558 



Proceedings of the Royal Irish Academy. 



available for analysis was not large, but tbe following results were 

 obtained with a quantity of material sufficient for an accurate analy- 

 sis : it should be said that "Oil gram silica bas been deducted fi'om 

 the amount taken for analysis : — 



Anal. 9. 



•4085 gram in vacuo over sulphuric acid 



lost 



gave 



•0125 



•0545 



■035° 



•164 



•225 



Ho.O = 3-0C% 



HoO in water-oven = 13-34% 

 HoO on ignition - 8-56% 

 AI0O3 = 40- 12% 



MgoPoO; = 35^23% PA- 



Ifow these results at the first glance might be taken to suggest a 

 wavellite formula, such as SAloOs, 2P2O5, 11 aq., which requires 

 almost precisely the percentage of water here given, and nearly the 

 percentages of AI2O3 and P2O5. But the veiy large loss of water 

 suffered by this Striegis mineral in the water-oven separates it from 

 the species wavellite at once. And if we consider the difficulty of 

 avoiding an excess of AI3O3 in an analysis, and the liability to count 

 as water the hydrofluoric acid which escapes on strongly heating a 

 phosphate of this group, I think it may be concluded that the ex- 

 perimental percentages of alumina and water are both above the truth, 

 while the phosphorus pentoxide should be supplemented, in critically 

 weighing the analytical results, by a small addition corresponding to 

 the lost fluorine. Prom these considerations it is clear that the speci- 

 men now under discussion differs decidedly fi'om wavellite in its slight 

 hold of most of the water present, however near it may be in its per- 

 centage composition. But there is reason to think that the mineral 

 under discussion, as well as wavellite itself, retains at ordinary tem- 

 peratui'es 2 or 3% of interstitial or mechanical moisture, which in 

 both cases is easily removed in vacuo over sulphuric acid. Recalculat- 

 ing the analytical results on this assumption, but without making any 

 of the other suggested corrections, we reach the following percent- 

 ages : — 





Experiment. 



Theorj-. 



Alumina, 



Phosphorus Pentoxide, . , 

 Water, 



41-41 

 36-34 

 22-60 



39-83 

 36-79 

 23-32 



100-35 



100-00 



The theoretical numbers here given are those required by the 

 formula 3ALO3, 2P2O6, 10 aq. If the excess in the analytical per- 



