346 Mr. Carl Barus on the 



a long range comparison between No. 1 and No. 4 is given in 

 Table IY. Five minutes were allowed per observation. In 

 the second part of Table IV. both gauges are standardized 

 by a comparison with a small, very sensitive Bourdon Gauge, 

 found identical with the large Bourdon Gauge, used above. 

 Fi and F 4 are the factors found. In other series similar to 

 Table III., the agreement of observations occupying like 

 positions in like series is always very close, the mean difference 

 being 1*2 atmospheres. Nevertheless the data of Table III. still 

 show decided hysteresis. See fig. 6, No. 1 (PI. X.) . The maxi- 

 mum distance apart of the "on" and "off" series is about 

 40 atmospheres. Indeed I was even able to grade the cycles. 



19. It is expedient to examine the chief sources of error 

 involved. The cycles cannot be attributed to an imperfect 

 Bourdon mechanism ; for in adjusting the pressure at any 

 given value just before observing, the actual pressures must as 

 frequently be incremented as decremented, both in the " on " 

 and the a off" series. Thus the aggregate error must be nil. 

 Liquid adhering to the walls of the capillary tube s s, would 

 show its largest discrepancy at zero atmospheres, following 

 high pressure. Similarly, slight changes of temperature of the 

 water in mm might displace the fiducial zero, but they would 

 produce no persistent cyclic effects. 



I think the clue for the interpretation of the apparent 

 hysteresis is given by the form of the cycles, which is that of 

 an archer's bow with the cord uppermost. Bearing in mind 

 that the liquid in tt is heated by compression during the 

 " on " series, and that even though gradually parting with 

 its heat increment, it is, nevertheless, continually hotter than 

 the surrounding medium ; bearing in mind, moreover, that in 

 the " off' " series the water in tt is cooled by expansion, but 

 that the water is hotter than the surrounding medium only 

 during the high-pressure stages, and colder than this medium 

 during the low-pressure stages (heat having been continually 

 dissipated) — an explanation of the bow-shaped cycles is sug- 

 gested. Tait gauges considered as thermometers are relatively 

 very sensitive instruments ; even my best gauges show dis- 

 placements of 1 or 2 centim. per degree. Kegarding the small 

 cycle (17 atmospheres) of Table IV., it will be remembered 

 that the error here occurs differentially. 



A bow-shaped cycle may also be conceived to result as a 

 viscous phenomenon, if strain (set) were gradually impressed 

 on the gauges during the " on " series, which strain asserts 

 itself during the "off" series. (Cf. § 34.) I must reserve 

 my opinion until further experiments have been made, in 

 which Amagat's " manometre " (I, c.) is employed. 



