the River Severn at Worcester. 263 



retained in solution by the sole additional agency of *142 grm. 

 of disodic sulphate, which, regarded as a solvent, appears wholly 

 inadequate. Parallel experiments made with magnesic sulphate 

 and chloride exhibit more striking results ; 1000 cub. centims. 

 of water may contain 1*3 grm. of magnesic sulphate, or 1*05 

 grm. of magnesic chloride, perfectly unaffected, so far as preci- 

 pitation is any evidence of chemical change, by the presence of 

 the atomic proportional, or 1*167 grm. of disodic carbonate. 

 The peculiar proneness of magnesium to form double salts doubt- 

 less conduces to this result, which is probably arrived at by par- 

 tial decomposition. 



In the presence of these facts it appears unreasonable to as- 

 sume that, in rivers which traverse gypseous deposits, decom- 

 position of the calcic sulphate which they dissolve must imme- 

 diately take place, if from another source they receive alkaline 

 carbonates. Such a decomposition would undoubtedly occur, 

 unless solvent agents intervened, whenever alkaline carbonates 

 passed into a river which contained more calcic sulphate than 

 •136 in the 1000 parts; but such rivers are of rare occur- 

 rence; a few small streams in the neighbourhood of Paris* 

 contain more than this quantity, but they are, I believe, 

 exceptional. The Thames, containing from *04< to *08 in the 

 1000, stands high in this respect among the great European 

 rivers, and '07 is the largest amount I have found in the Severn ; 

 these quantities, however, fall far short of what might coexist 

 in solution with an atomic proportional of disodic carbonate. 

 If calcic sulphate decomposed after its entrance into a water, it 

 would rather be a question whether it would not produce mag- 

 nesic in preference to any other sulphate : it is known that by 

 digestion f of magnesic carbonate with solution of calcic sul- 

 phate, double decomposition occurs ; and it seems by no means 

 improbable that the same change might take place in river- 

 water. It is worthy of note that, in all the specimens of Severn 

 water of 1866-67, the sulphuric anhydride is to the magnesia in 

 as nearly as possible the atomic proportion ; and when precipi- 

 tates form in the greatly concentrated waters, they contain but 

 little magnesia, the mass of which remains in solution together 

 with the whole of the sulphuric anhydride. 



But to assume changes of which no proof exists would be to 

 depart from the principle in accordance with which I have ar- 

 ranged the results of the present analyses. This is to assign to 

 the saline constituents those forms of combination in which they 

 may be considered to have entered the water. The saliferous 

 and gypseous marls of the New Red Sandstone have yielded the 

 sodic chloride and calcic sulphate ; and therefore I have expressed 

 in these forms of combination all the sulphuric anhydride and, 

 * Bisehof (Cav. Soc.) vol. i. p. //. f Ibid. p. 430. 



