26 
CHEMISTRY: LEWIS, EASTMAN AND RODEBUSH 
In the calculations of Lewis to which we have referred the value of C v was 
found to be 6.4 for sodium and 6.5 for potassium, but the difference between 
these values and the theoretical values might at that time have been ascribed 
to experimental error. More recently Dewar 4 has determined the mean 
atomic heats of a large number of elements between the boiling points of 
hydrogen and nitrogen. He found that at these low temperatures, where for 
most elements the atomic heat is only a small fraction of 3R, that of cesium is 
6.3. This is the value of C p , but it seems likely that at such a low tempera- 
ture the difference between C P and C v is very small even for an element like 
cesium. These facts led us to suspect that an abnormally high value of the 
atomic heat capacity might be a characteristic property of highly electro- 
positive metals. We were therefore interested to find that the experiments of 
Nernst and Schwers 5 on the specific heat of magnesium showed for this metal 
a change of heat capacity with the temperature which at low temperatures 
corresponds with that obtained for other monatomic elements, 6 but which at 
higher temperatures appears abnormally large. 
In order to test the supposition that all electro-positive metals would exhibit 
this kind of abnormality we have determined over a wide range of temperature 
the specific heats of two metals of the first group, sodium and potassium, and 
of two metals of the second group, magnesium and calcium. The detailed 
measurements and the experimental method will be described shortly in the 
Journal of the American Chemical Society. Here we shall give only the general 
results of the investigation . 
Before discussing the heat capacities found for the four metals it will be 
well to review briefly previous work on the heat capacities of other typical 
metals. Of these aluminum, copper, zinc, silver, mercury (solid), thallium 
and lead have been fully investigated over a wide range of temperature. For 
all of these metals C v has been found to be the same function of T/d, where 
T is the absolute temperature and 6 is the temperature at which C v is equal to 
3R/2. In other words, if we follow the procedure of Lewis and Gibson and 
plot C v against log (T/d) the individual values of C v for all these metals fall 
upon the same curve, which they speak of as the regular curve, or the curve of 
Class I. This may be expressed in still another way by the statement that if 
the values of C v as ordinates are plotted against the values of log T as abscissae 
the curves for the several elements may be made to coincide by horizontal 
displacement. 
The form of the regular curve is shown by the continuous curve of figure 
1 , where C v is plotted against the common logarithm of the temperature, and 
the individual points are those found for magnesiun. The five lowest points 
are taken from the measurements of Nernst and Schwers 7 and the remainder 
represent our own measurements. It is evident that at low temperatures the 
specific heat curve for magnesium coincides with the regular curve, but that 
at higher temperatures the specific heat rises more rapidly than in the case of 
any of the other metals to which we have referred. 
