OF ARTS AND SCIENCES. 275 



for example, of roseocobalt and luteocobalt are represented respect- 

 ively by the formulas Co(NH3)5Cl3 and Co(NH3),;Cl3. I have kept 

 the old formulas only to permit an easy comparison with those given 

 by Wing, and the matter is not one of consequence in this place. 

 It may also be remarked that, as cerium is at present the only one 

 of the group which forms a well defined oxide higher than R2O3, 

 the two formulas II. and IV. cannot be generalized by substituting 

 the symbols of other elements for that of cerium. I have en- 

 deavored, however, to prepare such compounds by adding a solution 

 of Lco(S04)3 to one containing the sulphates of oxides other than the 

 cerite oxides, and then adding some oxidizing agent, as for instance 

 potassic permanganate, chlorine, or bromine. No decisive results 

 were obtained. It is at least probable that all the earths the sul- 

 phates of which in cold solutions give only slightly soluble double 

 sulphates with potassic and sodic sulphates will also give insoluble 

 double sulphates with sulphate of luteocobalt. These earths are, so 

 far as now known, CejOa, La^Os, NdjOg, PS.2O3, SmoOs, ScoOa, while 

 the following give soluble double sulj^hates : EroOj, Y0O3, YboOa, 

 and TbaOs. The four cerite oxides cited are not the only ones 

 which give insoluble crystalline precipitates with sulphate of luteo- 

 cobalt m the cold, but I am not at present able to give more accurate 

 information on this point. On the other hand, we meet in the case of 

 luteocobalt sulphate with some of the relations which present them- 

 selves when the alkaline sulphates are employed. Thus sulphate of 

 yttria is not precipitated by sulphate of luteocobalt when alone, but 

 when mixed with the sulphates of the cerite group more or less of u 

 double sulphate of luteocobalt and yttria is always thrown down, and 

 the same appears to be true for the sulphates of some other earths. In 

 such cases the mixed sulphates of earths and luteocobalt should be 

 gently heated in a muffle until the cobaltamine is completely decom- 

 posed and the excess of sulphuric acid is expelled. The residual sul- 

 phates of cobalt and the earths should then be dissolved in cold water, 

 filtered, and again precipitated by sulphate of luteocobalt, allowing the 

 mixture to stand twenty-four hours. The supernatant liquid appears 

 then to contain all but the four cerite earths. This point is, however, 

 not yet sufficiently proved, and I reserve it for further investigation. 



The solution from which the double sulphates of the cerite earths 

 and luteocobalt have been separated by decantation or filtration usu- 

 ally gives a more or less abundant crystalline precipitate on boiling. 

 The filtrate from this again gives a precipitate with ammonia. The 

 above stated facts are precisely those which we meet with in employing 



