476 
SUMMARY AND GENERAL DEDUCTIONS 
BOOK VIII 
and sometimes in sills, the basic elements having tended to mass themselves 
towards the margins of the rock, leaving more acid material in the centre. 
The cases of Garabol Hill among the Dalradian schists of Scotland, of 
Carrock Fell among the Silurian strata of the Lake District, and of the 
Cramond pierite among the Carboniferous formations of Midlothian, with 
others that might be cited from various other regions and geological forma- 
tions in Britain, prove to what a considerable extent a separation of in- 
gredients may take place in a boss, and even sometimes in a comparatively 
thin sill before the molten mass consolidates. 
In the third place, there is good evidence that already before the 
magma is either intruded or extruded, and while it still lies within the 
internal reservoir, it may not possess a general uniformity of composition, 
but may have become more or less heterogeneous. In regard to intrusive 
rocks, the extraordinarily banded gabbros of the Tertiary series of Skye 
obviously proceeded from a magma in which the molten material consisted 
in some paits mainly ot felspar, and in others mainly of the ferro-magnesian 
minerals and iron-ores. Streams from these differently constituted parts 
of the magma were simultaneously or successively injected as sills into the 
older portions of the volcanic series, while, as the process of differentiation 
within the magma proceeded, still more felspathic liquid was left behind, 
to be thrust into cracks in the sills previously consolidated. 
Moreover, the banded basalts of the Tertiary plateaux show that this 
heterogeneity was not confined to internal intrusions, but maintained its 
place even when the molten material was ejected to the surface. The 
differentiation indeed is not so striking there as among the sills of gabbro ; 
but its presence, even in a less degree, proves that the separation of constituent 
minerals was not due to any general cooling of an erupted body of igneous 
lock, but was already developed in the reservoir from which the molten 
material was propelled to the surface. 
Attention has been called to the remarkable similarity of structure 
between these banded intrusive rocks and some of the ancient gneisses. The 
resemblance is so close that we may with every probability infer that the 
gneisses acquired their characteristic banding as intrusive masses of igneous 
rocks, discharged from heterogeneous magmas, like that which supplied the 
gabbros of the Guillin Hills. And as these gneisses belong to pre-Cambrian 
formations, we are thus led to the interesting result that the tendency to 
develop heterogeneity was already as characteristic of the magma-basins of 
the earliest geological time as it has been of those of later periods. 
The evidence of differentiation presented by superficial lavas, and by 
intrusive sills and bosses, acquires great interest when considered in con- 
nection with the changes which are seen to have occurred in the character 
ot the materials erupted during the course of a definite volcanic period. An 
attentive examination of the volcanic products of the various ages, so fully 
recorded in the geological structure of the British Isles, shows that a recog- 
nizable sequence in the nature of the materials erupted during a single 
volcanic period can be traced from the earliest to the latest times, and 
