CHAP. XXXVIII 
RIVERS OF THE PERIOD 
underlain, sometimes overlain, with similar material. It is quite obvious that 
their deposition was contemporaneous with volcanic action in the immediate 
neighbourhood, and that at least part of their finer sediment was obtained 
directly from volcanic explosions. In wandering over the coast-sections ol 
these coarse deposits, I have been impressed with the enormous size 
of many of the stones, their resemblance to the ejected blocks of the 
agglomerate, and the distinction that may sometimes be made with more or 
less clearness between their rather angular forms and the more rounded and 
somewhat water-worn aspect of the other boulders. It seems to me not 
Fig. 273. — View of Dun Beag, Sanday, from tlie north. The island ol Rum in the distance. 
(From a Photograph by Miss Thom.) 
improbable that some of the remarkably coarse masses of unstratified con- 
glomerate in Ganna Harbour consist largely of ejected blocks from the 
adjacent vent. 
The only instance which I have observed oi erosion of the basalt con- 
temporaneous with the operations of the river that spread out this conglo- 
merate is to be found in the striking stack of Dun Beag already alluded to . 1 
1 This pinnacle of rock is referred to by Macculloeh in his account of Canna, and is figured in 
Plate xix. Fig. 3 of his work already cited. But neither his description nor his drawing conveys 
any idea of the real structure of the rock. 
