128 



HARDWICKE'S SCIENCE-GOSSIP. 



Santini (" Gazetta Chimica Italiana," vol. xiv. 

 p. 274) has already shown that the flame of hydrogen 

 assumes all the colours of the spectrum, and now 

 replies to the objection that this coloration is due to 

 impurities of the gas, as at first prepared, by making 

 it from potassium formate heated with potash. He 

 still observes the same phenomena with this. To 

 show these colours, the hydrogen should be collected 

 in a bell jar about 8 inches long, and 2 inches 

 diameter, which should be then held with its mouth 

 downwards, a light applied and the vessel gradually 

 inclined. A flame pours upwards in which all the 

 prismatic colours may be observed as the jar 

 approaches the horizontal position. Carbonic oxide, 

 sulphuretted hydrogen, methane, and vapours of 

 alcohols, ethers, &c, display similar colours. 



In the current volume of " The Proceedings of the 

 Royal Society," W. N. Hartley states that the 

 sensitiveness of the spectrum in detecting magnesia 

 is practically unlimited ; that it is possible to obtain a 

 definite magnesium spectrum from a spark carrying 

 only a one thousand millionth part of a milligramme, 

 (a milligramme is a little more than T y o of an ounce). 

 Also that a solution containing one part of magnesia 

 in ten thousand million parts of water displays two 

 of the characteristic lines of magnesia. These 

 quantities are inconceivably small, but still the 

 substance is outspread and continuous ; no physical 

 indication of discrete molecules is displayed. Some 

 of my readers may know that I am a heretic in 

 reference to the actual physical existence of any 

 ultimate atoms or molecules, regarding all the 

 speculations concerning the limits of littleness as 

 vain and worthless, quite as vain as discussing the 

 boundaries of space. 



A very interesting and important paper was 

 recently read at the French Academy of Sciences, 

 (Comptes Rendus, vol. xcix. p. 1072) by J. Thoulet, 

 describing experiments which justify the conclusion, 

 that an attraction is exerted between a dissolved salt, 

 and an insoluble solid immersed in the solution, and 

 that the amount of this attraction varies with the 

 surface of the solid. Thus when marble, kaolin, 

 cjuartz, or other solid, is immersed in a solution of 

 barium or sodium chloride in which they are chemically 

 inert and insoluble, they nevertheless disturb the solu- 

 tion, and render it weaker by effecting a deposition 

 upon themselves of some of the dissolved salt. 



Important practical consequences follow from this. 

 One of the oft-repeated fallacies of the half-learned, 

 but not of the unlearned, is that of stating that a 

 filter can only remove mechanical impurities from 

 water, that it cannot remove matter which is there 

 dissolved. This statement has been disproved by 

 experiment. By repeated nitrations sea water may 

 be rendered less and less salt, until it becomes nearly 

 if not quite tasteless. This fact has hitherto been 

 rather puzzling, but' is now readily explained by the 

 adhesion of some of the salt to the filtering medium. 



It should be, however, understood, as a matter of 

 course, that in order to obtain this result, fresh 

 filtering material (sand for instance) must be used at 

 each filtration, since the sand thus used ultimately 

 takes up its utmost attainable supply of salt, and 

 then may rather give some back to fresh water than 

 take any more away from salt water. 



As far back as 1878 similar results were obtained 

 by Bayley, but in a different manner. He let fall 

 upon the white blotting-paper used for filtering, drops 

 of various solutions, and observed that generally the 

 salt remained near the centre, and that a ring of water 

 extended round this. By using solutions of metallic 

 salts which became blackened by hydrogen sulphide, 

 he was able, by simply applying this reagent, to 

 obtain a picture of the diffusion, and produced 

 similar pictures by staining the blotting-paper with 

 turmeric or litmus, and then adding allkaline or acid 

 solutions which change the colour of the stain. The 

 greater the dilution, the broader the water ring 

 surrounding the coloured spot and indicating the 

 position of the salt. Concentration of the solution, 

 heat, and looseness of the texture of the paper, 

 increase the mobility of the solution, i.e. the distance 

 to which the dissolved matter may stand before the 

 water leaves it. The mobility of different salts 

 varies, and in mixed solutions they act independently 

 of each other. 



These results have been recently confirmed by 

 J. U. Lloyd (" Chemical News," vol. li. p. 51), who 

 modifies the experiments by dipping chips of blotting 

 paper into various solutions, and observing how far 

 the substances in solution climb up the paper before 

 they are left behind by the water. Various solutions 

 were "thus tried ; in some cases, as with very dilute 

 solutions of ferric sulphate, the salt just creeps up 

 above the surface of the solution, while the pure 

 water travels to the end of the paper (five inches). 

 A concentrated solution of the same travels with the 

 waters, no separation taking place. Such a solution 

 of this salt is like a syrup. 



When solutions of ferrous sulphate, copper sulphate 

 and ferric sulphate were mixed, each salt showed a 

 limiting line at the same distance above the solution ■ 

 as when tested separately and of corresponding 

 strength ; the ferrous sulphate travelling farthest, the 

 copper sulphate next, and the ferric sulphate lagging 

 behind. Other salts behaved in like manner. Even 

 sulphuric acid is separated from water when the 

 solution is dilute ; water quite free from acid passing 

 onwards. By bending the blotting-paper over at a 

 height above the reach of the salt, and allowing the 

 further end to hang below the level of the solution, 

 a perfect filtration of ferric sulphate was effected, 

 drops of water free from iron salt falling from this 

 lower end. Quantitative experiments were made 

 on this and other solutions, showing their relative 

 distances of travelling in solutions of measured 

 strength. 



