C. Barus — Experiment with Liquid Carbon Dioxide. 3 



diameter, and placed about 65^"^ from the tube d^ and about 

 39()cm fj.Qjn the white wall or screen G. 



The whole experiment is, as stated, made in full daylight. 

 The displacement of the lens F^ when focusing the liquid or the 

 vaporous column respectively, is fully 5^°^ to 6°"^, and curiously 

 enough the extreme positions of the lens remain practically 

 unchanged with temperature. 



To supjDly the trough with warm or cold air, a bellows may be 

 used and the current passed through a coil of lead pipe con- 

 tained in a water bath kept at the boiling point. The length 

 of coil is easily regulated so that the available current of air 

 enters the trough between 30° and 40° C. The extreme 

 mobility of liquid CO2 makes these relatively high and un- 

 equal temperatures harmless, for in a serious case the liquid 

 merely simmers, boiling at the lower end of the tube, while 

 the bubbles condense before they reach the surface. When 

 no current of air passes, the air bath soon cools without special 

 appliances. 



The surprising part of the experiment, to me, is the obser- 

 vation that the marked displacement of the lens F^ necessary 

 to converge the linear foci of gas and of liquid, does not 

 vary appreciably (i. e., in an experiment conducted like the 

 above) with temperature ; one would naturally suppose that the 

 highly compressed gas and the liquid so near the point of con- 

 tinuously merging into the gas, would show similar refracting 

 properties ; one would, in other words, expect to find two 

 linear foci, one corresponding to the compressed gas and 

 the other to the liquid, at a distance apart which would grad- 

 ually vanish as the critical temperature is approached. 

 This means that the necessary displacement of the lens F 

 would grow continually less, becoming zero at the critical 

 temperature. Yet such is not at all the case, as the following 

 data show : 



Temperature : 



17° 



21° 



23-5° 



25-5° 



27° 



28° 



29° 



Column of gas left : 



6 



5 



4 



3 



2 



1 



Qcr 



Displacement of lens : 



5-5 



5-5 



5-5 



5-0 



50 



6-0 



5-5 



the error contained being the uncertainty of focusing. 



The linear focus due to the liquid column is of course real, 

 and in case of my tube (external diameter O'TS'^"^, internal 

 diameter unknown) lay about 3°°" in front of the tube."^ (i. e. 



* The theoretical treatment of this case is cumbersome, involving as it does an 

 extensive consideration of caustics in view of the wide aperture. The following 

 data, however, illustrate the final result for parallel rays. If tubes be taken 

 about 0-6, 0-8, 1-2, 1-4, 2-6<='" diameter, and partially filled with water, the liquid 

 focal line will as a rule lie within 1'°^ in front of the nearer wall of the tube. 

 The corresponding focus for the gaseous column will be equally sharp but virtual 

 and lie about 3, 4, 6, 9, 30*=°^ respectively behind the liquid focal line, showing a 



