224 



SCIENCE 



[Vol. LV, No. 1418 



clear that the exact order of the chemically 

 determined atomic weights is of little signifi- 

 cance and that the anomalies such as argon 

 and potassium and possibly too of tellurium 

 and iodine as well as nickel and cobalt are 

 merely due to the unequal relative proportions 

 of their constituent isotopes. 



From a consideration of the total abundance 

 of various elements Harkins^^ pointed out that 

 for the great majority of possible configura- 

 tions it would probaljly be found that even 

 atomic weight was associated with even atomic 

 number and odd with odd. The results given 

 in Table I, it will be seen, support this view. 

 Of the halogens (odd atomic numbers) all sis 

 isotopes are odd. Of the alkali metals (odd 

 atomic numbers) seven istopes are odd and 

 only one even. On the other hand, of the iso- 

 topes of the inactive gases (even atomic num- 

 bers) fifteen are even and but three odd. This 

 means that in the nuclei of most types of atom 

 the number of electrons is an even number. 



(d) The spectra of isotopes. 



In an attempt made by Harkins, Arouberg 

 and Gale^° to see whether any method of dis- 

 tinguishing between the isotopes of an element 

 could be obtained from a study of their spectra 

 it was found that the wavelength of the line 

 1 = 4058 A.U. as obtained from radiolead was 

 0.0044 A.U. greater than that from ordinary 

 lead. A similar result was also obtained by 

 Merton.^^ It has been pointed out however 

 that this difference is about one hundred times 

 greater than that predicted on the basis of 

 Bohr's Theory of Radiation. Loomis'^ also 

 has drawn attention to the unexpected satel- 

 lites which Imes^° found beside each line of 

 the HCl absorption band at 1.76fj., and which 

 measurements of his curves show to have an 

 average wavelength of 16.4 A.U., longer than 

 the lines which they accompany. These satel- 

 lites Loomis has shown can be accounted for by 

 assuming them to be due to the heavier of the 

 isotopes of chlorine of weights 35 and 37. On 



i" Harkins, Nature, April 4, 1921. 

 18 Harkins, Aronberg and Gale, Jl. of the Am. 

 Chem. Soc., July, 1920, Vol. 42, p. 1328. 



1' Merton, Proe. Roy. Soc, 96A, p. 388, 1920. 

 IS Loomis, Nature, Oct. 7, 1920. 

 1^ Imcs, Astropliys. Jl., Nov., 1919. 



this basis, his calculations show that the differ- 

 ence between the wavelength of the main line 

 and its satellite should be 13 A.U., which it 

 will be seen is in good agreement with observa- 

 tions of Imes. 



(e) Structure of atomic nuclei. 



By far the most important conclusion which 

 can be drawn from the results recorded in 

 Table I is that, with the exception of hydrogen, 

 the weights of the isotopes of all the elements 

 measured and, therefore ahnost certainly of all 

 elements, are whole numbers, within the 

 accuracy of the experiments — namely, about 

 one part in a thousand. This result carries 

 with it the possibility of greatly simplifying 

 our ideas of mass. The original hypothesis of 

 Prout, put forward in 1815, that all atoms 

 were themselves built of atoms of protyle, a 

 hypothetical element which he tried to identify 

 with hydrogen, has been established on a new 

 basis with the modification that the primordial 

 atoms are of two kinds — atoms of positive and 

 negative electricity. The unit of negative elec- 

 tricity, the electron, we have long been familiar 

 with, but the unit of positive electricity, which 

 also appears to he the real unit of mass, has 

 remained unidentified experimentally until now 

 as the positive nucleus of the atom of hydro- 

 gen. To this unit of mass and of positive elec- 

 tricity the name of "proton" has been given. 



This profound modification of our views of 

 the nature of mass has been very clearly set 

 forth by Aston. The Rutherford atom 

 whether in Bohr's or Langmuir's development 

 of it consists essentially of a positively charged 

 central nucleus around which are set planetary 

 electrons at distances which are great compared 

 with the dimensions of the nucleus itself. As 

 has been stated the chemical properties of an 

 element depend solely upon the atomic number 

 which is the charge on its nucleus expressed 

 in terms of the unit charge "e." A neutral 

 atom of an element of atomic niunber N has a 

 nucleus consisting of K-fN protons and K 

 electrons and around this nucleus are set N 

 electrons. The weight of an electron on the 

 scale we are using is 0.0005 so that it may be 

 neglected. The weight of the atom will there- 

 fore be K + N so that if no restrictions are 



