The layered syenites in the Klokken intrusion consist of horizons of fine-grained, granular-textured ferroaugite syenite showing inverted cryptic layering, interleaved with coarser, laminated, more fractionated hedenbergite syenite. Distribution of hydrous mafic phases indicates build-up of water in parallel with magmatic evolution, and druses and pegmatitic segregations in the laminated syenites are evidence for late development of a gas phase. Feldspar bulk compositions are close to the minimum on the Ab-Or binary, with An decreasing from An7 to An1 with fractionation, and normal zoning in cryptoperthite crystals. Feldspars in granular syenites are transparent coherent cryptoperthites or braid microperthites; An-content is probably the main control of fine-perthite coarseness. Laminated syenite feldspars are turbid, coarse patch microperthites with rare relics of braid textures. This non-coherent coarsening was caused by interactions between feldspars and water entrapped at magmatic temperatures which was retained within the original lithologies to low subsolidus temperatures. Fe-Ti oxides reflect this water distribution, with regular trellis ilmenite-titanomagnetite intergrowths in less fractionated rocks and ragged granule exsolution in more advanced syenites. The sharp change in exsolution textures at granular-laminated syenite boundaries implies steep water-gradients within these interleaved rock types. Water was unable to penetrate the granular layers and did not circulate freely in the cooling intrusion.

The extent of 64 glaciers that existed in the Lake District during the Loch Lomond Stadial is inferred from end, hummocky and fluted moraines and illustrated by detailed maps. Glacier surfaces are contoured and the equilibrium firn lines of the glaciers are calculated. Firn line altitudes were strongly influenced by snowfall, glacier aspect and the transfer of snow by wind from adjacent high ground. Snowfall was associated mainly with south to southeast winds and the mean direction of winds responsible for snow transfer was probably slightly west of south. It is inferred that mean July temperature at sea-level was slightly below 8°C. Protalus ramparts and other snowbed deposits indicate the former presence of numerous perennial snow beds down to altitudes of 200-300 m although the widespread occurrence of fossil gelifluction sheets and lobes points to large snow-free areas above these altitudes in summer. Fossil valley-wall rock glaciers indicate a mean annual temperature at sea-level no higher than 1°C. During deglaciation the great majority of the small glaciers retreated actively for 100-150 m and then probably decayed in situ, but some, at least, of the larger glaciers were active during much of the time their margins retreated.

A sequence of upward fining cycles contain braid bar structures in the coarse member. Bars, medially placed in the stream are recognised either by radiating large foresets or by identifying bar tail sediments with converging dips to their foresets. Lateral bars are recognised by the growth of major sandstone sheets oblique to the direction of channel trend and over slough mudstone. The fine member contains structures which indicate a direction of flow in the same direction as the channel, suggesting that only the unlevéed part of the floodplain is preserved.

The direction of transport for the Old Red Sandstone sediments is perpendicular to the direction of bar accretion. This is accounted for by assuming the section dealt with to be on the NNW margin of a cone which builds outward to the NE, and where the river building the cone swings from one side to the other. An upward increase in cycle thickness within the section is possibly the result of an increase in the size of the cone and the consequent amalgamation of the coarse members.

Garnet-bearing xenoliths from Letseng-la-Terae display a range of textures from coarse to granuloblastic. Equilibration temperatures and pressures of primary phases are in the ranges 950-1400°C and 27-50 kb, respectively. Deformed lherzolites equilibrated throughout this temperature range but coarse xenoliths are restricted to low temperature equilibration.

All garnets display coronas developed during the reaction:

In some rocks, reaction has completely eliminated garnet.

In rocks where garnet is disrupted, the corona minerals are strung out in the fluidal texture indicating that reaction occurred before deformation. Rocks transitional to, and of granuloblastic texture, contain garnet and aluminous spinel; in addition ‘pools’ of minerals originating by dynamic separation of corona fragments are observed.

Chemical comparison between the corona minerals and minerals in a garnet-spinel rock and two spinel granuloblastites, suggests that these spinel-bearing rocks may be derived from normal garnet peridotite by a complex sequence of reaction, followed by deformation, annealing and chemical homogenisation. The conclusion that reaction and deformation took place at high levels in the upper mantle is contrary to some earlier hypotheses of shearing within the low velocity zone in response to continental plate movement, but is consistent with mantle diapir models.

In spite of the structural complexity of polyphase deformed gneisses and migmatites and the absence of distinctive marker horizons, these rocks are amenable to structural investigation and interpretation with resolution of their tectonic history possible, particularly in regions of reasonable exposure. Although methods such as those employing interpretation of stereo-plots can seldom be used, direct observation of refold relationships on the outcrop using all available structural, igneous and metamorphic features will ultimately lead to the resolution of the deformational history. This is achieved by determining the relationships of fold sets observed at the various localities examined throughout the field of investigation and then combining these by correlation according to basic stratigraphical principles. The effects of weak deformation normally not discernible in complicated successions may be detected by the simple form of folded planar and linear structures in igneous masses emplaced in complex structures at intermediate stages in the deformational history.

Synthesis of the total succession and determination of the overall sequence of deformation in particularly complex terranes is made easier if at least one distinctive set of structures can be recognised throughout the area. These key structures have the effect of subdividing the long succession into two in which refold relationships are generally more easily determined. An example of the method is given using photographs from the migmatite terrane in southern Finland.