Reliable information about the γ-rays emitted by radium C' is particularly valuable since the main features of the nuclear level system are shown by the groups of long range α-particles and are hence accessible to direct investigation. These long range groups of α-particles have been measured recently by Rutherford, Lewis and Bowden by a greatly improved method which has not only brought to light several new groups, but in addition has given considerably greater accuracy in the determination of the energies of the groups than had been possible in the past. The energy of these groups in excess of that of the normal group is a measure of the excitation energy of the nucleus and the older measurements had indicated, as was to be expected, a close correspondence between these energies and the quantum energies of the γ-rays. The recent more accurate measurements of Rutherford, Lewis and Bowden provided the opportunity of a more rigorous test of this connection and showed the possibility, by the combination of the information from these two sources, of a direct experimental determination of the level system. However, the accuracy of the published data on the Ra C γ-rays deduced from the natural β-ray spectrum was subject to some doubts for the following reasons. The measurements dated from 1924 when the absolute energies of certain strong groups in the β-ray spectrum were measured and the energies of the remaining lines determined by relative measurements. The strong groups in question lay between 0.4 X 10
5
and 3.0 X 10
5
volts and the procedure of step-wise comparision up to energies of over 2 X 10
6
volts may have led to cumulative errors. More serious was that recent measurements on the Th (B + C) β-ray spectrum had thrown doubt on the correctness of the absolute values. Lastly, experiment gives values for H
ρ
, that is the momenta of the electrons in the groups, and the calculation of the energies involves
e
/
m
. The older data had been based on
e
/
m
= 1.769 X 10
7
, and while the change to the value 1.760 X 10
7
alters the energies proportionally far less, there was involved here an avoidable error which had to be removed.
The thermal conductivity of some gases at 0° C
