z 



6. Compression tubes and appurtenances. — The substance 

 to be tested is enclosed in capillary tubes 

 of glass of very fine bore (*03 cm and less) 

 and about *6 cm thick, such as are used for 

 thermometers. The length of these tubes 

 a be, figure 1, was about 50 cm or 60 cm . To 

 insert them into the trough, the solid glass 

 of the tube was swelled or bulged at h 

 about 5 cm from one end, and then pushed 

 through an axially perforated steel flange 

 £■£) A A BB, in which they fitted snugly. The. 

 bulged end and the flange faced the in- 

 side of the trough, so as to be acted against 

 by the pressure. To seal the tube a b c, in 

 the steel flange A A BB fusible metal was 

 poured around it at ff, and the joint was 

 then further tightened by a thick coating 

 of marine glue, gg. A strong hollow nut 

 of iron surrounding both the lower end of 

 the tube and the stem of the flange, fas- 

 tened it to the trough by forcing it against 

 a leather washer. When completely ad- 

 justed, the capillary tube projected about 

 40 ctn or 50 cm clear above the trough. Pres- 

 sure was transmitted through the oil of 

 the pump to the mercury in the trough, 

 w T hich communicated directly with the 

 open lower end of the capillary tube. 



7. A thread of the (solid) substance to 

 be acted on was introduced into the 

 cleansed capillary, between two threads of 

 mercury. This is a delicate operation, re- 

 quiring much care ; but the details of 

 manipulation can not be given here. I 

 will say merely that it is not advisable to 

 seal the upper end, a, of the tube a b o by 

 fusing the glass, because the strains intro- 

 duced during cooling greatly diminish the 

 strength of the tube. The end may be 

 sealed by a thread of paraffine 5 cm long, 

 properly inserted and kept solid by a cold 

 water jacket, Gr Gr. Under these condi- 

 tions the viscosity of paraffine is too great 

 to admit of its being forced out of the 

 glass by any pressure compatible with the 

 strength of the tube. Faultless adhesion 

 of the paraffine to the glass is however es- 

 sential. 



Passing from top to bottom of the 

 charged capillary, a be, one therefore en- 

 counters the following succession of threads: 

 with^app^rST Pfrafflne (5-), mercury (5-), substance 

 measuring compressibil- (20 cm ), mercury (25 cm ) m communication 

 ity of fluids. Scale £. with the mercury of the trough. Measure- 



-A 



Fig. 1. — Capillary tube 



