Changes in decomposition rates and Collembola densities during the forestry cycle in conifer plantations


J. Butterfield (fax: 01913742417; e-mail:


1. Conifers have a tendency to produce detrimental changes in the soil that suppress decomposition. It is also possible that clear-felling and the short plantation rotation (40–50 years), usual in Britain, will eventually cause nutrient depletion. Decomposition rates are important in determining the level of nutrients available, to both the crop and the soil invertebrate community, and this study has compared them at three stages during the plantation cycle.

2. Cotton cloth buried in the soil of open areas decomposed at twice the rate of cloth buried in the soil of closed canopy plantations.

3. The rate of decomposition in the soil layer 0–80 mm from the surface was 1·5 times that in the 81–160 mm layer.

4. In young plantations, the decomposition rate of cloth decreased with an increase in the age of the plantation. It is suggested that the decrease in decomposition rate with plantation age is associated with increased shading and a decrease in the amplitude of the diurnal temperature cycle.

5. Springtail (Collembola) densities were high in spring on an open area and decreased in summer. The opposite occurred in the closed canopy plantation. Drying on the exposed open area is suggested as the cause of the decreased summer densities.

6. On the site where Collembola densities were monitored, decomposition rates were higher in the upper soil layer of the closed canopy plantation than in an open area. It is probable that this was due to the upper layers on the exposed area drying, rather than to the high Collembola densities checking fungal decomposition.

7. The nutrient release that accompanies decomposition has potential importance for the rate of growth of the conifer crop. This preliminary study suggests that the decomposition rate is highest in the years immediately after felling, probably favouring the establishment of the next rotation.

8. This study has not detected suppression of decomposition on open or clear-felled areas in Hamsterley Forest (NZ0530), suggesting that present planting and harvesting methods are appropriate and should continue. However, further investigation of the interactions between fungi, meso- and macrofauna is needed to understand both nutrient cycling and the invertebrate community structure in plantation forestry.


Between 1919 and the present day, Forestry Commission planting has expanded woodland cover of the land surface of the British Isles from 5% to 10% (Forestry Commission 1994). Most of this planting is coniferous, predominantly of Sitka spruce Picea sitchensis Bong., and the bulk of planting has been in the uplands, on marginal farmland or moorland. Picea sitchensis plantations probably support a higher density and diversity of animal species than the upland pasture or moor that they replace (Moss, Taylor & Easterbee 1979; Newton 1986; Ratcliffe 1986; Young 1986). However, it has been argued, usually with reference to birds and mammals, that plantation forestry favours common woodland species of little conservation importance while adversely affecting some of the rarer species of open moorland (Thompson, Stroud & Pienkowski 1988).

One of the aspects of plantation forestry that promotes diversity in invertebrate communities is the forestry cycle itself. In Britain, plantation blocks are usually harvested when they are between 40 and 50 years and the present policy is to produce a patchwork of different aged blocks across a forest (McIntosh 1995). In P. sitchensis plantations, stages in plant succession grade from virtually unvegetated needle litter at clear-felling through to well developed herb or shrub layers. These are then suppressed by shading at canopy closure, about 15–20 years after planting. Depending on the thinning regime, there may be little ground vegetation, apart from sparse moss cover, from about 20 years after planting until felling. The habitats presented during succession are diverse and previous studies indicate that they also support a high diversity of insects (Welch 1986; Young 1986; Buse & Good 1993; Butterfield et al. 1995). At the soil surface, the high densities of predators such as ground beetles in young conifer plantations on peaty soils, 50–250 m−2 (Butterfield 1997) as opposed to 1–3 m−2 on treeless, blanket bog (Coulson & Whittaker 1978), suggest that the early regrowth stages of the forestry cycle, in particular, can provide productive habitats for invertebrates.

There is, however, anxiety over the sustainability of plantation forestry with the possibility that clear-felling will, in the long term, cause nutrient depletion (Anderson 1985). Conifers also have a tendency to cause detrimental soil changes, such as a decrease in pH, accompanied by a rise in aluminium concentrations, and podsolization (Gee & Stoner 1988). These conditions may lead to decomposition and mineralization rates in conifer woodland being lower than in soils under deciduous trees (Heal 1979; Miles 1986). In northern Finland there have been failures in reforestation (Huhta 1976) and it is suggested that these are the result of the formation of an organic mat that inhibits seedling establishment. Further, as nutrients are immobilized in the organic layer, growth in established trees also slows down (Miles 1986).

In the acid soils of upland conifer plantations in Britain, the soil macrofauna, such as earthworms, woodlice and millipedes, which contributes to decomposition by comminution of plant material, is largely absent and decomposition is predominantly microbial (Swift, Heal & Anderson 1979). The abundant mesofauna, dominated by mites and springtails, probably influences the decomposition rate more through grazing on fungal mycelia than by breaking down organic material (Swift, Heal & Anderson 1979; Seastedt 1984). For the greater part of the conifer afforestation cycle, the field layer is absent and detritivores and mycophages, such as springtails, form the main prey of many spider and beetle species (Clark & Grant 1968; Hengeveld 1980; Loreau 1983; Buse & Good 1993). The decomposition system not only contributes to tree growth by releasing nutrients, it also forms the trophic base supporting the soil fauna.

One aim of the present study was to compare decomposition rates at different stages in the conifer afforestation cycle to provide information on nutrient release in relation to tree growth. Decomposition rates of cellulose have been compared in open areas (1–5 years after tree removal), in young plantations (9–18 years after planting) and in closed canopy plantations (46–62 years after planting).

A secondary study comparing the densities of Collembola in open areas and in a closed canopy plantation was undertaken as a preliminary investigation of the interaction between microbial and invertebrate components of the soil community. At high densities, collembolan grazing can reduce the fungal standing crop to the extent that the decomposition rate decreases (Verhoef & De Goede 1985). Further, a significant positive correlation between springtail density and carabid density has also been demonstrated (Ernsting, Isaaks & Berg 1992). Despite the authors’ suggestion that the latter relationship was probably a mutual response to other environmental variables rather than that of predator to prey, collembolans provide a link between decomposition processes and the surface-active predators. In conifer plantations it is likely that they are a particularly important food resource for predators in soil and litter layers.


Hamsterley forest

The study was carried out in Hamsterley Forest (NZ0530), a Forestry Commission plantation in County Durham, north-eastern England. The forest is largely planted with conifers and covers c. 2000 ha lying between 150 and 400 m a.s.l., with moorland to the north and rough pasture to the south. The soils vary from comparatively fertile mineral soils to shallow peat over sandstone, with pockets of deep blanket peat. First rotation planting began in 1929, continuing until 1970, and felling began in the 1980s. Replanting is usually undertaken within 2 years of felling, although some areas have been left to natural regeneration.

Study sites

In 1989, decomposition comparisons were carried out in Sitka spruce plantations on peat and peaty podsols at three sites (Sites 1–3). Each site was chosen on the basis that an open area (clear-felled, wind-blown or replanted within the previous 5 years) lay adjacent to an old, closed canopy plantation (46–62 years) and to a young plantation (9–18 years). In 1991 a further comparison of decomposition rates was made between two adjacent clear-felled areas and a closed canopy plantation at a fourth site. Collembola were sampled on the clear-felled areas and in the old plantation at Site 4 in 1991 and 1992. Table 1 summarizes the soil and vegetation characteristics in each of the plantations at the four sites.

Table 1.  Characteristics of the three plantation stages (open, young and old) at four sites in Hamsterley Forest (CF = clear-felled, RP = replant, WB = wind-blown, with year indicated)
Plantation stageYears from tree loss/plantingSoil organic content (%)Soil pHLoose litter depth (mm)Vegetation % cover
Site 1 Open, WB901914·84210
Site 1 Young18334·23050
Site 1 Old49504·476<2
Site 2 Open, RP85584·2097
Site 2 Young13155·3591
Site 2 Old62814·226 4
Site 3 Open, RP855244·5095
Site 3 Young9274·2293
Site 3 Old57404·120 2
Site 4 Open, CF892814·22060
Site 4 Open, CF901414·725 5
Site 4 Old46724·945<2

Soil organic content and pH values in Table 1 are the mean measurements of five samples taken in the ‘A’ layer (sensuTakeda 1995) in each plantation. The organic content was calculated as the mean percentage loss of weight of sample (dried to constant weight at 90 °C) after 8 h at 400 °C. The pH values were determined for a 1 : 5 soil : water mixture.

The forestry cycle in relation to ground flora and shading

In the old plantations where canopy closure was complete, light levels were so low that vegetation under the trees was almost absent, consisting of 1–2% cover of bryophytes. The loose litter layer was 20–76 mm in depth (Table 1). In the first summer after tree removal, cover remained low on the open sites (5% on Site 4 CF90, clear-felled in 1990, and 10% on Site 1 WB90, wind-blown in 1990) but the vegetation was supplemented by the arrival of ruderals such as Chamerion angustifolium (L.) J. Holub and seedlings of species such as Calluna vulgaris (L.) Hull, surviving in the seed bank. The litter layer had decreased and plant cover increased to 60%, 2 years after felling, on Site 4 CF89. In the absence of brash, vegetation cover was about 95% 5 years after felling on the open replanted areas, Site 2 RP85 and Site 3 RP85. Plant species composition depended on soil type and previous land-use. Grasses were dominant on the clear-felled areas at Sites 2 and 3 and Calluna vulgaris was dominant on Sites 1 and 4.

As well as reducing light levels, the closed canopy of old plantations intercepts rainfall and reduces diurnal temperature fluctuation (Miles 1986). The development of the herb layer during the early stages of succession also modifies temperatures below ground, with temperature differences similar to those outside and inside the forest recorded in the soil below a sparse and a dense grass sward, respectively (Geiger 1965).

Measurement of decomposition rates

Decomposition rates were determined as the loss in tensile strength of cotton strips after burial in soil. Shirley Test Cloth (Walton & Alsopp 1977) was cut into 300-mm strips, frayed to 60-mm width, washed and autoclaved and inserted vertically into the soil, as described by Latter & Howson (1977). In some plantations, the underlying gritstone was near the surface and it was not possible to insert the strips beyond 200 mm. Therefore decomposition was measured in two depth zones only: 0–80 mm and 81–160 mm from the litter surface. In the closed canopy plantations, a large part of the first depth zone was occupied by litter, while in the clear-felled areas and young plantations the depth of litter was smaller and the upper section of the strips reached further into the soil below. At all sites, the lower depth consisted of a more homogeneous peaty soil, although the mineral content varied between sites (Table 1). In 1989 strips were in place from 11 May to 7 July, and in 1991 from 10 June to 4 August (57 and 55 days, respectively). During the periods the strips were in the ground, the mean air temperature was 12·1 °C and the rainfall was 57·6 mm in 1989, and 14·8 °C and 75·6 mm in 1991 (recorded at Durham University Observatory, 25 km to the north-east of Hamsterley).

At retrieval, the position of the soil, or compacted litter, surface was marked on each cloth strip. On return from the field each strip was washed and dried and marked at 80 mm and 160 mm from the soil surface. The rate of decomposition in the two depth zones was assessed by measuring the tensile strengths of the two sections of each strip by tensometer (type E; Monsanto plc, High Wycombe, UK). All strips were cut from the same batch of cloth and average tensile strength of the strips before decomposition was calculated from 16 control strips, reserved from burial but otherwise processed with each batch of experimental strips.

Collembola sampling

Site 4 was sampled for Collembola by 10−3-m cores, taken to a depth of 60 mm, divided into 30-mm segments and extracted in a high gradient extractor (Macfadyen 1961) in spring (May–June), summer (July) and autumn (October) in 1991 and in April and August in 1992. The pressure exerted when taking a core compacted the loose litter so that even in the old plantation (litter depth 45 mm) the core extended at least 30 mm into the ‘A’ layer. The extraction method followed Hale (1966), with a 3-day extraction period.

Statistical analysis

Analyses of variance for differences in decomposition rates were carried out on the untransformed tensile strength measurements of the buried cloth sections. The results are tabulated with the decomposition rates, presented as mean percentage loss in strength, based on subtraction of the tensile strengths of the buried sections from the mean of the reserved control sections.

Collembola densities were log-transformed before analysis of variance.



Despite selection of plantations at different stages in the forestry cycle in adjacent blocks at each site, there was variation in organic content of the soil and pH within sites and between sites (Table 1). A preliminary partial regression was carried out on the mean decay rates in each plantation in relation to plantation age and soil pH. The mean decomposition rates of sections of cloth in the upper, 0–80 mm, and lower, 81–160 mm, soil depths at each site showed a significant relationship between plantation age and decay rate (% loss in strength of the cloth), but no significant effect of pH (Table 2). Therefore pH was not used as a variable in further analyses.

Table 2.  Partial regression results showing the relationship between mean decomposition rates at two depths in the soil, plantation age and pH
Depth (mm)VariablePartial regression coefficientSEt9P
0–80Plantation age−0·4110·1283·21<0·02
 pH 7·387·770·95NS
81–160Plantation age−0·5120·1882·72<0·05
 pH 3·0411·400·27NS

Table 3 shows the mean percentage loss in strength of cloth buried at two depths in the plantations at Sites 1, 2 and 3 for three stages of the forest growth cycle. Two-way anova, based on the individual breaking strengths, of each cloth segment indicated significant differences between decomposition rates in plantations of different ages and at the two depths below the soil surface (F2,210 = 54·5 and F1,210 = 59·8, respectively, P < 0·001 in both cases). The average loss of strength declined from 62% in the open areas (clear-felled, wind-blown or replanted) to 39% in the young plantations to 32% in the old plantations, while, overall, the sections of cloth in the upper soil layer lost 54% strength compared with 35% at the lower depth. However, the anova also indicated significant interaction between plantation stage and depth of burial on the decomposition rates (F2,210 = 5·2, P < 0·01). The percentage loss in strength for the cloth in the upper 80 mm of soil showed progressive increases from old plantation to young plantation to open areas, whereas at the 81–160 mm depth there was no difference between young plantations and old plantations (Table 3). The decrease in decomposition rate with soil depth was greater in the young plantations than in the more open areas. As there was considerable heterogeneity between the ages of the young plantations, ranging from 9 years, where the habitat was still relatively open, to 18 years, where the canopy was virtually closed, each site was compared separately for each depth.

Table 3.  Percentage loss in tensile strength of cloth buried in soil at two depths below the surface in conifer plantations at three stages in the forestry cycle
Percentage loss in strength
Depth below surface (mm)Open (1–5 years)Young plantation (9–18 years)Old plantation (46–62 years)

The percentage decomposition losses in the sections of cloth in the upper soil layer on each site are shown in Table 4, and the losses of the sections at lower depth in Table 5. The results from Site 4, where clear-fell and old plantation stages only were compared, are included in each table. One-way anovas comparing the loss of tensile strength in the upper cloth sections at Sites 1–3 indicated very similar decay rates in the three old plantations (39–41%) and broadly similar rates in the open areas (60–73%) but a much greater variation in the young plantations (40–73%). Overall, the loss in tensile strength of the upper cloth sections in the open areas was 1·7 times that in the old plantations. Tukey HSD tests showed that the rates of decomposition in Sites 1–3 were significantly faster in the open areas than in closed canopy plantations. In the young plantations, the decomposition rate in the 9-year-old plantation (Site 3) did not differ significantly from that in the recently replanted area, whereas in the 18-year-old plantation (Site 1) the decay rate was almost the same as that in the old plantation. In the 13-year-old plantation (Site 2) the rate was intermediate and not significantly different from either the old plantation or the replanted area. This sequence suggests a gradual decrease in decay rate with the increase in plantation age between 9 and 18 years.

Table 4.  Mean percentage strength lost in cloth buried 0–80 mm below the soil surface for 55 days in different aged conifer plantations. anova results are based on the individual breaking strength of the samples. CF = clear-fell, RP = replant, WB = wind blown, YP = young plantation, OP = old plantation
% strength lost
Plantation stage
SiteOpenYoungOldd.f.FPTukey HSD test
  • *

    CF90, clear-felled 1990;

  • †CF89, clear-felled 1989.

17040412,3611·4<0·001WB ≠ OP = YP
26054392,397·3<0·01RP ≠ OP
37373392,3333·6<0·001CF = YP ≠ OP
Table 5.  Mean percentage strength lost in cloth buried 81–160 mm below the soil surface for 55 days in different aged conifer plantations. anova results are based on the individual breaking strength of the samples. Plantation stage indicated as in Table 3
% strength lost
Plantation stage
SiteOpenYoungOldd.f.FPTukey HSD test
  • *

    CF90, clear-felled 1990;

  • †CF89, clear-felled 1989.

15918112,3737·9<0·001CF ≠ YP = OP
25618322,3915·5<0·001RP ≠ OP = YP
35352212,3229·4<0·001RP = YP ≠ OP
455* 222,367·7<0·001CF90 ≠ OP

At the lower depth, 81–160 mm below the surface, the rate of decay in the open areas at Sites 1–3 was 2·6 times that in the old plantations, with the differences between the rates of decomposition being significantly higher in all three cases. In the young plantations, the decomposition rate in the 9-year-old plantation (Site 3) did not differ significantly from that in the open area but the rates in the 13- and 18-year-old plantations (Sites 2 and 1, respectively) were similar to those in the closed canopy plantations.

Figure 1 shows the mean percentage decay rate (loss in tensile strength) of the cloth in upper (1–80 mm) and lower (81–160 mm) soil layers in the plantations. The relationship was interpreted as decay rate declining with plantation age between clear-felling and canopy closure and thereafter remaining at the same level, indicated by the superimposed trend lines. The relationships between decay rate and plantation age up to canopy closure were:

Figure 1.

The mean percentage loss in tensile strength of upper (1–80 mm) and lower (81–160 mm) pieces of cotton strip placed vertically in the soil. The trend lines suggest a decreasing rate of decomposition up to canopy closure (y = 76·4 − 1·73x for upper sections and y = 66·9 − 2·85x for lower, where x is plantation age) and an unchanged rate thereafter.

y1 = 76·4 - 1·73x, r4 = 0·82, P < 0·05
y2 = 66·9 - 2·85x, r4 = 0·91, P < 0·05

where y1 and y2 are the percentage loss in tensile strength of cloth in upper and lower soil layers, respectively, and x is plantation age.

On Site 4, where decomposition rates were monitored in relation to Collembola sampling in 1991, the magnitude of the differences between clear-felled areas and the old plantation was not as marked as in 1989. The decomposition rates at Site 4 were measured a month later in the year than Sites 1–3. The mean temperature was higher over the period of the decomposition experiment and rainfall was lower than average in July. Higher temperature may have favoured the decomposition rates in the shade of the old plantation, while drying of the surface peat had an adverse effect on decomposition rates on the clear-felled areas. The decomposition rate in the upper depth layer in the old plantation at Site 4 was about a third higher than in the other old plantations, although decomposition at the lower depth was similar, and the decay rates at both depths in clear-fell CF89 were lower than in all the other open areas. The decay rates in the upper sections of cloth from clear-fell CF90 were similar to those in the other clear-fell sites, despite not being significantly different from those in the old plantation, and decay rates at the lower depth in CF90 were more than twice those in the old plantation, differing significantly (Tukey HSD test). It is possible that the slightly deeper litter layer in CF90, compared with CF89 (Table 1), protected the layers below from drying out.

Collembolan densities

Annual mean densities of Collembola on the two clear-fell areas and in the closed canopy plantation at Site 4 did not differ significantly in either 1991 or 1992, and densities on all three areas were higher in spring 1992 than in 1991, with a geometric mean density of 58 600 m−2 in 1991 and 184 200 m−2 in 1992 (F1,66 = 17·4, P < 0·001). Comparing spring 1991 and 1992, the differences between the clear-fells and the old plantation were not significant but interaction between the area and the year was (F2,66 = 3·37, P < 0·05). The seasonal sequence on the three areas is shown in Table 6. Significant differences between areas in spring and summer 1991 and in summer 1992 resulted from a significant summer increase in collembolan density in the closed canopy plantation in 1991 (t38 = 3·88, P < 0·001) coupled with significant summer decreases on CF89 in both years (t38 = 2·98, P < 0·002 and t22 = 2·47, P < 0·02, respectively). Table 5 shows that in 1991 spring densities on CF89 were higher and summer densities lower than in the old plantation, and in summer 1992 densities on both clear-felled areas were lower than in the old plantation (Tukey HSD test). Figure 2 shows the seasonal variation in densities on the clear-felled sites as a percentage of the density in the closed canopy plantation, and it can be seen that in the second year after felling, in 1992, CF90 showed a seasonal pattern similar to that of CF89.

Table 6.  Geometric mean Collembola densities (m−2) at Site 4 in 1991 and 1992; comparisons are between stages in the forest cycle and between successive seasons
SeasonClosed canopy plantationClear-fell 1989Clear-fell 1990d.f.FPTukey HSD test
  • *

    Comparison with preceding season: + significantly higher, − significantly lower, P < 0·05; ++ significantly higher, −− significantly lower, P < 0·005.

1991 Spring37 300
139 800
80 600
2,675·75<0·005CF89 ≠ OP
Summer141 800
43 600
91 900
2,574·17<0·05OP ≠ CF89
Autumn52 500
45 800
50 800
1992 Spring158 800
162 200
241 900
Summer177 10040 300139 1002,320·71<0·02OP ≠ CF90 ≠ CF89
Figure 2.

Seasonal variation in density of Collembola at Site 4; OP = Old plantation; CF89 = clear-felled 1989; CF90 = clear-felled 1990.

Relationship between collembola densities and decomposition rates

During the period that the decomposition rates were measured on Site 4, from June to August in 1991, Collembola were sampled in May–June and July. The spring densities of collembolans on CF89 were almost four times higher than in the old plantation and it is possible that the reduced decomposition rate in the clear-fell was due to grazing pressure on the fungi. However, collembolan density decreased on CF89 between the May–June and July sampling period with the result that the average densities, on the two clear-felled areas and in the old plantation, over the spring–summer period were not significantly different.


The increase in decomposition rates found soon after clear-felling in the present study agrees with Huhta's (1976) observations in Norway spruce Picea abies (L.) Karst. forests in Finland. In Finland, decomposition rates were significantly higher in clear-fells, from 2 to 12 years after felling, than in old plantations. Huhta attributed the increase, which was accompanied by an increase in total soil community respiration rates, to the higher day-time temperatures encountered at the soil surface in the clear-felled areas. In soils that retain adequate moisture it has been demonstrated that raised temperatures increase the rate of litter decay (Hobbie 1996), and Kochy & Wilson (1997) have demonstrated experimentally that artificial shading reduces decomposition rates on open prairie to the level of those found in adjacent aspen forest. With the exception of clear-fell CF89 at Site 4, in the present study decay rates in plantations between 1 and 5 years after tree loss were higher than in the closed canopy plantations. Decomposition rates at both the 0–80 mm and 81–160 mm depths in the relatively open 9-year-old plantation were similar to those in the 1–5-year-old plantations, while the rates in the 13- and 18-year-old plantations did not differ significantly from those in the old plantations. However, comparison between the decay rates of the upper cloth sections in the 18-year-old and 13-year-old plantations indicated a slower rate in the older plantation, suggesting a graded response to increased shading. Although it is not possible from the present study to define the response curves of decomposition rate to plantation age precisely, the data are consistent with a progressive decline in decomposition rate between clear-felling and canopy closure, with no change thereafter. It is probable that the increase in cover of the herb layer, as well as shading by trees, contributed to the modification of the temperature regime (Geiger 1965) leading to a decline in decomposition rate.

The decomposition rates of cellulose strips used in the present study do not necessarily reflect the complexity of decomposition processes at forest floor level. However, the raised rate of cellulose decomposition after clear-felling, together with the observed, relatively rapid disappearance of the accumulated litter, indicated enhanced decomposer activity and suggested that this is an important stage for nutrient cycling and release. Other studies have also shown that clear-felling leads to nutrient release, with increased concentrations of ammonium, nitrate and phosphate ions in surface water run-off (Likens, Bormann & Johnson 1969; Vitousek & Melillo 1979; Stevens et al. 1988). At Hamsterley, the decomposition rate was not lower on the wind-blown area at Site 1 compared with the other open sites, despite the high organic content (91%) and in the absence of soil cultivation. It seems likely that decomposer response under the climatic regimes encountered in most areas of plantation forestry in Britain will be adequate to remove the accumulated needle litter after clear-felling, even on peat soils. Unlike the situation in Finland (Huhta 1976), the inhibitory litter layer is likely to persist for a short time only and recycled nutrients will be available for a further rotation of trees. Moreover, as Forestry Commission plantations are usually replanted rather than left to regenerate, nutrients will become available when soil preparation breaks up and aerates the surface and redistributes minerals from the lower soil layers. Huhta (1976) noted short-term increases in collembolan densities after spruce in southern Finland had been clear cut, the differences disappearing after 4 years. However, in the present study, mean annual densities of Collembola on clear-fells and in the old plantation were not significantly different. Takeda (1987) has shown that habitat biomass is the most important determinant of the number of Collembola present, with a linear relationship between log numbers extracted and the depth of organic soil and, in general, the density of the mesofauna is inversely correlated with decomposition rate (Swift, Heal & Anderson 1979). In summer 1991, the density of carabids on CF89 was high, reaching a peak of 129 m–2 in the autumn (Butterfield 1997). There is a possibility that fungal activity was reduced by the high spring densities of collembolans, which were then suppressed by carabid predation in the summer. However, Takeda (1987, 1995) found summer drought to be a limiting factor for the Collembola community and in the present study the summer fall in density found on clear-fell CF89, but not in the old plantation, was more probably the result of the drying of the exposed upper litter and humus layers of the clear-fell rather than predation pressure from carabids. It is also likely that the low decomposition rates in CF89 were due to drying rather than high spring densities of mycophagous Collembola inhibiting the rate of cellulose decomposition. Equally, the high decomposition rates in the old plantation, relative to those in closed canopy plantations in 1989, were probably the result of the higher temperatures during the period of decomposition in 1991.

As the years immediately after clear-felling are a key period for the recycling of nutrients within plantation forestry, the effects of abiotic factors and the interactions of decomposers, detritivores and their predators at this stage are important. There has been a large number of studies on decomposer organisms (Swift, Heal & Anderson 1979), on the interaction between decomposition and mesofauna (Crossley 1977; Anderson, Rayner & Walton 1984; Seastedt 1984) and on the contribution of the fauna to nutrient recycling (Petersen & Luxton 1982; Verhoef & De Goede 1985; Brown 1992), but little on the interaction between predators and detritivores or mycophages and the consequences for decomposition. From the point of view of forest management, the factors that can contribute to the failure in litter breakdown after clear-felling need to be evaluated. The Gisburn experiment suggests that the growth of Norway spruce can be improved by interplanting with pine, and that this is due to improved organic matter turnover brought about by increased microbial and faunal activity at the forest floor (Chapman, Whittaker & Heal 1988; Brown 1992). Miles (1981, 1986) has shown that, in general, Norway spruce litter tends to cause a decrease in pH on poorly buffered soils, whereas birch Betula pendula Roth and B. pubescens Ehrh. agg. increases it. Mixed species planting, or tolerance of birch regeneration within replanted areas, could promote conifer growth in climatic zones where temperature does not favour rapid decomposition. However, McTiernan, Ineson & Coward (1997) have pointed out that further work is necessary to elucidate the interaction between leaf species mixtures and fungal activity. In particular, they suggest that where their microcosm results are at variance with Chapman, Whittaker & Heal (1988) the faunal component of the forest floor could be an important influence mediating decay rates. The observations on springtails in the present study are too limited to allow conclusions to be drawn but they suggest the possibility that the interaction between predators and mycophages, as well as that between mycophages and fungal activity, could influence decomposition rates. The decomposers regulate nutrient release, which is reflected in tree growth, and also contribute to and sustain the greater part of the soil fauna in conifer plantations. As these plantations now cover 10% of the land surface in Britain (Forestry Commission 1994) they contribute substantially to the available habitats.

At present, under the climatic regime in northern England, rapid decomposition of the Sitka spruce litter layer occurs after clear-felling. Further, soil preparation for planting redistributes leached minerals and nutrient depletion seems unlikely to become a major problem. However, greater understanding of the interactions between decomposers and mycophages and between mycophages and their predators is needed both for long-term forest management and for invertebrate conservation.


I am very grateful to the Forestry Commission for permission to work in Hamsterley Forest, especially to Gordon Simpson and Brian Walker for their enthusiasm, to Eric Henderson for operating the tensometer, and to Val Standen and John Coulson for commenting on earlier drafts of this paper.

Received 19 November 1997; revision received 5 December 1998