HERD TESTING

Hawke's Bay Herald-Tribune, Issue 199, 8 September 1937, Page 13

HERD TESTING

The Question of Accuracy in Returns « — — VITAL REQUIREMENTS It is often contended, Avhen comparisons between herd-test figures and butter factory returns are being made, that herd testers are not working under laboratory conditions, and therefore, their results cannot be considered reilable. "YYhue the writcr's own experiments show that this is not necessarily the case, it is true that much depends on the methods of individual testers and on their ability to adapt themselves to a variety of conditions. Time has a great bearing on a tester's accuracy. "iTsually he has to aim at completing his work, including the recording of results, in time to move to anqther l'arm for the evening milking. On some farms this is easy, even wlien tliere is a large number of samples to be tested. On others, owing to lateness in finishing milking, meal tinies wliich ao not fit in well with the work, and various otker forms of inconveni.ence, it may n« difficult to test a small numher lii the time avaiiable. Usually, when theso conditions seriously prejudice the tester's chances of completing his day's work, it is possible to come to some understanding with the farmers wliich ^vi-1 overcome the difficulty, but there is a definite danger that the junior, still siow in his methods and at' the same time feeling that he is among comparative strangers, may'be tempted to sacrifice accuracy for speed, and so form bad habits in the early stages of his testing career. There ai'e, of course, legitimate ways of saving time. The outfit should be arranged in such a manner that everything is within easy reach when vvanted, and, with the jiossible exception of the centrifuge, which necds a solid foundation, should be so placed as to eijminate the ueed for stooping. Where no other bench is avaiiable it is usually possible to find boards which can be placed across the rails separating tlie bails in the cow shed. Oid packing cases, empty petrol drums, and in fact, a wide variety of things, may be of help in the improvisation of benches. But perhaps thc bost way to attain speed is to keep one's mind on the job and plan every move ahead. The amount of time lost by a wandering mind is sometimes astounding. During sampling it is a good plan to note anytliing unusual about the milk yields and, if possible, find out its cause. I11 the rules an abnormal result is defined as oue iu whicb tbe production is more tlian 2-5 per cent above or below the normal production (ineidentally, tlie rule reads with some ambiguity as to tlie meaning of "production".) Testers usually find, however, that this rule cannot be applied rigidly aud indiscriminately, and that some judgment must be used if oue is to arrive, as uearly as possible, at a cow's true production. This can only be done when all the possible data is obtained and taken into consideration. A tester also has a good opporfunity to detocb tlie presence of a sjck cow in tlie herd, as ihe cliaraeteristic sweetisli smell of "the milk is very pronounced when it has been mixed for sampling.

Importanoe of Temperature. As in most cliemical processes, so in testing, temperature js of great importance, and in this matter great adaptability is required on the tester's part. Une day he may be working in strong sunlight, and have difficulty in obtaining water as low as 70 degrees. On another occasion he may be in a comparatively open shed, in the track of an east jvind, which quickly reduces everything to 40 degrees. Temperaturws are regulated by the use of liot water.' Where there is a co'pper and plenty of firewood the supply is at its best, but even then the tester is well advised to take an acfive part in attending to it. Otherwise, just as he wants a temperature of 140 degrees, he may find that some well-meaning person has added cold water, which must heat up before the work can proceed. The most difficult cases are those in wliich a small quantity of water is heated in kcttles on a sbove some distanco away. This is quite workable when tlie samples to he tested are i'ew, but when tlie number is large and tlie weather cold the tester's adaptability is taxed to tlie utmost. Farmers should understand that tlie hot waLer is not required all at cnee, but at intervals, and that tho temperaturo is a'H-important. Actually, the amount needed is not large, but n fair quantity should be heated so that it will retain its temperature for somu time if the fire liappens to burn low. Where water has to b& brought some distance from lvnere it is heated, it is a good plan to wrap bags round the bucket by way of heat insulation.

In the liiterests or accuracy some sfcandard ought to be fixed as to equipment. The writer's outfit has been, on the whole, very convenient, but he lias seen, and occasionally used, others which were lacking in many ways. The water bath in which the samples are warmed should be partitioned off into twenty-four spaces, and tbe bottles should stand on cross wires some distance above the bottom of the vessel s0" that the water warms them from beneatli as well as from the sides Verv often a cut-down kcrosene tin is used. In a simple tray of this sort so littlo water can be added without ovcrbnlaneing tlie bottles that oti a cold day it may fail to raise tho samples to tlm j propcr tcinperatures. 1 Tlie lieatiug water, wliich slioulfl be several degrees warnier tha 11 Ihe temperature required in ihe .samples, should | be mixed in a sepurale bucket, and l

afterwards sliould be taken first, aua the cold water added to it. If the hot is added to the cold, imperfoct mixing will result and the upper and lower layers of Avater may difler by as much as 20 degrees. Atmospheric Conditions. Milk and acid are normaily mixed at 70 degrees. In summer the milk may need cooling to that temperature ; in winter it may be necessary to keep it warmed until just before a'dding the acid. Where atmospheric conditions make it almbst impossible to kold tho right temperature, a correction can be made, with experience, by adding more acid for cold conditions and less for hot. If clear butter-fat columns result,- the amount can be considered correct. Some time ago the'writer tried to determine the error mcnrred in attempting to read curdy and charred colums. Three tests of the one sample were taken. In one the normal amount of acid— 17.5 c.c— was used; in ancther, considerably less, and in the tliird considerably more acid. Ail were mixed at 70 degrees. The result was that all three tests were clear, aml gave the same reading ! Evidently the amount of acid is fairly elastic if otlier conditions are normal. Another experiment was made. No 1 was the normal clear check, and read b.7 per cent. No 2, with considerably excess acid a,t 70 degrees, added tc milk at 90 degrees, was ratlier charred, and read, as far as it is possible to read a charred coluinn with an uneven lower end, 5.75 per cent. No 3, with less tlian normal acid at 70 degrees, added to milk at 50 degrees, was very curdy, and read oniy 5.5 per cent. This shows that to adhere as closely as possible to standard amounts and temperatures is the best course. It may be found advantageous, however, to add more acid when the samples are known to be of a very higii test, since the ncn-fatty solids are also liigh. It has been actually observed that in such samples there is a strong tendency for boiu» curd to remain undissolved if standard amounts are used. » Strictly speaking, the mixing of milk and acid by skaking in a certain way should be done as soon as the latter is ruu into tlie test flask, but .most testers seem to revert to the system of measuring the acid into all twentyfour fiasks, after which, in a suitably constructed holder, they can all be mixed in the one operation, with great saving of time. In some large herds only this method makes it possible for the tester to fit liis work in with the farmer's arrangements, and finish in the time avaiiable. Is accuracy being sacrificed? [Probably not, if the tests are not left standing uiimixed for too lciig, and the mixing is done carefully, so as to ensure a clear result. In adclition, there is less time for loss of temperature to occur before the first centrifuging — an important consideration, since everything depends on the fat being mobile enough to rise to the top. No argument, however, can justify the cutting down of the times of centrifuging, l'or, though a clear test may be obtained, there is nothing to indicate whether all ihe fat has risen in the shorter time aliowed. Ncr is it advisable, for the small time saving involved, to read tlie fests direct from the centrifuge unless compelled to by Some liitcli jn the hot water supply. Where tiventy-five or moro samples ar» to be tested, the water in which the first batch is placed for reading can, with a little cold water added, be used to warm the second batch of samples. Clean glassware has its diilieulties. In tlie sampling pipetto washing alone is not eff-octive, as the milk has so much opportunity to dry on its sides. The oecasional use of clean sand or lead shot will overcome this. The test fiasks, too, often become coated inter-1 nally with an albuminous deposit, but this can be prevented if they are cleaned by the following method, recommended by Wing: — "As soon as the bottles are read, and wbile they are still hot, the contents should be emptied out. The hot acid and water will carry out with it the larger part of the fat in the neck of the bottle. The emptied bottle should then be rinsed once in warm water and once in hot water containing some allcali, either washing soda or any of the various washing powders, and theu j rinsed with either warm or cold water f until they are perfectly clean." I Probably many testers make the 111181 take of using hot water for 'the first riusing, wliich evidently tends to eoagu* late the albuaiinous substances present. Thc outfit usually includes a copper vessel litted with a tap. This is very useful for running nnsitig water into the bottles. Speed and accuracy iu coinpuling butUr fat results come with, practice. A carefui inan will always check his figures, and his aim should be to perform the check in a manner wliieli difEers i'roni the origiual calculation, e.g., by starting at the opposite end in an addition, to avoid niaking the same mistake twice. Here is a plan which re'tuces checldug to a minimum: — 1. Add "daily milk" column. To check, add ''evening" and "morning" 1 columns. This also cheeks the cross additions. 2. Calculate milk for period and daily butter fat. From the latter work out butter fat for period by multiplying by number of days. All these figures can be checked at once by multiplying milk l'or period by test to get butter fat for period.

3. Add "daily butter-fat" column once from eacli end. 4. Calculate average (daily) intlk and butter fat and, l'rom them, average test. Then calculate average test from total milk and total butter fat. This cheeks all the herd avernge figures in one oper c.tion. | Paris of this plan will not apply !« J those who use ready rcckoners- Tlm j wj-itcr hns always used a slide rule, wliich is adaptable to all multiplicatiou

and division, afid provldeB a very i'ast means of checkiug. As a means of checking aeeuracy it is interesting to compare test figures and factory returns, provided the comparison is properly made. In one experiment the writer endeavoured to aceount for all the losses which occur on thc farm. A composite sample was taken of all milk, from "strippers," and the herd's total production was obtained. Loss in skim milk, household milk, and in milk used for calves, dogs and cats' liouehold milk, was calculated. The day 's test samples were deducted; also milk clinging to the sides of the buckets at the end of milkiiig (this amounted to '1/20 lb. milk per bucket for aach milking). To determine losses in other utensils, the washup water was weighed and a sample tested in a skim milk bottle. Its test was 0.14 per cent. Cream for household use was also aliowed for. The day's cream was sent to the factory in a separate can and the results compared. Later it was decided that some of the allowances were too small to justify their iuelusion, and only skim milk loss and milk for household use were deducted. For one herd the following results were obtained: —

These figures show that, with proper care, herd-test figures can very closely repTesent actual production on the day of test. The greatest difference ahown is 1.31b. Spread over about 40 cows, this only meaus a very small error per cow. If average conditions are encountered over the nine test days of a cfiw's lactation — and this should be so if the farmer is regular in his methods — one would not, therefore, expect a very large error in the season 's total. And, after all, the records of individual cows are of more* practical importance to the owner than is thc herd average.

— Butter FatHerd Factory Date. test. test. 10-1-37 33.49 32.30 6 2 37 27.35 27.40 8 3 37 32.73 31.40 10-4-37 days cream mixed by mistake 7-5-37 25.68 26.10 11-6 37 19.57 18.50