The Proposed Northern Course of the Rother: a Sedimentological and Microfaunal Investigation
Introduction
Despite being one of the largest areas of reclaimed marshland in Britain, it is only in comparatively recent times that detailed examination of the relationships between the valleys draining eastwards into Romney Marsh and the marshland sequences themselves has been carried out. This has been based on the initial impetus given by the Soil Survey of England and Wales in the late 1950s, and developed particularly by Green (1968), whose consideration of the deposits (mainly down to 42 inches, or 1.06m), combined with historical documentation and cartographic evidence allowed him to develop a model of the evolution of the marshlands. This provoked a series of detailed investigations into the deposits on the western side of the marsh: in the Pannel valley (Woodcock 1984); the Pett Levels (Marlow 1984); the Brede and Lower Rother valleys and across Walland Marsh to the Rhee Wall (Waller et al. 1988), and the Rother valley (Burrin 1988). A study of the sediments in the Horsemarsh Sewer Valley between Kenardington and Warehorne, one kilometre north of the Thrift Cottage site of this study, was undertaken by Tooley and Switzur in 1988.
The underlying deposits in the Marsh itself were described by Green (1968), with the deepest Midley Sand overlain by the Blue Clay, and covered by extensive peat deposits dated in the Appledore Dowels at 3020 ± 94 B.P. (Callow et al. 1964). The post-peat deposits, referred to as the "Young Alluvium", consist of (1) an area of Decalcified (Old) Marshland where creek ridge inversion has occurred, and (2) three areas of Calciferous (New) Marshland, which include a large area in the north-east representing a silted up inlet near Hythe (Green 1968, Fig. 10), as shown in the inset to Figure 4.1.
It is the area of Decalcified Marshland between Appledore and Hamstreet which is the subject of this study (Fig. 4.1). Its significance lies in the suggestion that it was crossed by a possible northern course of the Rother (or Limen), or a distributary, during the Saxon period. This channel, now buried, is said to have flowed north of the Isle of Oxney towards Appledore, then to have followed the line of the present-day Speringbrook and Sedbrook Sewers across the north of the marsh towards Newchurch, and to have reached the sea near Hythe (Green and Askew 1959).
Historical Background
The possible "Northern Course" of the Rother (or Limen) was based largely on historical sources, and has long been a source of controversy. Its possible evolution has usually been linked with enclosure and reclamation of various parts of the marsh itself.
A number of Anglo-Saxon charters discussed by Holmes (1907) and Ward (1931, 1933a, 1933b) supposedly refer to the river Limen passing south of Kits Bridge (TR 017325) and Warehorne village, and a Saxon chronicle referred to by Lambarde (1576) also refers to an incursion by the Danish fleet into the mouth of the Limen near Newchurch. The idea of a river flowing almost due east from Appledore was further endorsed by Holloway (1849).
However, neither Elliott (1862, 1874) nor Topley (1875) found any topographic evidence of this river, and Elliott further argued that the meandering elements of this course represented relic sea creeks. Such a non-fluvial origin was also preferred by Steers (1964). Elliott favoured instead a generally southward course from Appledore then heading north-east towards New Romney, with which Lewis (1932) agreed and which was also favoured by Green (1988).
In 1959 the northern channel was identified and mapped by Green and Askew. Green (1968) noted its correspondence with an area of inverted creek ridges on the Sedbrook Sewer south of Ruckinge and a narrow channel near Ham Mill Green (TR 002316), continuing across to the Dowels where it follows a section of the Blackmans Arm of the Speringbrook Sewer as far as the Rhee Wall. There is, however, a degree of uncertainty in Green's view here, since he referred to "historical records of this creek or river course" (1968,36), yet he mapped creek exits emptying east ofNewchurch into an area where subsurface sediments contained a marine molluscan fauna (1968, Figs. 14 and 16).
Recently Cunliffe (1980) suggested that in Anglo Saxon times the Rother probably had two outlets, near Hythe and at Romney, and he considered that the Danish invasion occurred at the latter. However, Brooks (1988) identified an inlet, estuary, or arm of the sea in the Warehorne and Ham area from an 11th century manuscript, and argued that only an extensive tidal inlet would have allowed the fleet to penetrate as far as Appledore. He envisaged Walland Marsh as an area of tidal mudflats in the ninth century, with the presence of "a tidal channel or an artificial river course" representing the branch of the Limen flowing north of the Isle of Oxney persisting until the 11th century, but already under threat from the sea in the south. Eddison (1983), arguing on the basis of channel size, considered that this course represented a tidal creek feeding into the Hythe inlet. Thus, although on balance current opinion favoured tidal rather than fluvial conditions for the Appledore to Hamstreet channel, much uncertainty still remained. The object of this study was to resolve that dichotomy of opinion.
The sediments of the channel were first examined eastwards to Warehorne (Wass 1987). Although lacking the detailed lithostratigraphy of the studies of Woodcock (1984), Marlow (1984), Waller et al. (1988) and Burrin (1988), that work indicated that the channel dimensions were inconsistent with a river the size of the Rother, and that the particle size distribution was at variance with that to be expected from Burrin's work. Limited foraminiferal evidence indicated the increasing presence of high marsh towards the west, leading to the suggestion that the channel was more likely to have developed as a tidal creek rather than a river.
The present paper synthesizes material from the 1987 work with additional microfaunal evidence resulting from analysis of almost 7,000 ostracods and foraminifera. It aims to provide insight into thedimensions, lithostratigraphy and sedimentary structures of the buried channel. It examines the evidence of the two groups of microfauna which, in turn, allows reconstruction of the depositional environments under which the channel developed.
General methodology
To facilitate identification of the channel deposits, a site at Bainbridge Farm (TR 016319), on the Sedbrook Sewer to the south-east of Hamstreet as shown on Figure 4.1, was used for an initial examination of the estuarine/ marine deposits. These were then traced south-westwards along Green's channel. Sections were established across the channel at four sites: near Bridge Farm (TQ 988315), by Thrift Cottage (TQ 981312), on the Blackmans Arm branch of the Speringbrook Sewer (TQ 974307) and in the Western Dowels (TQ 971301) (Fig. 4.2). These traverses were set at right angles to Green's plotted line, and hence the sections represent true dimensions and not oblique sections of the channel.
Samples were obtained from all sites by hand augering, using a 2.5cm gouge sampler. This method was effective for silts, clays and peat, although difficulties were encountered with wood fragments and compacted sands. At times, saturation of the deposits prevented their retention in the sampler, as at Thrift Cottage where the base of a sand unit could not be reached. Each borehole was then surveyed relative to Ordnance Datum, using a Kern G.K. IA level, with a maximum closing error of 3cm, from benchmarks located on the Rhee Wall and at Bridge Farm.
In total 36 boreholes were sunk at five to ten metre intervals, depending on the width of the channel and the existence of obstructions. As Green noted (1968, 109), the channel "meanders" across the line of the current creek ridge and drainage ditch. As a result the channel sediments are at times difficult to recognise under the present channel.
All cores sampled for particle size and microfauna were taken from the deepest part of the channel, in which maximum transportation of the microfossils took place. This insures that indications of salinity are preserved in this fossil record. Samples taken from the channel margins would have been more representative of bankside or bankfull conditions, and might therefore be slightly misleading.
Channel Morphology and Sediments
The deposits at the Bainbridge Farm site consist of a very dark grey(N4/)-black(5Y2,5/1) silty clay, showing parallel to sub-parallel laminations of fine to medium sand, on both a coarse and fine scale. They smell strongly of hydrogen sulphide, contain the occasional wood fragment, and show only very limited amounts ofbioturbation. At some depths, black sands dominate, and clay is subsidiary to them. The base of the deposit was not reached, as at depth the sands became so wet that they could not be retained in the corer. No other structures were visible, and the deposit completely truncated all lithologies found on the channel sides. The distinctive odour has been recognised elsewhere on salt marshes (e.g. Kellerhals and Murray 1969) and the pattern of laminations is similar to that described and photographed by Evans (1965, Plate 18 core 17 and Plate 20 core 22) for lower mudflat and muddy creek floor deposits in the Wash. Since this site was a preliminary investigation to the main work, no section has been drawn, but all other locations show variations on the same theme. The four complete sections are shown in detail in Figure 4.3, and presented in an area context in Figure 4.4. The channel deposits consist of black laminated silty clays and sands similar to those found at Bainbridge Farm. It can immediately be seen that the channel is both deepest and widest at Bridge Farm, narrowing and shallowing progressively towards the Western Dowels. Its width at the former site is approximately 30m, decreasing to under 25m at Thrift Cottage, to slightly under 20m at Blackmans Arm and to between 10m and 7.5m in the Western Dowels. Depths decrease from 2.7m to 0.75m in the same direction. In all cases these are maximum dimensions. The Western Dowels figures may be overestimated further, for the deposits around the channel margin are grey to dark grey depending on their degree of saturation, contain many vertical black organic streakings, and at time merge into the channel deposit itself.
Green (1968) indicated that the top of the peat across the marsh is usually flat, with only small undulations where relief inversion has occurred. The five-times vertical exaggeration used on the sections may over-emphasize the scale of these undulations, but the channel is incised into the peat to a greater or lesser extent at all sites except Bridge Farm. At Bridge Farm on the other hand the peat has been completely cut through and removed, and there has possibly been some incision into the underlying coarse black sand. The varying thickness of the peat around the channel in the other sections may be attributable to differential compaction and subsidence, but in all cases the peat still forms the base of the channel. It is not possible to estimate the amount of peat removed by the channel, but the lowest sediments contain both fresh brown and reworked black peat fragments.
Sedimentary Structures
The Bridge Farm deposit is the most variable, both structurally and in particle size. At first sight the lowest 15cm appeared to contain the coarsest material present, but subsequent analysis revealed no significant differences from the rest of the sample, which is described as a fine sand. It has a sharp junction with the underlying black sand, possibly suggesting an erosive contact, and contains isolated black peat fragments, either inwashed or due to channel bank slumping. Unfortunately it was not possible to compare the underlying black sand with the basal deposit mineralogically, but texturally the two are so different that the black sand is unlikely to be a major contributor to the channel deposit above it, which suggests minimal incision. Upwards, the most noticeable feature is the variable lamination in the deposit. Towards the channel margin, particularly above the fine sandy base in Borehole 5, silty-clay bands predominate with unbroken sequences up to 25cm thick recorded, whilst some sandier bands reach a maximum unbroken thickness of only 10cm. Occasionally finer sand laminations, usually between 3mm and 5mm thick, can be detected in the silty-clay bands, but the latter usually dominate.
On the other hand the lower part of the centre of the channel, in Borehole 7, does not show these thicker unbroken silty-clay bands but is frequently interrupted by sandy laminations which reach a maximum thickness of 6mm, separated by silty clay laminations generally less than 4mm thick. These patterns repeat frequently from the base to a depth of about 60cm, above which the repetitions disappear and are replaced by the thicker silty-clay bands of Borehole 5. Sandy material again becomes subsidiary to the silty clays.
These variations indicate quieter, more constant low energy conditions on the channel sides, which initially were more variable, and slightly higher energy towards the centre. Above 60cm a decrease in energy and a quieter regime is evident. No other sedimentary structures were apparent, and bioturbation was absent.
The Thrift Cottage and Blackmans Arm sites are perhaps too close together to show variation between their deposits, but they show important characteristics which differentiate them from Bridge Farm. The fine basal sand here contains both fresh and reworked peat fragments, brown and black in colour, in which wood (max. 5mm long) and reed (max. 3mm long) could be easily identified, passing upwards into a silty clay. The channel margins (Boreholes 3 at Thrift Cottage; 3, 4 and 7 at Blackmans Arm) are again of finer material with only small amounts of laminated sand, but the situation alters towards the centre (Boreholes 4 and 5 at Thrift Cottage; 5 and 6 at Blackmans Arm). Here, the number and frequency of laminations are markedly lower than at Bridge Farm, which has allowed the development of unbroken sequences of the silty clays up to 60cm, with thin sandy laminations of a maximum thickness of 30mm. Sandy sequences reach 20cm with silty-clay laminations up to 25mm. These thicker bands indicate less changeable conditions, and the higher proportion of clays compared to Bridge Farm suggests a lower energy environment of deposition. The laminations again disappear above 30cm, towards the top of the deposits, and are replaced by unbroken silty-clay sediments.
Fig. 4.3. Channel cross-sections.
Fig. 4.4. Variation in channel characteristics in a downstream direction.
At the Western Dowels site the laminations are absent, and deposits show no evidence of the sandy material recorded elsewhere. The characteristic hydrogen sulphide smell is still present, especially in the black wood fragments at the base. The black colouration is to be found mainly in the centre of the channel, but changes laterally to a very dark grey in which abundant vertical black organic streaks occur. Organics increase towards the base of the deposit, and include birch, unidentified wood, and reed fragments. Again the lowest 10cm were characterised by a higher silt/very fine sand content than the material above. A very low energy environment is suggested, where the fluctuating conditions recorded by the laminations are absent.
Particle Size Analysis
For comparative purposes particle size analysis was undertaken on samples from the Western Dowels and Bridge Farm sites which, being respectively at the south-western and north-eastern ends of the channel, were expected to show greatest variation. Samples were extracted at vertical intervals of 10cm through 0.70m of the channel deposit in the Western Dowels, and through 1.7matBridge Farm, the lowest sample at each site corresponding to the channel base deposit. Fine and coarse laminae were also analysed from depths of 45cm and 165cm at Bridge Farm, making a total of nine samples in all.
All samples were pre-treated with hydrogen peroxide to remove organic material, and subsequently gently washed through a 250µm sieve to remove remaining fragments. The residues consisted entirely of plant debris, and the analysis therefore refers to grains smaller than 250µm. During subsequent decanting of the sediments into evaporating dishes some of the finest material was lost, and hence the results may be slightly depleted in clay particles. The loss of such minimal amounts did not affect the particle size distribution significantly, and will be almost the same for all samples as they were treated similarly.
Seven of the samples varied in dry weight from 25.03gm to 27.90gm. Laminate samples were smaller, ranging from 10.9gm to 15.6gm. The percentage losses due to organic removal are shown in Table 4.1. It is immediately apparent that the Western Dowels samples contain noticeably higher proportions of organic material at all levels, particularly between 50cm and 70cm, than the top metre of the Bridge Farm deposit.
As suggested in the preceding section, this may be associated with reworking of the underlying peat, but the higher percentages above 40cm may possibly also suggest
that there was saltmarsh in the close vicinity. At Bridge Farm, the higher amounts towards the base of the channel (below 120cm) may be associated with slumping and reworking of the peat from the channel sides, and possibly inwash from elsewhere, and the upward decrease may possibly be attributable to different lithologies on the channel sides. The percentages in the 45cm laminates are broadly compatible with both overlying and underlying sediments, although the lower samples (as discussed later) may not be reliable.
The analysis itself was effected using an overlapping two-tube technique on an electronic Coulter Counter TA11. Only small samples, up to 5gms were used, which gave populations in excess of two million grains. The principles and uses of the system are described in Barfield (1982), and its validity in this work has been discussed elsewhere (Wass 1987). The method is regarded as particularly suitable for the small samples of fine-grained material used in this study.
Choice of samples for particle size analysis was controlled by two factors. Firstly, a preliminary single tube analysis indicated that the Western Dowels samples were so uniform that it was felt unnecessary to sample again at every 10cm with the more time-consuming two tube overlap. Samples from depths of 10cm and 70cm were selected as typical of the total deposit. Secondly, at Bridge Farm it was necessary to avoid using samples which were close to the boundaries of the laminations, and so only those which fell in the broader bands were included. These were those at depths of 10cm, 50cm, 80cm, 110cm and 170cm. To compensate, fine and coarse laminations at 45cm and 165cm were also sampled, although in the latter case the laminations were so fine that contamination occurred, and so these too had to be omitted.
Table 4.1. Percentage weight loss after removal of organic material.
The two-tube overlap produced 21 separate size divisions between 2.52µm and 256.02µm, which makes sample comparison difficult and not particularly meaningful. To facilitate interpretation, classes were combined and reduced to seven, corresponding approximately to clays, four grades of silt, and two of sand. The divisions do not correlate exactly with the Coulter Counter divisions, and therefore a degree of error occurs, but all samples have been similarly treated. As noted, some clay was lost during decanting and may therefore be slightly under-represented. The results are shown in Figures 4.5 and 4.6. Both Western Dowels samples show modes in the coarse silt range (4<!>) (40%+), with totals of 11% to 23% for the combined sands (3<j>,2<j>). This is noticeably different from the Bridge Farm deposits which below Im have modes in the very fine sand range, from 34% to 60% (3<!> to 4<j>), and observable proportions in the fine sand (2q>). Above this depth, modes fall into the coarse silt category, but again there are higher percentages in the sands than in the Western Dowels samples (23-34%). It was not possible to identify differences in the finer grades between the samples, except at 110cm at Bridge Farm, where the fines are virtually absent. The extreme variability of the sediments is shown in the nature of the laminations, where the coarse samples indicate 77% in the sands and 22% in the silt whereas the fine sample indicates 25% sand and 73% silt. Again, a higher percentage in the sand category is found in the fine laminae compared with the Western Dowels samples.
The Bridge Farm site also shows vertical variation with an apparent fining upwards. The deposits at 170cm and 110cm have higher proportions of fine sands, with modes in the very fine sand category. Upwards, these diminish, and are replaced by coarse and medium silts. Despite the variation recorded by the laminae, it is believed that this is a real trend because of the omission of any doubtful samples from the analysis, as described earlier.
Moments analysis (Table 4.2) confirms the above trends. Nearly all samples have mean grain sizes in the medium-coarse silt range, are poorly sorted, show very positive skews and are leptokurtic. The exceptions are at 110cm and in the coarse laminate at Bridge Farm, which both have higher skewness values and are moderately sorted. No further sedimentological difference between the Western Dowels and Bridge Farm deposits could be identified, and it would appear that the only significant variation is the higher percentage of sands in the latter, detectable even in the fine laminae. Thus a coarsening eastward, and a fining upward, can identified.
Table 4.2. Moments analysis (based on Folk & Ward, 1957).
Fig. 4.6. Grain size distribution: cumulative frequency curves.
Discussion
The channel at Small Hythe, north of the Isle of Oxney, which corresponded to the course of the Rother between the 14th and 17th centuries (Eddison 1985), is incised into bedrock, is some 125 metres wide, and is filled with medium grey sands (Waller et al. 1988). Burrin (1988) has shown that the Rother carried essentially silt-dominated sediments from the Weald to the developing marshland, and argued that any coarser deposits would have been supplied by estuarine and shallow marine processes rather than by the river itself. These characteristics allow some comparisons to be made with the channel which is the subject of this study.
If this channel was a downstream continuation of the Rother channel, it would have dimensions in excess of those of that at Small Hythe since the channel would widen seawards, and would contain medium to fine sands at its western end which would fine eastwards in a downstream direction.
On the other hand, if the channel was simply a marine creek, it would taper towards the west, possibly ending in marsh, and be significantly smaller than that at Small Hythe. This would produce reduced energy conditions allowing accumulation of fine sediment which would coarsen eastwards as the creek opened into an estuary or tidal inlet, where more vigorous tidal currents operated. The major differences between channel dimensions in this work and that at Small Hythe indicate little, if any connection between the two, and suggest that the channel here developed under very different conditions from the Rother. Further, the coarser nature of the channel sediments at Bridge Farm compared with those of the Western Dowels indicates the provenance of the coarser material was not from the west, but is more likely to have been derived from an estuarine or marine source in the east. The Wadden Zee tidal flat area behind the Friesian Islands has been described by Van Straaten (1961). He indicated that channels in dendritic creek systems were characterised by general decreases in width, depth, and maximum and average current velocities with corresponding decrease in size of bottom sediments upstream from the mouths of the creeks. Given that the channel near Bainbridge Farm is accepted as estuarine/ shallow marine, then the channel studied upstream shows the size and sediment characteristics expected of a tidal inlet. The pattern of creek ridges converging on this channel, particularly in the Western Dowels, further suggests such a dendritic system. Sediments in such a system would be expected to have marine rather than fluvial origins (Kukal 1971).
Laminations are due to tidal fluctuations in which coarser material is introduced as bedload, and finer material is dropped from suspension. The laminated nature of the deposits at the eastern end of the channel continues as far west as Blackmans Arm, but with decreasing frequency both upstream and towards the margins of the channel. This pattern is similar to those described in the Wash by Evans (1965). Thus it appears that a stronger tidal influence and higher energy environment at Bridge Farm has led to scouring of finer material and the accumulation of thicker sandier bands, particularly in the centre of the channel. In the Western Dowels, however, lower energy levels have failed to introduce the sandier components, and have allowed the accumulation only of silt. Variations in flow due to spring tides and storms would be more commonly recorded nearer the sea, and in the centre of tidal creeks, which accounts for the more rapid fluctuations at Bridge Farm, and in the centre of the sections at Thrift Cottage and Blackmans Arm. Fluctuations would be less noticeable towards the channel margins, and may not be recorded at all in the vegetated upper ends of creeks. The vertical black streakings recorded in the Western Dowels may possibly be interpreted as the decomposition of in situ plant debris. The fining upward sequence recorded particularly at Bridge Farm also indicates decreasing energy conditions, supported by the change in lamination pattern from a rapidly fluctuating to a more constant environment. This may have occurred as marine access to the inlet became reduced, or as the area silted up. The absence of bioturbation is surprising, but may be due to an unfavourable chemical environment which has produced the high levels of hydrogen sulphide in these sediments (Berner 1971, Kukal 1971).
Both the channel morphology and the sediments in it thus suggest clearly that this channel was a tidal creek rather than a downstream continuation of the Rother channel at Small Hythe.
Microfaunal Investigations
Methodology
In 1987 microfauna were used more as a support for lithostratigraphical evidence rather than as an analytical tool in their own right. Samples were small - 61 ostracoda, 209 foraminifera - but nevertheless it was clear from the assemblages that marsh and inter-tidal conditions were present (Wass 1987). New samples were taken with a 10cm gouge auger through the total thickness of the channel deposit, except at the Bainbridge Farm site where only the top three metres were sampled, since the channel was too deep for the bore to reach its base. Samples 10cm thick were extracted at Bainbridge Farm, and 30cm thick from the others. Pretreatments are summarised in Wass (1987). The samples were then collected on 250µm and 63µm sieves, and hand-picked under a low-power binocular microscope until reasonable populations were obtained. All specimens of both groups were extracted and their taxonomic composition determined.
Ostracoda
The Ostracoda are a class of small Crustaceans, usually between 0.5mm and 1mm long. Their soft-bodied parts are enclosed within a kidney- or bean-shaped carapace, consisting of two hinged valves which separate readily on death. They grow by moulting, in discontinuous stages, usually up to eight in number between egg, juvenile and adult. These stages, or instars, are designated A (=Adult), A-1, A-2, etc in descending order of size. Calcified valves or carapaces are readily fossilised, and are abundantly represented from the Cambrian period onwards. They inhabit all depths within the oceans, and also estuaries, lagoons, freshwater lakes and ponds, streams and salt lakes. They are most commonly benthonic (crawling on or through sediments) but may be phytal (living on plants) or free swimming. Their distribution is controlled by a number of factors, of which salinity, temperature and type of substrate are considered to be the most important. Consequently their ecological value is immense, and they are being used increasingly by biologists and micropalaeontologists as important environmental indicators.
In this work salinity is considered the most important characteristic in determining the environments under which this channel developed, reinforced by evidence from the substrate analysis. It is usual to divide ostracods environmentally into marine, brackish, and freshwater species. The term "freshwater" has been used elsewhere to include both saline lakes as well as freshwater ponds and streams, but its use in this work is strictly freshwater.
The interpretation of the term "brackish-water" also needs some explanation. These Ostracoda are of two types: true brackish-water species, which are only found in waters of less than normal marine salinity (ranges between 30-35%0 in British waters), e.g. Loxoconcha elliptica and Cytherois fischeri; and marine species which can tolerate reduced salinities to varying degrees e.g. Hirschmannia viridis and Hemicythere villosa. This allows them to penetrate varying distances into estuaries and tidal inlets. Some species can tolerate considerably lower or higher salinities than usual, e.g. Cyprideis torosa, which although regarded as a true brackish-water species, can tolerate salinities as high as 60 %0 in situations such as the Arai Sea. In this work freshwater, true brackish, marine/ estuarine, and marine forms are recognised.
Analysis
Overall a total of 1626 specimens of ostracods were collected. Their general distribution is shown in Table 4.3, and the species are shown with their salinity tolerances in Table 4.4.
Both the frequency and diversity of ostracods was found to increase eastwards (Fig. 4.7a). All the sites showed a high percentage of valves and low carapace content, with proportions of broken valves increasing slightly in the same direction. The western sites yielded fewer adults, but more small instars (Fig. 4.7b), and the eastern sites record lower numbers of freshwater, but increasing numbers of marine species.
Fig. 4.7. Distribution of ostracods by site.
Table 4.3. Distribution of Ostracoda according to site.
Table 4.4. Ostracods - salinity tolerances and distribution by site.
Western Dowels
Ostracods were relatively rare at this site (57 specimens belonging to 23 species) and were mainly obtained from the 63µm fraction. Only three freshwater species were recorded, consisting of single specimens of Cyclocypris globosa, Cypria opthalmica and undifferentiated Candona sp., comprising only 5% of the total.
True brackish forms were represented only by C.torosa and Leptocythere castanea, accounting for 11% of the total. The most common marine/estuarine forms included H.viridis (11%), Semicytherura angulata (5%), Leptocythere psammophila (5%) and Leptocythere baltica (3%). With only a single A-1 and two A-2 exceptions, these two groups yielded all A-5 and A-6 instars. The most common marine ostracods are Palmaconcha laevata (12%) and Heterocytherois albomaculata (5%). Semicytherura and Paracytherois species are also present (10% and 7% respectively) but cannot be differentiated due to their small size. Size distribution for the site is shown in Fig. 4.8(a), with only 8% adults, and 49% smaller than A-3. The extremely mixed nature of this assemblage, its size distribution and the broken nature of the valves of many of the marine and marine/estuarine species would indicate significant transport by tidal currents, producing a thanatocoenosis (death assemblage) which has been combined with the brackish and locally derived freshwater species.
The assemblage also includes undifferentiated marine Bairdiacean juveniles, although these have been omitted from the analysis as they are numerically insignificant.
Blackmans Arm
A total of 128 specimens belonging to 25 species was recorded at this site, and the assemblage is shown compared to that of Bainbridge Farm on Fig. 4.9. It is dominated by C.torosa (33%), which with L.elliptica, L.castanea and Leptocythere porcel/anea gives a total of 38% true brackish forms. The size distribution of C.torosa (42 specimens) is given in Fig. 4.8(b). The presence of both male and female adults, with high proportions of A-1 and A-2 instars, together with the existence of smaller juveniles is strongly suggestive of a life assemblage, and resembles the Type A histogram for the low energy biocoenosis of Whatley (1983,1988). In this case, however, some of the finest material may have been selectively removed by tidal currents to produce a slightly reduced percentage. Although only 7% of brackish forms have survived as intact carapaces, only 10% of the remaining valves have been broken, suggesting burial near life position. Freshwater species are again limited, with adults and A l instars of Cypria opthalmica, Cyclocypris globosa, and undifferentiated Candona sp. comprising 7% of the total. The most common marine/estuarine species recorded were H.viridis (4%) and H.villosa (3%) with S.angulata, Semicytherura striata and L.psammophila contributing a further 3%. A further 13 marine species were also recorded, with Paradoxostoma robinhoodi (12%), Pontacypris mytiloides (8%), H.albomaculata (4%) and Loxoconcha rhomboidea (3%) being the most common. Their size distribution is shown in Fig. 4.8(c). This histogram should be treated with caution, however. No adult specimens of P.mytiloides were recorded from this site, and the specimens were all A-1, A-2, and A-3 instars, smaller than the 1mm size to which the adult may grow. However, all adult forms present here are from species whose maximum size is 600µm (P.robinhoodi, L.rhomboidea, Paracytherois sp., Hemicytherura cellulosa and S.angulata). Thus although the histogram shows 22% adult forms, these are small species, and an analysis based on actual size rather than instars would show an increase in small material. Thus it appears that these marine forms have been introduced only by gentle currents into this environment. Generally both the size distribution, and the higher proportion of broken specimens of brackish forms than in the Western Dowels (25%), the reduction in percentage of small juveniles (from 49% to 25%), and the increase in adult forms to 18% (despite the cautionary note sounded above) suggest higher energy conditions than before.
Thrift Cottage
A total of 366 specimens belonging to 30 species were collected at this site, which showed a considerably increased and diversified marine fauna. Freshwater forms (3% of the total) consisted only of the pool dwelling Cypria exsculpta (2 adults, 1 A-2, 3 A-3 instars) and undifferentiated llyacypris sp. (1 adult, 1 A-1). True brackish specimens accounted for only 8% of the total, from three species, with L.castanea (4%) the most common, and C.torosa only 1%. The only significant marine/estuarine species is H.villosa (13%) which, when combined with other true brackish forms produce the histogram shown in Fig. 4.8(d). This compares with the Type B high energy biocoenosis of Whatley, where the energy is just sufficient to remove the younger dead valves whilst allowing larger instars and adults to remain. Twenty marine species were identified, dominated by P.mytiloides (25%), H.albomaculata (9%), P.robinhoodi
(8%), L.rhomboidea (7%) and Paradoxostoma ensiforme (6%). Table 4.3 indicates that 32% of the valves here are broken, higher than elsewhere. This figure may be misleading as 68 out of the 90 valves of P.mytiloides fall into this category. This benthonic ostracod is usually found living on a variety of coarse substrates, which only occur in laminations at these sites. This would indicate a longer period of transport from a higher energy marine environment, thus possibly giving a biased figure. Energy levels are higher than at Blackmans Arm, however, as the proportions of disarticulated valves has increased to 96%, with only 4% carapaces. 27% of the remaining valves are broken. This is reinforced by a size distribution for marine species of 57 adults, 73 A-1 instars, 60 A-2, and only 8 A-3, as shown on Fig. 4.8(e), and indicates the introduction of larger instars and adults, and the selective removal of juveniles to produce this mixed assemblage. Smaller freshwater and brackish forms may also have been flushed out, and redeposited away from these sites. Total size distribution gives an increase to 31% adults, and reduction to 3% of A-3 or smaller instars. No hidden bias is indicated in the size of the adult forms as shown at Blackmans Arm. This assemblage also included two possible juvenile valves of Cytherissa lacustris, Sars 1863, typically living in cold freshwater lakes. This may, however, be confused with juvenile C.torosa, or may indicate the existence of reworked Pleistocene material. It has therefore not been included in this analysis.
Fig. 4.8. Size distribution of ostracods. °V, '
Bridge Farm
A total of 370 specimens from 38 species were collected from here, with freshwater specimens almost absent, and a considerably increased marine fauna. Valves form 78% of the total, with increased adults (33%) and very low A-3 and smaller instars (4%). Only six freshwater valves were noted, of C.globosa and undifferentiated Ilyocypris spp., of which four were broken indicating inwash from elsewhere. True brackish forms accounted for 9% of the total, with L.castanea (4%) and C.torosa (3%) the most common. Marine/estuarine forms were dominated by H.villosa (15%) and H.viridis (3%). The size distribution of these two groups shows 21 adults, 64 A-1, 19 A-2, 6 A-3 and a single A-4 instar, and the histogram Fig. 4.8(f) again generally resembles the Type B biocoenosis of Whatley. A high proportion remain as carapaces (23%) but are mainly the smaller Leptocytherids, with 35% of the valves broken further, suggesting higher energy but a similar interpretation to the previous site.
The marine fauna is both more numerous (253 specimens) and more diverse (24 species) than previously. Although no single species dominates, H.albomaculata (13%), P.robinhoodi (9%), P.mytiloides (7%),
L.rhomboidea (6%), P.ensiforme (5%), P.laevata (5%) and L.psammophila (5%) are all common. The size distribution of these species (as a fairly typical sample) is shown in Fig. 4.8(g), and consists of 80 adults, 47 A-1, 49 A-2 and 7 A-3 instars. 22% of these specimens survive as carapaces of all sizes, possibly suggesting reduced transport despite the higher energy suggested by the size distribution, although 40 of the remaining valves have been broken further. These figures are within± 2% for the total marine fauna. The almost total absence of small juveniles is also indicative of their selective removal by currents.
Isolated juvenile Bairdiaceans were also found at this site, but are numerically insignificant and have been omitted. Similarly, single specimens of Finmarchinella finmarchica (Sars 1866), possibly from a reworked cold Pleistocene deposit, and Callistocythere currii Horne, Lord, Robinson and Whittaker 1990, from reworked interglacial deposits, have been omitted.
Bainbridge Farm
A total of 705 specimens from 40 species was collected from this site, composed of 2 freshwater, 4 true brackish, 6 marine/estuarine and 28 marine types. It shows the highest percentage of adult forms (34%), and the lowest of small juveniles (3%), with only 12% preserved as carapaces, and 26% of all valves broken. Numerically, only four specimens are freshwater, 39 are true brackish, 207 are marine/estuarine and 455 are marine. The species present are shown compared to those at Blackmans Arm in Fig. 4.9.
Freshwater forms are negligible here with only undifferentiated Cypria sp. and fragments of Eucypris sp. being present. True brackish forms are also reduced,
C. torosa and L.castanea accounting for 2% each. Marine/ estuarine forms are again dominated by H.villosa (20%) and H.viridis (3%). Size distribution for these two groups shows 47 adults, 131 A-1, 30 A-2, 1 A-3 and 1 A-4 (Fig. 4.8h), and again shows a Type B histogram. Only 14% have survived as carapaces, with 29% of the remaining valves having been broken further. No clear trends emerge in relation to size or species for the surviving carapaces.
The marine fauna is again more numerous and diverse than before, with H.albomaculata (16%), Aurila convexa (7%), P.robinhoodi (6%), P.laevata (6%), L.rhomboidea
(5%), P.ensiforme (5%) and P.mytiloides (5%) being the most common. These species give a size distribution of 127 adults, 97 A-1, 121 A-2, 7 A-3 and 1 A-4 instars, as shown on Fig. 4.8(i), although there is clear between species variation, with P.robinhoodi producing 43 adults to 1 A-1, and P.mytiloides no adults but 29 A-1 instars. Clearly, the relative sizes of each species is important here (500-600 µm against 1 µm). Of these specimens, only 11% survive as carapaces, with 23% of the remaining valves broken further. The more diverse marine, and more restricted freshwater and true brackish fauna are indicative of a location closer to true marine conditions. Higher energy levels than Bridge Farm are suggested by the size distribution of marine species, the high percentages of broken freshwater and brackish forms, which may have undergone further transport, and the relative absence of juveniles. It is otherwise interpreted in a similar way to the previous site.
As before, both isolated undifferentiated Bairdiaceans and Finmarchinellafinmarchica have been omitted from the analysis.
Discussion
If this channel represented a large river course, the western sites should record high percentages of diverse freshwater species, with little or no marine or marine/estuarine fauna. The flow of the river should have removed most small instars, and they should not have been redeposited, unless the sediment size fell below silt grade. This is the opposite of what occurs here. On the other hand, if this channel represents a restricted tidal creek opening to the east, it would be expected to show virtually no freshwater fauna, but a significant number of brackish or marine/estuarine species which would vary in aspect eastwards from marsh to fully marine. The restricted, low energy environment at the western end (head) of the creek would have allowed the accumulation of small instars, the survival of more intact carapaces, and less breakage of separated valves. All these trends are apparent here.
A useful comparison can be made with Christchurch Harbour. Whittaker (1981a) indicated that 13 species of freshwater ostracod were found in the sediments of the rivers Stour and Avon where they entered that harbour, yet only six have been found in the present study, with a maximum of three at any one site, and one of these is represented only by broken fragments. Only C.opthalmica is common to both locations, although Candona spp., which are present and clearly differentiated at Christchurch, are undifferentiated here. Their actual numbers are minimal, and do not indicate the presence of a large river feeding into this zone.
Moreover, the most common species found here, C.globosa, C.exsculpta and C.opthalmica usually inhabit the vegetation around the edges, and muddy bottoms, of permanent water bodies such as ponds, rather than rivers, and C.globosa in particular is not found under any conditions which might remotely be described as brackish. Their presence here therefore may be due to their being dislodged from their normal habitat (maybe on a marsh or floodplain surface) by a flood tide, which has caused them to contaminate a brackish assemblage rather than form an integral part of it.
Further, Kilenyi (1969) detailed a freshwater biocoenosis collected downstream from London Bridge, where nine species were present, comprising 96% of the assemblage. In total Candona spp. contributed 25%, Ilyocypris gibba 47%, and C.opthalmica 3% to the total, which is completely at variance with the proportions indicated here.
The true brackish water assemblages are almost identical with the more stable and protected creeks described from the inner part of Christchurch Harbour by Whittaker (1981a), where the presence of C.torosa is largely ascribed to extensive stands of Phragmites in the creeks, and the dominance of soft, muddy substrates, also found in the three most western sites here. Similar ecological factors are equally applicable to L.elliptica, L.castanea, and L.porcellanea. The large amounts of plant debris in the sediments would also have provided a habitat for the phytal and algal dwelling marine/estuarine species e.g. H.villosa, H.viridis, L.baltica, Cythere lutea, S.angulata and S.striata. The assemblages are very similar to those recorded by West et al. (1984) in the Earnley deposit, which was interpreted as representing the margins of the deep-water channel into Chichester Harbour, with slight variations in percentages of species indicating saltmarsh backwaters and creeks above the channel. Finally the Western Dowels, Blackmans Arm and Thrift Cottage assemblages bear marked similarities to the inner end of the Fleet lagoon, Dorset, a lower salinity brackish environment characterised by little influx of fresh water from rivers, and restricted tidal flow (Whittaker 1981b). The Bridge Farm and Bainbridge Farm assemblages are similar to those found in the outer parts of Christchurch Harbour, and include many marine species which have been inwashed, and which are included in the death assemblage of Whittaker (1981a). As indicated, they represent a brackish water biocoenosis from which juveniles have been removed, but into which marine forms have been introduced. These include benthonic varieties which would normally be found on sand substrates (L.psammophila, P.mytiloides, Pontocythere elongata, Urocythereis brittanica, Semicytherura acuticostata, Carinocythereis whitei, Cuneocythere semipunctata and Sahnicythere sp.) and phytal forms (H. cellulosa, H.albomaculata, Leptocythere tenera, L.rhomboidea, P.robinhoodi and Bythocythere bradyi). Sandy sediments are not present at these sites and the fauna must therefore have been inwashed from the sea to the east, whereas the phytal forms may have been introduced from littoral and sublittoral zones on floating seaweed by tidal flows. Further, the Bridge Farm assemblage is similar to that described by Whatley and Kaye (1971) at Selsey, comparable to the inner Solent and Southampton Water, and to the seaward end of the Fleet where a restricted marine fauna exists, characterised by many of the phytal species recorded here, and less common marine species which have penetrated from Weymouth Bay (Whittaker 1981b).
In summary, this analysis shows clearly that energy levels in this channel consistently increased towards the east. The ostracod assemblages indicate a brackish water environment of increasingly restricted circulation westwards, with no evidence of a large river in the vicinity. Neither of these is consistent with the hypothesis of a river flowing eastwards, and it appears that this channel represents the upper end of a sheltered arm of a tidal inlet.
Foraminiferida
The Foraminiferida are a group of protozoans, whose cytoplasm is enclosed within a chambered test or shell, which is usually either calcareous (sub-orders Miliolina and Rotaliina), or agglutinated (particles cemented together, sub-order Textulariina). Foraminifera are mainly benthonic but may also attach themselves to rocks, shells or plants. Many live in suspension in water (planktonic), but these species are oceanic and do not inhabit brackish or near-shore waters.
Growth is continuous, producing an adult shell usually 0.5 to 1mm in diameter. However, reduced size in a particular indigenous species is frequently observed. This can be due to several factors, particularly immaturity, or may be a response to high stress conditions such as a marked variation in salinity (common in estuaries), and under optimum conditions, reproduction at a younger and smaller stage. A further complicating factor is the frequent presence of small marine species inwashed in suspension by tidal currents. Interpretation of in situ assemblages on size alone is therefore difficult (J.E. Whittaker, personal communication).
The main ecological control on brackish water species is salinity, and the size and frequency of its variation, hence foraminifera are regarded as particularly useful indicators of sub-tidal and inter-tidal environments. Many species, however, cross boundaries - such as Protelphidium germanicum, a marine phytal species which also inhabits brackish tidal creeks and estuaries, and Miliolinella subrotunda, an inner shelf species which colonises estuaries when conditions are favourable. In this work hyposaline (low salinity), euryhaline (highly tolerant of salinity variation) and stenohaline (less tolerant of salinity variation) forms are distinguished. In terms of habitat, these three categories may be defined respectively as salt marsh, tidal flat, and marine, based on distinctive assemblages described by Murray (1971, 1979) (Table 4.5).
Analysis
A total of 5280 specimens of foraminifera were collected, and their distribution according to site is shown in Table 4.6. The frequency and diversity of species generally increased eastwards. The numbers of estuarine mudflat and marine species increase in the same direction (Figs. 4.10a, 4.11), as do the size of their tests.
Western Dowels
Considerable difficulty was experienced in obtaining a numerically large sample from the 250µm fraction without recourse to finer material. As a result 38% of the 327 specimens in this analysis were derived from the 63µm mesh, the typical sediment grain size for this site. Thus, species are excluded from discussion which could not be identified at sufficiently detailed levels, e.g. Elphidium sp., because the general species occupies many different environments, and its inclusion would not assist analysis. The percentages of small forms of the most common species are shown in Table 4.7. ("Small" is hereafter taken to indicate specimens whose size is below average for the collected material).
Two groups apparently co-dominate here, the euryhaline P.germanicum, Ammonia beccarii limnetes, and Elphidium williamsoni, and the hyposaline Haplophragmoides wilberti, Jadammina macrescens and Trochammina inflata, with significant proportions of the stenohaline species M.subrotunda and Elphidium gerthi. However, the hyposaline group consists of fully developed large tests, indicating that considerable areas of saltmarsh were adjacent to this site, whereas the other forms show noticeable size reduction. This may be due to inwash from other habitats, to immaturity, or to reduced growth under stressed conditions near their tolerance limits.
Of the remaining 13 species, 12 are typically found in open water outside estuaries, with 65% of all specimens picked from the 63µm fraction. The most common are M.subrotunda and Elphidium gerthi which have only penetrated this area to a limited extent and whose sizes are mainly reduced. They also include a number of species listed in Groups 1 or 2 by Murray (1968) in his study of Christchurch Harbour. The former consists of species introduced into the harbour by tidal currents, but never successfully colonising the area, including Quinqueloculina seminulum, Oolina williamsoni and Elphidium macellum. The latter includes Brizalina variabilis and Buccella frigida, fundamentally marine forms but capable of invading certain areas of the estuary at particular times of the year when conditions such as temperature and chlorinity are favourable. In general, these washed-in species with their small size would appear to represent a death assemblage introduced by tides and deposited under low energy conditions.
Blackmans Arm
A total sample of723 specimens was collected from here, with 178 being picked from the 63µm fraction to ensure adequate sample size (see Table 4.7).
The hyposaline forms here account for some 35% of the total assemblage, dominated by J.macrescens and T.inflata, although H.wilberti is considerably reduced. Euryhaline species are in approximately the same proportions (38%), but consist largely of P.germanicum. M.subrotunda is more plentiful (11%), but E.gerthi is reduced (1%). Although the proportion of small hyposaline forms has increased, there is abundant evidence again of adjacent salt-marsh, with possibly some selective removal of the smaller material to this site. The high proportion of small P. germanicum, M.subrotunda and E.williamsoni may again indicate inwash, or stressed conditions, although the environment would appear more suitable than before. M.subrotunda is both benthic and phytal, and may therefore either be an essential part of this assemblage, or possibly be introduced by tidal currents either as dead tests or attached to drifting weed. The reduced percentage of small specimens may indicate slightly more favourable conditions for it to live in, with increased penetration of sea water, although the dominance of the estuarine Ammonia beccarii limnetes over its marine counterpart
A. beccarii batavus should not be dismissed.
Of the remaining 21 species, 19 are marine and make up 16% of the total. In addition to those recorded in the Western Dowels, there is the appearance of Cibicides lobatulatus (Group 1 of Murray), and a sucessful Group 2 coloniser, Quinqueloculina dimidiata, whose presence in Christchurch Harbour coincided with the main passage of sea water into the estuary. Here it is only represented by three small specimens, suggesting either inwash of
Table 4.5. Principle environments indicated by foraminifera, and their distribution by site.
Table 4.7. Foraminifera. Distribution of small forms of some common species at two sites.
dead tests or an unfavourable environment, possibly towards the upper end of an isolated creek, which the abundance of hyposaline forms would support. There are also higher percentages of Lagenids to be found, especially Lagena sulcata, and more numerous and diversified Quinqueloculina spp., all of which characterise the inner shelf. This assemblage is again interpreted as mixed, with a marine thanatocoenosis combined with a biocoenosis of high marsh, upper estuarine mudflat forms. The marine influence is more significant than at the Western Dowels.
Thrift Cottage
The 1149 specimens from this sample, representing 24 species, was easily and entirely picked from the 250µm fraction, with the same species present in the finer fraction. The assemblage is dominated by euryhaline forms, with P.germanicum (47%) and E.williamsoni, the most common brackish species of Elphidium, becoming more apparent (8%). A. beccarii limnetes, although reduced in numbers, is still more common than A.beccarii batavus. Hyposaline forms are considerably reduced, with J.macrescens (11%) and T.inflata (4%) most common, and H.wilberti represented only by isolated specimens. A number of
marine species appear for the first time, including Lagena substriata and Lagena sulcata var. torquiformis, Patellina corrugata, Planorbulina mediterranensis and Triloculina trigonula. Further, there is a wider range of Quinqueloculina, including Q.cliarensis, Q.dimidiata, Q.oblonga and Q. seminulum, and E.gerthi is more common. Despite the greater diversity, however, marine species still only make up 12% of the total count once the euryhaline and hyposaline species have been deducted. M.subrotunda dominates with 18%.
This mixed assemblage is therefore interpreted as a more open part of an estuary or tidal inlet, a greater distance from high marsh environments, and closer to the sea, where currents are strong enough to inwash large marine forms.
Bridge Farm
A total of 1524 specimens from 28 species was obtained from the 250µm mesh, with similar material present in the 63µm fractions. The assemblage was again dominated by P.germanicum (45%) and M.subrotunda (22%) with
E.williamsoni (10%) and A.beccarii limnetes (2%). Hyposaline forms are present but further reduced,
Fig. 4.10. Distribution of common species offoraminifera along the Channel.
J. macrescens to 8%, T.inflata to under 4%, and only three isolated specimens of H.wilberti. 82% of T.inflata and all H.wilberti specimens are broken, which would suggest greater post-mortem transport to this site. Given the thinness of the test wall of J.macrescens and its tendency to collapse and break easily when dried it is however inadvisable to draw conclusions from the high proportion (50%+) of broken specimens of that species.
Stenohaline forms only contribute 9% of the total. There is no dominant species, although E.gerthi accounts for 3%. The five recorded Quinqueloculina species together represent only 0.9% with Q.oblonga being most common. A.beccarii batavus is again outnumbered by its brackish counterpart A.beccarii limnetes, as are the marine E.gerthi and E.macellum by E.williamsoni. The only new species recorded is Cyclogyra involvens (Group 1 of Murray), the others simply repeating those found at the earlier sites.
The increasing proportions of M.subrotunda and the slightly greater diversity of other marine forms suggest a further step seawards from the previous site, but still within the open water of an estuary or tidal inlet. Marsh would appear to be further distant, based on the decreased proportions of hyposaline forms, and the high degree of damage incurred in transport.
Bainbridge Farm
A total of 1557 specimens from 28 species was obtained from the 250µm fraction, with similar species found in
the finer material. The assemblage was again dominated by P.germanicum (39%), M.subrotunda (28%) and E.williamsoni (11%), with J.macrescens and T.inflata reduced again (4% and 3% respectively). More than 50% of both these species were broken, but there was also a notable reduction in their size, particularly relative to the Western Dowels and Blackmans Arm sites. Given their preference for high marsh locations, this may suggest that the unbroken specimens were living in stressed conditions with higher salinities, further away from the marsh, or may again indicate their selective removal and inwash to this location. The increased proportion of M.subrotunda also suggests the open sea in closer proximity.
Stenohaline forms account for 12% of the total, with E.gerthi up to 4%. Pateoris hauerinoides, although present at other sites as traces, now comprises 2% of the total, suggesting marine conditions. Quinqueloculina species have increased to 1.4% of the total, still with Q.oblonga being the most common. A.beccarii batavus is present as a trace, but still secondary to A.beccarii limnetes. Again, there is considerable diversity in the species present, and all forms are of normal size.
Finally, the presence of Heterohelix (a Cretaceous species) and Neogloboquadrina dutertrii (from the Quaternary) indicates reworking of older sediments.
Discussion
Although freshwater foraminifera do not exist per se, they do have an ecological equivalent, the Thecamoebina. These small protozoa, with finely agglutinated flask shaped tests made of sand grains, are absent from the samples studied. They were recorded by Murray (1968) as living in the rivers Avon and Stour which empty into Christchurch Harbour, and represent forms which migrated into the estuary itself when conditions were favourable. Given the comparable size of the Rother, these might have been expected to occur here, particularly at the Western Dowels and Blackmans Arm sites which, had a river existed across this part of the marsh, would be upstream. However, although they are quite common today on the sediment surface in lowland rivers, their preservation potential is poor, and they may not survive in the sediment as fossils. Nevertheless, their absence may be a further indication of the lack of a freshwater source in this zone.
It is apparent from the data that a small number of species are dominant, divided into three main groups, according to their salinity tolerances. They are summarized as follows:
Hyposaline (salt marsh):
J.macrescens, T.inflata, H.wilberti
Euryhaline (tidal flats):
P.germanicum, E.williamsoni, Trochammina haynesi, A.beccarii limnetes
Stenohaline (marine):
M.subrotunda and all other species listed in Table 4.5
A summary of results of analysis of these species appears in Table 4.8 and Figure 4.10b.
This clearly shows the diminishing presence of marsh, and the increased existence of marine conditions towards the east. The hyposaline forms achieve both their greatest size and frequency in the west, and become smaller, less numerous and more broken eastwards. This would suggest transport in that direction and more stressed conditions in higher energy open water, indicated by the increasing frequency of stenohaline forms. The assemblage in the Western Dowels and at Blackmans Arm is similar to that recorded by Murray and Hawkins (1976) in the Severn estuary and by Haynes and Dobson (1969) in the Dovey estuary for areas covered only by highest spring tides, and associated with Juncetum, Festucetum, Armerietum and Glycerietum communities.
All the sites in this study are dominated by P.germanicum, which thrives in conditions regarded as inhospitable for normal marine stenohaline forms (Murray 1968). Its salinity tolerances range from 0-35 %0, and it is abundant both in estuaries and lagoons, as described by Murray (1979). Its association with E.williamsoni and A.beccarii limnetes facilitates comparison with high and low tidal flats in the Severn estuary, and areas covered by all tides except the lowest neaps and vegetated by Salicornietum and Spartinetum grasses in the Dovey. The evidence of high marsh in the west implies longer periods of exposure and reduced periods of cover by sea water, which would produce a more hostile environment for these species, evidenced by the decrease in size of both P.germanicum and E.williamsoni, the decrease in percentage from 11% to 5% in the latter westwards, and the higher frequencies of the former in the middle and east of this zone. If the reduced size is due to the presence of inwashed, immature forms, however, then this is consistent with the low energy environments to be expected in high marsh locations. Although the sizes of A.beccarii limnetes are similarly reduced, the percentages are too low to draw meaningful conclusions.
Table 4.8. Summary of distribution of salinity groups of foraminifera.
Fig. 4.12. Environments indicated by main sub-groups offoraminifera.
Many of the stenohaline forms recorded here match the Group 1 assemblage of Murray, and represent washed in forms similar to those described in the Severn (Murray and Hawkins 1976), in the Thames (Boyd 1981) and in southern Cardigan Bay (Haynes, Kiteley et al. 1977). The figures in Table 4.8 are strongly influenced by the presence of M.subrotunda. The reduced presence of this, and the decrease in its size and those of all other marine forms may also indicate progressively less favourable conditions for the stenohaline species westwards in terms of salinity. Further, the size reduction indicates a decrease in energy levels, which is more indicative of the restricted circulation within a tidal embayment than in a river.
These assemblages are therefore regarded as mixed with clear evidence of biocoenoses at all sites; but are contaminated by the introduction of small dead marine tests. The application of statistical techniques such as the Fisher Diversity Index would appear therefore to be inappropriate as they need to be based on life or death assemblages. However, triangular plots of the relative proportions of Miliolina, Textulariina and Rotaliina indicate that all sites fall within the area of normal marine marshes, although Thrift Cottage, Bridge Farm and Bainbridge Farm sites also fall on the margins of the shelf seas zone (Fig. 4.12) (Murray 1973).
In summary, therefore, the foraminiferal evidence suggests an area of salt marsh in the west with extensive intertidal mudflats, and marine conditions towards the east. The absence of evidence of any freshwater input would therefore indicate that this area formed a sheltered arm of a tidal inlet dominated by low energy conditions.
Conclusions
The specific aim of this paper has been to present evidence which allows firm conclusions to be drawn as to whether the channel identified and mapped by Green (1968) across the Dowels was the northern course of the Rother or one of its distributaries, or whether it evolved as a tidal creek. The evidence is summarised as follows.
First, the channel dimensions are significantly smaller than upstream sections at Small Hythe, and show pronounced tapering westwards, inconsistent with the lower reaches of a large river. Secondly, particle size increases downstream, and is composed of material of coarser grade than the Rother has been shown to produce. Thirdly, the salinity inferred from the microfauna contained in the sediments, and the structures they contain, evolved under tidal rather than fluvial conditions. Fourthly, neither group of the microfauna examined shows evidence of any freshwater input within this area other than the possible existence of isolated surface pools. Fifthly, the environment indicated by the microfauna is one of tidal marsh in the west
with low energy levels, which increase eastwards towards broad intertidal tracts, culminating in near-marine conditions. The interpretation presented therefore is that this channel represents the upper end of a sheltered arm of a tidal inlet which opened eastwards. It is not consistent with any suggestion that the Rother, or a major distributary of it, has crossed this area since the formation of the peat.
It has not been the intention of this paper to attempt to fit this channel into any chronological sequence, but it is hoped that it will stimulate further investigation and re examination. Where, for example, would the watershed have lain between this creek system and that of Walland Marsh? What are the implications for the chronology of more favoured courses of the Rother such as the Romney outlets, and what are its implications for the timing and patterns of reclamation and settlement within this area, particularly in the ninth and eleventh centuries? Suffice it to say that the channel under investigation in this paper came into existence since the formation, some 3000 years ago, of the peat which it incises. It is open to further interdisciplinary investigations to integrate the geological evidence presented here with the historical data, and stimulate further discussion of the evolution of this complex area.
Acknowledgements
I would like to thank all who have helped in preparing this paper, particularly Dr David Horne of Thames Polytechnic whose enthusiasm and interest provided the impetus to initially commit pen to paper, and who with Dr John Whittaker of the Natural History Museum was invaluable both in identification and discussion of the microfauna. Their comments on earlier drafts of the manuscript were particularly helpful. Dr Peter Allen of City of London Polytechnic also provided much valuable comment. James Eddison showed considerable patience in drawing up the tables and most of the figures, and Jane Russell drew fair copies of Figures 4.1 and 4.12. I am also indebted to Jill Eddison for typing the script.
References
Athersuch, J., Horne D.J. and Whittaker J.E. 1989: Marine and Brackish Water Ostracods. E.J.Brill, The Linnean Society of London.
Berner, R.A. 1971: Principles of Chemical Sedimentology. McGraw Hill.
Boyd, P.D.A. 1981: The micropalaeontology and palaeoecology of medieval estuarine sediments from the Fleet and Thames in London. In Microfossils from Recent and Fossil Shelf Seas, Neale, J.W. and Brasier M.D. (editors), Ellis Horwood, Chichester. 274-292.
Brooks, N.P. 1988: Romney Marsh in the early Middle Ages. In Eddison, J. and Green, C. (editors), Romney Marsh, Evolution, Occupation, Reclamation. OUCA Monograph 24, 90-104.
Burrin, P.J. 1988: The Holocene floodplain and alluvial fill deposits of the Rother Valley and their bearing on the evolution of Romney Marsh. In Eddison, J. and Green, C. (editors), Romney Marsh, Evolution, Occupation, Reclamation. OUCA Monograph 24, 30-52.
Callow, W.J., Baker, M.J., and Pritchard, D.M. 1964: National Physical Laboratory radiocarbon measurements II. Radiocarbon 6, 25-30.
Cunliffe, B.W. 1980: The evolution of Romney Marsh: a preliminary statement. In Thompson, F.H. (editor), Archaeology and Coastal Change (London Soc. Antiq.), 37-54.
Eddison, J. 1983: The evolution of barrier beaches between Fairlight and Hythe. Geogr. Journ. 149, 39-75.
Eddison, J. 1985: Developments in the lower Rother valleys up to 1600. Arch. Cant. 102, 95-110.
Elliott, J. 1862: (maps and opinions reported in) Lewin, T., The invasion of Britain by Julius Caesar, London.
Elliott, J. 1874: (views reported in) Furley, R. The History of the Weald of Kent, Ashford.
Evans, G. 1965: Intertidal flat sediments and their environments of
deposition in the Wash. Quart. J. Geol. Soc. London, 121, 209-241.
Folk, R.L. and Ward, W. 1957: Brazos river bar: a study in the significance of grain-size parameters. J. Sedim. Petrol., 27, 3-26.
Green, C.P. 1988: Palaeogeography of marine inlets in the Romney Marsh area. In Eddison, J. and Green, C. (editors), Romney Marsh, Evolution, Occupation, Reclamation. OUCA Monograph 24, 167-174.
Green, R.D. 1968: Soils of Romney Marsh. Soil Survey Gt. Britain, Bull.4, Harpenden.
Green, R.D. and Askew, G.P. 1959: Kent. Rep. Soil Surv. Gt Britain 11, 22-27.
Harfield, J.G. 1982: BehaviorofConductingParticles in the Sensing Zone of the Coulter Counter. P.O.C. Particle Size Analysis, Loughborough Univ. 1981. Wiley-Heyden, 151-164.
Haynes, J. 1973: Cardigan Bay Recent Foraminifera. Bull. British Museum (Natural History) Zoology, Supplement 4, London.
Haynes, J. and Dobson, M. 1969: Physiography, forarninifera and sedimentation in the Dovey Estuary (Wales). Geol. Journ. 6, 217-256.
Haynes, J.R., Kiteley, R.J., Whatley, R.C. and Wilks, P.J. 1977: Microfaunas, microfloras, and the environmental stratigraphy of the Late Glacial and Holocene in Cardigan Bay. Geol. Journ. 12, pt.2, 129-159.
Holloway, W. 1849: The History of Romney Marsh. London. Holmes, T.R.E. 1907: Ancient Britain and the Invasion of Julius Caesar. Oxford.
Kellerhals, P. and Murray, J.W. 1969: Tidal flats at Boundary Bay, Fraser River Delta, British Columbia. Bull. Can. Petrol. Geol., 17(1), 67-91.
Kilenyi, T.I. 1969: The problems of ostracod ecology in the Thames estuary. In Neale, J.W. (editor); The Taxonomy, Morphology and Ecology of Recent Ostracoda, Oliver and Boyd, Edinburgh, 251-267.
Kuka!, Z. 1971: Geology of Recent Sediments, Acad. Press Inc., London and New York.
Lambarde, W. 1576: A perambulation of Kent; containing the description, historie and customes of that shyre. London.
Lewis, W.V. 1932: The formation of Dungeness foreland. Geogr. Journ. 80, 309-324.
Marlow, A.O. 1984: Stratigraphic andforaminiferal analyses of coastal Flandrian deposits, Pett Level, East Sussex. Unpublished M.Sc. thesis, Polytechnic of North London and City of London Polytechnic.
Murray, J.W. 1968: The living Foraminiferida of Christchurch Harbour, England. Micropalaeontology 14, (1), 83-96.
Murray, J.W. 1971: An atlas of British Recent Foraminiferids.
Heinemann, London.
Murray, J.W. 1973: Distribution and ecology of living benthic Foraminiferids. Heinemann, London.
Murray, J.W. 1979: British nearshore Foraminiferids. Acad. Press, London, New York and San Fransisco.
Murray, J.W. and Hawkins, A.B. 1976: Sediment transport in the Severn Estuary during the past 8000-9000 years. Journ. Geol. Soc. London, 132, 385-398.
Steers, I.A. 1964: The coastline of England and Wales. Cambridge U.P.
Tooley, M.J. and Switzur, V.R. 1988: Water level changes and sedimentation during the Flandrian Age in the Romney Marsh area. In Eddison, J. and Green, C. (editors), Romney Marsh, Evolution, Occupation, Reclamation. OUCA Monograph 24, 53-71.
Topley, W. 1875: The Geology of the Weald. Mem. Geol. Surv. U.K.
van Straaten, L.M.J.U. 1961: Sedimentation in tidal flat areas.
Journ. Alberta Soc. Petrol. Geol. 9, (7), 203-226.
Waller, M., Burrin, P.J. and Marlow, A.O. 1988: Flandrian Sedimentation and Palaeoenvironments. In Eddison, J. and Green, C. (editors), Romney Marsh, Evolution, Occupation, Reclamation. OUCA Monograph 24, 164-174.
Ward, G. 1931: Sandtunes Boe. Arch. Cant. 43, 39-47.
Ward, G. 1933a: The River Limen at Ruckinge. Arch. Cant. 45, 129-132.
Ward, G. 1933b: The Saxon ChartersofBurmarsh.Arch. Cant. 45, 133-141.
Wass, M.P. 1987: The proposed northern course of the Rother - an investigation of its morphological, sedimentological and microfaunal characteristics. Unpublished M.Sc. Thesis, Polytechnic of North London and City of London Polytechnic. West, R.G., Devoy, R.J.N., Funnell, B.M., and Robinson, J.E. 1984: Pleistocene deposits at Earnley, Bracklesham Bay, Sussex.
Phil. Trans. Royal Soc. London, B, 306, 137-157.
Whatley, R.C. 1983: The application of Ostracoda to palaeoenvironmental analysis. In Maddocks, R.F. (editor): Applications of Ostracoda, University of Houston Geoscience, Houston, 51-77.
Whatley, R.C. 1988: Populations structure of ostracods: some general principles for the recognition of palaeoenvironments. In: Ostracoda in the Earth Sciences, De Deckker, P., Coli, J-P., Peypouquet, J-P. (editors). Elsevier, Amsterdam, Oxford, New York, Tokyo, 245-256.
Whatley, R.C. and Kaye, P. 1971: The palaeoecology of Eemian (last interglacial) Ostracoda from Selsey, Sussex. In: Palaeoecologie des Ostracodes, Oertli, H. (editor), Bull. Centre Recherches Pau - S.N.P.A., 5 supplement, 311-330.
Whittaker, J.E. 1981a: The living Ostracoda of Christchurch Harbour, Dorset. In Neale, J.W. and Brasier, M.D. (editors) Microfossils from Recent and fossil shelf seas, Ellis Horwood, Chichester, 293-303.
Whittaker, J.E. 1981b: The distribution of Ostracoda in the Fleet. In Ladle, M. (editor): The Fleet and Chesil Beach. Structure and biology of a unique coastal feature. Dorset County Council, Dorchester, 61-64.
Woodcock, A.G. 1984: The sedimentological history of the Pannel Valley, Pett, East Sussex. Unpublished M.Sc. thesis, Polytechnic of North London.
