Earthworms influenced by reduced tillage, conventional tillage and energy forest in Swedish agricultural field experiments

Abstract

We compared earthworm density, depth distribution and species composition in three soil cultivation experiments including the treatments ploughless tillage and mouldboard ploughing. Sampling was done in September 2005 and for one experiment also in 1994. By yearly sampling 1995–2005, earthworms in an energy forest of Salix viminalis were compared with those in an adjacent arable field. Sampling method was digging of soil blocks and hand sorting and formalin sampling in one cultivation experiment. Both methods were used in the energy forest and arable land comparison.

In two soil cultivation experiments, highest abundances or biomass were found in ploughless tillage. Earthworm density was higher in the upper 10 cm, especially in the ploughless tillage. Earthworm density was significantly higher in the energy forest than in the arable field. Formalin sampling revealed c. 36% of the earthworm numbers found by digging in the energy forest and gave almost no earthworms in the arable field. In all treatments with soil cultivation, species living and feeding in the rhizosphere and soil dominated. One such species, Allolobophora chlorotica, was more abundant under mouldboard ploughing than ploughless tillage. Lumbricus terrestris, browsing on the surface and producing deep vertical burrows, was more common in the ploughless tillage. Species living and feeding close to the soil surface were almost only found in the energy forest, which had not been soil cultivated since 1984.

The findings support earlier studies pointing out possibilities to encourage earthworms by reduced soil cultivation. This is one of the first published studies that followed earthworm populations in an energy forest plantation during several years. Explanation of earthworm reactions to management and environmental impacts should be done with consideration of the ecology of species or species groups. Earthworm sampling by formalin must always be interpreted with caution and calibrated by digging and hand-sorting sampling.

Keywords:

• Ecological groups of earthworms
• energy forest plantation
• field sampling methods
• Salix
• soil cultivation

Introduction

Earthworms contribute to several aspects of soil fertility in agriculture. By ingestion and mixing of crop residues and soil they enhance decomposition and mineralization and increase soil aggregate stability and thereby nutrient uptake by plants (Edwards and Lofty 1972, Lee 1985). By creation of burrows that may reach down to far beyond plough depth they increase water infiltration and root growth (Clemens et al. 1991). Their activities may also contribute to soil health by disfavouring plant parasites and diseases (Friberg et al. 2005). Because of these documented ecosystem services delivered by earthworms, there are good reasons to develop cropping systems that conserve and promote their abundance and species diversity. Addition of organic matter to the soil from crop residues and amendment with manure and other organic residues will increase population densities, while intensive soil cultivation will in most cases drastically reduce earthworm densities (Boström 1988, 1995, Curry 1998, Curry et al. 2002). Great negative influence is reported from some agricultural systems (Pfiffner and Mäder 1997).

Reduced tillage is nowadays practiced to a growing extent. The reason is saving fuel, conservation of soil organic matter and water, maintaining soil structure and protection of nearby surface waters. It may also spare earthworms from mechanical destruction or from being exposed to predators when carried to the surface by ploughing (Cuendet 1983). During years of perennial crops in the rotation, earthworm numbers will also have a chance to increase, especially those species that are living close to the surface (epigeic species) and those that develop permanent channels from the surface to depths beyond the plough layer (anecic species), while species living and feeding in the soil (endogeic species) are less negatively affected (Lagerlöf et al. 2002). Long-lasting perennial crops, like energy forest of Salix, would give the earthworm community a chance to develop under rather undisturbed conditions for decades. When searching the literature, very few publications dealing with effects of Salix plantations on the earthworm fauna were found (Friis et al. 1999), while there is a rich literature on effects of other perennial crops, fallow and successions with woody plants (e.g., Decaëns et al. 2003, Gormsen et al. 2004).

In order to study effects of agricultural practices of earthworms one needs extensive and laborious field samplings. The most reliable sampling method is digging out soil blocks and hand sorting them, possibly followed by wet sieving in order to retrieve the smallest specimens and the cocoons (Schmidt 2001). Especially in humus soils, heat extraction in Tullgren funnels possibly after hand sorting is an efficient method for small specimens of earthworms (Malmström et al. 2009). These methods mentioned above are very laborious and time consuming and therefore other methods have been introduced, such as pouring a chemical solution over the soil that irritates the earthworms (vermifuge) and thereby forces them to the surface, where they can be collected. The most used vermifuge is formalin (Raw 1959), but also mustard and washing-up detergent can be used, which are more environmentally friendly and less hazardous to handle (Gunn 1992, East and Knight 1998). Exposing the earthworms to electricity is also an option (Thielemann 1986) that forces the earthworms to the surface. While the digging method can give almost 100% yield if sampling depth is enough, these other methods give less and the results will depend on several factors, e.g. soil humidity and structure, and therefore such methods need to be calibrated against the digging method in order to give comparable results.

The aim of this paper is to analyse the effects of reduced tillage on earthworm populations. This was studied with help of sampling in a number of Swedish long-term field experiments grown with annual crops. We also present the results of a long-term study comparing earthworm populations in energy forest soil with that under annual arable crops. For sampling, the digging method was used in most cases and in some cases formalin and sometimes both methods. Because of this, a comparison of the efficiency of these methods can be done.

The hypotheses tested were: (1) Reduced tillage will enhance earthworm abundance and species diversity (2) Long-term perennial crops, like energy forest on agricultural land, will increase earthworm abundance and species diversity even further than reduced tillage in annual crops (3) Different methods for earthworm sampling will influence the extraction efficiency differently in different treatments and thereby the results and conclusions.

Materials and methods

Site descriptions

During September 2005, earthworm populations were sampled in two field experiments with reduced tillage in southern Sweden (Charlottenlund and Väby) and one field experiment in south-central Sweden (Ultuna). In Ultuna, the same experiment was also sampled in late August 1994. During 1995 to 2005, the earthworm population development in an energy forest plantation and in an adjacent arable field in Ultuna was followed by sampling in May every year except in 1996 and 2000.

The soil cultivation field experiments at Charlottenlund (55^o25′N, 13^o40′E, sandy loam) near the town Ystad and Väby (56^o58′N, 12^o41′E, sandy loam) near the town of Falkenberg were started in 1994. Experimental set up of the two field experiments is identical. Treatments are randomized in three blocks. Plot size is 12 m×20 m. Sampling was made in the treatments: A = ploughless tillage and B = mouldboard ploughing.

Before the sampling for earthworms in 2005, the soil of treatment A at Charlottenlund had not been mouldboard ploughed since 2000 and at Väby not since 2001. The primary tillage in treatment A at Charlottenlund was carried out with a disc harrow to a depth of 5–10 cm. Cultivation in treatment B was carried out by sweep share mouldboard ploughing to a depth of approximately 20 cm. Seedbed preparation was done with a roller with crosskill rings (Väderstad Rollex). At Väby, primary tillage in treatment A consisted of two passes with a disc harrow. Mouldboard ploughing in treatment B was carried out to a depth of approximately 20 cm with the attachment of a furrow packer and no special seedbed preparation was done. Soil cultivation in both treatments, in Charlottenlund as well as in Väby, was done in August–September 2004.

Treatments A and B were managed identically regarding sowing, fertilization and crop protection. Crop residues were left in the field and incorporated by the tillage. The crop in 2005 at Charlottenlund was winter wheat (Triticum aestivum L.) and the preceding crop was spring barley (Hordeum vulgare L. var. disticon). The crop in 2005 at Väby was spring barley and the preceding crop was rye wheat (Triticale utile L.).

Ultuna (59^o49′N, 17^o43′E, heavy clay) is situated 5 km south of Uppsala city centre. The soil cultivation experiment R2-4007 was started in 1974 with the purpose to investigate long-term effects in soil of different soil cultivation depths and methods. Crop residues are incorporated into the soil. In 1994, the crop was spring barley preceded by oats (Avena sativa L.). In 2005 oats was grown and the preceding crop was winter wheat. The sampled treatments were:

A: Mouldboard ploughing every year to 20–25 cm depth.

B: Mouldboard ploughing every fifth year to 20–25 cm depth, last ploughed in 2003, in other years only superficial cultivation as in treatment D.

D: Ploughless tillage, only superficial cultivation each year, disc harrowing or chisel ploughing to 10 cm depth.

The treatments in the Ultuna experiments are randomized in four blocks. Plot size is 20 m×13 m. In 1994, treatments A and D were sampled and in 2005 treatments A, B and D.

Before the set up of the three field experiments Charlottenlund, Väby and Ultuna the whole experimental areas of each site, respectively, had been managed in the same way, i.e. possible differences between treatments should be due to treatment effects and not earlier field history.

In an energy forest plantation (Salix viminalis L.) at Ultuna and in an adjacent agricultural field grown with arable crops, earthworms were sampled in May during the period 1995–2005. The soil of both fields is heavy clay. The area of the Salix plantation was c. 200 m×100 m and the agricultural field was c. 280 m×400 m. The sampling was done in areas of 25 m×100 m in the two crops, bordering each other with a small road in between, and 25 m from the road on each side. The Salix plantation was established in 1984 and was intensively fertilized and irrigated for maximal production during 1984–1991; after 1991 no further fertilization and irrigation was done. Before 1984, the field was grown with agricultural crops, like the adjacent agricultural field. The Salix biomass was harvested at wintertime every fourth year; 1995/96, 1999/2000 and 2003/04.

Up to 2002, the adjacent agricultural field was grown with annual crops and conventional agricultural practice was used concerning ploughing, addition of fertilizers and stable manure and chemical plant protection and weed management. Thereafter, it was shifted over to organic farming. Harvested crops in the agricultural field were: green fallow 1995, rye 1996, winter wheat 1997, spring wheat 1998, spring barley 1999, winter wheat 2000, oats 2001, green fallow 2002, winter wheat 2003, green manure 2004, and winter wheat 2005.

Sampling procedure

Two alternative sampling methods were used: digging out and hand sorting of soil volumes (soil cultivation experiments at Väby and Ultuna and energy forest and agricultural field comparison at Ultuna) and the formalin extraction method (soil cultivation experiment at Charlottenlund and energy forest and agricultural field comparison at Ultuna).

Samples for hand sorting were taken with a steel frame measuring 25 cm×25 cm, which was forced into the soil with a sledgehammer down to 20 cm depth. In each plot of the three soil cultivation experiments, between two and four sample units were taken from which mean number and biomass was calculated and used as one replicate. In the energy forest and agricultural field comparison at Ultuna, eight samples were taken per treatment at each sampling occasion. The extracted soil volume was hand sorted and all earthworms were collected. In the samplings of soil cultivation experiments in 2005, the soil was divided into two depths, 0–10 cm and 10–20 cm, before the hand sorting.

Formalin sampling was done with formalin solution (50 ml of concentrated formalin in 10 l water = 0.2% formaldehyde) that was poured over a 0.5-m² square area that had been rinsed from above-ground plant material and debris. After application of formalin, the soil surface was covered with black plastic during c. 10 minutes while the earthworms came to the surface. Thereafter the plastic was removed and the earthworms collected with forceps, rinsed in clean water and put in vessels. The plastic prevented the earthworms from turning back into the soil when meeting the daylight. Also during the first ten minutes the soil surface under the plastic was inspected occasionally for collection of worms. The procedure was repeated once again after c. 15 minutes. The method is modified from Raw (1959).

The earthworms were brought to the laboratory and determined to species in the live state. They were classified as adults if they had a clearly visible clitellum, or as juveniles otherwise. We used identification keys in Lofs (1991) and Andersen (1997). With knowledge of the local fauna in Swedish agricultural land with few species, identification of earthworms to the species level is feasible and the used keys rely both on morphological and behavioural features of live specimens. Live identification was also used in earlier studies in the area (e.g. Lofs-Holmin 1983, Boström 1995, Lagerlöf et al. 2002). For less experienced persons, juvenile specimens can be problematic and therefore we combined them in some of the samplings. In the sampling of reduced-tillage experiments in 2005, the earthworm species of the genus Lumbricus (mostly L. terrestris) were put together, and so were specimens of the genera Aporrectodea and Allolobophora (mostly A. caliginosa). In the 1994 sampling and in the energy forest/agricultural field comparison, all species were considered separately, except Lumbicus castaneus (Savigny) and L. rubellus (Savigny) which were lumped since juveniles could not be accurately distinguished. In the sampling of the Ultuna soil cultivation experiment in 1994, and Charlottenlund and Väby in 2005 the earthworm biomass was estimated. All juveniles of each species from each sample unit were weighed alive together and the same was done with the adults. Fresh biomass of earthworms with gut content can approximately be transformed to dry biomass with empty gut by dividing by 10 (Petersen and Luxton 1982).

Statistical analysis

The results were analysed by ANOVA (GLM) for estimation of differences in abundance, biomass and depth distribution between treatments. When significant differences were found, pair-wise comparisons were made using Tukey's test.

In the analysis of Salix energy forest vs. arable agriculture at Ultuna, only values from the period 2002–2005 were used in the ANOVA while only values for mean could be calculated for the period 1995–2001. This study was done as an exercise at an ecology course and the initial intention was not to make it a long-term study. Therefore complete primary data were unfortunately not stored from the start.

Results

Soil cultivation experiments

The sampling with formalin at Charlottenlund in 2005 (Table I) revealed similar densities of total number of earthworms in the two treatments, 70 and 73 ind. m⁻² in ploughless tillage A and mouldboard ploughing B, respectively. Significantly higher density and biomass of Lumbricus spp. (mostly L. terrestris) were found in ploughless tillage while higher density of juvenile Aporrectodea caliginosa was found in the mouldboard ploughing treatment.

Table I. Earthworm mean abundance and biomass (SE within parentheses) in soil cultivation field experiment Charlottenlund, September 2005. Sampling by formalin, n = 3. Significant differences between treatments: ***, p<0.001, **, p<0.01, * = < 0.05. L. spp. =Lumbricus spp., A. c.=Aporrectodea caliginosa, ad.=adults, juv.=juveniles, tot.=total.

	Numbers ind. m^-2		Fresh biomass g m^-2
Earthworm spp./Treatment	Ploughless tillage A	Mouldboard ploughing B	Ploughless tillage A	Mouldboard Ploughing B
L. spp. tot.	17.0 (2.0)**	8.0 (1.3)	30.9 (3.2)***	10.0 (2.9)
L. spp. ad.	8.0 (1.5)	5.3 (1.4)	25.0 (2.9)**	9.0 (3.2)
L. spp. juv.	9.0 (2.1)	2.7 (1.4)	5.9 (1.3)*	1.0 (0.5)
A. c.+other spp. tot.	53.0 (10.3)	64.7 (7.7)	9.5 (1.4)	8.3 (1.2)
A. c.+other spp. ad.	11.0 (2.0)	6.7 (2.2)	3.9 (1.0)	2.3 (0.9)
A. c.+other spp. juv.	42.0 (8.5)	58.0 (6.0)*	5.5 (1.0)	6.0 (0.5)
All earthw. tot.	70.0 (9.8)	72.7 (8.3)	40.4 (3.7)***	18.3 (3.3)

The total abundance of earthworms in Väby in 2005, sampled by digging and handsorting (Table II), was 171 ind. m⁻² in ploughless tillage A and 145 in mouldboard ploughing B. The total of A. caliginosa and others (ad.+juv.) was 135 ind. m⁻² in ploughless tillage and 116 ind. m⁻² in mouldboard ploughing. These differences were not significant. Juveniles dominated among A. caliginosa and other earthworms while only adults of Lumbricus spp. were found. The depth distribution of A. caliginosa and others showed higher mean densities in the upper 10 cm than in the lower in both treatments, and more so in the ploughless tillage than in mouldboard ploughing (not significant). For Lumbricus spp., there was a tendency for more individuals in the lower stratum (Table II).

Table II. Earthworm mean abundance and biomass (SE within parentheses) in soil cultivation field experiment Väby, September 2005. Sampling by digging and handsorting, n = 3. No significant differences between treatments were found. L. spp.=Lumbricus spp, A. c.=Aporrectodea caliginosa, ad.=adults, juv.=juveniles, tot.=total.

	Numbers (ind. m^−2)		Fresh biomass (g m^−2)
Earthworm spp./Treatment/Depth (cm)	Ploughless tillage A	Mouldboard ploughing B	Ploughless tillage A	Mouldboard ploughing B
L. spp. tot. 0–20	26.7 (10.6)	18.7 (4.1)	28.2 (11.4)	34.6 (9.1)
L. spp. ad. 0–10	10.7 (3.7)	6.7 (3.0)	10.1 (3.2)	10.4 (4.8)
0–20	16.0 (7.5)	12.0 (3.4)	18.1 (9.1)	24.2 (8.3)
L. spp. juv. 0–10	0	0	0	0
10–20	0	0	0	0
A. c. tot.+other spp. 0–20	144.5 (24.2)	116.0 (15.7)	35.3 (4.0)	35.0 (5.9)
A. c. ad. 0–10	25.3 (7.5)	12.0 (4.5)	12.6 (3.7)	9.4 (3.5)
10–20	11.7 (6.8)	20.0 (5.9)	6.4 (3.4)	10.1 (3.4)
A. c. juv. 0–10	65.3 (17.0)	52.0 (10.4)	11.3 (2.4)	8.9 (2.1)
10–20	42.2 (10.2)	32.0 (10.7)	5.0 (1.4)	6.6 (2.1)
Tot. earthw. 0–20	171.2 (25.2)	134.7 (15.9)	63.5 (13.3)	69.6 (9.5)

In Ultuna 2005 (Table IIIa), sampling by digging and hand sorting gave significantly higher total earthworm mean densities in the ploughless tillage D (245 ind. m⁻²) than in the mouldboard ploughing A (80 ind. m⁻²) treatment with intermediate densities in the every 5th-year mouldboard ploughing treatment B (140 ind. m⁻²). A. caliginosa and other endogeic earthworms dominated strongly in all treatments while Lumbricus spp. (mostly anecic L. terrestris) contributed less than 10% of the total abundance. A high proportion of individuals of all species and life stages were found in the upper 10 cm in all treatments and the distribution of A. caliginosa and other earthworms and total earthworm numbers was significantly more superficial in no tillage than in the other treatments (Table IIIb).

Table IIIa. Earthworm density and biomass (SE within parentheses) in soil cultivation field experiment Ultuna, September 2005. Sampling by digging and handsorting, n = 4. Significant differences between treatments (p<0.05) indicated by letters; rows with different letters for a certain species, stage or depth differ significantly. L. spp.=Lumbricus spp., A. c.=Aporrectodea caliginosa, ad.=adults, juv.=juveniles, tot.=total.

	Number (ind. m^−2)
Earthworm spp./Treatment/Depth (cm)	A: Mouldboard ploughing	B: Mouldboard ploughing every 5th year	D: Ploughless tillage
L. spp. ad + juv 0–20	5.3 (5.8)	3.6 (6.3)	16.0 (10.5)
L. spp. ad. 0–10	5.3 (11.3)	3.6 (7.1)	14.2 (20.3)
10–20	0	0	1.8 (5.3)
L. spp. juv. 0–10	0	0	0
10–20	0	0	0
A. c.+other spp. ad.+juv. 0–20	74.4 (25.3)b	136.9 (30.9)b	229.3 (38.3)a
ad. 0–10	7.1 (11.6)	23.1 (25.4)	30.2 (30.4)
10–20	1.8 (5.3)	0	8.9 (21.3)
juv. 0–10	64.0 (44.4)b	104.9 (36.8)ab	172.4 (68.7)a
10–20	1.8 (5.3)	8.9 (19.5)	17.8 (28.2)
Tot. earthw. 0–20	80.0 (22.2)b	140.5 (35.1)b	245.3 (31.9)a

Table IIIb. ANOVA, p-values for factors treatment and depth in Soil cultivation experiment, Ultuna 2005 (as in Table IIIa). Values for adult + juvenile earthworms.

	L. spp.	A. c.+other spp.	Total earthw.
Treatment	0.11	0.002	0.001
Depth	0.02	<0.000	0.001
Treatment×Depth	0.36	0.02	0.01

The ploughless tillage D and the every year mouldboard ploughing treatment A in the soil cultivation experiment at Ultuna were also sampled in 1994 (Table IV). All four earthworm species found were determined and counted separately and biomass was estimated. Like in 2005, total abundance was significantly higher in the ploughless tillage than in the mouldboard ploughing treatment, with mean numbers of 161 and 95 ind. m⁻², respectively. L. terrestris was only found in the ploughless tillage treatment and in low numbers. A. caliginosa dominated in both sampled treatments and had significantly higher densities in the ploughless tillage treatment. Aporrectodea longa (Ude) was found with on average 8 ind. m⁻² in the ploughless tillage but only 1 ind. m⁻² in the mouldboard ploughing treatment (not significant). Allolobophora chlorotica (Savigny), however, was significantly more abundant in the mouldboard ploughing treatment (18 ind. m⁻²) than in the ploughless tillage (2 ind. m⁻²). The total earthworm biomass in ploughless tillage was 45 g m⁻², of which 31 g m⁻² consisted of A. caliginosa. In mouldboard ploughing, it was significantly lower with total biomass of 16 g m⁻² of which A. caliginosa made up 12 g m⁻².

Table IV. Abundance (no. m⁻²) and biomass (g fresh mass m⁻²) of earthworms in Ultuna field experiment with mouldboard ploughing (A) and ploughless tillage (D), n = 4. Sampling date 29 August 1994. Sampling by digging and handsorting. * or ** indicate significantly higher levels than in the other treatment, p<0.05 and p<0.01, respectively.

	Numbers ind. m^−2		Fresh biomass (g m^−2)
Earthworm spp./Treatment/	A: Mouldboard ploughing	D: Ploughless tillage	A: Mouldboard ploughing	D: Ploughless tillage
Lumbricus terrestris	0.0	1.0 (2.0)	0.0	0.2 (0.4)
Aporrectodea caliginosa	76.0 (11.7)	156.0 (11.8)*	12.1 (7.4)	30.7 (5.1)**
Aporrectodea longa	1.0 (2.0)	8.0 (11.8)	0.8 (1.7)	13.1 (17.8)
Allolobophora chlorotica	18.0 (13.3)*	2.0 (4.0)	2.8 (1.8)	1.1 (2.1)
Tot. earth-worms	95.0 (23.6)	161.0 (39.5)*	15.7 (8.1)	45.0 (20.0)*

Salix plantation vs. arable agriculture

Sampling of earthworm populations in a Salix plantation and an adjacent agricultural field with arable crops in Ultuna from 1995 to 2005 showed surprisingly constant abundances over the years (Figure 1) with no significant between-year variation for total number of earthworms or for any of the eight species found (p>0.05). The two sampling methods gave quite different results. In the soil of the agricultural field the formalin method gave almost no earthworms while in the energy forest it gave on average 36% of the numbers recovered by hand sorting. For anecic and epigeic species the yield was better and for endogeic it was worse than this average of 36%. So, the efficiency of the formalin method differed significantly between the cropping systems (interaction treatment×sampling method: p<0.001). In the energy forest, when sampled by digging and hand sorting the average abundance of total earthworms was 145 ind. m⁻² and sampled by formalin it was 51 ind. m⁻². The average abundance in the arable field was 80 ind. m⁻² when sampled by digging and hand sorting and 0.3 ind. m⁻² when sampled by formalin.

Figure 1.  Total abundance of earthworms in an energy forest plantation (filled bars) and in an adjacent agricultural field with annual crops (unfilled bars) at Ultuna in May 1995–2005. Sampling by digging and hand sorting, n = 8. No sampling in 1996 and 2000.The Salix biomass was harvested at wintertime every fourth year; 1995/96, 1999/2000 and 2003/04. Harvested crops in the agricultural field were: green fallow 1995, rye 1996, winter wheat 1997, spring wheat 1998, spring barley 1999, winter wheat 2000, oats 2001, green fallow 2002, winter wheat 2003, green manure 2004, and winter wheat 2005.

Eight earthworm species were found in the energy forest and the adjacent agricultural field (Table V). Among these, Aporrectodea caliginosa, A. longa and Lumbricus terrestris were significantly more abundant under energy forest than in arable soil, while for A. chlorotica and Aporrectodea rosea (Savigny) no significant differences could be found. Eiseniella tetraedra (Savigny) and L. castaneus/rubellus were found almost exclusively in energy forest soil, but no significant differences could be established due to mostly low numbers and great variations.

Table 5. Abundance of earthworms, total numbers and per species, in an agricultural field and an adjacent energy forest plantation in Ultuna. Mean numbers (ind. m^-2). Sampling in mid May 2002–2005, n = 8. Sampling methods: (1) digging and hand sorting and (2) formalin extraction. Significant differences between land uses for the two sampling methods, respectively: *** = p<0.001, ** = p<0.01, * = p<0.05.

Method:	Digging and hand sorting		Formalin extraction
Land use/Species	Agricultural field	Energy forest	Agricultural field	Energy forest
Lumbricus terrestris	0.5 (0.5)	7.5 (2.6)*	0	7.0 (1.7)***
Lumbricus rubellus/castaneus	0.5 (0.5)	6.5 (3.2)	0	4.0 (1.7)*
Eiseniella tetraedra	0.5 (0.5)	7.0 (4.1)	0	0.8 (0.5)*
Aporrectodea caliginosa	35.5 (6.4)	74.0 (10.5)**	0.25 (0.3)	20.3 (9.8)***
Aporrectodea longa	7.0 (2.7)	21.5 (4.3)*	0	10.2 ((2.6)***
Aporrectodea rosea	14.0 (3.3)	14.5 (3.9)	0	3.2 (1.3)***
Allolobophora chlorotica	21.5 (23.2)	13.5 (4.0)	0	6.0 (2.0)***
Tot. earthworms	79.5 (10.4)	144.5 (11.4) ***	0.25 (0.3)	51.5***

Discussion

Soil cultivation experiments

In comparison with other agricultural sites, earthworm abundances found in this study show rather normal values and they do not come close to either top or bottom notations for studies in northern Europe (Boström 1988, Lagerlöf et al. 2002). As we expected, significantly higher earthworm abundances and biomass were in most cases found in treatments with reduced tillage as compared with conventional ploughing. This agrees with earlier results from similar studies in different parts of the world (Hendrix et al. 1986, Freibe and Henke 1991, Rasmusen 1999, Terbrügge and Düring 1999, Chan 2001, Emmerling 2001, Nakamura et al. 2003, Rothwell et al. 2005).

The more superficial distribution of earthworms found in 2005 in the ploughless tillage treatment D of the Ultuna experiment as compared with the mouldboard ploughing treatment A coincides with significantly higher organic matter content in the 0–10 cm layer of D (2.45% C) as compared with A (1.92% C). In the deeper layers, the organic matter contents were reversed with significantly higher levels in treatment A. The same trend was also found for enzymatic activities of dehydrogenase, acid phosphatase and alkaline phosphatase with significantly higher levels in the superficial 0–10 cm level of treatment D, reflecting higher microbial activity. This was found in analyses of samplings during 1996 (Sjökvist et al. 1998). The distribution of carbon and microbial activity is an effect of the depth to which plant residues were incorporated by tillage methods. Subsequently, the distribution of plant residues affects the distribution of food for earthworms.

On the earthworm functional group and species level, our results show that the anecic species Lumbricus terrestris and Aporrectodea longa, were more adversely affected by ploughing than endogeic species (Aporrectodea caliginosa, A. rosea and Allolobophora chlorotica) and that epigeic species (Lumbricus castaneus/rubellus and Eiseniella tetraedra) were absent or rare in the ploughed as well as the superficially cultivated treatments and mostly found in the energy forest. The mean weight of the adult Lumbricus spp. indicates that there was a certain proportion of L. rubellus mixed with the L. terrestris.

These findings are also supported by earlier studies where the endogeic species always dominate in terms of density and biomass in all types of agricultural systems submitted to some type of cultivation (Chan 2001). Sometimes, this group is just as abundant in conventional tillage as in reduced-tillage systems or in soil under permanent vegetation, sometime even more abundant (Lagerlöf et al. 2002, Krogh et al. 2007). High inputs of crop residues, manure and other types of organic matter in combination with moderate soil cultivation can give favourable conditions for population increase of species that feed inside the soil (endogeic species) and are not dependent on permanent burrows, in contrast to the anecic species (Schmidt et al. 2001). Smith et al. (2008) found in a study of earthworms in field experiments in Michigan, USA, representing land-use gradients of soil cultivation intensity and input of organic matter, that earthworm density increased with reduced tillage and increased organic matter input. But, as in our study, higher species number was only found in forests plantations on former agricultural land that were a decade old or older. Species found were only introduced earthworms of European origin.

Nuutinen (1992) studied response of earthworm communities to ploughing and stubble cultivation in silty clay, silty clay loam and sandy soil in different long-term field experiments in Finland. He found that in silty clay the lowest abundances of earthworms were found in ploughed soil. In sandy loam, however, the total numbers were 1.5–2 times higher in ploughed soil than in stubble cultivated soil. This coincides with our findings where the most negative effect of ploughing was found in the heavy clay at Ultuna while smaller differences between ploughing and ploughless tillage were found in the sandy loam soils of Charlottenlund and Väby.

We found that A. chlorotica was significantly more common in the conventional tillage treatment than in the minimum tillage treatment at Ultuna in 1994, while the other species were less common or absent. According to Wilcke (1952), cited by Edwards & Lofty 1972) the development time from egg to adult in the field was 48 weeks for A. chlorotica and 74 weeks for A. caliginosa. The cocoon production was equal for the two species, 27 cocoons per year on average. The shorter development time of A. chlorotica gives it a competitive advantage over A. caliginosa in a more disturbed situation since it will take a shorter time to recover its numbers. This may be the explanation for its higher abundance in conventional tillage. A. chlorotica is known to prefer soil with higher moisture than A. caliginosa does and it can survive inundation to a greater extent (Zorn et al. 2008). We cannot see, however, how these traits could have been important for the explanation of the results. In addition to this, A. chlorotica has harder body tissue than other earthworms. It rolls up to a ball when disturbed or in the inactive stage. These abilities might give greater resistance to physical destruction. This is only speculative and we have not found any evidence in the literature.

A higher mortality rate in the conventional tillage is also a possible explanation for the lower individual biomass values found in Ultuna 1994 for A. caliginosa: 0.16 g/individual in conventional and 0.20 in minimum tillage, for total earthworms 0.16 g/individual in conventional and 0.28 in minimum tillage.

Salix plantation vs. arable agriculture

The densities in the energy forest plantation were, as expected, higher than in the adjacent agricultural field under annual crops. The digging and hand-sorting method gave on average for the period 2002–05 the total abundance of earthworms 145 ind. m⁻² in the energy forest and 80 ind. m⁻² in the agricultural field, while the formalin extraction gave around 40% of the yield in the energy forests and a very low yield in the agricultural field. This shows that the formalin method was not useful for this comparison and produced very misleading results. The only published study that we found on earthworm communities in energy forest with comparisons of the populations in agricultural land is Friis et al. (1999). They compared the earthworms in three Salix plantations in Denmark with the situation in adjoining agricultural fields with cereals. They found higher abundance and biomass of earthworms in all three studied Salix plantations (66–324 ind. m⁻²) than in the cereal fields (0.06–28 ind. m⁻²). Five earthworm species (anecic and endogeic) were found in the cereal fields. An additional two species of epigeic earthworms were found in the Salix plantations. The sampling was done with the formalin method, which in the light of our findings give some doubts about the correctness of the results and the significance of the differences found. However, the sampling period in November when the soil is wet and the worms have been able to work through the soil for a longer active period could mean that the comparison with help of formalin would be more accurate than in our study done in May each year.

The consistency in both abundance and species composition, over more than ten years in the energy forest and the agricultural field in Ultuna, is striking. No significant changes over years in either energy forest or in the agricultural field were found, irrespective of crop rotation in the agricultural field or harvest of biomass in the energy forest. Possibly sampling in autumn instead of in May would have given larger differences between years depending on weather conditions during the vegetation period and crops grown in the agricultural field. The extremely low abundances found in 1997 in the agricultural field could be due to weather conditions during the winter. The only unusual factors for this particular period were high precipitation in November 1996 with 110 mm (average is 46 mm) and mean temperature + 2.2 °C (average 1.1 °C). This combination of high precipitation and higher temperature than normal could have caused inundation and anoxic conditions that caused death of earthworms. The succeeding winter months December–March were somewhat milder than the average, with mostly periods of a thin snow cover (Ultuna Climate Station).

The reason for higher densities of earthworms in the energy forest that in the agricultural field is first of all the absence of tillage for many years, which especially favours the epigeic species since a superficial organic matter layer is formed. Cooler temperatures and higher soil moisture during dry summer periods should also be of importance (no data available from this experiment). Also the higher amount of organic matter input, forming the food resource for the decomposer food web of the soil, is crucial since soil food webs are in general resource controlled (Scheu and Schaefer 1998). In a Salix field experiment in the same geographical region as Ultuna the below-ground C input was estimated to 3 tonnes C ha⁻¹ yr⁻¹ and the input from leaf litter was 0.95 tonnes dry mass ha⁻¹ yr⁻¹ (c. 0.45 tonnes C ha⁻¹ yr⁻¹) These figures could be increased by 50% with optimal fertilization (Grelle et al. 2007). In a field experiment with field crops in the same region, input to soil litter from shoots and roots was estimated to between 1.4 tonnes C ha⁻¹ yr⁻¹ (barley without N fertilization) and 2.5 tonnes C ha⁻¹ yr⁻¹ (lucerne ley) (Paustian et al. 1990).

Hand sorting vs. formalin extraction

Field studies of earthworm communities in agricultural land that used formalin or some other vermifuge for extracting earthworms must be regarded with caution since the extraction efficiency is in general low as compared with digging and hand sorting (Klein and Knäbe 2007). This type of sampling should therefore always be calibrated to sampling by digging and hand sorting. The fact that extraction efficiency differs between treatments, because of differences in soil structure and thereby the ability of the earthworms to get to the surface, is an additional complication, which makes the method especially problematic when studying effects of soil cultivation. Barnes and Ellis (1979) concluded that comparable earthworm counts can be obtained with the formalin method from soils under different tillage treatments if the sampling is done in autumn before cultivation. Our studies of the energy forest plantation compared with arable agriculture were done in the spring, which can explain the poor efficiency of formalin sampling in the arable field compared with digging and hand sorting. But the efficiency was equally bad during years when the field had been ploughed just a few weeks earlier and sown with spring crops and when it was sown with winter crops and therefore had not been soil cultivated since the early autumn before, or it had been under green fallow or green manure for even longer periods without soil cultivation. The combination of digging and hand sorting and vermifuge extraction has been recommended by many authors (e.g. Smith et al. 2008) since only digging and handsorting may under-estimate the large and deep burrowing species. Others have found that vermifuge gives very little additional catch compared with the digging and hand sorting only, e.g. Christensen et al. (1987) who sampled agricultural soils in Denmark where the anesic species L. terrestris and A. longa were common.

In most agricultural soils, mouldboard ploughing every year results in lower earthworm abundance and biomass as compared with reduced soil cultivations. This is especially true for anecic species (feeding on the soil surface and producing deep burrows) and epigeic species (living superficially) whereas endogeic species (living and feeding within the soil) are less affected. Endogeic species (e.g. Allolobophora chlorotica) can even be more abundant in conventional ploughing than in soil under reduced tillage. Energy forest plantations of Salix on agricultural land that are not soil cultivated for several years enhance earthworm abundance and biomass even more than reduced tillage in annual crops. They can also sustain higher species diversity, including several epigeic earthworm species.

Methods for sampling of earthworms by means of formalin or other vermifuges function with different efficiency depending on soil cultivation. Therefore, such methods have to be used with caution when studying impact of agricultural methods on earthworms and should always be supplemented with sampling by digging and handsorting.

Acknowledgements

We thank a great number of technicians and student helpers who spent countless hours sampling and hand sorting soil samples. We owe great gratitude to all agronomy students that through the years studied earthworms in the energy forest plantation and in the adjacent arable field at Ultuna, which was done as an exercise during their basic ecology course.The studies were financially supported by the Faculty of Natural Resources and Agricultural Sciences, SLU.