Genth.] 698 (Oct. 17, 
To my four analyses I add for comparison those of Mr. Mackintosh and 4 
| Dr. Winkler. 
| | Ehren- 
| Stoneham, ‘a friedersdorf, ; 
{| I Ir Ais IV Mackintosh, Winkler, Winkler, | 
P.O, == 41.76 — 48.01 — 48.88 — 43.48 — 44.381 — 41.51 — 42.44 
| BeO = 14,60 — 15.01 — 15.17 — 15.04 — 15.76 — 14.84 — 8.61 
Al,O, = 017— 0.22 — 0.009— 0.20 -— —— — 2.26 — 6.58 
Fe,0; = 0.48— 0.81—- 0.49— 0.15 —- ——- — 118 — 1.77 \ 
MnO = 0.09— 0.08— 0.12— 011 — — — — — —- | 
| JaO = 33.96 — 34.06 — 33,74 — 33.65 — 33.21 — 33.67 — 34.06 
| H,O = — ——= 70.61 — 70.61 — — 659 — 6.54 
Fl= ——?6.04— ——— 8,93 11.82 ‘ 
| 102.12 104.06 
| LessO = 8.76 4.76 
| cae ' 
98.36 99.84 
Silicic acid and the alkalies belonging to the albite, etc., have not been 
determined, it should be remembered however that 0.20 alumina repre- 
sents one per cent of albite. 
The analysis made by Mr. Mackintosh and myself show that herderite 
is anhydrous beryllium-calcium phosphate and fluoride,—with traces only 
i of alumina and ferri¢ oxide and perhaps a little water. The traces of alu- 
| mina are owing to slight admixtures of traces of mica and albite. 
Somewhat doubtful is the exact quantity of fluorine which it contains, 
Mr. Mackintosh determined its quantity from the excess of lime which he } 
i found, <A determination which I have made in the same manner gave me 
| 
i 
a far lower result, instead of 11.32, only 8.93 per cent. My direct fluorine 
determination is probably too low, owing to the incomplete decomposi- 
tion of the mineral by fusion with silica and sodium carbonate and the 
difficulties in the separation of fluorine from such a solution. A doubt 
also exists as to the 0.61 per cent loss by fusion with plumbic oxide, 
whether it is water or lead fluoride. 
As all my material was used up I could not attempt any other determi- \ 
nation for clearing up these doubtful points. 
It is to be regretted that the results of Dr. Winkler’s two analyses are 
| so very unsatisfactory, and that he has sacrificed the very precious Ehren- 
friedersdorf herderite by employing incorrect methods for his analyses. 
By ignition he has volatilized the greater portion of the fluorine, then 
by evaporation with nitric acid the rest may have gone (although nitric 
acid is less liable to drive off hydrofluoric acid than hydrochloric acid 
would be); therefore when he subsequently tested for fluorine, there was 
no more left than sufficient to give a doubtful reaction. ¥ 
Although it is stated (Rose’s Qual, Analyse, Leipzig, 1867, p. 212), in- 
correctly as I believe, that hoiling with sodium acetate does not precipitate 
