]\gt ; On the Index of of Liquefied Gases .
339 when ]arge vessels of complicated form , like those used by me , have to be put together and exhausted .
All photographs were taken by him .
I have also to acknowled ge valuable assistance from Mr. Moss , a student at the Owens College .
On the Experimental Determination of the Index of Refraction of Liquefied Gases By Dr. L. BLEEKRODE .
Communicated by Dr. GLADSTONE , F.R.S. Received and read June 19 , 1884 .
I. Remarks .
On a previous occasion*I had the honour to present to the Royal Society some results of an investigation made about electrical conductivity of chemical compounds , and I then chiefly examined the liquefied gases , pointing them out as very bad conductors .
I have since been engaged in studying another property of these substances , and I have succeeded in determining in an experimental way their refractive power .
As in England and abroad several papers have been subsequently published bearing on the relation between the liquid and gaseous conduction of matter , and especially liquefied gases present themselves well adapted to kind of research , I hope the Royal Socisty will consider my paper not devoid of interest , the more so bscause our knowledge of their physical constants is somewhat limited .
And though we possess numerous determinations of the refractive power of a vast number of chemical compounds , still increasing daily , I have found only very little information concerning my subject , this being limited to sulphurous and prussic acid , that are readily liquefied by cold and present no difficulty in manipulating .
Faraday in his extensive paper on liquefied gases , published in 1823 , when describing the properties of several of them , compares only their index of refraction to that of water , calling it more or less , and Brewster in 1826 mentioned in a communication to the Society of Edinburgh the index of refraction of liquefied cyanogen as but without any remarks on the manner in which it was deduced .
I will commence this paper by describing the method I followed to obtain the numerical values of the index of refraction of several gases , liquefied either by pressure or by cold , and that enabled me to surmount the difficulties resulting from high tensions and small quantities of fluid substarlce , that may perhaps have kept back other experimenters from this field of research .
* ' ' Proc. Roy .
Soc vol. 25 , p. 322 .
As in my former investigation , I chose again in most cases Faraday tubes , but I had them modified in the proper way , so as * obtain the liquid enclosed between parallel glass sides , when I coffi apply the microscope bo make quantitative experiments .
After many preliminary arrangements that were successively tried I came to an equally effective and simple construction , that showed itself most convenient and is represented in figs. 1 and 2 , and I suppose it may prove useful also in other researches with liquefied gases .
A section of the apparatus is given in fig. 1 ; it consists of a short glass cylinder ( length 15 millims. , inner diameter 3 millims. , outer diameter 11 millims communicating on both sides 342 Dr. L. Bleekrode .
On the Determination of the with capillary tubes and ; this was transformed into a having two parallel side plates , and free from leakage even at high pressures , in the following manner .
The ends of the cylinder were both ground as flat as possible , ou of them a thin circular disk reed in the centre and cut from lead or tin , according to the liquid examined , was put , subsequently a plane glass plate that had been previously submitted to toughening process of La Bastide ( in some a quartz disk wn used ) followed , and lastly a perforated smooth leathern disk .
Ths whole was held together by two perforated steel plates forming the ends ( fig. 1 shows the construction , and showing the apertures and connected with another by three sorews turned on as strongly aa possible .
* A perfectly gas-tight vessel was thus obtained that did not allow leakages even at pressures above a hundred atmospheres , ag was tried with the compression apparatus of Cailletet .
And I must point out here especially the property of toughened glass to resist in the most effective way the always more or less unequal pressure , exerted by the three connecting screws , when even quartz plates being 3 millims. thick burst to pieces , if the screws happened to be turned on unequally .
The glass plates , having only a thickness of 1 to millims. , were cut out of very plane surfaces , and after being reduced to the proper dimensions , suffered the toughening operation , that was applied with much care .
Of course some of them got their surface more or less impaired and rendered less transparent ; from a great number the best ones were chosen , and they never became unfit for use .
In most cases the lead disks , procuring a very successful closing , were not attacked in a chemical way by the liquefied substance if no moisture accompanied it ( with hydrobromic acid tin disks resisted better ) , and thus commonly a perfectly clear liquid filled the cylindrical vessel .
In how effective a way leakages were prevented may be concluded from the fact that liquid carbonic acid was often kept during more than a year in such hermetically closed vessels , and even when heated during several hours above C. , without a perceptible loss occurring .
If the gas could be evolved chemically in a sufficient quantity to become liquid by its own pressure exerted , the cylindr.ical vessel wa8 fixed with the blowpipe on a stout bent glass tube ( fig. 2 ) , having a length generally of 40 centims. , ari inner diameter of6 millims. , an outsr diameter of 10 millims. The appropriate substances were put in the * The steel plates were 8 millims. thick ; in the figure only two screws could be represented entirely .
Organic substances , applied to obtain a gas-tight closure , should be when working with apparatus for liquefied gas , as these are absorbed by the liquids , rendering them often opaque .
Index of Refraction of Liquefied Gases .
lower part of the tube , in the upper one some compounds fit to absorb moisture present in the evolved gas , and both ends sealed with the blowpipe .
Afterwards the substances were put in contact by gently shaking the tube , and the operation conducted so as to have the liquefied gas distilling over into the vessel , directed with its axis in a horizontal position .
Enough liquid was always obtained to fill it compJetely and generally also the connected capillary tubes on both sides , and no gas or air remained ; the small dimensions of this part of the apparatus caused some drops of liquid to be sufficient .
The liquefying operation conducted the Cailletet pump procseded in the usual way ; the capillary part ( fig. 1 ) of the vessel was fixsd with the blowpipe on the ordinary gas tube of this , the other capillary part was closed after this tube was filled with the purified gas .
If the gas could be put readily into the liquid condition with the aid of low temperatures , the cylindrical vessel was put on a doubly bent glass tube , and this surrounded by a freezing mixture ; this was applied in the case of sulphurous acid , some amine compounds , \amp ; c. Lastly , I will observe , that after terminating a series of experiments with some gas , the whole apparatus could be liberated from its contents by cutting off a very small part of the lower end of the tube , and all remaining substances were removed by the pressure of the gas ; thus I could afterwards examins the glass disks and convince myself whether the liquefied gases , being perhaps not pure , had left behind some moisture or other deposits .
III .
The Method of Observation .
The reason why I had to provide for getting the liquefied gas enclosed between parallel glass plates was to have an opportunity to apply the microscope , which perhaps may be the only practical effective method under these circumstances .
The use of this instrument in determining the index of retraction was first proposed by the Duke de Chaulnes in 1767 , but the exactitude of the results being limited to the third decimal ( from the principle on which the method is based ) it is inferior to many other methods , and was , of course , neglected in physics .
* The mathematical expression for the index that has to be used in this case follows from very simple considerations .
* In 1849 Bertin published some modifications of this method , without improving , however , its exactitude , as appears from the values communicated for the index of glass and , the goniometer giving and More recently , in 1876 , Royston Pigott , when describing his refractometer , mented on adapting his instrument to the microscopioal method , and in 1878 .
Sorby , of Sheffield , made several valuable and interesting communications on the refractive indices of minerals , which were determined with the microscope , and he pointed out the practical use in the case of small ntities of liquids .
344 Dr. L. Bleekrode .
On the Determination of the Suppose in fig. 3 a liquid layer of a certain thickness limited parallel surfaces , its refractive power shall be such that any object placed in will be seen apparently in ; hence the relation\mdash ; when the angles of incidence and refraction are very small , calling the thickness of the layer , and the apparent displacement of the equation is obtained\mdash ; ' which gives the value of the index by measuring two linear dimensions .
Practically the case is not so very simple as is supposed , because the terminal glass plates , transmitting the rays , contribute to the displacement , as is shown in fig. 4 .
The effect caused by thess two disks together I call the " " glass effect and it had to be determined for each vessel itself before it was filled with liquid ; the amount was afterwards deducted as a constant correction .
In fig. 4 the vessel is represented by abdc , and it is obvious that the rays , which can bo transmitted through the liquid under the greatest angle of refi'action , have a direction along ac , when they will pass through the apertures in the centre of the steel plates laid on and In the greater part of experiments , and measured 20 and 21 millims. respectively ; the angle of refraction deduced from its tangent will not surpass a value of , and therefore the error , caused by taking the sine instead of the tangent , only amounts to ome unit in the third decimal .
The numerical values for and are easily obtained with the aid of the microscope , which was specially constructed for my purposs by Nachet of Paris .
The ocular tube ( that in most observations was placed in a zontal d was provided with a vernier , moving along a scale divided in halves of a millimetre , and allowing to determine a displacement of the tube to millim. ; a micrometer screw could also be used for a more adjustment to the millim. Of course , the exactitude that may be reached in focussing any object depends on the magnifying power applied ; unfortunately , I was limited in using very high powers by the necessary dimensions required by the glass cell to be strong enough and fit to the end in view .
I chose a high magnifying ocular power and a very feeble objec tive lens , getting by this combination a magnifying power of nearly ; it gave the advantage of having very well-defined and clear images to focus on of microscopic test-objects\mdash ; as , for instance , the scales of wings of Lepidoptera fixed on the glass plates of the cell merely by adhesion on the inner surfaces .
When the cell was constructed , I Index of Refraction of Liquefied Guses .
began with determining its height , putting it before the objective lens of the microscope , and focussing first the objects on one of the glass plates , then those on the other ; taking now the difference between the two readings the value of was obtained .
I remark that in this way the effect of the front glass plate ( next to the objective lens ) and also of absolute errors in the division of scale and micrometer screw were eliminated .
Putting successively the tube on other parts of the scale , I could take a range of readings numbering six or eight , from which the mean value was deduced .
The determination of , the apparent displacement , was somewhat more complicated , a part of it being caused by the refraction of the glass walls of the cell , and this was to be first ascertained .
I found it most convement for this purpose to have the objects to be focussed not on the glass plates of the cell , but on a separate glass slide placed immediately behind it ; otherwise in taking different readings , when working with a liquefied gas , I should have had the cell to be alternately filled and emptied again .
Yet the deduced mathematical expression may also be used in this modified case .
Considering in fig. 3 an object placed beyond the cell to be focussed through it , either empty or filled , then , by the effect of refraction , it will be seen apparently in ; and when the angle is very will be almost a straight lin , and the equation is obtained\mdash ; Therefore , when a complete observation of 7 ?
for any liquefied gas was to be made , a glass slide with test-objects was placed in front of tlJe microscope , the glass cell still empty ( whose height had been previously ascertained ) followed close to it , and , putting the tube of the microscope in a horizontal position , I focussed the objects first through the cell , then , moving this aside , I focussed again ; a set of such observations gave the mean effect of both glass plates of the cell together .
Now the liquefied gas was prepared in the glass tube connected with the cell , and I look great care to have it completely filled hout any remaining gas-bubbles .
I obtained this easily enough , when maintaining the axis of the cell in a horizontal direction .
To get the index of refraction of the liquid , I focussed through it the test-objects ; I determined the position on the scale of the microscope , corrected it for the glass effect , as described ; moving the ccll aside , I focussed again the test-objects directly ; the difference between the former and the atter reading gave the of , and the height or being already known , is calculated .
This method is dependent on the condition that the cell is provided with parallel * With the greatest value of the angle of refraition , , and with an index of refraction ranging between and , the obtuse angle amounts to 346 Dr. L. .
On the Determination of the .
sides ; it was therefore necessary to ascertain for each experiment what reliance could be placed upon it , the more so because , though tho .
glass plates themselves might have been very parallel , yet by a more or less unequal compression of the intermediate leathern and leaden disks , a somewhat prismatic form of ceJl could have ensued .
Whether this were he case , I always controlled with every new prepared tube by to determine in it the index of refraction of sulphurio ether , in he same way as had to be done afterwards with liquefied gases .
I compared the value obtained to that which has been found with the more exact methods ; if it agreed within the degree of exactitude that comports the use of the microscope , I was sure that no error was to be feared from want of parallelism between the ends of the if there was discordance , I reconstructed the cell , till a satisfactory was got .
To appreciate the influence of errors of observation on the obtained value , a differentiation with respect to and of the relation lvill give If a nifying power of 30 were applied , I could adjusb the testobjects with a displacement to or millim. , calling the errors made in the observation and , the following expressions for the errors in the values of and ensue\mdash ; In an experiment with sulphuric ether , for instance , measured millims. , millims. , millim. , millim. therefore , ; In the most unfavourable case , the total error may amount to , and probably it may be expressed A higher magnifying power\mdash ; as , for instance , one of 200 times-would have allowed of reading a displacement to millim. ; but mechanical difficulties in the construction of a cell of very reduced dimensions , required by the short focal distance , prevented the employ* Sulphuric ether was chosen as controlling substance , because it could be easily removed by evaporation , accelerated with the air-pump .
Index of Refraction of Liquuefied meant of one .
Its height should have been less than 0 centim. , and a cell which ought to resist strong pressure was not realisable in this case .
I will now adduce the ] of some determinations concerning the thickness of a plane parallel glass plate , made with the microscope , and also to verify the precision with a spherometer allowing to read to a of a millimetre .
Microscope .
Magnifying power ) Spherometer .
111 millims. 10745 millims. 107 Mcan . . . . . .
10746 Mean . . . .
1 Dffierence . . . . . . . .
millim. Calculating the probable error , its value is found for each determination as , and for the result of a whole series as Applying a much higher magnifying power , I could focus on the very clearly defined strips on the scales of Lepidoptera wings ; the results were still more satisfactory .
As proof , I present the determination of the height of a glass cell , consisting of a short cylindrical tube , having plane parallel plates on both ends .
Microscope .
( Magnifying power ) Spherometer .
540 millims. 5 millims. 5 .
5.429 5 .
5.413 5 .
5.427 5 .
5.426 5 .
5.432 5.42 5 .
32.540 540 Mean . . . . . .
5423 5.44 Mean . . . . . . . . . . . . . . . . . . . . . . . .
Difference . . . . . .
VOL. XXXVII .
2 348 Dr. L. Bleekrode .
On the Determination of the The alculated probable error of the result amounts to I will now adduce a complete series of observations taken from preliminary experiments , which I made to become familiar with the microscopical method .
They serve also as an elucidation of manner in which the results recorded into the next chapter were obtained .
The following example bears on the determination of index of refraction of sulphuric ether , and was effected with the glass cell mentioned above , and with a magnifying power of 140 .
Determination of the Height D. Focussing on the upper glass disk. . . . . . . . . . . .
Focussing on the inferior glass disk. . . . . . . . . . . .
5 .
o'.43 Mean . . . . . . . . . . . . . .
millims. Determination of the Correction for the Glass Effect .
Focussing through the cell Focussing without the cell Mean. . . . . . . . . . . . . .
millim. Determination of the Index Focussing through the glass cell and liquid . .
Correction for glass. . . . .
Focussing directly . . . . . .
Mean. . . . . . . . . . . . . .
Temperature The light employed for the determination was that from a Bunsen 's gas-burner , the flame containing some natrium chloride ; the value , fore , refers to the line .
If it is compared to the result obtained with other methods , for instance , that given by Lorenz* , at and at , leading by interpolation to at about , it is obvious that the concordance is very satisfactory , and values are taken from the recently ] ished ysikalische und Chemischo Tabelle\ldquo ; of Landolt and Bornstein .
The temperatures given in his } refer to the centesimal scale .
Index of Refraction of Gases .
it may be concluded that when some precautions are taken the micl'oscopical method may give reliable results .
Of course , a high magnifying power was used in the fore-mentioned experiment that could not be applied when dealing with liquefied gases , still good results are ossible even with a magnifying power of 30 , as is evident from the following table containing observations on sulphuric ether and water taken in cells of different heights . .
quid .
The determinations from the correct values are within the limits of the probable error , and I must observe that even with very accul.ate methods , attaining four decimals , the results given by different experimenters vary in the third decimal , for instance , with sulphuric ether at , Gladstone and Dale found the index ( line ) ; Kundt , however , .
Experiments on the Refi.action of Gases .
I will now proceed to give the r , sults of the experiments bearing on several liquefied gases , and made as described above .
Thel.efore I think it sufficient , after the explanations , to give for each gas the final results for the of , these having the same signification as in the former chapter .
Yet I must add that during a space of four years I repeated at different times the determinations for each gas with different vessels ; in this way the number of single observations has amounted- to some thousands , consisting of series of six or eight terms .
I commonly employed as source of light the flame of a Bunsen gas-burner , containing a piece of wovell asbestos , moistened with a solution of natrium , and pro- ducing an excellent monocbromatic light .
Sometimes I also experimented with ordinary daylight ; this was necessitnted by some calculations , as will be seen in the next chapter .
1 .
phnrous Acid .
The gas was prepared in the usual way , and liquetied in a tube connected with a ] -sided glass cell by means of a freetiug mixturc .
350 Dr. L. Bleekrode .
On the Determination of th observations led to millims. , with the natrium millims. , at , with daylight at The index of refraction of the easily liquefied gas was already determined by Ketteler in 1865 with a goniometer , and he found the value of for the D Iine at ; more recently Hichat* observed at .
It may be deduced from these quantihes that this quantiby ) nearly for one degree of changein the telnperature ; if , therefore , a quantity of is subtracted from the value obtained in my experiments , there resuIts , approach .
ing nearly the amount 1 344 , resulting from an exact and demonstrating again what may be expected from the microscopical method in such cases , and how far it is reliable .
2 .
Cyanogen .
The gas was prepared from very pure cyanide of mercury strongly dried before , and the liquefaction effected in a bent tube with glass provided .
The observations led to millims. , lvith the natrium light millims. , at , with daylight at ths same temperature .
In 1826 already D. Brewster the amount of the index as , yet without indicating the method by which ib was found , nor the to which it refers ; no comparison thus is possible .
3 .
ocyanic .
gas was prepared by decomposing the cyanide of mercury by sulphuretted hydrogen had been led through anhydrous phosphoric acid ) , and condensing with a freezing mixture in the glass cell .
It was thus obtained pure , as it showed no alteration after a long space of time .
The observations led to millims. , with the light millims. , at , with daylight at Determinations of the index were already made by Bussy and with the prismatic method , giving at for the ] Earlier still , in 1839 , Cooper also obtained , in the same way , for daylight , 4 .
Nitrous Oxide .
I could not succeed in preparing the gas obtained by the decomposition of ammonia nitrate in liquefied condition in glass tubes , as * Bichat , " " Journ Physique , 1880 , p. 279 .
" " Transactions of the Society of Edinburgh vol. , p. 407 .
In this paper there is only said on this subject that Dr. Turner prepared the liquid cyanogen by compression .
Bussy and , " " Annals de Chimie et de Physique\ldquo ; ( 4 ) , ' Index of Refraction of Liqu Gases .
as done by Faraday , as these alw.ays exploded .
The compression pump of Cailletet , therefore , was extremely useful in this case , and applied it with its glass tube modified as desclibed in the former chapter .
Enough liquid was produced to fill the cell at the top completely at a pressure of nearl .
sixty atmospheres ; a series of observations taken in 1879 led to millims. , with natrium light millims. , at , with daylight , also at .
Later on , in , when controlling my former experiments with another glass cell , I found millims. , with light millims. , and at , with daylight These resuIts are almost identical .
5 .
Phosphoretted Eydrogen .
The liquefacbion of this gas was also readily effected with the aid of the Cailletet pump .
It was prepared from a mixture of phosphonium iodide and a weak soIution of potash , *evolving by their reaction a thoroughly pure gas ; at a pressure of about thirty atmospheres a perfectly clear liquid filled the glass cells .
The experiments gave millims. , with natrium light millims. , , at the temperature with daylight at 6 .
Sulphuretted Hydrogen A series of observations was taken in 1879 with the aid of the compression pump , and the following values on experimenting with the liquefied substance ensued : millims. , with natrium ligLt millims. , at a temperature of Later on I repeated the experiments , using the faradaic tube , and liquefied the gas by its own pressure .
I obtained millims. , millims. , with natrium light at , with at the same temperature .
Instead of leaden disks tinfoil was applied to obtain a perfectly vessel , and to prevent any chemical reaction .
7 .
Hydrochloric ( HC1 ) .
This gas was evolved in a suitable bent glass tube from a mixrure of ammonia chloride and strong sulphuric acid the very little gas bubbles ( evolved under increasing exerted by the gas itself ) had to pass , of course , through a long columll of this acid , and were thus deprived of any moisture .
The liquefied gas did not at all attack the leaden disks used between the glass of the cell .
The observations led to millims. , with the natrinm light millims. , at .
Another set of observations * This method was first proposed by Hofmann , Ber .
Chem. Gesellsch 1873 , Bd .
, S. 100 .
352 Dr. L. Bleekrode .
the Determination of the were made with the gas , liquefied with the Cailletet pump , and gave similar resuIts .
I found millims. , with natrium light millims. , at , lvith daylight , also at 8 .
Hydrobromic Acid .
As mercury is attacked either by the gas or liquid I could not avail myself of the Cailletet apparatus in this case , and on this occasion the faradaic tubes served very well .
I obtained great quantities of liquefied gas from a mixture of moist potassium bromide and a small quantity of red phosphorus , separated by a paraffine stopper from a column of liquid bromine put in s/ lbsequently .
The gas evolved passed through granulated zinc ( to combine with any bromine vapours ) and through calcium chloride , and condensed afterwards into the glass cell as a colourless liquid .
Tin disks were also here applied for perfect closure , yet were slightly attacked by the liquefied gas .
The determinations led to millims. , with the light millims. , at , and with daylight also at 9 .
Hydriodic Acid ( HI ) .
This gas was liquefied in a similar way as above , it being evolved from a moist mixture of red phosphorus iodide and potassium iodide ( in a fixed proportion ) .
The observations led to millims. , with natrium light millims. , at 10 .
Chlorine ( Cl ) .
Here also only the bent glass tube could serve , as mercury would not have su#ered the contact of the gas without combining with it .
I prepared it from a mixture of pyrolusite and hydrochloric acid by gently heating and purifying the evolved gas by a long column of pyrolusite and caloium chloride that would absorb some hydrochloric acid and moisture escaped at the same time with the gas , thus a perfectly transparent and ight yellow liquid was obtained .
The observations led to millims. , with natrium light illims .
, at 11 .
Carbonic Acid .
I paid peculiar care to obtain as accurate as possible the index of refraction of liquid substance that has been studied in various vays .
I soon became aware in preliminary experiments that the temperature bad more influence than in the other investigations , as might be expected from the circumstance that the critical temperature of liquid carbonic acid is not very far above the surrounding temperature .
I preferred therefore the liquefaction a bent tube by its Index of Refraction of Liquefied Gases .
3.53 comprossion to the Cailletet apparatus , because the heat generated in becoming liquid lasts longer than with the other method , where it was always kept ready in liquid state .
Another disturbing influence is exerted by some residual air in contact with liquid carbonic acid ; discordant results led me to the fact that near to the critical temperaturs it becomes a very good solvent to air , and the index is then much lowered .
The gas was evolved in the tube from a mixture of ammonia bicarbonate and strong nitric acid , and had to pass through calcium chloride before entering into the glass cell .
I expelled the air as far as possible with an air-pump and allowing a little reaction between the ingredients a certain quantity of carbonic acid gas was liberated , which , pouring out from the capillary tube at the end of the cell , filled remaining interstices and expelled remanent air , while the tube was sealed .
From numerous series of observations obtained during the last years , communicate the following numbers taken with different tubes:\mdash ; millims. miUims . .
Temp. . . . . . . . .
II . . . . . . . .
III . . . . . . . .
I has been taken with , the two other observations with the natrium liglJt .
To illustrate the effect of the residual air , I can present the following experiments:\mdash ; 1 .
II .
millims. millims. Temp. . . . . . . . . . .
1601 . . . . . .
15.98 . . . . . . . . . .
1.193 . . . . . . . . . . . .
1.175 . . . . . .
1.184 . . . . . .
17.5 It is evident that tube I having some air included , and I ] being from air , the difference between the values at the same temperature is increasing as this becomes higher ( the difference is greater than the probable error ) .
It also appears in this table how the index is very sensibly affected by changes of temperature ; this influence is already marked in the second decimal , and the more as the criticaI temperature is approached .
* The influence of the temperature on the index of refraction of liquefied carbonic acid , and the changes occurring near and above the critical temperature , will probably form the subject of my next communication .
354 Dr. L. Bleekrode .
On the Determination of the 12 .
Ammonia .
I prepared this gas in its liquid condition in the most convenient way by filling a tube with strongly dried granulated calcium chloride ; whilst a current of pure ammonia gas was passing it for time , the gas became copiously absorbed , and the filled tube served afterwards , by means of a moderate heat , to deliver a large quantity of gas in the glass cell , connected with the tube , which by its own compression became liquid .
The experiments gave millims. , and with the natrium light millims. , at the temperature of , and with daylight at .
I turned my attention to some amine compounds , which are in chemical properties similar to ammonia , and are prepared from it by substitution .
Among them only methylamine requires liquefaction , as it boils at , and the liquefaction is easily effected by means of a freezing mixture , which is also desirable , when working with dimethy ] amine trimethylamine , having boiling points respectively at and .
I ordered these compounds from the renowned chemical workshops of Kahlbaum at Berlin , whence I recsived them in sealed glass tubes , and they were distilled in proper bent tubes connected as usual with the parallel-sided glass cells .
I obtained with the natrinm light\mdash ; .
Temp. Methylamine , . .
1342 . . . .
Dimethylamine , . . . .
17 Trimethylamine , . . . . . .
16 The determination of the index for other amines presents no peculiar interest , as their boiling point is high enough to admit the ordinary methods of observation .
13 .
Eihylene .
The history of the liquefaction of this gas is somewhat curious .
It was already liquefied in 1845 by Faraday on application of both inc.reased pressure and low temperatures ( 42 atmospheres at ) , but as the notion of critical temperature was then not fully understood , it remained unobserved for this gas , though it is very easily obtained .
In 1880 Amagat*published an experimental verification of Boyle 's for this gas , and my compatriot , Professor .
Waals of Amsterdam , deduced on theoretical grounds from the resuIts obtained the critical temperature of liquid ethylene , and had it *Amagat , ' Sir la Compressibilite des Gas sous For\amp ; s Pressions " " Comptes Rendus 1880 , t. 91 .
Index of Refraction of Gases .
confirmed actually by experiment as .
In 1882 Cailletet* strncted a special apparatus to prepare liquid ethylene in great quantities , and pointed it out as a means of obtaining a much lower degree of cold than was possible before , and this opened the way for the liqnefaction of oxygen and hydrogen .
For my purpose I prepared a snfficient quantity of the liquefied gas with the ordinary compression pump of Cailletet , taking care to use a very pure gas obtained in the usual way .
I collected the liquid in a tube divided in calibrated parts , and could thus determine the specific gravity at different temperatures after the method described by Ansdell , and applied by him to liquid hydrochloric acid and acetylene .
My experiments gave at at , at .
For acetylene Ansdell observes that its specific gravity being at , it is the lightest fluid substance known ; it is surpassed in this regard by liquid ethylene , that has now to be considered as such. .
A second tube provided at one with the parallel-sided vessel already described , at the other end with a ) specially constructed iron connecting piece , stopcock cemented on the tube , served observation of the index of refraction .
After the column of mercury in the pump had driven the liquefied gas into .
the vessel , the stopcock was firmly screwed down , and now the tube with its contents could be removed from the ratus , and was always ready for observation .
Determinations of the index of liquid ethylene could only be made of course when the temperature of the surrounding atmosphere was below ; I had them in free air , but the winter being very mild this year , I had no opportunity to observe below .
The effect , of a slight variation of temperature on the refraction is much more marked than with carbonic acid , as was to be expected , and renders an exact resuIt rather difficuIt .
I obtained under satisfactory conditions for daylight at Conclus.ions .
SoIne investigations of a mathematical character have been published in later years that treated of the relation of the refractive power of the same substance considered in different states .
The theol'etical * Cailletet : " " Comptes Rendus 1882 , t. 94 .
He mentions on p. 1224 as the critical temperature , and Sarrau ( p. 846 ) nearly or .
I found the same as by .
Waals .
AnsdeU : " " Proc. Roy .
Soc vol. 29 , 1879 , p. 221 .
Such a connecting piece is a most useful appendix to the compression pump , and is constructed also by Ducretet , at Paris .
It is very accurately wrought .
I kept for more than ten days the liquid ethylene at a pressure from Bixby to seventy atmospheres without perceptible leakage , in a tube having the stopcock turned down .
conditions of the former expression .
Especially has this relation been used in cbemistry for comparing the molecular refractive powers of organic compounds , and has again been applied in the extensive and .
important researches of Bruhl .
Far less are the attempts made to test both expressions for the different states of substances , and to ascertain which of them rees the best with the experimental resuIts .
The papers of Lorenz and of Prytz afford together seventeen compounds , nearly all belonging to ganic chemistry , and examined as fluids at and and compared to the vapours at .
The method used for determining the refraction allowed an exactitude to the fifth and sixth decimal .
I adduce from the paper of Lorenz some results in the table:\mdash ; deduced this expression from the theory of undulations in a paper published in the Danish language in 1869 .
A translation appeared in 1881 in the " " Wiedemann Annalen 11 , p. 70 .
Professor Lorentz came to the same relation from the electro-magnetic theory of light , and his estigation was translated in the same periodical in 1880 , Bd .
9 , S. 641 .
Dr. T. W. Bruhl , in " " Liebig .
Annalen .
Chemie Bd .
200 , 1879 .
Lorenz , , and Prytz , ' Wiedemann .
Annalen Bd .
11 , S. 204 .
Besides these papers , I know of no others bearing on this subjeot .
ff Index of Ref.action of Liquefied Gases .
To these results I add the following remarks .
1 .
The values are calculated from the indices determined with the ratrium flame , and not corrected for dispersion .
observes that the deduction of the index of refraction for undulations of infinite length is not possible with the same exactitude as is attained by direct observations with visible light , and prefers , therefore , to apply the resuIts in the formula .
2 .
As Lorenz has determined by experiment for each substance in its fluid and gaseous condition the amount of dispersion , and this remains nearly the same in both cases , he concludes that this justifies bhe use made of the expression for undulations of definite though it was deduced from those of infinite length .
3 .
The recorded oalculations point out that , though a great in specific gravity occurs when a substance is passing from one condition into another , yet the relation betwee1l the refraction and the density approaches closely to a constant , the greatest difference amounting to 5 per cent. with the acetate of ethyl .
It also appears that the exprsssion established on tbeoretical grounds answers far better to the experiments than the empirical one , as is evident from the three latter columns in the tables .
I was induced by these considerations to try how far my own experiments would support these esults , though less concordance is to be expected , as the exactitude is already impaired at the third decimal of the observed value .
The next table contains my determinations put together with other quantities belonging to the same substance and required for calculation by both expressions .
The indices of refraction of the are partly taken from the paper of and others , from more recent and accurate researches made by Ketteler , Lorenz , and Croullebois .
The densities of in liquid condition have been accurately determined by Andreef , that of liquid HC1 by Ansdell .
As in Dulong 's researches no special li is mentioned , it must be assumed that daylight was used in determining the refraction ; I therefore employed my own corresponding values , takell with common light , for comparison , inasmuch as no recent experiments on gases with natrium light have been published .
From the results in the table , it is again evident that , though no * Dulong : " " erches sir les Pouvoirs Refringents des Fluides Elastiques .
Croull bois : " " Annals de Chimie et de Physique 1870 , t. 20 .
Andreef : " " Annalen der Chimie .
Pharmacie lS59 , p. 1 .
Ansdell : " " Proc. Roy .
Soc vol. 30 , p. 221 .
He applied the compressionl apparatus of Cailletet , and I followed this very useful method with phosphoretted ydrogen , sulphuretted hydrogen , ethylene .
The specific weights of the other liquefied substances are taken from the " " Physik . .
Chem. Tabelle\ldquo ; of Bornstein ndolt .
Dr. L. eekrode .
On the Determination of the \mdash ; Index of Refraction of Liqnef Gases .
perfect concordance between tbeoretical relation and the experimental values is necessary , yet the -expression can be adapted in much higher degree to the observations than the empirically established -expression ; and the difference between the calcuIated and observed quantity is in most cases no more than that also in the very accurate expe ] of Lorenz , given in the table on p. 358 .
With the latter expression the value of the proportion is generally decreasing when the substance becomes gaseous ; with the former it is increasing ; yet there are some exceptions .
* Especially with nitrous oxide , there is a greater diffsrence than with others , and I know of no peculiar cause for it , as I repeatedly examined the index .
I must observe that also among the substances given in the paper of Lorenz , some of them may be remarked ( for instance , the acetate of ethyl ) as presenting a far greater difference than others .
The ethylene , also , is distinguished among the other substances by a far less concordanco , but I may perhaps account for it by the very conditions of the experiment , as this had to be conducted very near to the critical temperature of the liquefied substance .
I must add , however , that if , with the value of the proportion given in the table for the liquid ethylene , the equivalent of refraction is calculated , the result is most satisfactory , corresponding to that deduced from refraction equivalents of the composing elements , added together .
Accordingly we have\mdash ; Refr. equiv. Refr. equiv. Refr. equiv. Of course such verifications may be made also for the other substances of the former table , and calcnlated for their liquid condition , but as these may be deduced as well from the gaseous state , and generally with more accuracy , I think there is no peculiar interest in adding the results .
' In the experiments of Lorenz this value is for the examined compounds always increasing when they become gaseous ; in those of Prytz it is for some for others employed in his calculation the ession M , in which ents the molecu ] weight of the compound , and the ction ealents of , and of wera taken from the paper of Landolt Sitzungsber .
Preuss .
Acad. .
Wissensch 1882 , I , p. 43 ) .
If th and the corresponding values , and , are admitGed , bhe concordance is much less , as we obtain ( observed ) and ( calculated ) .
Professo ] .
Gladstone remarks that as the carbon atoms in this compound , according current theories of chemical constitution , are to be considered as do -linked , the calculated value is much higher ; the ence from the experi- meutal value is too great to be accounted for by errors of ervation .
1 Landolt , in the paper alluded to above , made such verifications Bith both expressions applied on nearly fifty organic compounds , and concluded that the formula gives more accurate lesults than the other when applied to the index of a from my experiments with the liquefied acids , I add here the results , as they can help to elucidate a question suggested by Professor Gladstone , in his former investigation on the efraction equivalents of the elements , and relating to that of hydrogen especially .
Refraction Equivalent of the Halogens and of their Combinations with Hydrogen .
The columns headed with the term " " calculated\ldquo ; contain the that are obtained by addition of the refractive equivalents of the combined atoms ; for and Cl they are given in the paper of LandoIt ( already alluded to ) , for both expressions ; for Br and I they were calculated by myself .
Next to this column are placed the corresponding values of the substances , deduced ectly f the , compared in liquid and gaseous condition , and that both were only approximative .
*The determination of the index of liquid will be described in the next chapter , and it was ascertained to be Gladstone : " " Pro Roy .
Soc vol. 16 , p. 443 .
calculated the value of iodine for the formula with the value given in the " " hysik . .
Chem. Tab. ' ' of Bornstein .
Landolt as the index for the line , and the density , at Index of Refraction of Liquefied Gases .
index of refraction and density* ( experimentally determined ) , and the atomic weight .
It is obvious that by subtraction of the proper experimental values we may the mean refraction equivalent of ydrogen , thus determined from observations on the liquid acids themselves , and not in solution .
We have for both expressions\mdash ; HCI\mdash ; CI. . . . . .
\mdash ; HBr\mdash ; Br . . . . .
16 Mean value of . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
, Calculated value from other substances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The conclusion is that the concordance between the value deduced from my experiments and that obtained from other observations is tolerably satisfactory , and the remark that the equivalent of refraction of hydrogen in an inorganic combination , and especially in an inorganic acid , amounts to , is not applicable to liquefied hydrochloric acid , nor hydrobromic and hydriodic acid .
They were till now only examined in their solutions in water , and it has already been pointed out by LandoltS and others that the dissolving does not exert the same influence on the index of refraction as on the density ; my experiments being free from this influence may be deemed appropriate to fix the value of the atomic refraction of hydrogen in these compounds .
*I had to determine the density of liquid hydrobromic and hydriodic acid , hitherto unknown .
I could not well apply the method of the compression apparatus , as the mercury is somewhat attacked by the liquefied gases .
I therefore employed the Faraday tubes , with glass cell , described before and provided at the end with a calibrated capillary tube .
The tube with its contents and the liquefied gas weighed ; then , by unscrewing the glass oell a little , a certain quantity of the liquid was allowed to escape .
The cell was closed and the whole tube weighed again .
The diminution in weight observed gave that of the escaped liquefied gas , and its volume was before noted in the capillary part of the tube .
I have found the specific weight of liquid HBr at , that of liquid at The difference between calculation and experiment in this case perbaps be accounted for by the difference existing between the values communicated by various observers for the same substance .
the table the refr .
equiv. of iodine given by Landolt as , is used ; if I had admitted the value given by Professor Gladstone as , the refr .
equiv. of hydrogen would have amounted to , according exactly .
with other observations .
Professor Gladstone , .
I have shown previously that also from an electro-chemical point of view these liquids differ essentially from their aqueous solutions .
S Bruhl , , p. 222 .
362 On the Index of Refraction of Liquefied Gases .
VI .
Determi , of the Indeip of Refrachon of Liquid Bromine and some Organo-Metallic Compounds .
I add to this paper the results of experiment , with some substances that , though they are liquid at ordinary temperature , yet present some difficnlties on being examined by the common methods ; their index of refraction has also now not been observed directly .
The practical efficiency of the microscopical method is illus .
traled in the case of liquid bromine , which is very opaque even in thin layers .
I had a glass cell constructed with parallel sides in a simplo way , by cementing ether two plane glass disks , one of tbem consisting of 'a glass cover ( as used in microscopical preparations ) , with a hole in the centre ; this forming the proper vessel when it was finally covered with a glass disk , a single drop of liquid bromine being previously put between .
Supplying a 140 magnifying power , enough light dayligh was transmitted through the layer to permit the focussing on scales of Lepidoptera wings .
This layer measured millim. , the apparent displacement millim. ( as a mean result of repeated observations ) ; from this the index of refraction is calculated at a temperature of I examined also some orgauo-metallic compounds , especially those which are spontaneously ignited on contact with air .
I employed doubly bent glass tubes , ending in the parallel-sided vessel and filled previously with carbonic acid gas .
Then I distilled into tbem from another tube the liquid compound prepared in usual way , a very small quantity .
of courso sufficient .
I add the following results , to the natrium light:\mdash ; .
Temp. Zinc ethyl . . . . . . . . .
1485 . . . . . . . .
12 Zino methyl. . . . . . . .
1474 . . . . . . . .
14 ethyl . . .
1480 . . . . . . . .
65 Aluminium methyl . .
1432 . . . . . . . .
12 these values are compared to those of other metallic compounds , as fcr instance the tetrachloride of tin , its index amounting to ( for the line of the spectrnm of hydrogen ) , we may remark that it is not necessary for a substance to be very combustible in ordel to have a very high index of refraction , an opinion that sometimes has been pronounced .
* Haagen , in " " Pogg .
Ann , Bd .
151 , S. 122 .
I may also add analogous compounds , as mercury ethyl line ) and mercury methyl .

