October 28, 1909] 



NA TURE 



5'9 



10 DETERMINE THE REFKACTIVITY OF 

 GASES AVAILABLE ONLY IN MINUTE 

 QUANTITIES. 



ON a former occasion ' I described a refractometer 

 capable of dealing with rather small quantities 

 (12 c.c.) of gas. The optical tubes, one of which 

 would contain the material under investigation and 

 the other air, were of brass, 20 cm. in length and 

 () mm. in bore, and were traversed by two pencils of 

 light from the same origin, subsequently brought to 

 interference in the observing telescope. For this pur- 

 pose the object-glass of the telescope was provided 

 with two parallel slits opposite the axes of the tubes. 

 The image of the original slit, formed in the focal 

 plane, was examined through a high-power cylindrical 

 lens, constituting the eye-piece of the telescope, and 

 exhibited the familiar pattern of interference bands 

 the position of which shifts with changes in the densi- 

 ties of the gases occup^'ing the tubes. With this 

 apparatus, and using pressures not exceeding one 

 atmosphere, it was possible to compare refractivities 

 (ju— i) with a relative accuracy of about one- 

 thousandth part. 



In recent conversation my son, the Hon. R. J. 

 Strutt, raised the question as to the minimum 

 quantity of gas upon which a determination of re- 

 fractivity could be made, having in mind such rare 

 gases as the radium emanation. Towards answer- 

 ing it I have made a few experiments dealing merely 

 with the optical side of the question. 



A reduction of volume in the gas tube implies a 

 reduction of length below the 20 cm. of the apparatus 

 Just referred to. and this carries with it a loss of 

 accuracy. A reduction to 2 cm. should leave possible 

 an accuracy of at least i per cent., and this was the 

 length chosen. As the inquiry was limited to the 

 optical conditions, it was unnecessary to close the 

 ends, and thus the tubes reduced themselves to two 

 parallel tunnels through a block of paraffin 2 cm. 

 thick. Thev were prepared by casting wax (from a 

 candle) round two similar sewing needles of suitable 

 diameter previously secured in a parallel position. 

 The rest of the apparatus was merely an ordinary 

 spectroscope arrangement (without prism). Sun- 

 light admitted through a slit, and rendered parallel 

 bv the collimating lens, traversed the double tunnel, 

 and was received by the observing telescope focussed, 

 as usual, upon the slit. It is necessary, of course, 

 that the length of the slit be perpendicular to the 

 plane containing the axes of the tunnels. 



The appearance of the bands as seen with a given 

 telescope depends upon the size of the apertures and 

 upon their distance apart. The width of the bands 

 is inverselv as the distance between the centres of 

 the apertures (tunnels), and the horizontal diameter 

 of the luminous field upon which the bands are seen 

 is inversely as the diameter of the apertures them- 

 selves. Since a large number of bands is not re- 

 quired, small and rather close apertures are indicated. 

 The onlv question is as to the amount of light. If 

 we suppose the apertures and their distance apart to 

 be proportional, we may inquire as to the effect of 

 linear scale L. Here a good deal may depend upon 

 the relative values of length of slit, focal length of 

 collimator, length as well as diameter of tunnels. In 

 mv apparatus the slit was short, and the height, as 

 well as the width of the field of view, was determined 

 mainly by diffraction. If we suppose the slit very 

 short, the calculation is simplified, though this 

 cannot be the most favourable arrangement. 

 With a given width of slit the whole light in 

 the field of view is then proportional to L'. 

 Since the angular area of the field practically 



1 Proc. Roy. Sor. . vol. Ixiv., p. 95, 1898 ; Scientific Pap«rs, iv., p. 3^4. 

 NO. 2087, VOL. 81] 



varies as L--, it would seem that the brightness 

 varies as L'. This would impose an early limit upon 

 the reduction of L ; but there are other factors to be 

 regarded. In order to secure an angular field of 

 given size, we must use an eye-piece the magnify- 

 ing power of which is proportional to L. This con- 

 sideration changes L' back to L-. Nor is this all. 

 With a given eye-piece the admissible width of 

 primary slit varies inversely as L, and thus, finally, 

 the brightness of a field of given angular width, and 

 containing a given number of bands, varies as L 

 simply. 



In the earlier experiments the tunnels were of 

 J mm. bore, and were too widely separated. In 

 order to see the bands well, a very powerful eye- 

 piece was needed. An attempt to gain light by sub- 

 stituting a cylindrical lens (very successful in the 

 former apparatus, whore the beams are limited by 

 slits) for the spherical lenses of the eye-piece showed 

 little advantage. Subsequently smaller tunnels were 

 prepared i mm. in bore, and so close that the dis- 

 tance of the nearest parts was rather less than the 

 diameter of either. These gave splendid bands with 

 the ordinarv eve-piece of the spectroscope, and I 

 estimated that there should be no difficulty in setting 

 a web correctly to one-twentieth of a band. 



The capacity of one of these passages is about 

 4 cubic millimetres, and I have no doubt a further 

 reduction might be efl'ected, so far as the optics is 

 concerned ; but the further such reduction is carried 

 the greater, probablv, would become the difficulties 

 of manipulation. The mere closing of the ends of 

 such small tubes with plates of glass would not be 

 an easv matter. In order to prevent encroachment 

 upon the course of the light, it might be necessary 

 to enlarge the ends so as to allow a little more room 

 for overflow of cement. For the present I content 

 myself with showing that it is possible to obtain 

 well-formed black bands on a sufficient angular scale 

 with light which has traversed tubes 2 cm. long and 

 i mm. in bore. R.^yleigh. 



G 



GAY-LUSSAC'S LAW— ITS CENTENARY. 

 .\Y-LUSSAC'S law regarding the composition of 

 gases by volume was made known about a 

 hundred years ago. The paper in which he elaborated 

 it, having been read to the SocietiJ philomatique on 

 December 31, 1808, was published in the M^moires 

 de la Society d'Arceuil in the following year. Since 

 then the law has come to have a history of its own. 

 Chemists were at a loss and made many efforts to 

 get it and the atomic theory to suit one another, and 

 the place of the law in science, though not now likely 

 to change, was for long unsettled and dubious. 



Ostwald puts the law in the following way: — " If 

 several gases appear or disappear in a chemical 

 change, they do so always in simple ratios by volume." 

 For example, one volume of hydrogen and one of 

 chlorine yield two volumes of hydrochloric acid, and, 

 again, one volume of oxygen and two volumes of 

 hvdrogen give two of steam. 



Tlie composition of nitrous and nitric oxides and 

 numerous other gases was discovered for the first time 

 by Gay-Lussac. But no one who has paid much atten- 

 tion to the history of science can be surprised that ob- 

 servations had been made earlier in the same field. 

 William Higgins knew that hydrogen and oxygen 

 combine, yielding water, in the proportion 2 : i, and 

 this was probably only a version of Cavendish's result. 

 He knew also Austin's experiment proving that 

 sulphuretted hydrogen contains its own volume of 

 hydrogen, and he had himself observed that sulphur 

 dioxide contains its own volume of o.xyyen. It must 



