November i8, 1920] 



NATURE 



375 



far more rapidly with a change of composition than 

 do the other properties, diffusivity, tensile strength, 

 and Young's modulus, which make up the thermal 

 endurance of the glass. The coefficient of expansion 

 of German laboratory glassware is something more 

 than 10 per cent, lower than that of English, and 

 this fact alone would account for the breakage of 

 the English glass with drastic heat treatment. 



There is no particular difficulty attached to the 

 manufacture of glass having a low coefficient of 

 expansion and high thermal endurance, and we may 

 take it that the English manufacturers have decided 

 that the maximum resistance to attack by reagents is 

 to be desired, and have, accordingly, sacrificed thermal 

 endurance to a small extent in order to obtain the 

 encouraging results outlined in Mr. Jenkinson's letter 

 in Nature of October 28. 



The whole question is largely one of general policy. 

 If the chemists of this country prefer to use a glass 

 of higher thermal endurance but with less resistance 

 to reagents, then I have little doubt that the British 

 inikers would supply it. E. .A. CoAD Pryor. 



Milford, Park Road, Teddington, November 8. 



The Separation of the Element Chlorine into Normal 



Chlorine and Meta-Chlorine, and the Positive 



Electron. 



In commenting on a letter under the above heading 

 from Prof, ilarkins in Natike of \or\\ 22 last, I 

 remarked that his assumption that " the hydrogen 

 nucleus or the positive electron has, according to these 

 papers, a weight, and presumably a mass, of 1000, 

 on the basis of oxygen as i6o<x)," was contradicted 

 bv experiment. 



Prof. Markins has pointed out to me that this 

 assumption was made with the reservation " when- 

 ever the positive electron is combined in a complex 

 atom," which I was careless enough to overlook. I 

 wish, therefore, to apologise for my remark and 

 withdraw it unreservedly. 



I very much repret that this apology comes so late, 

 but the delay is due to the fact that Prof. Harkins 

 led me to understand that he was himself publishing 

 a statement on the matter in NATirRE, to which 

 I could reply. He now tells me he has decided not 

 to do so. F. W. Aston. 



Trinity College, Cambridge, November 9. 



The Stereoscopic Appearance of Certain Pictures. 



In going round a picture gallery it will be noticed 

 that in certain pictures the objects delineated appear 

 to stand out in a similar manner to those seen with a 

 stereoscope. A picture of this kind has the effect of 

 making those surrounding it appear very flat by 

 comparison. This appearance is not the characteristic 

 of a particular artist, because in an exhibition, all 

 the paintings being by one man, only one or two may 

 hr found which have this stereoscopic appearance. 



The majority of the pictures will be iorrect in 

 drawing, perspective. light and sh.ide, but it will be 

 noticed that this will be correct only for one eve — 

 that is to say, the picture is quite correct for efther 

 eye when the other is closed. Those pictures, how. 

 ever, which have a stereoscopic appearance are 

 painted so that the representation is as nearly as 

 possible a delineation as seen by the combined retlnic, 

 anv disturbing element which would do away with 

 this illusion being eliminated. For instance, a 

 revolver pointed straight at n person so fh.i* the 

 centre of the barrel is pointing directly between the 

 two eyes wjll be seen quite difTerentlv with the two 

 eyes. When the right eve is closed the left-hand 



NO. 2664, VOL. 106] 



side of the barrel will be seen, the right-hand side '' 

 being invisible. When the left eye is closed the right- 

 hand side is seen, and the left-hand side of the barrel 

 is invisible. When, therefore, the barrel of the 

 revolver is foreshortened and both sides are visible, 

 the muzzle being pointed directly at the observer, the 

 appearance is such as could be seen only with both 

 eyes. The stereoscopic appearance is then very 

 striking, and the revolver appears to follow the 

 observer and to be pointed directly at him, no matter 

 what position he takes up with regard to the picture. 

 In all the pictures or portions of them which present 

 this stereoscopic appearance it will be noticed that 

 the appearance as seen by both eyes is represented, 

 the left-hand side of the picture being represented as 

 seen by the left eye, the right-hand side as seen by 

 the right eye. 



In a demonstration which I gave before the Physio- 

 logical Society (Journal of Physiology, vol. xlviii., 

 1914) I showed that the perception of binocular relief 

 is independent of double images and the stimulation 

 of disparate points, provided that the object presents 

 images to the two retinae similar to those which are 

 presentetl by an object in the field of vision. This 

 can be shown by taking a pair of stereoscopic photo- 

 graphs in which the point of sight is at the centre of 

 each and cutting them vertically in two, and then, 

 having pasted the left half of the left photograph on 

 the left side and the right half of the right photo- 

 graph on the right side on white or black cardboard 

 at an appropriate distance, so that there is no over- 

 lapping when placed in the stereoscope, a picture in 

 striking relief is obtained when combined together in 

 the stereoscope. In this case it will be noticed that 

 there is no portion common to both fields of view. 

 In each case the overlapping portion is combined with 

 white. It seems probable that this is how binocular 

 vision takes place in ordinary circumstances. If an 

 object in high relief — as, for instance, a vase or the 

 face of a person — be viewed at a short distance and 

 one particular point fixated, it \vill be noticed that 

 the right eye dominates the right side of the field of 

 vision and the left eye the left side. The image seen 

 is almost entirely that of the right eye for the right 

 side and that of the left eve for the left side, as mav 

 be proved by noticing the relation of surrounding 

 objects, and closing first one eve and then the other 

 alternately. F. W. Edridgk-Green. 



The Energy of Cyclones. 



It does not seem to me as though any reallv satis- 

 factory theory has yet been put forward to explain 

 the genesis and maintenance of cyclones; I fully 

 agree with Mr. Deeley (November 11, p. 345) that 

 thfv are not due to contiguous masses of air at 

 different temperatures, but, on the other hand, I do 

 not see how they can originate in an inert and stable 

 region like the stratosphere^ 



Were storms produced by contrasts of temperature 

 — or, in other words, by the so.calle<l polar front 

 surely they would be most violent where the contrast 

 was most marked. The stormiest parts of the world 

 are the great belt of the southern ocean from 40° to 

 60° S. lat. and that part of the .Atlantic which lies 

 north-west of Scotland, and neither of these regions 

 shows any exceptionally steep gradient of fempernturc. 



Observations in the upper air have shown a remark- 

 able uniformity in the mean temperature (mean with 

 regard to height) from o to 20 km. in everv place 

 where they have been obtained, and it follows as a 

 corollary that there is a verv uniform pressure at 

 20 km. heitjht over the globe, for the pressure at 

 20 km. is .nlmost independent of the surface pressure. 



