1855] WRITINGS OF JOSEPH HENRY. 363 



every vessel used to contain it having been tried at each of 

 the several fillings of the tube, which were made on the first 

 days of the experiment, when a leak required its discharge 

 for the purpose of tightening the joints. A portion of the 

 original liquid which had been set aside was also tried at 

 the end of the experiment, and at different temperatures. 



The readings of the hydrometer were made with as much 

 accuracy as possible under the circumstances, some of them 

 being taken late at night and exposed in the open tower to 

 a violent wind. No pains were spared to test the liquid 

 under every variety of circumstances. At first the windows 

 of the tower were open, but for the last two or three months 

 they were closed. Fifty-four readings were made; nineteen 

 of which were from the original liquid, and the remainder 

 on that drawn from the different cocks. The result may be 

 stated, as follows: 



On plotting the readings of indication and temperature 

 they all follow nearly in the same line, the deviations of 

 those taken from the original fluid being quite as great as 

 those taken from either the bottom or top, even after the 

 lapse of months. Or in other words, within the limits of 

 error (the extreme being but a portion of a degree of the 

 hydrometer), there is not the slightest indication of any dif- 

 ference of density between the original liquor and that from 

 the top or bottom of the column after the lapse of hours, 

 days, weeks, or months. The fluid at the bottom of the tube, 

 it must be remembered, was for five months exposed to the 

 pressure of a column of fluid at least one hundred feet high. 

 This pressure however is much within that at which inferior 

 champagne bottles are burst, and if pressure alone could 

 produce such an effect, wine of that kind should have long 

 ere this given instances of it. 



As the fact has been taken for granted, and chemists of 

 repute have made use of it, there seems good ground for 

 thus formally refuting an error which, at first sight, would 

 not appear worthy of being dignified by so much notice. 



