96 C. JBarus — Isothermals, etc., relative to Viscosity. 



a substance like marine glue, n is probably small as compared 

 with N. 



13. Measurement of excessively high pressure. — Let there 

 be given a tube of length I and radius p. Let tj = rj (l + bj>) be 

 the viscosity of the viscous liquid forced through it by the 

 pressure excess dp = 2j? (so that there is no pressure at one 

 end of the tube), and at the constant temperature d. The 

 length (X) of a cylinder of fluid issuing per unit of time (t) 

 will then be p~Ap/8ly} (l + bAp/2). Hence if negative pressures 

 be excluded, the function X/t is of a kind which continually 

 increases with dp, a state of the case which would not be true 

 if the expression of § 11 were applicable. 



In view of the observed property of X/t, it is worth inquir- 

 ing in how far the transpiration method is available for high 

 pressure measurement, when most other means fail. 



Take for example a tube T ^ inch in diameter and 1 inch 

 long. Then the mass m of the above marine glue which at 

 say 18° would exude per hour is 



mxi0 6 = 5970 7550 7810 7900 7950 grams. 



Ap — 1000 5000 10000 15000 20000 atm. 



where q = 10 9 (1 + -00285A^). 



Thus it appears that whereas a hole -^ inch in diameter may 

 be efficiently sealed by marine glue at 18° 0., pressure meas- 

 urement by aid of the exuding mass is impossible above 

 10,000 atm., whereas even between 5000 atm. and 10,000 atm. 

 the method is insensitive. To use a method like the present 

 for very high pressure measurement, a substance of smaller 

 pressure coefficient must therefore be sought, if such a one 

 with other necessary qualities, exists. It is with the object of 

 searching for such a body, as well as of finding the maximum 

 of hydrostatic pressure attainable in the laboratory that I had 

 a tinned screw and socket constructed,* and hope to be able to 

 report the results of my work at an early opportunity. To 

 my knowledge the only other gauge available under the cir- 

 cumstances is the one I based on the resistance of mercury. 

 It is sufficient, however, for making the comparisons in ques- 

 tion. 

 Phys. Lab. U. S. G-eolog. Survey, Washington, D. C. 



* Proceed. American Acad., xxv, pp. 94, 108, 1390. 



