071 the Crystallization of^alts. 45 



stance proves that t!ie solvent power of water does not depend 

 solely on atmospherical pressure. 



I poured mercury into f;\e barometrical tubes to within three 

 finger breadths of the top, taking care to remove all the small 

 particles which stuck u.vm the sides, and I filled them up with 

 a boiling and sn'i,-:.*. '.1 solution of sulphate of soda: having 

 instantly inverted them ovei a bath of mercury, the solution 

 crystallizfcd in each tube in proportion as it rose in its upper 

 part. 



I repeated i,his experiment upon five other tubes in which 

 mer-.iiry had been boiled, but the solution did not crystallize in 

 any. I iutroduced a small bubble of air occupying about the 

 two hundredth part of tlie tube, and by shaking", or frequently 

 even without this precaution, crystallization took place promptly. 

 The same eiTcct may be produced by using hydrogen, carbonic 

 acid or nitrous gas instead of air. It would seem therefore that 

 a very small quantity of any gas is sufficient to determine cry- 

 stallization. 



I introduced a concentrated and boiling solution of sulphate 

 of soda into five oarometrical tubes in which mercury had been 

 boiled. Twenty-fcur hours afterwards no crystals were ob- 

 served in any tube, although they had been shaken considerably 

 in the interval. Nevertheless, after some violent concussions 

 produced by suddenly dipping the tubes in the mercurial bath, 

 I succeeded in crystallizing the salt in three of them, in a few 

 minutes. It was always in the upper part of the tube, where 

 the shaking had collected some very small air-bells, that cry- 

 stallization commenced. The solution in the two other tubes 

 not having undergone any change under the same circumstances, 

 I operated the crystallization speedily by introducing into one a 

 crystal of sulph ito of soda, and into the other an air-bell. I 

 ought to observe that I never succeeded in crystallizing tlie solu- 

 tion by shaking it in a small apparatus ;' but I ascribe this 

 failure to my not being able to produce an agitation so violent 

 as in the baro-.netrical tubes. I did not succeed any better on 

 vibrating the tul)cs by means of the bow of a violin, after having 

 secured one of their extremities in a vice. 



It seems fiom the^e experimcuts, that simple shaking does not 

 produce crystallization unless there is air in the apparatus ; but 

 on the one harid I produced it by shaking barometrical tubes 

 freed from air witli the greatest care ; and on the other hand, 

 this precaution was almo-t alvvays insufficient for the small ap- 

 paratus, in which I found, on opening them under water, a volume 

 of air sometimes amountin.;- to one-thirtieth of the empty space 

 in the tiil>e. Besides, is it Mot \ cry remarkablo that the steam 

 ol water which is developed in the apparatus, and the tension of 



which 



