record of the moisture analyses of the three varieties of cheese has been computed. From this it is observed that the Friesian and Ayrshire cheeses commenced the season with an average moisture content fluctuating between 37 and 39 per cent., whereas that of the Jersey varies from 36 to 38 per cent. The difference between the two types of cheese tended to expand as the season advanced—for example, during the month of October the moisture content of the low testing cheese varied principally between 39 and 40 per cent., while that of the Jersey lay mostly in the region between 36 and 37 per cent. ■ Moisture contents of all types of cheese declined slowly and steadily from October, 1926, to the end of February, 1927, but for the majority of this period the difference of nearly 3 per cent, in the moisture contents of the low and high testing varieties tended to be maintained. By the end of February, however, the Jersey cheese was showing moisture contents varying between 34 and 35 per cent., while that of the Friesian and_ Ayrshire breeds lay between 36 and 37 per cent. The moisture figures all tended to increase towards the end of the season, hut the difference of approximately 2 per cent between the two extreme types of cheese was maintained for nearly two months. During the • latter portion of April, 1927, very large and irregular variations in the moisture contents of all three breeds of cheese occurred. These were probably caused by the fact that the supply of milk was now getting so low that accurate control of temperatures and of the cheesemaking process generally was becoming extremely difficult in the factory, and this is substantiated by the fact that the Ayrshire breed, which was the smallest in hulk at this time, showed the widest divergences.

Relationship Eetween Moisture Content And New Zealand Grading Figures.

Since the moisture content of cheese curd” is popularly supposed to have a profound hearing upon the body and texture, it will be interesting to compare the New Zealand grading returns lor batches of cheese showing divergent values for their moisture, for example, during the month of October, when the Friesian and Ayrshire cheese showed between 39 and 40 per cent, or niofcture, as against 36 to 37 per cent, in the Jersey. During October the lowtesting cheeses varied in grading between 90J and 9li points, whereas the high testing cheese graded 92 points uniformly throughout the month. The points awarded for body and texture to the three varieties of cheese are as follow: POINTS FOR BODY AND TEXTURE DURING OCTOBER, 1926. No. of Days Score was Awarded. Grading Friesian Ayrshire Jersey Points 39-40 39-40 36-37 per cent. per cent. per cent. Maximum 30 Moisture Moisture Moisture 254 12 28“ 7 31 12 271- 12 The two low testing Varieties of cheese were frequently reduced half a point for colour, being apparently considered too dead white, but the points for flavour were practically identical for the three varieties throughout. On those 12 d'ays when the Friesian 'scored only 274 points for body and texture, the comment w-as added that the make was “loose and open,” and the “body inclined weak.” Occasionally also, a little over-acid was developed and adversely commented upon by the 'graders.

The figures serve to prove that the body and texture of cheese is not quite so dependent upon the moisture content as is popularly supposed among cheesemakers, and .that, according to the fat test of the milk, the moisture content of the cheese may be high or low without appreciable variation in the body and texture.

In orther words, milk reasonably low in fat can be made up into cheese of high moisture content which may have just as good body and texture and be indistinguishable on these grounds alone, from another cheese containing much less moisture, and made from milk of higher fat content. The significance of this from the economic viewpoint is most important. Further reference to the question of moisture content will be made when considering the returns from London in conjunction with the composition of the cheese. IOHEESE. The actual percentage of fat present in cheese can be made to vary, almost. at the will of the clieescmaker, according to the percentage of moisture incorporated in the curd. Accordingly it will be more useful for us to consider, under this heading, the percentage of fat contained in the waterfree substance of cheese, as 'this figure is independent of the moisture 'cpnten. Generally speaking, the Jersey cheese ishowed throughout the whole season about 4 per cent, more fat in the water-free substance than either of the other two Varieties. At the beginning and' 'end of the season, the Friesian and Ayrshire breed's tended to be approximately equal in this respect, but from December to February, the Ayrshire showed a distinctly higher fat content in the dry matter than the Friesian, the difference varying from. 1 to 2 per cent.

FAT CONTENT AND N.2. GRADING RETURNS.

The fat convent of a cheese curd is considered .to exercise a profound influence upon the smoothness and “mellowness” of the texture of cheese, hut the returns in the present investigation indicate that, within the regions likely to he encountered in handling genuine whole milk for eheesemaking, this influence is not as strong as was formerly supposed. Of course, cheese made from heavily skimmed milk may exhibit harshness of texture, hut such a case has not fallen within the scope of the present investigation.

Innumerable instances can he found throughout the season when one or other or both of the low testing varieties ’of cheese, showing about 4 per ceiit. less fat in the dry substance than the Jersey cheese, received identical grading points for body and texture, while in other equally numerous instances, the difference in favour of the higher testing cheese was assessed at only half a point out of a possible 30. These results indicate that, as long as we are dealing with normal whole milk, identical body and texture can be secured with fairly widely divergent fat tests, 'and ’that, when slight differences are detected, they are not at all commensurate with the variations in composition which have given rise to them. 1

LEGAL LIMITS FOR FAT IN CHEESE. According to -tne regulations of the Now Zealand Department of Agriculture, all cheddar cheese exported form New Zealand must contain at least 50 per cent, of fat in the water-free substance, and must bo made from normal whole milk. If these l regulations have been wisely framed, we might be led to suppose that cheese which falls below the limit of 50 per cent, of fat in the water-free substance is an inferior article not fit for export, and that, as such, its inferiority might be readily detected upon the usual examination by exports.

Reference to the graphical record of the percentage of fat in the water-free substance of the experimental cheese will show that, in a very considerable proportion of the months from .September, 1926, to December, 1926, and in parts of January and February, 1927, the low testing (Friesian and Ayrshire) cheese fell below the New Zealand legal limit of 50 per cent, of fat in the dry matter. ’This also occurred once in the case of the Jersey cheese viz. on December 1, 1926. The lowest value ever reached was 4.6.0 per cent, by the Friesian on December 14, 1926, the next lowest being 46.4 per cent, by the Ayrshire on October 20, and by the Friesian On November 3. Actually out of a total of 543 analyses of the three types of cheese, 82 or 15.1 per cent of these fell below the 50 per cent, limit set by the New Zealand Department of Agriculture. It will be interesting to review the Now Zealand .Grading of the experimental cheese in the light of this information. In the case of the Friesian breed, not a single crate of cheese was graded “second” from 'September to December, 1926, although a large proportion of the Friesian cheese fell below the standard of 50 per cent, during this period. Seven crates of Ayrshire cheese were graded “second” in September on account of the occurence of a. rancid flavour and its perpetuation for several days owing to the carrying over of a little cold curd from one day to ’the next, but during the whole of October, November and December, the Ayrshire cheese continued to grade “first” even though, like the Friesian, it was frequently an “illegal” cheese and presumably unfit for export. The Jersey cheese made on December 1, 1926, also fell below the legal limit for fat, but nevertheless was graded equally as high (92 points) as other cheese showing over 54. per cent, of fat in the water-free substance. These facts, therefore, prove that cheese falling below the export standard cannot be detected by physical examination readily be defected by physical examination alone, • and that it is not characterised by any obvious inferiority in body and texture. It is important to note that the New Zealand regulations governing export cheese prescribe that the cheese shall be made from normal whole milk, and also that it shall contain not less than 50 per cent, of fat in the dry matter. Since this investigation shows that cheese- made from normal Friesian and Ayrshire whole milk frequently falls below the 50 per cent, standard during the spring months, the effect of the New Zealand regulation is to place undue restrictions upon these wellknown breeds in areas devoted to cheesemaking. It is ‘Significant that these breeds form a large proportion of the cheese herds of our Canadian competitors, 'and that the standard set for export, cheese in Canada is 45 per cent, of fat in the dry matter. In a recent bulletin published by the Department of Agriculture of the Dominion of Canada there may be found many examples of “First” arid “Special ”| grade cheese having less 'than 50 per | cent, of fat in .the watcr-ffcc substance, and considered to possess “good” or “perfect” texture.

Since Canadian cheese has at least during the last few years consistently commanded a premium over New Zealand cheese, we may well ask ourselves whether our regulation is wisely framed, and, if so, what return are we getting for the burden of the extra 5 per cent, of butter-fat in the waterfree substance with which we voluntarily load ourselves.

SHRINKAGE OF CHEESE DURING PERIOD OF CURING. . The measurements of weight made when the cheese was received into the curing room and again when it was being packed for export, served to furnish data for showing the ’comparative shrinkage of the three breeds of cheese during the period of curing. It must be noted that, since the 'cheese was exported in batches of approximately one week’s manufacture, not all the cheese had stood upon the shelves for exactly the same length of time, but that each batch contained some which had stood for not less than 14 days, and others which had stood for gradually increasing periods up to 21 days. Hence the figures given indicate the average shrinkage for a period averaging 174 days.

A number of important 'facts are brought 'to light by a consideration of the shrinkage figures. Firstly the low testing cheese, as was to be expected, has shown a greater shrinkage than the 'high ‘testing, since it contained more moisture. However, in the case of the two extreme breeds, namely tne Friesian and the Jersey, the difference was not nearly as large as might 'have been expected from the comparatively largo differences in their moisture con. tent. Actually the Friesian cheese, containing up to 4 per cent, more moisture than the Jersey, lost only an in-

significant trifle more in weight during the period of curing, the actual difference over 'the whole season averaging only 0.049 per cent. Tills proves that the* shrinkage of normal whole milk cheese is not. proportional to its moisture content., and indicates that the higher moisture percentage contained in low testing cheese is not necessarily lost during curing.

In fact, tlie results of tliis investigation prove that the high moisture content of the low testing cheese may not be reduced by shrinkage during the period of curing to an extent that has any commercial significance. This finding is of the utmost importance from the economic point of view.

Notwithstanding the low temperatures of the curing-room during September and October, higher shrinkages were recorded during these' months than during the height of the summer. This is due tcv the fact that medium chee.se were made during the whole of September and portion of October, after which full size export cheese were mad,3 wherever possible. The greater surface exposed by these small cheese wi'tl furnish an explanation of the greater shrinkage shown by all varieties of cheese during the cool months of September and -October. In the autumn months there is a tendency for progressively higher shrinkages to be recorded, notwithstanding a gradual fall in the average temperatures. This shows that shrinkage is not entirely a matter of temperature and size of cheese, but is connected also with the character of the cheese -itself and with other factors which are at present not fully understood. The gradual decline towards the end of the season in the grading points awarded in New Zealand to a',l the varieties of cheese for “body and texture” is no doubt closely eo-related with the increased shrinkage shown as the season advanced, thus demonstrating that a faulty -texture is highly conducive to a- loss of moisture, even though the average temperature he lower.

Some difficulty will be found in accounting for the fact that Ayrshire cheese throughout the whole season showed the highest loss by shrinkage during curing. If this loss were enen'tirely connected with the original moisture content of the cheese, we should expect the figures for the Ayrshire shrinkage to he practically identical with those, for the Friesian, and. if anything, a slight .improvement upon them. No explanation can lie offered based upon the location occupied bv tlie cheese upon the curing-room shelves, as no attempt was made to allot the different breeds to set positions, nor. indeed, wen? they ever kept separated. The cheese were simply placed upon the shelves in rotation -as received, and, over the whole .season, each variety must have received identical treatment in the matter of the number of times it occupied a given shelf or position.

It is suggested that the texture of the Ayrshire cheese was the feature which caused its shrinkage to be the highest of the three. For era sons already explained, the Ayrshire milk was less _ satisfactory baoteriologioally than either of the other two, and this must undoubtedly have had its effect upon the body and texture of the resulting ohee.se. More important still, the Ayrshire supply of milk was the smallest in bulk. Hence, in the matter of controlling temperatures in the factory, and. in the keeping of the curd warm during the clieddaring process and after mid-king, the Ayrshire must always have been atsome disadvantage when compared with either of the other two. breeds. Another fact worthy of mention is that the Ayrshire milk was invariably received into the factory about an hour later than the others, and that, in consequence, the manufacture of the Ayrshire cheese was always somewhat more hurried than that of the other two varieties. For these- reasons it will cast no .slight upon the Ayrshire breed to state that tlie texture of its cheese in this test could hardly be expected to be as good as that of either of the other two. In fact, it is quite probable that the actual difference would be greater than appears to be indicated by the following average figures for the season given for “body and texture” by the New Zealand graders : Jersey 28.25; Friesian, 28.03 ; Ayrshire 27.89. Summarising the -results of this section of the investigation, we may consider the following points proved:—1. Tb,at normal whole milk low

testing cheese containing a high percentage of moisture does not neces■sani'ly .lose appreciably more in weight than more fatty cheeses of lower moisture content. Actually the slight extra amount lost is not sufficient to discount to any appreciable degree, the economic advantage gained by the use of such milk for chee'semaking. . . , 2. The body and texture of clieese! is shown to be even more, important than temperature as a factor in determining the percentage of immediate shrinkage.

LOSSES OF FAT AND CASEIN OCCURRING DURING MANUFACTURE.

On 194 days throughout the eight months of the investigation, the losses of fat and casein occurring in the whey were determined, with a view to finding out whether any one variety of milk is more conducive to lasses than another. The samples were taken about the middle of the “run-off” from the wliey-tap, and, to’ avoid the inclusion of solid pieces of curd, were carefully strained by filters of cheese cloth placed over the tap.

in a table which lias been prepared the fab losses are shown to rise more or less steadily towards the end of the season, with an exception occurring in the month of September for all breeds. In this mouth the fat losses are all higher than those found in any subsequent month until February. This is explained .by tlie fact that the change from colostrum to normal milk is a gradual one, and that the emulsion of the fat is not at first perfect, the globules being large and irregular. For this reason they separate much more easily from the milk and from the curd, thus making the fat losses higher at the beginning of the season. The occurrence of the yellow fat floating freely on the surface of the milk in the early spring is extremely familiar to every eheesem aker.

The casein losses are relatively low, anid show comparatively, small Variation, but there is a distinct tendency to a reduction as the season advances. The a verage losses of casein in the whey from all breeds of millk for the months of February, March and April are all lower than the respective •seasonal averages. Moreover the lowest los.s of casein for the whole season occurs in either March or April for all breeds, thus proving that the loss of casein in the whey is not a function of the initial amount of casein present in the milk.

_ln other words, a milk rich in casein doe§i not necessarily lose more during the process of chees ©making than the same milk when relatively poorer in casein at an earlier period in, the season.

PROPORTION OF ORIGINAL FAT AND CASEIN PRESENT IN MILK LOST IN THE WHEY.

A study of the figures for the losses of fab and casein .in the wliey might lead Qiie to suppose that the lowest testing milk is necessarily the most economical in use, as it leads to smaller losses of fat and casein in tfi© whey. It must be remembered however, that tiie lowest testing milk originally contained the smallest percentage of fat and casein, and that therefore a true comparison of the losses ocourriug in the wliey can only be made by comparing the proportions of the original fat and casein present in tlie milk, which were lost in the wliey. In order to arrive at this comparison tlie percentages of fat and casein lost in the whey were expressed a.s a proportion of the fat and casein occurring in the milk on that particular day. Considering firstly the percentages of fa.t lost as the season advances, we observe that, not only does tlie actual percentage contained in the whey increase, but also that the loss is proportionately heavier as the fat test rises with the advance of the season. For example, during the spring months about 7 per cent, of the fat entering a cheese factory would, be lost in the whey, whereas, in the late autumn, over 10 pe.r cent, of this commodity is lost. Of courst the fat is ’.©covered by separation of the whey, but the butterfat is less valuable in whey butter than in cheese. This fact, coupled with the low yields of cheese per pound o.f butterfat experienced at the end of a milking season .indicate that a change over from the manufacture of cheese, t 0 that of butter and casein might be profitable rather earlier in the .season than is frequently supposed, assuming of course, no great disparity in market returns from the two systems. The figures for the- percentages of the casein supply which run to, waste ave very important, indicating as they do, that tins loss has .a tendency to- become lower as the season advances. In this respect, the casein losses are quite The opposite of the fat losses. Apparently end-season milk is more perfectly coagulated with rennet, this action being probably assisted by the increased proportion of mineral matter present in the milk as the lactation period draws to a close. Also, the relative smaller bulk of whey late in the autumn oilers loss facility for the loss of casein. Finally there is no tendency tor the casein, like the fat, to separate upon standing, or to be agglomerated into- particles by churning or agitation. Hence there is no counter-balancing tendency to overcome the above mentioned influences, all of which exercise come part in reducing the losses of casein in the whey to a minimum at the end of the milk-

ing season. When comparisons ate made between the losses of fat and casein if the milks of the different breeds it becomes obvious that the richness or poorness of the milk has little or no intiueuce upon the proportion of fat and casein list. At .corresponding periods of the season the losses are, with minor irregularities of the same order, of magnitude. Finally the seasonal averages- are, at least in the cases of the two extremes (Friesian and Jersey) practically identical. The fat lost ever the wnole season from the Ayrshire breed is somewhat in advance of that from either

of the other two, but sufficient has already been said of the disadvantages from which the Ayrshire breed suffered 1 in this investigation to- justify the opinion that, had conditions of handling in the factory been identical, even this small difference would have disappeared. The two extremes of high and low testing milk however, have lost, over the whole season, practically the same proportions of their criginai fat and oas-O'n, thus proving rhat there is no .artiouilar breed of milk, either nigh or low testing, wlr' h >s more conducive to losses in eheesemaking than anoth-r.

| SUMMARY OF RESULTS OF INVESTIGATION UP TO TIME OF ) EXPORT OF CHEESE.

(1) Large daily fluctutaions have been observed in all the analytical figures relating to the- three breeds of milk. Many of these variations are apparently correlated with meteorological conditions, as many of them occurred simultaneously with tlie three milks'.

(2) The lower testing milks (Friesian and Ayrshire) have on occasions, and the Jersey only once, been found to fall below the legal limit of 8.50 per cent for solids-not-fat without adulter t-ai-on with extraneous- water. The Friesian milk has also been found to fall naturally below the legal limit of 3.25 per cent for fat- in milk.

(3) Large daily fluctuations occured in the yield of cheese per pound of butterfat from each variety of milk in the factory.

The low testing milks (Friesian and Ayrshire) always showed, at the same period of the season, a higher yield of cheese per pound of butter-fat than the high testing (Jersey) milk. The Friesian, possessing the lowest fat test, showed to the greatest advantage. Gross yields of over 31b of cheese to lib of butter-fat were obtained by this breed of milk over three consecutive periods of 10 days each, and the lowest yields per pound of but-ter-fat (2.652) *by the Friesian breed was higher than the highest (2.621) secured by the Jersey breed during the whole season.

Expressed in percentage form, the actual average yields secured over the whole season show a strong preponderance in favour of the low testing milks, the Friesian producing over 13 per cent, more cheese and the Ayrshire over 9 per cent, more cheese per pound of butter-fat than the Jersey. (4) The cheese made from the three varieties of milk graded mainly “first grade” throughout the season. Out of 714 crates 36, or 5.04 per cent, were graded “second.” Individual amounts were:

Friesian, 10 crates out of 310, equal to 3.23 per cent. Ayrshire, 20 crates out- of 181, equal to - 11.05 per cent. Jersey, six crates out of 223, equal to 2.69 per cent.

The Jersey cheese, on the whole, graded better than either of the two low testing varieties, the premium in points being gamed mainly on the score of flavour, with ocoasionally half a point extra for colour and at times half or one point for body ami texture. »

The Jersey cheese secured 18 crates of “finest” grading (93 points) during January and February, 1927, hut thereafter declined rapidly in March and April owing to high 'fat test in the milk, and. finished the season “second grade.” During the same time the low testing Friesian cheese tended to improve its average grading.

Tiio Ayrshire cheese declined l to “second grade” after February, but this has been ascribed to an undue handicap placed upon the quality by the existence of the “bucket” type of milking plant,

(5) The two low testing varieties of cheese. (Friesian and Ayrshire) were found to contain practically three per (?ei)t more moisture than the high, testing Jersey cheese during the spring and summer months; and about two per cent more moisture in the autumn months, without any commensurate detraction from body and texture. (6) The Jersey cheese showed from three to four per cent more fat in the water-free, substance than either of the other two varieties throughout the whole season, and only once fell below the New Zealand legal limit of 50 per cent. The Friesian and Ayrshire cheese were very similar in respect to their fat contents at the beginning and end of season, with art average varying from one to two per. cent in favour of the Ayrshire during the summer months.

The differences in fat content were not uniformly perceivable in any commensurate advantage in grading points for body and texture gained by the Jersey breed. From September to December, tlie Friesian and Ayrshire cheese, frequently fell below the Xew Zealand legal limit of 50 per cent of fat in the dry matter without the slightest disadvantage to body and texture in numerous instances. At no time did the Friesian and Ayrshire; fall below the. Canadian standard of 45 per cent in the. dry matter, the nearest approach being in a single instance of 46 per cent by the Friesian breed.

( (7) The shrinkages of the three ! varieties of cheese were measured duriing the period of curing over the | whole season. Body and texture . were, shown to be even more import- | ant than temperature and . initial i moisture content in determining the j immediate shrinkage. The Friesian ' cheese, containing from two to four .per cent more moisture than the Jer- ! sey. was found to lose only 0.049 per cent more than the Jersey during the neriod of curing over the whole season. Hence the economic advantage of manufacturing low testing milk into cheese of high moisture content is not necessarily offset by any considerably greater shrinkage up to time of export, (8) The losses of fat and casein occuring in the wheys from the three milks were measured throughout, the whole season. For each individual breed, fat losses were shown to be actually and relatively high in the opning month of the season, owing to the physical state of the fat of the very earlv lactation period. Fat losses tended to decrease up to December, after which the amount of fat lost in the whey steadily increased, both actually and relatively, until April of the investigation, when over 10 per cent of the fat in the milk was being lost in t-lie whey. High casein losses occurred in all breeds in the whey during September and October, followed bv a sharp decrease in November. There was a slight rise in December and January, after which the casein losses, although increasing slighly in numerical value, steadily diminished in relation to the actual amount of casein present in the milk. Over the whole season the whey losses of fat and casein relative to the amount originally present in the milks were shown to be identical, thus proving that high and low testing milk alike lose the same proportion of their fat and casein during the process of eheesmaking. In other words, no one variety of normal milk is more conducive to losses than another.

(To he continued to-morrow.)