140 Professor Baphael 3Ieldola [May 16, 



century. The problem has been attacked by chemists again and 

 again, but its solution presents extraordinary difficulties. These 

 products are never formed — even under the most favourable con- 

 ditions of division and with prolonged periods of exposure — in quan- 

 tities beyond what the chemists would call " a mere trace." Their 

 existence appears to be determined by the great excess of unaltered 

 haloid with which they are combined. Were I to give free rein to 

 the imagination, I might set up the hypothesis that the element silver 

 is really a compound body invariably containing a minute percentage 

 of some other element, which resembles the comj)ouud which we now 

 call silver in all its chemical reactions, but alone is sensitive to 

 light. I offer this suggestion for the consideration of the speculative 

 chemist.* For the coloured product as a whole, i. e. the product of 

 photo-decomposition with its combined unchanged haloid, Carey Lea 

 has proposed the convenient term '• photosalt." It will avoid circum- 

 locution if we adopt this name. The photosalts have been thought 

 at various times to contain metallic silver, allotropic silver, a sub- 

 haloid, such as argentous chloride, &c., or an oxyhaloid. The free 

 metal theory is disposed of by the fact that silver chloride darkens 

 under nitric acid of sufficient strength to dissolve the metal freely. 

 The acid certainly retards the formation of the photosalt, but does 

 not prevent it altogether. When once formed the photo-chloride is 

 but slowly attacked by boiling dilute nitric acid, and from the dry 

 photosalt mercury extracts no silver. The assumption of the existence 

 of an allotropic form of silver insoluble in nitric acid cannot be 

 seriously maintained. The sub-haloid theory of the product may be 

 true, but it has not yet been established with that precision which the 

 chemist has a right to demand. We must have analyses giving not 

 only the percentage of halogen, but also the percentage of silver, 

 in order that it may be ascertained whether the photosalt contains 

 anything besides metal and halogen. The same may be said of the 

 oxyhaloid theory : it may be true, but it has not been demonstrated. 



The oxyhaloid theory was first suggested by Eobert Hunt f for the 

 chloride ; it was taken up by Sahler, and has recently been revived 

 by Dr. W. E. Hodgkinson. It has been thought that this theory is 

 disposed of by the fact that the chloride darkens under liquids, such 

 as hydrocarbons, which are free from oxygen. I have been rej)eating 

 some of these experiments with various liquids, using every possible 

 precaution to exclude oxygen and moisture ; dry silver chloride heated 

 to incipient fusion has been sealed up in tubes in dry benzene, 



* I have gone so far as to test this idea experimentally in a preliminary way, 

 the result being, as might have been anticipated, negative. Silver chloride, well 

 darkened by loug exposure, was extracted with a hot saturated solution of potas- 

 sium chloride, and the dissolved portion, after precipitation by water, compared 

 with the ordinary chloride by exposure to light. Not the slightest ditierence was 

 observable either in the rate of coloration or in the colours of the products. 

 Perhaps it may be thought worth while to repeat the experiment, using a method 

 analogous to the " method of fractionation " of Crookes. 



t *Etsearches on Light,' 2nd ed. 1854, ji. SO. 



