Earthworm Populations, Microbial Biomass and Coffee ... · Earthworm Populations, Microbial Biomass and Coffee Production in Different Experimental Agroforestry Management Systems

Earthworm Populations, Microbial Biomass and Coffee Production inDifferent Experimental Agroforestry Management Systems in

Costa Rica

YANIRIA SÁNCHEZ-DE LEÓN1*, ELIAS DE MELO2, GABRIELA SOTO2,JODI JOHNSON-MAYNARD1, AND JAVIER LUGO-PÉREZ1

1Plant, Soil and Entomological Sciences, Room 242, University of Idaho, Moscow ID 83844-2339, USA2Centro Agronómico Tropical de Investigación y Enseñanza (CATIE), Escuela de Posgrado, Sede Central, CATIE 7170,

Turrialba, Costa Rica.*Corresponding author: [email protected]

ABSTRACT.—Coffee management systems are diverse and can have significant impacts on soil biota.Earthworms are important members of the soil biotic community and may enhance soil fertility in agricul-tural soils. The impact of coffee management on earthworms and other factors related to productivity are notwell known. Our main objective was to determine the impact of coffee management on earthworm popu-lations, microbial biomass, and coffee production in Turrialba, Costa Rica. Three experimental coffee (Coffeaarabica ‘Caturra’) management systems were studied: full sun, shade with Erythrina poepiggiana and shadewith Terminalia amazonia. Within these systems there were three sub-treatments with different levels ofinputs: high conventional or medium conventional sub-treatments with chemical fertilizer, herbicide andfungicide additions, and organic inputs in the shade systems. We found the earthworms Pontoscolex core-thrurus and Metaphire californica in shade treatments, but M. californica was absent from sun systems andTerminalia high conventional systems. Mean earthworm density was lowest in the sun high conventionaltreatment (63 ind. m−2), higher in medium conventional treatments (from 108 ind. m−2 to 225 ind. m−2), andhighest in Terminalia organic systems (334 ind. m−2). In 2003, coffee production was highest in sun highconventional treatments (17.6 Mg ha−1), but in 2004 it was highest in the Terminalia organic treatments (9.2Mg ha−1). Microbial biomass was not different among treatments, but was correlated with total earthwormdensity and biomass. Results indicate that shade organic management favors high earthworm density andbiomass and that coffee yields under organic management can equal those of conventional systems.

KEYWORDS.—sun and shade coffee, organic, exotic earthworms, agroforestry, Costa Rica

INTRODUCTION

Coffee is one of the most important agri-cultural commodities worldwide occupy-ing 10,095,276 hectares of land (FAOSTATdata 2004), and providing economic sup-port for 20 to 25 million people (Aguilarand Klocker 2000; Somarriba et al. 2004). InCosta Rica and other tropical countries, cof-fee exportation is considered a critical com-ponent of the country’s economy (Aguilarand Klocker 2000; Somarriba et al. 2004).Because of its extensive cultivation and eco-nomic importance, there is considerable in-terest in improving the sustainability of

coffee production in Costa Rica (Aguilarand Klocker 2000).

Management of coffee-producing landsin Costa Rica varies widely; with specifictechniques depending on site conditionsand socioeconomic pressures (Beer et al.1998; Aguilar and Klocker 2000; Somarribaet al. 2004). Coffee management systemsoccur as a continuum from open-sunmonocultures to multistrata coffee polycul-tures with different levels of agrochemicalinputs (Somarriba et al. 2004). Sun coffeemanagement is recommended in cloudy ar-eas of elevation higher than 1000 m and inareas where agrochemical inputs are avail-able (Beer et al. 1998; Aguilar and Klocker2000). In areas that receive more intense ra-diation, the use of shade trees is considered

ms. received November 14, 2005; accepted June 1,2006.

Caribbean Journal of Science, Vol. 42, No. 3, 397-409, 2006Copyright 2006 College of Arts and SciencesUniversity of Puerto Rico, Mayaguez

397

beneficial because they decrease exposureof coffee plants to extreme heat or cold.Shade trees will also reduce wind speed,control erosion, and help maintain soil fer-tility (Beer et al. 1998; Aguilar and Klocker2000; Somarriba et al. 2004). Within shadecoffee systems, organic production repre-sents a new market for farmers and an op-portunity to increase income from coffeefarms (Beer et al. 1998; Lyngbæk et al. 2001;Somarriba et al. 2004). Factors such as bio-logical conservation, low environmentalimpact, valuation of environmental ser-vices (e.g., water and soil conservation, car-bon sequestration) and the fall of interna-tional prices have placed organic coffee in apreferential place in the market over coffeeproduced with conventional practices (Beeret al. 1998; Lyngbæk et al. 2001; Somarribaet al. 2004).

Agricultural management in tropicalagroecosystems has a significant impact onsoil macrofauna communities, microbialbiomass and soil organic matter (SOM)content (Woomer et al. 1994; Edwards andBohlen 1996; Lavelle et al. 2001). These soilproperties are essential for the maintenanceof soil fertility and sustainable agriculturalproduction in tropical regions throughtheir influence on soil biological, physicaland chemical properties (Scholes et al. 1994;Woomer et al. 1994; Lavelle et al. 2001). Ag-ricultural management practices that mini-mize negative impacts on soil macrofaunaand SOM are critical for maintaining thesustainability of agroecosystems (Woomeret al. 1994; Brown et al. 1994; Lavelle et al.2001).

Soil macrofaunal populations play an im-portant role in the degradation of organicmatter, nutrient cycling, carbon sequestra-tion, and soil aggregation and macroporeformation through their behavioral activi-ties (Edwards and Bohlen 1996; Sollins etal. 1996; Groffman and Bohlen 1999; DeGoede and Brussaard 2002). Earthwormsmay have a positive or negative impact ofvariable magnitude on agricultural produc-tion through a variety of biological, chemi-cal, and physical mechanisms (Brown et al.1999; Senapati et al. 1999; Lavelle et al.2001). As discussed by Brown et al. (1999),biological mechanisms include competition

between earthworms and plants for water,increase in microbial biomass and micro-bial activity, alteration of carbon and nitro-gen status, reduced damage from parasiticnematodes, and increased plant nutrientuptake. Chemical mechanisms include in-creased nutrient availability, recycling ofnutrients from dead worm tissues and ex-cretions, and accelerated nutrient releasefrom decomposing plant residues. Physicalmechanisms include changes in soil aggre-gation, porosity, aeration, water infiltrationand bulk density.

Maintenance of SOM is of critical impor-tance within tropical agroecosystems, be-cause it is the principal source of energy forthe soil ecosystem, and a nutrient sourcefor plants (Woomer et al. 1994; Carter 1996;Schroth et al. 2001; Coleman et al. 2004). Inaddition, SOM improves soil physical andchemical properties important for plantproductivity through increasing aeration,cation exchange capacity and water hold-ing capacity while reducing erosion(Woomer et al. 1994; Carter 1996). Concep-tually, SOM is described as separate poolsthat include the microbial biomass, mac-roorganic matter or coarse fraction, andmineral associated organic matter or finefraction (Carter 1996). Changes in soil mi-crobial biomass and SOM fractions in dif-ferent agricultural management systemshave been proposed as early indicators ofchange in SOM content (Powlson 1994;Woomer et al. 1994; Alvarez and Alvarez2000).

Few studies of coffee management sys-tems have focused on measuring and com-paring the soil biological properties men-tioned above in relation to coffee yields.Furthermore, Schroth et al. (2001) identifiedrelationships among SOM dynamics, soilmicroorganisms and soil fauna in tropicalagroforestry systems as a topic that de-serves more research attention. The objec-tives of this study were to describe andcompare earthworm populations, microbialbiomass and coffee production among suncoffee plantations and shade coffee systemswith Erythrina poepiggiana, a nitrogen fixer,and with Terminalia amazonia, a timber spe-cies in Turrialba, Costa Rica. These coffeeplantations had different levels of agro-

SÁNCHEZ-DE LEÓN ET AL.398

chemical inputs (high conventional, me-dium conventional and intensive organic).In addition, we wanted to describe the re-lationship among the soil biological prop-erties under these different managementsystems. We hypothesized that earthwormspecies composition, density and biomass,microbial biomass and SOM fractionswould be higher under shade-organic thanin the other management systems. We alsohypothesized that there would be positiverelationships among earthworms and mi-crobial biomass.

MATERIALS AND METHODS

Study sites

The study was conducted at the experi-mental coffee fields located at the “CentroAgronómico Tropical de Investigación yEnseñanza” (CATIE) facilities in Turrialba,Costa Rica (9°53’44”N, 83°40’7”W). Thestudy area was located at 600 m above sealevel, where annual precipitation was ap-proximately 2600 mm yr−1 and the meanannual temperature was 21.8°C (ZuluagaPeláes 2004). Soils were classified as TypicEndoaquepts and Typic Endoaquults usingthe USDA Soil Taxonomy classification sys-tem (Soil Survey Staff 1999), with extensivedrainage channels. The experimental fieldswere planted with two coffee cultivars,Coffea arabica L. ‘Caturra’, and in minor ex-tent, C. arabica ‘Costa Rica 95’. However,the present study was conducted in the ar-eas planted with Caturra exclusively. Theexperimental design was an incompletefactorial treatment arrangement, in a ran-domized complete block design with threeblocks and one replicate per block. Previ-ous to the development of this study, thesite had been in sugarcane (Saccharum offi-cinarum L.) for at least 5 years.

Among the coffee management treat-ments established were sun and shade witheither Erythrina poepiggiana (Walpers) O.F.Cook (nitrogen fixer), or with Terminaliaamazonia (J.F. Gmel.) Exell a timber speciesthat does not fix nitrogen. The number ofshade trees in the Erythrina treatments was49 trees ha−1 and in the Terminalia treat-ments was 63 trees ha−1. The total area of

the sun, shade with Erythrina and shadewith Terminalia treatments was 4000 m2,4088 m2 and 3200 m2, respectively. Thesemanagement systems were initially plantedin 2000, but organic treatments had to bereestablished in 2001. Initially, the organictreatments were established with a lowershade tree density and with young shadetrees that were too short to provide suffi-cient shade during the first year. The rela-tively high light intensity and low nutrientstatus of the soil resulted in a very highmortality of coffee plants. In 2001, a tempo-rary shade (Ricinus communis L.) was intro-duced in the organic treatments, and in2004 these trees were eliminated as appro-priate shade levels were obtained.

Sub-treatments within the main coffeemanagement treatments included: high,medium conventional and intensive or-ganic. There was no organic sub-treatmentwithin the sun coffee systems. Growers inthe area do not use this combination due tothe excessive cost of weed control. In highconventional treatments, agrochemical fer-tilizers and fungicides were used as in com-mercial fields and herbicides were appliedto eliminate all the herbaceous cover. In themedium conventional treatment, fertilizerapplications were half the amount of thosewithin the high conventional treatment,and fungicides were used. Herbicides wereapplied to eliminate the herbaceous vegeta-tion among individual coffee plants, but se-lected herbaceous species were left amongcoffee plants in a row. In organic treat-ments, organic fertilizers were used, noherbicides were applied, and undesiredherbaceous species were removed manu-ally and mechanically with a string trim-mer. Fertilizer, herbicide and fungicide ap-plication rates within each sub-treatmentare shown in Table 1. The annual inputs ofkey nutrients are shown in Table 2. In thehigh conventional Erythrina treatmentsonly, the branches of the Erythrina treeswere completely pruned twice a year as istypically done in the coffee farms of theregion. All the organic material was left onthe soil surface after pruning events. Inshade medium conventional and organictreatments, the shade was regulatedthrough pruning to control for irregular

COFFEE MANAGEMENT AND EARTHWORMS 399

branching and, trees were never com-pletely pruned. A detailed description ofthe sub-treatments and pruning regimes ispresented in Zuluaga Peláes (2004) andMontenegro Gracia (2005). Nutrient addi-tions to the soil through the decompositionof branches removed during pruning werenot included in Table 2.

Earthworm collection

Earthworms were sampled in each of thethree replicates of each treatment duringJuly 2003. Five randomly assigned soil pitsof 25 × 25 cm and 10 cm depth were locatedamong the coffee plants rows. Earthwormswere collected by hand sorting in the fieldand taken to the laboratory for countingand weighing. We obtained earthworm

fresh weight after earthworms were rinsedwith water and dried with paper towels onthe same day of collection. Individualswere classified into morphospecies basedon external morphological features, and asub-sample of individuals from each groupwas preserved. The eco-taxonomic guidefor peregrine earthworms developed byBlakemore (2002) was used to identify theearthworm species.

Microbial biomass and soil organic matter(SOM) fractionation

Soils were sampled during July and Au-gust 2004 in each replicate of the sun me-dium conventional, and Erythrina and Ter-minalia medium conventional and organictreatments. Microbial biomass and SOMmeasurements were not made in high con-ventional treatments. Twelve soil sampleswere collected from the upper 10 cm ofeach sub-treatment replicate and combinedinto a single representative sample (∼500 g)per replicate. Soil samples were collectedfrom the area between rows of coffeeplants.

Microbial biomass carbon (C) was mea-sured using the chloroform fumigation-extraction procedure using 0.5 M potas-sium sulfate (K2SO4) for extraction, and theNelson and Sommer’s method as described

TABLE 1. Fertilizer and herbicide applications and other inputs in the sub-treatments of the experimentalcoffee agroforestry systems

Sub-treatment Fertilizer application Herbicide application Other inputs

Highconventional

800 kg ha−1 yr−1 18-15-6-2 (N,P, K, Mg and B) 45 kg ha−1

yr−1 NH4NO3

Foliar application: B, Zn(three times per year)

10 ml l−1 Roundup� toeliminate all herbaceousvegetation

fungicides: 2.5 g l−1 H2Oper block of Atemi orCopper sulfate (twice ayear)

Mediumconventional

400 kg ha−1 yr−1 18-15-6-2 (N,P, K, Mg and B) 45 kg ha−1

yr−1 NH4HO3

Foliar application: B, Zn(once a year)

10 ml l−1 Roundup� toeliminate herbaceousspecies among coffeeplants within a row

fungicides: 2.5 g l−1 H2Oper block of Atemi orCopper sulfate (once ayear)

Organic 20 tons ha−1 yr−1 coffee pulp7.5 tons ha−1 yr−1 chickenmanure

200 kg ha−1 yr−1 KMAG200 kg ha−1 yr−1 Phosphoric

rock

No application ofherbicides. Weeds wereremoved manually andmechanically with astring trimmer

No application offungicides

TABLE 2. Fertilization rates for nitrogen (N), phos-phorus (P) and potassium (K) in the sub-treatments ofthe experimental coffee agroforestry systems. Nutrientinputs from decomposition of shade tree biomasswere not considered in these estimates.

Sub-treatmentN (kg ha−1

yr−1)P (kg ha−1

yr−1)K (kg ha−1

yr−1)

Highconventional 221 17 98

Mediumconventional 110 8 49

Organic 365 205 327


in Anderson and Ingram (1992). The factorfor microbial C calculation was 2.64(Anderson and Ingram 1992). Soil organicmatter fractions were isolated by particlesize fractionation (Cambardella and Elliott1992). Soils were dispersed with sodiumhexametaphosphate (NaPO3)6 and passedthrough a 53-�m sieve. The coarse fractionwas defined as SOM greater than 53 �m,and the fine fraction was defined as SOMsmaller than 53 �m. Organic carbon per-cent of each soil fraction was determinedusing the Nelson and Sommers’s titrationmethod for total organic carbon in soil ex-tracts (Anderson and Ingram 1992).

Coffee production

Coffee yield was defined as ripe coffeeberries that were hand picked from the cof-fee plants from June 2003 to January 2004,and from June 2004 to January 2005. Here-after we will refer to these coffee harvestingperiods as production 2003 and production2004, respectively. The coffee productionwas measured in fanegas per hectare, thestandard measurement for coffee bean pro-duction in Costa Rica. One fanega equals 2hectoliters or 0.2 m3. We used the standardaccepted conversion (1 fanega = 258 kg or0.258 Mg) to convert yields to a weight perunit area basis.

Data analysis

We used a one-way analysis of variance(ANOVA) using the Mixed Model to com-pare total earthworm density and freshweight, earthworm density and freshweight by species, and microbial biomass Cand SOM fractions (dependent variables)among the eight combinations of coffeemanagement systems (independent vari-able). Where the data violated the normal-ity and homogeneity of variance assump-tions, we performed a Friedman non-parametric ANOVA using the GeneralLinear Model to compare earthworm den-sity and fresh weight for the specific earth-worm species. Treatment means were com-pared using the Tukey’s Least SquareMeans analysis when statistical differenceswere detected. We used a repeated mea-

sure ANOVA using the General LinearModel to compare coffee production be-tween years 2003 and 2004 and amongtreatments. The data for coffee productionwere log transformed before the analysis tofulfill the normality and homogeneity ofvariance assumptions. Simple linear corre-lations were used to determine the relation-ship between total earthworm density andmicrobial biomass C, and the relationshipbetween earthworm fresh weight and mi-crobial biomass C.

To test the effect of shade and level ofagrochemical inputs on earthworm density,fresh weight and coffee production inde-pendently, we used a linear contrast test.We used this test instead of a two-wayANOVA due to the incomplete design. Totest the effect of shade, we contrasted thesun treatments results versus the shadetreatments, and Erythrina treatments versusTerminalia treatments. To test the effect ofagrochemical inputs, we contrasted the or-ganic versus the conventional treatments,and high versus medium conventionaltreatments. All the statistical analyses werepreformed using SAS� Version 9.1 statisti-cal software (SAS Institute Inc. 2004) andthe significance level was set at � = 0.05.

RESULTS

Earthworm populations

We found differences among coffee man-agement systems for earthworm total den-sity (Fig. 1a) and fresh weight (Fig. 1b).Mean earthworm density and fresh weightvalues were significantly lower in the sunhigh conventional treatments (63 ind. m−2;20 g m−2) than in medium conventional suntreatments (91 ind. m−2, 28 g m−2). UnderErythrina shade, earthworm density andfresh weight were lowest in the high conven-tional treatment (131 ind. m−2, 55 g m−2),intermediate in the medium conventionaltreatment (186 ind. m−2, 68 g m−2), andhighest in the organic treatment (265 ind.m−2, 99 g m−2). Similarly under Terminaliashade, earthworm density was lowest inthe high conventional treatment (73 ind.m−2, 24 g m−2), intermediate in the mediumconventional treatment (108 ind. m−2, 32 g


m−2), and highest in the organic treatment(334 ind. m−2, 118 g m−2).

Only the exotic earthworm species, Pon-toscolex corethrurus (Müller) and Metaphirecalifornica (Kinberg) were found in thestudy sites. Pontoscolex corethrurus wasdominant and represented 76-100% of thetotal earthworm density (Fig. 1c) and freshweight (Fig. 1d) in all treatments studied.Both species were present in shade treat-ments, however M. californica was absentfrom sun treatments and Terminalia highconventional treatments. Metaphire califor-nica mean values for density (Fig. 1e) andfresh weight (Fig. 1f) were lowest in Termi-nalia high conventional treatment (3.2 ind.m−2 and 0.73 g m−2) and highest in the or-ganic systems (51.2 ind. m−2 and 16 g m−2)(P < 0.001). The contrast analysis showed

that there was an effect of shade and levelof agrochemical input for both earthwormspecies densities (Table 3). However, typeof shade (Terminalia vs. Erythrina) andtype of conventional system (high vs. me-dium) had no effect on P. corethrurus freshweight.

Microbial Biomass C and soil organic matter(SOM) fractionation

Mean microbial biomass ranged from 583mg kg−1 in Terminalia medium conven-tional treatments to 913 mg kg−1 in Ery-thrina organic treatments (Fig. 2) and didnot differ significantly among treatments.Percentages of SOM fractions were statisti-cally different within management sys-tems. The fine fraction percent by weight

FIG. 1. Mean (±SE) earthworm density and fresh weight from total earthworms (a, b), Postocolex corethrurus (c,d) and Metaphire californica (e, f) under sun, Erythrina shade and Terminalia shade coffee management systems.Common letters indicate no significant differences among coffee management systems. Friedman non-parametric ANOVA analysis was used to determine significant differences among treatments for M. californicaonly.


was always higher than the coarse fractionwithin each treatment (Fig. 3a). Differencesamong treatments were observed withhigher mean percent of the fine SOM frac-tion in Erythrina medium conventionaltreatments and lowest mean percent in Ter-minalia organic systems (P = 0.035). The in-verse pattern was observed for the coarseSOM fraction with the highest mean coarseSOM fraction in Terminalia organic systemsand the lowest in Erythrina medium con-ventional systems (P = 0.042) (Fig. 3a).

Organic carbon content did not differ be-tween fractions across management sys-tems. Mean organic carbon percentages inthe fine SOM fraction ranged from 5.3% inErythrina medium conventional treatmentsto 6.3% in Terminalia organic systems (Fig.3b). Mean organic carbon percentages inthe coarse SOM fraction ranged from 4.4%in Terminalia medium conventional treat-ments to 5.3% in Erythrina organic systems(Fig. 3b).

Significant positive correlations werefound among some of the measured vari-ables in this study. Microbial biomass was

positively correlated with total earthwormdensity (y = 1.03x + 511.86, r = 0.65, P =0.008) and with total earthworm freshweight (y = 2.89x + 533.99, r = 0.63, P =0.011) across treatments. No significant cor-relations were found among SOM fractionsand carbon percent within fractions andmicrobial biomass, or SOM fractions andtotal earthworm density and total earth-worm fresh weight.

Coffee production

Coffee production mean values werehigher in the sun high conventional (17.6Mg ha−1) and sun medium conventional(14.7 Mg ha−1) compared to that measuredin the rest of the treatments for year 2003(P = 0.0045) (Fig. 4a). Coffee production un-der Erythrina shade was lowest in the or-ganic treatments (3.2 Mg ha−1) intermediatein medium conventional (9.9 Mg ha−1) andhighest in high conventional treatments(14.5 Mg ha−1). Similarly, coffee productionunder Terminalia shade was lowest for or-ganic treatments (4.5 Mg ha−1) intermediate

TABLE 3. Statistics of linear contrast tests for the density and fresh weight of the earthworm species Pontoscolexcorethrurus and Metaphire californica and coffee production 2003 (degrees of freedom for all test were equal to one)

Treatments contrasted F P

Earthworm densityPontoscolex corethrurus Sun vs. shade 29.55 <0.01

Erythrina vs. Terminalia 5.19 0.02Organic vs. Conventional 86.72 <0.01High vs. medium conventional 4.57 0.04

Metaphire californica Sun vs. shade 26.10 <0.01Erythrina vs. Terminalia 8.88 <0.01Organic vs. Conventional 40.78 <0.01High vs. medium conventional 7.66 0.02

Earthworm fresh weightPontoscolex corethrurus Sun vs. shade 28.82 <0.01

Erythrina vs. Terminalia 1.01 0.32Organic vs. Conventional 99.52 <0.01High vs. medium conventional 1.85 0.18

Metaphore californica Sun vs. shade 19.98 <0.01Erythrina vs. Terminalia 10.94 <0.01Organic vs. Conventional 21.61 <0.01High vs. medium conventional 3.20 0.01

Coffee production 2003 Sun vs. shade 4.80 0.04Erythrina vs. Terminalia 0.10 0.75Organic vs. Conventional 21.75 <0.01High vs. medium conventional 1.18 0.29


in medium conventional (10.1 Mg ha−1) andhighest in the high conventional (13.3 Mgha−1). The contrast analysis showed thattype of shade (Terminalia vs. Erythrina) andlevel of inputs (high vs. medium) did notsignificantly influence coffee productionfor year 2003 (Table 3).

In contrast to 2003 results, coffee produc-tion in 2004 was not statistically differentamong treatments (Fig. 4b). Coffee produc-tion ranged from in 5.1 Mg ha−1 in sun highconventional treatments to 4.5 Mg ha−1 inmedium conventional treatments. Coffeeproduction within Erythrina sub-treatmentswas not significantly different, however theorganic systems produced the most coffee(8.2 Mg ha−1), followed by high conven-tional (6.7 Mg ha−1), and medium conven-tional (4.7 Mg ha−1). Terminalia treatmentsproduced similar results to those found inErythrina systems, with organic systemsproducing 9.2 Mg ha−1, followed by highconventional (6.6 Mg ha−1) and lowest pro-duction in the medium conventional (2.9Mg ha−1). The repeated measures ANOVAalso showed a significant interaction be-tween treatment and time (P = 0.0049).

DISCUSSION

Our hypothesis, that earthworm speciesdiversity, density and biomass would behigher under shade-organic compared tothe other management systems was par-tially supported. We found the two earth-

worm species in the shade-organic treat-ments and higher earthworm density andfresh weight than in the remaining treat-ments. In addition, we found that earth-worm species composition was differentamong coffee management systems withMetaphire californica being absent from suntreatments and Terminalia high conven-tional treatments, and a clear dominance ofPontoscolex corethrurus. There are very fewpublished studies on earthworm speciescomposition in Costa Rican agroecosystems(Fragoso et al. 1999). Fraile (1989) described

FIG. 2. Mean (±SE) microbial biomass C under sun,Erythrina shade and Terminalia shade coffee manage-ment systems. Microbial biomass C was not measuredfor high conventional treatments. There was no sig-nificant difference among treatments for microbialbiomass.

FIG. 3. Mean (±SE) SOM fractions (a) and organiccarbon percent within SOM fractions (b) under sunmedium conventional (S-MC), Erythrina medium con-ventional (E-MC), Erythrina organic (E-O), Terminaliamedium conventional (T-MC), and Terminalia organic(T-O) coffee management systems. SOM fractions andorganic percent carbon were not measured for highconventional treatments. Common letters indicate nosignificant differences among coffee management sys-tems. There was no significant difference among treat-ments for organic carbon percent.


earthworm populations for non-grazedpastures and pastures associated with thetree species Erythrina poeppigiana and thegrass species Cynodon plectostachyus also inthe region of Turrialba, Costa Rica. Fraile(1989) found that P. corethrurus was largelydominant in all the agroecosystems stud-ied, M. californica was present in highestdensities in pastures associated with E. po-eppigiana, and only one native earthwormspecies Glossodrilus nemoralis was found inlow densities. In a more recent study,Lapied and Lavelle (2003) sampled earth-worms in banana plantations, primary for-ests and villages located along the Car-ibbean coast of Costa Rica. Lapied andLavelle (2003) found that P. corethrurus wasthe dominant species at all study sites ex-cept for those within the Tortuguero Na-tional Park primary forest where it was ab-sent, and densities were highest in banana

plantations (361 individuals m−2). Our re-sults, therefore, are consistent with paststudies showing that P. corethrurus is adominant earthworm in agricultural sys-tems in Costa Rica, and that this species iscapable of living under a variety of condi-tions.

Our results also show that M. californicais not tolerant of habitat conditions undersun and Terminalia high conventional treat-ments. In contrast, the presence of M. cali-fornica in the Erythrina high conventionaltreatments suggests a shade species effecton the earthworm species composition. Theecological classification of these earth-worms provides a possible explanation forour results. Metaphire californica is classifiedas an epi-anecic earthworm (Fragoso et al.1999) which indicates that it lives in the sur-face soil, but feeds on leaf litter (Barois et al.1999). Pontoscolex corethrurus is classified asa meso-humic earthworm which meansthat it lives in the upper 20 cm of the soiland feeds on soil in the upper 10 cm (Baroiset al. 1999). In sun treatments and in Termi-nalia high conventional treatments, leaf lit-ter quantity and/or quality is limited tothat provided by the coffee plants or coffeeand Terminalia plants, and this might not besufficient to maintain M. californica popula-tions. Furthermore, our contrast analysisshowed that the shade species (Erythrinavs. Terminalia) had a significant effect onboth earthworm species. These resultsshow that Erythrina can provide enoughaboveground and/or belowground inputsto maintain both M. californica and P. core-thrurus populations. Other indirect influ-ences of shade trees and their biomass in-puts, such as differences in microclimate,and protection from predators by the litterlayer, can also be affecting the presence ofM. californica in the Erythrina agroforestrycoffee systems.

Coffee management system had a signifi-cant impact on earthworm density andfresh weight. Organic systems had thehighest density and fresh weight of totalearthworms, P. corethrurus individuals andM. californica individuals followed by shademedium conventional treatments. The low-est mean total earthworm density and freshweight was in sun and shade high conven-

FIG. 4. Mean (±SE) coffee production (measured asripe berries) under sun, Erythrina shade and Terminaliashade coffee management systems for year 2003 (a)and 2004 (b). Common letters indicate no significantdifferences among coffee management systems. Therewas no significant difference among treatments inyear 2004.


tional treatments. These results suggestthat organic and shade medium conven-tional coffee management systems pro-vided more favorable conditions for earth-worm growth and reproduction than in sunand shade high conventional treatments.Brown et al. (1999) and Lavelle et al. (2001)suggested agricultural management prac-tices that promote a diverse earthwormcommunity in tropical agroecosystems.These suggestions included: mulching, re-stricted use of pesticides, incorporation oflegumes, returning plant residues to thesoil, and use of organic manures (Brown etal. 1999; Lavelle et al. 2001). In our study,the organic and shade medium conven-tional coffee management systems incorpo-rated most of these practices, meanwhilesun and high conventional management in-cluded few or none of them. The presenceof herbaceous plants could also influenceearthworm populations. Herbaceous plantsprovide aboveground and/or below-ground organic inputs that can be anotherfood source for earthworms (Edwards andBohlen 1996). In addition, the high use ofagrochemicals (e.g., herbicides and fertiliz-ers) can have a detrimental effect on earth-worm populations (Edwards and Bohlen1996; Brown et al. 1999) in the high conven-tional treatments when compared to or-ganic and medium conventional systems.

Contrary to the prediction of our first hy-pothesis, microbial biomass and carbon inthe fine and coarse SOM fractions did notdiffer among coffee sun and shade mediumconventional and organic treatments. How-ever, fine SOM fractions were higher in thesun and Erythrina treatments and lower inthe Terminalia treatments, and the inverseresults were observed with the coarse SOMfraction. These results show that havingErythrina or Terminalia as a shade speciescan influence SOM dynamics. Beer et al.(1998) and Schroth (2001) emphasize theimportance of biomass production by theshade species as a way to maintain highSOM levels, and the associated soil fertilitybenefits, in tropical agroecosystems. Mon-tenegro Gracia (2005) measured the bio-mass added to the soil-plant system afterpruning of the shade species Erythrina andTerminalia in these experimental fields, and

found a higher maximum contributionfrom Erythrina (11790 kg ha−1 yr−1) thanfrom Terminalia (3453 kg ha−1 yr−1). Fur-thermore, Beer (1988) showed that the in-fluence of nitrogen fixing trees, such as Ery-thrina, over SOM dynamics in fertilizedplantations is more important than their in-fluence on the soil nitrogen status. ZuluagaPeláes (2004) found higher total C and N inthe top 10 cm of the soil in the Erythrinaorganic treatments than in the Erythrinamedium conventional treatments, and sug-gested that the type of agricultural manage-ment (organic vs. agrochemical) had a sig-nificant effect over the C and N dynamicsof these experimental agroforestry systems.

Coffee production in 2003 was greater inthe high and medium conventional treat-ments in both sun and shade than withinany of the organic treatments. In 2004, how-ever, no significant difference was found.We also found a significant effect of the in-teraction between time and treatment inthese coffee agroecosystems. This could beexplained by the one-year lag in the estab-lishment of plants in the organic treat-ments. For year 2003 results, we are com-paring three-year-old coffee plants in highand medium conventional treatments withtwo-year-old plants in the organic treat-ments. The coffee plants in the organic sys-tems were too young to produce a mean-ingful coffee harvest in 2003; however theseplants were mature enough to have highcoffee production the following year.

When high levels of agrochemical inputsare used, sun coffee management systemsusually have higher yields than shadegrown coffee (Beer et al. 1998). A generalpattern of higher coffee yields in systemswith sun coffee however, has not beendemonstrated (Beer et al. 1998). Lyngbæk etal. (2001) performed a 3 year, paired com-parison of coffee production in organic cof-fee farms and conventionally managedfarms, and found that organic farms hadmean coffee yields 22% lower than conven-tional farms. However, yields for some or-ganic farms were higher than their conven-tional farm counterparts depending oninputs and labor of farm management(Lyngbæk et al. 2001). Our results for year2004 support Lyngbæk et al. (2001) findings


that intensive organic agroforestry systemscan have equal or higher coffee productionas compared to conventional systems if suf-ficient organic amendments and labor offarm management are provided. Our or-ganic management systems had higheryields than most of the organic coffee pro-ducers in the Turrialba region. The coffeeproduction in our shade organic manage-ment systems had mean values of 8.2 Mgha−1 and 9.2 Mg ha−1 in the Erythrina andTerminalia treatments respectively, whilethe organic producers of the area produceless than 3.9 Mg ha−1 due to their minimalmanagement practices. A recent studyshowed that more than 50% of the Turri-alba organic coffee farmers do not applyfertilizers at all, prune or replant coffeeplants in their farms (C. M. Porras personalcommunication).

Our second hypothesis was that therewould be positive relationships betweenearthworm populations and microbial bio-mass. We found a positive correlation be-tween microbial biomass and earthwormdensity and biomass. These results suggestthat there exist positive interactions be-tween earthworms and microbes, and/orthat the conditions that positively influenceearthworm populations also positively in-fluence microbial biomass in these manage-ment systems. Inoculations of P. corethrurushave been shown to increase microbial bio-mass in microcosm (Araujo et al. 2004) andpot (Pashanasi et al. 1992) experiments.Earthworms can have a direct positive in-fluence on microbial biomass through theproduction of mucus and other excretedcompounds that serve as nutrient-rich sub-strates for microorganisms (Pashanasi et al.1992; Araujo et al. 2004). Our data for earth-worms and microbial biomass were takenone year apart, during the same time of theyear (July). We expect, however, that in theabsence of disturbance, the relationship be-tween earthworm density and biomass andmicrobial biomass will be consistent fromyear to year. Although earthworm densitymay fluctuate from year to year, differencesbetween tillage management systems havebeen found to be consistent (Butt et al1999). For future studies, we suggest that

both variables are measured at the sametime.

The influence of microbial biomass andearthworm density and biomass on coffeeproduction is less clear than the influenceof earthworms on microbial biomass. A di-rect relationship between these soil biologi-cal parameters and coffee yields has notbeen described. Brown et al. (1999) per-formed a meta-analysis of earthworm ef-fects on crop productivity. Their analysisshowed that earthworm impact was mostlyon plant root growth, shoot growth andwhole plant growth and no significant ef-fect was observed in other aspects of plantproduction (e.g. grain production). Coffeeagroecosystems were not part of the Brownet al. (1999) analysis. However, an impor-tant consideration of the studies used inBrown et al. (1999) is that earthworm popu-lations were manipulated specifically tomeasure their effects on plant production.In our study, earthworm populations werenot manipulated. To elucidate the effect ofearthworm populations and microbial bio-mass on coffee yields, experimental ma-nipulation of these variables is necessary.We cannot demonstrate long-term benefitsof any of these systems since our data werecollected only 3 and 4 years after the treat-ments were established. This is especiallytrue considering that it normally takes 5 to6 years after coffee planting to get a profit-able harvest (Aguilar and Klocker 2000).However, a more intense management re-gime that includes fertilization, weed man-agement, and pruning, among other prac-tices for organic systems may increasecoffee production and earthworm popula-tions in the local organic coffee farms.

Acknowledgments.—We are grateful to theNational Science Foundation—IntegrativeGraduate Education and Research Trainee-ship Program no. 0114304, CATIE and theUniversity of Idaho for financial support.We are thankful to Luko Hilje, Jeremy Hag-gar, Charles Staver and Nilsa Bosque-Pérezfor logistical support, and to D. J. Lodgeand two anonymous reviewers for theircomments on an earlier draft of this paper.We also thank Carlos Cordero and LuisRomero for their help in the field.


LITERATURE CITED

Aguilar, B., and J. Klocker. 2000. The Costa Rican cof-fee industry. In Quantifying Sustainable Develop-ment: The Future of Tropical Economies, ed. C. S. Hall,P. Van Laake, G. Leclerc, C. Leon Perez, 595-627.New York: Academic Press.

Alvarez, R., and C. R. Alvarez. 2000. Soil organic mat-ter pools and their associations with carbon min-eralization kinetics. Soil Sci. Soc. Am. J. 64:184-189.

Anderson, J. M., and J. S. I. Ingram. 1992. Tropical SoilBiology and Fertility: A Handbook of Methods.Wallingford: CAB International.

Araujo, Y., F. J. Luizao, and E. Barros. 2004. Effect ofearthworm addition on soil nitrogen availability,microbial biomass and litter decomposition in me-socosms. Biol. Fertil. Soils 39:146-152.

Barois, I., et al. 1999. Ecology of earthworm specieswith large environmental tolerance and extendeddistribution. In Earthworm Management in TropicalAgroecosystems, ed. P. Lavelle, L. Brussaard, P.Hendrix, 57-85. New York: CABI Publishing.

Beer, J., D. Muschler, D. Kass, and E. Somarriba. 1998.Shade management in coffee and cacao planta-tions. Agrofor. Syst. 38:139-164.

Beer, J. W. 1988. Litter production and nutrient cyclingin coffee (Coffea arabica) or cacao (Theobroma cacao)plantations with shade trees. Agrofor. Syst. 7:103-114.

Blakemore, R. J. 2002. Cosmopolitan Earthworms—anEco-Taxonomic Guide to the Peregrine Species of theWorld. 1st ed. CD-ROM. Australia: VermEcology.

Brown, G. G., et al. 1999. Effects of earthworms onplant production in the tropics. In Earthworm Man-agement in Tropical Agroecosystems, ed. P. Lavelle, L.Brussaard, P. Hendrix, 87-148. New York: CABIPublishing.

Brown, S., J. M. Anderson, P. J. Woomer, M. J. Swift,and E. Barrios, 1994. Soil biological processes intropical ecosystems. In The Biological Management ofTropical Soil Fertility, ed. P. L. Woomer and M. J.Swift, 15- 46. New York: John Wiley & Sons.

Butt, K. R., M. J. Shipitalo, P. J. Bohlen, W. M. Ed-wards, and R. W. Parmelee. 1999. Long-termtrends in earthworm populations of cropped ex-perimental watersheds in Ohio, USA. Pedobiologia43:713-719.

Cambardella, C. A., and E. T. Elliot. 1992. Particulatesoil organic-matter changes across a grassland cul-tivation sequence. Soil Sci. Soc. Am. J. 56:777-783.

Carter, M. R. 1996. Analysis of soil organic matter inagroecosystems. In Structure and Organic MatterStorage in Agricultural Soils, ed. M. R. Carter, andB. A. Stewart, 3-11. New York: Lewis Publishers.

Coleman, D. C., D. A. Crossley Jr., and P. F. Hendrix.2004. Fundamentals of Soil Ecology. New York:Academic Press.

De Goede, R. M. G., and L. Brussaard. 2002. Soil zo-ology: an indispensable component of integratedecosystem studies. Eur. J. of Soil Biol. 38:1-6.

Edwards, C. A., and P. J. Bohlen. 1996. Biology and

Ecology of Earthworms. New York: Champman &Hall.

FAOSTAT data, 2004. http://faostat.fao.org/. Veri-fied on January 12, 2005.

Fragoso, C., et al. 1999. Earthworm communities oftropical agroecosystems: origin, structure and in-fluences of management practices. In EarthwormManagement in Tropical Agroecosystems, ed. P.Lavelle, L. Brussaard, and P. Hendrix, 27-55. NewYork: CABI Publishing.

Fraile, J. M. 1989. Poblaciones de lombrices de tierra(Oligochaeta: Annelida) en una pastura de Cynodonplectostachyus (pasto estrella) asociada con árbolesde Erythrina poeppigiana (poró), una pastura aso-ciada con árboles de Corida alliodora (laurel), unapastura sin árboles y vegetación a libre crecimientoen el CATIE, Turrialba, Costa Rica. Tesis MSci,Universidad de Costa Rica y Centro AgronómicoTropical de Investigación y Enseñanza.

Groffman, P. M., and P. J. Bohlen. 1999. Soil and sedi-ment biodiversity. Bioscience 49:139-148.

Lapied, E., and P. Lavelle. 2003. The peregrine earth-worm Pontoscolex corethrurus in the East coast ofCosta Rica. Pedobiologia 47:471-474.

Lavelle, P., et al. 2001. SOM management in the trop-ics: Why feeding the soil macrofauna? Nutr. Cycl.Agroecosys. 61:53-61.

Lyngbæk, A. E., R. G. Muschler, and F. L. Sinclair.2001. Productivity and profitability of multistrataorganic versus conventional coffee farms in CostaRica. Agrofor. Syst. 53:205-213.

Montenegro Gracia, E. J. 2005. Efecto del aporte denutrientes de la biomasa de tres tipos de árboles desombra en sistemas de manejo de café orgánico yconvencional. Tesis MSci, Centro AgronómicoTropical de Investigación y Enseñanza. Escuela dePosgrado. http://orton.catie.ac.cr/repdoc/A0295e/A0295e.html

Pashanasi, B., G. Melendez, L. Szott, and P. Lavelle.1992. Effect of inoculation with endogeic earth-worm Ponstoscolex corethrurus (Glossoscolecidae)on N availability, soil microbial biomass and thegrowth of three tropical fruit tree seedlings in a potexperiment. Soil Biol. Biochem. 24:1655-1659.

Powlson, D. S. 1994. The soil microbial biomass: Be-fore, beyond and back. In Beyond the Biomass. Com-positional and Functional Analysis of Soil MicrobialCommunities, ed. K. Ritz, J. Dighton, and K. E.Giller, 3-20. New York: John Wiley & Sons.

SAS Institute Inc. 2004. SAS OnlineDoc® 9.1.3. Cary,NC: SAS Institute Inc.

Scholes, R. J., R. Dalal, and S. Singer. 1994. Soil physicsand fertility: the effects of water, temperature andtexture. In The Biological Management of Tropical SoilFertility, ed. P.L. Woomer and M.J. Swift, 117-136.New York: John Wiley & Sons.

Schroth, G., J. Lehmann, M. R. L. Rodrigues, E. Barros,and J. L. V. Macedo. 2001. Plant-soil interactions inmultistrata agroforestry in the humid tropics.Agrofor. Syst. 53:85-102.

Senapati, B. K., et al. 1999. In-soil earthworm technolo-


gies for tropical agroecosystems. In EarthwormManagement in Tropical Agroecosystems, ed. P.Lavelle, L. Brussaard, P. Hendrix, 199-237. NewYork: CABI Publishing.

Soil Survey Staff. 1999. Soil Taxonomy. A Basic System ofSoil Classification for Making and Interpreting SoilSurveys. Washington D.C.: United States Govern-ment Printing Office.

Sollins, P., P. Homann, and B. A. Caldwell. 1996. Sta-bilization and destabilization of soil organic mat-ter: mechanisms and controls. Geoderma 74:65-105.

Somarriba, E., et al. 2004. Biodiversity conservation inneotropical coffee (Coffea arabica) plantations. InAgroforestry and Biodiversity Conservation in TropicalLandscapes, ed. G. Schroth, G. A. B. da Fonseca,

C. A. Harvey, C. Gascon, H. L. Vasconcelos, andA. N. Izac, 198-226. Washington: Island Press.

Woomer, P. L., A. Martin, A. Albrecht, D. V. S. Resck,and H. W. Scharpenseel. 1994. The importance andmanagement of soil organic matter in the tropics.In The Biological Management of Tropical Soil Fertility,ed. P. L. Woomer, and M. J. Swift, 47-80. NewYork: John Wiley & Sons.

Zuluaga Peláes, J. J. 2004. Dinámica de la materiaorgánica del suelo en sistemas agroforestales decafé con Erythrina poeppigiana (Walpers) O.F. Cooken Costa Rica. Tesis MSci, Centro AgronómicoTropical de Investigación y Enseñanza. Escuela dePosgrado. http://orton.catie.ac.cr/repdoc/A0281e/A0281e.html


Documents

Earthworm Populations, Microbial Biomass and Coffee ... · Earthworm Populations, Microbial Biomass and Coffee Production in Different Experimental Agroforestry Management Systems