'478 



SGIENCK 



[N. S. Vol. Vll. No. 171. 



many series of observations were made at 

 varying pressures all within the last mill- 

 ionth. The gauge could be relied upon to 

 measure these small pressures with very 

 great accuracy ; but it was diiiicult to main- 

 tain them long at an exactly constant value 

 on account of the continual, though slight, 

 evolution of gas from the glass of the ap- 

 paratus. 



As I desired only comparative results, no 

 correction was made for the probable slight 

 inequalities in the callibration of the ther- 

 mometer ; nor for heat conducted to or from 

 the bulb by the stem ; nor for the change of 

 zero point due to changing external pres- 

 sure. The mercury fell exactly to zero at 

 atmospheric pressure, and about one-fiftieth 

 of a degree lower at no pressure. The pres- 

 sure error due to differences of capillary de- 

 pression in the two barometers was ascer- 

 tained at high exhaustions, and foxind 

 nearly constant. It was always corrected. 

 The different gases used were carefully pre- 

 pared and dried, and were introduced quite 

 free from any admixture with air. 



My observations have extended over a 

 long period, and are far too voluminous to 

 be recorded here in detail. But I have em- 

 bodied their most salient features in a series 

 of curves which render them readily appar- 

 ent to the eye. In these curves the abscissae 

 represent the pressure, and the ordinates 

 represent the rate of heat transmission 

 through the gas, from the thermometer bulb 

 to the ice-cold envelope. The rate of trans- 

 mission at any particular pressure is ex- 

 pressed by the reciprocal of the number of 

 seconds required for the temperature to fall 

 through a given number of degrees. For 

 convenience of scale, all the reciprocals are 

 multiplied by 500. 



Fig. 2 shows the curve for air. The 

 heavy line represents the rate of cooling 

 from 1-5 degrees to 10 degrees. It is in three 

 sections, A, B and C Section A embraces 

 the whole range of pressure from nothing to 



atmospheric ; section B embraces the range 

 of pressure from nothing to .01 of atmos- 

 pheric ; and section C embraces the range 

 of pressure from nothing to .0001 of atmos- 

 pheric, i. e., 100 M. (one hundred mill- 

 ionths). Atmospheric pressure is taken at 

 760 mm. Thus it will be seen that section ' 

 B is the last hundredth of A, magnified a 

 hundred times ; and section C is the last 

 hundredth of B, magnified a hundred times. 

 This magnification of the abscissae without 

 change of the ordinates, enables us to study 

 every part of the curve with ease. The 

 small circles represent the points in the 

 curve established by observation. These 

 points are shown exactly as found, without 

 any attempt to smooth out rough places in 

 in the curve. The same is true of the curves 

 of other gases. The heavy dotted line par- 

 allel with the base indicates that portion of 

 the total heat transmission due to the ether ; 

 while all above it represents that due to the 

 air. 



Starting at the left-hand end of section A, 

 representing the rate of heat transmission 

 at atmospheric pressure, we observe that 

 the curve drops regularly at a rate faster 

 than the diminution of pressure during 

 ninety-five per cent, of the whole range of 

 pressure from atmospheric to zero. Beyond 

 this point the rate of heat transmission re- 

 mains substantially constant, as shown by 

 section B and the latter part of A, down to 

 a pressure of about .0003 — a range of nearly 

 ninety-nine and a-half per cent, of that re- 

 maining. Here the curve suddenly begins 

 to drop again, and falls steadily, as shown 

 by section C and the latter part of B, until 

 it meets the ether line at the zero of pres- 

 sure. 



Under the curve A, I have drawn curves 

 with finer lines, representing the rate of 

 heat transmission at smaller differences of 

 temperature between the thermometer and 

 ice bath. As before stated, A represents 

 the cooling from 15 degrees to 10 degrees. 



