October 24, 1907J 



NA TURE 



63: 



LETTERS TO THE EDITOR. 

 [The Editor does not bold himself responsible jor opinions 

 expressed by his correspondents. Neither can he undertake 

 to return, or to correspond with the writers of, rejected 

 manuscripts intended for this or any other part of Nature. 

 No notice is taken of anonymous communications.] 



Upper Air Research in Egypt. 

 The Hehvan Observatory (under Uie .Survey Department 

 of Eg-vptJ has recently acquired apparatus to enable it to 

 join in the study of the upper regions of the atmosphere. 

 As a commencement, some fifteen ascents of small " pilot " 

 balloons were made during the month of August. On 

 three occasions the balloons were watched to a height of 

 5000 metres, but rather a large percentage of the balloons 

 used burst at much lower altitudes. A better type of 

 balloon has been ordered, and it is hoped that observations 

 mav be regularly made with them up to 5000 metres or 

 6000 metres. The balloons are of 2 feet nominal diameter, 

 and are filled with dry hydrogen made from zinc and 

 sulphuric acid. They are observed as they ascend by two 

 observers at the ends of a base line i kilometre in length. 

 The theodolites are of the very convenient type made by 

 S. and A. Bosch, of Strassburg, specially for this 

 work. The result of a month's work shows that at this 

 season the surface wind (N. to N.W. as a rule) is from 



1000 metres to 2000 metres thick. Above this there is a 

 layer of varying thickness of winds from \V. to W.S.W., 

 whilst above 4000 metres other winds are reached, but 

 the number of observations is at present too few to 

 generalise about this region. An interesting ascent is 

 shown in the figure, which represents the horizontal pro- 

 jection of the flight of a balloon on August 27, with con- 

 tours showing the position of the balloon at intervals of 

 500 metres. In this case, above the stratum of S.W. wind 

 there was a layer of N.U'. winds, whilst above this a 

 S.W. current was again entered. 



Besides this study of the winds, kite ascents will very 

 shortly be commenced. The apparatus acquired for these 

 consists of a winch of the pattern designed by Mr. W. H. 

 Dines, F.R.S., driven by a Crossley petrol engine of 

 4 horse-power. The observatory is indebted to Mr. Dines 

 and also to Mr. J. E. Petavel,' F.R.S., who watched the 

 manufacture of the winch and introduced many minor 

 alterations which use of a similar machine at GIossop had 

 suggested. The machines are housed in an iron building 

 on the flat desert plateau behind the observatory. 



During the September international days {September 4, 

 5, and 6) five ascents of pilot balloons were made. The 

 height to which they were followed varied between 2500 

 metres and 3300 metres. 



Hehvan. B. F. E. Keeling. 



Newton's Rings in Polarised Light. 



An erroneous statement regarding the above-mentioned 

 subject is made in Preston's " Theory of Light " (p. 363, 

 1901 edition) and also in Edser's " Light " (p. 519, 1902 

 edition). As the error is a rather serious one, it seems 

 worth while to point it out. 



When the rings are seen between two lenses of the 

 same substance, by light polarised perpendicularly to the 

 plane of Incidence, reflected at an angle greater than the 

 polarising angle of the substance, it is stated that the 

 centre of the rings is bright. That this is wrong can be 

 seen. For : — 



(i) Stokes has shown from the principle of reversibility 

 that, whatever be the nature of light, the centre of the 

 rings seen between lenses of identical refractive indices is 

 black at all incidences of the light. 



(2) Since the centre of the rings is black at all incidences 

 for common light and for light polarised in the plane of 

 Incidence, it follows by resolution that It is also black 

 when the light is polarised in a perpendicular plane. 



(3) When the angle of Incidence is less than the polar- 

 ising angle, the coefficients of reflection in glass and in 

 air at the bounding surfaces of the two media are opposite 

 in sign. It is argued that, on increasing the incidence, 

 the coefficient of reflection in air changes sign as the 

 polarising angle is passed, and therefore at such incidences 

 the two coefficients agree in sign, and destructive Inter- 

 ference no longer takes place. Really, however, it appears 

 from Fresnel's formula (coefiicient= — tan (i-r)/tan (i-(-r) 

 that both the coefficients change sign as the incidence 

 passes through the polarising angle, and therefore con- 

 tinue to differ in sign, as can be directly shown from the 

 principle of reversibility. Destructive interference does, 

 therefore, take place. 



(4) I have shown by experiment that the statement is 

 not true. 



(5) .\lrv has shown (Lloyd's "Wave Theory," p. 178, 

 and Jam'in's " Optlque Physique," p. 503) that when 

 the two lenses differ in refractive index, the centre of the 

 rings seen in light polarised perpendicularly to the plane 

 of incidence is white only when the Incidence lies between 

 the angles of polarisation of the two media. Outside these 

 limits the centre Is dark. 



C. V. Raman. 

 Science Association Laboratory, Calcutta, 

 September 12. 



Mr. Raman's criticism of the statement made on p. 519 

 of mv " Light for Students " is quite justified. Some time 

 a<»o I noticed the error myself, and devised the follo^ying 

 experiment, to which the same objections cannot be raised, 

 while at the same time it is more easily performed than 

 that in which two lenses of different refractive indices are 

 used. 



An ordinary black tea-tray Is filled with tap water, and 

 the surface is then touched by the end of a glass rod 

 which has been diopcd in oil (I find that the heavy naraflfip 

 oil used for engine lubrication answers well). The oil 

 spreads over a falrlv large area, Newton's rings being 

 exhibited round the edge of this. On viewing the colours 

 through a Nlcol at an angle slightly greater than 45 _ no 

 change Is produced when the light transmitted Is polarised 

 In the plane of Incidence, but on turnlne the Nlcol through 

 a right angle the colours change to their complementaries. 

 In this rase the light is reflected from the lower surface 

 of the film at an angle slightly greater than the angle of 

 polarisation for that surface, while it Is reflected from the 

 unper surface of the film at an angle less than the polar- 

 ising angle. I presume that if Lloyd's sintrle mirror 

 fringes were observed through a Nlcol. a similar change 

 would occur on rotating the Nicol : I should he obliged 

 If anyone who has tried this experiment would let me 

 know whether this actually occurs. Edwin Edser. 



Thermodynamics of Diffu"^ion 



In his review of " Thermodynamics " (N.ature, July 25) 

 and again In the Philosophical Magazine for July, Mr. 

 Burbury directs attention to a result stated by me regard- 

 ing the gain of entropy resulting from slow diffusion of 

 gases at constant pressure and temperature. 



Mav I direct attention to the context In connection with 



NO. 1982, VOL. 76] 



