564 



NA TURE 



October 4, 1906 



A small circular dish is filled almost to overflowing 

 with water the surface of which will be convex. A 

 single floating magnet (with its axis vertical) placed on 

 this moves at once to the centre ; two or more such 

 magnets placed on it form regular equilibrium figures, as 

 in the usual form of the experiment. The chief interest 

 of the modification arises, however, from the fact that the 

 figures are not in general the same as in the ordinary 

 arrangement. This is instructive, because it brings out 

 clearly the necessity of knowing the exact law of force 

 between the parts of an atom before it can be possible to 

 predict its structure. 



The experiment is so easily tried by anyone that there 

 is no need to go into great detail here ; but it may be 

 mentioned that with the particular dish and magnets used 

 by me it is possible to arrange ten in a single ring with- 

 out any central nucleus, and that in a larger dish more 

 can, of course, be so arranged. ITiese ten also form stable 

 groups as a ring of nine with one in the middle, or a 

 ring of eight with two in the middle. But a ring of seven 

 with three in the middle is not possible ; if temporarily so 

 placed one of the three gradually moves out and joins the 

 seven. The first arrangement in three groups occurs for 

 eighteen magnets in all ; these are stable when placed with 

 twelve in an outer ring, five in an intermediate ring, and 

 a single one in the centre. Alfred W. Porter. 



University College, London, September 17. 



Chemical and Electrical Changes induced by Light. 



The issue of Nature for August 30 (p. 455) contains an 

 abstract of a paper read before Section A of the British 

 Association by Sir \Vm. Ramsay and Dr. J. F. Spencer 

 on the chemical and electrical changes induced by ultra- 

 violet light, in which the " fatigue " shown by certain 

 surfaces is discussed. I have for some time been engaged 

 in an investigation of the fatigue shown by metals for the 

 photoelectric effect, and have made a careful examination 

 of the rate at which the photoelectric current decays in the 

 case of a zinc plate, polished or amalgamated. A large 

 Nernst lamp supplied with current from storage cells was 

 used to give a steady source of light. The decay immedi- 

 ately after exposure to the light was very rapid, but after 

 about twenty minutes became much slower. For a change 

 taking place according to the " compound interest law," 

 as in the case of a monomolecular chemical reaction or a 

 single purely surface effect, we know that the curve can 

 be represented by an exponential term involving the time. 

 In the case of zinc, I find that the activity at any instant 

 can be represented with considerable accuracy by the sum 

 of two exponential terms. It is possible to interpret this 

 result somewhat on the lines followed by Rutherford in 

 explaining the decay of the excited activity of radium or 

 thorium, by supposing that a succession of changes takes 

 place. 



Similar results have been obtained in the case of 

 aluminium, and also with specimens of coloured fluor- 

 spar. In the latter case the colour is attributed to the 

 presence of particles of reduced metal. 



It is interesting to note that the longer waves of light 

 tend to produce a change in the opposite sense, so that 

 the rapid decay at first observed on exposure to light may 

 be followed by a small increase in activity unless the long 

 waves are absorbed by a solution of alum. 



These experiments were carried out partly in the labor- 

 atory of Lord Blythswood, to whom my thanks are due, 

 and partly in the Wheatstone Laboratory of King's College. 



H. S. .'\llex. 



King's College, London, September 21. 



The Rusting of Iron 



TiiK experiments made by Mr. J. Newton Friend, and 



described by him in Nature of .September 27, confirm 



similar experiments previously made by me, and furnish 



further evidence that the rusting of iron is primarily a 

 result of acid attack. That cast iron, a very complex 



NO. 1927, VOL. 74] 



material frequently containing a high percentage of sulphur 

 and phosphorus, decomposes hydrogen peroxide " with 

 astonishing rapidity," and that the metal becomes covered 

 with rust in a few minutes, is not, however, to be referred 

 to catalytic action, as Mr. Friend suggests, but is a conse- 

 quence of the formation of acids by the oxidation of some 

 of the impurities present in the iron, and of the subsequent 

 electrolytic action. -As Mr. Friend says, " the purer the 

 iron the less is the action of the pero.xide upon it," which 

 is another way of slating that the intensity of action will 

 be determined by the amount of acid formed on the surface 

 of each particular sample of metal when in contact with 

 the peroxide. 



Cast iron is known to oxidise in air more readily than 

 wrought iron, and this is probably due to the former con- 

 taining impurities which on oxidation yield acids. The 

 rust formed on cast iron exposed to air often contains 

 appreciable quantities of combined sulphur. 



The fact that cast iron is attacked by water in absence 

 of air, becoming darker in colour, whilst pure iron under 

 identical conditions remains unchanged, may also be re- 

 ferred to the production of a minute quantity of acid. In 

 this case the acid is not formed by oxidation, but it is 

 probablv hydrogen sulphide resulting from the interaction 

 of sulphides, such as silicon sulphide, contained in the 

 crude iron, with water. Gerald T. Moody. 



Central Technical College, October i. 



Remarkable Rainbow Phenomena. 



May I be permitted, with reference to Mr. Spence's 

 observation of a remarkable rainbow, described in your 

 issue of September 20, to direct attention to a number of 

 phenomena of the same kind observed in Holland during 

 the last ten years, and published by the Dutch Meteor- 

 ological Institute in Omiveders, Optische Verschijnselenens. 

 At Fort William, also, on August 16, 1887, a pheno- 

 menon of this sort was seen, a drawing of which is to be 

 found in Trans. Roy. Soc. Edin. (vol. xxxiv., p. xvii. 

 Fig. 17). Readers of Nature will find an observation of 

 a double rainbow, with drawing, similar to the oval de- 

 scribed by Mr. .Spence, made by Prof. Tait on .September 

 II, 1874, in the issue of October i, 1874, with a comment 

 by Maxwell upon it. 



The explanation of the phenomenon is simple, and seems 

 to have been first given by Rubenson. The upper of the 

 two ordinary and the two secondary bows is generated by 

 rays which enter the raindrops after reflection from a 

 level of water situated behind the observer. It is obvious 

 that the altitude of the ordinary rainbow being 42° — /i, the 

 altitude of the one generated by reflection will be 42°4-'!, 

 h being the sun's altitude; the same holds good for the 

 secondary rainbow. The centres of all the bows lying in 

 the same vertical, it is clear that the two ordinary bows 

 and the two secondaries touch each other at the horizon. 

 For further information see my " Meteorologische Optik " 

 (pp. 401 and 555). J. M. Pernter. 



\''ienna, September 28. 



Fugitive Coloration of Sodalite. 



\\ iTH reference to the properties of Indian sodalite shown 

 by Mr. T. H. Holland at the York meeting of the British 

 Association (September 27, p. 550), will you permit me to 

 point out that, although not generally noticed in the text- 

 books, the change of colour referred to is not peculiar to 

 the Rajputana mineral. The first sodalite discovered had 

 the same property, and Gieseck^, under date .August 28, 

 1806, records the occurrence of " pfirsichbliithenroth- 

 farbene " sodalite from Kangerdluarsuk, in Greenland, 

 " welche die hohe Farbe auf frischem Bruche sogleich 

 beinahe ganz verliert." The same observation was made 

 independently by Allan (Thomson's " Annals of Philo- 

 sophy," 1813, vol. i., p. 104); but I am not aware that 

 there is any record of a recovery of the lost colour, which 

 Mr. Holland appears to have observed. 



Jas. Currie. 



Edinburgh, October i. 



