The introduction of the image analysing computer (Quantimet 720) makes possible the rapid and accurate measurement of components in thin sections and extends the range of measurements possible. Two techniques for presenting material for analysis are outlined. Models are used to demonstrate the measurement of size, irregularity and orientation of components and their discrimination according to shape.

WITH TWO PLATES SummaryImage analysis was used to measure and characterise the voids in two sets of soil samples; (i) a surface water gley soil of the Deighton series and (ii) soils from a compaction experiment. The results show that the Deighton soil contains two impervious horizons separated by a much more porous horizon. In all the horizons most of the pore space is due to pores >180 pm diam. In the case of the soils from the compaction experiment compaction was shown to change the shape, orientation and size distribution of the pores considerably. The range of measurements demonstrates that image analysis is applicable to several areas of soil research.

The water in replicate samples from sub-surface horizons of a clayey and a fine-silty soil was removed either by oven-drying, freeze-drying or by acetone-replacement prior to resin impregnation. Pore space photograms (pores >60 pm diam.) from thin sections were analysed on a Quantimet 720 image analysing computer following each drying technique. During oven-and freeze-drying the clayey soil contracted considerably and subsequently satisfactory impregnation of these samples was either difficult or impossible. This macro-shrinkage caused a decrease in number and lengths of most planar pores and a decrease in size and number of all intra-aggregate pores. The loss in pore space and continuity obviously affected resin impregnation. The clayey samples in which water was replaced by acetone, and all samples of the silty soil, however pretreated, showed no measurable macro-shrinkage and all impregnated well. Oven-drying the silty soil appears to increase the porosity, causing an increase in size of all pores. A mechanism for this apparent enlargement is proposed. It is evident that acetone-replacement of the soil water prior to resin impregnation is the best of the three methods used here and should be seriously considered when image analysis is to be carried out on resin-impregnated blocks from soils of similar or related textures.Brown (IOYR 513) clay; many fine prominent strong brown (7.5YR 518) and greyish brown (IOYR 512) mottles; moist; moderately developed fine and medium angular blocky with smooth shiny brown (IOYR 513) ped faces; stoneless; abundant fine pores; common former fine root paths outlined in greyish brown; firm; very plastic. very sticky.Hamble (TL7 11673 1/30-39 cm, Eb horizon) Yellowish brown to brown (10-7.5YR 514) silt loam to silty clay loam; many medium dark brown (7.5YR 412) mottles; occasional inclusions of light yellowish brown and dark greyish brown (IOYR 412); moist; weakly developed medium angular blocky; few very small flints; abundant very fine macropores; rare fine roots; friable; common earthworms.The particle size distribution analyses of both horizons are as follows (sand 2 mm to 60 pm; silt 60-2 pm; clay < 2 pm):Windsor: sand 6 per cent, silt 35 per cent, clay 59 per cent. Hamble: sand 16 per cent, silt 66 per cent. clay I8 per cent. SamplingLedges of I x 2 m dimension were exposed at 46 cm depth in the Windsor soil and at 35 cm depth in the Hamble soil. Sixty Kubiena tins, each 7.5 x 6 x 4 cm were inserted into each ledge in groups of three so that large (7.5 x 6 cm) horizontallyoriented thin sections could be prepared. The samples from each group were x e d for oven-drying, freeze-drying and acetone-replacement respectively. Pre-treatmentOven-drying. The soil samples were placed in an oven set at 40°C and leR for 2 weeks before resin impregnation.