TRANSACTIONS OF THE SECTIONS. 359 



repeated these experiments, and as they appear hitherto not to 

 have satisfied philosophers, I shall here mention, that in all my 

 experiments upon the subject, I have invariably found that the 

 recession of the w^ater in the capillary tube is about 1 'b millionth 

 greater in bottles of lead or tin than in those of glass. Sup- 

 posing the compressibihty of the solid bodies to be so small 

 that it cannot be observed in those experiments, yet the heat 

 developed by the compression, feeble as it is, produces a small 

 augmentation of the recession of the water in the capillary tube. 

 If the dilatation of a rod of glass by 1 Th. is 0-0009, its cubical 

 dilatation is 0*0027, and the dilatation by an increase of 0-00025 

 is 0-000000675, or nearly 7 ten miUionths. The dilatation of lead 

 is about 3 times greater, and the bottle containing it must get an 

 increase of 0-00000225, which exceeds the former by more than 

 1 -5 millionth. The dilatation of tin should give only one mil- 

 lionth more than glass, but it seems to give a little more, yet 

 the quantity is not great. After all this, I think that the true 

 compressibility of water is about 46-1 miUionths, and that the 

 apparent compressibility depends upon the effect of the heat 

 developed by the compression, by which the liquid and the bot- 

 tle are dilated. 



" My continued experiments have confirmed my earlier re- 

 sult, that the differences of volume in the compressed water are 

 proportionate to the compressing power. I do not know if the 

 method I have made use of to try the effects of high compression 

 has been published in England. These experiments cannot be 

 made in a cylinder of glass ; one of metal is required. As, in 

 this case, the opacity prevents direct observations being made, 

 an index, nearly like that in Six's register-thermometer, is 

 placed in the capillary tube of the bottle. This tube is dilated 

 a little at the top, so as to form a minute funnel. Some drops 

 of mercury are poured into it, which being pressed, pushes the 

 index forward ; thus the recession may be seen when the bottle 

 is taken out of the large cylinder. The compression of the air 

 is measured in another way : a bent tube, of the form shown in 

 fig. 3, is fixed in a glass vessel F G H I containing mercury, and 

 exposed to the pressure together with the bottle. The pressure 

 of the piston upon the water in the cylinder is communicated to 

 the mercury, and pushes it into the wide part of the tube, as far 

 as the resistance of the air will permit. The weight of the mer- 

 cury driven into the wide part A B C D, together with that 

 which has filled D E, and which may be computed, compared 

 with the weight of mercury which the whole tube can admit, 

 gives the volume of the air compressed. By this kind of ex- 

 periment I have found that the decrease of volume produced by 



