COAL AND ITS CLASSIFICATION 269 
of forests or swamps. Many coal seams rest on a * seat- 
clay” or underclay, which is a high-grade fireclay as the 
constituents that act as fluxes in common clay have been 
extracted by plants that grew on it, when it acted as a soil. 
The clay often includes tree-roots (Stigmaria) which join 
trunks that rise through the coal seam, and around them 
the coal often includes fragments of fossil wood and other 
plant debris. The sandstone roof may contain fragments of 
tree-stems and leaves. Above this sandstone may occur 
another fireclay which is covered by coal; this recurrence 
may be many times repeated, the fireclay being always under 
the coal and containing tree-roots. The coal may be inter- 
rupted in places by bands or * horses” of sandstone; some 
of them join like the branches of a river, and were probably 
stream channels that have been filled with sand. Other 
coal seams, also formed im situ, may rest upon widespread 
sheets of clay without tree-roots, though the coal may con- 
tain tree-trunks; these seams grew in situations like the 
Dismal Swamp of Virginia or fenlands where a level sheet 
of clay has been covered with swamps and a growth of peat. 
The theory that coal has been formed from drifted vege- 
tation is true for other fields. Some coal seams are not 
underlain by fireclay, but may be interbedded with, or lie 
beneath it. For example, in the Kilmarnock field in Ayr- 
shire, the seam known as the Stone Coal is divided by three 
layers of which two are fireclay. The Hard Seam of Ayr- 
shire is covered by a bed of fireclay, 3 feet 6 inches thick, 
which rests on 4 feet 10 inches of coal, including 1 inch of 
shale, and the seam rests on a soft layer of coaly shale, 2 inches 
thick. Again in ‘the Patna Coal’ the seam consists in 
descending order of—fireclay 3 feet, top coal 6 feet, sand- 
stone 4 feet 9 inches. These fireclays moreover do not con- 
tain tree-roots, and tree-trunks in the coal may be hori- 
zontal or inverted, with the roots on the top, indicating that 
they drifted to their present position. Vertical tree-trunks 
rising from roots may occur in the Coal Measure sandstones, 
as well as in coal seams; thus in the fossil grove in White- 
inch Park, Glasgow, the Lepidodendron trunks are still 
attached to their roots; but they are in sandstone—not 
coal. Some coal seams which are shown by their vertical 
trunks and underlying fireclay to have been formed in situ,