406 On dejinite Proportions. 



the vapour of vinegar, they did not restore its colour : when 

 thrown int!) boiling water, they melted, without the extri- 

 cation oi any air; nor did the water become in any degree 

 alkaline. Was the teniperalure, or the want of affinity of 

 the compound For potassium, the cause that none was ob- 

 tained ? I at first sus|)ecled the former, and therefore ex- 

 posed, in the same apparatu-s, a solution of potass, with 

 quicksilver, ti) the heat of the same lamp. The quicksilver 

 at first ennltid much gas, but its quantity continually di- 

 minished as the solution becan)e more concentrated by the 

 effect of the heat: and the quicksilver, when poured out, 

 was found to be strongly impregnated with potassium. The 

 elevated temperature had consequently not prevented the 

 decomposition of the potass in this experiment, but only 

 in the beginnuig reduced the solution to the stale in which 

 it is found at common temperatures when less concen- 

 trated; and this experiment affords a positive proof that 

 the quicksilver, in the decomposition of potass, operates 

 not merely by collecting the product, but also by an affi- 

 nity, which is wholly wanting in Rose's fusible compound. 



B. Attempt lo ascertain the Composition of Potassium. 



It would be a fruitless labour to endeavour to add any 

 thing to the excellent essay of Davy on the properties of 

 the bases of potass and soda, even if it were possible to ob- 

 tain these substances as readily as he did without the assist- 

 ance of mercury. On the other hand, his attempts to Iht 

 vestigate the proportions of the»e compounds seem to re- 

 quire some confirmation; since they were made on too 

 small a scale, and the weight of the bases burnt was^ ascer- 

 tained indirecily, so that sinall errors may have had a very 

 material influence on the results. 



1 have attempted to perform the analysis of the alkalis 

 in such a manner as to obtain a result on which more de- 

 pendence may be placed, notwithstanding the tiiany diffi- 

 culties to wliich my niPthod is liable. 1 suffered a portion 

 of the amalgam of potassium of known weight to oxidate 

 itself in water; f saturated the potass thus ol)tained with 

 muriatic ;icid, and fused the salt that was formed. From 

 the weight wh.ich the mercury had lost, I inferred the 

 weight of the yxitassium, and from the analysis of the nm- 

 riate of ^lotass that was obtained, that of the potass formed. 

 In the beginning the ex|)eriments disagreed very much 

 among tiiemselves ; and in order to discover the cause 

 of the difference, 1 was obliged to repeat the same experi- 

 ment 20 or 30 limes. The Jlrst cause of the uncertainty 



was 



J 



