Church — On Native Phosphates of Aluminium. 559 



centages were simply thrown on the alumina, the discrepancy between 

 experiment and theory would not be unreasonably large, particularly 

 as the alumina did contain traces of ferric oxide. The present ana- 

 lysis is alluded to in my classification table (§ XT.) under the signature 

 "Coeruleolactite," for the published analyses of that mineral (by Peter- 

 sen and by Genth) from two localities, Eindsberg, jS'assau, and Ches- 

 ter Co., Pennsylvania, carefully studied, lead to the forn^ula SALOs, 

 2P2O5, 10 aq. above given. I do not, however, find that any experi- 

 ments have been made, by the analysts just named, to determine the 

 tenacity with which the water in coeruleolactite is held. 



§ XIII. Coeruleolactite. — It seemed important to secure some in- 

 formation on this point in order to learn what warrant there might be 

 for including these four minerals found respectively at Striegis, at 

 Nobiya, at Eindsberg, and in Chester Co. Pa., under the same specific 

 name. With this object in view, I made the following trials with a 

 picked specimen of coeruleolactite from the last-named locality : the 

 percentages have been recalculated after the deduction of the undis- 

 solved silica : — 



Anal. 10. -1235 gram lost in vacuo -004 gram 'K^O = 3-24% 



•1235 „ ,, in water-oven -003 „ HoO = 2-43% 



•1235 „ „ at 100° -007 „ HoO = 5-67% 



•1235 „ „ on ignition -018 „ H2O = 14.57%. 



It is worth while recalculating these results once more after de- 

 ducting the small percentage of water lost in vacuo. Then we find 

 that the vacuum-diied mineral lost 8*37% H2O at 100°C. in the air- 

 oven ; and 15 '06% more on ignition. This result confirms my 

 conclusions as to the position of the Striegis mineral referred to in 

 § XII., and tends to show that the group of minerals under discussion 

 may be referable to a single species, differing from wavellite not 

 merely in a lower percentage of water but also in constitution. It is 

 true that the analytical results, so far as the percentages of water lost 

 at different temperatures are concerned, are not alike in the two 

 minerals analyzed, but in both cases we have a notable proj)ortion of 

 water lost and a notable proportion retained at 100°. This sufiices to 

 suggest for these minerals a constitution in which both the normal 

 hydrate and the normal phosphate bear parts, being p-soeiated in 

 such a way as to retain in a measure their ordinary relations to 

 their own combined water. 



§ XIY. Evansite. — This species, from Zsetczik in Hungary, is at once 

 the most basic and the most highly hydrated of all the native aluminium 

 phosphates. It occurs in colourless and nearly transparent nodular 

 concretions, and was formerely mistaken for allophane. The late 

 David Forbes analyzed^^ it with such care that nothing remained to be 

 accomplished, save to ascertain the condition in which the water exists 



^^ Fhilosophieal Magazine, November, 1864. 



R. I. A. PROC, SEE. II. VOL. HI. SCIENCE. 3 A 



