Mr Dalton found the same thing. But he considers this diminution as apparent only, and not real, and owing to the expansion of mercury not being equable. According to him, water, mercury, and all liquids, expand as the square of the temperature, reckoning from the freezing point of the respective liquid. Ac cording to this notion, the expansion of air, (and in deed of all permanently elastic fluids,) is in geome trical progression to equal increments of temperature. The followilig Table exhibits the rate of expansion of air from 32° to 212°, according to Mr Dalton: The reader will observe, that the expansion of air in the second column of the Table constitutes a geome trical progression, the ratio of which is 1.0179. The third column exhibits the corresponding degrees of a Fahrenheit's thermometer graduated, according to Mr Dalton's notion of the expansion of mercury, accord ing to the square 'of the temperature. This notion of Dalton must be allowed to be very ingenious. Unfortunately we are not in possession of any mode of ascertaining how far it is correct. It is only sup ported by analogical reasoning, and cannot well be otherwise in the present state of our knowledge of heat.
Atmospherical air was long considered as a sim ple elementary body. But it is now known to con sist of at least four distinct substances, namely, oxygen, ' azote, carbonic acid, and aqueous vapour. The first two substances must be considered as its essential con stituents, and constitute in fact almost the whole of it. The other two are variable in their proportion, and exist only in minute quantities, which it is diffi cult to appreciate.
The first knowledge of the composition of the at mosphere must have been after the period of the dis covery of oxygen gas by Dr Priestley in 1774. La voisier, indeed, in his posthumous works, appears to insinuate a knowledge of it in 1772. But this claim cannot be admitted, as he gives no hint of any such knowledge in his volume of essays published after that period, and as he was entirely unacquainted with oxygen gas when Priestley shcwed him the way to prepare it at Paris, about the end of 1771. It is very probable that Lavoisier became acquainted with the composition of atmospherical air not very long after that period ; though some years elapsed before he made it known to the public. Whether he pre ceded Scheele in his knowledge of this important fact, we do not exactly know. But there is no doubt whatever, that Scheele's investigations were carried on without any assistance from abroad, and that it was in consequence of the publication of his Trea tise on Airand Fire, that the chemical world became acquainted with the nature and composition of at= mospherical air. This important work was printed at Upsal in 1777, with an introduction by Bergmann, and translated into English by Dr Foster in 1780. The experiments of Priestley indeed would have war ranted the conclusions respecting the composition of atmospherical air drawn by Scheele ; but those of Dr Priestley were different and more complicated. In Scheele's first experiments, he estimated the bulk of oxygen gas in air at 30 per cent. But in the year
1779, lie published a set of experiments continued for a whole year, in order to ascertain whether the bulk of oxycn in air be constant, or varies with the season of the year. He found it in general remarkably con stant, and amounting to 27 per cent. The smallest bulk was 21, and the greatest observed was 30 per cent. Dr Priestley had made similar experiments, and had estimated the bulk of the oxygen at nth of the air, or 20 per cent. Mr Lavoisier's experiments, which were very numerous and varied, almost coin cided with those of Scheele. He considered air as composed of 27 parts by bulk of oxygen, and 73 of azote. Mr Cavendish's experiments were published in the Philosophical Transactions for 1783. He proved decisively, that the proportion of the :azote and oxygen in the atmosphere does not vary; and by a very careful analysis, concluded, that 100 parts of air in bulk are composed of l 00.00 This opinion was not at first acceded to by chemists, misled by the previous conclusions of Scheele and Lavoisier ; and it was not till towards the commence ment of the 19th century, that the true proportion of these constituents was generally known. The experiments of Berthollet in Egypt and in Paris, seem to have led the way to it. These were almost immediately confirmed by those of Davy, Beddoes, and many other chemists. At present it is univer sally admitted, that atmospheric air never varies in its composition ; that it is the same in all places, and in all seasons; and that it consists in bulk of 79 azote 21 oxygen 100 proportions almost exactly the same with those ori ginally settled by Mr Cavendish.
Oxygen gas is undoubtedly the most important of the constituents of the atmosphere, and indeed one of the most remarkable substances in nature, and highly worthy of the investigation of the chemist. Dr Priestley, its original discoverer, gave it the name of dephIngiqicaled air, Scheele called it empyreal air, Lavoisier called it at first highly respirable air, then vital air, and at last oxygen gas, because he consi dered it as the acidifying principle. It possesses the mechanical properties of common air; combus tibles burn in it with great brilliancy ; and animals can breath it much longer than the same quantity of common air. If the specific gravity of common air. be reckoned 1.000, that of oxygen gas, according to the experiments of Kirwan and Lavoisier, is 1.103 ; according to Davy, 1.127 ; according to Fourcroy, Vauquelin, and Seguin, 1.087 ; and ac cording to Allen and Pepys, 1.090. These results do not differ much from each other, except that of Mr Davy. His oxygen was obtained from the black oxide of manganese, and might perhaps contain a little carbonic acid gas. If we exclude his, the ave rage of the other three is 1.093. This may liNcon sidered as near the truth as can well be attained. Rating its specific gravity at 1.093, 100 cubic inches of it, at the temperature of 60° when the barometer stands at 30 inches, will weigh 33+ grains troy.