Some sandstones contain sufficient felspar or calcium carbonate to form a flux when they are subjected to thermal metamorphism. At times a glass thus arises between the grains, and reacts upon the original quartz. When the igneous magma has melted up a sandstone or a quartzite, blocks of the sediment may remain surrounded by a mixed and recrystallised product from both rocks. Wright and Bailey[102] have studied an example in Colonsay, where a hornblende rock has partly dissolved a quartzite, the residual blocks being surrounded by "halos" of interaction, composed of quartz and alkali felspar. GNEISSES _Gneisses_ may be broadly defined as banded crystalline rocks in which felspar is visible to the unaided eye. Though this will include many igneous masses, it is doubtful if a more rigid description can be given. Numerous gneisses, in fact, owe their parallel structures to flow while in a molten state. Others are rocks that have been deformed by pressure, and their constituents have become drawn out along planes of solid flow. Where actual shearing has taken place, the minerals in the close neighbourhood of the planes of movement may become especially modified, ground down, and deformed. The foliated structure may then be marked by the appearance of differentiated bands. Such bands may also arise from the spreading out under pressure of certain large constituents, such as porphyritic crystals of felspar, which produce white bands, or of pyroxene, which will become modified into granular amphibole and will produce dark streaks through the rock. [Illustration: Fig. 19. Composite Gneiss. Gartan Lough, Co. Donegal. Fragments of mica-schist project from a gneiss, the banding of which follows the foliation planes of the schist. On the right the mass retains less schist and is more granitic.] Gneisses may also result from the intrusion of felspathic igneous rocks, in sheets of varying thickness, between the layers of a sediment or a schist (Fig. 19); or from the intrusion of one igneous rock into another, with varying degrees of interaction and absorption. It has often been presumed that the invaded igneous rock must have been in such cases in a plastic state. The supply of heat within the earth during such processes, and the action of the gases, corroding, as Doelter says, "like a blowpipe-flame," are, however, clearly sufficient to melt down large blocks, the residue being then carried forward as wisps or bands in the invader.