It is obvious from the method now described, that the altitude thus found must always be too small, and that a correction depending upon the quantity of water removed by the measure M should be made. This correction must be very small, if the diameter of the glass vessel is great. This instrument is obviously liable to a source of error, from the interposition of the glass vessel between the telescope and the fluid column, as the least inequality or difference of parallelism in the parts of the glass oppo site to S and II, would produce a sensible error in the result. This, however, may be easily obviated by cutting a piece out of the glass vessel, a.nd cementing upon it a plate of parallel glass. We would also recommend, in making very accurate observations, that the sides of the glass tube next the telescope should be either ground flat about S, or have any inequality of refraction removed, by cementing upon it a plate of parallel glass.
The method of measuring the cohesion of fluids, which we have described for the first time under CAPILLARY AT TRACTION, will, we have no doubt be found the most cor rect. The apparatus will be greatly improved by using two solids of the same kind and form instead of one. By this means the termination of the elevated circle of fluid will be more easily ascertained.
With the instrument shewn in fig. 3, M. Gay Lussac has made several experiments on the ascent of water and alcohol in tubes of glass, of which the following are the results. In these experiments, the tuhes were well wetted with the fluid. • Our readers will no doubt be much surprised at the great discrepancy among the results in the preceding table, and particularly between those obtained by Dr Brewster and Gay Lussac, which were made with instruments founded on the same principle. M. La Place has ascribed these differences to the greater or less degrees of humi dity on the sides of the tubes ; and he informs us, that, in Gay Lussac's experiments the tubes were very much wet ted. Now, it appears to us, that though by this method the water will always stand nearly at the same height in the same tube, yet it does not afford us an accurate mea sure of the height of ascent. Let us suppose that a tube of an inch in diameter is so perfectly cleared of all grease and extraneous matter, that the attractive force of the glass is allowed to exert itself without any diminution, and that the water stands at the height of 3.3 inches. Let us now suppose that by some means or other a film of water of the thickness of of an inch is introduced at the upper end of the tube, and equably diffused over its in terior surface, it is obvious that the water will rise above its former level, and consequently to a height greater than that which is due to the force exerted by the tube and the mutual action of the fluid. We conceive, therefore, that
M. Gay Lussac's measures err in excess ; and that the error increases inversely as the diameter of the tube.
The discrepancies in the table appear to us also to be partly owing to the different kinds of glass employed. The flint glass of which tubes are composed, varies very much in its density ; and there is every reason to believe, both from analogy and from some direct experiments we have made, that the substance of the tube has a very per ceptible influence on the height to which the fluid ascends.
The following were the results which M. Gay I.ussac obtained for alcohol.
The results of experiments 1 aid 2, when reduced to English inches, give .01798 and .01840 for the value of the constant quantity. The constant quantity for alcohol found by Dr Brewster is almost the same as this, namely .0173. Benjamin Martin makes the constant quantity 18, and Muschenbroek 10.
M. Gay Lussac obtained the following result for oil of turpentine.
M. Gay Lussac measured With great care the rise of water between two plates of glass ground perfectly fiat, and placed exactly parallel to each other. In order to do this with accuracy, he kept the plates separate by four very fine iron wires cut consecutively from the same piece, so as to have their diameters as equal as possible ; and in order to find the thickness of the wire, he placed a great number of them together, and measured the sum of their diameters. The following was the result of his obser vations.
It appears from the experiments mentioned by Newton m his Optics, (p. 366. edit. 3d, 1721) that water rose one inch between two plates of glass, whose distance was Th of an inch, and that water rose to the same height in a ca pillary tube, the semi-diameter of whose bore was equal to the distance of the plates, which gives .010 as the con stant quantity for the glass plates, and .020 as the constant quantity for capillary tubes.
The following experiments were made by M. Monge, on the rise of water between two plates of glass. The plates of glass were first cleaned with caustic alkali, and carefully washed, and, when separated to different dis tances, by the interposition of silver wires of different thicknesses, they were plunged in the water of the Seine, which had been previously filtered. The diameters of the silver wires, from which the distance of the plates was inferred, were obtained by rolling a wire round a tube of glass, and finding the of thicknesses which occu pied air: exact number of lines. By dividing the number of lines by the number of revolutions, he obtained the exact diameter of the wire, and consequently the distance of the plates. The following are the results which he obtained.