Giving a total of 77-263 cc. of carbon dioxide in 19 days, the temperature ranging between 14° and 19°5. Milk previously deprived of dissolved air by the Sprengel pump, then confined over mercury and submitted to an atmosphere of oxygen, rapidly absorbs the oxygen, the place of which is taken by carbon dioxide, provided the temperature is a fermentation one-that is, above 9° and below 60°. This continuous absorption of oxygen was well shown in an experiment of the author's, in which a litre of milk was submitted to the action of a Sprengel pump, and in which it was found there was a continuous slow diffusion of air through the india-rubber connections. It has long been shown by Graham, that air thus finding its way through the minute pores of thick rubber is very highly oxygenised; yet all oxygen rapidly disappeared from the gas, and after the second day pure nitrogen and carbon dioxide could alone be obtained :—

§ 123. If an animal is fractionally milked-that is, the whole of the milk received into three or four different vessels-it will be found that, on analysis, the several portions exhibit some difference of composition, more observable in the last and the first, than in the intermediate portions. This difference mainly affects the fat, the first portions of the milk yielding, as a rule, but little fat, while the latter portions, called "strippings" (in speaking of cows' milk), contain an excess of milk-fat. Thus, in a Devon cow milked in this way for the purpose of analysis, the writer found the two extreme portions to have the following composition -

In another experiment a Guernsey cow yielded the following:

Dairymen are perfectly aware of the poorness of fore milk in

fat, and more than once fraudulent milkmen have endeavoured

to defend themselves by having recourse to the strange expedient of partially milking a cow before such functionaries as aldermen or policemen, and delivering with all formalities the sample to be analysed. The analyst, not knowing its history (for in such cases it is transmitted as an ordinary commercial milk), and finding it on analysis deficient in fat, certifies accordingly, and until the matter is explained suffers in reputation. Such tricks have during the last few years been rather common, but so fully exposed that they are not likely to re-occur.

This difference in the first and last milkings is not confined to cows' milk, but has also been observed in the milk of other animals. Peligot had an ass milked in three successive portions, and found as follows:

Reiset has also found a considerable difference in the percentage of total solids in human milk in fractions taken before the child was applied to the breast and after.

It has been considered that this difference is merely due to the effect of a physical cause; that, in short, as regards cows' milk, the milk already secreted is in the same state as if it stood in a vessel, and the fat rising to the top is, of course, drawn last. This explanation cannot be altogether true, for the same phenomenon is observed in human milk, and here the breasts are horizontal, or nearly so. It is more probable that during the act of milking secretion goes on, and it would seem that the fatty contents of the milk-producing cells are set free before the more watery and albuminous. Hence, the strippings are, as the most recent portions of the whole secretion, rich in fat. This view is supported by an experiment of Reiset, in which it was proved that the longer the time elapsing between the partial milkings, the less the percentage of solids.

* J. Reiset: Annales de Chimie et de Physique, 3 ser., xxv., 1849.

HUMAN MILK.

§ 124. Woman's milk has been long an object of research, and numerous analyses of it are scattered through scientific literature. These analyses, in their quantitative results, show considerable discrepancies, so that we must either adopt the supposition that human milk is very variable, or, what is more probable, that the samples taken did not represent the average secretion. From experiment, the author has come to the conclusion that it is impracticable with any mechanical appliances to obtain a complete sample of human milk. In civilised life the nervous system assumes such a high and delicate state of organisation, that the secretion is far more dependent on the presence and contact of the offspring than among animals. Hence, samples of human milk taken by breast pumps, or other exhaust apparatus, can only be considered partial samples; and a study of partial sampling in the case of cows' milk (p. 253) has taught us how very widely the quantities of the fatty constituents in such samples differ from one taken from the whole bulk. Woman's milk contains milk-fat which has not been obtained in quantities sufficient for accurate investigation, and a knowledge of its exact composition is still a desideratum. It, however, certainly contains butyrin, for the author has succeeded in isolating a sufficient quantity of butyric acid from saponified human milkfat to identify it satisfactorily. Milk-sugar, casein, albumen, peptones, and a colouring-matter, with mineral substances, are also constituents of woman's milk.* The casein, like that of the ass, is peculiar in not separating in flocculent masses by the processes recommended at p. 282), and the analyst is under the necessity of adopting a different process. This difference is allimportant; for in artificial feeding with cows' milk, as soon as the milk reaches the stomach, the milk, in popular language, "curdles," and is often rejected by vomiting.

One of the earliest exact analysis of human milk was made by Meggenhofen + in 1826. His treatise scarcely appears to be known, yet it contains pretty well all that is known of the composition of human milk. The total solids of human milk Meggenhofen determined from twelve samples, the highest of which is 13:38, and the lowest 9-25 per cent., the mean being about 12 per cent. Probably for the first time Meggenhofen determined the

* Human milk on being shaken up with ether parts with its fat, the globules dissolving in the ether, the fluid therefore clears up and separates into two layers-an upper ethereal containing the fat in solution, and a lower layer consisting of a solution of the casein, albumen, and salts. This peculiarity is not shared by the milk of the herbivora. See P. Radenhausen, Zeit. f. physiologische Chemie, 5, 13-30.

+ Dissertatio Inauguralis Indagationem Lactis Muliebris Chemicam. C. Aug. Meggenhofen. Frankfort, 1826.

CP IT

THIVEA

albumen separately from the casein, and also weighed "materiæ animalis tinctura gallarum animalis precipitata." His view of the composition of human milk may be fairly stated thus

According to the writer's own experiments on human milk, and the quantitative analysis of samples taken as fairly and completely as can be done, it has the following composition:

With regard to other constituents, urea is often present; there is also an odorous principle. Human milk decomposes similarly to cows' milk, and yields similar gaseous and other products.*

§ 125. Milk of the Ass.-The author has investigated the milk of the ass. Milk was obtained under his personal superintendence from asses kept and fed in London dairies for the purpose of supplying the demand that still exists for asses' milk. The

* A. R. Leeds (Chem. News, 4, 263-267, 250, 281) has examined eightyfour samples of human milk, and gives the results as follows: