Macroinvertebrate distributions and the conservation value of aquatic Coleoptera, Mollusca and Odonata in the ditches of traditionally managed and grazing fen at Wicken Fen, UK

Authors


* Present address and correspondence: Duncan Painter, Land Use Consultants, 43 Chalton Street, London NW1 1JD, UK.

Abstract

1. Water-filled ditches are an important ecological feature of lowland Britain. Originally dug to facilitate wetland drainage, ditches often provide refuges for aquatic flora and fauna of high conservation value.

2. The ditches, ponds and major waterways of a traditionally managed undrained fen and the ditches of a previously drained cattle-grazed fen meadow were sampled at Wicken Fen National Nature Reserve for Coleoptera, Mollusca and Odonata, and for physical, chemical and biological variables.

3. Multivariate analysis showed a clear division between ditches on the two fens, and between larger and smaller waterbodies.

4. Individual ditches possessed distinctive faunas. Ditch age and bank profile were important factors influencing faunal species composition.

5. Invertebrate distributions were shown to be significantly correlated with macrophyte distribution.

6. Waterbodies were ranked in terms of their species quality, using a point-scoring system; there was no correlation between rankings for the three invertebrate groups; only ditch age and detritus levels were correlated with species quality score (for Coleoptera).

7. Ditch management recommendations are discussed in the light of these results.

Introduction

Ditches are ubiquitous features of Britain's lowland wetlands. In the past, networks of ditches have served to conduct water away from fens and marshy areas to permit exploitation of the land for agriculture. Today, however, remaining wetland landscapes are being conserved and, in some places, former wetlands being reclaimed; in these cases, the function of ditches may be reversed, to bring water onto the site and store it. The management of fenland ditches may be largely determined by this functional, hydrological requirement, in which case ditches should be kept clear by periodic removal of vegetation and sediment. However, ditches support floral and faunal elements that may not otherwise be represented in the area and may therefore be considered worthy of conservation (Foster et al. 1990). The composition of the ditch community is likely to be a function of the successional stage reached during the inexorable in-filling of the ditch by vegetation and sediment (Caspers & Heckman 1981, 1982). Despite an upsurge in interest in the ecology of ditches since the 1970s, the amount of published work on the aquatic macroinvertebrate fauna of ditches in Britain is sparse (Clare & Edwards 1983; Eyre, Foster & Foster 1990; Foster et al. 1990; Painter & Friday 1995), with most of the work being unpublished biological surveys (R.J. Driscoll in press) and ditch management recommendations being made on the basis of entomological intuition (Kirby 1992).

Any attempt to manage a wetland reserve must address the potentially conflicting requirements of keeping ditches open and allowing the hydrosere to progress. Two basic questions must be resolved: to what extent does the recent management history of a ditch determine the composition of its aquatic community; and what are the relative ‘conservation values’ of communities found under different management regimes?

Wicken Fen National Nature Reserve (NNR), Cambridgeshire, UK (Grid Reference TL5570) (Fig. 1) is owned and managed by the National Trust. During the first half of the twentieth century, cessation of cropping and falling water levels led to extensive invasion of Wicken Sedge Fen by fen carr, consisting principally of Frangula alnus Miller and Rhamnus cathartica L. In 1961, a management plan was drawn up to arrest systematically the Sedge Fen's decline and to restore its former habitats. Foremost among the objectives of the management was the restoration and maintenance of diversity appropriate to an East Anglian fen system with a long history of cropping (Friday 1994). Adventurers’ Fen lies south of Wicken Lode and is hydrologically isolated from the principal drainage (‘lode’) system (Friday & Preston 1997). In contrast to the Sedge Fen, Adventurers’ Fen was cut over intensively for peat up to the end of the nineteenth century and was drained for cultivation in 1940, although large areas were restored to grazing meadow in the 1950s. This is in contrast to the more traditional management of regular cropping that takes place on the Sedge Fen (Lock, Friday & Bennett 1997).

Figure 1.

Map of Wicken Fen National Nature Reserve (NNR)

Wicken Fen's water-filled ditches, ponds and lodes support rich assemblages of aquatic invertebrates, many of which are confined to low-lying fens in the east of England, and which are nationally rare and vulnerable (Foster & Eyre 1992; Painter 1994). The water-filled ditches on the Sedge Fen were originally dug in the seventeenth and eighteenth centuries (Rowell 1986), and since 1961 have been managed by periodic mechanical clearance of accumulated detritus and vegetation. The ditches on the Sedge Fen are now subject to a management regime that clears short (50 m) stretches of ditch on 4–12 year cycles, with some newly re-excavated old ditches being allowed to run through succession to dryness (Wicken Fen Local Management Committee 1992). The ditches on Adventurers’ Fen mostly date from the 1950s and most had not been managed up to the time of this study; of the others, Commissioners’ Drain was re-made in 1988 and one of the ditches surrounding the mere was cleared out in 1990. This study investigated the distributions of aquatic Coleoptera, Mollusca and Odonata on the traditionally managed Sedge Fen, compared them with those found on Adventurers’ Fen, and sought to identify which characteristics of the ditch habitat are most closely correlated with the number of species and occurrence of rare species in these three taxa. The taxa were chosen for study because of their diversity at Wicken Fen and because an extensive literature exists on the conservation status and distribution of individual species (Kerney 1977; Foster 1987; Foster & Eyre 1992; Merritt, Moore & Eversham 1996; Willing 1997).

Materials and methods

A variety of waterbodies consisting of ditches (12), brick-pit ponds (four) and lodes (three) within Wicken Sedge and Adventurers’ Fens were sampled for Coleoptera, Mollusca and Odonata; 67 samples were taken from 19 waterbodies (Fig. 1). A standard pond net (mesh 0·5 mm; GB Nets, Todmordon, Lancs., UK) was used from the bank side over 5 consecutive days in June 1994. At each sampling location (see Painter 1995 for exact locations), four vigorous 15-second sweeps along 1 m of bank side were pooled together to form one sample. The samples were preserved immediately with 70% industrial methylated spirits, and invertebrates were sorted from debris in the laboratory. The amount of detritus as a percentage of the whole sample composition by volume was estimated during the sorting stage. Molluscs, odonate larvae and Coleoptera were identified to species where possible using Macan (1977), Miller (1987) and Friday (1988).

At each sampling location the profile of the bank (the acute angle between the slope of the bank and the water surface) was measured. Water depth, the species of macrophyte present and their percentage cover were recorded. Conductivity (using a portable water meter, Jenway model 4200; Merck, Lutterworth, Leics., UK), dissolved oxygen (Jenway 9200; Merck, Lutterworth, Leics., UK) and pH (Jenway 3150; Merck, Lutterworth, Leics., UK) were measured in the middle of the water column. A filtered (0·5 μm) water sample was also taken from the middle of the water column at each sampling point and transferred to (150 ml) iodized polythene bottles and kept cool for transportation back to the laboratory. Samples were analysed for nitrate and phosphate (Dionex system 2000 ion chromatograph; Merck, Lutterworth, Leics., UK), calcium and magnesium (Perkin-Elmer 2280 Atomic Absorption Spectrophotometer; Merck, Lutterworth, Leics., UK), sodium and potassium (Corning 410 flame photometer; Merck, Lutterworth, Leics., UK) and alkalinity (the acid titration method described by Mackereth, Heron & Talling 1989). Ditch age in years since last excavation was calculated from management archives.

Classification

Classification of the invertebrate data was performed using twinspan (Hill 1979) on untransformed invertebrate abundance data. The number of individuals in each species is expressed as abundance categories 1–5 corresponding to 1; 2–4; 5–9; 10–19; >20 individuals, respectively. The analysis was terminated after the fourth division.

Ordination

Ordination was performed using canoco Version 3.10 (ter Braak 1987, 1990). Canonical correspondence analysis (CCA) was performed using default values and untransformed invertebrate abundance data. The CCA was first run with all variables. All variables (except pH) were logarithmically transformed before analysis. However, CCA produces a correlation matrix for the explanatory variables used in the ordination, and calculates an ‘inflation factor’ (IF) for each variable (ter Braak 1987). Variables that are almost perfectly correlated with each other, and therefore have no unique contribution to the regression equation, have high IF values >20; where necessary, ordinations were re-run after removing those variables with IF values >5. An unrestricted Monte Carlo permutation test was used to test the significance of the relationship of the first canonical axis to the environmental variables (ter Braak 1987, 1990).

Macrophytes

The macrophyte data collected at each sample site were not added as environmental variables to the ordinations because plant distributions are themselves likely to be partly determined by the environmental variables. The untransformed data for macrophytes were ordinated by detrended correspondence analysis (DCA) and the site scores on the first axis were tested for correlation with the corresponding site scores obtained by performing DCA of the invertebrate data.

A point scoring system for coleoptera, mollusca and odonata

Foster (1987) describes a point scoring system for evaluating and ranking wetland sites by the rarity status of the water beetle species present: each species is assigned a score based on a geometric progression of national abundance scores successively doubling from 1 (commonest species) to 32 (rarest); an aggregate species score for each waterbody is calculated, and a species quality score (SQS) calculated by dividing the aggregate score by the number of species sampled from the waterbody. This system was used to rank the Wicken Fen waterbodies. Aggregate scores were calculated using points assigned to East Anglian beetle species according to Foster & Eyre (1992) (Table 1). Beetles that were classed as Nationally Notable and those frequently associated with low-lying fen conditions (G.N. Foster, personal communication) are also shown on Table 1.

Table 1.  The scores awarded to species of Coleoptera based on Foster & Eyre (1992). All other beetle species caught in the survey scored 1 and are not shown. RDB 2 = Vulnerable: taxa believed likely to move into the endangered category (in danger of extinction RDB 1) in the near future if the causal factors continue to operate (Shirt 1987). RDB 3 = Rare taxa with small populations that are not at present endangered or vulnerable but are at risk (Shirt 1987). Na = ‘Notable A’ species thought to occur in <30 10-km grid squares (Hyman 1992). Nb = ‘Notable B’ species (30–100 10-km grid squares) (Hyman 1992). Fen = species that are frequently associated with low-lying fen conditions (G.N. Foster, personal communication)
 ScoreStatusScoreStatus
HALIPLIDAE  HELOPHORIDAE 
Peltodytes caesus4Nb/fenHelophorus griseus4
Haliplus confinis2 H. nanus4
H. flavicollis4 HYDROCHIDAE 
H. immaculatus2 Hydrochus elongatus8
H. lineolatus4 HYDROPHILIDAE 
H. obliquus4 Coelostoma orbiculare8
HYGROBIIDAE  Cercyon convexiusculus4
Hygrobia hermanni2fenC. sternalis8
NOTERIDAE  C. tristis4
Noterus clavicornis2 Anacaena bipustulata4
DYTISCIDAE  A. limbata2
Laccophilus hyalinus2 A. lutescens2
Hygrotus decoratus4Nb/fenLaccobius biguttatus2
H. versicolor2fenL. minutus2
Hygrotus impressopunctatus2 Helochares lividus2
Hydroporus erythrocephalus2 Enochrus coarctatus2
H. incognitus2 E. ochropterus8
H. neglectus8NbE. quadripunctatus8
H. striola2 E. testaceus2
Suphrodytes dorsalis2fenCymbiodyta marginella2
Graptodytes granularis4Nb/fenChaetarthria senunulum2
G. pictus2 Berosus luridus4
Porhydrus lineatus2 HYDRAENIDAE 
Stictotarsus duodecimpustulatus2 Octhebius bicolon4
Scarodytes halensis4Nb/fenO. dilatatus2
Laccornis oblongus4NaHydraena palustris16
Copelatus haemorrhoidalis2fenH. testacea4
Agabus didymus4 Limnebius aluta8
A. uliginosus16NaL. nitidus4
A. undulatus16RDB 2/fenL. papposus4
llybius ater2 ELMIDAE 
I. fenestratus4NbOulimnius major8
I. quadriguttatus4fenO. tuberculatus2
Rhantus exsoletus4 DRYOPIDAE 
R. grapii8Nb/fenDryops anglicanus16
R. suturalis2NbD. ernesti2
Hydaticus seminiger4Nb/fenD. griseus32
Dytiscus semisulcatus2   

This point-scoring system was adapted for Mollusca and Odonata. Data from Kerney (1977) and Bratton (1991) were used to award points to each molluscan species on the basis of its national distribution and rarity (Table 2). The common species of mollusc (present in >100 10-km grid squares) are subdivided according to the number of ‘pre-1950 only’ records: a large number of such records may indicate a decline in the species. Thus, common species with >30 ‘pre-1950 only’ are awarded two points, while those with <30 are awarded one point. Similarly, Shirt (1987) and the English Nature Invertebrate Site Register database for Wicken Fen were used to assign scores to species of Odonata (Table 3). Larvae of Coenagrion pulchellum and C. puella could not be distinguished; as both species are known to breed in some of the waterbodies at Wicken (Painter 1995) the larvae were awarded the score of the rarer species, C. pulchellum.

Table 2.  The scores awarded to each species of Mollusca based on species’ rarity and distribution. RDB categories as Table 1  
SpeciesScoreStatus
VALVATIDAE
Valvata cristata 2
V. macrostoma32RDB 2
HYDROBIIDAE
Potamopyrgus antipodarum 1
Bithynia leachii 2
B. tentaculata 2
LYMNAEIDAE
Lymnaea palustris 1
L. peregra 1
L. stagnalis 2
PHYSIDAE
Aplexa hypnorum 2
Physa fontinalis 1
PLANORBIDAE
Anisus vortex 2
Armiger crista 2
Bathyomphalus contortus 1
Gyraulus albus 1
Hippeutis complanata 2
Planorbarius corneus 2
Planorbis carinatus 1
P. planorbis 2
ANCYLIDAE
Acroloxus lacustris 2
UNIONIDAE
Anodonta anatina 2
SPHAERIIDAE
Sphaerium corneum 1
S. lacustre 2
PISIDIIDAE
Pisidium casertanum 1
P. milium 1
P. nitidum 1
P. obtusale 1
P. pseudosphaerium16RDB 3
P. subtruncatum 1
Table 3.  The scores awarded to each species of Odonata based on species’ rarity and distribution. Notable categories as Table 1. Local = Wicken Fen Invertebrate Site Register database (Joint Nature Conservation Committee, unpublished data)
SpeciesScoreStatus
COENAGRIONIDAE
Coenagrion puella/pulchellum4Nb
Enallagma cyathigerumm1 
Erythromma najas2Local
Ischnura elegans1 
Pyrrhosoma nymphula2Local
LESTIDAE
Lestes sponsa1 
AESHNIDAE
Aeshna cyanea1 
A. grandis1 
A. juncea1 
Brachytron pratense4Nb
LIBELLULIDAE
Libellula quadrimaculata2Local
Orthetrum cancellatum1 
Sympetrum sanguineum2Local
S. striolatum1 

Results

Classification

The first twinspan division split all of the Adventurers’ Fen sites (groups 3 and 4) from all the Sedge Fen and lode sites (groups 1 and 2) at the first division (Table 4). The Red Data Book (RDB; Shirt 1987) water beetle Agabus undulatus was recorded in high numbers in Adventurers’ Fen sites; only two specimens were recorded from any other site (the old ditch along Sedge Fen Drove). Laccobius minutus was recorded in very high numbers from Adventurers’ Fen and was recorded in low numbers from the Sedge Fen and lode samples. The mollusc Bathyomphalus contortus was recorded regularly in high numbers from Adventurers’ Fen samples, while it remained unrecorded from the Sedge Fen and lode samples. In contrast, the water beetle Hydroporus memnonius was not recorded from Adventurers’ Fen but was widespread and common in the Sedge Fen and lodes.

Figure 4.

Figure 4.

twinspan abundance table of sites and species sampled from the Sedge and Adventurers’ Fen: 1–4 = end-groups after two twinspan divisions; 1.1–4.2 = end-groups after four divisions. For site code letters (A–S) see Fig. 1. Lower case letters (a–b) denote species’ end-groups

Figure 4.

Figure 4.

twinspan abundance table of sites and species sampled from the Sedge and Adventurers’ Fen: 1–4 = end-groups after two twinspan divisions; 1.1–4.2 = end-groups after four divisions. For site code letters (A–S) see Fig. 1. Lower case letters (a–b) denote species’ end-groups

At the second division, four groups of distinctively different habitats emerged. On one side of the dendrogram, the lode samples, and samples from the large ditches and the main brick-pit (group 1), was split from the other remaining Sedge Fen sites, which were all narrower ditches or small ponds (group 2). Laccophilus hyalinus was recorded from every site in group 1, but occurred less frequently in the waterbodies that comprised group 2. In contrast, Hydroporus planus was recorded most frequently from group 2. Bithynia tentaculata was recorded from every waterbody; it was, however, most abundant in group 2. On the other side of the dendrogram, the second division split the 1950s ditches surrounding the mere (group 4) from the other Adventurers’ Fen ditches (group 3). Anisus vortex was absent from only two Adventurers’ Fen samples, but was most abundant in group 3. Haliplus lineatocollis was abundant in every sample in group 4, but was found in only one sample in group 3.

The ordered species-by-sample abundance table (Table 4) shows a strong diagonal pattern of species abundance. Species groups a–h represent the end-groups of the species classification. Groups a and b consist of taxa that were mostly or exclusively found in Sedge Fen and lode sites comprising groups 1 and 2 while, at the bottom of the table, species groups g and h contain taxa that were most abundant in the Adventurers’ Fen samples comprising groups 3 and 4.

Cca ordination

The physical and chemical data for all sites are shown in Table 5; the 14 site end-groups derived from four twinspan divisions are shown and the mean values given for each varied little between sites. Almost all the environmental variables were highly correlated. Sodium, calcium, potassium, magnesium and alkalinity were highly correlated with conductivity, so the CCA was run omitting these variables. Figure 2 shows the distribution of the four major twinspan groups and environmental variable arrows in the CCA ordination. Groups 3 and 4 (Adventurers’ Fen) were clearly separated from each other and from groups 1 and 2 (Sedge Fen and lode sites). Groups 3 and 4 consisted of ditches that were relatively shallow but of comparable low temperature, high in detritus and conductivity, with gently sloping bank profiles and low concentrations of nitrate (Table 5). Sites in group 1 (which included the lodes, major ditches and main brick-pit) had, in general, been recently cleared out and possessed small amounts of detritus in their substrata. They also had steep bank profiles, relatively high levels of nitrate but low conductivity. Sites in group 2 (narrower Sedge Fen ditches, Wicken Lode dead-end and the smaller brick-pits) were generally low in oxygen and nitrate, and were relatively warm. They were deeper than sites in groups 3 and 4 but shallower than sites in group 1, had steep bank profiles (although generally less steep than sites in group 1), and were high in detritus. The longest and most influential variable arrow in determining the position of sites in the ordination was bank profile. Ditch age was apparently not a major factor; however, the majority of sites in groups 3 and 4 were more than twice the age, in terms of time since last excavation, of the sites that comprised groups 1 and 2, with the exception of Sedge Fen Drove and the brick-pit sites.

Table 5.  Mean values of environmental variables for each sample site: 1–4 = end-groups after two twinspan divisions; 1.1–4.2 = end-groups after four divisions. For site code letters (A–S) see Fig. 1
 1.1
AC
1.2
B
1.3
FE
1.4
HDM
2.1
N
2.2
KJL
2.3
GD
2.4
D
3.1
P
3.2
Q
3.3
P,Q
3.4
R
4.1
S
4.2
S
Depth (cm)584071713538342582317224412
Bank profile (°)619292929057869311040271
Temperature (°C)13.113.214.314.214.41516.315.811.411.511.612.312.313.3
Conductivity (μs)602660644600726663691710100097311171202597687
pH7.57.57.47.47.37.77.67.57.47.37.47.47.17.1
Oxygen (mg L−1)11.97.27.78.45.81.87.27.711.98.48.36.23.08.8
NO3 (mg L−1)0.490.170.370.370.080.070.190.570.030.040.070.160.060.12
Ca (mg L−1)11210513013198123104130245246250256151133
K (mg L−1)3.52.83.94.02.52.83.33.6911.31010.89.19.8
Mg (mg L−1)4.14.14.34.455.24419.124.621.924.4910.2
Na (mg L−1)16.617.116.916.717.61717.11635.937.4344022.523
Alkalinity (meq L−1)4.74.94.14.23.93.23.74.67.98.587.65.95.9
Time since last management (years)497219320610638638261
Detritus (% composition)92903522016205627592751
Figure 2.

CCA showing sample sites, twinspan end-groups (numbered polygons) and environmental variables (arrows). For key see Fig. 1. *Recently excavated.

Axes 1 and 2 together explained 18·3% of the variation in the species data; axes 3 and 4 together contributed another 5·9%. The environmental variables nitrate, temperature, conductivity, detritus, depth, bank profile, age and oxygen together accounted for 30·3% of the variation in the species data. The relationship between the first canonical axis and all the environmental variables was significant (P < 0·01).

The CCA ordination scores of selected species of Coleoptera, Mollusca and Odonata were plotted in relation to the twinspan end-groups and environmental variables. Figure 3 shows the distribution of species recorded in numbers of three or more, with RDB status (Shirt 1987; Bratton 1991) or specifically associated with fen habitats (i.e. 24 Coleoptera, two Mollusca, no Odonata).

Figure 3.

CCA showing Coleoptera and Mollusca ordination scores. Only Coleoptera specific to low lying fen conditions (fen in legend) are plotted. Red Data Book species are shown in bold.

Of the RDB beetle species, Agabus undulatus (RDB 2) was widespread and abundant only in Adventurers’ Fen sites (groups 3 and 4), while Dryops anglicanus (RDB 3) was widespread in its distribution, being recorded from sites in all four groups. Dryops griseus (RDB 3) was recorded on only one occasion, from Malcarse Drain (group 1). Three specimens of Hydraena palustris (RDB 2) were recorded during the survey, two from the brick-pits (group 2) and one from an Adventurers’ Fen meadow ditch (group 3). Hydrochus elongatus (RDB 3) was recorded in low numbers in Adventurers’ Fen meadow ditches, and Oulimnius major (RDB 3) was recorded frequently in Wicken Lode and ditches connected to it (group 1).

Of the RDB molluscan species, Pisidium pseudosphaerium (RDB 3) was recorded in low numbers from four sites (Malcarse Drain, Drainers’ Dyke, Brick-pit 78d and Gardiner's Ditch), while a single specimen of Valvata macrostoma (RDB 2) was recorded (from Brick-pit 78f).

Classification of sedge fen sites alone

In order to examine the possible effects of the management cycle on the pattern of invertebrate distribution on the Sedge Fen, the classification was repeated after removal of the Adventurers’ Fen samples.

The first twinspan division split the lodes, the main brick-pit and larger ditches from the remaining brick-pits and narrower ditches. The second division split Wicken Lode and Monks’ Lode from the larger ditches (Cross Dyke, New Dyke, Malcarse Drain and Drainers’ Dyke) and the main brick-pit. On the other side of the dendrogram, the second division split the smaller brick-pits, Sedge Fen Drove, Gardiner's Ditch and Thomson's Ditch from North Dyke and New Dyke. At the third division, eight end-groups corresponding closely to the individual waterbodies were created and were identical to the groupings 1.1–2.4 in Table 5.

Cca ordination of sedge fen sites

Sites in end-groups 1.1, 1.2 and 1.3 were relatively deep, with steep bank profiles, and were high in oxygen and nitrate; the Wicken Lode dead-end sites (end-group 1.2) were lower in oxygen and nitrate than sites in end-groups 1.1 and 1.3. In end-group 1.4, Drainers’ Dyke sites (H) were also high in oxygen and nitrate, and were relatively deep with steep bank profiles; the main brick-pit (M) had a steep bank profile but it was low in oxygen and nitrate; New Dyke sample 2 was high in nitrate and oxygen, with a steep bank profile; both the main brick-pit and New Dyke (D) samples were generally about two degrees warmer than those from Drainers’ Dyke.

The brick-pits that comprised end-group 2.1 and Sedge Fen Drove (end-group 2.2) were the oldest sites (93 and 100+ years, respectively) and contained large quantities of detritus in their substrata; they were also low in nitrate and oxygen. However, Sedge Fen Drove (K) had a shallow profile while the brick-pits (N) had steep sides. Gardiner's (J) and Thomson's Ditch (L) samples (end-group 2.2) were relatively shallow and low in oxygen and nitrate. Gardiner's Ditch samples were approximately three degrees warmer than those from sites in end-groups 1.1–1.4; some of the Gardiner's Ditch sites had shallow bank profiles.

The North Dyke (G) sites (end-group 2.3) were some of the youngest, the ditch having been re-excavated totally in 1991. These sites were relatively low in nitrate, high in oxygen, and possessed steep bank profiles, while one New Dyke (D) site was high in nitrate and detritus. All samples in end-group 2.3 were about three degrees warmer than those in end-groups 1.1–1.4. The remaining New Dyke sites (end-group 2.4) had similar environmental characteristics to the New Dyke sample in group 2.3 but had smaller amounts of detritus in their substrata.

Macrophyte and invertebrate dca ordination

The DCA axis 1 site scores for macrophytes and invertebrates were highly correlated (Spearman's Rank: P < 0·001, r = 0·522, n = 67).

Ranking waterbodies in terms of invertebrate species

Table 6 shows sites ranked in descending order of species quality score (SQS) for Coleoptera; the ranks of SQS are also shown for Mollusca and Odonata. The waterbodies were ranked in a different order for each invertebrate group, and species quality scores of sites for different taxa were not significantly correlated (Table 7). There were, however, some sites, notably the brick-pits 78e and 78f, that had high SQSs for each of the three groups, and colleagues, such as Monks’ Lode, that had consistently low scores. For Coleoptera, the SQSs were significantly correlated with ditch age (P < 0·05) and detritus (P < 0·01); the oldest site, Sedge Fen Drove, had the highest SQS, and North Dyke (3·5 years old) had the lowest. There were no significant correlations between SQS and any environmental variables for either Mollusca or Odonata.

Table 6.  Ditches, ponds and lodes ranked in order of species quality score (SQS) for Coleoptera, Mollusca and Odonata. Also shown is the number of species recorded (S), the number of Red Data Book Species (RDB), the number of water beetles specifically associated with low-lying fen conditions (Fen) (see the Materials and methods), and the number of samples taken at each site
ColeopteraMolluscaOdonata
 nSQSSRDBFenRankSQSRankSQS
Sedge Fen Drove13.891925171.40101.50
Meadow Ditch Q43.873841191.5381.57
Malcarse Drain73.85283522.9081.57
Brick-pit 78e13.62211432.5822.17
Brick-pit 78f13.36252614.3013.00
Meadow Ditch R63.3046315111.47121.43
Mere Ditch63.174721371.60171.25
Commissioners’ Drain32.97382961.62161.33
Wicken Lode32.973224161.40191.00
Brick-pit 78d12.962624141.44101.50
Main brick-pit12.901011181.3871.75
Gardiner's Ditch92.734511052.3142.00
New Dyke42.712414131.4542.00
Cross Dyke52.652624151.4442.00
Thomson's Ditch22.63381981.60141.40
Monks’ Lode12.501811191.38191.00
Drainers’ Dyke42.20291442.54121.43
Wicken Lode dead-end42.203515121.4742.00
North Dyke41.711701101.50141.40
Table 7.  Rank correlations between site SQS for three invertebrate groups and (a) SQS for the other groups; (b) the mean environmental variables for each site; (c) macrophyte abundance and diversity
 Coleoptera
SQS
Mollusca
SQS
Odonata
SQS
  1. Significance levels: **P < 0·01, n = 19.

(a)Invertebrate SQS
Mollusca SQS 0·11 – –
Odonata SQS 0·16 0·34 –
(b)Variable
Depth−0·42−0·02−0·07
Bank profile−0·44 0·13 0·13
Temperature−0·40 0·20 0·47
Conductivity−0·44 0·50 0·21
Oxygen−0·31−0·08−0·41
Nitrate−0·37−0·26−0·06
Calcium 0·16−0·07−0·54
Potassium 0·22−0·03−0·34
Magnesium 0·51 0·29−0·07
Sodium 0·56 0·13−0·02
Alkalinity 0·42−0·32−0·38
Age 0·61** 0·08 0·32
Detritus 0·61** 0·10 0·15
(c)
Macrophytes % cover 0·08 0·04 0·05
Macrophytes no. of species−0·20 0·08−0·38

At least one ‘fen’ species of water beetle was recorded from each waterbody. The Adventurers’ Fen ditches possessed at least nine fen species each, with meadow ditch 96 containing 15. It should be noted, however, that the number of fen beetle species was positively correlated (P < 0·05) with the number of samples taken in each waterbody, as were the numbers of species (P < 0·01) (species quality scores are independent of sample size). For Odonata, the number of species was also correlated with the number of samples taken in each ditch (P < 0·01). This was not true, however, for Mollusca (P = 0·35).

Discussion

The survey reveals a clear distinction between the invertebrate assemblages of the Sedge Fen waterbodies and the ditches of Adventurers’ Fen. Some differences might have been predicted, given the contrasting management regimes to which these two sets of waterbodies have been subjected: the Sedge Fen ditches and the lodes are cleared of accumulated vegetation and detritus at regular intervals, while the majority of Adventurers’ Fen ditches have been relatively undisturbed since the 1950s. The land use in the two areas is also very different: regular harvesting of herbaceous vegetation on the Sedge Fen compared to grazing by cattle on Adventurers’ Fen. Furthermore, the Sedge Fen receives its water from the lodes largely via percolation through the peat body, while Adventurers’ Fen ditches receive water directly from the lodes, which themselves carry run-off from arable farmland. What is more unexpected, however, is the detail of differences in species composition and the distribution of species of high conservation value in the two areas of fen.

The sedge fen

The invertebrate communities of the Sedge Fen fall into two main groups: the lodes and the wide ditches connected to them; and the small ponds and narrow ditches receiving water by percolation and precipitation.

The macrophyte flora of the lodes are rich and distinctive, containing species that do not occur elsewhere on the reserve, e.g. Potamogeton perfoliatus L., P. freisii Rupr and Sparganium emersum Rehm (Friday & Preston 1997). The lodes are navigable waterways, kept clear by slubbing out sediment on a 10-year cycle (with the exception of the dead-end of Wicken Lode, in which vegetation is cut annually). The other large waterbody on the fen, Drainers’ Dyke, also maintains a slow flow, at least in winter, and is cleared out in alternating 50-m stretches on a 4-year cycle. These waterbodies are therefore the most mechanically disturbed aquatic habitats on the fen, but also the most floristically diverse. The faunas of these sites include some beetle species characteristic of running water (such as Laccophilus hyalinus) and slow-flowing drains (e.g. Hygrotus versicolor;Balfour-Browne 1940).

The distinction between the communities of the major watercourses and of the ponds and narrow ditches extends to the benthic fauna. Five species of unionid mussels occur in the lodes (D.C. Aldridge, personal communication) and Gammarus pulex L. is present in the lodes but is replaced by Crangonyx pseudogracilis Bousfield in the ditches (Lee 1988). The wide ditches that group with the lode sites are interconnected, virtually static, and are generally deeper than the internal ditches. Cross Dyke and Malcarse Drain are both cleared out on 8-year cycles and support very similar macrophyte communities; there is no barrier to dispersal between the ditches and their invertebrate faunas are very similar (Painter 1995).

The only pond site within this group, the main brick-pit, is far larger than the other brick-pits, being steep-sided and at least 3 m deep. Its waters stratify from May to September, producing a deoxygenated hypolimnion but a well mixed and oxygenated epilimnion (from which the samples for this survey were taken). This is in sharp contrast to the smaller, shallow brick-pits, in which even the surface waters can experience partial deoxygenation at times during the summer (L.E. Friday, personal communication).

All of the smaller waterbodies of the Sedge Fen are prone to drying out and to wide fluctuations in temperature and chemistry during the summer. In particular, the old Phragmites-dominated ditch along Sedge Fen Drove, which has not been cleared out for more than 100 years, dries out completely in most summers and has a distinctive invertebrate fauna: its beetle fauna is dominated by large dytiscids, including Agabus undulatus, and it has very few molluscs and breeding odonates. The rest of its invertebrate fauna also differs markedly from other sites in the fen: it is the only known site at Wicken for larvae of the nationally scarce caddis larvae Phacopteryx brevipennis (Curtis) (Painter & Friday 1995), a species that requires shallow, temporary aquatic conditions (Wallace, Wallace & Philipson 1990). The other small ditches and ponds are densely vegetated, fringed by Phragmites, and are dominated by Lemna trisulca L.

Adventurers’ fen

Most of the Adventurers’ Fen ditches have not been cleared for 38 years, with the exception of ditches 85 and 82, which were partly re-excavated in 1989 and 1994, respectively. In general these ditches are shallow, with abundant detritus and gently sloping margins which, in some areas, are trampled by cattle. Kirby (1992) and Drake (in press) both recognize the importance of livestock trampled, gently sloping ditches as a habitat that potentially can support a diverse invertebrate fauna. Furthermore, Galewski (1971) concludes that physical and structural aspects of the habitat may be more important than chemical variables in influencing Polish water beetle communities. Nilsson, Elmberg & Sjöberg (1994) also found differences between the invertebrate communities of steep-sided lakes and pools, and shallower waterbodies with gentle profiles, with the latter supporting richer communities of dytiscids. Foster (1995), studying the aquatic macroinvertebrates of peat pools, also found differences in invertebrate community structure between steep-sided deep pools and shallower pools with gentle profiles.

The reasons why sloping margins might influence invertebrate communities are not well understood. However, the range of microclimatic conditions, such as temperature and light, is likely to be greater in a ditch with gentle margins. Shallow margins can also provide areas for rooting of shallow-water plant species that would not be able to grow in steep-sided ditches (Newbold, Honnor & Buckley 1989).

For insect species that leave the water to pupate, a sloping margin may be easier to negotiate than a steep one. Ribera & Nilsson (1995), studying water beetle morphometrics, report that steep margins are the preferred habitat for more spherical, manoeuvrable beetles such as Hyphydrus ovatus and Hygrotus inaequalis, which were ‘habitually found in vegetated ditches with steep margins’ in the Pyrenees. However, in the current study, both species were recorded in equal abundance in both steep-sided and gently sloping ditches on both fens.

The coleopteran faunas of these ditches reflect the unkempt and trampled nature of the sites: Agabus undulatus, a fen species classified as RDB 2 (Shirt 1987), occurred in all of the Adventurers’ Fen ditches, while it was found on the Sedge Fen only in the oldest ditch; many of the other beetle species found primarily on Adventurers’ Fen are ‘detritus pond species’ such as Peltodytes caesus, Hydaticus seminiger and Hydrochus elongatus (Balfour-Browne 1940, 1950, 1958); and Hygrobia hermanii, which was found only on Adventurers’ Fen, is described by Balfour-Browne (1940) as ‘seeming to appreciate ponds frequented by cattle where the water acquires a high ammoniacal content’.

The molluscan faunas, too, appear to be influenced by the amount of detritus accumulating in older ditches: Bathyomphalus contortus is detritivorous (Callow 1973) and may require ditches near the end of the successional cycle (Caspers & Heckman 1981); Oldham (1926) remarked that B. contortus was found ‘in the lesser drains and swampy places’ at Wicken Fen. In contrast, the RDB mollusc Pisidium pseudosphaerium was recorded only from the ditches and ponds of the Sedge Fen that contained less detrital material than the Adventurers’ Fen ditches, and appeared to have increased growth rates in ditch sections that had been recently excavated (Painter 1995).

One further difference may be found between the waters of Adventurers’ Fen and the Sedge Fen: the Adventurers’ Fen sites have higher conductivity, with higher concentrations of Ca, K, Mg and Na than the Sedge Fen and lode sites. Since these ditches receive water direct from the lode, the explanation might be sought in the presence of Gault Clay, which forms a considerable part of the substratum in the Adventurers’ Fen ditches, but lies well below the surface peat over much of the Sedge Fen. It is unlikely, however, that the concentrations of ions found on Adventurers’ Fen are high enough, nor those on the Sedge Fen low enough, to affect invertebrate species’ abundance and distributions directly.

Macrophyte and invertebrate distributions

The significant correlation between DCA axis 1 site scores for macrophytes and invertebrates suggests that macrophyte vegetation structure is likely to be an important influence on the invertebrate assemblage present at a site. Large single-species beds similar to those studied by Downing (1986) are encountered only in the wide, recently excavated ditches associated with the lode systems. In general, narrow ditches, and those dominated by Phragmites, which reduce the level of photosynthetically active radiation (PAR) entering the water column (Painter 1995), possess rather patchy aggregations of a few macrophyte species.

Where chemical conditions are not limiting, the most important factor controlling mollusc species richness and abundance is the amount of macrophyte vegetation (Økland 1990). Macrophytes can provide shelter from predators; a substratum to which to attach eggs; a large surface area for periphyton, on which many snails feed; plant material that may be eaten alive or in a state of decay or as detritus; oxygen released through photosynthesis; reduced turbidity of the water; and a more stable bottom substratum (Økland 1990). Clare & Edwards (1983) found vegetational stage in the hydrosere to be an important factor influencing ditch macroinvertebrate species distributions in the Gwent Levels. Similarly, Scheffer, Achterberg & Beltman (1984) found vegetation pattern to be important in determining the spatial distribution of non-benthic ditch macroinvertebrates. Foster et al. (1990), studying the arable ditches in East Anglia, reported that ‘the extent of management by destroying vegetation through dredging was important in dictating the form of water beetle association’. Ditch excavation patterns have a profound influence on macrophyte vegetation structure (Painter 1995).

Management recommendations

The current study provides support for many of the statements and recommendations on ditch management for invertebrates made by Kirby (1992).

Ditches can support invertebrate communities of high conservation value

The ranked species quality scores for Coleoptera, Mollusca and Odonata highlight the national importance of Wicken Fen's waterbodies as habitats for rare invertebrate species. Foster & Eyre (1992) list the SQSs for water beetles from 521 sites in Lincolnshire and Northern East Anglia: the mean SQS was 1·87, with a maximum of 6·2 (West Mere) and a minimum of 1·0 (51 sites); ‘Wicken Fen’ scored 2·5 and ‘Wicken Fen brick-pits’ scored 2·8. All sites sampled in the present survey, except North Dyke, had SQSs greater than 2, with Sedge Fen Drove scoring the highest (3·89). The significant correlations of site coleopteran SQS with ditch age and detritus content indicate the importance of old ditch habitats as sites for rare water beetles. However, the recently excavated Malcarse Drain had a SQS of 3·85, largely due to the presence of one RDB species, Dryops griseus.

The ranking of waterbodies on the basis of mollusc species’ rarity reflects the distribution of RDB molluscs across the fen, with the first five ranked waterbodies each yielding one RDB species. Malcarse Drain is ranked highly because of the presence of Pisidium pseudosphaerium. The old ditch along Sedge Fen Drove that dries up in the summer is, unsurprisingly, one of the worst sites for Mollusca. This, however, illustrates one of the problems of studying single groups of invertebrates: if the molluscs alone had been surveyed then the ecological ‘value’ of Sedge Fen Drove would have been overlooked (Painter & Friday 1995).

The ranking of ditches in terms of Odonata species’ rarity suggests that the brick-pits and smaller ditches are important habitats for odonate larvae. The low ranking of Wicken and Monks’ Lodes is difficult to explain, as both waterbodies are good sites to observe adult Odonata. However, adult sightings do not always predict larval distribution at Wicken (Painter 1998).

The greater the range of sizes and conditions of a ditch on a site, the greater will be the range of invertebrates the site can support

The current study has shown that the physical characteristics of a site are an important influence on the invertebrate assemblage found there, with ditches on highly disturbed former fen supporting rather different invertebrate assemblages than the more intensively managed waterbodies on the original fen. The physical nature of a ditch, such as bank profile and width, will also influence the site's emergent and submerged vegetation structure, which in turn can influence the site's invertebrate assemblage.

The neglected and cattle-grazed Adventurers’ Fen ditches contained, on average, more species of Coleoptera, Mollusca and Odonata than those on the Sedge Fen, and, in terms of the rarity of their species, were comparable in conservation value. It is also clear from these results that each of the aquatic habitat types, from the lodes and wide open ditches, to old, semi-permanent waters, have distinctive contributions to make to the overall diversity of the fen's fauna.

All stages of the hydrosere support distinct communities of invertebrates

The current management plan for ditch excavation on Wicken Sedge Fen will complete its full clearance cycle in 2004, and will generate a good patchwork of diversity, with a variety of ditches of different ages (Painter 1995). By ensuring that all stages of the hydrosere are represented on the fen at one time, the continued conservation of fen invertebrate species with specific habitat needs will be facilitated.

Should ditches be excavated in their entirety or in short stretches?

Classification of the samples from short (50 m) alternating excavated and unexcavated sections of Cross Dyke and Drainers’ Dyke revealed that the macroinvertebrate assemblages in the excavated sections were typical of the rest of the ditch, rather than of the ditches recently cleared in total (e.g. North Dyke; Fig. 1) This is in spite of the fact that a separate study of the phytophilous invertebrates associated with the common ditch macrophyte Myriophyllum spicatum in an excavated and unexcavated section of a wide ditch has shown that invertebrates were most abundant in the more recently excavated section, where plants were growing vigorously. In addition, notable differences in invertebrate composition on plants in the two sections were also recorded (Painter 1995). Another study of the biological and physical effects of excavating a narrow ditch in short stretches revealed that mechanical excavation directly removes benthic invertebrates, and significantly affects physical (bank profile) and biological (primary production) conditions (Painter 1995). For rare species with low powers of dispersal and specific habitat requirements, wholesale ditch excavation could result in local extinction. For example, it has been shown that the RDB water beetle Hydraena palustris favours sites in the middle of their hydroseral development (Painter & Friday 1995); by excavating the ditches where it occurs in short stretches, populations of H. palustris can be maintained from which recently cleared stretches can be colonized when conditions are suitable (Painter 1995). Kirby (1992) adopts a similar cautionary approach to ditch excavation, and recommends that only one side of the ditch should be managed at one time.

Acknowledgements

I am grateful to Richard Preece for taxonomic assistance and Garth Foster and Michael Kerney for checking invertebrate identifications. In addition, I thank Laurie Friday, Norman Moore and Montserrat Real for their helpful comments. The study has been funded by a Balfour Studentship; I am also grateful for financial assistance from Emmanuel College and the A. J. Keith Fund.

Received 20 June 1997; revision received 17 November 1998

Ancillary