300 MAINE AGRICL'LTURAI, KXTERIMKNT STATION. I913. 



weather conditions, etc. In addition there are undoubtedly 

 many internal factors involved, such as, for example, the ner- 

 vous condition of the cow, the general state of metabolism, etc. 



There is a widespread belief, which has indeed found its ex- 

 presson in legislation, that in spite of the variation in the milk 

 of the individual, if all the milk of the cows of a large herd 

 be put together and thoroughly mixed, the resulting compo- 

 site will not var}^ significantly from day to day in its fat con- 

 tent. It seems to be a matter of considerable importance, as 

 a basic datum in milk production studies, to know the actual 

 facts regarding the daily fluctuations in fat content of the 

 mixed milk of a large herd. 



The problem has recently been studied by Klose" for a herd 

 of 70 cattle at the Milchwirtschaftliches Institut in Proskau. 

 Onlv rather crude and inadequate statistical methods are used 

 by Klose in the analysis of his data. Under these circum- 

 stances, and because of the valuable character of the raw data 

 which are fully given in the paper, it has seemed desirable to 

 apply biometric method's to these figures with a view .to get- 

 ting an accurate and trustworthy measure of the degree of 

 variability shown. It is the purpose of this paper to present 

 the results of such an analysis of Klose's data. 



It may be said briefly that the records analyzed come from a 

 herd of 70 animals (breeds not specified) and cover the milk 

 of 30 consecutive days in each of four periods of the year. The 

 cows were milked three times a day. morning, mid-day, and 

 evening, and a separate record kept of the milk from each 

 milking. The data include die folio-wing items; (a) The Aveight 

 of milk produced at each milking, in kilograms: (b) the 

 specific gravity of the milk; and (c) the fat percentage of 

 milk. All of these figures are given for the three milkings of 

 each day separately, and then for the total milk of the whole 

 day. The four periods chosen for the records comprised 

 practically the calendar months of Alarch, ]\lay, July and Oc- 

 tober. The reason for choosing these particular months was 

 to get the greatest possible contrast in regard to feeding and 



^Klose. Untersuchungen iiber die taglichen Schwankungen im spezi- 

 fischen Gewicht und im Fettgehalt der Milch einer grosseren Herde. 

 Milchw. CentralbI.-]\Iilch-Zeitg. Jahrg. 42, pp. 385-392. 1913. 



