31 ABSTRACTS OF CHICRIIICAL PAPERS PUBLISHED IN BRITISH AND FOREIGN JOURNALS. General and Physical Chemistry. The Absolute Weight of Atoms. By J. ANNAHEIM (Deut. Clienz. Ges. Be,.. ix 1151-1152) .-To show the divisibility of matter 0.0007 gram rosaniline hydrochloride C,,H,,N,,HCl was dissolved in alcohol and the solution diluted to 1litre. When a burette of 1cm. diameter is filled with this solution the red colour is clearly seen on a white background; and even if one drop be allowed to fall into a small test-tube which is held obliquely on a white paper the colour can still be perceived. Now as I cm. of the solution gives 35 drops it follows that 0~00000002 gram of the colour is perceptible and on dividing this number by the molecular weight of ihe hydrochloride we find that the absolute weight of 1 atom of hydrogen cannot be larger than 0~000000000059 gram.By using a solution of cyanine C28H35XJ,it was found that 0*0000000285 could be seen or that the weight of the hydrogen atom is not greater than 0~000000000054. c. s. Specific Heats of Saline Solutions. By C. MARIGNAC (Am. Chim. P7q.s. [5] viii 410-430).-This paper gives the numerical results of a long series of experiments undertaken with the view of determining. whether any relation could be traced between the specific heats of saline solutions and the nature of the contained acids and bases. The results show a certain parallelism in the various series of salts; the bases for example often range themselves in the same order.The exceptions however are numerous ; and calculations founded on the mean differences between the members of various series of acids and of bases do not fairly accord with the experimental figures in more than half the numberof cases. No relation was found between the specific heat and the greater or lesser tendency of the dis- solved salt to form definite and crystallisable hydrates. The experi- ments fully confirm an observation previously made by others as to the fact of the specific heat of a saline solution being usually less than the sum of the specific heats of the water and of the dissolved salt. In some cases however the inverse condition was noticed. The author admits the existence in the solution of definite and dissociatecl hydrates the proportions of which vary with the quantity of water and xxith the temperature ; and that these changes of constitution are accompanied by a disengagement or by an absorption of heat which diminishes or increases the apparent specific heat.R. R. Coefficient of Expansion of Gases. By D. MENDELEJEFF and N. KAJANDER (Deut. Chem. Ues. Ber. ix 1311).-Gases having ABSTRACTS OF CHEMICAL PAPERS. the same molecular weight have the same coefficient of expansion which increases with the increase of the molecular weight. C. S. Compression of Gases by Low Pressures. By HEMILIAN and BOGUSKY MENDELEJEFF (Deut. Chew,. Ges. Bey. ix 1312).-Air carbon dioxide and sulphur dioxide under a pressure of more t-han one atmosphere are more compressible than required by Boylc's law but less compressible under a diminished pressure ; such positive deviations are shown by air between 20 and 650 mm.by carbon dioxide between 20 and 180 mm. and by sulphur dioxide between 20 and 60 mm. Hydrogen always exhibits positive deviations from the law or contracts less than the law states. c. s. The Laws of the Compressibility and the Coefficients of Dilatation of certain Vapours. Py L. TROOST and P. HAUTE-F E u IL L E (Cowpi. rend. lxxxiii 333).-The vapours of silicic chloride carbonic chloride and phosphorous chloride even at temperatures far above their boiling points contract more under pressure than might be expected from Boyle's law. The following summary shows the contraction for each vapour that is to say the difference between the volume calculated from Boyle's law and that observed when the pressure is increased from + to 1atmosphere.Vapour. Contraction at 100". Contraction at 180°. Silicic chloride ...... 2.07 per cent. .... 0.455 per cent. Carbonic chloride .... 1.38 , .... 1.3637 , Phosphorous chloride. . , .... 1.548 ,, 7 From this it will be seen that the density determined at one tem- perature will vary according to the pressure for each of these vapours. The mean coeffieients of expansion as determined from experi- ments at different temperatures but at the same pressure are as follows :-Mean coefficients of expansion. n --. / Vapour. From 100" to 125O. From 125" to 180". Silicic chloride ......... 0.00449 ........ 0.00399 Carbonic chloride........ 0.00470 ........ 0.00414 Phosphorous chloride .... 0.00489 ........ 0.00417 These variations of the density of a given vapour with the tempe- rature at a given pressure and with the pyessure for a given tempe- rature exhibit the difficulties to be encountered when it is desired to calculate the elastic force which a vapour would acquire in a mixture. C. H. P. Migration of Gases. By F. BELLAMY (Compt. rend. lxxxiii 669-671).-The author has designated by the name of migration the passage of gases through minute channels or capillary tubes such as exist in a bundle of cotton threads a cotton cord a deal sharine &c. When a connection of this kind is made between vessels containing gases at different pressures the gas is aspired into the vessel where GENERAL AND PHYSICAL CHEMISTRY.the pressure is the less. Thus a cotton cord having one extremity passed up into a tube containing mercury inverted over the mercurial trough and the other extremity exposed to the air affords a cmduit for the passage of air into the tube. R. R. Determination of Vapour-densities. By L. T RO o sT and P. HAUTE FEUILLE (Cowpt. rend. lxxxiii 220-223).-The authors criticise the method which some chemists have adopted of deducing the vapour-density of a substance from results obtained with mixtures of the vapour with air or with some other inert gases or vapours. They point out that several causes of error affect these determinations arising from the known inexactitude of Dalton's Iaw the depctrture of the vapour from the law of compressibility and the variation of the coefficients of expansion.They animadvert upon the experiments of Wurtz for the determination of the density of pentachloride of phos- phorus-vapour in mixture with the vapour of the trichloride of which neither the coefficient of expansion nor the law of compressibility was accurately known. R. R. Determination of Vapour-densities in the Torricellian Va-cuum. By A. W. HOFMANN (Deut. Chern. Ges. Ber. ix 1304-1308).-Calibrated and graduated glass tubes are very liable to crack when exposed to a sudden change of temperature. The author has there- fore replaced these by plain tubes and determines the volume of vapour as follows. As soon as the mercurial column remains con-stant the pendulum cathetometer is brought to a level with it the apparatus allowed to cool and after removing the outer glass tube the volume of vapour is marked by a strip of paper.The barometer-tube is now inverted and filled to the mark with mercury the volume of which is ascertained by weighing it. In the usual form of the apparatus the whole column of mercury is not surrounded by the vapour and therefore two corrections have to be made on reducing the height to O" and in so doing it is assumed that the lower part of the column has the temperature of the sur-rounding air and the higher one that of the vapour which is not correct as the temperature of the mercury near the cork gradually changes from the lower to the higher temperature.This source of error is of little conscquence if the vapour-density be taken at a low temperature but at a high temperature it may influence the results. Wichelhaus has therefore proposed a modification of the apparatus by converting it into a syphon barometer which allows the immersion of the whole mercurial column in the vapour. The objection to this is that only one determination at a given temperature can be made whereas by using the conimon form the vaponr density can be ascer- tained at varying temperatures. To retain the latter advantage t,he barometer-tube may be surrounded by a long outer tube which dips into the mercury and to which at about 2-3 cm. above the level of the mercury a tube is sealed t,hrough which the vapour escapes.A much simpler method however is to retain the original form and to rest the barometer-tube on an india-rubber plate which is fixed on VOL. XXXI. D 34 ABSTRACTS OF CHEMICAL PAPERS. an iron disk to which an iron handle is attached On the side of the plate a groove is cut through which during the experiment the mer-cury can flow out. As soon as the height of the column remains con- stant the india-rubber plate is moved so that the groove is removed from the opening which is thus closed. The apparatus is now allowed to cool and the height of the column read off at the ordinary tem- perature. In the origisal form of the apparatus the steam is admitted at the top and this arrangement answers very well if the boiling point of the sub- stance does not exceed 150".But for higher temperatures it is more con- venient to allow the vapour to enter from below. For this purpose the cork of the outer tube as well as that of the copper boiler is provided with two holes. The tube through which the vapour enters begins just under the cork of the boiler and ends about 44-5 cm. above the cork of the outer tube while a second tube through which the condensed liquid flows back begins just above the cork of the outer tube and goes down nearly to the bottom of the boiler. Thus 100-150 ccms. of aniline or any other liquid are sufficient and if the outer tube is about 40 cm. longer than the tube a perfectly constant temperature is obtained in 20-25 minutes by using aniline ethyl benzoate or amyl benzoate.The connections between the outer tube and the boiler are conveniently made of metal and then the tube through which the vapour enters may be provided with a stop-cock which is closed when the boiler is heated. Thus the warm liquid is forced into the space between the outer tube and barometer-tube which is therefore gradually heated. As soon as the boiling point is nearly reached the cock is opened to allow the vapour to enter. In this way the tempera- ture can be kept constant for hours. c. s. A Method of Determining the Vapour-density of Substances boiling at High Temperatures. By VICTOR MEPER(Deut. Chern. Ges. Ber. ix 1216-1228).-The contents of this paper cannot be made intelligible in an abstract without the aid of the accompanying draw- ings.J. R. On the Separation of Mixed Liquids. By E. DUCLAUX (Ann. Chirn. Plays. [51 vii 264-280).-When two liquids dissolved one in the other are induced to separate by an external influence such as cold neither of the layers thereby formed consists of one of the liquids exclusively the other being a mixture but two new mix- tures are formed in which the two substances are distributed in other proportions. At the temperature of 15-20" glacial acetic acid and benzene may be regarded as soluble in all proportions one in the other; but when mixtures of these substances are cooled down to temperatures varying with the proportions of the constituents separa- tion into two layers takes place as shown in the following statement where the volume of the upper layer is placed above that of the lower :-Acetic acid.10 C.C. acetic acid separate at 20.1 c.c. containing 33.3 per cent. 15 , benzene } 15" into { 4.9 , 7 62.8 9 GENERAL AND PEIYSICAL CHEMISTRY. 35 Acetic acid. 10 c.c..acetic acid containing 33.6 per cent. 7, 10 , benzene 63.5 , 15 C.C. acetic acid containing 35 , 10 , benzene ? 62.5 9 A similar separation is effected by the addition to the above mix- tures of a small quantity of water. Mixtures of acetic acid and petro- leum also behave in the same manner. When amyl alcohol ethyl alcohol and water are mixed together in such proportions that the mixture just remains homogeneous at 20° the addition thereto of a slight excess of water causes separation into two layers having the composition shown in the following table where the letters U and L denote the upperand lower layers :--Amyl alcohol I Ethyl alcohol U *my1 ~thy' Water.-. 1 I alcohol. alcohol. L in U. in L. in U. in L. 1 I -1 -5.8 33 p. c. 33 '0 p. c. 2 -5 33 '8 ) 15% ) 30'3 , 30.9 , 0 *67 30.6 , 15.2 , 29'0 , 29'2 , 0 *40 31 *O , 14.4 ) 29.0 , 29'0 , -0 *18 33-0 , 14-8 , -0 -11 35-0 , 12'0 , 26.0 , ' 28.0 , The separation is effect'ed equally by the addition of excess of amyl alcohol instead of water. Such mixt'ures are very sensitive also to slight changes of temperature. The following table exhibits the tem- peratures at which mixtures of the liquids in the proportions there given separate into two layers :-~ ~~~_______ Amyl Ethyl Temp.of U Amjl alcohol alcohol. alcohol. Water. separation. -L' in U. in L. --~-I-100 C.C. 133 C.C. 246 C.C. 30" 0 *44 29.3 p. c. 16.1 p. c. 100 ,> 133 9 219 ?> 20 0.67 30% , 15-2 , 100 9) 133 191 9 10 1.2 . 31.3 , 14'2 ,, 39 100 7 133 9) 164 2 0 1-5 31% , 13.5 , looJ> 133 133 9 -14 1-8 34.4 , 10'3 ,, 97 Mixtures of methyl and am$ alcohol and water or of amyl alcohol acetic acid and water behave in exactly the same manner. A mixture of 5 parts of alcohol of 86" 10 parts of ether and 6 parts of water is perfectly homogeneous; but on adding a slight excess of water (A) the liquid becomes turbid and separates into two nearly equal parts both cont'aining the three liquids. This mixture resembles in its behaviour so far those previously described but it possesses the remarkable property of becoming turbid when heated.A difference of temperature of 0.1" at the critical point is sufficient to cause a sudden separation of the liquid into two laavers having the D2 ABSTRACTS OF CHEMICAL PAPERS. same composition as those produced by the addition of water. This property is exhibited also by a mixture of ether acetic acid and water. The author recommends the employment of mixtures such as are described above for thermometric purposes. The indications of ther-mometers made on this principle may be rendered more apparent by adding to the liquid a drop of red ink. So long as the liquid remains homogeneous it is uniformly coloured ; but when separation takes place the colouring matter becomes concentrated in the lower layer leaving the upper nearly colourless.Such mixtures might be used most advantageously as maximum and minimum thermometers. It is easy to prepare mixtures of nmyl and ethyl alcohol and water which will separate into two layers at any given temperature between -15" (or even lower) and 30". The two layers after separating do not mix again when the temperature rises owing to the difyerence in their densities. Such thermometers need be rery small only 1C.C.of the liquid sufficing for each so that a large number could be arranged in a small compass. J. R. Salt Solutionsand Attached Water. By PR E D E R I c K Gu T H RI E (PkiZ. Mag. [S] i 354-369 ; 446-455 ;and ii 213-225).--When a solution of a salt at or a little below 0" is further cooled one of three things must happen and which of them happens is determined by the strength of the solution.1. In all solutions weaker than the cryohydrate ice is formed at temperatures which are lower according as the solution is richer in salt. 2. In solutions of a certain strength (namely that of the cryohydrate) combination of the salt and water takes place in definite ratio and at a coilstant temperature. The solution is therefore a melted cryohydrate and solidifies as a whole. 3.When solutions stronger than the cryohydrate are cooled below O" either the anhydrous salt or some hydrate richer in salt than the cryohydrate separates. It follows therefore that the cryohydrate is ultimately obtained by cooling either a weaker solution or a solu-tion stronger than the cryohydrate since in the one case ice separates and the solution strengthens while in the other case anhydrous salt sepamtes and the solution becomes weakened.It is proposed (1)to trace the history of solutions weaker than the cryohydrates as they yield ice on cooling ; and (2) to examine the separation of such anhydrous salts or hydrates which separate when solutions richer in salt than the cryohydrates are cooled. A solution of a salt below O" which is weaker than the cryohydrate may be regqrded as a solution of ice in the cryohydrate and just as a given weight of water dissolves as a rule more of a salt t-he higher the temperature so a given weight of the cryohydrate dissolves more ice at higher temperatures below 0" than at lower ones.Again since at any given temperature the quantity of ice dissolved depends on the quantity of cryolijdrate present that is on the strength of the solution tlle stronger a solution is the lower must its tempera- ture be reduced before it can be made to yield up ice. In a similar GENERAL AND PHYSICAL OEIEMISTRY. manner anhydrous salt may be supposed to dissolve in the liquid cryohydrate and to separate therefrom at O" or at temperatures very little below. Solutions of various salts were prepared of known strength by dis- solving a weighed quantity of the salt in a weighed or measured quan- tity of water. Beginning with a 1 per cent. solution it was cooled the temperature noted and the nature of the solid which separated examined.The strength of the solution was then increased regularly by 1 per cent. examined as before axid the operation continually repeated until the solution solidified as a whole ; that is until the cryo- hydrate was obtained. While the solution is weak solidification begins and ice separates out at moderately low temperatures but when a certain degree of con-centration is reached the lowest temperature is obtained and t'he regular cryohydrate crystallises. On continuing to increase the strength of the solution a very slight alteration of temperature is sufficient to cause a deposit of anhydrous salt or of some hydrate containing less water than the cryohydrate so that as the concentra- tion proceeds the temperature at which salt separates may be 40" or more above 0" ; in fact the deposition of salt has then become only a case of ordinary crystallisation.Twenty-three complete tables are given containing the results of examination of as many different salts ; it is not however necessary to reproduce them here as the only point of special interest is the tem- perature at which the cryohydrate forms. The phenomenon of a solution undergoing fractional solidification by cooling may be considered physically as the homologue of the con- centration of a solution by boiling. Thus :-1. A solution weaker than the 1. A non-saturated solution re- cryohydrate loses heat; ice is ceives heat ; vapour is formed. formed. 2. Ice continues to form and 2.Vaponr forms and the tem- the temperature to fall until the perature rises until saturation is cryohydrate is reached. reached. 3. Water is thus withdrawn in 3. Water is thus withdrawn in the form of ice. the form of vapour. 4. At the point of saturation 4. When the solution is satu- ice and salt separate simulta- rated vapour and salt separate neously and the solid and liquid simultaneously and the same portions are identical in composi- ratio exists between the vapour tion. formed and the salt precipitated as existed previously between the water and the salt it held in solu- tion. Separation. of Ice from Mixtures of Salts.-20 per cent. solutions of silver nitrate and ammonium nitrate were mixed together in varying proportions; also 10 per cent.solutions of' ammonium nitrate arid ammonium sulphate and the temperature at which they began to give np ice wits noted. This appeared to be about a mean between the glacia- ABSTRACTS OF CHEMICAL PAPERS. tion temperatures of the constituents. It may be calculated from the equation-t=t,+ where n = grams of A m = grams of B of equal percentage strength and t, tz the temperatures at which A and B give out ice respec- tively. A few organic siibstances were experimented with. Cane-sugar gave a definite cryohydrate at -8.5",but from aqueous solutions of glycerin nothing but ice could be obtained. The results with tartaric acid were rather indefinite but like sodium iodide it appeared to yield two cryohydrates at different temperatures.Dry gum arabic was found to be powerless as a cryogen and in solution it also failed to produce a cry&alline compound on cooling. Albumin in form of white of egg began to separate ice at O" and froze into a solid mass at -0.5". No satisfactory results were obtained by cooling solutions of gelatin but ih was noticed that a strong solution of 50 per cent. boiled steadily at 97". J. W. On the Elasticity of Metals at Various Temperatures. By G. PISAT I (Gazzetta chi?nicaitaliana vi 23-32) .-After noticing the researches of Coulomb Wertheim Kupffer and Kohlrausch and Loomis the author describes the apparatus employed in which the elongation of two equal portions of the same wire under various ten- sions and at various temperatures could be accurately measured by means of a cathetometer.On examining an iron wire which had pre- viously been thoroughly annealed by heating it to dull redness and allowing it to cool very slowly in this instrument it was found that with a tension of lS00 grams and on gradually raising the tempera- ture to 300" the two halves expanded very irregularly and a similar phenomenon was observed on cooling However after the operation of alternate heating and cooling had been repeated several times the wire was reduced to the normal state ;that is the rates of expansion of the two halves of the wire coincided perfectly. The weight on one of the halves was now reduced by 1000 grams so that one was loaded with 800 the other with 1800 grams and the difference in elongation observed at 18.1" and at intervals of 50" from 50" to 300' ;the length of each half of the wire when stretched by a weight of 800 grams was 1778.22 millimeters and its diameter 0.4236 millimeters at 28.1".Similar experiments were made with a steel wire the two halves being loaded with 1000 and 4000 grams respectively the length of each half when stretched by a weight of 1000 grams was 1782.69 at 23*3" and its diameter at 14.8" was 0.4940 millimeter. The results obtained in these two sets of experiments were as follows :-Increase in length in millimeters. Temperature. Iron. Steel. ........ - 18" ................ 0.588 22 ................ -........ 1.510 50 ................ 0-590 ........ 1.519 100 ................ 0.594 ........1.589 GENERAL AND PHYSIOAL CHEMISTRY. Increase in length in millimeters. Temperature. Iron. Steel. 150 ................ 0.602 ........ 1.543 200 ................ 0.615 ........ 1.563 250 ................ 0.633 ........ 1.582 300 ................ 0.656 ........ 1.601 the readings giving the increase of length produced by the increased load on the one wire at the various temperatures. From the results it will be seen that the modulus of elasticity diminishes as the tempera- ture rises up to 300" C. whilst Wertheim (AY~ Chirn. Phys. [3] xii 385) found that it increased between -15" and ZOOo which is per- haps to be attributed to the latter not having brought the wire into the "normal state" previous to making measurements.If K be the modulus of elasticity of tension that is 10000 times the weight necessary to produce an elongation of 0.001 in a wire 1meter long and of section equal to 1millimeter square- I(=-P.L s.1 where the weight P produces an elongation Z in a wire whose length is L and section s. If moreover the wire be cylindrical and at the tem- perature 0" have a length of Lo,and its section a radius r0,its coeffi- cient of linear expansion being a we have- which gives the modulus of elasticity at the temperature to. Adopting Fizeau's values for the coefficient of expansion namely- a = 0.00001228 for iron a = O*OOOO1112for steel the following are the moduli of elasticity for iron and steel deduced from the author's results :-Temperature (cor.).Iron. Steel. 20" .............. 21441 ...... 18481 50 .............. 21364 ...... 18416 100 .............. 21212 ...... 18232 150 .............. 20895 ...... 18052 200 .............. 20458 ...... 17820 250 .............. 19871 ...... 17593 300 .............. 19175 ...... 17372 C. E. G. Elasticity ofTorsion. By G. P ISATI (Gazzetta chimica italiana vi 57-88).-After alluding to the labours of Kupffer Weber and others in this subject and to the fact that they had not examined the effects of torsion at temperatures above loo" the author describes the apparatus employed by himself and also gives the details of the methods. The apparatus a figure of which is given is very similar ABSTRACTS OF CHEMICAL PAPERS. to that used for the experiment on tension but the oil-bath and the tube containing the wire is only about 0.8 met.in length ; the wire for the experiment being about 0.65 met. long and 0.25 to 0.50 mm. in diameter. The met,hod employed was that of oscillations measuring the number and duration of the oscillations occurring whilst the ampli- tude of the oscillations gradually decreased from 90" to 10". In the first experiment a silver wire 0.468 mm. in diameter and 643.38 long was employed stretched by a weight of 309 grams obser- vations of amplitude being registered at the end of each 50 oscilla- tions. On repeating the experiment with the same wire a second third fourth time &c. it was found that the time occupied whilst the amplitude of the vibrations decreased from 90" to lo" .gradually aug- mented whilst the duration of a single oscillation diminished.After the twentieth experiment however no further change took place (the temperature remaining constant at 26.5"). From this it will be seen that a phenomenon took place similar to that observed in the experi- ments on tension namely that during the oscillations the wire gradually underwent alteration until finally it arrived at a normal state for the special conditions under which the experiment was made. In these experiments it was found that at first the time required to reduce the amplitude of thevibrations from 90" to 10" (vzt) was 15' '22.9"; n = 350 being the number of oscillations in this time and therefore the mean duration of an oscillation (t) was 2-637" whilst after the wire had reached the normal state these numbers were n = 450 vzt = 19' 42-1" and t = 2.6269'' respectively so that the mean duration of a single oscillation had been diminished in the proportion 1.0038 1.A fresh set of experiments was then made. The wire was heated to 100" for an hour allowed to cool and examined when it was found that the time required to reduce tlhe amplitude had again increased whilst the duration of a single oscillation had diminished. On repeat-ing this treatment several times these numbers at length became constant. An oscillation observation was then taken at loo" and also after the wire had cooled to 26". The numbers obtained at 100" were n = 100 m! = 4' 26*8" and t = 2.668" ; and at 26" n = 650 mt = 28' 09.7" and t = 2.5995".On now repeating the experiment several times the results obtained each time were found to be identical so that the wire had reached a second normal state differing from the first normal state in the same character and in the same direction in which the latter differed from the natural state. On increasing the temperature to 200° similar phenomena were observed the numbers for the normal state under these conditions being :-n. t. nt. At ........ 200" 30 2.7'50" 1',22.5" Cooled to . . 25.4" 1000 2.512" 41',52-0" It was then heated to 305" cooled to 26" and a series of oscillations observed these operations being repeated until constant results were obtained. The numbers at 25.8" for this third normal state were n = 1450 t = 2.417" (about).It wits found moreover that when the wire was relieved from the GENERAL AND PHYSICAL CHEMISTRY. tension to which it had been subjected even for so short a time as 30" an alteration in the elastic state was produced the duration and number of the vibrations being temporarily altered ; but that after the wire had been caused to oscillate for a long time stretched w-ith dif- ferent weights and at different temperatures it was finally reduced to a normal state such that wen when the tension was varied the same number of oscillations n were always observed the temperature remaining constant. A series of experiments was then made to ascertain the variation in the elasticity corresponding to variation of temperature the details of which are given.Under these circumstances the number of oscilla-tions 12 decreased rapidly up to 160" and then much more slowly up to 300° while the duration of a single oscillation gradually increased. On measuring the wire it was found to have slightly increased in length and diminished in diameter its dimensions now being 647.48 and 0.4661 millimeters. Modulus of the Elasticity of Torsion at Various Temperatures.-If E is the modulus of elasticity of torsion referred to a square millimeter in section where M is the moment of inertia of the tension (equal to 13608 grams per square millimeter in the experiment) ; t the duratioii of a single oscillation ; L and r the length and radius of the wire ; and g the accel- L leration of gravity expressed in millimeters.As -varies with the r4 temperature if a is the coefficient of linear expansion and T the tem- perature (1)becomes I?-M Lo 1 E=-.-. -g t" r,l' (1 +cxT)~ where Lois 647.48 and yo 0.233 millimeters and a = 0*000*0187+ 0*000~000~010 T. In the following table the number of oscillations during which the amplitude was reduced from 90" to 10" is given and also the modulus. Number of Modulus of Temperature. oscillations. torsion. 0" .......... -1644.5 10 .......... -1637.8 20 .......... -1630.9 30 .......... 1620 1623.8 40 .......... 1275 1616.4 50 .......... 1000 1608.6 60 .......... 780 1600.3 70 .......... 590 1591.5 80 .......... 445 1582-1 90 .......... 348 1572.0 100 ..........277 1560.9 110 .......... 218 1548.5 120 .......... 172 1535-0 ABSTRACTS OF CHEMICAL PAPERS. Number of Modulus of Temperature. oscillations. torsion. 130 .......... 135 1520.5 140 .......... 104 1502.5 150 .......... 80 1481.0 160 .......... 62 1456.0 170 .......... 48 1429.0 180 .......... 40 1400.5 190 .......... 32 1371.0 200 .......... 30 1341-2 210 .......... 28 1311.4 220 .......... 26 1281-6 230 .......... 24 1251.8 240 .......... 22 1222.0 250 .......... 80 1192.2 260 .......... 18 1162.4 270 .......... 16 1132.6 280 .......... 14 1102-8 290 .......... 12 1073.0 300 .......... 10 1043.2 Curves corresponding with these numbers and with their differences accompany the paper.C. E. G. Retardation of Chemical Reactions by Indifferent Sub-stances. By G. LUNGE(Deut. Chem. Ges. Ber. ix 1315-1316).-A mixture of equal volunies of fuming hydrochloric acid and glycerin (a) acts on ultramarine only after 45 seconds and bleaches it in 3 minutes while a mixture of equal volumes of water and acid (b) begins to act in 10 seconds and destroys the colour in 35 seconds. Mixture a dissolves zinc and iron much more slowly than b. Thus 10 C.C. of the latter dissolved 0.5 gram of nails in less than 24 hours while a left after 24 hours 86.2 per cent. undissolved and after 14 days 1.3 per cent. still remained. The cause of this is not that ferrous chloride is less soluble in glycerin than in water because the salt readily dissolves in the former and during the experiment none separated out.Mixtures of sulphuric acid and glycerin show a similar inactivity and in- stead of glycerin gum may be used. A mixture of acid and soot scarcely acts on metals but on removing the soot by filtration the filtrate acts like fresh acid. The retardation reaches a maximum by using a mix- ture of strong acid and glycerin with 5 per cent. of soot; iron nails lost in it-1. 2. In 3 days ...... 10.8per cent. ...... 21.2 , 6 days ...... 25.4 , ...... 13.0 , 14 days ...... 51.0 , -c. s. The Point of Combustion. By A. MITSCHERLICH (Deut. C7~em. Ges. Ber. ix 1171-1178).-By this term the author understands the temperature at which a body begins to take up free oxygen. This may be a simple oxidation or a decomposition may take place at the same INORGANIC CIFEMISTRY.time and the process may proceed only gradually with a slow evolu- tion of heat or quickly with a violent evolution of heat and light. The author describes the apparatus and the method which he used for these determinations but these can only be understood with the help of a drawing. The results which he obtained will be published afterwards. c. s. Contributions to the Theory of Luminous Flames. By KARL H E u MA N N (Liebig's Annalen 183 1-29) .-In this second contribu- tion the author considers specially the circumstances which determine the distance existing between a luminous flame and the orifice whence it issues. His principal conclusions are as follows :-1.The fact that a space occupied by unburned gas exists between a gas flame and the burner or between a candle flame and the wick as also the fact that a dame does not actually touch a cold body placed within it is due chiefly to the cooling action of the surroundings of the flame whereby the temperature of the gases is reduced below the ignition point. 2. The great distance existing between the flame and the burner in the case of a gas issuing under high pressure as in the case of a gas largely diluted with indifferent gases is to be traced partly to the cooling action already mentioned but more especially to the fact that the rate of propagation of ignition in t'he neighbourhood of t,he burner is less than the rate at which the gaseous stream issues.3. In order to remove other conditioning circumstances the rate of propagation of ignition shonld be maintained equal to the rate at which the gas flows at that point situated some distance from the burner at which the flame begins. The rate of propagation of igni-tion will be determined under such conditions for many gases and vapours ; and inasmuch as this magnitude is a function of the differ- ence between the temperatures of ignition and combustion of the com- bustible body it is hoped that much light m7ill be thrown on the relations existing between these two. 4. The rate of propagation of ignition may easily be determined for solids and liquids and comparative quantitative expressions for the liability to ignition of combustible substances may thus be obtained. M. M. P. M.