RECENT ADVANCES IN SCIENCE 191 



indicate that the density of the fraction which remains in the 

 diffusion tubes is increasing at the rate which is predicted by 

 the Raleigh theory of diffusion, if the two atomic weights of 

 chlorine are 35 and 37. 



Triatomic Hydrogen. — Previous reference has been made 

 to the preparation of triatomic hydrogen (Science Progress, 

 1919, 53, 34). Wendt and Landauer {J.A.C.S., 1920, 42, 930- 

 46) have published a further paper, in which several methods 

 of preparation are given. The production of triatomic hydro- 

 gen is possible by three methods, the actions of a-rays, the 

 alternating discharge at low pressure, and passage through an 

 ozoniser at atmospheric pressure. Schumann light is without 

 action. The activated hydrogen does not resemble the atomic 

 hydrogen of Langmuir {J.A.C.S., 191 5, 37, 417) in its general, 

 physical, and chemical properties. 



Its activity is not due to ions or atoms, and the contraction 

 in volume under the action of an electric discharge indicates 

 a triatomic form. It is not yet clear that the atom of atomic 

 weight 3, observed by J. J. Thompson {Proc. Roy. Soc, 191 3, 

 A, 89, i) and by Rutherford {Proc. Roy. Soc, 1920, A, 97, 374), 

 is the same substance, for the form prepared by Wendt appears 

 to be more reactive. 



The chief theoretical interest in this work lies in the 

 fact that no hydrogen molecule larger than H3 is pos- 

 sible on the older conception of valency. One suggestion 

 put forward by the authors is that of a triatomic ring of 

 monovalent atoms with the alternation of the atom and the 

 electron in a six-membered ring. Bohr, however, regards the 

 structure of this molecule as a system of three electrons rotat- 

 ing at equal angular distances in a circular orbit, the three 

 nuclei being placed respectively at the centre of the orbit 

 and at two points on its axis equidistant from the centre. 



The Specific Heats of Gases by the Explosion Method. — ^The 

 radiation from the explosion of gases reaches very high values 

 in the case of coal-gas and air and of hydrogen and air (David, 

 Phil. Mag., 1920,39,66-95, 551-3); 26-1 and i6'i per cent, 

 of the total chemical energy respectively being emitted as 

 radiation in the two cases. It is concluded that part of the 

 radiation is due to chemical action, and part due to a purely 

 thermal emission from the hot gases during the period of 

 heating and cooling. It is suggested that the specific heat 

 of a gas is the greater, the greater the mean free path of the 

 molecules, and that the specific heats of the gases at 2,000° C. 

 and over depend to a large extent on the density and the 

 volume. If this suggestion be correct, the values of the specific 

 heats of gases determined by Pier, Bjerrum, and others would 

 require serious modification. 



