TRANSACTIONS OF SECTION B. 



689 



silver chloride was collected on a Gooch's filter, and dried in an air-bath at a tem- 

 perature of about 200° C. The results obtained gave a rather low percentage of 

 chlorine ; calculated v?ith Clarke's atomic weights it is 60-598. If the atomic 

 weight of boron is determined from the chlorine, it is rather higher than that given 

 Jrom the weight of salt. 



Table IH. 



Fspt. No. 



22 

 23 



24 



25 



26 



Aggregate 



Weight of 

 fused borax 



5-3118075 

 4-7805583 

 4-9907395 

 4-7231225 

 3-3137921 

 23-1200199 



Weight of 



silver chloride 



AgCl 



7-525869 

 6-7794186 

 7-0804317 

 6-6960242 

 4-6931271 

 32-7748706 



Chlorine 



per cent. 



in sodjuni 



chluride 



60-493 

 60-515 

 60-516 

 60-514 

 60-479 

 60-505 



Atomic 



weight of 



boron 



from 



chlorine 



11-071 

 11-024 

 11-003 

 11-039 

 11-091 

 11084 



Mean atomic weight of boron 10966. 

 Probable error ± -005. 



Mean atomic weight of 

 boron 11 052. 

 Probable error ± -010. 



Since no special precautions were taken to obtain pure silver for the chlorine 

 determinations, but ordinary silver nitrate was employed, it is probable that the 

 distillation results yield the more correct atomic weight of boron, which is there- 

 fore 10-966. 



On the Preparation of pure Olucina, and the Atomic Weight of Olucinum. 



By Dr. J. Gibson. 



I 



3. On the Assumed Potential Difference heticeen a Metal in the Molten and 

 the Solid State. By Professor "W. Ostv?ald. 



If two pieces of the same metal are immersed in a suitable electrolyte con- 

 taining this metal as an ion, and the whole is maintained at the melting-point of 

 the metal, it is possible to keep one of the pieces in the solid, the other in the 

 liquid state. It is generally admitted that such a system would show an electro- 

 motive force, the amount of which can be calculated fi-om the latent heat of the 

 fused metal. 



But if there were indeed a potential difiereuce between both the electrodes, it 

 would be possible to gain electric energy, and therefore also mechanical work, 

 from heat at a constant ^cw^jw-rt/j/re, contrary to the second law of thermodynamics. 

 A potential difference is therefore impossible in this case. 



To control this conclusion several experiments were executed at my request by 

 Dr. Lash Miller in my laboratory. They led to the result that no potential 

 difference exists between the same metal in the solid and in the liquid state at the 

 melting-point. The cells consisted of tin in a molten mixture of sodium and 

 jjotassium nitrates, of lead in molten zinc chloride containing some lead chloride, 

 and of mercury in sulphuric acid of 60 per cent., and mixed with mercurous 

 sulphate. In no case was it possible to detect a potential difference of 001 volt. 

 The above-mentioned calculation from the latent heat would give some centivolts. 



It is easy to deduce from the second law of thermodynamics that at the 

 melting-point there must be a break in the E.M.F., although not a jump. The 

 case is fully analogous to the case of vapour pressure upon water and ice. Indica- 

 tions of the break were observed, but no measurements were made. 



If the metal is combined -with another, farming a cell of the Daniell pattern, 

 and the chemical energy of the combination is converted by the current generated 



1892. T T 



