138 GEOLOGY" OF THE HIGH PLATEAUS. 



horizontally in the amount of rise as we pass from point to point. It may 

 also rise more rapidly in the lower part of the field than in the upper ; and 

 as between many fields, local circumstances may accelerate beyond the 

 mean rate the fusion and expansion of one class of rocks or retard the 

 same effects in others. Thus, while there is a normal or typical order of 

 eruptions, it may become liable to not infrequent exceptions arising from 

 want of exact homogeneity of conditions. 



There are several sub-groups of rocks which present difiiculties some- 

 what greater and have the appearance at present of being somewhat anom- 

 alous. These are principally quartz-propylite and quartz-andesite or 

 dacite. These rocks are much more siliceous than the other members 

 of the groups to which they are mineralogically most nearly allied, being 

 about as siliceous as the more acid trachytes. They have apparently had 

 their epochs of eruption coevally with the hornblendic members of their 

 respective major groups, while according to the theory their epochs should 

 have fallen much later. I am unable to harmonize these apparent anomo- 

 lies with the main theory upon any considerations which at once carry with 

 them a conviction of intrinsic probability and an obvious reason for their 

 exceptional relations. They are comparatively rare rocks, and do not 

 occur in very extensive masses ; their physical constitution and properties 

 are much less known than their chemical and mineralogical. Their infe- 

 rior bulk, however, does not break the force of the anomaly if it be real. 

 Considerations like the following, suggest themselve's : The theory assumes 

 that the physical properties (density and fusibility) have a definite rela- 

 tion and dependence upon the proportion of silica which a rock contains. 

 Although this is approximately true, it is in all probability not rigorously 

 so, and indeed the probabilities, so far as fusibility is concerned, are that 

 the variations from definiteness in the dependence of fusibility upon the 

 percentage of silica are in some cases very notable, though these varia- 

 tions may not impair the general law as an approximate expression of the 

 truth. In spite of their high percentage of silica, therefore, these rocks 

 may turn out to be exceptional in having a degree of fusibility correspond- 

 ing very closely to that of the hornblendic members of the major groups 

 to which they belong. While, therefore, we cannot claim the dacites and 



