lable at ScienceDirect

Soil Biology & Biochemistry 43 (2011) 1333e1340

Contents lists avai

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lb io

Soil bacterial growth and nutrient limitation along a chronosequencefrom a glacier forefield

Hans Göransson a,b,*, Harry Olde Venterink a, Erland Bååth c

a Institute of Integrative Biology, ETH Zurich, Universitätstrasse 16, CHN 8092 Zurich, Switzerlandb School of the Environment and Natural Recourses, Bangor University, Bangor LL 57 2UW, UKcMicrobial Ecology, Department of Biology, Ecology Building, Lund University, SE-223 62 Lund, Sweden

a r t i c l e i n f o

Article history:Received 4 October 2010Received in revised form1 February 2011Accepted 6 March 2011Available online 21 March 2011

Keywords:Leucine incorporationCarbon qualityMicrobial activityChronosequenceC:N:P stoichiometrySoil development

* Corresponding author. School of the EnvironmBangor University, Bangor LL 57 2UW, UK. Tel.: þ44354997.

E-mail addresses: h.goransson@bangor.ac.uk (H.Görenv.ethz.ch (H. Olde Venterink), Erland.Baath@mbioeko

0038-0717/$ e see front matter Crown Copyright � 2doi:10.1016/j.soilbio.2011.03.006

a b s t r a c t

Resource availability and limiting factors for bacterial growth during early stages of soil development(8e138 years) were studied along a chronosequence from the glacial forefield of the Damma glacierin the Swiss Alps. We determined bacterial growth (leucine incorporation) and we investigated whichresource (C, N or P) limited bacterial growth in soils formed by the retreating glacier. The latter wasdetermined by adding labile sources of C (glucose), N and P to soil samples and then measuring thebacterial growth response after a 40 h incubation period. Bacterial growth increased with increasingsoil age in parallel with the build up of organic matter. However, lower bacterial growth, whenstandardized to the amount of organic C, was found with time since the glacier retreat, indicatingdecreasing availability of soil organic matter with soil age. Bacterial growth in older soils was limitedby the lack of C. The bacteria were never found to be limited by only N, only P, or N þ P. In theyoungest soils, however, neither the addition of C, N nor P singly increased bacterial growth, whilea combination of C and N did. Bacterial growth was relatively more limited by lack of N than P whenthe C limitation was alleviated, suggesting that N was the secondary limiting resource. The avail-ability of N for bacterial growth increased with time, as seen by an increased bacterial growthresponse after adding only C in older soils. This study demonstrated that bacterial growthmeasurements can be used not only to indicate direct growth effects, but also as a rapid method toindicate changes in bacterial availability of nutrients during soil development.

Crown Copyright � 2011 Published by Elsevier Ltd. All rights reserved.

1. Introduction

Glaciers in the Alps have been withdrawing since the little iceage ended in 1850. This has resulted in chronosequences, withincreasingly older soils with distance from the ice, that can be usedto study soil formation, and microbial and vegetation succession(Matthews, 1992). Immediately following the withdrawal of theglacier, the rocky ‘soil’ is poor in organic material. The concentra-tion of carbon (C) and nitrogen (N) is low, whereas phosphorus (P)can be higher but will be present mainly as apatite (Matthews,1992). During the first 100 years after deglaciation, the concen-tration of organic matter, and thus also those of total C and N,

ent and Natural Recourses,1248 383056; fax: þ44 1248

ansson), harry.oldeventerink@l.lu.se (E. Bååth).

011 Published by Elsevier Ltd. All

increase exponentially in the soil, after which they stabilise andeventually reach plateaus (Jacobson and Birks, 1980; Matthews,1992; Walker, 1993; Chapin et al., 1994). In contrast, total Pusually declines with soil development through weatheringprocesses and leaching (Walker and Syers, 1976). Thus, the limitingnutrient for plant growth may shift from N to P with time, althoughonly after several hundred to thousand of years (Chapin et al., 1994;Wardle et al., 2004).

Even though the changes in total amounts of C, N and P in thesoil during the early soil development, and changes in limitingnutrients during plant succession, are rather well known, changesin the available amounts of nutrients for microorganisms have beenless studied. However, it is well known that the size of themicrobialbiomass and activity will be determined by the amount of organicmatter in soil (Wardle, 1992). The increasing concentration oforganic matter with soil age along a retreating glacier transect willthus likely result in increased microbial activity, e.g. respiration,and increased microbial biomass (Insam and Haselwandter, 1989;Wardle et al., 2004). However, at the same time as the soil

rights reserved.

Fig. 1. The Damma glacier forefield with marked sampling sites and glacier advancesand retreats.

H. Göransson et al. / Soil Biology & Biochemistry 43 (2011) 1333e13401334

organic matter concentration will increase, a larger proportion of itwill become older and more resistant to decomposition, resultingin soil organic matter becoming less available for microbial use.Microbial activity normalized to soil organic matter, that is,measured as respiration per soil organic matter or soil organic C,will therefore likely decrease with soil age along a glacier chro-nosequence (Schipper et al., 2001; Tscherko et al., 2003).

The amount of available nutrients for soil microbes is difficult toestimate. There are no standardised extractionmethods available todetermine microbially available C, N and P fractions. Methods usingrespiration as a proxy for microbial growth to indicate available Nor P usually rely on initial additions of large amounts of glucose(Nordgren, 1992), and thus the relation between C availability andavailability of other nutrients cannot be studied. An alternativeapproach is to directly estimate bacterial growth in soil. Aldén et al.(2001) described such amethod, which later has been used to studylimiting factors for bacterial growth in different soils (Demolinget al., 2007, 2008; Rinnan et al., 2007). The method is based onthe addition of single nutrients or of nutrient combinations, fol-lowed by the determination of bacterial growth using leucineincorporation. The primary limiting nutrient is identified as thenutrient that increases bacterial growth in comparison to anunamended control soil, at the same time as there are no or onlya minor increase in bacterial growth adding any other nutrients.Also, by combining the addition of different nutrients in a factorialdesign, it may be possible to determinewhich nutrient is secondaryin limiting growth; i.e., the nutrient that further increases growthwhen limitation by the primary limiting nutrient is alleviated(Demoling et al., 2007; Rinnan et al., 2007). Thus, besides identi-fying which nutrient has the lowest availability, you also get anindication of the relative availability of other nutrients (Demolinget al., 2008).

Yoshitake et al. (2007) found that close to a retreating glacier,additions of both C and N were necessary to increase microbialrespiration, whereas further away from the glacier only C additionwas necessary. This suggests that microbially available N builds upfaster in soil than microbially available C, and agrees with obser-vations that microbial growth is primary limited by C andsecondary limited by N in most developed soils (Paul and Clark,1996; Demoling et al., 2007). P limitation of microbial growth hasmainly been found in soils with high P sorption capacity (Aldénet al., 2001; Ilstedt and Singh, 2005; Ehlers et al., 2010), butmostly P was only the third limiting nutrient (Demoling et al.,2007).

We investigated resource availability and limiting nutrients forbacterial growth along a chronosequence of soil development inthe forefield of the Damma glacier in Switzerland. To our knowl-edge, this is the first study of bacterial growth limitation along sucha chronosequence. We measured bacterial growth and factorslimiting bacterial growth by adding C, N and P to soil samples. Wetested five hypotheses:

1) Bacterial growth per gram of soil will increase with soil age inparallel with increased soil organic C concentration.

2) Bacterial growth calculated per organic C will decrease withsoil age, due to a smaller proportion of the accumulated organicmatter being of an easily available form.

3) Bacterial growth will be primarily limited by C. This will bereflected by increased bacterial growth upon C addition, but noincrease upon single additions of N or P.

4) Bacterial growth will be secondarily limited by N. This will beseen as a larger increase in bacterial growth after adding C incombination with N, than after C in combination with P.

5) The ‘extent of C-limitation’will increasewith soil age, due to andecreasing proportion of the accumulated C being in an

available form, and also while N will build up with time. Thiswill be seen as an increasing growth response of adding only Cwith age of the soils.

2. Methods

2.1. Site description

The Damma glacier (8� E 2703000, 46�N 3800000) is situated in theSwiss Alps (Fig. 1). The altitude is 2054 m at the front of the glacierand 1920 m at 1 km from the ice, where the oldest sampling sitewas situated. From October 2007 to October 2009 the mean annualtemperature measured on the forefield was 2.2 �C, and the meanannual precipitationwas 1781mmyear (J. Magnusson, unpublisheddata). The Damma glacier has retreated approximately 1000 msince about 1850 (data from the Swiss Glacier Monitoring Network;http://glaciology.ethz.ch/swiss-glaciers/glaciers/damma.html).During this period the retreat has been interrupted, and the glacieradvanced twice, once between 1911 and 1928 and once between1972 and 1991. Thus, the forefield can be divided into three areas:(1) less than 19 years, (2) 58e81 years, and (3) 109e150 years, each

H. Göransson et al. / Soil Biology & Biochemistry 43 (2011) 1333e1340 1335

forming a gradient in time with the youngest soil nearest the endmoraine/ice and becoming older with distance from it (Fig. 1). Thus,the Damma glacier forefield can be seen as 3 repeated gradients,with the distance to each end moraine/ice as a proxy of soil age.

The study was carried out with soil collected at 21 sites used inthe Big-Link project (Bernasconi et al., 2008) along a chronose-quence of 8e138 years after deglaciation. Common plant species ofthe young soils are Agrostis gigantea, Rumex scutatus and Cerastiumuniflorum; the vegetation in the soil of intermediate age is domi-nated by A. gigantea, Salix sp. and Deschampsia cespitosa, and in theold soils by Rhododendron ferrugineum, A. gigantea and Festucarubra. The last site (no. 21) of the old soils is dominated by thenitrogen fixing tree Alnus viridis.

2.2. Soil sampling and chemical analyses of soils

Soil samples were collected in September 2009 from the top5 cm of the mineral soil after the aboveground vegetation and, ifpresent, the organic layer had been removed. Four samples weretaken in the area less than 19 years old (sites 1e4), 12 samples inthe area of intermediate age (sites 5e16) and 5 in the oldest area(sites 17e21), thus forming 3 gradients in soil age (Fig. 1, Table 1).

One part of each soil sample was frozen for chemical analysesand one part sieved fresh (2 mmmesh size) and stored at 4 �C up to2 months during the analyses of limiting factors for bacterialgrowth (see Section 2.3). The frozen soils were thawed and wet-sieved (2 mm). One part was dried at 70 �C for 4 days and analysedfor C and N by combustion (LECO CNS-2000). One part was freshlyextracted for available P with 0.5 M NaHCO3, filtered with activatedcoal to remove colour and analysed with the molybdenum bluemethod (John, 1970).

2.3. Bacterial growth

Bacterial growthwas determinedwith the leucine incorporationmethod described by Bååth et al. (2001). 20 ml of distilled waterwas added to 50 ml centrifuge tubes with 2 g of soil and vortexedfor 3 min. After centrifugation at 1000 � g for 10 min, 1.5 ml of the

Table 1Soil nutrients in the 21 sites of the chronosequence of soil development after retreatof the Damma glacier. Age is expressed as years since deglaciation. Distance to themoraine is the proxy for actual soil age that takes into account that the glacier retreatwas interrupted twice, and in these periods the glacier advanced again.

Site Age Distance toend moraine/ice (m)

N (% dw) C/N ratio(w/w)

NaHO3

ExtractableP (mg g-1 dw)

Total N/extractableP (w/w)

2 8 75 0.002 12.4 3.3 6.21 9 96 0.002 14.5 5.2 4.54 13 130 0.007 13.0 7.1 9.83 15 150 0.016 11.9 11.0 14.45 60 30 0.004 14.1 2.4 16.07 63 90 0.012 11.4 6.5 18.16 64 105 0.022 12.4 7.2 30.38 66 150 0.011 13.4 18.1 6.19 68 200 0.034 11.8 7.4 46.210 69 225 0.016 15.8 9.0 17.911 72 260 0.033 13.2 5.0 66.912 74 315 0.105 12.9 20.1 52.513 77 365 0.042 12.4 10.9 38.614 77 340 0.030 11.7 5.6 52.516 79 420 0.039 15.4 5.3 74.515 80 430 0.120 12.6 19.2 62.417 112 30 0.060 15.1 10.0 60.318 119 120 0.025 15.0 8.2 30.019 122 155 0.062 13.1 5.9 104.820 130 250 0.073 13.0 6.6 111.321 138 350 0.094 11.8 8.2 113.6

supernatant was transferred to 2 ml micro-centrifuge tubes.L-[4,5-3H]-Leucine (2 ml, 155 Ci mmol�1, Amersham, UK) was addedtogether with nonlabelled leucine resulting in 275 nM leucine inthe bacterial suspension. The tubes were then vortexed and incu-bated for 2 h. Removal of non-incorporated leucine, and prepara-tion for scintillation counting was performed according to Bååthet al. (2001), except that after adding the scintillation cocktail(Ultima Gould, PerkinElmer), the samples were centrifuged at13000 rpm for 3 min, a procedure which decreased quenching bysoil particles. Bacterial growth is expressed in pmol leucine incor-porated into extracted bacteria per gram dry soil and hour unlessotherwise stated.

2.4. Determination of limiting nutrients for bacterial growth

For measurements of the microbial growth response afteradding C, N and P, i.e. determination of the liming nutrients forbacterial growth in soil, a modification of the method developed byAldén et al. (2001) was used. Two grams of fresh soil was put into50 ml centrifuge tubes. Glucose, NH4NO3 and KH2PO4 was added indifferent combinations (designed hereafter as C, N, P, CN, CP, NP andCNP) and one no nutrient addition control (No add.). The nutrientswere added as water solutions (totally 40 ml g�1 soil). The tubeswere closed with lids and incubated for 40 h at 20 �C. Bacterialgrowth was then determined as described above.

The design of each assay of limiting factors is basically a factorialdesign by adding C, N and P in combinations. Initially the amount ofCNP needed to achieve a growth response was determined in allsoils. The CNP addition was used as a positive control to indicatethat sufficient amounts had been added to remove the nutrientlimitation (Demoling et al., 2007). However, too high concentrationof added nutrients resulted in very low bacterial growth, either dueto altered osmotic conditions or that soils became anaerobic withproduction of organic acids decreasing pH during the 40 h incu-bation (seen as a drastic decrease in pH). The amounts of nutrientchosen for the final assays varied between 0.03 and 0.8 mg g�1 forC, 0.008e0.015 mg g�1 for N and 0.02e0.1 mg g�1 for P dependingon the soil. The lower amounts were needed at the younger siteswhere higher additions caused toxic effects leading to low bacterialgrowth. The higher amounts were needed at the older sites withmore organic matter and thus higher microbial biomass in order toachieve any growth response in the CNP treatment compared to theno addition control. At the chosen concentrations of CNP the meanbacterial growth response (CNP/No add. control) was 6.3 timeshigher after CNP addition than in the no addition control, with allsoils except one having a ratio >3.7.

Then all the soils were analysed using a full factorial designwithonly one replicate for each nutrient treatment. We found no indi-cation of a bacterial growth response adding N, P or NP (tested bya two-way ANOVA without replication) in any site (see alsoResults). Each soil sample was then analysed again (totally 2e4times) using only the additions No addition, C, CN, CP and CNP. TheCNP treatment was only used as a positive control, while the otherswere used for further analyses (see below).

2.5. Statistics

Mean values for the 2e4 analyses made for each soil werecalculated. For each soil a number of ratios of bacterial growth werethen calculated and regressed against distance from the endmoraine/ice as a proxy of soil age (n ¼ 21). We regressed theresponse variables both against the entire dataset, but also for eachseparated gradient. The C/No add. control ratio indicated C limita-tion for the bacterial community, with lower values when anothernutrient was close to being limiting. The CN/C and CP/C ratio was

0

20

40

60

80

100

120

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

Leu

cine

inco

rpor

atio

n / g

ram

soi

l

(pm

ol h

-1 g

-1)

Soil carbon content (%)

R2=0.59P<0.001

A

Young gradientMiddle gradientOld gradient

3 104

4 104

5 104

pora

tion

/ gra

m c

arbo

n

ol h

-1 g

C-1

)B

H. Göransson et al. / Soil Biology & Biochemistry 43 (2011) 1333e13401336

used to indicate the extent of N or P being secondary limitingnutrient. All the ratios were log-transformed before the regressionwas calculated.

3. Results

3.1. Chemical analyses

Soil C concentration increased exponentially from the youngestsite near the ice edge (site 2) towards the 1992 end moraine (site 3,Fig. 2). On the other side of the end moraine (site 5), the soil Cconcentration was set back to a value close to the one of theyoungest site and then increased towards the 1928 end moraine,reaching values up to 1.5% C of dry soil. Soil C then decreased on theother side of the 1928 end moraine, followed by increased soil C upto 1% in the oldest sites (sites 20 and 21). Total soil N increased ina similar way as soil C (Table 1), and soil C and N were closelycorrelated (r ¼ 0.996, p < 0.001). The soil C:N ratio was notcorrelated with the age since deglaciation or the distance to theclosest younger end moraine/ice, except for the old part of theforefield (sites 17e21), where the C:N ratio decreased with distanceto the 1928 moraine (Table 1, R2 ¼ 0.85, p < 0.05). NaHCO3extractable P increased significantly in the youngest part (site 1e4)with distance from the ice (p < 0.05), but not in the middle or oldpart (Table 1). However, total N: extractable P increased with thedistance to the closest endmoraine/ice (Table 1, R2¼ 0.28, p< 0.05).

3.2. Bacterial growth

Bacterial growth increased with increasing concentration oforganic C in the soil (Fig. 3A, R2 ¼ 0.59, p < 0.001). Thus, lowerbacterial growthwas found at the young sites low in organic matternear the endmoraine/ice. However, when standardized per organiccarbon content, significantly higher growth was found in theseyoung sites (Fig. 3B, R2 ¼ 0.59, p < 0.001), decreasing with distancefrom the end moraine/ice for the young and middle gradient,

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 20 40 60 80 100 120 140

Soil

carb

on (

%)

Years since deglaciation

R2 =0.99p<0.004

R2 =0.71p<0.001

Fig. 2. Total soil carbon in relation to soil age, as indicated by the number of years afterdeglaciation. Each transect was fitted to an exponential equation (only significant onesshown).

0

1 104

2 104

0 100 200 300 400 500

Leu

cine

inco

r

(pm

Distance from morain/ice (m)

R2=0.59P<0.001

Fig. 3. Bacterial growth, measured by means of leucine incorporation into bacteriaextracted from soil in relation to (a) soil carbon content, or (b) distance to the moraine/ice as a proxy for soil age (see Section 4.1). Note that in (a) bacterial growth iscalculated per gram of soil and in (b) bacterial growth is standardized per gram oforganic C.

whereas it was equally low all along the old gradient of theforefield.

3.3. Growth limitations

There was no major increase in bacterial growth after singleadditions of N or P, or of N and P in combination (NP) in any of thesoils. The mean bacterial growth response to additions of N, P or NPin comparison to the no addition control were, calculated for allsoils, N/No add. control¼ 0.91 (SD¼ 0.24), P/No add. control¼ 1.22(SD¼ 0.35), and NP/No add. control¼ 1.23 (SD¼ 0.50). The bacteria

H. Göransson et al. / Soil Biology & Biochemistry 43 (2011) 1333e1340 1337

in most of the soils responded to C addition by increasing growthcompared to the unfertilized control (Fig. 4A). Only in the youngersoils, nearest to the end moraine/ice, we observed no or only a verysmall growth response to C addition (sites 1, 2, 5, 7 and 17), while atthe two sites furthest away from their respectively end moraine(sites 15 and 21), bacterial growth increased 3e4 times after addingC compared with the no addition control. Thus, the growthresponse of the bacteria due to C addition in comparison to theunfertilized control (C/No add. control ratio) increased withincreasing distance from the closest younger end moraine/ice(Fig. 4A, R2 ¼ 0.26, p < 0.05). An increase was also found for the 3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0 100 200 300 400 500

Log

C/N

o ad

d. c

ontr

ol

Distance from morain/ice (m)

R2=0.26P=0.018

A

Young gradientMiddle gradientOld gradient

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

11 12 13 14 15 16

C/N in soil

Log

C/N

o ad

d. C

ontr

ol

B

Fig. 4. Bacterial growth upon C addition compared to that in the unamended control(C/No add. control ratio) in relation to (a) the distance to the moraine/ice as a proxy forsoil age (see Section 4.1), or (b) the C:N ratio in the soil.

gradients treated separately, being significant (p < 0.05) for theyoungest and oldest gradient.

Overall, the bacterial growth response to C addition was onlyweakly related to the C:N ratio in the soil (Fig. 4B). However, in soilswith a C:N ratio above about 13 bacterial growth never respondedstrongly to C addition, while in most soils with a lower C:N ratiousually bacterial growth was limited by the lack of C.

The bacterial growth response induced by CN addition ascompared to C addition (CN/C ratio) decreased with increasingdistance from the closest end moraine/ice (Fig. 5, R2 ¼ 0.34,p < 0.01) A decrease was also found for the 3 separate gradients,although only significant for the oldest one. The CP/C ratio did notchange along the gradient (Fig. 5).

4. Discussion

4.1. Soil development

The increases in soil C and N concentrations, resulting from thebuild up of organic material, were not directly correlated to the agesince deglaciation. The ice recession has been interrupted twicesince the 1850-ies (Fig. 1). Judging from C and N in the soils justbelow the 1992 moraine (sites 5 and 7 in the middle gradient),these soils were of similar age as the sites at the young gradient ofthe forefield of less than 15 years and not like a 50 year old soil,which is the time since deglaciation (Table 1). At the sites justbelow the 1928 moraine (old gradient), the C and N concentrationsin the soil were similar to those for the sites just above the 1928moraine. The young and the intermediately aged gradient thus startwith almost no organic material in the soil, whereas the oldgradient starts with a more mature soil richer in organic material.Still, the variation in soil C and N content strengthened the conceptof 3 separate gradients, with distance from the end moraine/ice asa reasonable proxy for soil age, and not one gradient related to agesince deglaciation.

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0 100 200 300 400 500

log CN/C

log CP/C

Dictance from morain/ice (m)

R2=0,34

p<0.01

Log

CN

/C o

r C

P/C

Fig. 5. Bacterial growth after C þ N addition or after C þ P addition compared to thatwith only C addition (CN/C and CP/C ratios) in relation to the distance to the moraine/ice as a proxy for soil age (see Section 4.1).

H. Göransson et al. / Soil Biology & Biochemistry 43 (2011) 1333e13401338

4.2. Bacterial growth

Bacterial growth increased with increasing concentration oforganic matter in soil, as evidenced by the positive correlation to %organic C (Fig. 3A). Thus, higher total bacterial growth was found inolder soils when calculated per gram of soil (according to hypoth-esis 1). A similar result was found at the Damma glacier usingfluorescein diacetate, where a higher activity was found withincreasing soil age (Sigler et al., 2002). However, to calculatebiomass and activity per gram of soil is not necessarily the best wayof presenting data from a microbiological point of view, sincedifferences in microbial availability of the organic matter will beconfounded by the amount of organic matter. Examples of this areaddition of metal contaminated sludge, where the negative effect ofheavy metals is obscured by the positive effects of increasedconcentration of organic matter from the sludge amendment(Chander and Brookes, 1991; Witter et al., 2000). This confoundingeffect of organicmatter concentration and availability was also seenin the present study. When expressing bacterial growth per unitorganic C the organic matter in the younger soils was moreconducive for bacterial growth than in older soils (hypothesis 2),although the concentration of organic C per gram of soil was verylow in young soils (Figs. 2 and 3B). This is in accordance with earlierstudies using respiration or biomass to indicate the availability oforganic matter for microbial growth and activity in a glacier formedchronosequence (Insam and Haselwandter, 1989; Wardle et al.,2004). Sigler and Zeyer (2004) also found more fast-growingbacteria in younger soils at the Damma glacier, also suggestinghigher availability of organic matter in these soils.

4.3. Primary limiting nutrient for bacterial growth

There was a bacterial growth response to C addition ascompared to the no addition control (C/No add. control ratio) in allcases for older soils (Fig. 4A), while there was no major effect ofadding N and P singly or in combination (NP). The slightly lowergrowth response after adding N (N/No add. control) and the slightlyhigher to P (P/No add control.) as compared to the control isprobably related to changes in pH. Adding NH4NO3 slightlydecreased pH (also found by Aldén et al., 2001) and adding KH2PO4slightly increased pH in the soils. Bacterial activity is very sensitiveto pH changes, where also a small increase in pH can increaseavailability of DOM in soil (Kennedy et al., 1996), thereby affectingbacterial growth. We thus conclude that except in the youngestsoils (see Section 4.5), bacterial growth was limited by the lack of C(hypothesis 3). Our results are in accordance with bacteria inmature soils commonly being C limited (Paul and Clark, 1996). Thishas been demonstrated in various mature soils using the samemethodology as in the present study (Demoling et al., 2007), andhas also been suggested by measuring respiration after addingnutrients in many soils (Duah-Yentumi et al., 1998; Yoshitake et al.,2007).

Growth limitation by a nutrient is an absolute trait in that it iseither limiting or not. However, by adding the limiting nutrient (inthis case C) in excess, bacterial growth increases until it is limitedby another nutrient. Thus, the extent of growth response afteradding C indicates the availability of the secondary limitingnutrient. Hence, the increasing response upon C addition (C/Noadd. control ratio) with soil age suggests that the relative avail-ability of the secondary limiting nutrient increased compared to theC availability with increasing soil age (see Section 4.4). Conse-quently, the extent of C limitation for bacterial growth increasedwith soil age (hypothesis 5), which is in accordance with soil Cbecoming less available with soil age (see Section 4.2).

4.4. Secondary limiting nutrient for bacterial growth

In our youngest soils, the bacterial growth response to CNaddition in comparison to only C addition (CN/C ratio) was higherthan that in response to CP addition (CP/C ratio) (Fig. 5). Thisindicated that Nwas the secondary limiting nutrient (hypothesis 4),that is, the limiting nutrient when the limiting conditions of themain nutrient, C, was alleviated by adding it in excess. This isconsistent with available P being relatively high in young soils,since P is present in the parent material, while N will be scarce(Richardson et al., 2004). This was also the case here, since lowestratios of soil total N:exchangeable P were found in the youngestsoils (Table 1).

The lack of a stronger response to CN addition than to C additionin more developed soils could suggests that N availability forbacteria increased with soil development, which if it continueseventually may result in another nutrient, like e.g. P, becoming thesecondary limiting nutrient. It is well known that with timeweathering and leaching will result in old soils becoming low in P,with vegetation becoming P limited (Wardle et al., 2004). However,this is unlikely to have happened at the Damma glacier, even at theoldest sites, since this process usually is important on a muchlonger time scale than for soils in the present study (Richardsonet al., 2004). If the bacteria would have become P limited therewould also be an increase in the CP/C ratio in the older part of thegradient, which was not the case (Fig. 5). Nitrogen beingthe secondary limiting nutrient is supported by the N:P ratio of theplants being around 10 along the entire gradient (H. Göransson,unpublished), which indicates N limitation (Olde Venterink et al.,2003; Güsewell, 2004). Moreover, the amount of available N inthe soil, measured as adsorption to resin, was also correlated to theN concentration of the vegetation (H. Göransson, unpublished).

One explanation for the low response to CN addition ascompared to C addition (CN/C) on the older soils could be technical.Adding N in high amounts often results in decreased bacterialgrowth (Aldén et al., 2001), which also was the case in thepreliminary stage of the present study, when different concentra-tions were tested (see Section 2.4). Eventually, the chosenconcentrations of nutrients added were a trade-off between addingenough to introduce maximum growth response, but still avoid thenegative growth effects of a too high addition of NH4NO3. We cantherefore not rule out that particularly in the oldest soils with highorganic matter contents, a higher dose of N could have induceda higher growth response in the CN treatment.

According to our fourth hypothesis, the extent of a growthresponse upon addition of surplus amounts of the primary limitingnutrient is an indication of relative availability of the secondarylimiting nutrient. Assuming that C and N were the primary andsecondary limiting nutrients in our soils, this hypothesis was sup-ported: With increasing distance from the end morain/ice wemeasured both an increasing availability of N, as well as a signifi-cant increase in bacterial growth response upon C addition incomparison to the unamended control (C/No add. control ratio,Fig. 4A).

4.5. Co-limitation for bacterial growth

In the youngest soils in each part of the forefield there was no oronly a small positive response of bacterial growth after adding Cand N singly, but in combination they stimulated a growthresponse, suggesting C and N co-limitation. Co-limitation by C andN for bacterial growth is rare, but it was reported in young soils neara retreating glacier on Svalbard (Yoshitake et al., 2007), and forsome N-poor arctic soils (F. Demoling, R. Rinnan, E. Bååth, unpub-lished), the latter study also using the leucine incorporation

H. Göransson et al. / Soil Biology & Biochemistry 43 (2011) 1333e1340 1339

technique. Co-limitation of bacterial growth has also been reportedfor aquatic ecosystems (Elser et al., 2009).

According to theory, a shift from C limitation to N limitationshould occur at a C:N ratio of the organic matter between 8 and 16;i.e., assuming a bacterial C:N ratio between 4 and 8 and that 50% ofthe carbon is incorporated into biomass and 50% respired (Kaye andHart, 1997). We found clear C limitation in most soils with a C:Nratio up to just above 13 (Fig. 4B), but not in soils with higher C:Nratio (up to 16). However, even soils with C:N ratios of 28 haveearlier been found to be C limited for bacterial growth (Demolinget al., 2007). In old soils, physically and chemically recalcitrantsoil organic matter has been accumulated, resulting in ratios ofavailable C and N being much lower than that reflected by total Cand N concentrations measured in the soil. Moreover, we did notdetect any growth increase upon C addition in some soils with lowa C:N ratio (11e13, Fig. 4B), which further emphasises that the soilC:N ratio is not a fully reliable measure to infer limiting substancesfor bacterial growth in soil. This also underscores an importantdifference between aquatic and terrestrial systems, where in theformer system C/N ratios in DOM appears to be a good indicator ofwhich factor that limits bacterial growth (cf. Sterner and Elser,2002).

Another difference between aquatic and terrestrial systems isthat fungi is an important part of the decomposition processes insoil. Fungal biomass has a higher C:N ratio (about 5e17) thanbacterial biomass (about 4e8; Cleveland and Liptzin, 2007). Thus,in an N limited situation, fungi would have an advantage comparedto bacteria. This could imply that inducing N limitation by C addi-tion in our youngest soils could first have increased fungal growth.A subsequent quick exploitation of available nutrients by fungicould have resulted in the observed lack of bacterial growth. Hence,where a positive growth response after adding a nutrient is clearevidence of a nutrient limitation, the lack of response is far moredifficult to interpret, particularly if another group of organismswith similar function is present.

5. Conclusions

By combining direct bacterial growth measurements withresponses upon factorial nutrient supply, we were able to showhow soil formation, as induced by the retraction of a glacier,regulates bacterial growth in soil. At the least developed soils wefound evidence of C and N co-limitation for bacterial growth.Accumulation of organic matter and C and N in it, increasedbacterial growth, but at the same time accumulated more recalci-trant C. The accumulation of organic matter was faster then thedecrease in carbon quality resulting in that the extent of C limita-tion for bacterial growth increased with soil age.

Since the bacterial growth was mainly limited by lack of C, whileplants in this gradient were mainly limited by lack of N, one mightconclude that bacteria will not compete with plants for N. However,also C-limited bacterial growth will have a demand for N. Theextent of C limitation could thus be important in determiningcompetition between bacteria and roots. The results of this studysuggest that bacteria will be more important competitors for N inyoung soils, but that their influence as N competitors decreaseswith soil age. However, to complete this viewwe need also to studythe other large decomposer organism group, the fungi, in thisrespect.

Acknowledgements

The study was funded by the Competence Center Environmentand Sustainability of the ETH Domain (CCES, project Big-Link), andalso supported by a grant from the Swedish Research Council

(Project No. 2009-4503) to E.B. We thank Dr. Johannes Rousk forconstructive criticism of the manuscript and Dr. Rienk Smittenbergfor Fig. 1.

References

Aldén, L., Demoling, F., Bååth, E., 2001. Rapid method of determining factors limitingbacterial growth in soil. Applied and Environmental Microbiology 67,1830e1838.

Bååth, E., Pettersson, M., Söderberg, K.H., 2001. Adaptation of a rapid andeconomical microcentrifugation method to measure thymidine and leucineincorporation by soil bacteria. Soil Biology and Biochemistry 33,1571e1574.

Bernasconi, S.M., Biglink Project Members, 2008. Weathering, soil formation andinitial ecosystem evolution on a glacier forefield: a case study from the DammaGlacier, Switzerland. Mineralogical Magazine 72, 19e22.

Chander, K., Brookes, P.C., 1991. Effects of heavy metals from past application ofsewage sludge on microbial biomass and organic matter accumulation in sandyloam and silty loam U.K. soil. Soil Biology & Biochemistry 23, 927e932.

Chapin III, F.S., Walker, L.R., Fastie, C.L., Sharman, L.C., 1994. Mechanisms of primarysuccession following deglaciation at Glacier Bay, Alaska. Ecological Monographs64, 149e175.

Cleveland, C.C., Liptzin, D., 2007. C:N:P stoichiometry in soil: is there a “Redfieldratio” for the microbial biomass? Biogeochemistry 85, 235e252.

Demoling, F., Figueroa, D., Bååth, E., 2007. Comparison of factors limiting bacterialgrowth in different soils. Soil Biology & Biochemistry 39, 2485e2495.

Demoling, F., Nilsson, L.O., Bååth, E., 2008. Bacterial and fungal response to nitrogenfertilization in three coniferous forest soils. Soil Biology & Biochemistry 40,370e379.

Duah-Yentumi, S., Rönn, R., Christensen, S., 1998. Nutrients limiting microbialgrowth in a tropical forest soil of Ghana under different management. AppliedSoil Ecology 8, 19e24.

Ehlers, K., Bakken, L.R., Frostegård, Å, Frossard, E., Bünemann, E.K., 2010. Phos-phorus limitation in a Ferralsol: impact on microbial activity and cell internalP pools. Soil Biology & Biochemistry 42, 558e566.

Elser, J.J., Kyle, M., Steger, L., Nydick, K.R., Baron, J.S., 2009. Nutrient availability andphytoplankton nutrient limitation across a gradient of atmospheric nitrogendeposition. Ecology 90, 3062e3073.

Güsewell, S., 2004. N:P ratios in terrestrial plants: variation and functional signif-icance. New Phytologist 164, 243e266.

Ilstedt, U., Singh, S., 2005. Nitrogen and phosphorus limitations of microbialrespiration in a tropical phosphorus-fixing acrisol (ultisol) compared withorganic compost. Soil Biology & Biochemistry 37, 1407e1410.

Insam, H., Haselwandter, K., 1989. Metabolic quotient of the soil microflora inrelation to plant succession. Oecologia 79, 174e178.

Jacobson, G.L., Birks, H.J.B., 1980. Soil development on recent end moraines of theKlutan glacier, Yukon Territory, Canada. Quaternary Research 14, 87e100.

John, M.K., 1970. Colorimetric determination of phosphorus in soil and plantmaterials with ascorbic acid. Soil Science 109, 214e220.

Kaye, J.P., Hart, S.C., 1997. Competition for nitrogen between plants and soilmicroorganisms. Trends in Ecology and Evolution 12, 139e143.

Kennedy, J., Billett, M.F., Duhie, D., Fraser, A.R., Harrison, A.F., 1996. Organic matterretention in an upland humic podzol; the effects of pH and solute type. Euro-pean Journal of Soil Science 47, 615e625.

Matthews, J.A., 1992. The Ecology of Recently Deglaciated Terrain: A GeoecologicalApproach to Glacier Forelands and Primary Succession. Cambridge UniversityPress, Cambridge.

Nordgren, A., 1992. A method for determining microbially available N and availableP in an organic soil. Biology and Fertility of Soils 13, 195e199.

Olde Venterink, H., Wassen, M.J., Verkroost, A.W.M., Ruiter, P.C., 2003. Speciesrichness-productivity patterns differ between N-, P-, and K-limited wetlands.Ecology 84, 2191e2199.

Paul, E.A., Clark, F.E., 1996. Soil Microbiology and Biochemistry, second ed. AcademicPress, San Diego.

Richardson, S.J., Peltzer, D.A., Allen, R.B., McGlone, M.S., Parfitt, R.L., 2004. Rapiddevelopment of phosphorus limitation in temperate rainforest along the FranzJosef soil chronosequence. Oecologia 139, 267e276.

Rinnan, R., Michelsen, A., Bååth, E., Jonasson, S., 2007. Mineralization and carbonturnover in subarctic heath soil as affected by warming and additional litter.Soil Biology & Biochemistry 39, 3014e3023.

Schipper, L.A., Degens, B.P., Sparling, G.P., Duncan, L.C., 2001. Changes in microbialheterotrophic diversity along five plant successional sequences. Soil Biology &Biochemistry 33, 2093e2103.

Sigler, W.V., Zeyer, J., 2004. Colony-forming analysis of bacterial communitysuccession in deglaciated soils indicates pioneer stress-tolerant opportunists.Microbial Ecology 48, 316e323.

Sigler, W.V., Crivii, S., Zeyer, J., 2002. Bacterial succession in glacial forefield soilscharacterized by community structure, activity and opportunistic growthdynamics. Microbial Ecology 44, 306e316.

Sterner, R.W., Elser, J.J., 2002. Ecological Stoichiometry: The Biology of Elementsfrom Molecules to the Biosphere. Princeton University Press, Princeton, NJ.

H. Göransson et al. / Soil Biology & Biochemistry 43 (2011) 1333e13401340

Tscherko, D., Rustemeier, J., Richter, A., Wanek, W., Kandeler, E., 2003. Functionaldiversity of the soil microflora in primary succession across two glacier forelandin the Central Alps. European Journal of Soil Science 54, 685e696.

Walker, T.W., Syers, J.K., 1976. The fate of phosphorus during pedogenesis.Geoderma 15, 1e19.

Walker, L.R., 1993. Nitrogen fixers and species replacement in primary succession.In: Miles, J., Walton, D.W.H. (Eds.), Primary Succession on Land. BlackwellScientific Publications, Oxford.

Wardle, D.A., Walker, L.R., Bardgett, R.D., 2004. Ecosystem properties and forestdecline in contrasting long-term chronosequences. Science 305, 509e513.

Wardle, D.A., 1992. A comparative assessment of factors which influence microbialbiomass carbon and nitrogen levels in soil. Biological Reviews of the CambridgePhilosophical Society 67, 321e358.

Witter, E., Gong, P., Bååth, E., Marstorp, H., 2000. A study of the structure andmetal tolerance of the soil microbial communitz six years after cessationof sludge application. Environmental Toxicology and Chemistry 19,1983e1991.

Yoshitake, S., Uchida, M., Koizumi, H., Nakatsubo, T., 2007. Carbon and nitrogenlimitation of soil microbial respiration in high artic successional glacier forelandnear Ny-Ålesund, Svalbard. Polar Research 26, 22e30.