RUM. GEOLOGY. 497 



trope for which the island of Rum has long been cele- 

 brated. It is partly basaltic and partly amygdaloidal ; 

 a judgment only formed from the inspection of fragments, 

 since it is I believe every where inaccessible. These 

 masses have a nearly common character, both mineral 

 and geological, and may safely be considered as one 

 deposit ; but I doubt much if it is possible to obtain any 

 evidence by which to determine the relation which they 

 bear to the trap described before. The differences are 

 as great as those already mentioned between the mountain 

 trap and the stratified trap of Sky. Did they approximate 

 in any place, some assistance might be procured in solving 

 this question, but I believe that there is every where a 

 geographic interval between them. It is fruitless to accu- 

 mulate conjectures, but it is not useless to confess igno- 

 rance, since it will point out to other travellers the blanks 

 to be filled, and the errors to be corrected ; while I may 

 state in the words of St. Augustin my conviction that in 

 this science also " Melius est dubitare de occultis quam 

 litigare de incertis." 



In concluding this description of the rocks which form 

 this island, I do not imagine that any material points 

 have been omitted; but, if there are such, they will be 

 found in a curved line between Loch Scresort and Giurdil, 

 at the inner edge of the sandstone. They are thus pointed 

 out for the correction of future observers. I have indeed 

 been informed by the inhabitants that coal similar to that 

 of Canna has been found in this part of the island. I did 

 not see the place, but it probably exists in some of the 

 trap rocks last described. 



IT only remains to describe the rocks out of place, and 

 the independent minerals. 



On the shores of Harris there are some fragments of 

 porphyry, with a felspar basis of a brown purple colour, 

 which by their position seem to have descended from 



VOL.1. KK 



