far the limits generally adopted by public analysts in the examination of butter were to be relied on.

The Royal Agricultural Society of Denmark, which, in many respects, acts as a branch of the Danish Home Office, now became fully alive to the importance of the question at issue, for Denmark being practically an enormous dairy-farm-a large proportion of the produce of which is sent to England-it was felt that the abnormality which had been observed in the case of the Swedish butter might very possibly extend to certain Danish butters, and hence it was very important to know the variations in composition which were natural to butter under exceptional circumstances. In support of this view, and at the suggestion of Mr. Harald Faber, Agricultural Commissioner abroad to the Danish Government, I was formally invited to visit Denmark, inspect thoroughly the whole process of the manufacture of butter at certain dairy-farms, and analyse the product obtained under these crucial conditions. Of course, it would have been impossible to visit and watch the process of butter-making at even a tithe of the numerous dairy-farms existing in Denmark, but the investigation was simplified by the fact that Professor Stein, who practically occupies the position of Public Analyst for the whole of Denmark, receives, under the Danish Margarine Act, three samples of butter daily, three inspectors being constantly employed in visiting the farms and taking samples of the products. Hence, Professor Stein was able at once to point to certain farms whence very abnormal butter had been obtained, and consequently our attention was specially directed to the product from these places. Accordingly, early in last month I visited Copenhagen, and met in conference at the Government laboratory Professor V. Stein; Professor Th. R. Segelcke, Lecturer on Dairy-products in the Copenhagen College of Agriculture; Mr. B. Böggild, Consulting Dairy Chemist to the Royal Agricultural Society of Denmark; and my travelling companion, Mr. Harald Faber, Commissioner Abroad of the Agricultural Department of the Danish Government. The following day we (with the exception of Professor Segelcke, and with the addition of Mr. A. Stewart MacGregor, British Vice-Consul at Copenhagen) proceeded to a distant part of Denmark, where was situated one of the farms ("B") already referred to. There we witnessed the milking of thirty-one cows, the product from each of which was weighed separately. The mixed milk was then put into a receptacle for ripening, and the lid securely sealed and fastened by Mr. MacGregor and myself. The next morning, having noted the conditions of the seals, we opened the vessel, and watched the transferance of the contents to the churn. The usual addition of butter-colouring was omitted. The churning being completed, we watched the butter worked and placed in a number of tin boxes, which boxes were then put into strong linen envelopes, which were closed and sealed by Mr. MacGregor and myself. We then proceeded to another farm ("O"), where we saw sixty-five cows milked, and watched the manufacture of butter in the same jealous manner as before, the product being duly sealed as on the previous occasion. The milking of the cows, as well as the treatment of the milk and butter, was conducted at both farms under the unremitting supervision of the whole party, and especially of Mr. Macgregor and myself. I am, therefore, in a position to give a most positive guarantee that the samples of butter in question were absolutely genuine products.

The butter made at "B" farm was the produce from milk yielded by cows which had been in the stalls for six days, but which previously had been fed on grass land reclaimed from the sea. Most, but not all, of the cows were in calf, some being pretty far advanced, but others not expected to calve until May next. At "O" farm the cows had been longer in the stalls, but were all expected to calve within a few months. I may say here that I was greatly struck with the extremely methodical manner in which the operations of milking and butter-making, and in fact the whole work of the farms, were carried out. Everything was scrupulously clean, and the whole manufacture was conducted on the most scientific principles, a thermometer being employed at every stage of the operation, the milk and other products weighed, and a definite proportion of salt added. It is impossible to conceive a manufacturing operation of the kind carried out in a more perfect, systematic, and cleanly manner, and it certainly is a matter of regret that many of our English dairies are so far behind those in Denmark.

On returning to Copenhagen, Professor Stein and I opened one of each of the sealed packets and jointly made an analysis of the contents. The method on which we placed the most reliance was Wollny's modification of the Reichert Process for determining the volatile acids," a translation of which by Mr. O. Hehner was published in the ANALYST last winter. The process as there described (Vol. xiii. page 40) is strictly that used by Professor Stein, but I may say that he lays special stress on the importance of ensuring the entire absence of carbonic acid. The water used is hot or recently boiled, and, by means of a T-piece in the tube by which the flask is connected with the condenser, the alcohol can be distilled off, and water added to the residual soap without opening the flask and so exposing the contents to the air. Another important point to which great attention is directed is to heat the acidulated liquid very gradually, and not to attempt to commence the distillation until the layer of fatty acids has become clear and the aqueous liquid has almost entirely lost its original milky appearance. The time of thirty minutes for the distillation of 110 c.c. is strictly adhered to, with a variation of not more than two minutes more or less.

My previous experience of Wollny's method, which I have used habitually for the analysis of butters ever since the appearance of the translation in ANALYST a year ago, had shown me that the results were very satisfactory, and, if anything, somewhat higher than those I used to obtain by the Reichert process as worked by myself, and described in my "Commercial Organic Analysis" (Vol. II. p. 45). The higher result is probably due to two or three causes. Thus, I had been in the habit of distilling off 50 c.c. out of 75, which is a smaller proportion than 110 out of 140. This difference probably more than compensated for any error caused by the presence of carbonic acid in the old way of operating. As worked by me, I do not think I ever lost volatile acids in the form of butyric ether, as I always used a flask fitted with a proper condensing arrangement. Professor Stein's experience of a number of samples of butter of known purity shows that, in his hands, and worked in the exact manner I have described, the volume of decinormal alkali neutralised by the volatile acids from 5 grammes of butter-fat ranges from 31.59 c.c. to 25.08 c.c. with an average of 28 64. I should myself have fixed the average at 28.00, or a little below it, so that our experience is fairly concordant in that respect.

The experiments on the samples of butter-fat in question, which were marked respectively "B" and "O," gave in Professor Stein's laboratory the following results :— "B" required 22 70 c.c. which, with the correction of 0·17 shown by a blank experiment, gives 22.53 c.c. as the final result. The butter-fat from the sample marked "O" gave a result of 24.76 c.c.

4

I may say that I took the most rigid precautions to avoid the introduction of any possible source of error-such precautions as might well have been resented by a brother chemist, had he not been well aware that their object was to meet all possible criticisms, and not because there was any real doubt in my mind respecting the conditions of working. Thus, I proved the accuracy of the weights of the balance employed; I personally took every weighing and measurement; I proved the neutrality of the distilled water, alcohol, and phenol-phthalein employed; I verified the capacity of the burette, pipettes, and measuring flasks; and I set the baryta solution. This was done by titration, with a sample of quadroxalate of potassium obtained from Professor Wollny, of Kiel. I brought some of this reagent home with me, and have since verified its substantial, though not absolute, accuracy, by titration against solutions made up in England. I have also made for myself some of this acid oxalate of potassium which has the formula KHC,O,, H2C2O,+2 aqua, and seems to offer certain practical advantages over oxalic acid, as it is of definite composition, easily prepared and purified, and not efflorescent. I have prepared it by making a saturated solution of re-crystallised oxalic acid in water and filtering the liquid. One-fourth of this liquid was then heated to boiling, and neutralised with pure potassium carbonate, employing litmus paper as an indicator of neutrality. This neutralised solution was then added to the main quantity, whereupon a copious deposition took place of the quadroxalate in the form of a granular powder. This powder is filtered off, washed with a little cold water, and dried over sulphuric acid. In using the quadroxalate for setting the baryta solution, about 25 of a grm. is exactly weighed and dissolved in a little warm water. Phenol-phthalein is then added and the baryta water run in until the pink colour becomes permanent. During the first part of the titration the solution remains quite clear, but subsequently a white precipitate of oxalate of barium is produced, and this forms an extremely delicate background for viewing the end of the reaction. In fact, I think it would be a distinct advantage in such titrations to aim at the presence of a finely-divided white precipitate in the liquid, rather than to employ a white background to the vessel. Of course, baryta possesses a manifest advantage for this purpose over caustic soda, for a clear solution of baryta must be free from carbonic acid, the presence of which affects any titrations in which phenol-phthalein is employed as an indication.

The time at my disposal in Copenhagen did not allow of any further analytical examination on the spot of the butter-fats in question, with the exception of the determination of the specific gravity of the sample marked "B." This was ascertained by a specific gravity bottle of a kind I have been unable to obtain in England. It was a conical bottle, furnished with a thermometer having a range of a few degrees above and below 100° C. There was also a second orifice, consisting of a vertical capillary tube ending at the top in a small funnel. The bottle having been filled with the butter-fat and the thermometer stopper inserted, it was immersed in a beaker of water, the tem

perature of which did not rise more rapidly than 1° C. per minute, so that there was ample time for the contents of the bottle to acquire the temperature of the surrounding water. When an external thermometer shewed this to be at 100° F., the fat which had passed by expansion into the little funnel surmounting the capillary tube was removed by means of filter paper, so that the measure of the fat was accurately adjusted to a mark just at the top of the capillary tube. The bottle was then removed and weighed with its contents, after which it was returned to the water and the heating continued, until the internal thermometer indicated that the fat had reached a temperature of 100° C. It was then removed and weighed as before. We had previously observed the weight of the bottle when dry and empty and when filled with water at 15° C. Hence the experiment gave us the means of ascertaining the density of the butter not only by Bell's method at 100° F., but also at 100° C.

The following were the figures obtained, water at 15° C. being unity:

Specific gravity of "B" butter-fat at 100° F. (=37.8° C.) ·9055.

This gives an apparent expansion coefficient of 0.00067 for 1° C.

In concluding the description of the experiments made at Copenhagen, I cannot omit to give expression to my sense of the great courtesy shown me by Professor Stein and his assistants, and my high appreciation of the methodical manner in which the work there is evidently done, and which is in itself so good a guarantee of its accuracy.

It must be remembered that in order to ripen the milk in time to allow of its being churned in twelve or sixteen hours after milking, it is necessary to add to it a certain proportion of buttermilk. Hence my critical examination of the butters in question would not have been complete had we omitted to analyse the butter-milk added, with a view of proving that it was not a medium of introduction of any foreign fat. The examination was made by rendering 100 c.c. of the butter-milk alkaline with ammonia and agitating with ether. The ether was allowed to separate and the agitation repeated with a fresh quantity of ether. The ethereal solutions were mixed and evaporated, the residual butter-fat being weighed. It amounted to barely 0.1 per cent. of the butter-milk, and hence even had the fat so extracted consisted wholly of foreign fat, it could not have appreciably affected the results of the analysis. As a matter of fact, however, we made an observation of the index of refraction of the extracted fat by means of Abbé's Refractometer, and found it to be 1.4620-a result which agrees with that of genuine butter-fat. I may here say that this is my first experience of the use of Abbé's Refractometer, which Professor Stein informs me is capable of distinguishing with certainty butter from other fats, but is not of much service for mixtures containing butter.

I brought away with me some of the sealed packets containing samples of the butter in question, while others were left with Professor Stein and Mr. Faber. I also brought a portion of the butter-fat rendered from the butters in question in Professor Stein's laboratory. Since my return to England I have analysed these samples of butter-fat as well as the butters brought home in their original condition, and have not limited myself merely to the Reichert-Wollny process, but have examined them by other well-known processes.

Before proceeding to the discussion of the other results, some explanation is required respecting the above determinations of specific gravity. The figures quoted are calculated to a temperature of 100° C., though actually observed at a temperature of 99.5° C. The coefficient of expansion of butter-fat according to the experiment already recorded being 007 for 1° C., it follows that the gravity of, say "N" sample at 99.5° was 86535, and similarly the other observations were ⚫00035 higher than the figures recorded in the tables.

In my work on "Commercial Organic Analysis," Vol. II., I state the average specific gravity of butter-fat as being 0-868 at 99°, equivalent to 8673 at 100°. This determination, like nearly all the other densities of fats at 100° C. published by me, was obtained by the means of a Westphal balance. The accuracy of this had been ascertained by comparing the gravities of proof-spirit and milk shown by it with determination of the gravities of the same samples made by the specific gravity bottle, and the accuracy of the rider weights was also verified-a precaution which was by no means