C.downward the divergence of such a thermometer from one of hydrogen is very marked.
The measurement of very high temperatures is not open to the same theoretical objections as that of very low temperatures; but, from a practical point of view, it is as difficult to effect with an ordinary gas thermometer.

It becomes impossible to guarantee the reservoir remaining sufficiently impermeable, and all security disappears, notwithstanding the use of recipients very superior to those of former times, such as those lately devised by the physicists of the _Reichansalt_.

This difficulty is obviated by using other methods, such as the employment of thermo-electric couples, such as the very convenient couple of M.le Chatelier; but the graduation of these instruments can only be effected at the cost of a rather bold extrapolation.
M.D.Berthelot has pointed out and experimented with a very interesting process, founded on the measurement by the phenomena of interference of the refractive index of a column of air subjected to the temperature it is desired to measure.

It appears admissible that even at the highest temperatures the variation of the power of refraction is strictly proportional to that of the density, for this proportion is exactly verified so long as it is possible to check it precisely.

We can thus, by a method which offers the great advantage of being independent of the power and dimension of the envelopes employed--since the length of the column of air considered alone enters into the calculation--obtain results equivalent to those given by the ordinary air thermometer.
Another method, very old in principle, has also lately acquired great importance.