October 20, 1923] 



NA TURE 



589 



178 and 180 with some 182, and has no odd isotopes. 

 Tantalum is 181 and 183, tungsten 184 only, and 

 element 75 would have 185 or 187 if they existed. 

 Iridium appears to have 191 as well as 193 ; platinum 

 has 194 and 196, possibly 192, but 198 is unlikely. 

 Mercury is 198, 199, 200, 202, and 204 and not 201 

 as Aston finds, but I cannot add 197 as he thinks 

 possible. 



Gold is 197 and simple if Aston is right about 

 mercury ; otherwise it should have 199 also. 

 Thallium is 203 and 205, lead principally 208 and 

 206, the former in excess. Bismuth is simple and 209. 

 Polonium is of course 210, and the only member of 

 element 84 with a chance (and that a very remote 

 one) of being isolated. Thorium is simple and 232. 

 One isotope of element 89 is too unstable ever to be 

 isolated. Element 91 has 231 and 233, the former 

 bemg probably protoactinium. Uranium, which is 

 complex, has been discussed in my paper in the 

 Philosophical Magazine. 



The order of intensity of the isotopes cannot be 

 given accurately from these considerations, but a 

 rough sorting into major and minor isotopes is not 

 difficult to make. Mass-numbers which belong to no 

 atomic number are difficult to estimate. At present 

 I feel sure of ten even ones and thirty - one odd 

 below polonium, most of which Aston has found. 

 All but one of the former are of the form 4« +2, and 

 more than three-fifths of the latter of the form ^n + i. 

 There appear to be at least thirty simple elements 

 if my predictions be added to Aston's certainties. 

 Fourteen of these have accepted atomic weights 

 within 0-05 of a unit, and as many fall short of a unit 

 by this amount as exceed it. I have assumed Aston's 

 whole-number rule in all the numbers given in this 

 letter. 



After these mass-numbers had been deduced I found 

 that the complexity of an element was apparently a 

 simple function of the atomic number i6n. Thus 

 there is a probability that elements of atomic numbers 

 i6n-f-7, i6n-l-io,and i6n+ 11 are simple; that i6»-f-3, 

 i6n + 5, i6w-M3,and i6w 4-15 have two isotopes ; that 

 i6w-f8 and 16^-1-14 have no odd isotopes; and that 

 I 6m, i6n + 2, and i6n+ 12 have odd isotopes. If this 

 deduction be substantiated by experimental work it 

 should throw light on the constitution and stability 

 of the nucleus. A. S. Russell. 



Dr. Lee's Laboratory, Christ Church, 

 Oxford, October 3, 1923. 



The Measurement of Very High Temperature. 



In 1914 Lummer ^ described some experiments on 

 an arc burning in a gas at high pressure. His method 

 of determining the temperature based on the increase 

 of surface brightness of the positive crater is extremely 

 unsatisfactory, and his figures, using his own values 

 of surface brightness, appear to be nearly three 

 thousand degrees too low. It seemed desirable, 

 therefore, to repeat and extend the experiments and 

 determine the temperature more precisely. A very 

 accurate way of doing this would be to determine the 

 ratio of the intensity of the light at two wave-lengths 

 as far as possible apart, which would define the 

 temperature if the positive crater were a complete 

 radiator. This a.ssumption need not be made if ratios 

 of the intensities arc determined at two different 

 temperatures, one of which is known. Thus, for 

 exa.mple, in the region in which Wien's law holds, if 

 OA is a constant proportional to the emissivity, the 

 intensity is given by 



' Lummer " Verflassigung cler Kohlc und Hcrstellung dcr Sonncti- 

 tcmperatur." (Sammlting Vicwcg.) 



NO. 28 t6, VOL. I 12] 



E\T = 





and log|^1^-c(l- ^)a -1), 



*= Ea.t, . Ea,t, VXj Xa/VTi Tg/ 



which determines Tj in terms of T^. Since a\, which 

 may also contain the sensitivity of the measuring 

 instrument, disappears from the final equation, this 

 method is very convenient and may be made very 

 accurate. 



Two methods were used for determining the in- 

 tensity, one by the use of a wedge as suggested by 

 Prof. Merton," the other making use of the photo- 

 electric effect. The first method is more convenient 

 in many cases ; the second is probably more accurate. 



The main difficulty is to make sure that one is really 

 observing the hottest part of the crater. It is very 

 difficult to keep the arc constant at high pressures, 

 and obviously too low a temperature will be found if 

 the arc shifts during the exposure so that part of the 

 measurement is carried out on the colder parts sur- 

 rounding the crater. If this has been avoided, com- 

 parison of the intensity at any two wave-lengths at 

 atmospheric pressure and at high pressure enables the 

 temperature at the high pressure to be calculated in 

 terms of the known temperature of the normal arc. 

 A check in the method is given by the constancy of 

 the temperature found using various wave-lengths. 

 The divergence from the mean is within the limits of 

 experimental error. 



Owing to the difficulty outlined above, observations 

 at the same pressure do not repeat very accurately, 

 though the highest values are fairly consistent. The 

 following table summarises the provisional results for 

 an arc in nitrogen : 



As already stated, these are minimum tempera- 

 tures ; and indications on one plate (10,000° at 50 

 atmospheres) seem to justify the suspicion that they 

 may be considerably underestimated. 



Further experiments making use of a number of 

 improved methods are now in progress, and it is hoped 

 shortly in a fuller publication to give more accurate 

 values for the temperature of the crater as a function 

 of the pressure and nature of the gas. The fact seems 

 certain, however, that one can by this means reach 

 temperatures in the laboratory considerably higher 

 than the temperature at the surface of the sun. 



It may be interesting to note here the strong 

 reversal of some of the cyanogen bands shown on the 

 plates within certain limits of pressure and tempera- 

 ture. The phenomenon is most noticeable between 

 30 and 40 atmospheres, and it should be possible to 

 locate these limits more definitely in the course of 

 the experiments. 



I have in conclusion to acknowledge a deep debt of 

 gratitude to Prof. Lindemann for much helpful 

 criticism and encouragement. I. O. Griffith. 



Clarendon Laboratory, Oxford, 

 September 22. 



» Merton and Nicholson (Phil. Trans. Roy. Soc. A. 217). Prof. Merton 

 kindly lent me the spectrometer and wedgo which be used in his own 

 investigations, and I take this opportunity of thanking him for the loan of 

 the apparatus and for his assistance in initiatin me into the details of his 

 method. 



