200 THE ROYAL SOCIETY OF CANADA 



There was no change in the appearance of the material after it had 

 been heated to this temperature. It was placed, together with a dish 

 containing water, under a bell jar, and allowed to remain for a day, 

 when it was found to have regained very nearly its original weight 

 (all but 0-3 per cent.)- This material was then ignited; almost 

 immediately it began to darken in colour, and soon the entire mass 

 appeared nearly black, but with further heating the mineral became 

 white again, and remained so on cooling. The total loss in weight was 

 22-61 per cent. From the change in colour it is evident that this loss 

 is not entirely due to the water expelled, but the cause of the blacken- 

 ing was not determined; it is apparently a characteristic feature of 

 all saponite. A direct determination of the water by the Penfield 

 method gave 20-42 per cent., but this is probably low, for the last 

 traces of water appear to be very tenaceously held, and by the ignition 

 method it is only after very prolonged heating that a constant weight 

 is attained. After complete dehydration in this way, the substance 

 does not re-absorb water when exposed to a moist atmosphere. 



The mineral is essentially a hydrous silicate of magnesium, 

 containing also aluminium, and although the analysis shows rather 

 less of the latter constituent than is usual in the recorded analyses 

 of saponite, the mineral is best referred to this species, which it further 

 closely resembles in its general physical and blowpipe characters. 



The recorded analyses of saponite vary between very wide limits 

 for all the principal constituents, and this is no doubt to be explained, 

 in part at least, on the assumption that the material analysed has in 

 most cases been impure, or has consisted of a mixture of two or more 

 substances. However, the analyses without exception show alumina 

 to be present, and in most cases ferric oxide also, these two sesquioxides 

 averaging more than 10 per cent., and even amounting to as much as 

 20 per cent, on occasion. Generally speaking, the sesquioxide and 

 magnesia contents vary more or less inversely. The mineral appears, 

 in every case, to lose about two-thirds of its water content at 100°C. 



It may be doubted whether saponite, considered as a hydrous 

 silicate of magnesium and aluminium, should be regarded as a definite 

 species. Many analyses would approximate very closely to mixtures 

 in varying proportions of, say, deweylite and kaolin. Clarke^ suggests 

 that "saponite is perhaps normally H4(MgOH)2Si207, although the 

 analyses all show admixtures of some aluminous compound." The 

 percentage composition calculated for Clarke's formula is given in the 

 first column of the table above. It is perhaps a curious chance that 

 the Montreal material gives almost exactly the ratios required by this 



' The constitution of the Natural Silicates, U. S. Geological Survey, Bulletin 

 588, 1914, p. 109. 



