JULY 1938 Vol. 63 No 748 THE ANALYST PROCEEDINGS OF THE SOCIETY OF PUBLIC ANALYSTS AND OTHER ANALYTICAL CHEMISTS Microchemical Investigations on Siliceous Dusts" BY JANET W. NATTHEWS PHD. D.I.C. F.I.C. (Read at the Veetiag April 6 1938) I. GENERAL.-Akfter the war in South Africa when the gold mines were re-opened it was observed that many of the white miners who had previously been employed failed to return. On making inquiries it was discovered that they had all died of some disease of the lungs which was identified as silicosis; thus it was discovered that silicosis instead of being a rare disease was extremely common. -4s knowledge of the characteristics of the disease spread throughout the medical profession and the Public Health Authorities it was realised that it was wide-spread in industry.Apart from the very high incidence of silicosis among the South African gold-miners where the cost to the industry has been as high as one million pounds a year in England it was found that coal-miners in certain districts were liable to contract the disease as were also tin-miners workers in the potteries and in flint-grinding mills quarrymen sand-blasters and others. Asbestosis though allied was held to be a different disease owing to some clinical differences notably the presence of specific asbestosis bodies in the lungs, and the fact that asbestosis was caused by breathing the dust of a silicate asbestos, whilst silicosis was considered to be due to inhalation of pure silica. * This is a short summary of work carried out during the past five years a t the Imperial College under the direction of Professor H.V. A. Briscoe and in collaboration with Miss P. M. Sanderson B.Sc. D.I.C. and Mr. P. F. Holt BSc. It comprises work published as follows:-J . W. Matthews and H. V. A. Briscoe Tram. Inst. Min. & Met. Nov. 1034; H. V. A. Briscoe, J. W. Matthews P. F. Holt and P. M. Sanderson ibid. April 1937; H. V. A. Briscoe J. W. Matthews P. F. Holt and P. M. Sanderson ibid. June 1937; H. V. A. Briscoe and J. W. Matthews, Mikrochimica Acta 1937 1 266; papers read before meetings of the Microchemical Club J. W, Matthews 1934 (2 papers) P. M. Sanderson 1936 P. F. Holt 1936 P. M. Sanderson 1938. 46 468 MATTHEWS MICROCHEMICAL INVESTIGATIONS OX SILICEOUS DUSTS Up to 1929 all the research work on asbestosis and silicosis had been carried out by medical men and at a World Congress it was stated that silicosis was caused by breathing-in the dust of free silica.The conclusions were based on microscopic examination and on some chemical analyses of the ash of the silicotic lungs. As is well known the chemist expresses an analysis as the oxides of the various elements present (6.g. Fe,O, A1,0, SiO, P,O,) but this does not mean that the oxides are present as such in the sample; it appears that this had been assumed without recognition that a high percentage of A1,0, K,O and Na,O implies that most of the SiO present occurs as silicate and not as free silica. Further the microscopical examination of the lung was carried out under such conditions that quartz crystals were usually detectable but other silicate minerals were unobserved.Dr. W. R. Jones of the Royal School of Mines was the first to point out in 1933 these facts and to indicate that from the analysis figures, silicates were as likely as free silica to cause silicosis; he was also the first to show the general prevalence of the mineral sericite in the ash of the silicotic lungs after acid treatment. He stated that the relative proportions of the different minerals in the dust of a rock are not necessarily the same as the proportions in the rock itself. Since at that time in England the law of compensation for silicosis was based on the analysis of the rock being worked and not on that of the dust his statement, if accepted would render necessary an alteration in the law.Dr. Jones caused the whole subject to be re-examined from the beginning, and at this stage we felt that some purely chemical work on the actual dusts might be helpful for until the chemical composition and properties of a dust and the chemical behaviour of the different minerals composing it are known a true knowledge of its behaviour in the lung is impossible, Pathologists are generally of the opinion that it is not the physical properties of the particles but the substances which dissolve out of the dust in the lung that cause the damage-whether the silica alone as is generally assumed or together with some other contributory factor. Therefore we felt that there were two im-portant subjects for investigation :-( 1) A chemical investigation of dangerous dusts which are known to cause silicosis and a comparison of their properties with those of siliceous but non-dangerous dusts; (2) an investigation of the solubility of the dusts and of finely divided particles of pure minerals under standardised conditions as no accurate information on this subject was hitherto available.METHODS FOR THE COLLECTION OF DUST.-In the investigation of the properties of dangerous dusts we had in the first place to devise suitable methods for collecting the dust for analysis as no good methods were available for obtaining for quantitative analysis a sample of dust in an unaltered condition. Since the dust concentration in a dangerous atmosphere is often only of the order of 1 mg. per cb. metre micro-methods of analysis were essential as the difficulties of obtain-ing samples large enough for ordinary methods are very great.For a complete chemical analysis it is desirable to obtain a sample of dust of the order of 50mg. but for an investigation of the solubility of a dust and an examination of the extract considerably larger samples are required-of the order of a few grams or even more. 11 XATTHEWS MICROCHEMICAL ISYESTIGATIONS ON SILICEOUS DUSTS 469 We have devised two different types of apparatus for collecting dust. The 6rst type uses filtration of the air for the collection with a filter-pad made either of a volatile solid (such as benzoic acid naphthalene or anthracene) removable by sublimation or a filter-pad made of a porous solid removable by solution in a Ron-aqueous medium and separation by centrifuging.This latter type of filter, with a filter-pad of porous solid has proved the most useful to us in our particular work in taking samples in mines and factories where the dust concentration is low. The filter is primarily intended to obtain samples of the order of 20-500mg. for quantitative analysis the samples being collected in a short time (about 1 to 6 hours) so that working conditions temperature humidity etc. are fairly constant and the differences in the composition of the dust due to these factors can be determined if necessary. This type of filter is designed to give almost complete retention of the dust so that the mass concentration of dusts may be determined. FIG. 1. Salicylic Acid Filter-holder. The filter medium we found most suitable was salicylic acid AR quality, sieved through a 40-mesh sieve-the mean crystal length being of the order of l o p .Finer crystals are more resistant to flow and coarser crystals are not so retentive of dust. The crystals are packed on the stainless steel gauzes A of the filter-holder shown in Fig. 1. The three stainless steel gauzes consisting of two coarse mesh gauzes with a fine mesh gauze in between (the coarse gauzes are essential to give rigidity) are screwed into position by means of the spoked support B made of ebonite. A filter-cell 7 cm. in diameter requires 10 g. of salicylic acid to give a pad thickness of about 4 mm. which has been found to be sufficiently thick to provide good retention without obstructing the flow. The efficiency of the filter was tested by using fine elutriated copper tartrate dust and catching on a filter-paper behind the pad any that had penetrated.The copper on the filter-paper was detected by Feigl’s rubeanic acid test. In a typical test with copper tartrate dust, with a rate of flow of 60 litres per minute the average penetration was about 0.05 per cent. Further experiments have shown that the penetration while appreciable at the beginning of a run sinks to a negligible amount after a short time as the dust The holder itself is also made of ebonite 470 MATTHETT-s MICROCHEMICAL ISVESTIGATIOSS 0s SILICEOUS DUSTS itself acts as a filter. For this reason the filtei- is unsuitable for the collection of more than a few grams as dust forms such an efficient filter that the flow is cut down to almost nil.For transport the filter-cell is covered with a fiat cellophane lid and may be handled freely or transported to any part of the world. Before the dust is sampled the crystals are shaken back over the whole surface and gently pressed into position with the finger-nail. The cell is fitted to the support horizontally the cellophane is removed and suction started. The greater the suction the more efficient are the filters. It is convenient to use an ejector operated by compressed air at 60-75 lb. per sq. in. When the suction is on the filters are very stable and sampling may be carried out in any position. FIG. 2. Cylindrical Labyrinth Dust Collector. When sampling is completed the pads are covered with cellophane and the whole filter returned to the laboratory where the whole of the salicylic acid pad, together with the dust is transferred to a centrifuge tube-a 50-ml.tube for 10 g. of salicylic acid-and the salicylic acid is removed by treatment with absolute alcohol. The dust is then air-dried and is in a suitable condition for chemical analysis or petrographic examination. The second apparatus for collecting larger samples of dust over longer periods of time we call the “labyrinth.” This may be used for samples of about 1 to 400g. and for a sampling time usually of 1 to 8 weeks though the times may be varied according to the dust concentration. This apparatus (Fig. 2 ) is based on deposition instead of filtration. It consists of a number of copper plates held together by a long brass tube B and kept in place by short equal lengths of brass tube.This is convenient as it shows the size of the sample collected without the apparatus being taken to pieces. The second model (Fig. 2) is brass and cylindrical in shape and is designed for safe travelling. To permit of the ready and complete removal of the collected sample, the baffle assembly is wrapped with an overlapping sheet of celluloid F held in place by a spiral winding of woollen yarn G of such thickness that the whole assembly fits in the copper casing. The suction used for tests in factories was that of a “high-pressure” fan ventilating the plant. To minimise the risk of large particles being caught we sometimes use an elutriator improvised from an iron drum. The efficiency of the labyrinth may best be determined from the graph of the weight distribution of the dust collected in the different sections.This varies considerably both with the rate of flow and the type of dust as is shown in Figs. 3 and 4. The efficiency is There are 32 plates and l& cm. space above each. The first experimental model had glass walls. Most of the sample is collected in the first few sections MATTHESVS MICROCHEMICAL INVESTIGATIOXS OX SILICEOUS DUSTS 471 high for such a dust as flint in a flint factory and low for a dust like asbestos. While it is interesting to know that the mass efficiency can be very high for our i' MASS D ISTRl B UTION ~ OF DUSTS IN 1 LABYRINTHS FIG. 3. purpose it is unimportant for we require only a representative sample of dust for examination.When a knowledge of the actual weight of dust in a given volume of air is required the salicylic acid filter is better as its mass efficiency is very high. Flint Dust from Laboratory Disintegrator in Cylindrical and Rectangular Labyrinths Ratio of air-velocities approximately Cyl. Rect. 4:l. 2 5 10 LABYRINTH SECTION NUMBER FIG. 4. ?he chief advantage of the labyrinth is that it enables us to obtain a large sample of uncontaminated dust that requires no further treatment such as e.g. alcohcpli extraction. A further advantage is that the labyrinth sample is graded in par.ticle size according to the different sections. Figs. 5 and 6 show flint dust collected in the first and last sections of the labyrinth from a flint-grindin 472 MATTHEWS NICROCHEMICAL INVESTIGATIONS ON SILICEOUS DUSTS factory.The particle size falls off so that in the last sections all the particles are 5p and under and it is these particles that are considered to be most dangerous in the lung Thus with the labyrinth samples we have been able to investigate the variation in properties with particle size. 111. METHODS OF AKALYSIS.-AS our work is largely comparative and con-cerned with fairly large percentage differences we need methods that are primarily rapid and accurate to t-0.2-0-5 per cent. rather than the highest accuracy ( 10-1-0.2 per cent.). Up to the present we have been chiefly concerned with the determination of water silica iron aluminium calcium magnesium sodium and potassium as these occur most frequently in the dusts we have examined.Apparatus.-For the determination of silica we use platinum crucibles weighing about 2 g. of 2 to 3 ml. capacity and with a base 6 mrn. in diameter. For the alkalis we use platinum micro-beakers of 5 ml. capacity and for the other elements the usual Emich micro-beakers of Jena glass of 5 t o 6 ml. capacity. For filtration we use porcelain filter-sticks and the Emich filtering technique.2 For the filtration of silica we use the simple micro-immersion filters described by King3 made of such a size that the filtering surface conveniently fits the base of the platinum crucible. After the silica has been filtered off the filter-paper pad is left behind in the crucible the contents of which can then be ignited and treated with hydrofluoric acid without danger.,4 single sample of dust 5 to 10 mg. is used for five to seven determinations: namely the loss of weight at 100" C. loss on ignition silica iron and aluminium, and if necessary calcium and magnesium. Sodium and potassium are determined on a separate sample. Silica.*-After determination of loss a t 100" C. and loss on ignition the dust is fused with three to four times its weight of sodium carbonate. The melt is taken up in 1 ml. of water and neutralised without danger of spirting by exposing it to acid vapour from a small dish of hydrochloric acid warmed slightly on the top of an air-oven run a t 100" C. The crucible is supported in a small triangle resting on the dish of acid and both are covered with a crystallising dish or small hell-jar. The silica is rendered insoluble by evaporation three times with hydrochloric acid followed each time by 20 minutes' drying at 105" C.Finally the residue is taken up in water containing about 0.25 ml. of hydrochloric acid and filtered through a King? filter-stick3 into a glass micro-beaker for the iron and aluminium determinations. The insoluble matter is ignited and weighed and the silica is volatilised by treatment with hydrofluoric acid and evaporation in the presence of water con-taining dilute sulphuric acid. The crucible is finally ignited for two minutes and re-weighed. We have found no difficulty in this method which gives excellent results. Nitric acid or perchloric acid may also be used for the dehydration o f the silica. For the determination of silica in aqueous extracts of the dust we use King's5 * The method is similar except in small details to that used by Thurnwald and Benedetti-Pichler .t When a platinum filter-stick is used for this determination careful analysij wiil give results differing from the calculated values by only 0.02 to 0-03 per cent.-( Private Communication from IT'. L. Stephenson Medway Technical College. ~ T I G S . .T a i i t l (i. 1;liiit (lust collcctctl froiii first ant1 last sections of Labyrinth, showing decrease. in particle. sizc. S;calc- Ixngtli of sitlw of stluares = 2Op MATTHEWS MICROCHEMICAL INVESTIGATIONS ON SILICEOUS DUSTS 473 colorimetric method comparing the yellow ammonium silico-molybdate with a standard picric acid solution. Iron and Aluminium.-As in our samples of silicate minerals iron is never present in large amounts compared with the aluminium we are able to use the hydroxyquinoline method as described by Benedetti-Pichler4 for the R,O deter-mination.The precipitate is dried at 140" C. weighed and dissolved in sufficient hydrochloric acid and the iron is determined colorimetrically without decom-position of the organic matter. Calcium and magnesium are determined as the oxalate and the ammonium phosphate respectively after ignition of the filtrate from the oxine precipitation, The oxalate is dried at 100"-105" C. and the phosphate at room temperature.* Sodium and Potassium.-We have tried a considerable number of methods for sodium and potassium in the hope of finding one more rapid than the classical methods or one in which the alkalis could be determined in the same sample as the other elements; partial success was obtained for potassium only by a method in which perchloric acid is used instead of hydrochloric acid for the dehydration of the silica and the solution is evaporated after removal of the silica and separation of the potassium as perchlorate before determining the other elements.Of the methods tried in which the sodium and potassium are determined in a separate sample a simplified micro-Berzelius procedure is proving very satisfactory.6 The sodium and potassium are determined in aliquot portions after removal of the silica, the sodium directly and the potassium after removal of sulphate from the solution. In aqueous solutions the zinc uranyl method for sodium and the platinic chloride or cobaltinitrite method for potassium used gravimetrically give excellent results, Water.-Water lost at 100" C.is usually determined by simple drying in an air-oven. Loss on ignition is determined by igniting the sample for 3 to 4 minutes and rapidly cooling and weighing after 1 minute or cooling in a closed container. This is necessary as the dusts are extremely hygroscopic. To distinguish between loss on ignition and water we have also determined water in some samples using the Pregl micro-combustion apparatus. Titratable alkali is determined in aqueous extracts of the dusts using a capillary micro-burette of the Rehberg type with a metal screw and N/10 acid or alter-natively a Pregl burette and N/100 acid. REsuLTs.-The first dust we examined was that of a tin mine known to be extremely dangerous.The rock being drilled contains considerable amounts of felspar and some quartz. The samples were collected with the salicylic acid filter and examined gravi-metrically with the results shown in Table 11. The most striking factor is the extremely large "loss on ignition"-much greater than in the original rock. On reporting this to our colleague Dr. Brammall of the Department of Geology he said that it confirmed his theory as to the extreme reactivity of finely divided particles of minerals which are inert in the mass; these small particles readily combine with water from the atmosphere and also from the moist lung and in this hydrated form the mineral is much more soluble and both caustic alkali and silica are liberated in water.We were able to confirm this liberation of alkali, using a small sample and the Rehberg burette. We then repeated the sampling, IV. General sampling results are shown in Table I 474 MATTHEWS 1 MICROCHEMICAL INVESTIGATIONS ON SILICEOUS DUSTS this time with both the labyrinth and the salicylic acid filter to confirm the results and obtained figures for loss on ignition varying from 10 to 25 per cent. The actual water was determined on two samples by the Pregl method and found to be 9 and 10 per cent. the rest of the loss being carbon dioxide from traces of carbonate present and from the burning of oil and other organic matter in the dust. Thus we were able to confirm Dr. Jones's hypothesis that the dust of a TABLE I SAMPLING RESULTS Distance Total from Rate Sampl- volume Date of Filter working of ing of air No.sampling type face flow time passed 1 21st Dec. 1936 Coarse 10 15.3 2.5 33 2 21st ) , Fine 20 4.2 2.5 11 3 22nd , , Coarse 10 19-2 2.5 48 4 22nd 3 , J 20 15.8 2.75 43 5 22nd , , Fine 40 9.1 2.5 23 6* 22nd , , Coarse 145 about 10 2.5 2% 23rd 9 ) 2 J 20 17.8 1 18 8 23rd , , ,f -t 16.7 3.3 55 23rd J > J J > I 145 15.6 3 47 10" 23rd , , 9 145 15.4 2.1 32 f t . cb. metres/hr. hrs. cb. metres. Weight col-lected mg* 19 158 89 36 103 16 26 38 105 7.6 Dust concen-tration mg./cb. m. 0.49:: 3.3 2.1 1-6 4.1 0.9 0.5 0-8 3.3 0.731 * These samples were run over blasting. t Nearest drilling likely to affect sample estimated to be 9 mile away.$ These low and irregular results are probably due partly to the fact that on the first day considerably more water was fed to the drill and sprayed on the face than on subsequent days, and partly to the fact that no driving had been done on the previous day a Sunday. mineral may be very different in properties from the mineral in bulk the difference being in the capacity to take up water and to dissolve in water to form a solution not of the mineral as a whole but of selected ions. The increased solubility of finely divided particles is well known but it was not realised that the finely divided particles could be so rapidly hydrolysed this hydrolysis being the cause of the increased solubility of the silica in the dust. The unexpectedly high silica solubility of the dust together with the liberation TABLE I1 ANALYTICAL RESULTS IN PERCENTAGES Total loss weighing Loss at Loss on immediately Ferric Sample 100' C.ignition after ignition Silica oxide Alumina 0.76 7-8 1 - - - 9.6 Filter left overnight 3.7 13.6 No. 3 10 ft. from face 0-06 3.7 3'8 {tt: 4.1 14-1 ) ) 4,* 20 , , ,) 0.24 5.7 6.4 64.2 5-5 14.6 3 ) 5J 40 ) > > # ) 9 OS5 8.4 __. 62.6 5.6 13-8 (damaged) A sample of the rock gave a loss on ignition of 0-60 per cent. * All the samples contained threads (apparently from clothes) and this sample contained human hairs MATTHEWS MICROCHEMICAL INVESTIGATIONS ON SILICEOUS DUSTS 475 of alkali seemed to us to be two most important factors and after some work on quartz dust we decided on the conditions for comparative determinations of the solubilities of the different dusts in the preliminary survey.Using quartz powder we found that as is to be expected the silica solubility increases with time and temperature. On similar samples the same solubility was attained in 5 hours at 100" C. as in 100 days at 20" C.-namely 5 mg. of SiO, per 100ml. of water, The solubility also varies with the proportion of solid to solute-that is there exists no definite solubility in the ordinary sense which is normally attained when the solid is present in excess but we find that within limits the more solid present the more silica goes into solution. For example a suspension containing 1 per cent. of finely divided quartz gave 1.4 mg. of SiO in 100 ml. of water and one containing 4 per cent.gave 3.8 mg. of SiO,. The limiting amount when increase in the amount of solid shows no increase in solubility differs for different dusts and for the same dust depends on the particle size distribution. Therefore it is necessary rigidly to define the conditions of extraction and in the results given we used 1 g. of dust and 100 ml. of water (or 0-2 g. in 20 ml.) at 100" C. for 5 hours. We use nickel tubes to avoid contamination with silica from the glass; where glass is used a correction for the blank which is very small for monax glass is applied. Table I11 shows the enormous variation in the solubility of silica with decreas-ing particle size on samples of flint dust obtained with the labyrinth the same samples shown in the photomicrographs (Figs.5 and 6). Flint dust is known to be extremely dangerous ; if the silica solubility possibly accompanied by alkalinity, is a dangerous factor then it is striking to see how the danger of a dust must increase with its fineness. TABLE I11 SOLUBILITY OF CALCINED FLINT DUST Particle size-distribution I 3 Soluble Labyrinth 3-2p 2-lp < 1P silica Section No. Per Cent. Per Cent. Per Cent. mg./100 ml. 2 15 18 65 4.5 4.9 4 5.1 6 A - - -- - -Alkali ml. of N/100 per 5 ml. of extract 1.18 1 *20 1.12 1-05 - - - 12.1 - 14 16 0.6 9 90 12.1 1.25 Comparative analyses of the solid flint dust itself from the two extreme sections of the labyrinth show that the only significant difference in composition with decreasing particle size is an increase in loss on ignition.We next took samples of another extremely dangerous dust-asbestos-and obtained figures for the silica solubility ranging from 5 to 30 mg. of SiO in 100 ml. of water and for alkali from 0.3 to 1.3 ml. of N/100 in 100 ml. Our next step was to compare these dusts with two siliceous dusts which are considered to be non-dangerous viz. cement and sillimanite from factories wher 476 MATTHEWS MICROCHEMICAL INVESTIGATIONS ON SILICEOUS DUSTS no particular ventilation precautions are taken whereas both the flint and asbestos factories had most efficient ventilation. On extracting samples of these two harmless dusts in water we found that the silica solubility was less than 1 mg. per 100 ml. which is of the order of 5 per cent.of that given by the dangerous dusts. We analysed the cement dust extract and found it to be precisely the same as that given by ordinary cement consisting to the extent of 95 per cent. simply of calcium sulphate and lime; the dust thus appears to have no properties differing from those of the material giving rise to the dust. On determining the comparative silica solubility of a number of powders, freshly ground in the laboratory and sieved through silk with particle size practically all under lop we obtained values that show clearly that raw flint and asbestos, which are by reputation more dangerous give very much more soluble silica in distilled water than minerals such as sericite kaolin and felspar. A very interesting point is a difference observed between the solubility of calcined flint (which is liable to contain lime) and pure flint the latter being so much more soluble than the former.This led us to test the effect of lime or cement (from which lime is liberated in water) on other dusts. As is shown in Table IV the silica solubility of asbestos mixed with an equal weight of lime or cement is reduced to a very low figure; the effect on flint dust on the tin mine dust and on quartz dust is equally striking. The effect of other calcium compounds is not nearly so great as that of the oxide. TABLE IV EFFECT OF LIME ON SILICA SOLUBILITY OF ASBESTOS Soluble silica yielded by 1 g. of asbestos in 100 ml. water in 5 hours at 100” C. mg. Sample Asbestos alone . . * . 29 , + lime (CaO) 1. 0.3 , $- cement 1.3 Exactly how the lime effects this reduction in solubility is not certain; possibly it is the reverse of the process involved in the preservation of eggs (shell-CaCO,) in water-glass (sodium silicate).This reduction of “silica solubility” by certain metals has been recently confirmed by King,’ who is also of the opinion that the action is purely chemical. Denny Robson and Irwin* have further noted the reducing effect of powdered aluminium on the solubility of quartz and have carried out animal experiments which indicate that whilst rabbits breathing freshly formed fine quartz dust contract silicosis when a small amount of aluminium powder is also present they remain healthy. These workers also tested the effect of traces of lime (insufficient to precipitate the calcium silicate) and found that the solubility of quartz was increased.This is of course to be expected owing to the alkalinity of the solution and confirms our theory as to the importance of traces of alkali in increasing the silica solubility. We have also tested the effect of sugar charcoal on silica solubility which might be expected to reduce the solubility by adsorption but found that the problem was very complex. The solubilities of the various powders and dusts teste MATTHEWS MICROCHEMICAL INVESTIGATIONS ON SILICEOUS DUSTS 477 were differently affected and further the different samples of charcoal varied in effect according to their past history. In conclusion I should like to put forward the tentative hypothesis that a dust is dangerous in so far as it dissolves to give soluble silica and that when alkali is simultaneously liberated it may increase the damage in the lung either by increasing the solubility of any quartz present or by actual local damage caused by the high pH or by both means.Further it may be possible to render a dangerous dust innocuous by mixing it with a second dust which reduces the silica solubility such as perhaps cement or iron or charcoal. We have to carry out animal experiments together with a great deal more laboratory work before we can confirm this hypothesis. Our work indicates the complexity of the problem of the solubility of silica; the solubility of quartz and silicates depends on a large number of factors-PH, particle size time temperature degree of hydration and presence or absence of other elements and compounds-so that a very large range of values may be obtained for every pure siliceous mineral and the behaviour of mixtures is even more complex.We hope however that it may be possible to find conditions in vitro by means of which the danger of a dust may be assessed in terms of some quantitative chemical determination even though the actual conditions in the living lung may be vastly different. This work has been carried out with the aid of grants from the Institute of Chemistry the Institution of Mining and Metallurgy and the Medical Research Council to whom thanks are due. This matter is under investigation. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. W. R. Jones J . Chem Met. and Mining Soc. S . Africa 1933 99; Abst.ANALYST 1934, Inst. Min. and Metall. Bull. No. 352 1934; Abst. ANALYST 1934 59 191. F. Emich Laboratory Manual of Microclzemistry. E. J. King ANALYST 1933 58 525. H. Thurnwald and A. A. Benedetti-Pichler Mikrochem. 1932 11 200. E. J. King J . Biol. Chem. 1928 80 25; Abst. ANALYST 1929 54 52. P. M. Sanderson Abst. Chem. and Ind. 1938 57 351. E. J. King and M. McGeorge Biochem. J. 1938,32 417. J. J. Denny W. D. Robson and D. A. Irwin J . Can. Med. Assoc. 1937 37 1. 59 124. IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY SOUTH KENSINGTON LONDON S.W.7 May 1938 DrscussioN Dr. J. H. HAMENCE asked if Dr. Matthews could give any idea of the proportion of suspended matter in the air of the average mine. How much air would one have to suck through the filter to obtain sufficient dust to work on? Mr.R. C . CHIRNSIDE remarked that it was difficult to get information as to which silica dusts were dangerous. For instance in blowing quartz lamp bulbs the operation took much longer than with glass and the quartz bulb became coated with a soft white deposit of silica which the glassblowers declared pro-duced an unpleasant taste in their mouths. One would like to be able to assure them that they ran no risk of silicosis. Was there anything in this connection in the regulations for testing the stone dusts for mines? If the regulations covered only the amount of combustible matter might not one be dusting the mine with dust which was dangerous from the point of view of silicosis? He wondere 478 RAYMOND THE DETECTION AND ESTIMATION OF OUABAIN AND STROPHANTHIN whether Dr.Matthews had seen the paper by R. F. Sykes in the Journal of the Society of GZass Technology 1936 which described a modified Berzelius method for the determination of alkali. Mr. F. W. EDWARDS suggested that for determining the actual mineral matter of some of these particles Dr. Matthews might find micro-analysis with a petro-logical microscope useful; he had found it so for examining particles in substances such as jam. Dr. MATTHEWS replying to Dr Hamence said that in the average mine where precautions were taken the amount of suspended matter in the air was of the order of about 1 milligram per cubic metre. One would have to suck 50 or 60 cubic metres of air through the salicylic acid filter to obtain a sufficient sample of dust. With regard to the use of the petrological microscope the chief difficulty was that it was very hard to identify particles as small as 5p and almost impossible to identify those less than 2p. It was the size of the particles that was the trouble. For determining the titration value methyl red was used as indicator and the solution was made just alkaline She had no information about quartz-blowing, but if the men were getting the dust in their mouths they were swallowing it not breathing it which was the dangerous procedure. As regards regulations for testing standard dusts for mines she did not know whether standard tests were made or whether such dusts were in themselves dangerous. She thanked Mr. Chirnside for the reference; she had not seen the paper. So far she had not made any determinations of titanium in clay dusts. Neither oxides nor sulphides of arsenic had as yet been found in the dusts examined