514 Sir W. Ramsay and Dr. B. D. Steele on the 



The methods of preparation of the three samples of ether 

 have ah-eady beeu mentioned on p. 496. Sample A was 

 distilled from phosphoric anhydride ; sample B was left in 

 contact with phosphoric anhydride for some days, and had 

 been shaken with metallic mercury before distillation; while 

 sample C had been recrystallized, and the crystals alone used. 



Ethe 



r. 



Temp. 



Pressure. 



'Pv/mT. 



A. 



B. 



C. 



Difference. 



129-60° 



99-71° 



500 

 400 

 300 

 250 



500 

 400 

 300 



250 



833-70 

 835-94 

 838-20 

 839-36 



830-03 - 

 833-03 

 836-00 

 837-50 



830-25 

 833-24 

 836-19 



837-68 



833-78 

 836-03 



838-28 

 839-43 



0-08 = 1 part in 10000 

 0-09=1 part in 9200 

 0-08 = 1 part in 10000 

 007 = 1 part in 12000 



0-22 = 1 part in 3800 

 0-21 = 1 part in 4000 

 0-19=1 part in 4300 

 0-18=1 part in 4600 i 



i 



These results show that the ether of all the samples may 

 be regarded as pure. The maximum difference is about 1 

 part in 4000. 



Sample A of toluene was prepared from paratoluidine : 

 sample B from bromobenzene and methyl iodide ; sample B 

 was subsequently recrystallized, as described on p. 495. 



Toluene. 



Temp. 



Pressure. 



Vv'mT. 



A. 



B. 



Difference. 



129-60° 



600 

 480 

 360 

 270 



658-61 

 662-62 

 666-43 

 66912 



658-93 

 662-94 

 666-':8 

 669-36 



0-32=1 part in 2000 

 0-32 = 1 part in 2000 

 0-35 = 1 part in 1900 

 0-24=1 part in 2800 



The greatest difference is observable wdth toluene ; it is 

 probable that the recrystallized sample w^as purer than the 

 other. But after all, the difference is by no means large 

 enough to affect the conclusions to be drawn. 



The next step is to deduce from the measurements given 

 in Table II. the values of pv/mT at zero pressure, when, if 

 Daniel Berthelot's contention is correct, equal volumes may 

 be expected to contain equal numbers of molecules ; a com- 

 parison of the true molecular weights should then be possible. 



