﻿110 Penjield and Harper — Chemical 



The ratio of P 2 6 : BeO : CaO : (F 2 +H 2 0) = 1:2:2:1 nearly. 

 The high temperature at which the water is driven off indi- 

 cates that it is not water of crystallization, but is very firmly 

 united in the mineral as hydroxvl, every H 2 representing two 

 hydroxyl groups, and the OH being probably isomorphous 

 with F. The ratio of F : OH is nearly 1 : 1, or more nearly 

 3 : 4 in our analysis. In the calculated analysis we have 

 used the ratio F : OH=l : 1, but recognize that it is prob- 

 ably simply a case of isomorphism. The composition of 

 herderite is, therefore, an isomorphous mixture of CaBeFP0 4 

 with CaBe(OH)P0 4 , which may be written CaBe(FOH)P0 4 , or 

 a salt of phosphoric acid, two of whose hydrogen atoms have 

 been replaced by a bivalent element, and the third likewise 

 by a bivalent element whose other free affinity has been satis- 

 fied by a fluorine atom or hydroxyl. This is the same compo- 

 sition as that proposed by Mackintosh, except that he regarded 

 the mineral as simply the fluorine compound and did not 

 detect the water. Chemically herderite is closely related to 

 the three minerals — wagnerite, triplite and triploidite, whose 

 compositions are respectively Mg 2 FP0 4 , (FeMn) 2 FP0 4 and 

 (FeMn) 2 (OH)P0 4 . These three minerals offer the best illus- 

 tration we have of the isomorphism of F and OH, and we feel 

 that we have in herderite another strong proof of the correct- 

 ness of this interesting relation. In crystallization the minerals 

 vary, herderite being orthorhombic, triplite questionable, and 

 wagnerite and triploidite monoclinic. The latter two have, 

 like herderite, a prismatic angle of nearly 120°. 



2. Analysis of beryl. 



In 1884 one of us* published a series of analyses which 

 showed that alkalies are present in beryl sometimes to quite a 

 large extent, and that water is always given off by ignition. 

 At the time the analyses were made the author had had little 

 experience in the determination and separation of beryllium 

 and used the ammonium carbonate method, which seemed to 

 him to be the most accurate. Our experience has taught us 

 that the method used in the manner described in the above 

 mentioned article gives too low results for beryllium, and the 

 analyses are only of value as showing to what extent alkalies 

 and water are present in beryl. Since the publication of the 

 article it has always been the author's wish to make a series of 

 experiments on the separation of aluminum and beryllium and 

 renew the investigation. 



We have considered it best not to make a long series of 



* S. L. Penfield. This Journal, III, xxviii, 25. 



