Chapter 4 – Online Appendices

Appendix 4.1 Mull humus topsoils and effects of burial at the Overton Down Experimental Earthwork, Wiltshire, UK

As noted earlier, mull formation (a L or litter or an Oi organic horizon; Table 1.4) reflects high biological activity and humification ((Brady and Weil, 2008)499-513; (Duchaufour, 1982)37-38). At Overton Down, Wiltshire, UK these conditions occur in a rendzina (rendoll with mollic epipedon; formed under grassland on chalk (Table 1.4). Such soils are the result of long-term use of the downland landscape as pastures, which are recorded through stratified landsnail sequences found in prehistoric ditch fills (e.g.(Carter, 1990; Evans, 1972). At Overton Down, the 1992 Control Profile is situated outside the fence-enclosed experimental ditch and bank (Bell et al., 1996), and is well studied using bulk analyses and soil micromorphology; the latter included painstaking counts of 0.5 mm squares across 60x70 mm size thin sections – see section 3.5. Details of the soil are essentially quoted from (Crowther, 1996b; Crowther et al., 1996; Macphail and Cruise, 1996), with some additions and unpublished material from Macphail.

The grass-covered control soil comprises a 1.5-2.0 cm thick root mat, an Ah mineral horizon, and a very stony AhC horizon. The Ah horizon averages 18 cm in thickness and has a high silt content (maximum, 53.0%), which is typical of the Icknield soil series and is partly derived from loess deposition during the late Devensian (Avery, 1990). The Ah horizon is rich in organic C and N, with mean concentrations of 11.0% (equivalent to c. 18.9% organic matter) 1.16%, respectively. These data are typical of rendzinas under permanent grassland. The low C:N range of 8.07-10.7 is indicative of a well-humified organic fraction. In thin section, the soil displays a total excremental fabric (Fedoroff, 1982), with roots concentrated in the top 2 cm. Here the soil has a very open compound packing porosity around crumb structures (porous or 'spongy' soil aggregates produced by earthworms) and other finer organo-mineral excrements, and an average void space of 55%. Below the rooted layer at between 2 and 8 cm more coarse crumbs and fine subangular blocky structures become more common as a result of coalescence through wetting and drying, and the mean void space is reduced to 32%. As is typical of mull horizons (Babel, 1975) and other grazed rendzinas with high base status brown earths (e.g., Martin Down, Hampshire (Carter, 1987); modern Maiden Castle, Dorset (Macphail, 1991) the soil is dominantly earthworm-worked. The micro- and mesosize excrements noted above are present throughout the thin section studied, and many of these are probably derived from enchytraeid activity, in particular through working of earthworm excrements. While this is a typical feature of mull humus formation (Babel, 1975; Bal, 1982), the actual biomass of enchytraeids is likely to be small compared with earthworms (Wallwork, 1983)."

This control profile is useful in order to evaluate the effects of 1) fencing that excluded animals from the site for 32 years, and 2) burial for the same period by 2a: rendzina turf forming the turf core of the bank and 2b: chalk dug up to create a ditch which forms the upper layers of the bank.

Soil within enclosure

It is interesting to compare this control profile with the modern profile within the enclosure, an area from which grazing stock had been excluded since 1960 to protect the ditch and bank from animal disturbance. The chief difference between this profile (Profile 1) and the control profile outside the enclosure (Profile 2) is the presence of a 50 mm-thick grass litter mat and evidence of associated development of a less well-drained environment. The soil was sampled from the base of this 50 mm mat. Typical soil diatoms (Hantzschia amphioxys and Pinnularia borealis; Nigel Cameron, UCL, pers. comm.) are present in this litter mat ('Laminated Mull Humus' (Barrat, 1964)), and dark brown coloured, slowly decomposing grass stems are releasing numerous phytoliths. The relatively high C:N ratio of 13.3 recorded immediately below the grass litter mat reflects the incorporation of poorly decomposed litter in the uppermost part of the Ah horizon, which generally has a C:N range of 9.72-10.3. Another chemical indicator of poorer drainage is the relatively high concentration of iron complexed with organic matter (0.153%), which is an indication of slow organic decomposition under poorly aerated conditions (Crowther et al., 1996). These results are consistent with those of the soil micromorphology.

Below the litter mat, the soil exhibits a 60 mm-thick unstructured layer that contains few separate peds (soil structures). Such soil horizons are termed 'massive'. They have a homogeneous appearance, do not break apart in any preferred direction and can be best described from their porosity pattern. The soil inside the enclosure shows both a decrease in porosity and change in type of pore space compared to that of the control soil. For example, the porosity is a mean 40% compared to 50-70% porosity in the control soil. The very open compound packing pores of the well-structured control soil have been replaced by fine vughs (spherical to elongate irregular voids) and channels giving rise to a spongy microfabric. In addition, this massive soil contains abundant micro- and mesosize iron and manganese nodules, whereas these features are quite rare outside the enclosure. Occasionally, plant remains have been (pseudomorphically) replaced by iron and manganese. Lower in the soil, the structure is coarse subangular blocky, formed of aggregated earthworm excrements down to 120 mm below the litter mat, and again this typical of the lower part of mull horizons.

Fencing and the exclusion of grazing animals has led to an accumulation of grass litter, which holds water and results in atypically moist topsoil conditions and a sharp decline in biological activity, especially by earthworms; the latter is recorded in the type and number of excrements counted. This resulting poorer surface drainage and possible localised waterlogged conditions (Bloomfield, 1951) have allowed the mobilisation of iron and manganese at numerous anaerobic micro-sites (Wiltshire pers. comm.; (Kelly and Wiltshire, 1996). Evans (Evans, 1990) has discussed the long-term effects of 'Festuca mats' on rendzinas, such as those leading to their decalcification, decrease in biological activity - especially from earthworms - and general development of poorly structured soils.

Effects of burial on rendzinas

Soil investigations were complemented by environmental studies, which included landsnails (these were essential absent because the rendzina is decalcified), 'pollen' (Lycopodium spore movement), microbiology, the migration of phosphate and buried material, including burned, cooked and uncooked bone; although bones remained their long-term survival is doubtful (Bell et al., 1996). Lycopodium spores, which had been scattered across the soil's surface in 1960, were mixed 90 mm upwards into the chalk bank (Crabtree, 1996) (see 3b). (Carter, 1990) also found that the original stratigraphic position of landsnails can also be affected by burial. Phosphate studies revealed that even in a chalk downland environment - generally considered favourable for bone preservation - it was possible to detect phosphate migration from weathering bone in both the buried soil and chalk bank (Crowther, 1996a); this observation further demonstrates bone deterioration at the experimental site.

Buried soils below the turf stack and chalk bank

Under the turf-stack: Measurements of the thickness of the buried Ah horizon beneath the turf stack showed that its thickness was reduced by 50% in places, with a loss inorganic C (from an average of 11.0% to 7.59%), and a decline in pH from a mean of 6.9 to a mean of pH 5.6 (pH 5.4 immediately under the turf stack)(see pH below chalk bank for comparison). It is therefore clear that leaching and decalcification have continued under the turf stack. This observation is surprising as it was expected that the chalk bank overlying the turf stack would have contributed more overtly to alkaline seepage; the decalcified turf stack soils have therefore continued to neutralise drainage water here. The increased acidity recorded in the buried soil is attributable to three factors: higher carbonic acid concentrations (from biogenic carbon dioxide) as the result of impeded aeration; an increase in organic acid secretion by microbial activity when anoxic conditions led to fermentation (Wiltshire, pers comm.); and reduction in the amount of fine chalk being incorporated into the Ah horizon from the underlying Ah/C horizon and bedrock as earthworm mixing declined.

In thin section the reduction in thickness can be partly accounted for by a change in the amount and number of voids, and loss of organic matter (as also recorded chemically). The original open-structured soil is replaced in the uppermost 7cm by a massive or poorly formed fine prismatic structure, with burrowed open vughs, loosely infilled with micro- and mesosized excrements (spongy microfabric). A few mammilated excrements are also present. In all cases, these are organo-mineral excrements. On this evidence, it would seem that the earthworm population declined quite rapidly after burial, and any annelid worms present today are almost exclusively enchytraeids (Mücher, 1997)(Mücher, pers. comm. to RIM) because of their wider tolerance of a lowered pH and sporadic anaerobic conditions. Quantitative analysis of pore space showed a marked reduction. In one thin section, void space is as little as 14% at 1 cm depth, compared to 55% at the same depth in the control soil. Also in comparison to the control soil, this buried soil contains very few in situ roots, and these seem to occur as 'ghost' features comprising pale yellow parenchymatous and/or brownish epidermal remains, or as fine fragments within the mineral soil. Again, this is consistent with soil chemistry at the experiment.

Microsize Fe and Mn nodules are ubiquitous, and concentrations of occasional to many mesosize nodules occur at 3 and 4 cm below the old ground surface. In addition, there are a few >500 µm-size nodules, some of these being pseudomorphic of plant material. In view of the high porosity of the chalk bank and the marked hydrological discontinuity between the turf bank and buried soil, the buried soil has inevitably been subjected to waterlogging. This concentration of nodules reflects mobilisation of Fe and Mn under anaerobic conditions, and is clearly an indicator that under more waterlogged conditions an iron (and manganese) pan could form. Such ironpans are common along and below buried old ground surfaces, and have been recorded from numerous sites (Evans, 1972; Limbrey, 1975; Macphail, 1987). Overton Down demonstrates how rapidly such features can potentially develop. When individual turfs in a turf stack and old ground surfaces are picked out by strongly-formed ironpans at more acid and wetter sites, it is important to differentiate between iron mobilisation through waterlogging (hydromorphic iron) and panning influenced by podzolisation ((Runia, 1988), see below).

Under the chalk bank: The nature of the buried soil under the chalk bank is essentially similar to that under the turf stack, but with important differences. In marked contrast to the turf-buried soil, the pH of the bank-buried soil has increased from pH 6.9 to a slightly alkaline pH 7.9. This change is understandable from the soil micromorphology, which records downward mixing of chalk and chalky soil from the chalk bank into the once-decalcified buried soil. Calcite biospheroids (earthworm granules; (Armour-Chelu and Andrews, 1994; Canti, 1998) are also present in both the chalk bank and buried soil, as evidence of their burrowing. However, the presence of secondary micrite in voids may suggest that earthworm working has mainly ceased, although earthworms were encountered during the excavation of the bank.

Appendix 4.2 Moder and Mor humus topsoils and effects of burial at the Wareham Experimental Earthwork

The podzols at the Wareham Experimental Earthwork, Hampshire have formed in infertile Eocene sands under a lowland heath (Ericaceous) vegetation, and surficial disturbances have removed the original long-formed mor humus (L, F, H horizons); currently the soils have a relatively recently developed, dominantly moder (LF[H]) organic horizon (Evans and Limbrey, 1974; Hillson, 1996; Macphail et al., 2003). Details of the results are broadly quoted from Macphail et al (2003) and are consistent with other studies of other 'lowland' podzols from mainland Europe, for example (De Coninck, 1980; De Coninck and Righi, 1983).

Three control profiles were studied employing soil micromorphology (including both optical quantitative analysis and image analysis) and chemistry; a fourth profile was also studied by chemistry alone.

Soil micromorphology: At the three control profiles this Moder-Mor superficial humus horizon is composed of 40-70 mm of litter (L - Oi) and fermentation (F - Oe) layers, with in some cases, a thin discontinuous humus (H – Oa) horizon. "These and all topsoils included in this study were classified according to humus form after Babel (Babel, 1975)(pp 446): L v_: macromorphological alteration of litter and laminated fabric; _F r: plant residues predominate over fine substances and breakdown of laminated fabric; F m_: equal proportions of plant residues and fine substances, finely fragmented plant residues and highly penetrated by roots; _H r and H f_: few and no plant residues, respectively, and predominance of fine substances, with enrichment of plant substances resistant to decomposition; _A h: humus-rich A horizon." "In thin section, the LF(H) can be seen to have an overall blocky structure that is strongly characterized by broadly laminated coarse plant residues (max. 20·1%) including in situ ericaceous roots, and increasing proportions of dark reddish brown organic, excrement-dominated fine organic matter down-profile (humus types L v_ and _F r /F m). This excrement-rich microfabric reflects the activity of arthropods and Enchytraeids in such soils (cf(Van der Drift, 1964; Zachariae, 1964)Abb. 1). In contrast, the underlying medium sand-size quartz-rich Ah horizon contains many fewer plant fragments, and is massive with a microaggregate structure and a brownish black microfabric attributable to Enchytraeids reworking arthropod faeces and microbiological breakdown of humus (A__h, Babel, 1975, 446).

Image analysis and optical counting: Similar to the field and chemical results (below), image analysis (by Tim Acott) revealed wide variation between four samples of the control Ah horizons. They contain 19·1–30·0% voids, 2·0–4·0% organic fragments, 9·2–33·0% dark organic matter and 36·0–33·8% mineral grains. Frequencies of small (<1000 pixels), medium (1000–2000 pixels) and large (>2000 pixels) voids also vary. Some 54–116 (mean 81) large, 91–107 (mean 99) medium and 3547–6452 (mean 4715) small voids are present. Optical counting also detected variability, but plant residues and organic excrements are consistently most frequent in the LF(H) horizon. Organo-mineral excrements are more common in the mineralogenic Ah horizon. Passage features (biogalleries), especially large (>500 µm) ones, are more common in the LF(H) than in the Ah horizon.

Chemistry: Averaged over the four profiles, John Crowther found that the L layer has a higher organic C concentration (38.8% organic C) and a higher C:N ratio (40.8) than the F layer (29.9% organic C; C:N=33.4). It is also less acidic, with a mean pH 4·0. The Ah horizon has a mean organic C concentration of 6·76%.

Effects of burial on podzols at Wareham

The buried LF(H)

Measurements of the probable diminution in thickness of the LF(H) horizon between 1963 and 1980 (apparently from 40–70 mm to 1·6–3·6 mm), and from 1980–1996 (turf buried: 1–4 mm, sand buried: 1–3 mm), show that the bLF(H) exhibits little further evidence of thinning since 1980 (Evans and Limbrey, 1974; Macphail, 1996)(This layer was only described in the field in 1972.)

Soil micromorphology: The two post-1972 monitoring excavations at Wareham in 1980 and 1996 revealed that individual laminae within the laminated fabric of the LF(H) of the control profiles had become equally reduced in thickness, from 5–10 mm to 0·4–1·2 mm. The greatest decrease is found in the 1996 sand-buried LFH horizon (0·4 mm). Although organic soil compaction is one factor, it is only a minor one. More importantly, the bLF(H) has been transformed from a Lv and Fr/Fm humus form, to a Fm/Hr form, in which the laminated fabric is more fragmented and where fine organic matter, including excrements, predominates over plant residues. This transformation from an open laminated microfabric characterized by simple and compound packing void patterns, to a microaggregated microstructure with compound and complex packing voids and developing vughy porosity, also contributes to the compacted nature of this horizon.

In the 1980 bLF(H) an open dark reddish brown excremental fabric is still visible, but by 1996 a dense, blackish brown amorphous organic matter had become dominant; this is ascribed to continued decomposition, for example by bacteria and fungi. This transformation, and attendant diminution of pore space, resulted in additional compaction. As organic matter decomposition has continued, relative amounts of plant substances resistant to breakdown (e.g. bark) have increased, and charcoal has become markedly concentrated. At Wareham, in contrast to the Overton Down experiment, wood charcoal was not deliberately scattered on the old ground surface and so this charcoal concentration in the bLF(H) is purely the result of natural differential preservation of organic matter – charred material being highly resistant to breakdown.

Burrow-mixing of organic soil into the overlying bank and into the bAh horizon below, is another causative factor influencing overall compaction. This is well illustrated in sand-buried sample 726. Here, some rare <500 µm-size loose organic excrement infilled passage feature can be followed 4 mm upwards into sand-layer C, but more significantly coarse (>500 µm) and fine (<500 µm) passage features and excrements penetrated from to 10 mm and 20 mm below the bLF(H), respectively. This small amount of soil mixing is consistent with Professor Dimbleby's pollen studies of Lycopodium spores (1964, 1965, 1968 and 1972) that had been deliberately sprinkled across the site in 1963. He found that these marker spores were concentrated approximately 25 mm (1 inch)-thick zone either side of the old ground surface.

Microprobe studies: At Wareham, some Fe (and Al) replacement/coating of organic matter in the ''ferruginized'' bLF(H) is given both by microprobe data (mean 0·28% Fe, 0·66% Al in sand-buried sample 781, n=13) and their reddish colours observed under OIL. Fe and Al are both found in the immediately underlying bAh horizon (mean 0·015% Fe, 0·59% Al, n=20) and in the immediately overlying sand bank (mean 0·025% Fe, 0·24% Al, n=10). Unweathered sand in the bank and continuing podzolization (see below) are probably two causal factors in this enrichment process.

Sesquioxide ratios (Fe2O3/Al2O3) for Layer D sand, the ''ferruginized'' bLF(H) and bAh are 0·06, 0·32 and 0·19, respectively. Similarly, except for the innermost cells, ''ferruginized'' root material and associated ''ferruginized'' Oribatib mite excrements in the turf bank have sesquioxide ratios of 0·27–1·13 (sample 429). Such low ratios for ''ferruginized'' organic material are more consistent with podzolic (spodic) B horizon formation (0·15: polymorphic Bh horizon; 0·40–2·50: monomorphic Bh horizon) than with hydromorphic ironpan (placic) development (ratios of 5–49), where much larger proportions of Fe are encountered (De Coninck, 1980; Righi et al., 1982). On the other hand, very localized and minor Fe deposition induced by oxidation-reduction cannot be ruled out, as exemplified by a ''ferruginized'' bLF(H) (sample 726: mean 3·49% Fe, 4·23% Al, n=16) below a pottery fragment (mean 3·54% Fe, 4·74% Al, n=7) (Bouma et al., 1990; Breuning-Madsen et al., 2001) and see below). The bLF(H)/old land surface is likely acting as a barrier to soil water movement, and allows little ''ferruginization'' of the immediately underlying bAh (mean 0·07% Fe, 0·52% Al, n=7). These observations reveal the complex nature of organic matter preservation in the buried soils, under essentially acid, podzolic and aerobic conditions, where 'good' preservation of the bLFH in terms of an identifiable microfabric, appears at least in part due to mineral replacement/impregnation/coating. Findings from the 1980 excavation indicate that this relict bLF(H) horizon was already undergoing ''ferruginization''.

1b: Changes to the buried Ah horizon

It can be suggested that there has been a reduction in organic content in the buried soil, but because the control soils show such a wide variation in organic content this statement has to be regarded with great caution (Macphail et al., 2003). The marked reduction of roots may reflect their being an important source of easily decomposable organic matter in the buried soil. Roots, with leaf litter, made up an important (12–20%) part of the, now transformed, bLFH horizon (see below). In contrast, in the bAh, amorphous organic matter, which is apparently little changed in appearance (brownish black under PPL) has probably undergone much less reduction. This observation reflects the high proportion of very dark amorphous organic matter in podzolic topsoils that is already ''old'' and which contributes to older organic matter being present in podzol Bh horizons (Guillet, 1982)pp 243).

Image analysis and optical counts: We can note that pore space is especially diminished in the sand-buried soil compared to the control soils (from 24·5% to 17·9%), and this is matched by a marked decrease in the number of large voids, for example. This modification occurs alongside the intergrain microaggregate (enaulic) microfabric of the control Ah horizons developing a more coated (chitonic) microfabric. In addition, possibly because the turf stack holds more soil water, the original microfabric has become strongly transformed to coated (chitonic) and bridged (gefuric) related distributions. In addition to microfabric types which could not be accurately quantified through image analysis, it was decided to area-count excrement types optically. Nine types were differentiated, five organic and four organo-mineral. Different size passage features were also counted.

Firstly, insignificant mixing into sand layers C and D, with downward mixing only extending some 20 mm down-profile was recorded (see Crabtree 1996 for Lycopodium spore analysis from previous years). This is quite different from the noticeable earthworm mixing of spores at Overton Down where these spores were mixed 80-90 mm either side of the old ground surface). The lack of mixing at Wareham results from a probably smaller biomass of arthropods and Enchytraeids; in contrast major burrowing was by earthworms (Lumbricus terrestris) at Overton Down. In addition, total and type of organic excrements diminish, probably because of decreasing access to fresh organic matter such as roots and litter. This change is matched by a relative increase in organo-mineral excrements.

Chemistry: In general it was found that podzolisation and leaching have continued, as measured by a decrease in available potassium, and as evidenced by increases in alkali soluble organic matter and pyrophosphate extractable C in the buried soil. Such continuing podzolisation in and under barrows has been reported previously (Fisher and Macphail, 1985; Runia, 1988). Microprobe analysis of a 'ferruginised' root and an example of an Oribatid mite organic excrement it contains, also identified the effects of this continuing podzolisation. Moisture retention in the turf-stack also indicates a potential for gleying even whilst conditions of burial at Wareham were identified as essentially aerobic (Lawson et al., 2000). Finally, a slight rise in pH was recorded in the sand-buried soil, presumably because it was influenced by an overburden of relatively unweathered subsoil (with a pH as high as pH 5.6) containing rare calcareous sand.

Appendix 4.3

Neolithic Carn Brea, Cornwall; outer rampart-buried soil – chemistry:

• Leached bEa horizon (0-60mm): 2.1% org. C; 0.77% dith. ext. Fe; 0.32% pyro. ext. Fe; 9.4 me e/100 gms CEC) over,
• Illuvial bBhs (60-130 mm) (5.3% org. C; 4.11% dith. ext. Fe; 1.02% pyro. ext. Fe; 34.2 me e/100 gms CEC).