microbial community structure at different depths in disturbed and undisturbed semiarid...

MicrobialEcology

Microbial Community Structure at Different Depths in Disturbedand Undisturbed Semiarid Mediterranean Forest Soils

M. Goberna1, H. Insam2, S. Klammer2, J.A. Pascual3 and J. Sanchez1

(1) Centro de Investigaciones sobre Desertificacion (CIDE), CSIC, Universitat de Valencia, Generalitat Valenciana, Camı de la Marjal, s/n, 46470 Albal,Valencia, Spain(2) Institut fur Mikrobiologie, Technikerstr.25, A-6020 Innsbruck, Austria(3) Centro de Edafologıa y Biologıa Aplicada del Segura (CEBAS), CSIC, P.O. Box 4195, 30100 Murcia, Spain

Received: 2 September 2004 / Accepted: 11 January 2005 / Online publication: 24 November 2005

Abstract

Metabolic abilities and micrfiobial community structurewere investigated through three semiarid Mediterraneansoils of SE Spain. The soils were (1) a Typic Calcixereptunder an adult pine plantation (PP), growing on aban-doned agricultural terraces; (2) a Typic Calcixeroll undera native pinewood (NP); and (3) a Typic Haploxereptcovered with a grass steppe (GS). PP and NP were similaras regards their genesis, but the former used to be tilled.NP and GS were undisturbed and supported natural andseminatural vegetation, respectively. Seven samples in10-cm depth increments were taken in triplicate alongeach soil profile. Community-level physiological profilesbased on sole-C-source use were determined to charac-terize the metabolic abilities. A 16S rDNA polymerasechain reaction-denaturing gradient gel electrophoresisanalysis was performed to investigate the microbial ge-netic structure. Plant cover and land-use history weremajor determinants of microbial community structure.Microbial communities residing in soils under a nativepinewood, the most diverse and stable plant cover, werethe most complex both metabolically and genetically.The microbial community structure distinctly changedwith depth, related to the quantity and quality of totalorganic carbon. Both undisturbed soils showed fallinggradients of metabolic and genetic complexity, whichwere invariably of a greater magnitude in the maturewoodland than in the grass steppe. In the planted pine-wood, however, the substrate-use diversity increased withdepth, apparently a response to the depleted metabolicabilities within its upper layer (0–30 cm). Tilling and plantcover removal might be responsible for such a perturba-tion. In the same profile, molecular fingerprint patterns

of the topsoil layer (0–10 cm) indicated a disturbed ge-netic structure that might underlie the loss of metabolicabilities. However, the genetic structure of the deeperlayers of the planted and native pinewoods was not dis-similar, revealing that equivalent genetic resources per-form different environmental functions under changingsoil scenarios.

Introduction

Microorganisms residing in the soil’s shallow subsurfacehave received little scientific attention until the pastdecade. Vertical variations in the soil microbial activityand biomass have been documented, but little researchhas been done concerning the soil layers below 30–40 cm[6, 11]. Even less is known on the community structureof these soil layers [12, 17]. However, the microorgan-isms living in the soil subsurface are known to be directlyinvolved in soil formation, carbon biogeochemistry [25],and the regulation of water flows and groundwaterchemistry [43]. Consequently, the subsurface microbiotaensures the proper performance of a soil’s ecologicalfunctions, by playing an active role in the decomposition,transformation, and filtering of chemicals. Therefore, themetabolic and genetic structure of the microbial com-munities need to be better understood, because suchknowledge would help in the assessment of the fate ofnatural and xenobiotic compounds in their downwardmovement. This subject is of particular importance insemiarid Mediterranean areas, where groundwater con-stitutes a major source of water supply to satisfy humanneeds. In these lands, the low soil organic mattercontents are counterbalanced with an excessive use offertilizers, which is threatening the sustainability ofgroundwater quality [18].Correspondence to: M. Goberna; E-mail: [email protected]

DOI: 10.1007/s00248-005-0177-0 & Volume 50, 315–326 (2005) & * Springer Science+Business Media, Inc. 2005 315

Among the factors that may influence the differentialtrends among microbial communities living throughoutthe soil profile, most available surveys point to thevariations in soil organic carbon (SOC) [2, 11]. Plantcover and land-use history are two major factorsconditioning the amount, composition, and distributionof SOC inputs in semiarid Mediterranean areas. Theglobal objective of this survey was to shed some lightonto the influence that the mentioned factors have onsurface and subsurface microbial communities. In par-ticular, we aimed (1) to compare the metabolic andgenetic complexity of the microbial communities of soilsdiffering in their plant cover and land-use history, and(2) to analyze the depth patterns of the microbialcommunity structure in the same soils.

To test the hypothesis that due to their role in thecarbon dynamics the microbial communities within thesoil profile would mainly differ depending on the above-ground vegetation and the land-use history, three soilprofiles, all representative of the semiarid Mediterraneanarea, were studied at different depths. The profiles understudy comprised a gradient of disturbance in their plantcover structure and composition, including a nativewoodland, a seminatural grass steppe, and an artificialpine plantation. Plant species identity determines thepattern [40] and rate of C decomposition [1], as well asthe dissolved organic C leaching pattern [27]. A greaterdiversity of plant species, therefore, entails a wider rangeof debris and metabolites of different decomposability,and consequently, a higher number of microniches avail-able for soil microorganims. This might promote micro-bial species coexistence and stimulate their diversity[10, 44]. Furthermore, cooccurring plant species differin their rooting habits and root penetration pressured bycompetition for soil resources [36]. The stratification ofthe root network might introduce different microhabitatsthroughout the profile that would select for those micro-bial communities physiologically adapted to the metabo-lism of the particular C substrates [11]. Consequently,the higher aboveground diversity of the more matureplant cover was expected to be reflected by (1) a greatercomplexity of its soil microbial communities as regardstheir metabolic abilities and genetic structure and (2) asharper stratification with depth of these variables.

The metabolic abilities of the microbial communitywere investigated by community-level physiologicalprofiling (CLPP) using Biolog Ecoplates\ [13, 23]. Theinterpretation problems of this technique [8, 20, 38]limit its use as a stand-alone approach. It can be used,however, as a screening method that in combination withmolecular data permits linking structure and function[15]. The genetic structure of the soil microbial com-munities was analyzed by separating polymerase chainreaction (PCR) amplified genes coding for 16S rRNA,using denaturing gradient gel electrophoresis (DGGE).

PCR/DGGE was shown to be a useful technique forstudying the changes in the microbial communitiesthrough the soil matrix [21, 31, 34] and assessing soildegradation [35].

Materials and Methods

Soils, Plant Cover, and Land Use. The study wascarried out in the Crevillent mountain range, which islocated in the semiarid Mediterranean region of Alacant(southeast Spain). Annual rainfall is below 300 mm andannual temperature is about 20-C. The main characteristicsof the soils studied are summarized in Table 1 and below:

Profile 1: a Typic Calcixerept [42] under an artificialAleppo pine (Pinus halepensis Mill.) plantation witha perennial grass understory dominated by Brachy-podium retusum (Pers.) Beauv., growing on aban-doned agricultural terraces. The latter species is aclonal perennial grass whose root system constitutesa dense net in the uppermost 5–20 cm [49].

Profile 2: a Typic Calcixeroll [42] under a nativepinewood with a woody and diverse understory ofMediterranean maquis dominated by kermes oak(Quercus coccifera L.), but also with a major cover ofPistacia lentiscus L., Erica multiflora L., Juniperusoxycedrus L., Rhamnus alaternus L., and other speciestypical of Mediterranean shrublands. Scandentshrubs, grasses, and mosses constitute other plantlayers. Such a complex plant cover develops anintricate root system through the profile including awide range of root sizes.

Profile 3: a Typic Haploxerept [42] covered with aseminatural grass steppe dominated by the tussock-forming species Stipa tenacissima L. Alpha grass hasbeen harvested for fibers during millennia, butagricultural practices were abandoned in the 1960s[5]. Alpha grass develops a rather superficial rootsystem, which may go down to 50–100 cm depth,producing a large amount of fine roots [19].

Hereafter, these soils will be referred to as ‘‘pineplantation’’ (PP), ‘‘native pinewood’’ (NP), and ‘‘grasssteppe’’ (GS).

Sampling Procedure. Samples from 0 to 70 cmdepth were taken in triplicate in 10-cm increments fromthe three soil profiles in April 2003, when these wereevenly moist (data not shown). This sampling includedmost of the pedogenetic horizons, which were not mixedwith each other. Visible plant residues and roots wereremoved and fresh soil sieved G2 mm and divided intothree subsamples, one of which was air-dried forchemical and physical analyses. Another subsample waskept at 4-C during CLPP analysis and the third onestored at _20-C for DNA extraction and subsequent

316 M. GOBERNA ET AL.: MICROBIAL COMMUNITY STRUCTURE IN MEDITERRANEAN FOREST SOILS

Tab

le1.

Pro

file

des

crip

tio

nan

dm

ain

soil

char

acte

rist

ics

Soil

profi

leH

oriz

onT

hic

knes

s(c

m)

Col

oura

Stru

ctu

reb

Con

sist

ency

cP

last

icit

ydR

oots

epH

EC

(dS

m_

1)

Tex

ture

fA

S(%

)� B

(gcm

_3)

PP

O+

2A

p0–

2110

YR

4/3

2fgr

-sb

km

vfr,

wss

wsp

3f,

2m8.

340.

137

lc

22.6

0.96

2Ah

21–

4510

YR

3/3

2fgr

-sb

km

vfr,

wss

wsp

2f,

2m8.

240.

161

lc

35.8

2Bw

k45

–72

10Y

R3/

42v

f-f

sbk

mvf

r,w

ssw

sp1f

,1c

8.35

0.13

0l

c32

.32B

Ck

72–

9710

YR

5/4

2vf-

fsb

km

vfr,

wss

wsp

8.43

0.12

2l

c19

.12C

k9

9710

YR

5/4

8.48

0.11

0l

c10

.7N

PO

+8

Ah

0–23

10Y

R3/

33f

-mgr

mvf

r,w

ssw

sp3v

f-f-

m,

2m-c

8.06

0.33

7l

35.7

0.76

AB

23–

5110

YR

4/3

3f-m

grm

vfr,

wss

wsp

2f-

m-c

8.32

0.14

4l

s25

.6B

/A51

–95

10Y

R5/

43v

f-fg

rm

vfr,

wss

wsp

1m8.

450.

128

ls

19.4

Cm

k9

95G

SA

h1

0–22

7.5Y

R4/

31v

f-f

gr-s

bk

mvf

r,w

ssw

p3f

8.53

0.11

5l

25.0

1.01

Ah

222

–45

7.5Y

R4/

31v

f-f

gr-s

bk

mvf

r,w

ssw

sp2f

,1m

8.50

0.15

8l

24.0

Bw

45–

657.

5YR

4/6

1vf-

fsb

km

vfr,

wss

wp

1m8.

180.

388

l20

.4C

k65

–90

1vf-

fsb

km

fr,

wss

wsp

8.38

0.27

8l

cR

990

aM

ois

ten

and

cru

shed

.b1:

wea

k;2:

mo

der

ate;

3:st

ron

g;f:

fin

e;m

:m

ediu

m;

c:co

arse

;sb

k:su

ban

gula

rb

lock

y;gr

:gr

anu

lar.

cm

:m

ois

t;w

:w

et;

vf:

very

fria

ble

;f:

fria

ble

;ss

;sl

igh

tly

stic

ky.

dw

:w

et;

sp:

slig

htl

yp

last

ic;

p:

pla

stic

.e1:

few

;2:

ple

nti

ful;

3:ab

un

dan

t;f:

fin

e;m

:m

ediu

m;

c:co

arse

.fl:

loam

;c:

clay

;s:

san

d;

si:

silt

.p

H(1

:2.5

H2O

);E

C:

Ele

ctri

cal

con

du

ctiv

ity;

AS:

Agg

rega

test

abil

ity;� B

:B

ulk

den

sity

.

M. GOBERNA ET AL.: MICROBIAL COMMUNITY STRUCTURE IN MEDITERRANEAN FOREST SOILS 317

PCR-DGGE analysis. Furthermore, three unaltered soilsamples were taken from the upper horizons using acylinder (5.0 cm high and 4.77 cm diameter), 48 h after aheavy rainfall, for bulk density determination. Allanalyses were made in triplicate.

Soil Chemical and Physical Analysis. Soil pH andelectrical conductivity (EC) were measured in aqueousextracts (1:2.5 and 1:5 solid–liquid, respectively), exceptthose samples exceeding 0.2 dS m

_1 whose EC wasdetermined in the soil saturation paste. Soil carbonateswere titrated with HCl (1:1) and the volume of the CO2

produced was measured using a Bernard calcimeter [37].The total organic carbon (TOC) was determined byoxidation with 1 N potassium dichromate in acidic me-dium and back titration with 0.5 N ammonium ferroussulfate, as described by Walkley and Black (1934) [37].The granulometric analysis was made using a Robinsonpipette and the textural class was established using theUSDA classification [29]. Stationary aggregate stabilitywas determined as the difference between the sonicationresistant aggregates (40 Hz for 180 s) and the coarse sands[22]. Bulk density was determined as the oven-dried(105-C) weight of each unaltered sample divided by thevolume of the cylinder. Soil physical and chemical para-meters are provided in Table 1 and Fig. 1.

Community Level Physiological Profiles. CLPPswere determined by the use of Biolog Ecoplates\ [13,23]. Soil suspensions in Milli-Q water (1:10 w/v) werephysically dispersed using a Waring blender (13,500 revmin

_1, 3 cycles � 1 min, 4-C). Bacterial cells wereseparated from the soil particles according to theirbuoyant density in a high-speed centrifugation step,using a commercial solution of Nycodenz (NycoPrepiUniversal, Axis Shield, Austria) [39]. After carefully

pipetting 7 mL of NycoPrep\ (60% w/v) below the soilsuspension, the density barrier was established bycentrifugation at 10,000�g (30 min, 4-C) using aswing-out rotor [3]. A volume of 10 mL of thesupernatant, which included the cells, was transferred to250-mL centrifuge bottles containing 100 mL of Milli-Qwater. These were centrifuged at 16,000�g (60 min, 4-C)to obtain Nycodenz-free bacterial pellets [30]. The cellswere resuspended in 1/4-strength Ringer solution(Merck, Darmstadt, Germany) and diluted to stan-dardize the inoculum density. Dilution factors were de-termined according to microbial biomass carbon data[24]. The Ecoplates were inoculated with 130-mL sus-pension and incubated at 28-C in the dark. Optical den-sity (OD592) was measured every 12 h for 7 days using anautomated plate reader (Tecan, Grodig, Austria).

16S rDNA-PCR-DGGE Analysis

DNA Extraction and PCR Amplification. DNA wasextracted using the Fast DNA\ Spin Kit for soil (BIO101, Carlsbad, CA, USA). Extracts were diluted 1:100 andused as templates for PCR amplification of 16S rDNAfragments. For amplification, 1 mL of diluted templateDNA was added to 24 mL of PCR mixture, which includ-ed the following components: 2.5 mL 10� buffer Y, 5 mL5� reaction enhancer, 0.625 U Taq Polymerase, 200 mMeach deoxynucleotide triphosphate (all reagents fromPeqlab, Eramgen, Germany), 0.2 mM primer F984 (50-GC-clamp-AAC GCG AAG AAC CTT AC-30), 0.2 mMprimer R1378 (50-CGG TGT GTA CAA GGC CCG GGAACG-30) [21]. PCR amplification was carried out in aPCR Express (ThermoHybaid) thermal cycler, using aninitial denaturation step (94-C for 5 min), followed by 30cycles consisting of 30 s denaturation at 94-C, 30 s ofprimer annealing at 56-C, and 1 min of elongation at

Figure 1. Carbonate (A) and totalorganic carbon (B) content vssoil depth. Bars indicate standarderror, n = 3. For each profile,values with the same letters arenot significantly different (P G 0.05).Symbol shape indicates soilprofile: diamonds (PP), squares(NP), triangles (GS).


72-C, and a final extension step of 4 min at 72-C. Afteramplification, the PCR products were verified on a 1%agarose gel and stained with ethidium bromide.

DGGE of Amplified DNA. DGGE was performedusing a Bio-Rad DCodei Universal MutationDetection System (Hercules, CA, USA). A mixture of10 mL of PCR product and 4 mL loading dye (0.05%bromophenol blue, 0.05% xylene cyanol, 70% glycerol,and sterilized water) was loaded onto 7% acrylamide–0.1% bisacrylamide gels in 1� TAE (40 mM Tris base,20 mM glacial acetic acid, 1 mM EDTA, pH 8.0). Adenaturing gradient of 50–70% (7 M urea plus 40% w/vdeionized formamide, for 100% denaturing solution)was used. The electrophoresis was run at 60-C, first for20 min at 40 V, and then for 16 h at 70 V. The gels weresoaked for 20 min in a fixation buffer (9% ethanol–0.5%glacial acetic acid–90.5% Milli-Q water) prior to silverstaining (0.1% silver nitrate) for 40 min, and thendeveloped in a freshly made developing solution [0.1 gsodium borohydride L

_1–1.5% (w/v) NaOH–0.15% (v/v)formaldehyde] for 45 min. Finally, they were fixed for 5min and soaked in a preserving solution (24% ethanol–9% glycerol–67% Milli-Q water) for 10 min. Gels werephotographed with a digital camera.

Data Processing and Statistical Analysis. Foranalysis of the CLPPs, raw OD data were corrected byblanking each response well against its own first reading[24]. Shannon’s diversity index (H0) and evenness (E) [28]were calculated as a means of evaluating diversity in Csource consumption [47], using the following equations:

H0 ¼ �X

pi ln pið Þ and E ¼ H0

H0max

¼ H0

log S

where pi is the ratio of the corrected absorbance valueof each well to the sum of absorbance value of all thewells, and S (substrate richness) is the number ofdifferent substrates used by the community (countingall positive OD readings). Shannon’s diversity index(H0) and the substrate evenness (E) were calculated con-sidering: (1) the 31 single C sources in the Ecoplates,and (2) the six categories of substrates according to theirchemical nature (carboxylic acids, polymers, carbohy-drates, phenols, amino acids, and amines) [23, 47]. Two-way analysis of variance (ANOVA) was used for statisticaltesting of the effects of profiles and sampling depths oncarbonates, TOC, diversity (H0), and substrate evenness(E). Komolgorov–Smirnoff’s test was used to test thenormality, and Levene’s test to test the homogeneity ofvariances. Logarithmic transformations were applied tothe data to meet the assumptions of ANOVA [48]. Incases of significant F statistics, Tukey’s post-hoc test wasselected to separate the means.

CLPP data analysis was further elaborated byestimation of kinetic parameters by fitting the curve ofOD592 vs time to a density-dependent logistic growthequation [26]:

Y ¼ OD592 ¼K

1þ e�r t�sð Þ

where K denotes the asymptote or maximum degree ofcolor development (OD592), r the exponential rate ofOD592 change (h

_1), t the time following inoculationof the microplates (h), and s the time at the midpointof the exponential portion of the curve (i.e., when y =K / 2) (h). Kinetic analysis was carried out for the above-mentioned categories of substrates, and not for the Csources individually, because data treatment is less time-consuming and allows easier interpretation. The rate(r) and maximum degree (K) of color developmentwere used to classify the microbial communities usingprincipal component analysis (PCA). Multivariate anal-ysis of variance (MANOVA) applied on PCA factors wasused for statistical testing with SPSS 11.5.

Analysis and comparison of the different DGGEpatterns was carried out with the GelCompar 3.1program (Applied Maths, Kortrijk, Belgium). After con-version of the pictures into the program, the bands werenormalized and a reference position was defined to alignthe patterns for comparison after associating them withthe standard. Analysis was performed by comparing thesimilarity between their densitometric curves using thePearson correlation coefficient. Ward’s clustering meth-od was applied to calculate dendrograms. Distinct bandswere identified applying a position tolerance of 2%, min-imum profiling of 2%, 0.2% gray zone, and a shouldersensitivity of 5 to identify shoulders with no local maxi-mum, as well as doublets of bands with one maximum.

Results

Community Level Physiological Profiles

Shannon’s Diversity Index (H0) and Substrate Evenness(E). Shannon’s diversity index (H0), calculated con-sidering the single C substrates in the Biolog plates,ranged from 1.14 to 1.32 (Fig. 2A). This variable, whichreflects both the richness and evenness of substrateconsumption of the C sources [28, 47], differed signif-icantly in the three profiles (F2 = 46.5, P G 0.001).Functional diversity (H0) also varied with depth (F6 =5.3, P G 0.001), with all the profiles showing twohomogenous layers: (1) from the surface to 30 cm deepand (2) from 30 to 70 cm deep. PP showed lowermicrobial diversity in the upper layer, whereas NP andGS had the opposite pattern (Fig. 2A). These differencesaccount for the significant interaction between the twofactors—profile and depth (F12 = 8.0, P G 0.001).


Substrate evenness (E) indicates the homogeneity ofsubstrate consumption by microbial communities [28].E ranged from 0.85 to 0.92 (Fig. 2B) and was significantlydifferent among the soils studied (F2 = 12.0, P G 0.001).Evenness depth pattern was dependent on the profile(F12 = 2.1, P = 0.034). PP exhibited similar substrate even-ness values throughout its depth (Fig. 2B). However,E decreased toward the deeper layers of NP and GS,although only NP showed significant differences betweenthe upper (0–30 cm) and lower layers (30–70 cm).

Functional diversity, calculated for particular guilds(i.e., the six categories of substrates), showed differencesamong the soils studied due to their different abilities toconsume carboxylic acids (F2 = 25.9, P G 0.001),polymers (F2 = 15.2, P G 0.001), carbohydrates (F2 =39.5, P G 0.001), and amines (F2 = 14.9, P G 0.001) (datanot shown). Tukey’s test indicated that GS was the mostdifferent, except for the diversity of the amino acid con-sumption, whereas PP and NP differed slightly, particu-larly in their specific capability of degrading polymers.Each profile presented its particular guild depth pattern,which explained the significant interaction found be-tween the factors ‘‘profile’’ and ‘‘depth’’ for all the vari-

ables except for phenols and amino acids. One-wayANOVA indicated that the diversity of the carboxylicacid (F6 = 4.2, P = 0.012) and amine (F6 = 9.2, P G 0.001)consumption varied with depth in PP. NP had muchmore of its variation related to guild diversity, the use ofcarboxylic acids (F6 = 3.3, P = 0.029), polymers (F6 =5.6, P = 0.004), carbohydrates (F6 = 25.1, P G 0.001), andamines (F6 = 9.6, P G 0.001) differing significantlythroughout the profile. Lastly, H0 calculated for carboxylicacids (F6 = 6.1, P G 0.001) was the only variable showingsignificant differences with increasing depth in GS.

Rate (r) and Degree (K) of Consumption of the Groups ofSubstrates. The OD592 data vs time obtained forparticular guilds were fitted to the density-dependentsigmoidal curve proposed by Lindstrom et al. [26]. Theseauthors assessed that two of the kinetic parameters (therate, r, and maximum degree, K, of color development)were able to characterize the microbial communitiesgrowing in the Biolog plates, regardless of inoculumdensity. This statement was partly contradicted byGarland et al. [14], who suggested that correct classi-fication of the samples can only be achieved if Biologplates are inoculated with equivalent microbial densities.This requirement was complied with in this studybecause the bacterial extracts were diluted according tothe microbial biomass data.

The maximum degree (K) and rate (r) of consump-tion were estimated for each group of substrates. Data arepresented as the means of the previously defined upper(0–30 cm) and lower soil layers (30–70 cm) to simplify thedepth analyses (Table 2). PP and NP differed significantlyfrom GS in their degrees and rates of consumption of allthe guilds considered. The degree (K) of consumption ofcarbohydrates and amino acids, and the rates (r) of con-sumption of carbohydrates and polymers were the onlydifferences between PP and NP. As regards the variationsbetween the two layers within each profile, no differenceswere detected through PP, whereas NP showed different Kof polymers and carbohydrates and r of carbohydrates andamines. GS had different K for carbohydrates and r forpolymers between the upper and lower layers (Table 2).

Principal component analysis on the rate (r) ofconsumption of particular guilds within the upper soillayer (0–30 cm) identified two components thataccounted for 73.3% of the total variance (Fig. 3A). Therewas a significant separation of all profiles along the PC1axis (F2 = 57.6, P G 0.001). PCA on the r determined forthe lower soil layer (Fig. 3B) identified two componentsthat accounted for 80.5% of the total variance. Also, inthis case, there was a significant discrimination betweenthe profiles (F2 = 95.2, P G 0.001), but the post-hoc testclassified together PP and NP when PC-1 was analyzed. Asregards the maximum degree (K) of consumption, thesame patterns were found (Fig. 4).

Figure 2. Diversity of the microbial communities in C sourceutilization. (A) Shannon’s diversity index (H0) and (B) substrateevenness (E) vs soil profile. Values with the same letter are notsignificantly different (P G 0.05). Closed symbols: samples 0–30 cm,open symbols: samples 30–70 cm.


Carbohydrates (CH) and amino acids (AA) were thetwo major categories of substrates responsible for thedifferences between samples. CH and AA had strongpositive loading scores (higher than 0.90 in all cases)with PC1 when the sources of variation in the maximumextent of color development (K) in the upper and lowersoil layers were analyzed. As in [41], the CH/AA ratiowas calculated, although in this case K data were usedinstead of the average absorption value (Fig. 5). CH/AAsignificantly differed between the profiles studied (F2 =96.9, P G 0.001), which did not show similar variationsbetween the upper (0–30 cm) and lower (30–70 cm)

layers (F12 = 28.7, P G 0.001). PP showed no variations inthe CH/AA ratio between the upper and lower layers.Contrarily, CH/AA declined significantly through NPand GS.

16S rDNA-PCR-DGGE Analysis. The geneticstructure of soil microbial communities was investigatedby separating PCR-amplified genes coding for 16S rRNA,using DGGE. All the profiles presented complex DNAbanding patterns, with some strong bands, some of lowerintensity and a number of faint bands, which occasion-ally resulted in blurred images. Sketches depicting the

Table 2. Maximum degree (K) and rate (r) of consumption of the groups of substrates in the upper (0–30 cm) and lower (30–70 cm)layers of the profiles studied

Group of substrates Soil profile

K (OD592) r (h_1)

Upper layer Lower layer Upper layer Lower layer

CA PP 1.09 (0.14) a 1.15 (0.23) a 0.53 (0.03) a 0.57 (0.14) aNP 1.31 (0.06) a 1.24 (0.12) a 0.56 (0.06) a 0.60 (0.09) aGS 0.67 (0.08) b 0.91 (0.12) b 0.43 (0.25) b 0.27 (0.05) b

PL PP 1.56 (0.11) a 1.49 (0.49) a 0.67 (0.06) b 0.64 (0.20) bNP 1.78 (0.02) a 1.27 (0.10) b 0.85 (0.04) a 0.79 (0.05) aGS 0.60 (0.10) c 0.62 (0.23) c 0.47 (0.01) c 0.34 (0.08) d

CH PP 0.93 (0.11) b 1.05 (0.26) b 0.53 (0.04) c 0.52 (0.08) cNP 1.75 (0.08) a 1.01 (0.05) c 0.66 (0.02) a 0.57 (0.03) bGS 0.46 (0.07) d 0.37 (0.05) e 0.44 (0.05) d 0.38 (0.06) d

PH PP 0.54 (0.18) a 0.77 (0.39) a 0.50 (0.07) a 0.81 (0.65) aNP 0.99 (0.06) a 0.63 (0.13) a 1.61 (0.06) a 0.55 (0.08) aGS 0.30 (0.06) b 0.26 (0.03) b 0.41 (0.05) b 0.36 (0.10) b

AA PP 1.49 (0.06) b 1.57 (0.35) b 0.60 (0.02) a 0.49 (0.03) aNP 1.92 (0.12) a 1.65 (0.08) a 0.48 (0.04) a 0.59 (0.04) aGS 1.01 (0.15) c 1.00 (0.09) c 0.40 (0.10) b 0.31 (0.02) b

AM PP 8.71 (3.09) a 6.05 (2.23) a 0.35 (0.09) b 0.24 (0.10) bNP 4.65 (1.52) a 6.98 (2.96) a 0.18 (0.01) b 0.42 (0.04) cGS 0.24 (0.10) b 0.27 (0.04) b 0.41 (0.08) a 0.44 (0.06) a

In parentheses, standard deviation for n = 3. For each kinetic parameter and substrate group, values with the same letter are not significantly different(P G 0.05). Groups of substrates, CA: carboxylic acids; PL: polymers; CH: carbohydrates; PH: phenols; AA: amino acids; AM: amines.

Figure 3. First and second principalcomponents derived from the rateof color development (r) of thegroups of substrates. (A) Samples0–30 cm. (B) Samples 30–70 cm.The explained variation is given inbrackets. Symbol shape indicatessoil profile: diamonds (PP), squares(NP), triangles (GS).


strongest bands, which were used for data analysis, aregiven instead of the photographs of the gels (Fig. 6).Band numbers in the DGGEs were maximal in NP,which showed 20 bands in its top layer compared to amaximum of 12 and 11 bands found in PP and GS,respectively (Fig. 6). An absolute decrease in bandnumbers was observed with increasing depth for NP(20 to 15 bands from 0 to 70 cm) and GS (11 to 5 bandsfrom 0 to 70 cm). In PP, the least number of bands (6)was found in the top layer (0–10 cm), whereas ahomogeneous distribution was detected with depth,ranging from 9 to 12 bands.

Cluster analysis was performed to infer similaritiesamong the different banding patterns generated and theresulting dendrogram is presented in Fig. 6. The densito-metric curves were compared using the Pearson correlationcoefficient and Ward clustering. This procedure clusteredtogether nearly all samples from GS (except the 0–10 and20–30 cm layers) and the deepest layer of PP, which showed

only 16.25% similarity with the other patterns. The toplayer (0–10 cm) from PP was also distinguishable withrather low similarities to the other clusters (46.5%). Nofurther resolution in separating the different profiles couldbe obtained, indicating similar community genetic struc-ture in PP and NP (data not shown).

Discussion

The microbial community structure of three semiaridMediterranean forest soils was analyzed at differentdepths, the main differences between them being theirplant cover and land-use history. Both factors were majordeterminants of the microbial community structure.Aboveground diversity consistently promoted below-ground diversity, as reflected by the metabolic andgenetic fingerprint of the microbial communities studied.These followed different depth gradients, which appearedto be associated with the quantity and quality of TOCdistributed throughout each soil profile.

Metabolic Community Profile. The specific abilityof soil microbial communities to consume a range ofcarbon sources available in a commercial kit (BiologEcoplates) was tested as a means to assess their metabolicprofile. Greater diversity (H0) and evenness (E) in Csource utilization in Biolog Ecoplates was observed insoils under a pinewood with maquis (NP) than in soilsunder a pine plantation (PP) or a grass steppe (GS)(Fig. 2). The former soils also showed the highest diver-sity (data not shown), degree (K), and rate (r) of con-sumption of all the particular guilds analyzed (Table 2).This suggested that the greatest availability of differentC sources occurred in soils under a native pinewood [15],which sustained the richest plant community and heldthe highest TOC contents (Fig. 1).

Figure 5. CH/AA ratio vs soil profile. Bars indicate standard error,n = 3. Values with the same letter are not significantly different(P G 0.05). Closed symbols: samples 0–30 cm, open symbols: samples30–70 cm.

Figure 4. First and second principalcomponents derived from the max-imum extent of color development(K) of the groups of substrate. (A)Samples 0–30 cm. (B) Samples30–70 cm. The explained variationis given in brackets. Symbolshape indicates soil profile: dia-monds (PP), squares (NP), triangles(GS).


Differential patterns in the utilization of C substratesthroughout the studied profiles indicated distinct shiftsin the community metabolic potential. NP and GSunderwent falling gradients of metabolic complexity.Functional diversity (H0) and substrate evenness (E)decreased with depth through both profiles (Fig. 2).Furthermore, some microbial guilds experienced varia-tions with depth in their diversity (data not shown) andactivity (Table 2). Also, the CH/AA ratio declined fromthe upper to the lower soil layer (Fig. 5), indicating ashift in the relative abundance of readily oxidablesubstrates [41].

Although NP and GS followed similar qualitativedepth trends, changes were of a greater magnitude inthe former profile. It is postulated that the differentroot systems stretching throughout these soils accountfor the different degrees of stratification. The multi-layered aboveground structure of the native pinewoodwas reflected in a stratified belowground structure ofthe rhizosphere (Table 1). The great variations in theTOC content and the differential C sources released tothe soil matrix throughout the mentioned profile mightbe responsible for the stratification of its microbial com-munities [11]. The depth trends of the metabolic com-munity structure in soils under a grass steppe (GS) wereless complex, probably because the shifts were exclusivelyassociated to changes in alpha grass root density. In this

profile, carbonates may have further contributed to themetabolic similarity of its surface and subsurface micro-bial communities. It is widely accepted that the presenceof CaCO3 reduces the solubility of organic substances,preventing them from microbial attack through the pre-cipitation of carbonate coatings and formation of Ca-organic linkages [4]. The predominance of CaCO3 in thesoil matrix of GS (Fig. 1) may have greatly homogenizedthe resources available for the microbiota.

In soils under a pine plantation (PP), the naturaldepth pattern was inverted, as indicated by the rise infunctional microbial diversity (H0) in the deeper layers(30–70 cm) (Fig. 2A). Substrate evenness (E) was alsoaltered in this profile, through which E remainedinvariable (Fig. 2B). Moreover, these soils presented aneven distribution of all the groups of consumers, asreflected the degree (K) and rate (r) of consumption ofthe guilds analyzed (Table 2) and had a constant CH/AAratio with depth (Fig. 5). The depth trends detected in PPwere probably provoked by the shift in the metabolicfingerprint of the microbial communities inhabiting theupper layer (0–30 cm), as those of the deeper layers (30–70 cm) bore much resemblance with their counterpartsin soils under a natural pinewood with maquis (NP).This suggests that agricultural practices and depletion ofthe native plant cover caused a narrower range ofavailable C resources in soils under a pine plantation,

Figure 6. Dendrogram and DGGE banding patterns of the three soil profiles. Samples are identified with the code of the profile and thesoil layer.


this perturbation being confined to the plow layer. Ourfindings agree with those of Yan et al. [45] and Yao et al.[46], who used CLPPs to analyze the microbial commu-nities in soils with changing land uses.

PCA on r and K successfully discriminated the profilesstudied (Figs. 3 and 4). PCA on the rate (r) and maximumdegree (K) of consumption in all groups of substratessignificantly separated all the profiles when the upper soillayer (0–30 cm) was analyzed (Figs. 3A and 4A). However,soils under planted and native pinewoods (PP and NP)were grouped together when the data from the lower soillayer (30–70 cm) were tested (Figs. 3B and 4B). Thesefindings strengthen the above hypothesis that the physi-ological diversity of soils under a planted pinewood wasonly damaged within the plow layer.

Genetic Community Structure. The geneticstructure of the microbial communities inferred by theDGGE of 16S rDNA (Fig. 6) revealed complex bandingpatterns reflecting the high microbial diversity expectedfor a forest soil [2]. Despite the high-resolution powerof DGGE, it was not possible to estimate the totalnumber of 16S rDNA molecules present. It should bementioned that these data must be analyzed with cau-tion because species richness cannot be directly inferredfrom the number of bands in the DGGE. It is known thatsingle bacterial types might produce more than oneband [33] and bands at similar positions might consistof different sequences with the same melting temper-ature. Another uncertainty associated to the analysis ofthe banding patterns deals with the selection of thestrongest bands in the gels, which correspond to the dom-inant phylotypes.

Soils under a pinewood with maquis (NP) had themost complex banding patterns, both considering thestrongest (Fig. 6) and other fainter bands (personalobservation). Genetic data, therefore, supported thefindings concerning metabolic abilities (CLPPs). Ourresults agree with those of Øvreas and Torsvik [34], whoreported that the microbial communities residing in soilsunder a more diverse and stable plant cover exhibitedgreater physiological and genetic diversity.

Molecular fingerprint patterns of amplified 16SrDNA fragments were classified into two main clusters(Fig. 6), which clearly reflected the effect of grasslands vspinewoods in the microbial genetic structure, as follows.

Cluster 1 (GS). A rather homogeneous distributionwas observed within the profile covered with grass steppe(GS). In these soils, however, some variations were foundwith depth—samples from 0–10 and 20–30 cm beingclustered independently from the others in the same pro-file. This shift of soil microbiota as a function of depthhas been mostly described in the literature [2, 6, 9, 11].

Cluster 2 (NP + PP). The genetic fingerprint of themicrobial communities living throughout the soilunderlying the native pinewood (NP) was equivalent.This unexpected result interestingly coincides with thefindings of Bundt et al. [7], who did not find a depthgradient when analyzing the bacterial communities inSwiss forest soils, using PCR/restriction length polymor-phisms (RFLPs) based on small subunit ribosomal RNAgenes. Most samples taken under planted pinewoods(PP) were grouped together with those from NP.However, its upper soil layer (0–10 cm), in which theminimal band numbers was detected, was dissimilar toall the other samples, as reflected by the cluster analysis.These results suggest that the community structure of thetopsoil layer under a pine plantation was different to thatof the native pinewood, probably because of its formeragricultural use. DGGE banding patterns of the layersbelow 10 cm in PP were clustered together with those ofNP (except for the 60–70 cm layer). These findingssuggested an undisturbed community, which was partic-ularly remarkable in the case of the soil layers between 10and 30 cm considering their depleted metabolic abilities.Either these layers (10–30 cm) did not undergo a geneticperturbation despite plowing or, as what seems morereasonable, experienced a bottom–up genetic recoveryafter land abandonment. This process could respond tothe colonization exerted by the microbial communitiesfrom the deeper (30–60 cm) undisturbed layers. What-ever the case, idiosyncrasy appears as the prevailingmechanism relating biodiversity and function [32],because equivalent genetic resources performed differentenvironmental functions under disturbed and undis-turbed soil scenarios.

Conclusions

Plant cover and land-use history were major determi-nants of the TOC pool size and depth gradient throughsemiarid Mediterranean profiles. The amount and distri-bution of the organic compounds throughout the soilsstudied reflected the structure of their microbial com-munities. Soils under a native pinewood sheltered themost complex microbial communities, defined both byphenotypic (CLPPs) and genotypic (PCR-DGGE 16SrDNA) characteristics. From the study of the depthpatterns, two distinct trends emerged. First, soils under anatural pinewood and a seminatural grass steppe pre-sented falling gradients of metabolic complexity, whichwere more pronounced in the former profiles. Also, thenumber of bands in the DGGE pointed to geneticallysimpler communities inhabiting the subsurface of bothprofiles, whereas analyses of banding pattern similaritydid not discriminate any layer of soils under a pinewoodwith maquis. Second, the soil under a pine plantation


had an altered metabolic and genetic fingerprint, partic-ularly within the upper soil layer (0–10 cm). Therefore,the system cannot be said to be resistant to the persistentstresses applied (e.g., tilling and deforestation), becausethese altered the community structure. A certain recoveryof its genetic resources was detected, however, within thesubsurface 10–30 cm layer. It can be, therefore, assertedthat the soil system is apparently resilient [16]. Whether ornot, and, when the disturbed ecosystem will have its func-tions completely restored remains unsolved. The answer tothis question needs further insight into the relationshipbetween microbial diversity and ecosystem functioning.Long-term observations and hypothesis-driven experi-ments are needed to gain knowledge on this issue.

Acknowledgments

Marta Goberna was supported by funding from General-itat Valenciana, Valencia, Spain. Helpful comments bythree anonymous referees are gratefully acknowledged.

References

1. Aber, JD, Melillo, JM, McClaugherty, CA (1990) Predicting long-term patterns of mass loss, nitrogen dynamics and soil organicmatter formation from initial fine litter chemistry in temperateforest ecosystems. Can J Bot 68: 2201–2208

2. Agnelli, A, Ascher, J, Corti, G, Ceccherini, MT, Nannipieri, P,Pietramellara, G (2004) Distribution of microbial communities ina forest soil profile investigated by microbial biomass, soilrespiration and DGGE of total extracellular DNA. Soil BiolBiochem 36: 859–868

3. Bakken, LR, Lindahl, V (1995) Recovery of bacterial cells fromsoils. In: Trevors( JT, Van Elsas( JD (Eds.) Nucleic Acids in theEnvironment. Methods and Applications. Springer-Verlag, Berlin,pp 9–27

4. Baldock, J, Skjemstad, J (2000) Role of the soil matrix andminerals in protecting natural organic materials against biologicalattack. Org Geochem 31: 697–710

5. Barber, A, Cabrera, MR, Guardiola, I (1997) Sobre la cultura del’espart al territori Valencia. Fundacio Bancaixa. Valencia, pp 252

6. Blume, E, Bischoff, M, Reichert, JM, Moorman, T, Konopka, A,Turco, RF (2002) Surface and subsurface microbial biomass,community structure and metabolic activity as a function of soildepth and season. Appl Soil Ecol 20: 171–181

7. Bundt, M, Widmer, F, Pesaro, M, Zeyer, J, Blaser, P (2001)Preferential flow paths: biological Fhot spots_ in soils. Soil BiolBiochem 33: 729–738

8. Campbell, CD, Grayston, SJ, Hirst, DJ (1997) Use of rhizospherecarbon sources in sole carbon source tests to discriminate soilmicrobial communities. J Microbiol Methods 30: 33–41

9. Ekelund, F, Rønn, R, Christensen, S (2001) Distribution withdepth of protozoa, bacteria and fungi in soil profiles from threeDanish forest sites. Soil Biol Biochem 33: 475–481

10. Ettema, CH, Wardle, DA (2002) Spatial soil ecology. Trends EcolEvol 17: 177–183

11. Fierer, N, Schimel, JP, Holden, PA (2003) Variations in microbi-ological community composition through two soil depth profiles.Soil Biol Biochem 35: 167–176

12. Fritze, H, Pietikainen, J, Pennanen, T (2000) Distribution ofmicrobial biomass and phospholipid fatty acids in Podzol profilesunder coniferous forest. Eur J Soil Sci 51: 565–573

13. Garland, JL, Mills, AL (1991) Classification and characterization ofheterotrophic microbial communities on the basis of patterns ofcommunity-level sole-carbon-source utilization. Appl EnvironMicrobiol 57: 2351–2359

14. Garland, JL, Mills, AL, Young, JS (2001) Relative effectiveness ofkinetic analysis vs single point readings for classifying environ-mental samples based on community-level physiological profiles(CLPP). Soil Biol Biochem 33: 1059–1066

15. Grayston, SJ, Griffith, GS, Mawdsley, JL, Campbell, CD, Bardgett,RD (2001) Accounting for variability in soil microbial communi-ties in temperate grassland ecosystems. Soil Biol Biochem 33: 533–551

16. Griffiths, BS, Ritz, K, Wheatley, RE (1997) Relationship betweenfunctional diversity and genetic diversity in complex microbialcommunities. In: Insam H, Rangger( A (Eds.) Microbial Commu-nities. Functional versus( Structural Approaches. Springer-Verlag,Berlin, pp 1–9

17. Griffiths, RI, Whiteley, AS, O’Donell, AG, Bailey, MJ (2003) In-fluence of depth and sampling time on bacterial community struc-ture in an upland grassland soil. FEMS Microbiol Ecol 43: 35–43

18. Grove, AT, Rackham, O (2001) The Nature of Mediterranean Europe.An Ecological History. Yale University Press London, pp 384

19. Guerrero, J. (1998) Respuestas de la vegetacion y de la morfologıade las plantas a la erosion del suelo. Valle del Ebro y PrepirineoAragones. Publicaciones del Consejo de la Proteccion de laNaturaleza de Aragon, p 257.

20. Haack, SK, Garchow, H, Klug, MJ, Forney, LJ (1995) Analysis offactors affecting the accuracy, reproducibility, and interpretationof microbial community carbon source utilization patterns. ApplEnviron Microbiol 61: 1458–1468

21. Heuer, H, Krsek, M, Baker, P, Smalla, K, Wellington, EMH (1997)Analysis of actinomycete communities by specific amplification ofgenes encoding 16S rRNA and gel-electrophoretic separation indenaturing gradients. Appl Environ Microbiol 63: 3233–3241

22. Ingelmo, F, Molina, MJ, Ruiz, M (2001) Water stable aggregates insimulated eroded soils affected by fire intensity. In: Villacampa Y,Brebbia( CJ, Uso( JL (Eds.) Ecosystems and Sustainable Develop-ment III. Wit Press, Southampton, UK, pp 311–320

23. Insam, H (1997) A new set of substrates proposed for communitycharacterization in environmental samples. In: Insam( H, Rangger(A (Eds.) Microbial Communities. Functional versus StructuralApproaches. Springer-Verlag, Berlin, pp 259–260

24. Insam, H, Goberna, M (2004) Use of Biolog\ for the communitylevel physiological profiling (CLPPs) of environmental samples.In: Akkermans( ADL, Van Elsas( JD, DeBruijn( FJ (Eds.) MolecularMicrobial Ecology Manual, 2nd ed. Kluwer Academic Publishers,The Netherlands, pp 1–8

25. Lavahun, MFE, Joergensen, RG, Meyer, B (1996) Activity andbiomass of soil microorganisms at different depths. Biol Fertil Soils23: 28–42

26. Lindstrom, JE, Barry, RP, Braddock, JF (1998) Microbial commu-nity analysis: a kinetic approach to constructing potential C sourceutilization patterns. Soil Biol Biochem 30: 231–239

27. Magill, A, Aber, JD (2000) Dissolved organic carbon and nitrogenrelationships in forest litter as affected by nitrogen deposition. SoilBiol Biochem 32: 603–613

28. Magurran, AE (1989) Ecological Diversity and Its Measurement.Vedra, Barcelona, pp 200

29. Maranes, A, Sanchez, JA, de Haro, S, Sanchez, ST, del Moral, F(1998) Analisis de suelos. Metodologıa e interpretacion. Universi-dad de Almerıa. Almerıa, pp 184

30. Mayr, C, Winding, A, Hendriksen, NB (1999) Community level


physiological profile of soil bacteria unaffected by extractionmethod. J Microbiol Methods 36: 29–33

31. Muyzer, G, De Waal, EC, Uitterlinden, AG (1993) Profiling ofcomplex microbial populations by denaturing gradient gel elec-trophoresis analysis of polymerase chain reaction-amplified genescoding for 16S rRNA. Appl Environ Microbiol 59: 695–700

32. Naeem, S, Loreau, M, Inchausti, P (2002) Biodiversity and ecosystemfunctioning: the emergence of a synthetic ecological framework. In:Loreau( M, Naeem( S, Inchausti( P (Eds.) Biodiversity and EcosystemFunctioning. Oxford University Press, Avon, pp 3–11

33. Nubel, U, Engelen, B, Felske, A, Snaidr, J, Wieshuber, A, Amann,W, Ludwig, W, Backhaus, H (1996) Sequence heterogeneities ofgenes encoding 16S rRNAs in Paenibacillus polymyxa detected bytemperature gradient gel electrophoresis. J Bacteriol 178: 5636–5643

34. Øvreas, L, Torsvik, V (1998) Microbial diversity and communitystructure in two different agricultural soil communities. MicrobEcol 36: 303–315

35. Øvreas, L, Jensen, S, Daae, FL, Torsvik, V (1998) Microbialcommunity changes in a perturbed agricultural soil investigated bymolecular and physiological approaches. Appl Environ Microbiol64: 2739–2742

36. Parrish, JA, Bazzaz, FA (1976) Underground niche separation insuccessional plants. Ecology 57: 1281–1288

37. Porta, J, Lopez-Acevedo, M, Rodrıguez, R (1986) Tecnicas yExperimentos en Edafologıa. Collegi Oficial d’Enginyers Agronomsde Catalunya, Barcelona, pp 283

38. Preston-Mafham, J, Boddy, L, Randerson, PF (2002) Analysis ofmicrobial community functional diversity using sole-carbon-sourceutilisation profiles—a critique. FEMS Microbiol Ecol 42: 1–14

39. Rickwood, D, Ford, T, Graham, J (1982) Nycodenz: a newnonionic iodinated gradient medium. Anal Biochem 123: 23–31

40. Rovira, P, Vallejo, R (2002) Mineralization of carbon and nitrogenfrom plant debris, as affected by debris size and depth of burial.Soil Biol Biochem 34: 327–339

41. Sharma, S, Rangger, A, Insam, H (1998) Effects of decomposingmaize litter on community level physiological profiles of extractedbacteria. Microb Ecol 35: 301–310

42. Soil Survey Staff (1998) Keys to Soil Taxonomy. NRCS, USDAWashington, DC, pp 326

43. Taylor, JP, Wilson, B, Mills, MS, Burns, RG (2002) Comparison ofmicrobial numbers and enzymatic activities in surface soils andsubsoils using various techniques. Soil Biol Biochem 34: 387–401

44. Wardle, DA, Giller, KE (1996) The quest for a contemporaryecological dimension to soil biology. Soil Biol Biochem 27: 1601–1610

45. Yan, F, McBratney, AB, Copeland, L (2000) Functional substratebiodiversity of cultivated and uncultivated A horizons of vertisolsin NW New South Wales. Geoderma 96: 321–343

46. Yao, H, He, Z, Wilson, MJ, Campbell, CD (2000) Microbialbiomass and community structure in a sequence of soils with in-creasing fertility and changing land use. Microb Ecol 40: 223–237

47. Zak, JC, Willig, MR, Moorhead, DL, Wildman, H (1994)Functional diversity of microbial communities: A quantitativeapproach. Soil Biol Biochem 26: 1101–1108

48. Zar, JH (1996) Biostatistical Analysis. Prentice-Hall, New Jersey,pp 662

49. Caturla, RN (2002) Efecte del foc i la recurencia d’incendis encamps abondonats dominats per Brachypodium retusum (Pers.)Beauv. a la Comunitat Valenciana. PhD Thesis, Universitatd’Alacant, 188 pp


microbial community structure at different depths in disturbed and undisturbed semiarid...

Documents