J THE CHEMICAL VOLUME LXIV. NEWS. EDITED BY WILLIAM CROOKES, F.R.S., &c. No. 1649.-JULY 3, 1891. NOTE ON SOME OBJECTIONS TO THE WORK By SPENCER U. PICKERING, F.R.S. THE importance of the discussion which took place at the Chemical Society on June 4th (CHEMICAL NEWS, Ixiii., 305), on the subject of the hydrate theory of solution, and on my method of examining series of experiments, must be my excuse for answering some of the objections which were raised, and which my already published answer does not adequately meet. It is scarcely necessary, perhaps, to point out that Prof. Armstrong is quite in error when he interprets the 102 breaks in the figures representing various properties of sulphuric acid solutions as meaning the existence of 102 different hydrates, for the same hydrate will account for as many breaks as there were properties examined. The actual number of hydrates of which I found indications in the case in question was only twenty, and I must again protest against the unscientific nature of any objection to their existence on the grounds that their number is greater than we might expect; for we have no right to expect anything at all, being at present entirely in ignorance of the nature or number of those loosely combined and partially dissociated compounds which can exist in the liquid condition. It is certainly surprising to find Prof. Armstrong producing as a novel discovery, and as the result of his own experience, the conclusion that "changes of composition" (? constitution), "whenever they occurred, were indicated on the curves drawn directly to represent the experimental results"; for this is the very point on which I have insisted all along. Before I had published any of my experimental results, I said (Nature, 1889, p. 343), "Differentiation will sometimes bring about the recognition of breaks which might be overlooked in the original curve, for, though the differential curve can show no breaks which do not exist in the original curve, it may often, as a consequence of its very nature" (i. e., being more nearly rectilinear than the original curve), "show breaks clearly which would be recognised only with difficulty in the original curve"; while at the commencement of my first paper on Sulphuric Acid (Chem. Soc. Trans., 1890, p. 68) I said, “I would, with the experience I have gained, generally trust as much to the drawing of the original curve as to differentiation, for showing whether that curve is continuous or not, and where the changes occur in it." In my second paper on Sulphuric Acid I used differentiation to a limited extent only, and as affording supplementary evidence to that derived from a study of the experimental curves (Chem. Soc. Trans., 1890, p. 345); while in all the work of this nature which I have done during the last eighteen months I have not used differentiation once (see B. A. Report, 1890, 312, and CHEM. NEWS, Ixii., 185), although there are undoubtedly some advantages to be obtained through its proper use. Thus Professor Armstrong's experience on this point agrees entirely with my own; so also does Professor Rücker's, though I think that there must be some mistake about his statement that it was a study of my density results which led him to this conclusion, for the original density curve cannot, as a matter of fact, be plotted out at all on any workable scale. The fact that, in my paper on Sulphuric Acid, I gave a list of the scales which I had found most suitable for plotting out my results seems to have led Prof. Rücker to conclude that the position of the breaks may be made to vary with the scale used. This is certainly not the case (although of course we can select a scale so badly proportioned that it would be utterly useless). I mentioned in my paper (p. 68) that "Drawings on several different scales, and with several different points as origins, were made in all cases, and the labour entailed in the treatment of the results has by far exceeded that of the determinations themselves." Prof. Rücker's conclusion, that any break which is recognised only in the differential figure deduced from the experimental figure, and not in the experimental figure itself, is "of the most untrustworthy character," is precisely my own conclusion. In the two or three instances in which such indications were obtained I regarded them merely as suggestions (see p. 77 of my paper), and in no single instance did I accept them alone as proving the existence of a hydrate. Prof. Rücker says that the fact that a series of results may be represented more closely by two curves than by one, may not imply a discontinuity, because, by admitting a sufficient number of discontinuities, the results may be represented graphically with absolute fidelity. I certainly fail to see the force of this argument. A drawing which represents experimental results as being entirely free from experimental error is obviously incorrect, and the mere fact that it is physically possible to make an illegi timate drawing of this sort does not throw any discredit on other drawings which are in harmony with the ascertained magnitude of the experimental error. The problem of the correct depiction of results graphically is not a very difficult one; we must adopt the simplest drawing which is consistent with the known experimental error, and discard any drawing which attributes an error to the experiments, either greatly in excess or greatly below that of the observed or probable experimental error. As an instance I may take the diagram which was exhibited at the meeting of the Chemical Society, and which will be found in the July number of the Phil. Mag. (xxxii., 94), in my as the apparent error of the points. Drawn as 5 or 4 curves, this apparent error would be too far below the experimental error; the 3 curve drawing is legitimate as far as the magnitude of the error is concerned, but, as we can simplify the drawing still further without any increase -indeed with a slight decrease-of the apparent error, we are bound to make this simplification, and to accept the 2 curve drawing in preference to the 3 curve drawing. But no further simplification is possible; for simplifying it one degree further produces a sudden tenfold increase in the apparent error, and makes this ten times greater than the ascertained experimental error. Thus the 2 curve drawing is the simplest, and the only possible representation. Dr. Morley's objection is, I think, a mistaken one. A break at a particular hydrate by no means implies that this particular hydrate has only just begun to form there. Supposing there to be three hydrates: CaCl2 with 7, 8, and 9 H2O respectively; then, as water is added to CaCl27H2O, its effect will be to form some of the CaCl28H2O, and, as more water is added, the relative proportion of CaCl28H2O present will be increased: as soon, however, as more water is added than is necessary to entirely form the CaCl28H2O, another hydrate, CaCl29H2O, will begin to make its appearance, and the effect of adding more water now will be to increase the relative proportion of the latter and diminish the relative proportion of the CaCl28H2O present: thus the effect of adding water to solutions stronger than CaCl28H2O will. will be different to that produced by adding it to solutions weaker than CaCl28H2O; in other words, there will be a change in the rate of the effect at the composition of the CaCl28H2O itself: this is just what is actually observed It is generally acknowledged that the Davy safety-lamp cannot with certainty detect less than 3 per cent of firedamp in the air of the mine. Gas-indicators of much greater sensitiveness have been invented; amongst these the electrical apparatus of Liveing and the spirit safety. lamp of Pieler take first rank. The objection to these special forms is, however, a serious one. They do not serve for illuminating purposes, and therefore it becomes necessary to carry an ordinary safety-lamp together with the testing apparatus. Many attempts have been made to obviate this inconvenience by producing a safety-lamp which shall serve the double purpose of illumination and of detecting minute percentages of firedamp. The discovery of such a lamp would be of great value to the miner, in view of the fact that very low percentages of firedamp have been proved to be dangerous in the presence of coal-dust. The following apparatus has been devised to render easy the process of testing the sensitiveness of different forms of safety-lamps when used for detecting firedamp. * A Paper read before the Royal Society, June 18, 1891. To enable satisfactory tests to be made in the laboratory, it was necessary to insure (1) the easy and rapid production of mixtures of fired amp and air in known proportions; (2) to insure economy of the artificially prepared methane, which represented tiredamp; and (3) to examine the flame of the lamp under conditions as satisfactory as those existing in the nine. A wooden cubical box of about 100 litres capacity, was constructed so as to be as nearly gas-tight as possible. It was then made absolutely gas-tight by painting it over with melted paraffin wax, which was afterwards caused to penetrate more perfectly by passing an ordinary hot flat-iron over the surface. This testing chamber was furnished with a small inlet tube at the top, and with a similar outlet tube below. It had a plate glass window in front for observing the lamp in the interior, and a flanged opening below for introducing the safety-lamp. This opening was closed by a water-seal consisting of a small zinc tray supported by buttons, and containing about two inches depth of water, into which the flange dipped. A mixer was arranged, which consisted of a light flat board, nearly equal in dimensions to the section of the chamber, and suspended by an axis from the upper corner of the chamber. The mixer was moved rapidly ibackwards and forwards from the side to the top of the Interior of the chamber, by grasping a handle projecting through the front of the chamber. When a mixture of air with a certain definite percenand purified by ordinary chemical methods, was intro. tage of firedamp was required, the methane, prepared duced into the chamber in the requisite quantity by the top inlet. It displaced an equal volume of air which escaped through the lower outlet, the exit end of which surface. was sealed by being immersed just beneath a water mixture of gas and air throughout the interior of the A vigorous use of the mixer secured a uniform chamber in the course of a few seconds. The lamp was then introduced into the chamber, and placed in position behind the glass window. The simplicity of arrangement of the water-seal rendered the necessary opening of the chamber very brief, and the introduction and removal of the lamp many times in succession was not found to produce any appreciable effect upon the composition of dimensions of the "cap" over the flame were noted as the atmosphere inside the chamber. The appearance and soon as the cap underwent no further change. A lamp was left burning in the chamber for a considerable length of time, and its indications underwent no change owing to the large capacity of the chamber and the very limited amount of air required to support the combustion of the small flame always used in gas-testing. The whole interior of sthe chamber and mixer were painted dead black, so a to render visible pale and small caps against a black ground. The methane was introduced from an ordinary gasholder. A volume of water, equal to that of the methane to be displaced, was poured into the top of the gasholder. The gas tap of the holder was then momentarily opened so as to produce equilibrium of pressure between the methane and the atmosphere. The gas tap having then been placed in connection with the upper inlet of the chamber, the water tap was opened and the measured volume of water was allowed to flow down and drive air ceased to appear through the water at the outlet, the the methane into the chamber. As soon as bubbles of chamber was closed; the mixer was then vigorously worked for a few seconds, and the mixture of gas and air was ready for the introduction of the lamp. Before introducing the methane for a fresh mixture, the atmosphere of the chamber was replaced by fresh air by removing the water-tray from beneath the opening at the bottom of the chamber, and blowing in a powerful stream of air from a bellows to the top of the chamber. The chamber was supported on legs, which were arranged so as to place it at a convenient height for observations through the window, and also for the introduction and removal of the safety-lamp. The accuracy of this method was tested by introducing the Pieler lamp into the chamber, which was charged successively with a series of mixtures containing proportions of methane varying from 0.5 to 4 per cent. The height and appearance of the cap over the flame absolutely corresponded with a series of standard tests already published and made by a different method in which firedamp was used instead of methane. The observations were usually made in a darkened room, but the flame caps were easily seen in a lighted room, provided direct light falling on the eye or chamber was avoided. The capacity of the chamber was 95,220 c.c.; accordingly the following volumes of methane were introduced : -for per cent mixture 476 c.c., for I per cent 952 c.C., for 2 per cent 1904 c.c., for 3 per cent 2856 c.c., for 4 per cent 3808 c.c., and for 5 per cent 4760 c.c. It will be seen that a series of tests, in which the above mentioned percentage mixtures were employed, involves an expenditure of only 15 litres of methane, a quantity far smaller than that required by any other method of testing as yet described. PURIFICATION OF CHEMICALS BY EXPOSURE TO INTENSE COLD. By H. THOMS. THE author was informed by Pictet, who has taken out a patent for the process, that from the purest chloroform of commerce, if cooled down to about -70°, there separated out a crystalline body, which is then drained from the part which remains liquid. This liquid portion is then exposed to a temperature below - 100°, when the chloroform itself crystallises out, and can be separated from an impurity which remains liquid. Chloroform purified in this manner is a colourless liquid, having at 15° the sp. gr. 151. It is in course of an investigation by O. Liebrich. It is indefinitely permanent on exposure to light. Whilst the so-called pure chloroform of commerce takes a greenish colour if shaken up with solution of potassium dichromate and sulphuric acid, if Pictet's chloroform is similarly treated the chromic mixture retains its yellowish colour. - Chemiker Zeitung and Pharm. Central-Halle. THE FAT IN MILK. By FRANK T. SHUTT, M.A., F.I.C., F.C.S. Of many forms of safety-lamp tested in the above apparatus, the one which most satisfactorily fulfilled the THE BABCOCK METHOD FOR ESTIMATING two purposes of efficient illumination and delicacy in gas testing was Ashworth's improved Hepplewhite-Gray lamp. This lamp is of special construction, burns benzoline from a sponge reservoir, and its flame is surrounded with a glass cylinder, which is ground rough at the hinder part; this latter device prevents the numerous reflected images of the flame, and the generally diffused reflections which are seen from a smooth glass surface, and which render the observation of a small pale flame cap very difficult, if not impossible. The wick of this lamp, when at a normal height, furnishes a flame of great illuminating power. When lowered by a fine screw adjustment the flame becomes blue and non-luminous, and does not interfere therefore with the easy observation of a pale cap. The following heights of flame cap were observed, which fully bear out the unusual sensitiveness of this flame. With o5 per cent of methane 7 m.m.; with 1 per cent 10 m.m.; with 2 per cent 14 m.m.; with 3 per cent 20 m.m.; with 4 per cent 25 m.m.; and with 5 per cent 30 m.m. The cap, which with the lower proportions was somewhat illdefined, became remarkably sharp and definite when 3 per cent and upwards of methane was present. But even the lowest percentage gave a cap easily seen by an inexperienced observer. It appears from the above record of tests that the problem of producing a lamp which shall serve both for efficient illuminating and for delicate gas testing purposes has been solved. The solution is in some measure due to the substitution of benzoline for oil, since the flame of an oil flame cannot be altogether deprived of its yellow luminous tip without serious risk of total extinction, and this faint luminosity is sufficient to prevent pale caps from being seen. From further experiments made in the above testing chamber with flames produced by alcohol and by hydrogen it was found to be true in practice, as might be inferred from theory, that if the flame was pale and practically non-luminous, the size and definition of the flame cap was augmented by increasing either the size or the tempera. ture of the flame. It is quite possible by attending to these conditions to obtain a flame which, although it is very sensitive for low percentages of gas, becomes unsuitable for the measurement of any proportion of gas exceeding 3 per cent. This must, for the general purposes of the miner, be looked upon as a defect; but it is not a fault of the lamp already referred to. It is of interest to note that with the Pieler spirit lamp a flame cap an inch in height was seen in air containing only o'5 per cent of methane. PAYMENT of money according to value received is a sound principle recognised and acted upon as far as possible in every branch of commerce. With a view towards its application in the buying and selling of milk, it is necessary to know at the outset the elements in the composition of milk that determine the value of the latter for use in creameries and cheese factories, and, secondly, to apply some method whereby those valuable elements may be easily and cheaply determined. The true commercial value of a milk depends upon its solid constituents. The variation in the total amount and composition of these constituents that exists between samples of genuine milk (due to food, breed, and individual characteristics), and the extreme ease with which water may be added to or cream abstracted from pure milk without detection by the unaided senses, make it highly desirable that there should be some rapid and accurate means, capable of being used by intelligent labour, for ascertaining within small limits the amount of such solids. The butter fat is so much the more important and valuable constituent in the solids that the knowledge of its percentage in a milk not only protects the buyer from fraud, but at the same time renders it possible for the seller to be paid according to the quality as well as the quantity of his milk-a plan that will commend itself as an equitable and satisfactory one to both parties. The Babcock Method. The usual methods of chemical analysis, while giving exceedingly accurate results, are lengthy, and must be conducted by experts in properly equipped laboratories. Numerous processes, however, have been devised of late by which the fat may be determined with a greater or less degree of accuracy in the creamery or factory. Such processes may be worked by the man in charge, and require but little time. Prominent among these methods is the Babcock Test, which has for its principle the holding in solution of the solids other than fat by means of sulphuric acid, the fat at the same time separating as an oily layer. By the addition of water and the aid of centrifugal motion, the latter is brought into the graduated neck of the vessel in which the test is made and its amount at once noted. It is the object of the present report to bring before the farmers and dairymen of the Dominion our experience in the laboratories of the experimental farm regarding the NEWS reliability and accuracy of this process for examining | knowledge of which will assist in obtaining satisfactory milk. Examination as to Reliability and Accuracy. Thirty-two samples of milk have been examined in duplicate by (1) the Babcock test, and (2) by gravimetric analysis, the results by the latter being taken, for the sake of comparison, as correct. These operations gave us 128 independent fat determinations. As the results obtained throughout the whole series are uniformly close, it will suffice to tabulate here a few of them as examples : Milks: "Morning," "Evening," and "Mixed." results. 1. The milk should be well sampled. If the sample has stood several hours it should be thoroughly, though not violently, shaken before the pipette full is taken out. By this means any cream that has risen is once more uniformly suspended throughout the milk. 2. When a number of tests are to be made the milk of all the samples should be first carefully measured out into the bottles, the acid being added afterwards. When this is completed the two should be thoroughly mixed by shaking the bottle in the hand with a rotatory motion. The bottle should then be placed in the machine, which, when all the samples are placed, is immediately set in motion as directed. The thorough mixture of the acid and the milk tends to prevent the formation of curd or scum, which would interfere with the correct reading of the percentage of fat later on. If the bottles are whirled while still hot from the action of the acid, the fat separates more easily and quickly. 3. When the water is added, care should be taken that it is hot. No error can be made by having it too hot. 4. Hot water should also be placed in the machine so that the temperature of the mixture be not allowed to fall. 5. The amount of fat should be noted immediately at the close of the operation, and before the fat has had time to contract through loss of heat. 137 An inspection of this table will reveal (1) the limits within which the amount of fat will vary when the Babcock test is made in duplicate, and (2) the approximation of such results to those given by carefully conducted chemical analyses. Of the thirty-two samples tested in duplicate by the Babcock method, only two gave a difference between their duplicates amounting to three-tenths (0.3) of 1 per cent; two varied in their duplicates two-tenths (0.2) of I per cent; fourteen differed to the extent of one-tenth (01) of 1 per cent; and thirteen gave results identically the same.* The greatest difference between fat determinations by the Babcock test and gravimetric analysis on the same milk was (0.25) a quarter of 1 per cent. This occurs in three instances only. Where the results are not identical the variation is usually between one-tenth and two-tenths of 1 per cent. Conclusions. From these data therefore we may safely conclude that when the Babcock test is made according to the instruc. tion given with the machine, strictly reliable results are obtained, and that the percentage of fat so found, allowing for the greatest error possible under such circumstances, will be well within one-quarter of 1 per cent (0.25) of the amount of fat actually contained in the milk. Recommendations. Full instructions for operating the Babcock tester accompany each machine. Our own experience, however, leads me here to emphasize one or two particulars, the * Since the above was done the Babcock tester has been in constant use for examining the milks produced at the farm. The samples have been tested in duplicate, and the results so obtained show no greater variation than those recorded above. ON THE AMMONIA-SODA PROCESS. By H. SCHREIB. In the year 1876 appeared a report on a proposal of Mr. Thorwald Schmidt respecting the ammonia-soda process. Mr. Schmidt proposes to decompose the liquor of the ammonia-soda process containing chlorides of sodium and calcium, by a solution of the ashes of seaweed. The sulphates of potash, soda, and magnesia contained in the ash of the seaweed are decomposed so far that hydrated sulphate of lime and hydrated magnesia are precipitated in a form which may be available for paper making, as "pearl hardening." By further adding certain admixtures to the said liquor, at last a pure solution of chloride of sodium is obtained, while, at the same time, iodine, nitrate, and icdides are produced. The latter solution is then treated again by the ordinary ammonia-soda process for the production of bicarbonate of soda and white alkali. The proposal of Mr. Thorwald Schmidt cannot be generally used, as the ash of seaweed is not to be had everywhere, not to count for the somewhat circumstantial way of proceeding. But the method is based upon a sound idea, viz., the permanent regeneration of the liquor by simultaneously producing useful materials. I have become induced, by the article of Mr. Schmidt, to introduce to your readers another proceeding practically confirmed for a number of years. My way of treating the ammonia process is, as far as I know, the only one which has been applied hitherto to practically realise the products out of the said liquor, though I must, however, remark that I have been favoured by special opportunities, and that my idea of treating may scarcely be generally adopted. A very large German manufactory, making an indispensable article for household consumption, possesses as auxiliary annex a small factory, producing per annum about 1000 tons of ammonia soda required for the main works. The establishment in question requires, as well, large quantities of cardboard for packing purposes. This latter material, being of a cheap and grey-coloured quality, is also prepared by the same works in quantities of about 10 tons daily. Owing to its unattractive appearance the cardboard is |