Pressure in Temperature, F. Rise of Temperature Atmospheres ° for each additional Atmosphere 1 212 2 249·5 37·5 3 273·3 23·8 4 291·2 17·9 5 306·0 14·8 6 318·2 12·2 7 329·6 11·4 8 339·5 9·9 9 348·4 8·9 10 356·6 8·2 11 364·2 7·6 12 371·1 6·9 13 377·8 6·7 14 384·0 6·2 15 390·0 6·0 16 395·4 5·4 17 400·8 5·4 18 405·9 5·1 19 410·8 4·9 20 415·4 4·6 It may be seen from the above that, with the exception of one irregularity, there is a continual diminution of the additional temperature which is required to overcome an additional atmosphere of pressure, and if this goes on as the pressure and temperatures advance, we may ultimately reach a curious condition—a temperature at which additional pressure will demand no additional temperature to maintain the gaseous state; or, in other words, a temperature may be reached at which no amount of pressure can condense steam into water, or at which the gaseous and liquid states merge or become indifferent. But we must not push this mere numerical reasoning too far, seeing that it is quite possible to be continually approaching a given point, without ever reaching it, as when we go on continually halving the remaining distance. The figures in the above do not appear to follow according to such a law—nor, indeed, any other regularity. This probably arises from experimental error, as there are discrepancies in the results of different investigators. They all agree, however, in the broad fact of the gradation above stated. Dulong and Arago, who directed the experiments of the French Government Commission for investigating this subject, state the pressure at 20 atmospheres to be 418·4, at 21 = 422·9, at 22 = 427·3, at 23 = 431·4, and at 24 atmospheres, their highest _experimental_ limit, 435·5, thus reducing the rise of temperature between the 23d and 24th atmospheres to 4·1.