Abstract

Key Words

Phytolith, podzol, Podosol, bioturbation, pervection, Phytolith Depth Functions

Introduction

Phytoliths are microscopic opaline silica particles that develop in plants and are delivered to the surface of the soil by way of the litter layer. They enter the soil in a variety of ways but chiefly via (i) particles in faecal material that is deposited within the bioturbated zone, (ii) transportation by water movement through soil pores and cracks (pervection), (iii) falling (vertical tumbling) into larger cracks, and (iv) slow or rapid burial. Over time they are also subjected to dissolution leading to a reduction in size.

In previous work, Hart and Humphreys (2003) considered the distribution patterns of phytolith concentration within the soil and identified three main Phytolith Depth Functions (PDFs) which were related to soil characteristics and mixing processes. In particular, the PDF is related to soil texture, which, in turn, influences the efficacy of bioturbation, pervection and vertical tumbling. Interpreting these processes, however, requires an investigation based on phytolith morphology and their weathering characteristics and an understanding of faunal interactions within the soil. All of these are little understood at present and it is considered that an understanding of these relationships is essential to the interpretation of PDFs.

Of particular interest is the Type-2 PDF, the key feature of which is the spike of greater phytolith concentration at depth. This pattern has often been used as evidence of a buried soil, especially in loessic materials or where burial by ash or alluvium is faster than revegetation. For this to occur it requires that phytoliths remain immobile after deposition and burial and as a consequence it was referred to as the “static hypothesis” of Hart and Humphreys (1997). This interpretation is often implied in archaeological and paleoenvironmental research and sometimes considerable effort is expended in denying the possibility of post-deposition mobilization. The alternative explanation regards phytoliths as potentially mobile particles that can be moved by bioturbation, pervection etc. until a barrier is reached, where they will accumulate, thus giving rise to the zone of abundance or spike (the “mobile hypothesis” of Hart and Humphreys 1997).

Such a situation was recorded in a podzol at Oxford Falls, New South Wales, Australia, where phytolith content decreased through the A₁ and A₂ horizons until just above the pans (Bhs) where it increased (Hart and Humphreys 1997, Simons et al. 2000). A cluster analysis showed that high phytolith morphological diversity was encountered at the pans which implied that pervection was likely to be a major mixing mechanism. In addition the weathering patterns of phytolith morphologies selected for high pervection potential, such as spherical phytoliths which were found to be less weathered at depth in comparison with other morphologies, confirmed this interpretation (Simons et al. 2000).

The aim of the present study was to extend this examination by focusing on changes in phytolith size down the podzol profile. If pervection is important it is expected that size trends should be largely controlled by the size of the interconnecting voids which in turn must relate to packing density. Hence, the maximum phytolith size is expected to occur above the layer with the smallest interconnecting voids. Conversely, if phytolith distribution was due to bioturbation no such trend would be expected to exist providing that the phytolith bearing faecal pellets are larger than the interconnecting voids.

Methods

Study site

The field site is located near Narrabeen about 18 km north of Sydney, on Middle Creek. The examined pit (Figure 1) dislays the strongest development of a Podzol (WRB 1998) at the site (thickest and most conspicuously bleached A₂ horizon over a distinct double pan consisting of organic (Bh) and iron (Bhs) horizons (Field and Humphreys 2002). The soil classifies as a Semiaquic Podosol (Isbell 1996), Spodosol (Soil Survey Staff 1998), or Podzol (Stace et al. 1968). In Northcote’s (1979) classification it best fits a Uc2.36. Vegetation on the podzol is a woodland comprising Eucalyptus gummifera, E. piperita and Angophora costata, with an understory of Banksia serrata, Ceratopetalum gummiferum, Xylomelum pyriforme, Xanthorrhoea arborea, Ricinocarpos pinifolius and Gompholobium latifolium (Buchanan and Humphreys 1980).

Figure 1. The podzol at Narrabeen.The positions of the pans are noted.

Scale is 1 µm. P axis (see text) is 20 µm.

Figure 2. Scanning Electron Micrograph of a spherical phytolith from the podzol organic pan.

Sampling

Soil samples were obtained at depths of 5, 34 (A₁); 54, 66, 78, 87 (A₂); 94 (Bh); 99 (Bhs); 111 and 164 (C) cm for phytolith analysis as adopted by Simons et al. (2000). Bulk density data was obtained from a previous study (Field and Humphreys 2002) using metal rings with a volume exceeding 100 cm³. In this study bulk density is used as a proxy for the interconnecting void size: the greater the density the smaller the void size. Spherical phytoliths were selected for the study for ease of determining size and because they tend to be robust and hence are less affected by dissolution. Sphericals are part of the Spheroidal Class, a major shape category of solid plant opal silica bodies ranging from spherical (surface at all points equidistant from the centre) through slightly flattened bodies to ovoid (egg-shaped) (Hart et al. 2003). The spherical is a single, symmetrical phytolith where the Polar (P) and Equatorial (E) axes are equal in length (Figure 2).

Phytolith extraction, counting and analysis

Phytoliths in the 2–63 µm range were extracted using the method outlined in Hart (1992) and Simons et al. (2000) with sodium polytungstate as the heavy liquid. They were then mounted on glass slides using Eukitt as the mounting medium. In this study one slide for each soil depth sampled in the podzol was examined under a Leitz Petrological (polarising) binocular microscope at a magnification of 400×. A total of fifty single sphericals were measured to the nearest 2.5 µm in most samples and their P axis (Polar, longer axis which in this case was equal to the E or Equatorial axis) measured (Figure 2). Smaller sample yields were obtained in two layers (Bhs and the deep C) due to the relative paucity of spheroids compared to other shape classes in these layers. Phytolith morphology is described using the terminology of Hart et al. (2003).

Results

The sampling and measurement strategy biased the size distribution. Phytoliths in this soil profile are known to range in size up to about 50 µm. In only a few situations have phytoliths been recorded as being larger, but these were not spheroids (Hart 1992). The maximum phytolith size possible from our study is limited by the 63 µm sieve mesh. The fact that <2% of the spherical phytoliths (n=444) are larger than 30 µm, with a maximum sized of 42.5 µm, indicates that the upper sampling size was effectively obtained at this site (Figure 3). However, the strategy cannot sample clay sized (<2 µm) phytoliths as there is no effective means of separating them from clay minerals. This, together with a measuring resolution of about 2.5 µm, limited the smallest measured fraction to 5 µm. The distribution is skewed to the smaller size with half of the sampled phytoliths between 5 and 10 µm diameter. Individual samples are mostly not normally distributed even after standard transformations. For this reason we report the data showing median values, means and standard deviations to provide some sense of the distributions (Table 1).

Figure 3. Size distribution of spherical phytoliths (n =444).

Depth functions of phytolith concentration, soil bulk density and median spherical phytolith diameter (Figures 4-6) do not always correlate. Phytolith concentration decreases to the base of the A₂ horizon, spikes at the Bhs horizon and declines again with depth. It is this spike in concentration that charactertises a Type-2 PDF. In contrast, the bulk density is greatest at the top of the A₂ horizon, least in the Bh pan and increases again to the deeper C horizon. These data are neither in nor out of phase. The depth trend of the median (and average) size of spherical phytoliths most closely matches the bulk density trend. In particular, the maximum median diameter occurs at the top of the A₂ horizon, as does the greatest bulk density, and the smallest median diameter occurs in the pans matching the smallest bulk density. Paradoxically, perhaps, the smallest median phytolith diameter coincides with the spike in phytolith concentration, which was not expected.

Table 1. Spherical diameter at each sampled layer in the podzol.

Depth (cm)	Horizon	Sample size n	Median diameter (µm)	Largest diameter (µm)	Mean (µm)	Standard deviation (µm)
5	A1	50	12.5	35.0	12.6	6.1
34	A1	50	11.3	37.5	12.4	7.1
54	A2	50	20.0	42.5	18.9	8.8
66	A2	50	15.0	35.0	15.4	7.7
78	A2	50	10.0	27.5	12.1	6.0
87	A2	50	12.5	27.5	12.0	5.5
94	Bh	50	7.5	25.0	10.3	5.5
99	Bhs	21	7.5	15.0	8.3	3.2
111	C	50	10.0	22.5	10.6	4.9
164	C	23	10.0	20.0	10.2	3.7

Figure 6. Median spherical diameter with depth in the podzol

Discussion

Very little research has been conducted into the range of sizes found in each phytolith morphological class such as spheroids. In this study the maximum recorded spherical diameter was 42.5 µm, but >75% are smaller than 15 µm. The increase in proportion of phytoliths with decreasing phytolith size in sediments highlights that a high proportion of phytoliths are likely to be <2 µm in size. It is often assumed that the clay-sized fraction represents about 50 % of the total phytolith mass, but this is possibly a gross underestimate.

It might be expected that the movement of large diameter sphericals is inhibited where soil bulk density is large (smaller interconnecting void size) and thus some accumulation of larger diameter sphericals might occur here. This appears to occur in the upper A₂ horizon where there is a strong correlation of large diameter sphericals with bulk density (r² = 0.568). After this point the spherical diameter drops, apart from a small rise immediately above the pans, to the Bhs (iron pan) and correlation with the small bulk density is poor. Smaller bulk density is often assumed to indicate greater void space, but only when the density of all materials is similar. An examination of thin sections of the pan material indicates a reduced void size even though there is a greater spacing between the framework quartz sand grains. It is the binding matrix material, whose composition is dominated by carbon, iron and aluminium (Field and Humphreys 2002), that fills much of the void space and clearly must have a much reduced density compared to the quartz sand grains. Thus, the apparent reduced void size, yet to be quantified, appears to provide an effective trap that accounts for the spike in concentration levels. If this is the case, all morphological classes should reveal a similar decrease in phytolith size in the pan horizons – but this has yet to be examined.

In terms of the focus of this study it appears that pervection plays a leading role in the re-distribution of spherical phytoliths in the soil. The increase in size to the most dense horizon (top of the A₂) and the decrease in size with depth to the pans, as well as the spike in concentration, may be interpreted as developing as phytoliths are trapped in the pan because of a decrease in interconnecting void space. If spherical phytoliths were largely distributed by soil animals, as faecal material by earthworms and some beetles or as crop by termites, a more uniform trend might be expected if the soil or topsoil was deeply and thoroughly mixed (a Type-3 PDF) or, as is more often the case, a depth function expected that displays an exponential decline with depth (a Type-1 PDF). Neither occurs. Similarly, if there was no downward movement of phytoliths there should be no particular trends with interconnecting void size, which is not the case.

In the C horizon the maximum spherical median diameter rises and exceeds those of the pans. That the Bhs layer does not permit phytoliths >15 µm to pass through it implies that these larger grains got there by another means. Previously, Field and Humphreys (2002) reported that the age of the C horizon material exceeded 20 ka whereas that of the A₁ and A₂ was up to 10 ka. Thus, the C horizon could be part of an older surface in which the larger phytoliths represent part of a Late Pleisticene soil. Alternatively, it could indicate movement and accumulation before podzolization commenced with the pans eventually blocking pervection of larger phytoliths at some stage during the Holocene.

Conclusion

As soil bulk density increases with depth, the reduction in soil pore diameter progressively blocks the movement of phytolith sphericals which have migrated from the surface leading to an accumulation of larger diameter sphericals in the A₂. This process continues down the profile with finer and finer phytolith material being blocked by narrowing soil pores. The seemingly anomalous lower pan bulk density concomitant with a rise in small diameter phytolith concentration is due to the reduced density of the matrix in the material comprising the pan. Larger spherical diameters in the C horizon may be remnants of an earlier deposit.

References

Buchanan RA, Humphreys GS (1980) The vegetation on two podzols on the Hornsby Plateau, Sydney. Proceedings Linnean Society of New South Wales 104, 50-71.

Field RJ (1997) ‘The Middle Creek podzol: An investigation of temporal, stratigraphic and mass-balance relationships’. BSc Hons thesis, Macquarie University, Australia.

Field RJ, Humphreys GS (2002) The development of a podzol using stratigraphic-age relationships and a mass-balance approach. In ‘Regolith and Landscapes in Eastern Australia,’ Ed IC Roach pp. 28-32. CRC LEME.

Hart DM (1992) ‘A field appraisal of the role of plant opal in the Australian environment’. PhD thesis, Macquarie University, Australia.

Hart DM, Humphreys GS (1997) The mobility of phytoliths in soils: Pedological considerations. In ‘The State-of-the-art of Phytoliths in Soils and Plants’. Eds A Pinilla, J Juan-Tresserras, MJ Machado pp. 93-100. Centro de Clencias Medioambientales Monograph: Madrid.

Hart DM, Humphreys GS (2003) Phytolith depth functions in surface regolith materials. In ‘Advances in regolith’. Ed IC Roach pp. 159-163. CRC LEME

Hart DM, Lentfer C, Wallis LA (2003) A universal phytolith key 5: Spheroidal class. In ‘Poster Abstracts from the Australian Archaeological Association Conference, Jindabyne’ #16. http://car.anu.edu.au/AAA_poster_session.html

Isbell RF (1996) ‘The Australian soil classification’. Australian Soil and Land Survey Handbook. CSIRO publishing: Melbourne.

Northcote KH (1979) ‘A Factual Key for the Recognition of Australian Soils’ 4^th Ed. Rellim Technical Publications: Adelaide.

Simons NA (1998) ‘Interpreting phytolith depth functions in relation to the processes of pervection and bioturbation’ BSc Hons thesis, Macquarie University, Australia.

Simons NA, Hart DM, Humphreys GS (2000) Phytolith depth functions in a podzol and a texture contrast soil, Sydney. In ‘Soil 2000: New horizons for a new century’. Eds JA Adams, A. Metherell 3, pp.193-194. Australian and New Zealand Second Joint Soils Conference.

Stace HCT, Hubble GD, Brewer R, Northcote KH, Sleeman JR, Mulcahy MJ, Hallsworth EC (1968) ‘A handbook of Australian soils’. Glenside South Australia: Rellim Technical Publications.

Soil Survey Staff (1998) ‘Keys to Soil taxonomy’. 8^th Edition Washington DC: United States Department of Agriculture.

WRB (1998) World Reference Base for Soil Resources, World Soil Resources Reports No. 84, ISSS/ISRIC/FAO, Rome.