40 THE VOYAGE OF H.M.S. CHALLENGER. 



Steel Gauge. Challenger Gauge. 



Millimetres. Cwts. per sq. in. Ratio. 







5 0-0 



9 1-2 013 



15 87 0-58 



20 139 ()-69 , 



30 23 6 078 



40 35-0 0-87 



50 47-0 0-94 



60 58-7 098 



70 71-7 1-02 



The comparison was repeated several times with almost exactly the same results. 



It is quite clear that the Challenger gauge does not follow Hooke's law. It lags behind the steel 

 gauge at first (does not give any indication, in fact, till the pressure is nearly 50 atmospheres), then 

 gradually gains on it ; and, at pressures greater than 3£ tons, appears to leave it rapidly behind. The 

 instrument is, however, graduated up to 4 tons only. My very first experiments with this Challenger 

 instrument, in which I used a simple form of manometer, showed that it was not trustworthy, and 

 led me to make various trials for the purpose of getting a proper mode of measuring high pressures. 



[Inserted, July 8, 1881. — After recently examining a number of gauges of the Bourdon pattern, 

 some constructed to read to 600 atmospheres, I again tried the old Challenger gauge. The result was 

 very remarkable. Four successive trials agreed very closely with each other in giving 



Steel Gauge, Challenger Gauge, 



Millimetres. Cwts. per sq. in. Ratio. 







10 10-6 1-06 



20 20-3 1-02 



30 31-0 103 



40 42-8 1-07 



50 54-1 1-08 



60 657 1-09 



70 78-7 1-12 



A comparison of these, with the numbers of the former trials, shows that all the readings are 

 increased by somewhere about 7 cwts. This seems to show a definite slip of one of the bearings, or 

 possibly a new arrangement of the teeth gearing with one another in the two toothed arcs. But 

 whatever be the cause, the untrustworthiness of the gauge is obvious.] 



Columns 10 and 11 give the pure pressure effect on each of the thermometers, as calculated from 

 the measured dimensions of each instrument and of its principal aneurism by the help of the formula? in 

 Appendices A and B. The numbers given for the maximum side of each instrument are all slightly 

 too small, as I have not allowed for the effects of the (comparatively trifling) aneurisms at the bends 

 of the tube. Those given for the minimum side are the only ones of real importance ; and, in calcu- 

 lating these, all accessible details have been carefully attended to.' One or two of the instruments 

 were entirely smashed, so that no trace of the main aneurism was left. In such a case the correction 

 has a + inserted after it. In the other broken instruments the aneurism was still measurable, and 

 the correction has been adequately determined. 



The remaining columns of the table give the scale errors of the thermometers at 50° F. These were 

 determined casually in the course of the work, by comparing with a Kew Standard the thermometers 

 for trial next day, which were (for this purpose) kept for some hours in a steady stream of water. 



