able errors of his manual operations. and authorize him to eliminate
from the essential clements of a compound those products of an ana-
lysis whose quantity cannot be reduced to any admissible proportion,
and may therefore be considered as extraneous.
The author, at the close of his paper, controverts the opinion of those who think that crystallization requires a previous state of solu- tion in the matter crystallized ; and contends, that as long as any quantity of fluid is present in a solution, no crystallization can pos- sibly take place. ,
After a short recapitulation of what has of late been done by' Mr. Cavendish, Dr. Priestley, Dr. Nooth, and others, respecting the impregnation of water with diiferent gases, our author observes, that the circumstance of the different degrees of temperature and pressure had not been as yet sufficiently attended to. Dr. Priestley, indeed, had long since remarked, that, in an exhausted receiver, Pyrmont water will actually boil at a common temperature, by the copious discharge of its air; and that hence it is very probable, that by means of a condensing engine, Water might be much more highly impreg‘ nated with the virtues of thc Pyrmont spring: but this conjecture remained as yet to be proved by experiments; and this is the task our author has undertaken in the present paper.
This paper consists of two sections; the first treating of the quan- tities of gases absorbed by water under the usual pressure of the atmosphere ; and the second, of the influence of pressure in promoting the absorption of gases. The apparatus contrived for these experi— ments may be described as a siphon, of which one side, or leg, is a glass vessel of comparatively a considerable diameter, and the other a long glass tube of about a quarter of an inch bore; the junction of these two parts at the bottom being a short pipe of India rubber, well secured by proper integuments of leather, thus forming a joint, which admits of the vessel being briskly agitated. This vessel has a stop-cock both at top and bottom, in order to insert and emit fluids and gases; and both the vessel and tube are accurately graduated. It may now be.understood, that a known quantity of water and of a certain gas being put in the vessel, and the tube being filled to a certain extent with mercury, the absorption of the gas will be accurately measured by the column of mercury in the tube. Those who are particularly interested in this inquiry will find in the paper various precautions and additional contrivances, all tending to insure the success and accuracy of the investigation.
The first experiments were made on the absorption of carbonic acid gas by water: and here a singular disagreement was observed in the first trials made under exactly the same circumstances. It
