December i, 192 i] 



NATURE 



433 



in many cases returns when the pressure is reHeved. 

 The appearance then suggests a lamp alternately ex- 

 tinguished and relit. 



The same precautions as to the smallness of the 

 test-pieces and the absence of a comparatively soft 

 bed under the coloured layer have to be observed in 

 the case of metallic feathers. 



The colours of anilin films are not affected by pres- 

 sure otherwise than by the optical contact between 

 them and the upper quartz lens. 



The same may be said of the colours of tempered 

 steel, etc. A simple experiment is to heat a piece of 

 polished steel and to remove by polishing a narrow 

 band of the coloured part. Then form a series of 

 Newton's rings, having this band as a diameter. It 

 will be found that in the straw-yellow of the tempered 

 part the rings are scarcely displaced as they cross 

 the band. Further, it may be noticed that when any 

 of the colours are gradually polished off the colour 

 does not change as the thickness of the layer is 

 reduced but merely becomes fainter. 



From this it may be gathered that the colours of 

 tempering are not due to interference, at any rate 

 in the ordinary sense in which the word is used. 



A. Mallock. 



The Tendency of Elongated Bodies to Set in the North 

 and South Direction. 



The letters of Sir Arthur Schuster and Col. E. H. 

 Grove-Hills in Nature of October 20 and Novem- 

 ber 24 under the above heading are interesting, as is 

 also the article by Mr. W. D. Lambert in the 

 Atnerican Journal of Science for September last, to 

 which reference is made by the latter, but the ex- 

 tremelv weak gravitational force dealt with cannot 

 possiblv have any appreciable effect on my apparatus, 

 and is certainly quite inadequate to account for the 

 results of experiments made with it. Indeed, I 

 gather from some of Sir Arthur Schuster's remarks 

 that .he feels that this is the case. The force that 

 I have found to exist is another altogether, as I am 

 confident anyone who has watched the experiments 

 under favourable conditions would admit. 



Mr. Lambert in his article is referring to a sus- 

 pended rod of 40 cm. length, weighted at both ends, 

 and finds for this that the force was only i '400,000th 

 of a dyne or i '4oo,ooo4pooth of a gram. Now, 

 what conceivable effect can such a force have 

 on a vertically suspended circular ring or disc 

 of the thinnest paper, or other light material, 

 of only -^5 cm. diameter? And yet it is with these that 

 some of the most decisive results of my experiments 

 have been obtained. Under favourable conditions the 

 N. and S. directive force acting on the disc is by 

 no means insignificant, and, indeed, when the skv 

 is clear, the barometer high and steady, and there is 

 no wind, it is frequently strong enough, when the 

 apparatus is placed on high level ground in the open, 

 to cause the disc to come to rest in the true N. and S. 

 direction in less than five minutes, or to oscillate 

 fairlv rapidlv (about five seconds interval a few 

 degrees on eith^^r side of this line. Moreover, other 

 conditions are inconsistent with any purely gravita- 

 tional theorv being the explanation. I have used both 

 susoended discs and pointers floating on oaraflfin oil, 

 and obtained about the same results from both ; 

 not that one gives the E. and W. direction and the 

 other the N. and S. The only difference noticed was 

 that the floating pointer took longer to come to the 

 N. and S. line. Further, it is not necessarily the 

 elongated diameter that turns N. and S., as Sir 



NO. 2718. VOL.'loSl 



Arthur Schuster assumes. For instance, if a strip 

 of aluminium about i in. wide is coiled round into 

 the form of a ring and then flattened so that, the 

 sides almost touch each other, when suspended ver- 

 ticalh* the flat-surface sides turn N. and S., as in 

 the case of the ordihar\' discs. Now, if these flat sur- 

 faces are pulled out so as to form a ring with the 

 elongated horizontal diameter several times as long 

 as the breadth of the strip, it is not this elongated 

 diameter that turns N. and S., but the surface breadth 

 of the strip as before. In fact, there is no alteration 

 in the direction, but the surface sides of the aluminium 

 strip take up the same direction as when they were 

 flattened close together. 



In rainv weather, or when the sky is clouded or 

 the barometer is falling, no satisfactory results can 

 be obtained, and the disc will then usually turn ap- 

 proximately towards the bearing of the thickest 

 clouds. From repeated experiments it has been found 

 that when the apparatus is placed in an iron bucket 

 covered with an iron lid, the N. and S. directive force 

 ceases to act upon the disc, although it may be fairly 

 strong in the open at the same time. The force 

 appears to be strongest in still, frosty weather, when 

 the sky is clear, even though there may be some 

 ground mist. 



Since the apparatus was exhibited at the Royal 

 Society's conversazione in May of last vear many 

 more experiments have been made bv obser\'ers in 

 different parts of the wcwrld. extending from Spits- 

 bergen in latitude 80° N. to the top of the Cameroon 

 Mountain (13, -553 ft.) close to the f^quator. and from 

 Canada and the United States to the Red Sea. all of 

 which give much the same results that I and others 

 have obtained in this country: so whatever may 

 happen in other parts, there can be no doubt that 

 over this wide area of the earth'*; surface this N. and 

 S. directive force does exist. In one or two other 

 cases the results have been indefinite, owing apparently 

 to the observers having no proper apparatus or the 

 weather conditions being unfavourable. It is perhaps 

 too earlv vet to soeak positively about the nature of 

 this N. and S. directive force, but^ the conclusion I 

 have come to is that, in the main, it must be electric, 

 and is probablv due to the earth rotating in an electric 

 field. E. A. Rekves. 



Roval Geographical Society, 

 Kensington Gore, S.W. 



Table for the Duration of Sunset. 



The subioined table was made thus : — From 

 Table VI. in the American Nautical Almanac, for 

 each date and latitude indicated, the hour of sunset on 

 the meridian of Greenwich is taken and corrected for 

 equation of time, giving an hour-angle precise within 

 im. or 15' (as both local mean time and equation of 

 time are rounded off to the nearest minuteV From 

 latitude, declination, and this hour-angle the semi- 

 duration of sunset in arc minutes is computed b}' the 

 differential formula. 



(i) dP,=cos h d/i/cos * cosS sin Pj, 



in which P,=hour-angle, 7i = altitude of sun's centre, 

 (6 = latitude, 8 = declination, and dh = sun''? semi- 

 diameter. As the hour-angle found from the Nautical 

 Almanac is for the end of sunset, it is corrected by 

 subtracting this approximate semi-duration, and the 

 final value in mean time seconds is found by 



(2) dP = 8 cos h dhjcos ^ cos S sin P, 

 in which all sines and cosines refer to mid-sunset, 



