ORIGINAL PAPER A. Winding · R. Rønn · N.B. Hendriksen Bacteria and protozoa in soil microhabitats as affected by earthworms Received: 7 January 1996 A. Winding ( ) ) · N.B. Hendriksen Department of Marine Ecology and Microbiology, National Environmental Research Institute, PO Box 358, Frederiksborgvej 399, DK-4000 Roskilde, Denmark Tel.: 46 30 12 00; Fax: 46 30 1114; e-mail: [email protected]R. Rønn Department of Population Biology, Zoological Institute, University of Copenhagen, Denmark Biol Fertil Soils (1997) 24:133–140 ° Springer-Verlag 1997 Abstract The effects of incorporation of elm leaves (Ul- mus glabra) into an agricultural sandy loam soil by earth- worms (Lumbricus festivus) on the bacterial and protozoan populations were investigated. Three model systems con- sisting of soil, soil with leaves, and soil with leaves and earthworms, respectively, were compared. The total, vi- able, and culturable number of bacteria, the metabolic po- tentials of bacterial populations, and the number of proto- zoa and nematodes were determined in soil size fractions. Significant differences between soil fractions were shown by all assays. The highest number of microorganisms was found in microaggregates of 2–53 lm and the lowest in the <0.2 lm fraction. A major part of the bacteria in the latter fraction was viable, but non-culturable, while a rela- tively higher number of culturable bacteria was found in the macroaggregates. The number of colony-forming units and 5-cyano-2,3-ditolyl tetrazolim chloride (CTC)-redu- cing bacteria explained a major part of the variation in the number of protozoa. High protozoan activity and predation thus coincided with high bacterial activity. In soil with elm leaves, fungal growth is assumed to inhibit bacterial and protozoan activity. In soil with elm leaves and earth- worms, earthworm activity led to increased culturability of bacteria, activity of protozoa, number of nematodes, chan- ged metabolic potentials of the bacteria, and decreased dif- ferences in metabolic potentials between bacterial popula- tions in the soil fractions. The effects of earthworms can be mediated by mechanical mixing of the soil constituents and incorporation of organic matter into the soil, but as the earthworms have only consumed a minor part of the soil, priming effects are believed partly to explain the in- creased microbial activity. Key words Soil bacteria · Protozoa · Earthworms · Viability · Aggregates Introduction The structure of soil microhabitats is, in addition to physi- cal and chemical factors, affected by bacteria, fungal hy- phae, and macrofauna (Robert and Chenu 1992). As a function of these environmental conditions, soil bacteria show varying activity in different microhabitats (Kanaza- wa and Filip 1986; Lensi et al. 1995). Earthworms are an important element of the macrofauna and are reported to influence soil structure by creating burrows, which facili- tate water and gas transport (Zhang and Schrader 1993), by mixing soil minerals with organic material (Hendriksen 1991), and by comminuting and incorporating litter into the soil and thereby creating macroaggregates (Lee 1985). Earthworm activity leads to increased mineralization and microbial activity in bulk soil, and in earthworm casts the number of culturable bacteria is increased while the size of fungal and protozoan populations are decreased (re- viewed in Brown 1995). A change in bacterial community structure occurs along the earthworm gut (Pedersen and Hendriksen 1993). These effect have been proposed to be mechanical effects as well as priming effects of the earth- worms excreting and altering organic substrates (Lavelle and Gilot 1994). The aim of our experiments was to study the effect of earthworms on the number and activity of bacteria and protozoa in microhabitats of an agricultural sandy loam soil. The total, viable, and culturable number of bacteria and the metabolic potentials of bacterial populations were determined in five different size fractions of the soil and in earthworm gut material. The number and activity of protozoa were determined in soil fractions as well as in unfractionated soil. In the latter, the number of nematodes was also determined.
Transcript
ORIGINAL PAPER
A. Winding ´ R. Rùnn ´ N.B. Hendriksen
Bacteria and protozoa in soil microhabitats as affected by earthworms
Received: 7 January 1996
A. Winding ()) ´ N.B. HendriksenDepartment of Marine Ecology and Microbiology,National Environmental Research Institute, PO Box 358,Frederiksborgvej 399, DK-4000 Roskilde, DenmarkTel.: 46 301200; Fax: 46301114; e-mail: [email protected]
R. RùnnDepartment of Population Biology, Zoological Institute,University of Copenhagen, Denmark
Abstract The effects of incorporation of elm leaves (Ul-mus glabra) into an agricultural sandy loam soil by earth-worms (Lumbricus festivus) on the bacterial and protozoanpopulations were investigated. Three model systems con-sisting of soil, soil with leaves, and soil with leaves andearthworms, respectively, were compared. The total, vi-able, and culturable number of bacteria, the metabolic po-tentials of bacterial populations, and the number of proto-zoa and nematodes were determined in soil size fractions.Significant differences between soil fractions were shownby all assays. The highest number of microorganisms wasfound in microaggregates of 2±53 lm and the lowest inthe <0.2 lm fraction. A major part of the bacteria in thelatter fraction was viable, but non-culturable, while a rela-tively higher number of culturable bacteria was found inthe macroaggregates. The number of colony-forming unitsand 5-cyano-2,3-ditolyl tetrazolim chloride (CTC)-redu-cing bacteria explained a major part of the variation in thenumber of protozoa. High protozoan activity and predationthus coincided with high bacterial activity. In soil withelm leaves, fungal growth is assumed to inhibit bacterialand protozoan activity. In soil with elm leaves and earth-worms, earthworm activity led to increased culturability ofbacteria, activity of protozoa, number of nematodes, chan-ged metabolic potentials of the bacteria, and decreased dif-ferences in metabolic potentials between bacterial popula-tions in the soil fractions. The effects of earthworms canbe mediated by mechanical mixing of the soil constituentsand incorporation of organic matter into the soil, but asthe earthworms have only consumed a minor part of the
soil, priming effects are believed partly to explain the in-creased microbial activity.
The structure of soil microhabitats is, in addition to physi-cal and chemical factors, affected by bacteria, fungal hy-phae, and macrofauna (Robert and Chenu 1992). As afunction of these environmental conditions, soil bacteriashow varying activity in different microhabitats (Kanaza-wa and Filip 1986; Lensi et al. 1995). Earthworms are animportant element of the macrofauna and are reported toinfluence soil structure by creating burrows, which facili-tate water and gas transport (Zhang and Schrader 1993),by mixing soil minerals with organic material (Hendriksen1991), and by comminuting and incorporating litter intothe soil and thereby creating macroaggregates (Lee 1985).Earthworm activity leads to increased mineralization andmicrobial activity in bulk soil, and in earthworm casts thenumber of culturable bacteria is increased while the sizeof fungal and protozoan populations are decreased (re-viewed in Brown 1995). A change in bacterial communitystructure occurs along the earthworm gut (Pedersen andHendriksen 1993). These effect have been proposed to bemechanical effects as well as priming effects of the earth-worms excreting and altering organic substrates (Lavelleand Gilot 1994).
The aim of our experiments was to study the effect ofearthworms on the number and activity of bacteria andprotozoa in microhabitats of an agricultural sandy loamsoil. The total, viable, and culturable number of bacteriaand the metabolic potentials of bacterial populations weredetermined in five different size fractions of the soil andin earthworm gut material. The number and activity ofprotozoa were determined in soil fractions as well as inunfractionated soil. In the latter, the number of nematodeswas also determined.
An agricultural sandy loam from Roskilde, DK, sampled in Septem-ber and stored at 4°C (maximum 3 weeks), was passed through a 5.6-mm-mesh sieve before use. Leaf-fall elm leaves (Ulmus glabra) werecollected and dried at 65°C overnight. Elm leaves were chosen asthey are palatable to detritivore earthworms (Satchell and Lowe1967). Before addition to the soil, the leaves were rewetted overnight(0.8 ml g±1). Earthworms (Lumbricus festivus) collected atBùgemosen, DK, were acclimatized at 15°C on leaf-fall elm leavesand the Roskilde soil for 2±5 weeks before transfer to experimentalpots. Portions of 1.2 kg wet weight (ww) of soil were distributed intonine plastic pots (diameter 15 cm, height 13 cm) with lids. Three potscontained only soil (controls). In three pots 6.0-g dw elm leaves wereplaced on top of the soil. To three pots six worms (7 g ww) wereadded and 6.0-g dw elm leaves were placed on top of the soil. Thepots were incubated for 3 weeks at 15°C. Once a week the weight ofthe entire pot was checked, and millipore water was added to keepthe water content of the soil at 18±20% of dw, corresponding to fieldcapacity. By the end of incubation the leaves and worms were gentlyremoved. Earthworm gut material was obtained by forcing gut materi-al through the anus.
Soil fractionation
The soil was thoroughly mixed and fractionated into five size classes:(<0.2 lm particles, 0.2±2 lm clay particles, 2±53 lm and 53±250 lmmicroaggregates, and >250 lm macroaggregates) by a combination ofgentle wet-sieving and centrifugation based on the method of JocteurMonrozier et al. (1991). Separation of the three smallest fractions wasbased on Stoke's law and clay density of 1.8±2.6 g cm±3 (Weast andAstle 1982). Particles 2±53 lm in size were separated from particles<2 lm in size by four centrifugations at 90 g with intermittent resus-pension in Winogradsky salt solution (W) (Holm and Jensen 1972).Particles 0.2±2 lm in size were separated from particles <0.2 lm insize by four centrifugations at 2400 g with intermittent resuspensionin W. During the entire fractionation procedure, all solutions werekept at 0±4°C. Samples of soil fractions and earthworm gut materialwere either placed in a water bath sonicator (Branson 5200) for20 min (bacteriological analyses) or shaken for 20 min at 250 rpm(microfauna analyses).
Abiotic measurements
Dry matter content was determined in the unfractionated soil andeach soil fraction by drying at 105°C for 24 h. Carbon content wasdetermined by a Perkin-Elmer CHN-analyser, type 240C. As no air-bubbles developed when acidifying soil slurries with 30% HCl, thecarbon content is equivalent to the organic carbon content (Nelsonand Sommers 1982).
Bacteria
The total number of bacteria was determined by acridine orange di-rect counts (AODC) of a minimum of 400 bacteria as described byHobbie et al. (1977). For viability assays, appropriate dilutions of soilfractions were filtered through black 0.2 lm polycarbonate filters(Nuclepore; Costar, Cambridge, USA). Bacteria able to reduce CTC(5-cyano-2,3-ditolyl tetrazolium chloride) were enumerated after incu-bation with CTC and staining with 4'-6'-diamidino-2-phenylindole(DAPI) as described by Winding et al. (1994). Formation of microco-lonies was determined after incubation on W in sterile acid washedglass Petri dishes for 63 days at 15°C. The number and sizes of mi-crocolonies were determined after staining with acridine orange as de-scribed by Winding et al. (1994). Dilutions of each soil fraction andearthworm gut material were plated on seven replicate plates of W-
agar medium (15 g agar l±1 W) and the number of colony-formingunits (CFUs) was counted after 63 days at 15°C.
Biolog fingerprint
Dilutions of the soil suspensions and earthworm gut material (3´106
to 6´107 cells ml±1) were inoculated onto GN-Biolog microtitre plates(Biolog Inc., Hayward, CA). After 3 days of agitation at 15°C, theformazan formation in microtitre plates was measured by a video-image analyser measuring grey level values (GLV) (Winding 1994).In order to compensate for varying degree of colour development inthe wells, the GLV of each well was divided by the GLV of the wellwith the highest GLV.
Microfauna
Protozoa were enumerated by a modified version of the most prob-able number method (Darbyshire et al. 1974; Rùnn et al. 1995). Soilfractions and earthworm gut material were diluted in sterile amoebasaline (Page 1976), and subsamples (100 ll) of each dilution weretransferred to eight wells in a microtire plate together with 50 ll of awashed Pseudomonas chlororaphis culture. After 6, 12, 30±35 and47±49 days at 10°C, protozan growth was examined using an in-verted microscope (Olympus CK2). The free-living nematodes wereextracted from unfractionated soil with a Baermann funnel technique(Hooper 1986) from five replicates of 5 g (ww) soil.
Statistics
Two-way ANOVA was used to test for differences in dry matter con-tent, carbon content and in the estimates of viable, culturable and to-tal number of bacteria between soil fractions and soil treatments bySystat, ver. 5.0 (Systat, Evanston, IL, 1990). When no significant dif-ference was found in the estimates of bacteria between soil treat-ments, they were considered as replicates, and one-way ANOVA wasperformed to test for differences between soil fractions. The samesoftware was used for a stepwise regression analysis of microbiologi-cal parameters in a multivariate linear model explaining the variationin the number of protozoa. Principal component analysis, averagelinkage cluster analysis and ANOVA on the Biolog data were accom-plished using SAS, ver. 6.08 (SAS Institute, 1989).
Results
Abiotic parameters
The macroaggregates (>250 lm) contained the greatestpercentage of material with gradually lower amounts inthe smaller fractions (Table 1). Soil treatments did not sig-nificantly alter the amount of material found in each sizefraction. The carbon content (milligrams per fraction)showed significant differences between size fractions(P<0.001) and indicated differences between treatments(P<0.093). Microaggregates of 2±53 lm contained thehighest amount of carbon followed by macroaggregates(>250 lm).
Bacteria
Bacterial abundance is presented relative to the totalamount of soil fractionated in order to describe the distri-
bution of bacteria in soil structures in a given volume ofsoil. Total bacteria (AODC) number showed significantvariation among soil fractions (P<0.001), but not amongsoil treatments (Fig. 1). The abundance of bacteria in the0.2±2 lm and the 2±53 lm fractions was significantlyhigher than the abundance in the 53±250 lm fractions,which again was significantly higher than the abundancein the <0.2 lm fraction.
Microcolony-forming units (micro-CFUs) varied signifi-cantly among soil fractions (P<0.001), but not amongtreatments, though there was a tendency of lower numbersof micro-CFUs in soil with leaves and worms (Fig. 1). Im-mediately after filtering, the initial number of microcolo-nies was counted. These microcolonies are insufficientlyseparated cells and always accounted for less than 10% ofthe final number of micro-CFUs. There was no significantdifference between the number of micro-CFUs found inthe fractions of 0.2±2 lm, 53±250 lm, and >250 lm.Higher numbers of micro-CFUs were found in the 2±53 lm fraction, while the bacteria in the fraction <0.2 lmaccounted for the lowest number of micro-CFUs.
The microcolonies were divided into three size classesbased on the number of cells per microcolony (Fig. 2). Soilwith leaves contained the highest number of two to three-celled microcolonies in the <0.2 lm, 2±53 lm fractions,and >250 lm fraction. Soil with leaves and worms con-tained a higher number of microcolonies consisting of morethan 12 cells, especially in the microaggregates of 2±53 lm.
CTC-reducing bacterial abundance varied significantlyamong the size fractions (P<0.001), but not among soiltreatments, though the lowest numbers were found in soilwith leaves in all fractions except the <0.2 lm fraction(Fig. 1). The highest number of CTC-reducing bacteria pergram of soil was found in the microaggregates of 2±53 lm, the lowest number in the <0.2 lm fraction, whilethe remaining three fractions did not differ significantlyfrom each other.
CFU abundance showed significant differences betweenfractions (P<0.001) (lowercase letters in Fig. 1) and be-tween treatments (P<0.001) (capital letters in Fig. 1). Thehighest number of CFUs was found in the 2±53 lm frac-tion, while the lowest number was found in the <0.2 lmfraction. The difference between soil treatments was signif-
135
Table 1 Dry matter and organic carbon content in soil fractions andunfractionated soil
Soil fraction Dry matter Organic C% of intactsoil mg C
Fig. 1 Number of bacteria insoil fractions determined by: ac-ridine orange direct count(AODC), microcolony-formingunits (micro-CFU), reduction ofCTC to fluorescent CTC-forma-zan (CTC), and colony-formingunits (CFU). n control soil, `soil incubated with elm leaves,n soil incubated with elm leavesand earthworms. Identical lettersindicate no significant difference(P> 0.05) while no letter indi-cates significant differences byone-way ANOVA (AODC, mi-cro-CFU and CTC), and by two-way ANOVA (CFU). Lowercaseletters refer to soil fractions, andcapital letters (in plot of CFU)to soil treatments
icant within all size fractions. Amont the different treat-ments, soil with leaves had a significantly lower numberof CFUs in the <0.2 lm, 2±53 lm, and 53±250 lm frac-tions, while soil with leaves and worms had a significantlyhigher number of CFUs in the 0.2±2 lm, 53±250 lm, and>250 lm fractions (Fig. 1). When the number of CFUs inthe soil was calculated as the sum of CFUs in the indivi-dual fractions divided by the soil fractionated,5.0´108 CFU g±1 dw were found in the soil with leavesand worms. By contrast, 1.6´109 CFU g±1 dw was foundin earthworm gut material.
Viability and culturability
Bacterial viabilities, determined as number of CTC-redu-cing bacteria or micro-CFUs relative to total number ofbacteria, and culturability, determined as number of CFUsrelative to total number of bacteria, were calculated (Ta-ble 2). The micro-CFU assay generally detected the high-est number of viable cells (Fig. 1) and therefore also thehighest viabilities, while viabilities detected by the CTCassay were comparable to culturabilities detected by theCFU assay. Irrespective of soil treatment, micro-CFU-de-termined viability was highest in the <0.2 lm fraction,CTC-determined viability was highest in the 2±53 lmfraction, and culturability was highest in the macroaggre-gates, except in soil with leaves and worms. Comparingsoil treatments, micro-CFU-determined viability was high-est in soil with leaves and lowest in soil with leaves andworms. In contrast to this, culturability was highest in soilwith leaves and worms and lowest in soil with leaves. Inthree of the fractions culturabilities were even higher thanthe micro-CFU-determined viability. CTC-determined via-bility was lower in soil with leaves than in both the con-trol soil and soil with leaves and worms.
Biolog fingerprint
The oxidation of 95 different carbon sources in GN-Biologmicrotitre plates formed the basis of a characterization ofbacterial communities by principal component analysisand cluster analysis (Fig. 3). With principal componentanalysis earthworm gut material showed the greatest dis-tance to the other fractions along the first principal compo-nent and grouped together with the fractions with leavesand worms. Macroaggregates in the control soil groupedtogether with the <0.2 lm and >250 lm fractions of soilwith leaves. Finally, the remaining samples of control soiland soil with leaves grouped together. Cluster analysisshowed that samples from soil with leaves and wormsclustered together, and samples from control soil clusteredtogether with samples from soil with leaves. Two-way
136
Fig. 2 Distribution of the micro-colony-forming units (micro-CFUs) in three size classes as afunction of the five soil fractionsand three soil treatments. Sym-bols as in Fig. 1
Table 2 Culturability and viability calculated relative to AODC (%)
ANOVAs on each of the 95 carbon sources revealed sig-nificant differences (P<0.05) between soil treatments for18 carbon sources, 3 carbon sources showed significantdifferences between size fractions, while no significant in-teraction between treatments and size fractions was found.Significant differences between gut material and size frac-tions of soil with leaves and worms were found for justthree carbon sources. The metabolic potentials of bacterialpopulations thus showed larger differences between treat-ments than between size fractions.
Microfauna
The highest number of protozoa was found in the 2±53 lm fraction and the lowest number in the <0.2 lmfraction (Fig. 4A). Significant differences were found be-tween fractions (P<0.001), but not between treatments,though there was a tendency of a lower number in the soilwith leaves. A stepwise regression analysis indicated thatthe number of CFUs and CTC-reducing bacteria explainedmost of the variation in protozoa abundance (r=0.67),since these two estimates contributed significantly to amultivariate linear model relating protozoa to the four dif-ferent estimates of bacteria.
The rate at which protozoa appeared in the wells wasassumed to represent their activity at the time of inocula-tion. In Fig. 4B the MPN obtained after 6 days of incuba-tion is shown as a percentage of the total MPN obtainedafter 48 days. This percentage was higher in soil withleaves and worms for all fractions, except the fraction 0.2±2 lm, where the control soil showed a similar percentage.
This indicates that protozoa in soil with leaves and wormswere more active by the time of soil fractionation thanprotozoa in the other soils.
The dominant protozoa in the wells were naked amoe-bae and flagellates and their relative number varied be-tween fractions with a higher number of flagellates in thetwo smallest size fractions (data not shown). In the gut
137
Fig. 3 Principal component plot of the metabolic fingerprints of thesoil fractions after incubation in GN-Biolog microtitre plates. The var-iation along the first two principal component axis is shown. Signifi-cant clusters formed by average linkage cluster analysis (P<0.05 bypseudo t2-test) are encircled. · contol soil, s soil incubated with elmleaves, n soil incubated with elm leaves and earthworms, ` earth-worm gut material. Figures indicate size of soil fractions in micro-metres
Fig. 4 A Number of protozoa determined by MPN in the five soilfractions of the three soil treatments. B Fraction of the total numberof protozoa appearing during the initial 6 days of incubation. Sym-bols as in Fig. 1
Table 3 Protozoa and nematode abundance in unfractionated soilsamples of three soil treatments (number g±1 dw)
* Significantly different (P<0.001) from control soil
material the number of protozoa was 1.6´105 g±1 dw, cor-responding to approximately 33% of the number found inunfractionated soil with leaves and worms. There was noeffect of leaves or worms on protozoa abundance in theunfractionated soil, but the number of nematodes was sig-nificantly higher in soil with leaves and worms (Table 3),and they were primarily bacterial-feeding (data notshown).
Discussion
Diversity among soil microhabitats
The significant difference in carbon content between sizefractions is expected to be due to a higher clay and siltcontent in the 0.2±53 lm fractions. Christensen (1985) hasalso found higher carbon contents associated with clay andsilt than with sand (Table 1). The relative amount of car-bon in the fractions is comparable to the results of JocteurMonrozier et al. (1991), who found the highest carboncontent in the macroaggregates (>250 lm) followed bymicroaggregates of 2±20 lm.
The density of indigenous soil bacteria has been esti-mated to 1.07 g cm±3 and to be 1.29 g cm±3 for endo-spores (Bakken and Olsen 1983; Bakken 1985). Bacteriawith a diameter smaller than 0.5±0.8 lm and spores witha diameter less than 0.3 lm will be included in the soilfraction <0.2 lm, unless the bacteria are adsorbed to or in-tegrated with particles, or surrounded by extracellularslime. We found the lowest number of bacteria in the<0.2 lm fraction, and these bacteria showed high micro-CFU-determined viabilities and low culturabilities (Ta-ble 2). This population of bacteria thus consisted of manysmall viable, but non-culturable bacteria. This agrees withthe findings that small bacteria (<0.4 lm) have lower cul-turability than larger bacteria (Bakken and Olsen 1987).
Assuming a protozoan cell density of 1.04 g cm±3
(Roberts 1981) and sphericall cells, protozoa smaller than5.6±10.4 lm will appear in the <2 lm fractions, whilecells smaller than 0.6±1.1 lm will appear in the <0.2 lmfraction. The size of the smallest flagellates is 2±3 lm.The occurrence of protozoa in this fraction shows that pro-tozoa do not behave exactly according to the assumptionsof the critical diameter. The protozoa found in the twosmallest size fractions (Fig. 4A) must be unattached organ-isms, and the relatively larger importance of flagellates inthe two smallest size fractions is probably due to a smallersize of some of the flagellates compared to amoebae, andamoebae being more firmly associated with soil particles.The relatively high number of protozoa associated with the2±53 lm microaggregates (Fig. 4A) is in accordance withthe hypothesis of Hattori (1988) that most protozoa are inthe outer zone of the aggregates. The high number of pro-tozoa in this fraction coincide with a significantly highernumber of viable bacteria as detected by the microscope-based assays (Fig. 1), relatively high viability and cultur-ability (Table 2) and the highest carbon content for all soil
treatments (Table 1). Protozoan predation has been sug-gested to stimulate bacterial activity and production (Pus-sard and Rouelle 1986; Kuikman et al. 1990), as a reduc-tion in bacterial biomass increases the amount of availablesubstrate for the remaining bacteria (Hunt et al. 1977), andas equal removal of inactive and active bacteria will createa dominance of active bacteria in the soil (Habte andAlexander 1978). It is thus probable that bacteria in micro-aggregates of 2±53 lm are exposed to the highest preda-tion pressure which had led to a higher activity. In accor-dance with our observations, Postma et al. (1990) foundthe majority of added flagellates associated with particlesof 2±50 lm, and they suggested the predation pressure bythe flagellates to be larger in the 2±50 lm size fractionthan in other fractions.
Culturabilities in the macroaggregates in the controlsoil and in soil with leaves were higher than in the otherfractions (Table 2). This is probably due to more labileand less mineralized organic matter in larger aggregates(Gupta and Germida 1988). The culturabilities (Table 2)were higher than found in an earlier study on the samemedia and soil (sampled in January) (1±2%) (Winding etal. 1994), but lower than the culturability found in thesame soil (sampled in November) on a medium with 0.1 gtryptic soy broth per litre (8±22%) (Winding 1994). Asmall amount of nutrients in the media thus seems to in-crease culturability, but the time of year the soil issampled might also be important. Richaume et al. (1993)found lower culturabilities (0.5±2.2%) in >2 lm size frac-tions from a silt loam on nutrient agar, with a culturabilityof 14% in the fraction <2 lm. They explained the low cul-turability by either a large fraction of non-culturable cellsor the colonies appearing on plates originating from morethan one cell. Viabilities observed by the CTC assay arewithin the range observed earlier in the same soil, whileviabilities detected by the micro-CFU assay are high com-pared to earlier observations (Winding et al. 1994).
Effects of elm leaves and earthworms
The number of CFUs was higher in soil affected by earth-worms, while the number of micro-CFUs was higher inthe undisturbed soil, particularly in the <0.2 lm fraction.Both assays detect the ability of bacteria to divide, but de-tect different physiological groups in the bacteria popula-tions. The micro-CFU assay provides less nutrients forbacterial growth by omitting the agar that is included inthe CFU assay. We believe that, compared to the CFU as-say, the micro-CFU assay detects more oligothropic bacter-ia, which probably often are viable, but non-culturable(Winding et al. 1994). The CTC assay includes incubationin a nutrient-rich medium and therefore it might detectless oligotrophic bacteria than the micro-CFU assay. Earth-worms thus changed the microbial community towardsmore culturable and active, eutrophic bacteria in accor-dance with observations by Hendriksen (1996).
At harvest, it was clear that earthworms had had a sig-nificant impact on the leaves, which were partly degraded
and had been dragged down into burrows. Measurable im-pact on the abiotic parameters of the soil was restricted toorganic carbon content (Table 1). The leaves on top of soilwithout worms were not visibly degraded during the 3weeks of incubation, but fungal hyphae were seen amongthe leaves and in the top few millimetres of the soil, asalso noticed by Zhang and Hendrix (1995). This fungalgrowth was not seen in soil with leaves and worms. Earth-worms are expected to disrupt fungal hyphae mechanicallyor to feed selectively on the hyphae (Moody et al. 1995).Incubation of soil with leaves led to lower culturabilityand viability of the soil bacteria as determined by thenumber of CFU, CTC-reducing bacteria, and two- tothree-celled micro-CFUs (Figs. 1, 2). We believe this low-er viability is caused by nutrient competition between fun-gi and bacteria or by fungal excretion of antibiotics(Wicklow 1992). Polyphenols leaching from leaves havethe ability to inhibit bacteria (McClaugherty 1983), but asweathered elm leaves do not contain polyphenols (Satchelland Lowe 1967), this is unlikely to have occurred in ourexperiments.
CFU abundance in earthworm gut material was 3 timeshigher than the sum of CFUs in the fractions of the soil,which is a common observation (Brown 1995). Earth-worms have been shown to selectively feed on particlessmaller than sand (Zhang and Schrader 1993), and in ourstudy the particle composition of gut material appeared tobe comparable to the composition of the microaggregatesof 2±53 lm. The number of CFUs was of the same mag-nitude in this size of microaggregates and in the gut mate-rial.
Metabolic potentials of bacteria in fractions of soil withleaves and earthworms, including gut material, were moreclosely associated to each other and differed from the me-tabolic potential of size fractions in control soil and soilwith leaves (Fig. 3). Earthworm casts are shown to havelower aggregate stability than aggregates of bulk soil(Zhang and Schrader 1993), and the relatively unstablecasts produced during 3 weeks incubation might havebeen disrupted during size fractionation and distributedamong some of the soil fractions. This could explain thereduction in differences between metabolic fingerprints ofsize fractions (Fig. 3).
The MPN method does not discriminate between activeand encysted protozoa, and a large proportion of protozoaare probably encysted in most situations in the soil (Eke-lund and Rùnn 1994). If a stimulation only led to in-creased excystment and not increased multiplication, thestimulation would not be detected by the MPN method,but by the rate of appearance in the wells as shown inFig. 4B. Compared to soil without earthworms, protozoanactivity was higher in soil with leaves and earthworms aswas the culturability of bacteria, while the number of vi-able, but non-culturable bacteria was lower (Figs. 1, 4B,Table 2). Metabolic potentials of the microbial commu-nities changed, and the number of nematodes increased(Table 3). This shows that earthworm activity had signifi-cant impacts on soil organisms. The increase in protozoanactivity (Fig. 4B) is probably caused by the higher number
of active bacteria, which is also suggested by the modelexplaining the variation in the number of protozoa includ-ing CFUs and CTC-reducing bacteria. Such increased mi-crobial activity was also found in various soils with Apor-rectodea caliginosa (Wolters and Joergensen 1992), in soilwith leaves (Went 1963), and in soil with dung and L. fes-tivus (Hendriksen 1996). In our experiments only approxi-mately 5% of the soil in the pots was consumed by earth-worms as calculated on basis of cast production by L. ru-bellus (Shipatalo et al. 1988). The number of CFUs in thecast was 3 times higher than the number of CFUs in thepooled size fractions, but in the soil fraction of compar-able size, no increase was detected. It is thus unlikely thatthe effects of earthworms are only due to mechanical mix-ing and incorporation of organic matter into the soil. En-zymes and other labile organic substrate in the form ofurine and mucus are excreted along the gut and from thebody surface (Zhang et al. 1993; Lee 1983), and castshave higher contents of polysaccharides (Zhang and Schra-der 1993). This excretion in combination with alteration ofsubstrates has been proposed to prime microorganisms(Lavelle and Gilot 1994), and we propose that the ob-served effects of earthworms on number and activity ofbacteria and protozoa in our experiments are caused to alarge extent by priming.
Acknowledgement C.T. Agger is acknowledged for statistical ad-vice, J.C. Pedersen for critical comments on the manuscript, and D.Conley for comments on the language. V. Mikkelsen provided valu-able technical assistance. The work was supported by the Danish En-vironmental Protection Agency (file no. M 6045-0016), the DanishCenter for Microbial Ecology, and the Danish Environmental Re-search Programme, 1992±1996.
References
Bakken LR (1985) Separation and purification of bacteria from soil.Appl Environ Microbiol 49:1482±1487
Bakken LR, Olsen RA (1983) Buoyant densities and dry-matter con-tents of microorganisms: conversion of a measured biovolumeinto biomass. Appl Environ Microbiol 45:1188±1195
Bakken LR, Olsen RA (1987) The relationship between cell size andviability of soil bacteria. Microb Ecol 13:103±114
Brown GG (1995) How do earthworms affect microfloral and faunalcommunity diversity? Plant and Soil 170:209±231
Christensen BT (1985) Carbon and nitrogen in particle size fractionsisolated from Danish arable soils by ultrasonic dispersion andgravity-sedimentation. Acta Agric Scand 35:175±187
Darbyshire JF, Wheatley RE, Greaves MP, Inkson RHE (1974) A ra-pid micromethod for estimating bacterial and protozoan popula-tions in soil. Rev Ecol Biol Soil 11:465±475
Ekelund F, Rùnn R (1994) Notes on protozoa in agricultural soil withemphasis on heterotrophic flagellates and naked amoebae andtheir ecology. FEMS Microbiol Rev 15:321±353
Gupta VVSR, Germida JJ (1988) Distribution of microbial biomassand its activity in different soil aggregate size classes as affectedby cultivation. Soil Biol Biochem 20:777±786
Habte M, Alexander M (1978) Mechanisms of persistence of lownumbers of bacteria preyed upon by protozoa. Soil Biol Biochem10:1±6
Hattori T (1988) Soil aggregates as microhabitats of microorganisms.Rept Inst Agr Res Tohoku Univ 37:23±36
Hendriksen NB (1991) Gut load and food-retention time in the earth-worms Lumbricus festivus and L. castaneus: A field study. BiolFertil Soils 11:170±173
Hendriksen NB (1996) Earthworm effects on respiratory activity in adung-soil-system. Soil Biol Biochem (in press)
Hobbie JE, Daley RJ, Jasper S (1977) Use of Nuclepore filters forcounting bacteria by fluorescence microscopy. Appl Environ Mi-crobiol 33:1225±1228
Holm E, Jensen V (1972) Aerobic chemoorganotrophic bacteria of aDanish beech forest. Oikos 23:248±260
Hooper DJ (1986) Extraction of free-living stages from soil. In:Southey JF (ed) Laboratory methods for work with plant and soilnematodes. Ministry of Agriculture, Fisheries and Food, Her Ma-jesty's Stationary Office, London, pp 5±30
Hunt HW, Cole CV, Klein DA, Coleman DC (1977) A simulationmodel for the effect of predation on bacteria in continuous cul-ture. Microb Ecol 3:259±278
Jocteur Monrozier L, Ladd JN, Fitzpatrick RW, Foster RC, RaupachM (1991) Components and microbial biomass content of size frac-tions in soils of contrasting aggregation. Geoderma 49:37±62
Kanazawa S, Filip Z (1986) Distribution of microorganisms, total bio-mass, and enzyme activities in different particles of brown soil.Microb Ecol 12:205±215
Kuikman PJ, van Elsas JD, Jansen AG, Burgers SLGE, van Veen JA(1990) Population dynamics and activity of bacteria and protozoain relation to their spatial distribution in soil. Soil Biol Biochem22:1063±1073
Lavelle P, Gilot C (1994) Priming effects of macroorganisms on mi-croflora: A key process of soil function? In: Ritz K, Dighton J,Giller KE (eds) Beyond the biomass. Wiley-Sayce, Chichester, pp173±180
Lee KE (1983) The influence of earthworms and termites on soil ni-trogen cycling. In: Lebrun P, Andre HM, De Medts A, GreÂgoire-Wibo C, Wauthy G (eds) New trends in soil biology. Proceedingsof the VIII. Intl Colloquium of Soil Zoology. Louvain-la-Neuve(Belgium). Dieu-Brichart, Ottgnies-Louvain-la-Neuve
Lee KE (1985) Earthworms. Their ecology and relationships withsoils and land use. Academic Press, Sydney
Lensi R, Clays-Josserand A, Jocteur Monrozier L (1995) Denitrifiersand denitrifying activity in size fractions of a mollisol under per-manent pasture and continuous cultivation. Soil Biol Biochem27:61±69
McClaugherty CA (1983) Soluble polyphenols and carbohydrates inthroughfall and leaf litter decomposition. Acta Ecologica EcolGener 4:375±385
Moody SA, Briones MJI, Piearce TG, Dighton J (1995) Selectiveconsumption of decomposing wheat straw by earthworms. SoilBiol Biochem 9:1209±1213
Nelson DW, Sommers LE (1982) Total carbon, organic carbon, andorganic matter. In: Page AL, Miller RH, Keeney DR (eds) Meth-ods of soil analysis, part 2, Chemical and microbiological proper-ties. 2nd edn. American Society of Agronomy, Soil Science So-ciety of America, Madison, Wisconsin, pp 539±579
Page FC (1976) An illustrated key to freshwater and soil amoebae.Freshwater Biological Association, Special Publication No. 34.Wilson & Son, Kendal, UK
Pedersen JC, Hendriksen NB (1993) Effect of passage through the in-testinal tract of detritivore earthworms (Lumbricus spp.) on thenumber of selected Gram-negative and total bacteria. Biol FertilSoils 16:227±232
Postma J, Hok-A-Chin CH, Van Veen JA (1990) Role of micronichesin protecting introduced Rhizobium leguminosarum biovar trifoliiagainst competition and predation in soil. Appl Environ Microbiol56:495±502
Pussard M, Rouelle J (1986) PreÂdation de la microflore effect desprotozoaires sur la dynamique de population bacteÂrienne. Protisto-logica 22:105±110
Richaume A, Steinberg C, Jocteur Monrozier L, Faurie G (1993) Dif-ferences between direct and indirect enumeration of soil bacteria:the influence of soil structure and cell location. Soil Biol Biochem5:641±643
Robert M, Chenu C (1992) Interactions between soil minerals and mi-croorganisms. In: Stotzky G, Bollag J-M (eds) Soil biochemistry,vol 7. Marcel Dekker, New York, pp 307±404
Roberts AM (1981) Hydrodynamics of protozoan swimming. In: Le-vandowsky M, Hunter SH (eds) Biochemistry and physiology ofprotozoa, vol. 4. 2nd edn. Academic Press, New York, pp 5±65
Rùnn R, Ekelund F, Christensen S (1995) Optimizing soil extract andbroth media for the enumeration of naked amoebae and hetero-trophic flagellates in soil. Pediobiologia 39:10±19
Satchell JE, Lowe DG (1967) Selection of leaf litter by Lumbricusterrestris. In: Graff O, Satchell JE (eds) Soil organisms. NorthHolland, Amsterdam, pp 102±119
Shipatalo MJ, Protz R, Tomlin AD (1988) Effect of diet on the feed-ing and casting activity of Lumbricus terrestris and L. rubellus inlaboratory culture. Soil Biol Biochem 20:233±237
Went JC (1963) Influence of earthworms on the number of bacteria inthe soil. In: Doeksen J, van der Drift J (eds) Soil organisms.North Holland, Amsterdam, pp 260±265
Wicklow DT (1992) Interference competition. In: Carroll GC, Wick-low DT (eds) The fungal community. Its organization and role inthe ecosystem. Marcel Dekker, New York, pp 265±274
Winding A (1994) Fingerprinting bacterial soil communities usingBiolog microtitre plates. In: Ritz K, Dighton J, Giller KE (eds)Beyond the biomass. Wiley-Sayce, Chichester, pp 85±94
Winding A, Binnerup SJ, Sùrensen J (1994) Viability of indigenoussoil bacteria assayed by respiratory activity and growth. Appl En-viron Microbiol 60:2869±2875
Wolters V, Joergensen RG (1992) Microbial carbon turnover in beechforest soils worked by Aporrectodea caliginosa (Savigny) (Oligo-chaeta:Lumbricidae). Soil Biol Biochem 24:171±177
Zhang BG, Rouland C, Lattaud C, Lavelle P (1993) Activity and ori-gin of digestive enzymes in gut of the tropical earthworm Pontos-colex corethrurus. Eur J Soil Biol 29:7±11
Zhang H, Schrader S (1993) Earthworm effects on selected physicaland chemical properties of soil aggregates. Biol Fertil Soils15:229±234
Zhang QL, Hendrix PF (1995) Earthworm (Lumbricus rubellus andAporrectodea caliginosa) effects on carbon flux in soil. Soil SciSoc Am J 59:816±823
