210 



THE CIVIL EXGIXEER AND ARCHITECTS JOUIIXAL. 



[August, 



height ill both tul)e> ; the communication with the ilryintj apjiaratus 

 was dosed, tlie barometer and tlie temperature of the water around 

 the tubes noted. The ijlobe was then brouerht to loo", tlie mercury 

 in tlie vertical tube was of course depressed, and in order to keep 

 that in tlie barometric tube at about the same level with it, its 

 stop-cock had to be opened and the mercury sutlered to flow out ; 

 the two columns were thus kept nearly at the same hei^fht, that in 

 the tube in which the air was dilatinar, beinj; brou^^ht to a second 

 mark /3, and the exact difference in the heights of the two columns 

 was carefully noted, as well as the heij^ht of the barometer, and 

 the temperature of the water in the surrounding vessel. In order 

 from this experiment to determine the dilatation of the air, it is 

 only necessary to know the ca])acity of the globe, and of its stem 

 as far as the mark o, and that of tlie vertical tube between a and 6 ; 

 these are all easily determined by the weight of pure mercury 

 necessary to fill tliem. 



Four experiments tried in this way gave a mean dilatation of 

 l'.'J6706: the maximum being 1'36718; the minimum 1'36693; 

 difference j^^j of the mean. The co-efficient of dilatation given by 

 this fifth method is sensibly greater than that got from the others. 

 This circumstance is not accidental, as in the second part of the 

 memoir similar differences are shown for other gases, and in cer- 

 tain cases these differences are very considerable.* 



M. Regnault then proceeds to the discussion of his formulae, for 

 the purpose of determining the probable error in his results, and 

 he shows that in the first three series — principally owing to the 

 uncertainty of the readings of the barometer within ,'5 millimetre, 

 the maximum probable error is about ,431 which is aliout the greatest 

 difference between the maximum and minimum results in any one 

 series. The two last seiies include the same source of error, and 

 another arising from the uncertainty of the temperature of the air 

 in the capillary tube, which, however, he believes may be altogether 

 neglected in his experiments, the apparatus having been carefully 

 arranged so as to make this a very small fraction of the whole 

 volume of air under experiment. 



He finally assumes 0003(>65 as the mean co-efficient of dilatation 

 of dry atmospheric air as determined by the first four series of 

 experiments, and remarks that the number 0'003(i7 given by the 

 fifth series must be adopted in experiments where the gas dilates 

 freely and preserves its original elastic force. He also gives as a 

 fraction easily recidleeted, the remark of M. Babinet that 

 0-00366666 should be expressed by ^. 



Paut II. — On the Dilatation of some other Gases under Pressures 

 near that of the Atmosphere. 

 Physical philosophers admitted that all gases had the same co- 

 efficient of dilatation, but since so serious an error in the numerical 

 value of this co-efficient had been shown, it was necessary to submit 

 this law also to verification, the result of which was to show its 

 incorrectness. The experiments were tried chiefly by the methods 

 I. and IV. under constant volumes, and V. under constant pressure. 

 It is not necessary to describe them in detail, as M. Ilegnault has 

 done, nor to give the methods by which the gases were purified ; 

 suffice it to say that aU necessary precautions were taken, and the 

 general results were as follows : — 



Co. efficient of Dilatation from 0° to lOOo. 

 Under constant volume. Under constant pressure. 

 Hydrogen, 3667 0-3661 



Atmospheric Air, 03665 0-3670 



Nitrogen, 3668 



Oxide of Carbon, 0-3667 0-3669 



Carbonic Acnl, 0-368S 0-3710 



Nitrous Oxide, 0-3676 0-3710 



Sulphurous Acid, O.'iSlJ 03903 



Cyanogen, 0-3829 0-3877 



He also describes an apparatus, an easily-imagined modification 

 of method IV., by which the difference in the co- efficient of dilata- 

 tion of any two gases may be at once shown. 



Part III. — On the Dilatation of Gases under Different Pressures. 



It has been generally admitted that the dilatation of gases is 

 constant between the same limits of temperature, no matter to 

 what pressure they may be submitted ; consequently, that it is 

 altogether independent of their initial density. But it is difficult 

 to cite conclusive experiments upon which this law is founded. 

 Several observers having obtained the same value for the co-efficient 

 of dilatation of air, under different barometric pressures, concluded 



* M. Reenault describes in a note an indepeniient series of experiments tried by him 

 according to tlie metliod of M. Giy Lussac— th.it is, by observinij the dilatations of a 

 quantity of dry air coiit.iined in a true thermometer and separated from the external 

 atmosphere by a small index of mercury. (Blot. 'I'rait^ d ■ Phys. torn. l,p. IH-i.) The 

 results obtained did not agree at all, and were all feebler than by any of the other 

 methods i the highest result recorded was l-;j647. 



that the co-efficient of dilatation of gases w-as constant under all 

 pressures ; but the barometric variations in any place are not sufB- 

 ciently extensive to permit so general a conclusion to be thus 

 deduced. 



Sir Humphrey Davy is the only philosopher who has studied the 

 dilatation of gases under very different pressures. (Phil. Trans. 

 18a3, vol. ii, p. 20t.) 



He states that he found the same dilatation for air taken with 

 the densities n, i, .1, 1, and 2; but his experiments were not made 

 by a sufficiently delicate method to allow his results to be con- 

 sidered exact. 



The experiments of M. Regnault upon this subject were tried 

 witli apparatus of the same character as those before described, as 

 methods IV. and V., with such modifications as the peculiar cir- 

 cumstances of the experiments rendered necessary: and the con- 

 clusion at which he arrived was that '"'' the air di/ate.s; within the 

 same litnits of temperature, bij quantities vhich are greater in propor- 

 tion as the densiti/ of the yas is greater : that is, in proportion as its 

 molecules are brought nearer to each other." 



The following taldes exiiibit the results of his experiments upon 

 air, carbonic acid, and sulphurous acid : 



Dilatation of Gases wider different Pressures, determinid by the method of 

 Constant Volumes. (II. and IV.) 



ATMOSPHERIC AIR. 



Density of the airat 

 Pressure at 0°. Pressure at lllO". (F, that of air at 0', Volume of the air «t 



flliUimetres. Mdlimetres. under a pressure of lOU^, that at U° = 1. 



760 inlllinetres = 1. 



lOn-72 IW.-il 014+1 1-36482 



174-:il) 237 17 2-.'ll4 136513 



2'.6-l)l) 39.5 07 «b5nl 1 ■36542 



374 67 510-35 41130 1-365S7 



37.V23 610-97 0-4937 1-36572 



760-00 — 1 Ouon 1 36650 



167S-4I) 2-286 09 2-2C84 136760 



16D2 53 2306-23 2 2270 1-^,6800 



2144 IS 2924-14 2-"2l3 136894 



3655-56 4992-09 4-SlOO 1 37091 



753 47 

 901 119 

 1742-73 



3589-1)7 



CARBONIC ACID. 



11134-54 1 0000 



12311 37 1-IS79 



2187 72 2-2976 



4759 03 4 7318 



1 .16856 

 1 36943 

 1 37523 

 1-38698 



Dilatation of Gases under different Pressures, determined by the method of 

 Constant Pressures, (V.) 



Atmospheric Air. Carbonic Acid. 



H\di-ogen. 



Pressure Volume 



Willi, at 100°. 



760 1 -.167116 



25-25 1-36944 



2620 136964 



Pressure Volume Pressure Volume 



Wdli. at lOlF. Jlllli. at l(«l°. 



760 1 37099 760 1-36613 



2520 1-38455 2645 136616 



Sulphurous Acid. 



I * > 



Pressure Pressure Volume 

 at 11°. at 100"^. at lull". 

 '60 00 760-110 1-39020 

 982-73 987 64 1-39804 



The general conclusions of this memoir are as follows : — 



1st. The co-eflicient of dilatation of air, 0-375, heretofore ad- 

 mitted by philosophers from the experiments of M. Gay Lussac, is 

 much too great for dry air under the ordinary atmospheric pres- 

 sure. The co-efficient 0-36-t5, which is the mean of the experiments 

 published by M. Rudberg, is too small. When the co-eflicient of 

 dilatation of air is deduced by calculation, from the changes of 

 elastic force w hich the same volume of gas undergoes when carried 

 from 0^ to 100^, its value is 0-3665. But when this co-efficient is 

 deduced from the changes of volume of the same mass of gas in 

 passing from 0^ to 100, its elastic force remaining constant, we 

 find a value rather higher : that is — 0-3670. 



2nd. The co-efficients of dilatation of the different gases are not 

 equal, as has been hitherto admitted ; they present on the contrary, 

 notable differences, as may be seen by the numbers before cited. 

 There is often obtained for the same gas, very difl"erent values for 

 its co-efficient of dilatation, according as this is deduced immedi- 

 ately from the observation of the change of volume which the same 

 mass of gas undergoes between 0' and 100% its elastic force remain- 

 ing the same, or calculated from the variation in the elastic force 

 of the gas between 0' and 100', its volume remaining constant. 



3rd. The air and all other gases, except hydrogen, have greater 

 co-efficients of dilatation in proportion as their density increases. 



4th. The co-efficients of dilatation of the different gases approach 

 nearer equality as their pressures are lighter; so that the law 

 which is thus expressed, ^' all ga.ies have the same co-efficient dilata- 

 tion" may be considered as a limiting law which is applicable to 

 gases in a state of extreme dilatation ; hut which is farther from 

 the truth in proportion as the gases are more compressed, or, in 

 other words, as their molecules are brought nearer together. 

 (To he continued.) 



