l v t On Trtbasic Phosphoric Arid. 



after cooling dissolved in water. The residue, after washing 

 out. was dark-green and crystalline like hornblende. 



|- was dissolved in hydrochloric acid, and after separating 

 silicic acid by evaporating to dryness three times, precipitated 

 by ammonia. The fluid containing the phosphoric acid was 

 also separated from silicic acid and then boiled for Beveral 

 hours with concentrated sulphuric acid, in order to convert 

 pyro-phosphoric acid into the tribasic modification. After 

 dilution to 150 c. c. phosphoric acid was precipitated by mag- 

 nesia mixture. This method is very (dean, accurate, and. 

 under some conditions, preferable to the \I< > method. 

 The results are as follows : 



(:,,—!. 3567 (b)— 1.5683 (c)— 1.577 



PO, 36.63 36.71 36.59 



Fe,0 3 33.41 33.51 33.48 



IK) 29.96 k J'.«. 7 s 29.93 



mean 36.64— PO, 

 33.47— Fe.O 

 29.89— HO. 



III. Method with SiO, (Fuchs' by boiling with water-glass). 



A- this method was and still is often used in the analysis of 

 alumina compounds, 1 tried t<> introduce il here and compare 

 its results. It is an exceedingly troublesome and disagreeable 

 method, especially when taking much silicic acid. The results 

 are good. I am, however, aide to reporl only two analyses, 

 the only ones thai oul of nine I could Bafely bring to an end. 

 The substance was dissolved in hydrochloric acid, precipitated 

 with caustic potassa, and then boiled for a considerable time 

 with a Bolution of water glass, which in this case was prepared 

 from pure silica and pure potassa, the Bilica having been pre- 

 pared from soluble glass. The greatest care has to be 

 taken not to add too much of the silicic acid Bolution, as with 

 an i of it. the mechanical difficulties of the analysis will 



