Isonitrosoacetone semicarbazone,

CH3.C(=N.NH.CO.NH2).CH=NOH,

was prepared, and gave white crystals melting at 219220° C. It gave Liebermann's nitroso-reaction, and was insoluble in concentrated sodium hydroxide and soluble in dilute sodium hydroxide. When heated on a water-bath with acetic anhydride acetylisonitrosoacetone_semicarbazone was obtained, which, when re-crystallised from glacial acetic acid, gave cubical crystals melting at 186° C. It dissolved in dilute sodium hydroxide, and when reprecipitated with dilute sulphuric acid, needle-shaped crystals melting at 218-219° C. separated. Hence it was reconverted into the original semicarbazone by the splitting off of the acetyl group. When isonitrosoacetone semicarbazone was treated with nitrogen peroxide at a low temperature, the pseudonitrole of isonitrosoacetone semicarbazone, CH3C(=N.NH.CO.NH2).CHNO. NO2, was formed. It melted at 163-164° C. with sudden decomposition, and gave Liebermann's nitroso-reaction. It was soluble in dilute sodium hydroxide, giving a yellow solution which was decomposed by dilute sulphuric acid with evolution of nitrogen and oxides of nitrogen.

When isonitrosoacetone, CH3CO.CH NOH, in absolute ether solution was acted on by nitrogen peroxide at a low temperature, and the residue left on evaporation of the ether heated with benzene, red fumes were evolved and diacetylglyoxime hyperoxide,

CH3.CO.CNO

CH3CO.C

FURTHER INVESTIGATION ON THE DETECTION AND APPROXIMATE ESTIMATION OF MINUTE QUANTITIES OF ARSENIC IN MALT, BEER, AND FOOD STUFFS.*

By WILLIAM THOMSON, F.R.S.E., F.I.C.

SINCE my paper on this subject, read before the Manchester Section of the Society of Chemical Industry on May 2nd, 1902, and published in the British Food Journal, 1902, vol. iv., Nos. 44 and 45, and in the Medical Chronicle for October, 1902 (also in CHEM. NEWS, vol. lxxxvi., p. 179), I have continued the investigation with the view of improving on the process there described for detecting and estimating minute quantities of arsenic.

Wrapping the Heated Portion of the Tube in Copper Wire Gauze.

It was recommended by the Joint Committee of the Society of Chemical Industry and of the Society of Public Analysts, that a piece of copper wire gauze about one inch square be wrapped round the tube which receives the mirror at the point at which it is heated (immediately before the drawn-out portion). I directed my attention to find what advantage this offered over the heating of the naked glass tube, and ascertained that the use of the copper wire gauze was a positive disadvantage, the mirrors being less distinct with, than without, the wire gauze. Fig. 2 (a) shows photographic production of the mirrors in three tubes, which were obtained from 50 c.c. of a solution containing of a grain per gallon of arsenic trioxide,

shows photographic productions of mirrors in three tubes in the production of each of which 50 c.c. of the same solution were used, but where the heated portion of the tube was wrapped in one layer of copper wire gauze, as recommended by the Joint Committee. It will be seen on comparing these two series of tubes, that in each case more distinct and reliable mirrors are obtained by heating the naked tube than by heating it when wrapped in wire gaze.

I tried to determine the temperature at which small quantities of arsenic would be deposited from arseniuretted hydrogen, mixed with free hydrogen by placing the tube for receiving the mirror at right angles through an iron tube, underneath which was a Bunsen flame which was

* From the Memoirs and Proceedings of the Manchester Literary and Philosophical Society, xlvii., Part 6.

Free

FIG. 2.

(a).

(b).

& Zinc.

FIGS. 5 and 6. Hydrogen.

gradually raised, but I found that when heated to the full safety range of my thermometer, viz., 393° C., no trace of arsenic mirror was produced on the cooled part of the drawn-out portion of the tube. This result does not confirm the statement made in D. Mendeléeff's "The Principles of Chemistry" (English edition, 1897), vol. ii., p. 182, where he says, in speaking of arseniuretted hydrogen :-" But its presence in the most minute quantities may be easily recognised from the fact that it is easily decomposed by heat (200° C. according to Brunn) into metallic arsenic and hydrogen."

The gas containing arseniuretted hydrogen requires to be

heated to a very high temperature before all the arsenic is deposited as a mirror, and it appears evident that the wire gauze prevents the attainment of the necessary temperature for the complete decomposition of all the arseniuretted hydrogen present.

Influence of Temperature at the Drawn-out Portion of the Tube on which the Mirror is Deposited.

With the view of obtaining information as to this, I used for each of the following tests 50 c.c. of a solution containing of a grain of arsenic trioxide per gallon.

The tubes used are rather wider at the drawn-out parts

4

It is remarkable that the temperature of boiling water prevented the formation of the mirror altogether. The small ring of mirror formed just outside the beginning of the steam jacket, whilst the uncondensed arsenic passed for half-an-inch along the tube outside the steam jacket without depositing, and only made its appearance as a black deposit at the point (c) where it came in contact with the cooling effect of the stream of cold water.

The outside of the tube No. 3 in the series was covered with one layer of tissue paper from (d) to (e), and a stream of water at 50° C. was kept flowing over it. It is remarkable that at this temperature the bulk of the arsenic was deposited at the point where the comparatively hot water flowed over the tube, and it is curious to observe that on the further portion of the tube heated to this temperature no further mirror formed, but when the gas passed along the tube to the part which was not heated (to 50° C.) a second faint black deposit was formed, and, after leaving an interval of the tube free from deposit, a third, still fainter, black deposit made its appearance.

The tube No. 4 in this series was cooled by placing over it between the points (f) and (g) a single layer of tissue paper, which was cut into a triangular shape at the bottom, and over which water at 15° C. was kept rapidly dropping, as shown in Fig. 1. In this tube only one mirror was formed at the point where it came in contact with the cooled tube.

It was evident, then, first, that the cooling of the tube was an important condition for obtaining the largest mirrors; secondly, that when the tube was cooled to about 15' C. only one mirror formed; thirdly, that that mirror had a metallic lustre, and no second or third black deposit ever formed on any other part of the tube.

Since making these experiments I find that two other chemists have drawn attention to the importance of cooling tie portion of the tube arranged for receiving the mirror, the first being Gabriel Bertrand, who employs a strip of filter-paper 4 to 5 m.m. wide wrapped two or three times round the tube, and fed with water drop by drop; and the second A. Gautier (Bull. Soc. Chim., 1902, p. 27; 20, 21), whose method consists in applying a brass rider, the lower part of which is kept immersed in crushed ice.

My experience has shown that the best results have been given by a piece of tissue paper 3 or 4 inches long by-inch wide, folded in the centre, and hung over the tube for receiving the mirror, over which water is allowed to drop rapidly, the two hanging folds being cut to a point to allow the water to run off in a single stream into a glass placed underneath. A roll of several layers of filterpaper is not so effective for cooling, as the cold water takes some time to penetrate the several layers of paper. Brown-metallic looking Mirrors and Black Arsenic Deposits. It has been suggested by the Joint Committee above mentioned, and by other chemists, that the presence of oxygen or air, if mixed with the hydrogen from the generating apparatus, tends to produce black deposits rather than metallic arsenic mirrors; this seems to be so,

I have studied somewhat more minutely these two forms of arsenic. The first, or brown form, with metallic lustre, is that which is crystalline and firmly adherent to the glass tube; the second, or black form, is amorphous, and can easily be removed from the tube by gentle rubbing. I was led to believe that the amorphous form contained occluded gaseous matter, and I have spent some time in collecting a quantity of it (several grms.), which was placed in a tube from which the air was exhausted by a Topler pump, which is the method employed by Sir William Ramsay for removing helium from various minerals. After the tube was exhausted till no further gas could be drawn from it, the black arsenic was heated, and it volatilised and condensed in the crystalline form, but no trace of gaseous matter was liberated from it.

Internal Diameter of the Tube upon which the Arsenic Mirror is Deposited.

As the tubes used by me are all drawn out in the same manner from previously selected tube, they are found to be very closely the same in the internal diameter of the bore, but, as these tubes are slightly conical, I devised a simple measuring arrangement for obtaining the mirrors on exactly the same internal diameter of tube. This consists of an iron wire with the one end thinner than the other; the thicker end is first put into the drawn-out portion of the tube and then the thinner end, to see that the difference in the distance between which they enter is about of an inch, which it generally is, although the distance of the wider portion from the beginning of the drawn-out part varies slightly; the tissue paper for cooling is then placed so that the edge nearer the flame is exactly at the point where the entrance of the thicker end of the wire is arrested, and the mirrors, being all deposited at this point in the different tubes (where the cold water comes in contact with the glass), more accurate measures of the quantities of the deposits are obtained.

With the view of keeping the tubes hot to the point at which the cold water comes in contact with the narrow glass tube, to prevent the formation of a mirror before the portion of the tube is reached on which it should be deposited, I put a small coil of copper or platinum gauze of an inch square around the tube, so as to cover the rounded part of the drawn-out portion and part of the narrow tube to within 1 m.m. of the paper. I use a flat Bunsen flame one inch wide, the flame acting on about of an inch of the uncovered glass tube, with the end of the flame playing on the wire gauze.

Fig. 4 shows mirrors obtained by the cooling process on. the tubes of accurately measured internal diameter. The apparatus employed for carrying out the process is shown in Fig. 1.

Fading of the Arsenic Mirrors and Black Deposits. In my previous papers on the approximate estimation of arsenic, I have mentioned that I had not observed that any of my arsenic mirrors had faded even when exposed to air and light, but, as other chemists had had experience of fading, and had recommended the sealing of the mirrors in an atmosphere of hydrogen to prevent it, I considered it desirable to carry out this process, and have since done so,

CHEMICAL NEWS,

Nov. 6, 1903.

I was much surprised, however, to find that the mirrors which were made by the old process, and the original photographs of which are seen in Fig. 5, faded after exposure to the light for six weeks; a second photograph of the same, taken after exposure, is shown in Fig. 6.

It is difficult to imagine why these arsenic mirrors and deposits should have wholly or partially disappeared after being sealed up in an atmosphere of hydrogen. It is

curious, however, that, whilst the mirror from the middle tube has entirely disappeared, those in the first and third tubes have only partially faded.

Other instances of fading are shown in the photographs Figs. 7 and 8, all the deposits being obtained by the old process, in which the tubes were not cooled with water. The tubes Nos. 1 and 2 were unsealed, 3 and 4 were sealed in air, 5 and 6 in hydrogen, 7 and 8 in nitrogen, and and 10 in carbon dioxide, the mirrors being from the same amount of arsenic solution. The tubes 2, 4, 6, 8, and IO were detached from the card and exposed to the light for one month, the others being kept in the dark; they were then re-mounted, and again photographed. It will be seen in Fig. 8 that the mirrors in tubes 2 and 4 (exposed to the air) and in 8 (exposed to pure nitrogen prepared by heating ammonium nitrite, and passing the same through a tube containing copper gauze heated to redness) have faded considerably, while in tubes 6 and 10 (in hydrogen and in carbon dioxide) the black or second portion of the deposit only has faded. The mirrors formed by the cooling process appeared to have suffered no change in hydrogen.

My modified and new process affords a much more accurate method of approximately estimating minute quantities of arsenic, and it is much more delicate than the process previously employed; it is, in fact, so delicate that I have now failed to get any zinc which is abolutely free from any trace of arsenic. A very distinct mirror is formed with the th of a grain of arsenic trioxide per gallon, when working with 50 c.c.; that is one part in 140,000,000 parts of liquid, and half that amount can be detected; that is, one part in 280,000,000, which is equivalent to 1 grain of arsenic trioxide dissolved in 4000 gallons of beer; i.e., I grain dissolved in 111 barrels of beer of 36 gallons each, which is equivalent to 18 tons weight of beer.

EXPLANATION OF FIGURES.

FIG. 1.-Apparatus employed.

FIG. 2.-Photographic production of mirrors.

per gallon of As406 when using 50 c.c.

both gr.

(a) With the naked tube, heated directly with Bunsen flame.

(b) With the heated portion of the tube wrapped in wire gauze.

FIG. 3.-Showing the effect of temperature on the formation of the mirrors and arsenic deposits in a solution containing th of a grain per gallon of As406, when using 50 c.c,

1. Ordinary Marsh test method (without special cooling.

2. Enclosed in a steam jacket from (a) to (b), and cooled with cold water at (c).

3. With current of water at 50° C. passing over the outside of the tube between (d) and (e).

4. With current of water at 15° C. passing over the outside of the tube between (f) and (g). FIG. 4.-Standard bore tubes, using 50 c.c. with different quantities of As406. Cooling method.

FIG. 5.-Arsenic mirrors before exposure to the light. 26th gr. per gallon As406 using 50 c.c.

FIG. 6.-The same mirrors seen in Fig. 5, after exposure to the light for six weeks.

FIG. 7.-Photographs of five pairs of arsenic deposits, made by the old method, respectively unsealed, and sealed in air, hydrogen, nitrogen, and carbon dioxide. 10th gr. per gallon using 50 c.c.

FIG. 8.-Photographs of same after the right-hand one of each pair had been exposed to the light for six weeks.

231

THE VOLCANIC DUST OF MONT PELÉE.

By A. B. GRIFFITHS. Ph.D., &c.

MARTINIQUE is an island situated in the Caribbean Sea (14 degrees north latitude and 16 degrees west longitude), and its volcano, Mont Pelée, was extremely active last year. Mr. W. Wilson, Pharmaceutical Chemist, 281, Brixton Road, S.W., kindly sent me a specimen of the volcanic dust which he received from friends in the island. I have submitted the dust to analytical and microscopical examination, and a micro-photograph is enclosed with this note. It distinctly shows the crystalline nature of the dust. analysis of the dust (dried at 120° C.), yielded the following result:

Silica

Aluminium oxide.. Calcium oxide Ferric oxide.. Ferrous oxide Manganous oxide Magnesium oxide Potassium oxide Sodium oxide Titanium oxide Sulphuric oxide

55 ΟΙ

20'50

9:00

3.20

4.86

O'25 3:06

..

0.65

2'ΟΙ

0.68

0'42

O'20

0.16

traces

Phosphoric oxide..

Chlorine

Copper, nickel, lithium

An

Mont Pelée dust has a specific gravity of 2'7211, and contains 3.24 per cent of magnetic material.

Mr. W. J. Townsend and others believe that the wet summer has been caused by Mont Pelée. The black steam-clouds travel in the path of the trade winds over the Gulf Stream, and "there are thousands of tons in the air still waiting to trouble us. and snowy winter."

So look out for a wet autumn

PROCEEDINGS OF SOCIETIES.

PHYSICAL SOCIETY.

Ordinary Meeting, October 23rd, 1903.

Dr. R. T. GLAZEBROOK, F.R.S., President, in the Chair.

A PAPER on "The Bending of Magnetometer Deflectionbars" was read by Dr. CHREE.

A theoretical paper contributed to the Society by the present author in May, 1901, proved amongst other results that the bending of the deflection-bar of an ordinary magnetometer, under the combined weight of the bar and its load, must increase the distance between the deflecting and deflected magnets, during a determination of horizontal force, to an extent which is not negligible. This conclusion has been borne out by direct observations made at the National Physical Laboratory on a number of magnetometer-bars, including specimens from the leading makers. The mean results thus obtained are recorded in the present paper. The great majority of the data were derived from direct observations made with a pair of microscopes. But a method is described whereby the necessary information can be deduced without the aid of microscopes from the magnetometer's own readings. The magnetometer is set up exactly as in an ordinary deflection experiment, with the deflecting magnet on its carriage at any convenient position on the deflection-bar, and equal weights are hung up symmetrically one on each arm. The consequent increase in distance between the two magnets diminishes slightly the deflecting force on the deflected magnet, and a slight change of

Prof. S. P. THOMPSON said that Dr. Chree had pointed out an important matter in magnetometry, and had told us of the existence of various sources of error. He asked why instrument makers could not design a better instrument, referring especially to the deflection-bar itself, which he said was not of a good form from an engineering point of view. Could not the bending be considerably reduced by making the bar in the form of a girder? Would it not be an advantage to lessen the weight of the carriage by making it of aluminium or of some light alloy? Prof. Thompson said that although the error referred to by the author of the paper seemed large, yet it could not be greater than errors arising from inaccuracy in the setting of the carriage or from other causes inherent in a deflectionexperiment.

Dr. W. WATSON expressed his interest in the paper, and said he had never before appreciated the magnitude of the error due to the bending of the deflection-bar. In the many determinations of horizontal magnetic force which he had carried out, he had increased the effect by placing the thermometer near to the magnet on the same side of the magnetometer. He did not think it would be an advantage to lighten the carriage too much, because of the wear and tear which an ordinary magnetometer must be capable of withstanding. With regard to the accuracy of the setting of the carriage, he said that it was not easy to set it to o' m.m., but that inaccuracies of setting to a certain extent compensated each other. Dr. Watson also pointed out that the position of the magnet in the carriage was not always absolutely the same.

The CHAIRMAN pointed out that the method described in the paper might be used to determine Young's modulus by bending the bar with magnets of known weight. A large number of errors entered into the determination of horizontal force, and they were being investigated one by one at the National Physical Laboratory. Dr. CHREE, replying to Prof. Thompson's remarks, said that various attempts had been made to get rid of the bending effect. The difficulties were due to the fact that it was necessary that the bar should lie in one vertical plane, and that it should pass through the centre of the magnetometer. The error might be reduced by bending the bar so as to make the distance of the magnet from the neutral axis small, but the introduction of corners rendered the theoretical treatment of the problem difficult, and a bent bar was difficult to measure practically. In some of the magnetometers used on the United States Coast and Geodetic Survey, the deflection-bar consisted of two parts pinned together at the centre of the instrument, and although this arrangement had advantages, a great deal depended upon the pin fitting tightly. He contended that if the bending effect could be measured accurately, its magnitude was of minor importance. In the magnetometers used on the Indian Survey, the carriages were much heavier than those in use on ordinary unifilars, and the deflection-bar was constructed something like a girder. There were many causes of error in a force experiment. The formulæ used were complicated, and much depended upon an accurate knowledge of the constants P and Q.

A paper "On the Magnetism of Basalt and the Magnetic Behaviour of Basaltic Bars when heated in Air" was read by Dr. G. ALlan.

Bars cut from basalt obtained from Rowley Regis and

metric method to determine their magnetic properties at temperatures from 15° to 800° C. Hysteresis curves are given, and the temperature-permeability curves show that whilst the English basalt has, in general, a maximum permeability near 500° C. followed by a minimum about 550° C., the temperature of maximum permeability in the case of the German basalt lay in the neighbourhood of 50 C., there being a subsequent gradual loss of strength with rise of temperature. Sections of heated and unheated rocks are given, showing evidence of chemical change in some of the rock constituents, and a table of values of susceptibility of the specimens is appended.

The CHAIRMAN expressed his interest in the paper, and said that Dr. Allen had given interesting data which would be of value in the subject of terrestrial magnetism.

Prof. W. F. BARRETT, in a letter sent to the Secretaries, said that Dr. Allan's paper was one of great interest and considerable importance, and he wished to congratulate the author on the results of his laborious investigation. Several years ago he (Prof. Barrett) examined the magnetic properties of various specimens of columnar basalt which had been taken from the Giant's Causeway in Co. Antrim, and a note on the results of this investigation was published in the Proceedings of the Royal Dublin Society for Dec., 1899. Each block was found permanently magnetised with a strongly marked north and south pole, the magnetic axis running diagonally through the block, and inclined to the horizon approximately at the angle of dip. As the blocks formed part of vertical columns in the Causeway, their magnetisation was undoubtedly due to the earth's magnetic field, the concave end of each block-for the ends are not plane but slightly concave or convex-was (in all the blocks examined) found to be a north-seeking pole, and_must therefore have been downwards when in situ at the Causeway. The weight of the blocks varied from 24 to 37 kilogrms., and the specific gravity of a fragment of one of them was found to be 2.86. With such large irregular and feebly-magnetised masses only a very approximate estimate of the magnetic moment was possible by the ordinary method. This was, however, attempted, the blocks being placed on a turn-table with their centres 60 c.m. distant from the reflecting magnetometer. As might be expected, the magnet moment per grm. was found to be very small and almost alike in each side. The volume of the blocks being known, their permanent intensity of magnetisation was found to be less than that estimated by Everett for the earth, regarded as a uniformly magnetised body, viz., 0079. Nothing could be inferred from this, as the blocks had been removed a long time from the Causeway and had been lying about in different magnetic positions to that which they originally occupied. The long retentivity of the direction of their original magnetisation in spite of rough usage was, however, somewhat remarkable.

Dr. WATSON then showed some Experiments with Electrical Oscillations.

ROYAL SOCIETY OF NEW SOUTH WALES. General Monthly Meeting, September 2, 1903.

F. B. GUTHRIE, F.I.C., F.C.S., President, in the Chair.

"The Separation of Iron from Nickel and Cobalt by Lead Oxide (Field's Method)." By T. H. LABY. The accurate separation of iron from nickel and cobalt is peculiarly difficult.

Review of Methods. Ammonium Hydrate and Chloride.-The ferric hydrate precipitate carries down so much of the nickel and cobalt present as to require three re-precipitations.

Ammonium Carbonate.-This much-recommended separation is long and tedious.

Basic Acetate.-Using a small quantity of acetate to