WHY VARIATIONS OCCUR

Southland Times, Issue 23610, 10 September 1938, Page 19

WHY VARIATIONS OCCUR

Herd And Factory Tests INACCURATE SAMPLING BLAMED In 1929, Dr H. E. Annett _ (Matangi) carried out some investigations which indicated the possibility of a presence of a fat-destroying ferment in milk m the Waikato, possibly derived from feed. More recently Dr Annett has been studying closely the variation between factory tests and herd tests, and it appears as if that original work, which has never been followed up, may provide the solution. In an article in The Dairy Exporter Dr Annett summarizes this recent work which, says the Exporter, seems to maleate that errors occur in the factories under the present methods of sampling milk. The article is as follows: — . A common complaint among dairy farmers in New Zealand is that the factory test is usually lower than the fat test shown by herd test associations. The fact that there is such a difference is frequently held out as an argument against herd testing. In discussing the matter with representatives of the dairy factory interests, the writer has been struck by the fact that they usually dismiss the matter by assuming that the herd-testing figures give high results, and are almost inclined to ridicule the idea that there can be anything wrong with the factory test. This attitude is damaging to thq interests of herd-testing, a movement which has done so much in this and other countries towards increasing herd production. Why the factory side of the industry assumes that the herd-testing figures are at fault, rather than their own figures, is difficult to understand.

FACTORY SAMPLING MORE CRUDE The writer has now nearly 10 years’ experience of comparison between herd-testing and factory figures for butterfat test. Under the herd-test-ing association test, the testing officer weighs the milk of each cow in the evening, and takes a measured quantity of milk into a sample bottle labelled with the cow’s alloted number. The size of the sample depends on the quantity of milk given by each cow, a minimum of 20 c.c.’s being desirable. At the morning milking the same proportion in c.c.’s a pound of milk is taken as was taken the previous evening. The evening’s and morning’s samples of each cow are mixed together for testing purposes, and placed in a bottle with the cow’s appropriate number. As a check on the individual sample from each cow, a composite sample of the milk of the whole herd is also obtained by taking an additional sample (morning, and evening) consisting of 1 c.c. for every pound of milk produced by each cow. This composite sample of milk is analysed, and provides a figure for the total fat produced for the day by the whole herd which can be compared with the figure obtained by adding up the fat for each individual cow. Figures obtained in this way show extraordinarily close agreement, and it is difficult to see where there can be any constant error in determining the average butterfat test for the herd by the above methods. The factory, however, has in the past taken its samples in a far more crude manner. The milk is tipped from the cans into a vat at the factory and is weighed. A dipper containing one ounce is thrust into the milk in the weighing vat, and is poured into a sample bottle. Separate bottles are kept for morning and evening milk, and the samples are usually analysed for fat every 10 days, a tablet of potassium bichromate being placed in each bottle as a preservative. During the lower production periods of the year, when once-a-day delivery is practised at the factory, morning and evening milk is mixed and samples are only analysed twice a month.

COMPARISONS BETWEEN TE&TS From the results accruing over the ten-year period Dr Annett has compiled a very comprehensive table, the results of which he has carefully tabulated. He considers it a fair thing to assume that where the factory figure differs no more than .25 per cent, from the herd test, the agreement is satisfactory. He, found that in 108 comparisons shown between factory and herd test figures there were only 14 cases where the herd test proved to be lower than the factory figures. On the other hand there were 62 cases where the herd-testing figure exceeded the factory figure by more than .25 per cent.

There was considerable variation from year to year, but Dr Annett was unable to relate rainfall conditions with differences between the factory and herd-testing figures. There seemed to be a distinct indication of a greater difference between the factory and herd test figures from February onwards. During this period differences up to 1 per cent, were observed, and in general it seemed that the greatest discrepancy occurred during the colder months of the year.

Several points which are frequently made about testing are cleared up by Dr Annett’s work. For instance, it is frequently stated that as the milk of a herd varies day by day, and the herd-testing samples are taken only once a month, some difference should be expected between factory and herd test results. The doctor points out, however, that in that case it could be expected that the herd-testing figures would be as frequently lower than the factory tests as they would be above it. Similarly, the point sometimes made that the milking into the bucket on herd-testing day gave a higher figure is disproved because Dr Annett himself used a bucket type of plant until 1934, and this in no way affected the amount of difference between factory and herd tests. This also disposed of the idea that washing water going through the machines is responsible, for with the bucket plant no washing water was going through.

WHAT IS EXPLANATION?

“What then is the explanation of the difference observed between factory and herd tests?” asks the doctor. “One obvious cause suggests itself, namely, inaccurate sampling at the factory. The herd-testing officer samples .warm milk, a representative sample of which is easily obtained. Before the milk goes to the factory it usually, though not always, passes over a cooler. It stands for some time in the can, and during the journey to the factory undergoes a mild process of churning.

In consequence lumps of fat separate, and since the tendency is for fat to rise to the surface, any sample from beneath the surface of the milk must necessarily be lower in test than that of the whole milk. “As one who has been associated with sampling of factory raw material, the writer has been amazed that more detailed study has not been paid to the method of sampling in our dairy factories. The proof or disproof of the assertion that defective sampling is at the root of this trouble, is easy. AU that is necessary is to collect the milk from a supplier’s shed straight from the cows, using a bucket plant. The milk must be mixed warm in a large vat, and samples taken for analysis at least in triplicate. This milk/should then be put over the cooler in the shed, distributed in cans, taken to the factory in the ordinary way, and sampled. “A further subject worth investigation is the possibility of the presence of a fat-destroying ferment in milk, possibly derived from feed. In a paper in The New Zealand Dairyfarmer (March 20, 1929) the writer showed that white clover was the cause of a very bad feed flavour in milk in the Waikato. Although at the time this idea was not accepted by other workers in New Zealand it is interesting that they generally agree with it at present, though giving no acknowledgment to the original work. In that work, however, the writer produced ; suggestive evidence that the feed flavour was produced by a ferment in the clover which, at the same time, caused a destruction of fat in the milk. Long experience of biochemical work convinced the writer that the results obtained were worthy of further investigation. Practically nothing is known of the relationship between enzymes in fodder plants and their effect on the quality of the milk. Here is a field worthy of investigation. It is quite possible that at certain times enzymes may play a part in the destruction of fat during the storage of milk samples. The writer is convinced that in the case of clover taint, we have to deal with a biochemical problem, and not with the simple presence of an essential oil.”