Showing papers in "European Journal of Soil Science in 1968"

Distribution of pyrophosphate-extractable iron and organic carbon in soils of various groups

Abstract: Summary Potassium pyrophosphate (0.1m) removes very little Fe from crystalline Fe oxides at pH 10, but peptizes finely divided hydrous amorphous oxides and organic matter in soils. Fe and C contents of extracts from each horizon of twenty-six British soil profiles show distinctive patterns, independent of the residual dithionite-soluble Fe. Thus extracts of humus Fe podzols have maximum Fe and C in the B horizon, peaty gley podzol has maximum Fe in the B horizon but maximum C in the surface. These groups are differentiated from non-podzols which have maximum pyrophosphate extractable Fe and C in the surface horizon, decreasing with depth. Intermediate patterns help to quantify differences in soils of classes having properties of more than one soil group.

Changes in soil properties associated with the removal of iron and aluminium oxides

Abstract: Summary Red soils representing the major groups in which iron oxides are thought to exert a favourable effect on physical properties, have been treated with Na-dithionite and citrate, or Na-dithionite and 0.05 n HC1. These treatments removed the free iron oxides and some aluminium oxides and silica. Changes in the physical properties of the soils were measured by permeability, wet-sieving, mechanical dispersion, and swelling methods. These were compared with the changes induced by corresponding control treatments in which the soils were treated with Na2SO4 and Na-citrate, or Na2SO4 and 0.05 n HC1. The control treatments extracted some aluminium oxide and minor amounts of iron oxides and silica. In general, the control treatments decreased the stability of aggregates, increased the proportion of clay and silt but had little effect on the swelling. The treatments, which removed all the free iron oxide, caused no changes in excess of those due to the controls. The specific surface areas of samples treated with dithionite and acid, or sulphate and acid were determined by nitrogen adsorption methods. Removal of free iron oxides led in all but one instance to a reduction in the specific surface area, whereas sulphate and acid invariably increased the specific areas of the samples. The results show that most of the free iron oxides in these soils are present as small particles having little effect on the physical properties of the soil. The changes in physical properties brought about by dilute acid treatment indicate that the ‘active’ oxides in these soils are composed partly or wholly of the aluminium, iron, and silicon oxides which are readily soluble in dilute acid.

Variability of soil chemical properties in two uncultivated brown earths

Abstract: Summary The variability in pH, loss-on-ignition, exchangeable cations, extractable phosphate, and soil moisture, has been studied for the 0.15 cm horizon of uncultivated, unfertilized Brown Earths at two upland grassland sites in North Wales. The results show that there is a high degree of variability in samples taken at the same time on circles of 1 m radius at all sampling locations. For example, coefficients of variation for exchangeable cations (except sodium) average 33 per cent for twenty-two samples taken on the same occasion at three locations. This spatial variability prevented detection of any additional seasonal trends during a 22-week summer sampling period. It is essential to consider the importance of spatial variability in any pedological or ecological study.

The measurement and mechanism of ion diffusion in soils

Abstract: Summary Diffusion coefficients of soil phosphate calculated from flux to chloride form of anion-exchange resin-paper, assuming total depletion of the labile pool at the boundary, were too small and unrealistic. Thus not all the phosphate in the labile pool contributed to the diffusion process while being desorbed at constant pH in the presence of an indifferent anion—Cl—from the resin-paper. Desorption relationships under these conditions, using CaCl2 at atmospheric CO2 pressure, were markedly different from the relationship between the long-term 32P-exchangeable phosphate and the corresponding pore-solution concentrations. In the same soil containing different amounts of labile phosphate, a different desorption relationship was found for each phosphate level. The constant proportionality between amount of phosphate diffusing into resin-paper and square root of time was indicative of a constant concentration at the boundary. The resin-paper: soil system was therefore considered as an infinite composite medium in which diffusion through both resin-paper and soil were rate-limiting. The constant boundary concentrations were estimated by the use of the diffusion coefficients in the resin-paper, the phosphate adsorption isotherm for the resin-paper and the desorption relationship for each phosphate level in the soil. Diffusion coefficients, calculated using the boundary concentration appropriate to each phosphate level, were related to the slopes of the corresponding desorption relationships, resulting in values of the impedance factor similar to those found for K diffusion under similar conditions. The resin-paper method, however, does not provide an accurate enough measure of the diffusion coefficient of soil phosphate to be of any practical use. Until better and simpler methods are found, the diffusion coefficient may be calculated using the slopes of the desorption relationship and the separately determined value of the impedance factor.

The association of manganese and cobalt in soils—further observations

Abstract: Summary Soil Mn was mineralized either as birnessite or lithiophorite in five out of seven widely distributed soils taken from Bermuda, Europe, and the Middle East countries. The majority of the soil Co was associated with these minerals. This agrees with earlier work on Mn in Australian soils and suggests that these two minerals are the most common soil forms of secondary Mn minerals.

Fundamental objections to the 7th approximation

Abstract: Summary The 7th Approximation suffers from three fundamental ills. It demands an unattainable degree of precision and its exactly defined, mutually exclusive classes lead to inconsistency or absurdity when the system is used. Logical division is used out of context to construct a hierarchical structure, while the dispersed nature of soil distribution does not lend itself to hierarchical arrangement in any form. Its search for genetically significant properties on which to define classes involves circular reasoning, since the genetic significance of a property can usually be discovered only after the soils have been classified. Mutual exclusion and rigid adherence to the hierarchy result in the system failing to serve adequately the requirements of soil survey. The 7th Approximation explores to the limit the potential of logical division for soil classification yet remains unsatisfactory. The nature of soil distribution and the requirements demanded of soil classification suggest that a co-ordinate system would be more profitable, and a study of such systems is urged.

Electron‐microscope observations of iron oxides in some red soils

Abstract: Summary Electron-microscope studies of the clay fractions of red soils have shown that free iron oxides are present as small (50–100 A diam) discrete particles frequently clustered to form larger irregular aggregates. Although no evidence of iron-oxide coatings on clay minerals was found, an association of some of the oxide particles with the edges, but not the faces of kaolinite plates was observed.

Content of inositol penta‐ and hexaphosphates in some canadian soils

Abstract: Summary The combined amounts of inositol penta- and hexaphosphates in a number of Canadian soils of differing origin have been measured. The esters were precipitated as barium salts from alkali extracts and purified by anion-exchange chromatography; their identity was confirmed by paper-partition chromatography. An alternative method involving precipitation of the esters as ferric salts in acid medium was found to give much lower values, probably because of incomplete precipitation. Values for eighteen surface soils ranged from 20 to 71 and for twelve subsoils from 18 to 43 ppm P. The amounts found were related to the contents of both total phosphate and total organic phosphate, and accounted, on average, for 6 per cent of the former and 17 per cent of the latter. A correlation of +0.67 (P < 0.01) was found with orthophosphate retention capacity but correlations with soil N and C contents were poor. Amounts of the esters were higher in forest soils than in grassland soils.

Characterization of the inositol penta‐ and hexaphosphate fractions of a number of canadian and scottish soils

Abstract: Summary The nature of the inositol pentaphosphate and hexaphosphate isomers in a number of contrasting Canadian and Scottish soils has been examined. The mixed esters were extracted from the soil with alkali and separated from other soil phosphates by anion-exchange chromatography using HCOONH4 as eluent. The composition of the mixture was established by anion-exchange chromatography using a gradient of HC1 as eluent, followed by paper chromatography of the esters thus separated, and by paper chromatography of the hydrolysis products. Esters of myo- and scylloinositol together constituted more than 90 per cent of the mixture in most cases. Relatively small amounts of dl-inositol and neoinositol were detected in hydrolysates and it was estimated that esters of these cyclitols did not exceed 10 per cent and 1 per cent, respectively, of the total. The ratio of myo-+dl-inositol hexaphosphates to scylloinositol hexaphosphate ranged from 1.1 to 2.7 in the Canadian soils and 1.8 to 4.6 in the Scottish soils. The ratio of hexaphosphates to pentaphosphates ranged from 0.9 to 2.4 in the Canadian soils and 3.0 to 4.3 in the Scottish soils. The three soils with the highest pH values contained relatively large amounts of scyllo- relative to myoinositol hexaphosphate, but one very acid soil also contained a high proportion of this isomer and no consistent relationship was noted between the constitution of the inositol polyphosphate fraction and any other soil property.

Studies on the decomposition of plant material in soil. III. The distribution of labelled and unlabelled carbon in soil incubated with 14C labelled ryegrass

Abstract: The organic matter in soils containing decomposing 14C‐labelled ryegrass was fractionated chemically. Earlier work on these soils had shown that they contained a small fraction, heavily labelled relative to the rest of the soil organic matter, that was mineralized when the partially sterilized soils were incubated. Reagents effective in extracting heavily labelled‐C included cold o.in HC1, boiling saturated CaSO4 solution, and o.in Ba(OH)2, but neither these nor any other reagent tested could extract material as heavily labelled as that mineralized when partially sterilized soil was incubated. Reagents that extract heavily labelled‐C are poor extractants for humified material and are not strongly hydrolytic: the more vigorous the hydrolysis the smaller the proportion of labelled‐C in the hydrolysate. The amounts of labelled‐C dissolved by Ba(OH)2 from soils sampled after different periods in the field were directly proportional to the amounts of labelled‐C mineralized by those soils when partially sterilized (by exposure to CHC13 vapour), inoculated and incubated. Balance sheets are presented for the distribution of labelled and unlabelled‐C in fractions separated by hydrolysis with 6N HC1, by NaOH extraction, by neutral pyrophosphate extraction, and by oxidation with H2O2. The fraction remaining after hydrolysis with 6N HC1 was the most lightly labelled and had the widest C/N ratio. The percentage of labelled‐C in the material dissolved by alkali or by pyrophosphate was little more than in the material not dissolved, despite the presence in the soil of fractions differing at least twenty‐fold in intensity of labelling.

A dune catena on the clay plains of the west central gezira, republic of the sudan

Abstract: Summary Fixed dunes occupy 400 000 acres in the west central Gezira. Isolated dunes occur sporadically through the central Gezira, and trend NW from Sennar on the Blue Nile to El Geteina on the White Nile. They probably originated as channel deposits laid down by former distributaries of the Blue Nile. Minor aeolian re-sorting took place, and was followed by plant colonization and dune stabilization. Erosion of the dunes is locally active, resulting in exposure of the underlying clays. The catena comprises leached sands (Psammustents) on the dune crests and illuvial loams and clays (Orthustents) in the swales. Swale soils contain pedogenetic and relict lacustrine carbonate. The catena is often superimposed upon former soils, resulting in polygenetic profiles.

Pedogenic gradients of the polar regions

Abstract: Summary Soils of the polar regions are divided into three zones—Arctic Brown, Polar Desert, and Cold Desert. The Arctic Brown soil coincides approximately with the arctic tundra belt, the Polar Desert zone is penecontiguous with the high arctic, and the Cold Desert zone is represented by the ice-free sectors of Antarctica. The zones are broad and diffuse. Well-drained soils of the southern sectors of the Arctic Brown soil zone have embryonic podzolic affinities whereas soils of the northern portions of this same zone exhibit a low temperature calcification process akin to a cold steppe condition. Vascular plants form a continuous mat in the Arctic Brown soil zone but within the Polar Desert soil zone they are patchy or at times completely lacking. In the Polar Desert soils algae and diatoms appear to be the main contributors of soil organic matter. The Cold Desert soils of Antarctica are virtually void of organic matter. The above three soil zones, in effect, are sectors of a pedogenic gradient. Progressing from the northern fringes of the forested land, to colder climates, precipitation decreases, organic matter content of the soil decreases, and soil pH values tend to increase. In the Polar Desert soils, alkaline reactions are common whereas the Cold Desert soils of Antarctica are usually saline. Polar Desert soils and Cold Desert soils have many morphologic affinities. Tundra and Bog soils dominate areally in the main tundra belt but are confined to the local swales and depressions in the high arctic. Both Tundra and Bog soils tend to have higher pH values as one progresses to colder, drier climates. Neither Tundra nor Bog soils have, as yet, been recorded in Antarctica.

Relationships between soil volume used by roots and nutrient accessibility

Abstract: Summary Increasing the volume of soil available to plants growing in pots increased yields of oats, ryegrass, kale, and tomatoes and decreased the intensity of rooting; it increased the supply of N more than that of P, as P uptake was correlated with root intensity and N uptake was not. Responses to P fertilizer were greater in the larger volumes of soil, but responses to applied N were less than in small volumes. Isoquants showed that, when soil volume is restricted, it is more important to apply a mobile nutrient (nitrate) than an immobile nutrient (phosphate), although the value of either nutrient in compensating for inadequate soil volume depends on the supply of the other.

Properties and composition of the bacterial flora of a pine forest soil

Abstract: Summary A study was made of the qualitative and quantitative distribution of aerobic heterotrophic bacteria in the two mineral horizons of a developing podsol soil under Pinus. There was no significant difference in the total numbers or in the kinds of bacteria between the two horizons. In the acidic A1 horizon (pH, 3.6) Bacillus spp. comprised 58 per cent of the population, the remainder being composed of Gram-negative rods (15 per cent), pleomorphic rods (14 per cent), Gram-positive cocci (10 per cent) and streptomycetes (3 per cent). The corresponding figures for the C horizon (pH, 8.6) were Bacillus (23 per cent), Gram-negative rods (37 per cent), pleomorphic rods (10 per cent), Gram-positive cocci (17 per cent) and streptomycetes (13 per cent). These differences in distribution could not be related to any one environmental factor. The data from this investigation lend support to the view that different soil types can be distinguished by their bacterial floras. Bacillus subtilis, B. cereus and the Gram-positive cocci were biochemically and fermentatively active and many of the streptomycetes were able to degrade complex organic molecules. The Gram-negative and pleomorphic forms were apparently metabolically inactive.

A study on soil series: their correlation with the intensity and capacity properties of soil potassium

Abstract: Summary The Intensity (I) or availability of soil K exhibits considerable variations within a Soil Series, presumably as a result of local differences in management and fertilization. The contrast is most marked between woodland and cultivated examples of the same Series. Earlier work had shown similar variations in the Quantities (Q) of labile K. Those Capacity properties (dQ/dI) of soils for labile K which derive from the planar surfaces of the clay crystals in the mineral skeleton of the soil show rather less variability. The Capacity properties of different samples of soil from one Series are not equal, but become increasingly so as the non-clay fractions, the amphoteric oxides, and the organic matter held in the < 2μ fraction, are successively removed. It is inferred that the active surfaces available for K–(Ca + Mg) exchange in the soils studied are fundamentally similar for the soils in any one Series, but that they may be reversibly blocked by organic and amphoteric cations, to varying extents according to their previous management. The differences are greater than the differences commonly observed between different experimental plots in the same field. Duplicate samples of three Series out of the four examined showed close or moderate agreement in their K-fixation properties. While soil organic matter has some influence on dQ/dI, the fraction with most effect is not that which makes the most substantial contribution to the CEC of the soil.

Water movement in dry soils I Physical factors affecting sorption of water by dry soil

Abstract: Summary Semi-infinite columns of dry soil closed at one end had the other exposed to a turbulent atmosphere at constant relative humidity (0.98) and a range of constant temperatures. After varied times the water content of the columns was measured gravimetrically, in 1 cm layers, from which the total quantity taken up Q and the distribution of water content O with time t and distance x were found. Assuming that the Boltzmann transform, λ=xt−t can be applied to the standard diffusion equation, two soil parameters are derived. A test of the assumption is that O should be uniquely dependent on λ, and then the diffusivity is calculable from where Ot is the initial uniform water content. The second parameter—the sorptivity, S= Qt−½—and is not independent of Ot or the value maintained at x= 0. Results show that in a clay soil, between pF 5.8 and pF 4.2, water moves predominantly as vapour: in a non-swelling silicate mineral (sepiolite) there is significant liquid movement between pF 5.1 and pF 4.2. Long-continued use of organic manure has little effect on either D or S; aggregate size has some effect, in opposite senses for a sand and a clay; a mulch or a still atmosphere affects S but not D; evaporation suppressants decrease S; ignition decreases S and greatly alters D; and degradation of structure causes small changes in D.

A study of activities of enzymes hydrolysing sucrose and starch and of oxygen uptake in a sequence of soils under tussock grassland

Abstract: Summary Biochemical activities were studied at three sampling dates in a sequence of four soils, ranging in altitude from 275 to 1,520 m, on schist parent material in a tussock-grassland association on the Old Man Range, Central Otago. Activities were similar under different tussock species. The uppermost part of the soil appeared to be most affected by seasonal changes because fluctuations at different sampling dates were greater in samples of topsoil from depth 1 to 7.5 cm than in samples of A, AB, and B horizons at depths approximately 4, 13, and 28 cm. Activities of enzymes without added substrate and of enzymes hydrolysing sucrose or starch generally increased with increasing altitude of the soils and were strongly correlated with organic C content and, to a lesser extent, with soil moisture content. In previous pasture surveys, enzymes hydrolysing sucrose had not been significantly correlated with organic C. Although oxygen uptakes were also positively correlated with these factors, and with numbers of viable bacteria, consistent differences between soils were not found. Differences between A and AB horizons were usually greatest in the soil at the highest altitude but, otherwise, rates of regression of biochemical activities with depth of soil were generally similar in all soils. Results may be indicative of biochemical activities in other mountain-soil sequences.

Calcium: aluminium exchange equilibria in clay minerals and acid soils1

Abstract: Summary Exchange reactions between 0.0in AlCl3 solutions of different pH and Ca-saturated montmorillonite, vermiculite, illite, and soils from the Park Grass Experiment at Rothamsted and the Deerpark Experiment, Wexford, Ireland, showed that Al3+ and Al(OH)2+ were adsorbed from solutions of pH > 4.0 and Al3+ and H+ from solutions of pH < 3.0. When Al was adsorbed, the cation exchange capacity of Ca-saturated soils and clays increased. Conventional Ca: Al exchange isotherms showed that Al3+ was strongly preferred to Ca2+ on all soils and clays. The equilibrium constant for Ca: Al exchange, K, was identical for soils before and after oxidizing their organic matter and did not vary, for any exchanger, with Al-saturation or the initial pH of the AlCl3 solution. This proved the validity of the procedure used for calculating exchangeable Al3+. K values for Ca:Al exchange favoured Al3+ in the order: vermiculite > Park Grass soil > Deerpark soil > illite > montmorillonite. The influence of surface-charge densities of the clay minerals on this order is discussed and a method proposed and tested for calculating the K value of a soil from its mineralogical composition.

Genesis, mineralogy, and related properties of west indian soils

Abstract: Summary Soils of the Maracas series from two sites in the Northern Range of Trinidad differ only in minor aspects. With the exception of cheluviation of iron and formation of goethite the soils preserve the mineralogy and to some extent the structure of the parent phyllite. Weathering of calcic and sodic plagioclases is rapid. Mica is well preserved even in surface horizons and resists physical weathering, but when reduced to clay-size, it weathers rapidly to kaolinite throughout the solum with little accumulation of expandable minerals. Exchangeable K is very low in spite of the micaceous nature of the soil parent material. Low levels of availability of Ca, Mg, and particularly P are also characteristic of these soils. Total inorganic P is very low, the main fraction being reductant-soluble iron phosphate; other forms of inorganic P are unimportant. Organic P, derived from the carbonaceous matter of the parent rock, increases with depth. From the results obtained, the Maracas series may be classified as a latosol. For these soils located in a humid tropical environment, the paradoxical situation exists that they do not undergo sufficient weathering to release and make available plant nutrients before the solum is carried downslope by erosional processes. As a result, they are relatively infertile.

Rate‐controlling processes in the release of soil phosphate

Abstract: Summary The isotopic exchange of phosphate ions in a deep river gravel soil, maintained at different pHs from 35 to 67 in a field experiment, established by the NAAS at Shardlow, was governed initially (<20 h) by simultaneous ‘first-order’ and ‘bulk diffusion’ kinetics, but later by ‘bulk-diffusion’ kinetics only The ‘bulk-’ or ‘intracrumb-’ diffusion coefficient was independent of phosphate manuring, was least at pH 55 and increased on either side of this critical pH The rate constant for the phosphate component exchanging initially with ‘first-order’ kinetics increased significantly with phosphate manuring at pH 35, 44, and 55, although the constants for unmanured and manured soils did not change with soil pH But at pH 67, the rate constants for the unmanured and phosphate-treated soils were greater than those in the more acid soils, although the constant for the manured soil was just significantly greater than that of the unmanured soil Residual phosphate (ie the difference between manured and unmanured soils) adsorbed on the soil was greatest in the more acid soils but water-soluble residual phosphate increased with soil pH to a maximum at pH 55 Residual phosphate, exchanging initially, decreased to nothing above pH 65

The genesis of some basaltic soils in new south wales

Abstract: Summary Field and laboratory investigations of 182 soil samples and their underlying basalts in the Inverell area indicated that not all of the red soils in the area are relicts of Tertiary laterization as was previously thought. The soils on basalt were divided morphologically and mineralogically into shallow (black, brown/ chocolate, and red) and deep (krasnozems and lateritic soils) profiles. The mineralogy of the sand fractions of the shallow soils, which were similar in all respects except colour and location, suggests that they are young soils as the heavy minerals in the sand fractions are dominantly pyroxene, olivine, and the black opaque minerals which weather quite rapidly. The shallow black and brown/chocolate soils occupy whole slopes under grassland vegetation or grassland with trees, while the shallow red soils are restricted to steeper well-drained slopes and crests where trees tended to dominate grass. In similar areas throughout the New England region red soils are found on basalt in areas receiving 75 cm or less average annual rainfall, although previous workers have assumed that red soils develop from basalt at present only in areas receiving more than 140 cm average annual rainfall. The deeper soils, which were characterized by a lower pH, occupied upper or upper and middle slopes. The dominance of the resistant altered opaque minerals in these soils suggests that they are relict soils. The krasnozems showed little horizon differentiation apart from a slight darkening of the upper 8–10 cm. The lateritic soils had marked horizons, with laterite overlying mottled and pallid zones, the whole sequence being deeper than the krasnozems. The lateritic soils contained a high percentage of heavy minerals in the sand fractions in contrast to all the shallow profiles. Only one period of laterization is indicated in the area, although the laterite levels were not always accordant.

The effects of drying and ageing tree leaves on the ability of their aqueous extracts to dissolve ferric oxide

Abstract: Summary The effect was studied of drying and ageing leaves on the ability of their water extracts to dissolve hydrous ferric oxide. Extracts of freshly collected picked and autumn-fallen oak, beech, and larch leaves were compared with those of leaves from the same collections that were either freeze-dried or dried at 105°C. The iron-dissolving properties of the fallen leaves were not significantly affected by either method of drying, but drying at 105°C greatly increased the activities of picked larch and beech leaves. Freeze-drying had relatively little effect. Larch needles lost their activity very quickly when aged aerobically for 8 months, while the activity of beech leaves decreased slowly. The activity of picked and fallen oak leaves increased dramatically during the first 1–2 weeks, and then decreased, rapidly at first; after 8 months the activity was still comparable with the initial value. There was no direct relationship between the activities of the extracts and the pH or the concentrations of polyphenols.

A simple available soil nitrogen index

Abstract: Summary Mineralizable-N (Δ mineral-N) was correlated more with total-N than with organic-C in topsoils from maize fertilizer trial sites. In the absence of N fertilizers, Δ mineral-N was correlated with maize yield and per cent N in the leaf, in both seasons. Tests showed that a single regression equation calculated from data for two seasons, could be derived to express both these relationships. Maize yield on plots treated with fertilizer-N was not correlated with mineralizable-N in the first season and was correlated only at the lower rate of application (45 kg N/ha) in the second season. But maize-yield response to both rates of fertilizer-N was correlated with mineralizable-N in both seasons. Tests on the regression equations showed that yield responses were subject to seasonal variation. The Δ mineral-N value determined on conditioned soil samples (see Part 1) provided a good index of available N. It may be used to predict maize yield without N fertilizer and to rank soils, in terms of likely response to fertilizer-N, for advisory purposes.

USE OF RADIOACTIVE 59Fe FOR TRACING SOIL PARTICLE MOVEMENT

Abstract: Summary The field studies reported in Part 1 were complemented by laboratory investigations on the labelling of soils and the movement of soil particles by simulated rainfall splash. Of various labelling techniques tested the most uniform absorption of 59Fe, as shown by autoradiographs, was secured by adding DTPA chelating reagent to the FeCl3 solution used for wetting soil peds under vacuum. Studies on the specific radioactivity of the Atterberg textural fractions, separated from clayloam, sandy loam, loam and peat soils before and after labelling, showed in all cases that the clay fraction was dominant in retention of 59Fe. The radioactivity was so firmly absorbed that its movement could be used to trace both the horizontal and vertical displacements of labelled soil particles by scintillation counting or autoradiographic techniques. Tests with a rainfall simulator showed a close correlation between the weights and radioactivities of splashed soil particles. Gravimetric data obtained from unlabelled clay loam and sandy loam soils subjected to ‘rainfall’ of differing intensity, drop size and height of fall gave a linear relationship on a logarithmic scale between mass of particles and distance of splash. Calculation of the kinetic energy lost by water drops on impact with the soil and that imparted to the splashed soil particles show that the relationship depends principally on the rainfall intensity and the texture of the soil.