472 Dr. Carl Barus on the Absorption of 



The continuous, very slow contraction current toward the 

 top of the tube was a marked feature. At 12 h 20 m pressure 

 increment merely produced marked acceleration in the 

 velocity of this current, removal of pressure leaving the 

 mercury in place. On reducing pressure below a certain 

 value, however, the mercury retreated quite out of the cavities, 

 as above, to refill them again in the lapse of time. Figure 3 

 (p. 464) gives a view of a partially emptied cavity of this kind. 

 In figure 4 a globular empty cavity formed during this process 

 is shown, which was made to vanish symmetrically to a point 

 over an advancing mercury meniscus. 



The absorption of water in tube 5 virtually subsided after 

 about two hours of boiling. In the fine-bore tube, § 5, like 

 subsidence set in after but one hour of reaction. Differences 

 of chemical composition are not excluded. 



On cooling, bubbles appeared in homogeneous parts of the 

 thread, throughout its length, spontaneously, and expanded 

 often to over half the diameter of the core of water-glass. 

 Moreover, ovoid cavities filled with mercury when hot are 

 found partially empty after cooling, as shown in fig. 5, 

 due to the contraction outward of the walls. The final 

 breakage of this tube was confined to the core, which split 

 across axially, allowing the mercury to enter as a thin sheet 

 without breaking the envelope of igneous glass. 



The relations of ft and (v/Y) for this tube may be stated 

 from the following interpolation : — 



Thread-length (cm.) at 210° . . 15-8 15-4 14-8 14-2 13-6 12-8 

 0X10° 140 J 60 200 250 340 440 



Compressibility increases at an accelerated rate with con- 

 traction. Leaving out the final very large values, the mean 

 rate is 93xlO _6 per centim. of contraction. As the cold 

 length of thread is about 14 centim., this is a change of J3 of 

 about 13xl0~ 6 for each per cent, of volume-contraction 

 referred to the cold volume. The result is thus again of the 

 same order as the data from tubes 1, 2, and 5. 



J. The tube No. 6 was observed specially for volume-con- 

 traction, its large bore not being adapted for /3-measurements. 

 The tube when cooled down after corrosion showed a fine, 

 empty, axial canal, about '0063 centim. in diameter, left in the 

 core of aqueous silicate and forming a channel of communica- 

 tion between fourteen large axial bubbles, distributed with fair 

 uniformity along the length of the core. The constants of 

 the tube were (vapour-bath at 210°) 



