552 Proceedings of the Royal Irish Academy. 



Anal. 1. Gibbsite, from Eicbmond, Massachusetts — 



•309 gram lost -003 gram HoO in vacuo = 0-97% 

 •309 ,, ,, •001atlOO°C. = 0^30% 



•309 ,, ,, nothing in water-oven 

 •309 „ „ -106 on ignition = 34-30%. 



Anal. 2. Gibbsite, fi'om Yilla Eica, Brazil — 



•213 gram lost -002 gram H.O in vacuo = 0^93% 



•213 ,, „ nothing up to 100 C. 



•213 ,, ,, -069 gram on ignition 32-39%. 



Both the specimens were of great purity, neither containing any 

 silica, and a mere trace of phosphate being present in Xo. 2 only ; 

 this latter specimen was uniformly crystalline. The mean percen- 

 tage of combined water, corrected for accidental moisture, amounted 

 to 33-73 — a figure closely agreeing with the number demanded by 

 theory, namely, 34-48. This percentage of water is that required by 

 the formula AJ2O3, 3H2O = Alo HgOe, the normal aluminium hydrate. 



§ IV. Variscite. — "W^e have now to consider the mode in which the 

 water present in such of the native aluminium phosphates as are free 

 from excess of base is held. For the solution of this problem it might 

 have been thought that several native species would have served. But 

 there proved to be but one mineral of sufficiently definite character, 

 obtainable for this purpose in adequate quantity : this was variscite, a 

 phosphate found at Messbach, Saxon Voigtland, and described by 

 Breithaupt in 1837. It is clearly identical with the callainite of 

 Damour, a mineral the provenance of which is unknown, but of 

 which some polished beads and ornaments were found in a Celtic 

 grave at llorbihan. The two other normal aluminium phosphates 

 which have been described are the gibbsite of Hermann, and zepha- 

 rovichite. Of the latter, I have been unable to obtain an authentic or 

 adequate supply : of the former mineral nothing definite is known ; its 

 separate existence is perhaps doubtful. Thus my work has been per- 

 force limited to variscite. 



On submitting a carefully selected and prepared specimen of vari- 

 scite to the desiccating processes before mentioned, it was evident that 

 the water in this mineral was held far more loosely than that in 

 calaite, or true turquoise, and in wavellite. The analytical results 

 are here given : — 



Variscite (spec. grav. 2-24), from Messbach. 

 Anal. 3. -119 gi 



Anal. 4. -371 

 Anal. 5. -32 



1 Containiug some FesOa ; vide analysis 5. 



gave 



•084 gram 



JSIgoPoO, = 45-15% P0O5 



5J 



•0388 





AlA = 32-60% AI2O3 



lost 



•01 





H,0 in vacuo = 2-64% 



; J 



•08 





H2O at 100° = 21-11 



j> 



•004 





HoO on ignition = 1-08 



j; 



•0795 





HoO on ignition = 24-84 



gave 



•005 





PeA = 1-56. 



