580 Scientific Proceedings, Royal Dublin Society. 



of sugar. In an investigation extending over a short period, these changes 

 are not obvious, particularly in the case of solids-not-fat where the percentage 

 remains a practically constant quantity. But the proportions of solids-rint- 

 fat and of water are subject to slight daily variations. This can be observed 

 on diagram I, and is better illustrated in diagram III. 



II ere 'the percentage variation in the milk of each day above or below that 

 of the previous day is represented by a series of vertical lines above or below 

 a horizontal one marked off in days. Beside the line representing the 

 variation in milk there &re three others for each day, one for the water, 

 one for the solids-not-fat, and one for the total fat, each representing, as in 

 the 'line for milk, the percentage variation on the previous day's yield of that 

 constituent, whether an increase or decrease. In this way the variation of 

 the four quantities can be compared by observing the comparative lengths of 

 the four lines for each day. The returns from cow 155 for the period April 

 24th to May 12th are represented on the left side of the diagram (p. 578), and 

 those from cow 128 for the period June 1st to June 17th on the right (p. 579). 

 The first group (left side) is taken from a portion of diagram 1 where the line 

 representing fat approximates most closely to the other three (milk, solids-not- 

 fat, and water), whereas the second group (right side) represents a much more 

 representative section of the curves. From diagram III it is obvious that 

 the percentage increase or decrease of water and solids-not-fat on the 

 previous clay is not always exactly equal to the increase or decrease in milk, 

 the variations being at least sufficiently great to prevent the utilization of 

 formulae involving these factors in the determination of a result which can 

 claim to a greater accuracy than the nearest whole number. For instance, 

 the addition of water to milk cannot be determined accurately to a decimal 

 when the variation in solids-not-fat is more than this quantity. When, 

 however, the variations in solids-not-fat, water, and fat are compared with 

 the variation of the total milk, we realize that while the first two constituents 

 vary approximately as the milk the total fat varies in a fashion peculiar 

 to itself. 



That is to say, fat is the great variable constituent of milk. In the mixed 

 milk of a herd much smaller daily variations appear, the solids-not-fat pro- 

 bably remaining, for a short period, nearly constant. This is due to the 

 fact that diverging fluctuations in different cows balance each other in the 

 composite sample. A lactating cow produces water, and solids-not-fat, in 

 proportions which are practically constant, and produces fat in proportions 

 which bear no constant ratio to the total milk, thus indicating that fat- 

 production is influenced by factors 1 which are not identical with those 

 influencing the production of- the other constituents of the milk. 



1 To be dealt with in a later paper. 



