quantity of arsenic, and the lady had suffered from an obstinate skin eruption and other arsenical symptoms. Another lady who had bought some coloured "Indian" muslin at a large establishment in the West End, and worked at it with her maid, had noticed that both of them were beginning to suffer from inflammation of the eyelids. The muslin was brought to him for analysis, and he found large quantities of arsenic. The vendor had said that he would eat all the arsenic that could be found in it, and he (Mr. Cassal) was sorry he did not do so. The sulphide was at his service. In another case five or six children were being kept in a room with a large fire in it, the walls being covered with a green arsenical paper. The symptoms had been mistaken for those of a violent common cold. Dr. MUTER said he had had some similar Indian muslin sent to him, and he also found arsenic in it. Dr. HARVEY stated that he frequently met with arsenical wall papers. Only on the previous day a case was brought to his notice of serious illness traced to a drab-coloured bedroom paper, which upon analysis was found to contain a considerable amount of arsenic. He was acquainted with numerous other cases where injury appeared to have resulted from the use of such papers. Mr. ADAMS said that more than twenty times he had had patients suffering from arsenical poisoning. It was a common result that the eyelids became affected; they had a peculiar red appearance which a practical oculist at once knew to be due to arsenic. He had a case quite lately showing what a very minute amount would do the mischief. Of course, in the case of the candles it must have been a very small quantity that each person could have breathed. The poisonous effects of arsenic depended in a measure on two things-first, the idiosyncracies of the person (some could take it with impunity whilst others are injuriously affected by the smallest amount), and secondly, if it gets into the system by the air cells of the lungs it is absorbed much more quickly and injuriously. A lady patient of his had some fur which he had examined and found to contain arsenic. She wrote to him about it on the previous day as follows:-"I am very pleased to tell you anything I can about the fur you found contained arsenic. Directly I began to wear it I had every symptom of a very bad cold in the head, and later on this was accompanied by sore throat and diarrhoea. Naturally in hot rooms I always felt worse." The lady put the fur on one side and at once recovered. When she resumed wearing it, all the symptoms recommenced. This experience was repeated again and again till she came to associate the symptoms with the wearing of the fur, which was then sent to him, and he found out the cause. THE METHODS FOR DETERMINING FAT IN MILK. BY DR. P. VIETH, F.C.S., F.I.C. (Read at the Meeting, March, 1889.) On looking over the chemical literature of the last decade, we find that much attention has, during that period, been paid to the determination of fat in milk. So numerous are the methods brought out, and the alterations and improvements suggested, that it seems to me, that the supply far exceeds the demand, and that it becomes necessary to classify the various methods in divisions and subdivisions, when attempting to briefly review them in anything like a lucid manner. I will with your permission try to do so in the remarks which I am going to make. I. Methods which can be worked outside the chemical laboratory, and by persons having no analytical training. 1. Processes confined to the dairy. (a) Control apparatus worked in connection with the Danish Centrifugal Cream Separator. No chemicals required. The milk samples are simply subjected to centrifugal force in suitable vessels, and the volume of cream read off. This volume indicates the percentage of fat sufficiently near for all practical purposes. (b) Lactocrite, worked in connection with Swedish Cream Separator. The milk is treated with strong acids, whereby the fat is, so to speak, set free, and under the influence of centrifugal force unites, so that its volume can be determined. It is generally admitted that results thus obtained agree very closely indeed with those arrived at by determining the fat gravimetrically. At the same time it is rather curious that some experimenters, e.g., W. Blyth, compare it with Adams' method, while others-among them, Soxhlet compare it with the latter's method; both parties find agreeing results, and still paper and plaster method do not give results agreeing among themselves. It should be mentioned that a smaller machine has been constructed, which can be used in the laboratory of the analyst, although, I am afraid, the price will not facilitate its introduction. 2. Optical Methods.-They are based upon the assumption that the opacity of milk stands in direct relation to the amount of fat the milk contains. Not only is this supposition wrong, but the test is also influenced by a number of uncontrollable circumstances. The results are, therefore, always doubtful, frequently quite fallacious. Feser's Lactoscope is, relatively speaking, the best of all the optical tests. 3. Soxhlet's Areometric Process.-Milk is mixed with potash solution, and then shaken with ether. The ethereal fat solution is allowed to separate, and its specific gravity determined. A table gives the corresponding percentage of fat. The results may be considered to be identical with those arrived at by extracting milk dried up on plaster of Paris. 4. Cronander and Liebmann shake a mixture of milk and potash solution with ether. After complete separation they allow the ether to evaporate and measure the volume of the fat left behind. I do not think these methods recommend themselves, and cannot find that they have been taken up. 5 Marchand's Lactobutyrometer.-Milk is shaken in a gauged tube with ether to dissolve the fat; shaking is continued after the addition of alcohol, and the tube then placed in warm water. The mixture separates into three layers; the volume of the upper one, containing the fat, is read off. Every one-tenth c.c. indicates 2 per cent. of fat, in addition to 1.2 per cent. which is kept in solution. The results are fairly correct in the majority of cases, when dealing with milk which is not too rich. 6. Short heats milk with strong alkali in a boiling water-bath for two hours, decomposes the soap by the addition of a mixture of sulphuric and acetic acid, allows the liberated fatty acids to rise, and determines their volume, which is then enlarged in the proportion of 87 to 100. This method seems to be rather troublesome, and open to grave objections. The results are said to be lower than those obtained by Adams' method. II. Methods which must be considered as confined to the chemical laboratory, and the hands of the analyst. 7. Extracting the fat from the residue left when milk is evaporated without the addition of an inert substance. (a) No care is taken to disintegrate the residue (Wanklyn). (b) Pains are taken that the residue should be in a readily exhaustible condition (Dr. Jas. Bell, Carter Bell). The difficulty of completely exhausting such milk residues, more especially if the milk is poor in fat, is now generally recognised and the methods may be looked at as abandoned. 8. Muter proposes to precipitate the casein which carries down the fat with it; to wash the precipitate first with water, then with alcohol, and at last with ether, which is collected and evaporated, when the fat is left behind. 9. Milk is dried up on some inert substance and then extracted. The substances suggested are glass powder, sea sand, plaster of Paris, asbestos, pumice stone, filter and blotting paper, wood fibre, sponge; and these various modifications are connected with the names of Abraham, Adams, Babcock, Gantter, Johnstone, Macfarlane, Soxhlet, Storch, and others. This group certainly contains the most exact methods for the determination of fat in milk. 10. Morse and others suggest to dry milk on anhydrous sulphate of copper, extract the fat by means of light petroleum, saponify, and determine the quantity of alkali required for saponification. I find only one account of experimenting with this method, in which it is said that the results fall below those arrived at by using Adams' method. 11. Roese shakes milk first with alkali, then with a mixture of ether and light petroleum, reads off the volume of the upper layer after complete separation, and determines in an aliquot part of it the fat gravimetrically. An allowance is made for fat retained in the aqueous layer. 12. W. Schmid first heats milk with strong hydrochloric acid, after cooling shakes with ether, notes the volume of ethereal layer, and ascertains the fat in an aliquot part. Determining the fat in an aliquot part of its solution appears to me very undesirable, considering that one has to deal with ether, which is not only highly volatile, but the volume of which is also much influenced by temperature; besides, experimental errors are much increased. This is a rather long list of methods for the determination of fat in milk, and still, I am convinced, it is very incomplete, containing only those methods which have come under my notice and impressed themselves on my memory. Of some of the methods enumerated there exist modifications with regard to the chemicals and apparatus employed. Very generally acknowledged as standard methods are Adams' paper and Soxhlet's plaster processes; the indications of other methods which do not give direct results are compared with and gauged by either of these processes. Methods like those of Roese and Schmidt can claim no other advantage but speed combined with a near approach to correctness. Speed, no doubt, is a valuable item, and approaching correctly results may suffice under certain circumstances, but certainly not with regard to the official work of a Fublic Analyst. He ought to employ the most exact methods which are at his disposal. As an apology for the deficiency of new matter in the foregoing remarks I may say, that they were originally not intended to be read as a separate paper, but meant to be thrown into a discussion which, however, had to be abandoned for want of time. DISCUSSION. MR. DYER Suggested that calcined gypsum might give erroneous results by causing slight saponification of the fat, for gypsum usually contained some carbonate of lime which became caustic on calcination, and might make a lime soap. DR. MUTER noticed that Dr. Vieth had referred to the method of coagulating the milk with acetic acid, collecting the curd, washing the same first with water and then with spirit, and finally percolating with ether. This was in certain cases a very good way of working as he had had occasion to before mention to the society. MR. HEHNER said it would be very interesting to get a full catalogue of the different methods. As Dr. Vieth had gone so far perhaps he would go farther and complete it. DR. VOELCKER corroborated Mr. Dyer, and said that only last week he had such a sample of gypsum sent to him. Mr. Allen said that a good many years ago the late Mr. W. W. Stoddart, of Bristol, exhibited what was known as Horsley's process and tube, at a meeting of the British Pharmaceutical Conference, and also described a modified method by which the test was said to be applicable to butter. He, Mr. Allen, was very much interested in the description given, and consequently read a paper on the same subject at the following meeting of the conference, in which he expressed a very unfavourable opinion respecting the method. As a matter of fact, the process did not originate with Horsley, but had been described by Marchand many years previously. The whole subject had gone so completely out of his mind, that when a question was asked him in the "English Mechanic," as to the nature of the lacto-butgrometer, which was simply a Marchand's or Horsley's tube, he said he had never heard of the instrument. Dr. Vieth thereupon pointed out that the instrument had been advertised on the front page of the ANALYST, for a whole year previously, which he was afraid showed that advertisements were apt to be wasted on some people. DR. VIETH in reply said, that he considered plaster of Paris when employed in milk analysis, an inert substance inasmuch as its desired action was a purely mechanical one. Plaster of Paris was alkaline, but they must remember that milk was not neutral, but displayed besides alkaline, also acid reaction. A number of experiments had shown him that the acidity of milk, was in excess of the alkalinity of the plaster of Paris which he used. He had no difficulty in getting good plaster from Hopkin and Williams. With regard to the method of extracting the caseine, he had followed that himself in a great many instances, extracting the precipitate after it had been dried in the air bath; he found the results agreed very well indeed with the plaster extraction. The Marchand's process gave very fair results in the majority of cases, of course, the results were not exact enough for the work of public analysts; but, for instance, in the hand of a farmer who wanted to pick out his best and worst cows, it gave very good results indeed. He had made thousands of determinations; when the fat was within the limits of 3 per cent. and 3.5 per cent. the results agreed very closely. When dealing with richer milks, the results were not so reliable. He would never recommend the process for the use of public analysts, especially at the present day, when the amount of fat could be calculated with great accuracy, from the figures for specific gravity and total solids. (Conclusion of the Society's Proceedings.) ON SOURCES OF ERROR IN DETERMINATION OF NITROGEN BY SODA LIME, AND MEANS FOR AVOIDING THEM. By W. O. ATWATER. (Continued from page 76.) Effects of long heating and of open space in tube upon amounts of nitrogen obtained as ammonia from casein containing 12.43 per cent. of nitrogen. When the tubes were closely packed so as to insure the maximum of contact between the gases produced and the heated soda-lime (water-vapour at high temperature), and the combustion was conducted at a moderate heat and kept within the usual time of about three quarters of an hour, the full amount of nitrogen, 12:43 per cent. of the water-free casein, was obtained as ammonia. But when the combustion proceeded very slowly so as to occupy two and a half hours, the other conditions remaining the same, only from 12.27 to 12.08 per cent. of nitrogen were obtained as ammonia, making a loss of from 0.16 to 0.47 per cent. reckoned on the weight of the water-free casein, or from 1.3 to 3.8 per cent. of the total nitrogen. It seems reasonable to ascribe this loss to dissociation during the long time that the ammonia was exposed to the heat, though possibly some of it may have been due to incomplete ammonification of nitrogenous decomposition products. It is easy to understand how the ammonia passing rapidly through the interstices between the particles of soda-lime in the closely packed tube, in the ordinary analyses, would be so mixed with water-vapour and other gases that the nitrogenous distillation products would be completely changed to ammonia, and in the very brief exposure to heat the resultant dissociation would be too small to be noticeable. On the other hand, when the operation extends through two and a half hours instead of three quarters of an hour, it would seem by no means improbable that the anterior layer of soda-lime might, before the end of this long period during which it is kept hot, cease to give off any considerable amount of water-vapour. But it is probably this water-vapour which yields the hydrogen to form ammonia with the nitrogen of the volatile distillation products, and it is not impossible that the same vapour may also tend by its presence to prevent dissociation of ammonia. Add the fact that, when the operation goes on slowly, the time of sojourn of the ammonia in the heated tube is of |