L. A. Hill—Notes on Argon and Helium. 367 
Many of the metals are volatile if at all at such high tem- 
peratures that they have not been accurately measured, but 
generally the difference between the melting and boiling points 
is very great for metals. Apparently the rule is that the 
metalloids have a small while the metals have a large difference 
between their melting and boiling points, that is, that those 
elements (metals) whose chemical affinity is overcome by a 
small degree of heat energy, have large differences between 
their boiling and melting points, but that those (metalloids) 
whose affinity is only masked by a large degree of heat energy, 
have small differences between these points. 
Thus compare Hg and Cd and the other metals above with 
iON, O, Cl, Br and I. 
Now if mercury at 800° is devoid of affinity, then the dif- 
ference between its melting point and boiling point (395°) is 
about the same as between its boiling point and its upper limit 
of chemical combination (443°). But in the case of the 
metalloids, the difference between the melting and boiling 
points is small 66° to 70° for the halogens, from 17° to 20° for 
nitrogen, methane and carbonic oxide, and 13:4° for nitric 
oxide ; while in these cases the difference is very large between 
the boiling points and the upper limits of combination, (by 
which I mean the temperatures at which chemical affinity is 
no longer manifest,) and the disproportion between these dif- 
ferences tends to become greater as the element approximates 
more and more closely to the characters of the ideal metalloid. 
Thus iodine begins to dissociate at 700° and at 1500° the 
process is complete, so that the upper limit of combination 
must be above 1500°, and instead of a ratio of differences like 
that in mercury of 395° to 443° we should have instead the 
ratio of 70° to 1416°+. Inequality of differences character- 
azes the metalloids, equality the metals. Now if argon is a 
metalloid (and is it not one if an element?), ought it not also 
to have this inequality of differences, and to have its upper 
limit of combination at a very high rather than low tempera- 
ture? But if this limit be very high, argon can not at ordinary 
temperatures be like mercury at 800°, that is too hot to com- 
bine with other elements. 
Again, if, as is the case, the melting and boiling points of 
argon are very close to each other, ought we not to expect that 
its melting point and upper limit should be separated by an 
interval correspondingly great ? 
In Br and Cl the characteristics become less metallic than 
_ with J, and- approach more nearly to those of the ideal metal- 
loid; and we find that while the difference between the melt- 
ing and boiling points is but little changed, the differences 
between the boiling points and the upper limits are very greatly 
