In tracing the geographical distribution of the puy-eruptions we are first impressed with the force of the evidence for their extremely local and restricted character ([Map IV.]). Thus in the area of the basin of the Firth of Forth, which may be regarded as the typical region in Britain for the study of this form of Carboniferous volcano, traces of them are abundant to the west of the line of the Pentland Hills. To the east of that line, however, not a vestige of puy-eruptions, save a few sills of uncertain relationship, can be detected, though the same series of stratigraphical horizons is well developed on both sides of the Lothian coal-field. Again, to the westward of the Forth basin over the area of Stirlingshire, Lanarkshire and Renfrewshire lying to the north of the great volcanic plateau, no record of puy-eruptions has been noticed. Immediately to the south of that plateau, however, these eruptions were numerous in the north of Ayrshire. Yet the rest of the Carboniferous area in that large county has supplied no relics of these eruptions save at one locality—the Heads of Ayr. Lastly, while no trace of any younger display of volcanic activity occurs in the Merse of Berwickshire, east of the plateau series of that district, the ground immediately to the west abounds in puys, and contains likewise extensive sheets of tuff and beds of basic lavas connected with these vents.
Another fact which at once attracts notice in Scotland is the way in which the puy-vents have generally avoided the areas of the plateaux, though they sometimes approach them closely. As a rule, it is possible to distinguish the tuffs and agglomerates which have filled up these vents from those that mark the sites of the eruptive orifices of the plateaux. There are, no doubt, some instances, as in Liddesdale, where puys have appeared on the sites of the older lavas, but these are exceptional collocations.[442] On the other hand, many examples may be found where puys have risen in the interspace between the limits of the eruptions of two plateau-areas. Thus the tract between the Clyde plateau-eruptions on the west and those of the Garleton Hills on the east was dotted over with puys. Again, the southern margin of the Clyde plateau in Ayrshire, from Dalry to Galston is flanked with puys and long sheets of their lavas and tuffs.[443]
[442] A means of definitely placing some of these vents in the series of puy-eruptions is stated further on, at [p. 476].
[443] Reference may again be made here to the remarkable similarity between the Scottish Carboniferous puy-vents and those of older Tertiary time in the Swabian Alps so fully described by Professor Branco in the work already cited [p. 46]. Denudation in that region has bared the cones and exposed the structure of the necks which, down to even minute details, repeat the phenomena of Carboniferous and Permian time in Scotland.
ii. NATURE OF THE MATERIALS ERUPTED
A. The Lava-form Rocks
We have now to consider the nature of the materials erupted by the volcanic activity of the puys. The geologist who passes from the study of the plateau lavas to those of the puys at once remarks the prevalent more basic character of the latter. The great majority of them are basalts, generally olivine-bearing, in the various types embraced in the table on the following page. The olivine-free dolerites are generally found as intrusive bosses, sills and dykes. Such more acid rocks as andesites occur only rarely, and still more seldom are quartziferous masses met with in some of the bosses.
Dolerites and Basalts.—The great majority of the lava-form rocks connected with the puys are basic in composition, and belong to the family of the Dolerites and Basalts. They graduate, on the one hand, into ultra-basic rocks such as limburgite and picrite, and on the other, into compounds that approach andesites or trachytes in composition. A large series of specimens from Central Scotland was studied a few years ago by Dr. Hatch, who proposed a petrographical classification of the rocks, and arranged them in a number of types which he named after localities where they are well developed.[444] More recently the rocks have again been subjected to microscopic investigation by my colleague Mr. Watts, who, confirming generally Dr. Hatch's discriminations, has made some modifications of them. He has furnished me with a revised classification ([p. 418]), based on purely petrographical considerations. The doleritic and basaltic series may be grouped into two divisions, one with, and the other without, olivine, and each division may be further separated into a dolerite group, which presents an ophitic or sub-ophitic structure, and a basalt-group in which the groundmass is made up of felspar and granular augite, and possesses the "intersertal structure" of Rosenbusch, or consists of idiomorphic augite embedded in felspar substance. The term "sub-ophitic" is employed by Mr. Watts "to imply that the augite grains are neither very large nor very continuous, optically, and that they rarely contain entire felspar-crystals imbedded in them, merely the ends of a group of these crystals as a rule penetrating into them."
[444] This classification was given in my Presidential Address to the Geological Society, 1892, Quart. Journ. Geol. Soc. vol. xlviii. p. 129. See Report of Geological Survey for 1896.
Transitional forms occur between many of the following types by the increase or diminution in the relative proportions of the constituents. Thus it is not easy to draw a line between 2b and 2c; the latter again shades into 2d and 2e by the decrease of the felspar.