KEPOET ON THE PRESSURE ERRORS OF THE THERMOMETERS. 



15 



the danger from the bursting of a hydrostatic press has been usually very much exaggerated. 

 The bursting of the cylinder itself would probably be unattended with danger ; but some 

 of the nuts and connecting pieces had occasionally been projected with great violence. 



A slight numerical calculation shows that a cubic foot of water at a pressure of one 

 ton weight to the square inch is capable of doing only about 1210 foot Ibs. of work in 

 expanding, the reason being that although the pressure is intense, the amount of com- 

 pression it produces is exceedingly small. But a cubic foot of air at a pressure of a ton 

 weight to the square inch is capable of doing nearly 1300 times as much work in expand- 

 ing. Hence the danger of having large quantities of air in the press before the com- 

 pression is begun. 



Another defect of the apparatus was the comparatively small interior bore, which did 

 not admit of the proper carrying out of my scheme for measuring pressures the Bourdon 

 gauge having shown itself quite untrustworthy. Besides, two thermometers, at most, 

 could be exposed to pressure simultaneously, even when no gauge was inserted along 

 with them. 



The apparatus which Sir Wyville Thomson finally obtained from the Woolwich gun 

 factories, through the intervention of the Admiralty, was in fact a Fraser gun with a few 

 adaptations made to suit it to the purposes of the investigation. The gun, which is 

 shown in Plate I., on a scale of one-eighth the full size, was made of a cylinder of mild 

 steel, round which were shrunk two successive wrought-iron coils. The effective interior 

 is 4^ inches in bore, and nearly 4 feet long. 



This cylinder was guaranteed to be safe under pressures up to 18 or 20 tons weight 

 per square inch, and we have for various purposes already worked up to pressures of 1 1 and 

 12 tons. [The official memorandum concerning this apparatus is given in Appendix D.] 



The rest of the apparatus, to fit it for our immediate purpose, consisted of a tightly- 



fitting steel plug which was forced into the upper end of the cylinder after the ther- 

 mometers and other apparatus had been inserted, and the whole had been filled with 



