]\gt ; 1890 .
] iments on density 45 Figs. 7 and 8 represent these values of , and , respectively , the curves marked 1 being those obtained with tin at 65 ; those marked 2 with tin at 800 ; and those marked 3 with silver at so ; the value of being abscissae , and those of and ordinates .
have much pleasure in acknowledging the assistance afforded us by Mr. T. M. Wyatt in carrying out a considerable portiolt of the analytical resuIts above detailed .
Il .
' Experiments on Vapour-density.\ldquo ; By E. P. , B.Sc. , Clothworkers ' Exhibitioner at University College , London .
Communicated by Professor RAMSAY , F.R.S. Received April 17 , 1890 .
The of The main purpose of my work on this subject was to discover if bromine had any tendency to dissociate at low pressures , i.e. , dowm to about 15 mm. of mercury , and moderate temperatures .
The ubject was ested by Professor Ramsay , to whom I am greatly indebted for constant advice and assistance in carrying out the work .
The method adopted for determining the vapour-densities was a modification of the Dumas method , from which it differed in three essential points : The globe was not immersed in a bath , but had a vapour-iacket .
2 ) The weight of the vapour in the globe was not found by direct weighing , but by running in an liquid , and estimating it volumetrically .
( 3 ) A series of vapourdensity determinations at different pressures was made with the same quantity of vapour by lowering the pressure , absorbing the vapour drawn off and estimating it quantity .
As to previous work on this subject , Jahn has shown ( ' Wien , Akad .
Sitzber vol. 85 , 2 .
Abth 1882 , p. 778 ) that the vapour-density becomes normaF at about ; Meier and Zublin ( ' Deutsch .
Chem. Ges .
Berichte , ' vol. 13 , 1880 , p. 405 ) and Crafts ( ' Comptes Rendus , ' vol. 90 , 1880 , p. 183 ) have shown that partial dissociation takes place at very high temperatures ; and Professor J. J. Thomson states that " " vapour-density determinations showed that bromioe vapour is dissociated if it is heated for a time at a low pressure , even though the temperature is not very high\ldquo ; ( ' Roy .
Soc. Proc vol. 42 , 1887 , p. 345 .
His chief results .
l\mdash ; Apparatus used tor determining the Vapour-density of Bromine .
( Section .
) globe , B. The liquid which produced the vapour.jacket was boiled in a smalI globe , , connected with the outside globe , and condensed in the tube rollnd which , when necessary , a condenser was fixed .
Further to the right came an absorption-tube , , with a small bulb-tube , , ( iontaining a few drops of the same liqnid as vras introduced iuto the absorption-tube , viz. , potash solution during boiling-out process , and afterwards a strong solution of potassium 1890 .
] Experiments on Vapour-density .
47 iodide ; then came an air-reservoir , , and a pressure-gauge , arranged as shown in the diagram .
There were stopcocks at ; rubber joints at , and ; and a screw-clip at on the tube , leading to a water-pnmp .
Method of Procedure .
\mdash ; The capacity of the globe A was urately determined by weighing the apparatus empty , and then full of water ; it was filled from a wash-bottle , was also weighed as a check on the other result .
Pure bromine was specially prepared by boiling commercial bromine with potassium bromide for a day , distilling it throngb a tube containing red-hot manganese dioxide , and redistilling with phosphoric anhydride ; after a small portion had distilled over , the point was constant at ( reduced to 760 mm Asbestos plugs were found very useful im working with bromine ; by their use a good joint can be secured if liquid bromine is not allowed to come into contact with the asbestos .
The thermometer in the distillingflask was secured by a plug of asbestos , and its bulb was sulrounded with asbestos to prevent superheati-ng .
To commence an experiment , the globe was rinsed with ether , and dried by repeated exhaustion and admission of air ; some bromine was run in without admitting air , the tube above was fnsed on to the absorption-tube , , by means of a portable blowpipe , the liquid in was boiled , the soopcock cautiously turned till the pres- sure within the globe was equal to the atmospheric pressure , and then opened from time to time till the bromine ceased to blow ouf .
All the bromine was absorbed in , which at this stage contained potash solution .
The stopcock and the tube above and below it were gently heated by means of a Bunsen 's burner .
The absorption-tube was then connected with the , and the air was partially exhausted from the whole apparatus ; to complete the exhaustion as far as possible , the globe and the absorption-tube were directly connected with the pump .
The stopcock was then turned off , the tube above it was cracked and removed , more bromine was admitted , the tube fused together again , and the bromine driven out as before till the pressure in A was equal to the atmospheric pressure .
The globe was then full of bromine vapour at the atmospheric pressure , and at the cemperature of the jacketing vapour , and was ready for a vapour-density determination , or for a series of determinations .
The absorption-tube was rinsed out , partly filled with a strong solution of potassium iodide ; it was then clamped in its place , and connected with the globe , and with the reservoir M. The water-pump was then set to work till the reservoir was exhausted sufficiently\mdash ; the degree of exhaustion depending upon the number of nsity determinations which were to be made in the series .
The pump was then cut off by the screw-clip , and Mr. E. P. Perman .
[ May 8 , he stopcock turned very cautiously , letting out the bromiue vapour little by little ; when it ceased to come over , the globe and its contents were left for two or three minutes in order to regain the heat lost by the adiabatic expansion ; it was necessary to repeat this process till the bubble in the bulb-tube remained undisturbed on quickly and the stopcock E. recorded by the gauge was then carefully read , the stopcock was closed , the absorption-tube removed , and its contents washed into a stoppered bottle .
The whole process was repeated at successively lower pressures till the lowest desired was reached ; a good fvater-pump will reduce the pressure to the vapour-pressure of water at the temperature of the water passing through the pump .
residual bromine in the globe was mated by running potassium iodide solution , and titrating with a standard solution of sodium thiosulphate .
The successive of broInine collected in the absorption-tube were also estimated in the same way .
The hiosulphate solution was standardised with the specially prepared bromine , and also with pure iodine ; the results agreed within per cent. , and the mean value was used .
By adding together the residual quantity of mine and the quantities removed , the weight of bromine in the globe at each pressure was found .
The capacity of the globe being accurately known , and the temperature of the jacket from the tables of Professors Ra , msay and Young ( ' Chem. Soc. Journ vol. 47 , 1885 , p. 640 ) , the data requisite for the determination of a series of vapour-densities at different pressures were available .
Resnlts.\mdash ; After some practics in conducting the experiment , satisfactory results were obtained ; they are tabulated below:\mdash ; Series I.\mdash ; Temperature , ( Alcohol ) ; Capacity of Globe , Some bromine escaped absorption , causing the first vapour-densities to be too low .
1890 .
] iments on -density .
Series II.\mdash ; Temperature , The only noticeable point in this Qeries is the highness of the last vapour-density at the lowest pressure .
A single experiment was then made in order to try the effect of continued heating on bromine vapour at a low pressure .
After three hours ' heating at and mm. pressure , the vapour-density came out abnormally high , .
Professor J. J. Thomson found vapour-densities and 70 at much higher pressures and somewhat higher temperatures than these .
Series III.\mdash ; Temperature , ( Bromobenzene ) ; Capacity of Globe , Here there is still no sign of dissociation , and what seems somewhat remarkable , even at low pressures the vapour-density is above the normal density .
The bromine being then exhausted , a 1fresh sample was prepared from the residues and from some commercial bromine treated with potash ; the whole was evaporated to dryness , ignited , distilled with acid and potassium bichromate , and dried by shaking with sulphuric acid , allowing to stand , and -distilling ; the greater portion came over at ( corrected to 760 mm it commenced to boil at , but the boiling point soon rose to , and then remained constant .
A determination was then 1nade of the density of bromine vapour as nearly as possible saturated at .
The globe was immersed in VOL. XLVIII .
Mr. E. P. Perman .
[ May 8 , a pan of water , and the pressure was made so near the saturationpressure that the bromine took three hours to vaporise completely .
Series Temperature , ; Capacity of Globe , The vapour-pressure of bromine at is mm. , according to the tables of Professors Ramsay and Young ( ' Chem. Soc. Journ vol. , 1886 , p. 445 .
These results show that , on approaching the liquid state , bromine has no tendency to form molecules with more bhan 2 atoms .
This agrees with the results of Paterno and Nasini Deutsch .
Chem. Raoult 's method , which indicated molecules Br : in aqueous and acetic acid solutions .
Series .\mdash ; Temperature , 27 5 thalene ) ; Capacity of Globe , The temperature employed in this series was much higher than the temperatures used by Professor J. J. Thomson in his experiments ; some of the pressures were much lower , and the total time of heating was about five hours , yet there was no sign whatever of any dissociation .
Finally , an experiment was made with bromine in presence of air .
The was boiled out of the globe at exactly half the atmospheric pressure , and air was then slowly admitted until the pressure inside and out was the same .
1890 .
] xperiments on Vapour-density .
Series .\mdash ; Temperature , ( Chlorobenzene ) ; Capacity of Globe , .
; Volume occupied by Bromine Vapour , c.c. The results are negative , as before , but no doubb dissociation would be produced if the temperature were carried high enough , and I hope to carry on my experiments with a modified form of apparatus until dissociation takes place .
I must ask my fellow-workers to kindly leave to me the completion of this work .
Remarks.\mdash ; One of the most striking features in the series of vapour-densities is the invariable increase in density at pressures ; this may be accounted for partly by the unavoidable in reading low pressures , and partly by the irregularity in the position of the liquid in the bulb-tube ; but thess errors together would never amount to more than 1 mm. of mercury , seldom so much .
As the chief cause of the anomaly , I would suggesb that a film bromine may adhere to the glass , and thus cause the residual amount of bromine to be too large .
Vapour-density of Iodine .
Previous Work . .
Meyer obtained vapour-densibies closely corresponding to a molecule , at a temperature which he estimated at ( Deutsch .
Chem. Ges .
Ber vol. 13 , 1880 , p. 394 ) .
Crafts and Meier found a greater amount of dissociation than V. eyer at about the same temperature , which , however , they estimated at ] 39 instead of ; the lowest vapour-density they obtained was ( ' Comptes Rendus , ' vol. 90 , 1880 , p. 690 ) .
Deville and Troost fouud the vapour-density normal at and 1034o ( ' Annals de Chimie , ' vol. 58 , 1860 , p. 257 ) .
Crafts and Meier , in further experiments , obtained complete dissociation of iodine vapour , and they give a curve showing the amount of dissociation at differeri pressures and temperatures ( ' Comptes Rendus , ' vol. 92 , 1881 , p. 39 ) .
Crafts obtained partial dissociation by a modification of .
Meyer 's method ( ' Comptes Rendus , ' vol. 90 , 1880 , p. 183 ) .
Troost obtained a very considerable amount of dissociation by the Dumas method at ; the lowest number obtained was ; he also found the Mr. E. P. Perman .
[ May 8 , vapour-density at and mm. pressure to be ( ' Comptes Rendus , ' vol. 91 , 1880 , p. 54 ) .
The dissociation of iodine has thus been worked out with fair completeness , although the resuIts of FIG. 2 .
FIG. Forms of Apparatus used for Iodine and Aqueous Hydrochloric Acid .
( Section .
) Deville and Troost seem very much variance with fhose of other experimenters .
The object of my work was to discover the density of iodine vapour as ] as possible saturated .
A modified form of apparatus was devised by ProfessoI ' Ramsay so that the iodine could be conveniently introduced into the globe .
The modification consisted of the introduction of a stopper at the neck of the globe with a capillary tube throug.h the centre .
The method of procedure was the same as with bromine , except that it was necessary to heat the Temperature , ( Chlorobenzene ) ; Capacity of Globe , 21. .
1890 .
] Experiments on Vapouz-density .
tube and stopcock next the globe to a much higher ( emperalure , in order to volatilise the iodine ; consequently there was great risk of cracking the stopcock , and this indeed happened several times .
The first ( trial ) series gave resuIts pointed towards normal In the next series an attempt was made f , o find the vapour-density as nearly as possibIe at the saturation-pressure .
The vapour-pressure of iodine at 13 is mm. , from the results of Professors Ramsay and Young ( ' Chem. Soc. Journ vol. 49 , 1886 , p. 455 ) , and several determinations made at about this temperature and at about 150 mm. pressure .
Constant resuIts were not obtained , although much time and trouble were expended on the experiments .
The numbers obtained were : , and .
The iodine was allowed to blow out until none could be seen on the bottom of the globe , but it was extremely difficnlt to decide whether it had all volatilised or not , and no reliance can be placed on the results , notwithstanding the close agreement of the first two numbers with 254 .
Another plan was then adopted .
-density of IodrinDeterminations of the Velocity of Sound by 's Method ( ' Poggendorff 's Annalen , ' vol. 135 , 1868 , p. 337 The apparatus is best understood from the ( fig. 4 ) .
The glass tube , about 2 feet long and inch in diameter , was supporbed inside a larger tnbe , , by corks , as shown ; CD was slightly FIG. 4.\mdash ; Sectiou of Apparatus .
inclined one way , and GH the orher .
The bulb contained chlorobenzene ; is a condenser ; AB is a small-sized tube sealed into the tube CD at ; it was enlarged at , so as to fit the tube not to touch the walls .
lfethod of Procedure.\mdash ; Some solid powdered iodine was introduced Mr. E. P. Perman .
[ May into the tube CD , and shaken down to the end C. Some finelydivided precipitated silica was then introduced , distributed along .
The air was expelled by heating the tube with a Bunsen 's burner , the same time it by means of a water-pump ; the end was then sealed off .
The tube CD was then placed inside the vide tube , as shown in the figure ; the chlorobenzene was and when the whoJe tube had been heated for a few minutes , the projecting tube AC was rubbed with a cloth moistened with The whole tube AB vibrated , the iodine vapour in BD was thrown into vibrations , and the silica on the lower part of the tube formed little heaps at the nodes , and streaks perpendicular to the length of the tube at the loops .
The tube was then allowed to cool , withdrawn from the larger tube , and placed on a horizontal millimetre scale ( an etched mirror scale was used ) .
It was easier to determine the position of the loops than of the nodes when silica was used .
The mean distance between two loops was then calculated .
Similar ex. periments were made with air .
Let mean distance between two loops for air .
, , , , dine vapour .
density of air .
, , iodine vapotlr .
for air .
, , for iodine vapour .
Then it follows from the laws of the propagation of sound that Results .
was found to be ( 1 ) , ( 3 ) , eleven adings being taken in each experiment .
The mean of ( 2 ) and ( 3 ) was taken as the correct value of , these being the most trustworthy experiments .
The results obtained for iodine vapour are here tabulated : \mdash ; 1890 .
] Experiments on Vapou -density .
Experiments ( 1 ) and ( 2 ) are not reliable , as the vibrating glass tube gave a different note from that produced in the experiments on air ; it gave one of two notes , according to the way in which it was clamped .
The interval between these notes was estimated by ear as a fourth , but it could not be determined accurately owing to the highness of the notes , which appeared also to vary slightly in pitch .
The mean of experiments ( 3 ) , ( 4 ) , and ( 5 ) is .
In the calculation , was taken as ( Rontgen ) , and as ( Strecker ) .
These resuIts show conclusively that saturated iodine vapour , and therefore liquid iodine also , are composed of diatomic molecuIes , .
This result agrees with the late work of Ernst Beckmann ( ' Zeitschrift fur Physikal .
Chem vol. , 1890 , p. 76 ) on the molecular weight of iodine in solution in ether and carbon bisulphide ; the numbers varied from 235 to 261 for solutions in ether , and from 263 to 283 for solutions in carbon bisulphide .
A single experiment was performed with bromine in the same manner ; the mean distance between the heaps of silica , mm. at a temperature of about .
Compared with air at the same temperature , this gives a vapour-density .
This points to a normal density , as would be expected .
There was considelable uncertainty as to the temperature , because no heaps were formed while the [ )romine vapour was saturated , and the tube was therefore heated with a Bunseu 's burner .
Iodine vapour refused to give heaps in the same way when left long enough in contact with liquid iodine to become saturated .
It was predicted by Professor Fitzgerald , in a private letter to Professor Ramsay , that sound would not be propagated through saturated vapour , because the rate of propagation of a compression would not be the same as that of an expansion , owing to the difference in elasticity in the two cases .
Either the expansion or the compression of a sound-wave would produce condensation in any saturated vapour ; in the case of saturated steam , expansion causes condensation , except at high temperatures ( above the " " temperature of inversion At present there are not sufficient data to calculate the effect of a wave of compression or expansion upon saturated iodine vapour ; but condensation would take place either during a compression or an expansion , and the ratio \mdash ; and therefore the elasticity , \mdash ; would be quite different in the two cases .
The sound-waves would interfere with one another , and become confused , so that no sound would be propagated through the vapour .
Induction Spark through Iodine \mdash ; An expel.iment was then made determine if the passage of the spark through iodine vapour eflected its dissociation .
Professor J. J. Thomson obtained vapour-densities 1.37 and 130 for unsparked iodine , and 110 , 115 , 84 , and 86 fcr sparked iodine , the last vapour-density being determined twenty-four hours after sparking .
Professor Thomson .
E. P. Perman .
[ May 8 , says Proc , " " The appearance of the dissociated iodine is not greatly different from that of the unsparked ; its colour , however , is , I think , a little lighter , and not so uniform .
I was not able to detect any change in the absorprionspectrum produced by the sparking .
The electric strength of the sparked was , however , less than that of the unsparked The same apparatus was used , platinum wires , fused into the glass ( see fig. 4 ) .
The jacket used was methyl salicylate vapour , temperature 5 ; the iodine was boiled out at the atmospheric pressure , and the tube then sealed at D. The mean distance between the ps without sparking was mm. , and after sparking fifteen minutes , and cooling for three minutes ( to allow the vapour to regain the temperature of the jackeb , the mean distance was mm. Another experiment was then made , so that the heaps were formed during the sparking ; the mean distance between the heaps was mm. on the side of the piston mear the vibrator , and mm. on the other side ; the diiference is probably due to the heating effect of the induction sparks .
The results may be tabulated thus : Mean distance between heaps .
Before sparking .
After sparking .
During sparking .
14.2 Suppose that complete dissociation of the iodine vapour took place on sparking , then the pressure would be doubled , and therefore the elasticity would be doubled ; bnt the density ( taking no account of the pressure would remain unaltered .
Let velocity of sound in the undissociated iodine vapour .
, , completely dissociated ditto .
elasticity of the undissociated vapour .
density of the iodine vapour ( same in both cases ) .
Then and But the distance between the heaps of silica ( half a wave-length ) is directly proportional to the velocity of sound in the vapour ; there- fore , if any dissociation took place , there would be a marked increase in the distance between the heaps on sparking .
The results given above agree within the limits of the errors of 1890 .
] Experiments on Vapour-density .
experiment , and it is concluded , therefore , that sparking does not produce any permanent dissociation , altbough very probably a momentary dissociation is caused near the tsrminals when a spark passes .
The length of the spark in the iodine vapour was about 1 inch ; the coil was capable of giving a 3-inch spark in air .
By opening the tube under water , it was found to contain about 5 per cent. of air ; this would tend to aid dissociation , and would in no way lessen the value of ths experiment .
The probable experimental error abour 1 per cent. , and was caused by the difficulty in reading the distance between the loops as indicated by the arrangement of the silica .
-density of Sulphuric dride .
The apparatns used to determine the vapour-density of sulphuric anhydride was similar to that used for iodine .
The only work that has been done on this subject is apparently that by SchultzSellack ( ' Poggendoffi 's Annalen , ' vol. 139 , 1870 , p. 480 ) and Mitscherlich ( ' Watts 's Dictionary who both found the vapourdensity normal .
Some of the trioxide was introduced into a small flask , the neck of which was then sealed on to the tube from the inner globe ; it was then sublimed into the globe , which had been previously exhausted , by heating the flask .
The vapour-density was then determined in the usual way .
The trioxide was absorbed by water , and estimated by means of a standard ammonia soIution .
Using cochineal as indicator , it was found possible to titrate accurately to one drop of a decinormal solution in the following The relative value of seminorma,1 ( approximately ) soIutions of acid and ammonia was found , and these solutions were then diluted to decinormal strength ; a neutral solution was then made by using decinormal acid and ammonia in the proportion found for the seminormal solutions , and this was used as a standard neutral solution , with which the solutions titrated were compared .
The chief difficulty was the formation of sulphuric acid above the stopcock ; it was partially obviated by introducing a horizontal bulb-tube above the \mdash ; this retained most of the acid .
Only one series of experi- ments was made ; the first result is very nearly at saturation-pressure .
Calc.normalV .
on Vapour-density .
[ May It was useless to estimate the residual amount of trioxide , as sulphuric acid had collected in the lower part of the globe while the trioxide was being sublimed into it .
The vapour-density at the lowest pressure was therefore assumed to be normal ( corresponding to , as it wouId be in all probability , and on this assumption the other vapour-densities are also normal .
If the third vapoul-density were actually above normal , the second would be less above and the first still less ; while , if the third vapour-density were actually below normal , the others wouId also be below normal ; these alterna- tives are very improbable , and I it as proved that the formula of sulphuric anhydride is , and not of Aqueous Hydrochloric Acid .
Commercial hydrochloric acid was distilled until the boiling point became constant , at ( pressure , 745 mm The distillate was then collected and used for vapour-density determinations ; to find the percentage of real acid , weighed quantities were titrated with a standard soda solution .
The acid used in Series I contained per cent. ; that in Series II , per cent. ; and Series III , per cent. The last two samples were distilled from pure acid .
The vapour-density was found by means of the apparatus used for iodine .
The acid coming over on lowering the pressure was absorbed by a standard soda solution in the absorption-tube ; much of it , how ever , condensed above the stopcock , and caused some difficulty .
ResuIts .
Pressu .
Theor .
1890 .
] On Barometric Oscillations during hunderstorms .
There was some loss in Series I and Il in removing the acid from above the stopcock ; this was done by aspirating air through the which was heated by a Bunsen 's burner , and absorbing the acid in a soda solution ; but an acid fume was , which refused to be absorbed .
In Series III , the apparatus with the jacketed stopcock was used , and the acid which collected above the stopcock washed out .
The vapour-density in the extreme right-hand column is calculated for a mixture of hydrochloric acid and water in the pro- portions found by titration .
These results show that no compound of acid and water is formed , at least at the temperature employed ; they confirm the resuIts of Bineau , who found a vapour-density at the atmospheric pressure Annals Chimie , .
III .
" " On Barometric Oscillations during Thunderstorms , and on the Brontometer , an Instrument designed to facilitate their Study By G. J. SYMONS , F.R. .
Received Apri124 , 1890 .
The fact that a rise of the barometer occurs during thunderstorms has been supposed by many to be newly discovered through the general establishment of self-recording barometers ; but Dr. Hellmann has shown that it was noticed by J. J. Planer as far back as 1 In , Rosenthal epitomised the facts as follows:\mdash ; ' ' Whell a thunderstorm approaches the place where a barometer is situated , the mercury in the tube begins to rise ; the nearer the thundel'-cloud comes to the zenith of the observer , the higher does the mercury rise , and it reaches its highest point when the storm is at the least disbance from the observer .
As soon , however , as the cloud has passed the ze1lith , or has become more distant from the observer , the weight of the atmosphere begins to decrease and the mercury to fall A few years later , Toaldo determined the amount of the rise in several storms , and found it to be between 1 and 2 lines in .
to in Professor Strehlke ( in 1827-30 ) made several sets of observations , and found the rise to be from in .
to in .
, and was probably the first to point out that the highest point of the barometer is not absolutely synchronous with the passage of the centre of the stormcloud , but seems rather to be always at a certain distance it .
Kaemtz , in his Lehrbtlch ' ( 1832 ) , suggests that the rise is produced by the inrush of air towards the site of the storm , is accumulation causing the rise of the barometer as the storm nears the zenith .
Although Luke Howard had a recording barometer at work in the early part of this century , he seems to have failed to notice the

