Introduction

Volcanic landscapes typically consist of a mosaic of successional states (Kitayama et al., 1995). Deposit age, degree of weathering, precipitation, temperature and availability of potential colonists interact with other variables to determine patterns of ecosystem development at local-to-regional scales (Chadwick et al., 1999; Morisada et al., 2002). Analyses of these patterns have contributed to the formulation of ecological theories and to a greater understanding of biogeochemical controls of ecosystem dynamics (Hooper and Vitousek, 1997; Vitousek and Farrington, 1997; Chadwick et al., 1999). For instance, the roles of phosphorus, nitrogen and eolian nutrient inputs have been clearly documented through long-term observations of Hawaiian chronosequences (Crews et al., 1995; Chadwick et al., 1999; Neff et al., 2000).

Although they are among the first arrivals on volcanic deposits, relatively little is known about the temporal and spatial distributions of microbial colonists generally, about specific microbial functional groups or about interactions among microorganisms and animal and plant communities. Vitousek and co-workers have addressed some of these issues through their research on nitrogen fixation and nitrogen mineralization on Hawaiian volcanic deposits (Riley and Vitousek, 1995; Vitousek, 1999; Crews et al., 2001; Pearson and Vitousek, 2001). Nüsslein and Tiedje (1998), Nanba et al. (2004) and Dunfield and King (2004) have documented various aspects of microbial community structure and function on recent Hawaiian volcanic deposits (44–300 years old), and shown in particular that relatively young deposits harbored very distinct assemblages, especially where plants were present. Analyses of quinone profiles on Mt Pinatubo mudflows also suggested possible associations of Actinobacteria with developing plant communities (Ohta et al., 2003).

Other groups have reported variable relationships among plant community development, heterotrophic evenness, respiration, and C, N and P contents for successional transects on volcanic soils (Schippers et al., 2001). A survey of older soils of Mt Etna has also shown that respiration varied independently of soil carbon content, and that plant litter decomposition varied as a function of soil development and plant litter source (Hopkins et al., 2007).

King (2003a) showed that microbial uptake of atmospheric CO and hydrogen accounted for a significant fraction of respiratory metabolism on volcanic deposits ranging from about 21–210 years in age, but that these processes were relatively unimportant at a 300-year-old site supporting a mature forest. Results from Dunfield and King (2004, 2005) and Gomez-Alvarez et al. (2007) also showed that microbial community structure changed markedly along this chronosequence, with both CO oxidizers and the microbial community as a whole dominated by Proteobacteria in the forested site, while other phyla, including many unidentified lineages, dominated the remaining sites.

The observed patterns along the chronosequence may partially reflect factors such as physical weathering, accumulated inputs of microbes from the atmosphere, and various stochastic and deterministic processes independent of plant succession. They also undoubtedly reflect plant-dependent interactions. These may include species-specific interactions, such as those between actinorhizal or legume species (Ohtonen and Väre, 1998; Smolander and Kitunen, 2002; de Neergaard and Gorissen, 2004) and their symbionts, or more general interactions, such as those resulting from inputs of organic matter (Kent and Triplett, 2002). Similar conclusions have been drawn for successional patterns associated with glacial retreat (Ohtonen et al., 1999; Sigler et al., 2002; Sigler and Zeyer, 2002).

Results presented here document interactions between plant succession and CO dynamics on volcanic deposits independent of substrate age. CO uptake rates at atmospheric and elevated concentrations and comparisons of activities across gradients of plant development on volcanic deposits of a single age show that atmospheric CO is consumed most rapidly in deposits with little or no rooted vegetation. In contrast, maximum potential uptake rates, an index of CO oxidizer biomass, are greatest where plant development and organic matter are greatest. These results suggest that after initial colonization of newly formed substrates, CO oxidizer populations likely depend on organic inputs from plant communities for significant growth. Some of these CO oxidizers may contribute to increases in acetylene reduction rates (King, 2003a), an index of nitrogen fixation, that were observed with increasing root biomass.

Materials and methods

Site description

Sites were established near Kilauea caldera on a 1959 tephra deposit (Pu’u Puai—PP) and a 1969 lava flow (Mauna Ulu—MU), aspects of which have been previously described (King, 2003a; Dunfield and King, 2004; Nanba et al., 2004; Gomez-Alvarez et al., 2007). The PP site consists of islands of closed-canopy woody vegetation (mixed Metrosideros polymorpha (Ohia) and Morella faya (also Myrica faya or fire tree) interspersed with extensive unvegetated patches comprised of tephra (or volcanic cinders); the cinders were approximately 1 cm in diameter. At PP, samples were collected within tree islands (PP-canopy), at the island edges 3–5 m from canopy centers (PP-edge) and 5 m distant from islands in unvegetated tephra (PP-bare). A pronounced litter layer (approximately 5–10 cm thick) occurred within and at the edge of tree islands, while at bare sites, little or no litter was present. At MU, samples were collected from cinder deposits that had accumulated on the surface of pahoehoe lava. Vegetation at MU is more patchy and sparse than at PP. Sample sites were colonized by a pioneering fern, Sadleria cyatheoides (MU-veg) or were unvegetated (MU-unveg).

Bulk physical and chemical analyses

Aluminum core tubes (7.2-cm inner diameter) were used to collect material to 5–10 cm depths. Water contents were estimated by drying at 110 °C for >24 h. pH was determined using slurries of cinders and deionized water (1:2 mass ratio of samples and deionized water). Total organic carbon and nitrogen were analyzed using a Perkin–Elmer model 2400 elemental analyzer for ground material from the upper two centimeters. Root biomass was determined by harvesting fine roots (<2-mm diameter) from the upper 10 cm of material at each site.

Phospholipid phosphate analysis

Phospholipid phosphate contents were analyzed as a surrogate for microbial biomass (Brinch-Iversen and King, 1990). Triplicate 0.5- to 1-gram fresh weight (gfw) root-free samples from various depths were transferred to screw-cap glass tubes containing 5 ml of a 1:2 mixture of dichloromethane /methanol. The tubes were sealed, transported to the laboratory on dry ice and subsequently stored at −20 °C prior to further processing. Dichloromethane (1.67 ml) and deionized water (5 ml) were added to each tube prior to centrifugation (1000 g). Dichloromethane (1–2 ml) phases were transferred to glass ampoules and evaporated to dryness at 30 °C with nitrogen. Phospholipid phosphate was liberated by persulfate digestion and measured colorimetrically (Brinch-Iversen and King, 1990).

Gas flux analyses

Carbon monoxide and hydrogen fluxes were measured in situ using aluminum collars (about 67 cm2) fitted with a 1-l quartz chamber forming a static headspace (see King, 2003a). Gas exchange assays were initiated by sealing a chamber to a collar, covering the chamber with aluminum foil, and removing 3-cm3 headspace samples using a needle and syringe at 3- to 4-min intervals for up to 15–20 min. Samples were assayed in the field using a Trace Analytical RGD gas chromatograph (King, 1999). Starting ambient CO and hydrogen values were typically about 100 and 550 p.p.b., respectively.

Additional assays were conducted using triplicate intact cores (7.5-cm inner diameter × 20-cm length). Triplicate cores from PP and MU sites were returned to a field laboratory within a few hours after collection and incubated at ambient temperature. After sealing the core tubes, core headspaces were subsampled for CO and hydrogen analysis as above. Analyses of CO2 production were performed using headspace samples that were injected into an SRI flame ionization gas chromatograph equipped with a stainless steel column (3.2-mm outer diameter × 2 m) containing silica gel and a methanizer to convert CO2 into methane. Samples for CO2 production were collected at intervals over a 48-h period during incubations at ambient temperature. Instrument responses to CO2 and methane were determined using standards prepared from pure gases. Responses to CO and hydrogen were determined using standards at near-atmospheric levels or greater; they were prepared by diluting pure gases in CO- and hydrogen-free air.

Maximum potential CO and hydrogen uptake rates were estimated using fresh material from the upper 2-cm depth interval of each PP and MU site. Triplicate 5-gfw samples were transferred to 110-cm3 jars that were sealed with neoprene stoppers and incubated in darkness at ambient temperature. After adding CO or hydrogen to a final concentration of about 100 parts per million (p.p.m.; 1 p.p.m. is equivalent to about 0.1 kPa), jar headspaces were subsampled at intervals to determine uptake rates. Maximum potential CO uptake rates were also estimated for selected depth intervals of PP sites. Similar assays were conducted with freshly collected, washed, soil-free fine (<2 mm) roots from M. polymorpha, M. faya and S. cyatheoides; triplicate samples of approximately 1 gfw of each root type were incubated in darkness in sealed jars with an ambient atmosphere (0.1 p.p.m. CO) and with 100 p.p.m. CO as described by King and Crosby (2002).

Acetylene reduction rates

Triplicate intact cores (about 10 cm deep) from each of the PP and MU sites were collected using aluminum core tubes (7.2-cm inner diameter). Acetylene was added to the core headspaces to a final concentration of about 5%. Sealed tubes were incubated at ambient temperature for about 72 h. Headspace samples were obtained at intervals and transferred to evacuated 3-cm3 gas collection tubes containing 0.5 ml of ammoniacal silver nitrate to precipitate acetylene. Ethylene (C2H4) concentrations in subsamples were analyzed by flame ionization gas chromatography using a Shimadzu GC-14A. Rates of C2H4 production provided a surrogate estimate of N2-fixation rates.

Results

Bulk deposit properties

Root biomass varied dramatically among sites (Table 1) with a maximum in vegetated areas of 652±165 gram dry weight (gdw) per m2 for PP-canopy and minimum of 50±4 gdw m−2 for MU-veg. Unvegetated PP and MU sites contained negligible identifiable root mass. Total carbon and nitrogen contents paralleled root distributions with values up to 45.7 and 1.7%, respectively, in PP-canopy. C/N ratios were least in PP-bare (15.7), increasing by almost twofold in PP-canopy (27.2). In contrast, pH values decreased substantially from 6.9 in PP-bare to 3.9 in PP-canopy. Water contents displayed the opposite trend, with PP-bare material holding very little water, while PP-edge and PP-canopy were relatively moist by comparison. Trends between MU-unveg and MU-veg were comparable (Table 1).

Table 1 Selected properties of Pu’u Puai and Mauna Ulu volcanic deposits
Full size table

Phospholipid phosphate content

Phospholipid phosphate concentrations varied substantially among sites and with deposit depth (Figure 1). PP-canopy surface material contained phospholipid concentrations more than an order of magnitude higher than surface material from any other site (P<0.001, analysis of variance). Phospholipid phosphate concentrations were similar, however, for PP-edge and PP-bare sites at all depths. At all PP sites, phospholipid phosphate concentrations decreased by an order of magnitude with increasing depth up to 10 cm; these changes were statistically significant (P<0.01, analysis of variance). At MU-veg, phospholipid phosphate concentrations were an order of magnitude greater than at MU-unveg (P<0.01), and about 20-fold lower than at corresponding PP sites.

Figure 1
figure 1

(a) Depth profiles of phospholipid phosphate concentration (closed symbols) and maximum potential CO uptake rates (open symbols) for Pu’u Puai (PP)-bare (triangles), -edge (squares), and -canopy (circles) sites. All data are means of triplicate determinations ±1 s.e. (b) Correlation between maximum potential CO uptake rates and phospholipid phosphate concentrations for all sites, excluding the 0- to 2-cm interval of PP-edge (square).

Full size image

Gas exchange

Volcanic deposits at three of five sites consistently consumed atmospheric CO: PP-bare, MU-veg and MU-unveg (Table 2). CO was emitted consistently from PP-canopy, but results were more variable for PP-edge, with both uptake and emission observed (Table 2). Atmospheric CO uptake rates ranged between about 1.2 and 8.1 mg CO m−2 day−1, with no clear trends among sites. CO emission rates ranged from about 1.3 to 8.1 mg CO m−2 day−1 (Table 2) for PP-canopy, with intermediate PP-edge values. In situ CO exchange rates were similar to those for values obtained from ex situ incubation of intact cores (Table 2). Estimates for steady-state CO exchange concentrations (concentrations at which CO uptake equals CO production) ranged between 0 and 21 p.p.b. (parts per billion) for all but the PP-canopy site (Table 2). For the latter, the steady-state concentration was 207±43 p.p.b., well above the typical ambient atmospheric concentrations (75–125 p.p.b.). CO uptake rate constants were lowest in PP-canopy and MU-unveg (0.1 min−1) and significantly higher for the remaining sites (Table 2).

Table 2 Gas exchange rates for Pu’u Puai and Mauna Ulu sites
Full size table

For the MU sites, maximum potential CO uptake rates (0- to 2-cm depth interval) were significantly greater (P<0.01) at MU-veg than MU-unveg. The MU values, however, were substantially less than values for comparable PP sites (Table 2). At PP-canopy and PP-edge sites, maximum potential CO uptake rates decreased significantly (analysis of variance, P<0.01) with deposit depth (Figure 1a). For the 0- to 2-cm and 2- to 5-cm depth intervals, PP-canopy and PP-edge rates were similar, but PP-canopy rates were significantly greater for the 5- to 10-cm interval. In contrast, maximum potential CO uptake rates were least for PP-bare, and did not decline significantly with depth. With the exception of a single apparent outlier for the 0- to 2-cm depth interval of PP-edge, maximum potential CO uptake rates were strongly correlated (r=0.99) with phospholipid phosphate concentrations (Figure 1b); a strong correlation (r=0.74) remains even if the apparent outlying point is included.

All sites consumed atmospheric hydrogen at rates generally greater than those for CO (Table 2). Trends in uptake versus vegetational development among sites were not statistically significant (P>0.5, analysis of variance), although activity at PP-canopy was least with the highest threshold value (Table 2). Hydrogen was also consumed during ex situ assays to determine maximum potential uptake rates. Maximum potential hydrogen uptake rates were greater than the corresponding maximum potential CO uptake rates at all sites, and increased from MU-unveg to PP-canopy. A plot of paired maximum potential hydrogen uptake and maximum potential CO uptake rates indicated that the capacity for hydrogen uptake rose exponentially to an asymptotic maximum as a function of increased CO uptake capacity (Figure 2).

Figure 2
figure 2

Maximum potential hydrogen uptake versus maximum potential CO uptake rates for the 0- to 2-cm depth interval of all sites. All data are means of triplicates ±1 s.e. Curve represents a nonlinear best fit to an exponential rise to an asymptote (r2=0.98).

Full size image

Respiration rates, measured as CO2 production, increased with increasing vegetational development, and ranged from 3.2±0.6 to 103.7±22.8 and from 2.2±0.5 to 74.2±26.5 mmol m−2 day−1 for PP and MU sites, respectively (Table 2). Maximum potential CO uptake rates for the 0- to 2-cm interval of all sites were positively and linearly correlated with respiration rates (r=0.71); maximum CO uptake potentials integrated over 0- to 10-cm depth for the PP sites were highly linearly correlated with respiration rates (r=0.99; Figure 3). The relative contribution of atmospheric CO uptake to respiratory reducing equivalent flow for PP and MU was estimated by assuming a 1:1 stoichiometry for CO2 production and oxygen uptake, and two and four reducing equivalents formed per mole of CO oxidized and CO2 produced, respectively. The percent of total respiratory reducing equivalent flow due to CO was calculated as:

Since both PP-canopy and PP-edge sites emitted CO during the time the respiration assay was conducted, the relative contribution of CO to respiratory reducing equivalent flow was zero; for PP-bare, a value of 9.1% was obtained. For MU sites, the values ranged from negligible at MU-veg (0.2%) to 6.4% at MU-unveg (Table 2).

Figure 3
figure 3

Correlation between maximum potential CO uptake rates and root biomass. All data are means of triplicates ±1 s.e. Curve represents a linear fit (r=0.99).

Full size image

Acetylene reduction

Acetylene reduction rates were least in MU-unveg and PP-bare sites (0.4–0.8 μmol C2H2 reduced per m2 per day) and reached a maximum for PP-canopy samples (21.5 μmol C2H2 reduced per m2 per day; Table 2). Acetylene reduction rates were positively and linearly correlated with root biomass across all sites (r=0.98), and positively and exponentially correlated with respiration rates (r=0.95). In addition, acetylene reduction rates increased exponentially with increasing maximum CO uptake potential rates integrated over 10 cm depth for the PP sites (r=0.99); comparable depth profiles were not available for the MU sites.

Root CO production and oxidation

Live, fine roots from the fern, S. cyatheoides, and the trees, M. polymorpha and M. faya, each produced CO when incubated under ambient conditions. CO production rates for S. cyatheoides roots obtained from two distinct sites (MU and a 1982 lava flow; 1.9–3.3 nmol gfw−1 h−1, respectively) were comparable to rates for M. faya, 3.5±0.3 nmol gfw−1 h−1 and M. polymorpha, 1.2±0.1 nmol gfw−1 h−1 (Table 3). Maximum potential CO uptake rates were considerably higher than CO production rates, and varied significantly among the three taxa. Essentially, identical uptake rates were observed for the two sets of S. cyatheoides roots, 12.8±1.6 nmol CO oxidized per gfw per h. Uptake rates for M. polymorpha (9.6±0.6 nmol CO oxidized per gfw per h) were somewhat lower, while rates for M. faya (24.4±1.6 nmol CO oxidized per gfw per h) were somewhat higher. In general, there was a positive correlation between maximum potential root CO oxidation rates and root CO production rates (r=0.9; not shown).

Table 3 CO production under ambient conditions and CO uptake during incubation with 100 p.p.m. CO by washed, soil-free fine roots of Sadleria cyatheoides, Metrosideros polymorpha and Morella faya
Full size table

Discussion

Analyses of receding glaciers and chronosequences of volcanic systems have revealed several aspects of microbial community responses to ecosystem succession, including temporal and spatial changes in functional diversity, phylogenetic diversity (evenness and richness) and cell-specific activities. For example, Sigler et al. (2002) observed increases in cell number and total activity with successional age, but maximal activity per cell and greater relative culturability occurred at intermediate stages, and phylotype richness decreased with age. Their results confirmed an ‘r-‘ to ‘K-strategy’ shift during succession, but indicated that successional changes were dynamic and occurring at both individual and community levels. Others have documented interactions between microbes and vegetation, including changes in carbon utilization and community structure (for example, Ohtonen et al., 1999; Schipper et al., 2001).

Previous analyses of Kilauea volcano based on multiple sites that constituted a chronosequence from 21- to 300-year old showed that atmospheric CO and hydrogen utilization occurred early in successional development and contributed significantly to respiratory metabolism (King, 2003a). Molecular ecological analyses also revealed significant differentiation among CO-oxidizing and lithotrophic communities colonizing deposits of different ages and successional development (Dunfield and King, 2004, 2005; Nanba et al., 2004). In this study, we report interactions between successional development, specific functional processes (CO and hydrogen oxidation and acetylene reduction) and related microbiological and biogeochemical variables for Kilauea volcano deposits with uniform ages and climatic regimes.

Data from both PP and MU sites show that atmospheric CO uptake contributes significantly (6–10%) to total respiratory activity (measured as CO2 production) in unvegetated areas characterized by very low respiration rates and organic carbon concentrations (Tables 1 and 2). Atmospheric CO uptake rates at PP and MU unvegetated sites are similar to previously reported estimates (King, 2003a); CO2 production rates fall within ranges reported for other organic carbon poor systems (Hopkins et al., 2007). These observations support a role for CO as an energy source in carbon-limited systems, even when it is available at only very low concentrations (75–125 p.p.b.).

In contrast, atmospheric CO contributes little or nothing to respiratory activity at PP-canopy and -edge and MU-veg sites (Table 2). Even accounting for the possibility that root respiration might contribute a significant fraction of total respiration at these sites (up to about 50%; Andrews et al., 1999; Wang et al., 2005), CO oxidation remains a minor fraction of bacterial activity. Instead, at these sites rhizodeposition and leaf litter no doubt represent the primary sources of organic matter for respiration. These carbon sources support respiration rates for the PP and MU at levels similar to those reported for other vegetated systems with moderate-to-high organic carbon concentrations (Campbell and Law, 2005; Vogel et al., 2005; Bernhardt et al., 2006).

All of the sites assayed consume atmospheric hydrogen, but there is little consistent variation among them (Table 2). Relative to carbon-based respiratory metabolism, atmospheric hydrogen uptake supports considerable activity (Table 2). At sites MU-unveg and PP-bare, where organic carbon is limited, hydrogen uptake could contribute to biosynthesis, but CO2 fixation efficiencies are relatively low with hydrogen lithotrophy (Bowien and Schlegel, 1981), and some hydrogen uptake appears due to exoenzymatic activity (King, 2003a).

Although atmospheric CO contributes significantly to bacterial metabolism at unvegetated sites, CO-oxidizing bacteria were active at all sites. This is evident from time courses of headspace CO concentrations during both in situ and ex situ analyses (not shown). As in previous studies (King, 1999), a model for CO change based on a zero-order production term and a first-order uptake term explained patterns of CO uptake and emission. In addition, cinders from all PP and MU sites consume exogenous CO added at elevated concentrations without a lag. Maximum potential CO uptake rates based on these assays correlate linearly with estimates of microbial biomass (Figure 1b), and depth integrated maximum potential rates for PP correlate linearly with respiration rates (Figure 3). To the extent that maximum potential CO uptake rates reflect CO oxidizer biomass, these relationships indicate that CO oxidizer biomass accounts for an approximately constant fraction of total biomass, and that maximum potential activity likely responds to the same variables that determine respiratory activity.

In particular, it appears that CO oxidizer biomass may become initially established on barren deposits through inputs of atmospheric CO and exogenous organic matter (for example, in precipitation). Subsequently, CO oxidizer biomass expands with endogenous organic matter production, especially that from rooted vegetation. This is consistent with both the preference of CO-oxidizing bacteria for organic substrates, and with their expression of CO oxidation when substrate availability is low, as is typically the case in soil systems even when bulk organic concentrations are high (Morita, 1997).

The distribution of maximum potential hydrogen uptake rates among sites parallels that for maximum potential CO uptake rates (Table 2), suggesting some common controls for the two activities. This could be due in part to the fact that some CO oxidizers also oxidize hydrogen (King and Weber, 2007). In addition, many hydrogen oxidizers, like virtually all CO oxidizers, are facultative lithotrophs that grow preferentially with organic substrates (Bowien and Schlegel, 1981). However, when plotted as paired data, maximum potential hydrogen uptake rates increase exponentially to an asymptote as a function of maximum potential CO uptake rates (Figure 3). This may reflect a shift from populations that oxidize both CO and hydrogen to populations that oxidize primarily CO as maximum potential CO uptake rates increase.

The fact that CO oxidizers remain active across a gradient of increasing ecosystem development, even though atmospheric CO decreases in significance as a substrate, indicates that they may contribute to a variety of functions other than trace gas dynamics. Many CO-oxidizing isolates contain nitrogenase (King, 2003b). Some, such as CO-oxidizing rhizobia (for example, Bradyrhizobium japonicum USDA 6) contribute to symbiotic nitrogen fixation; others, such as CO-oxidizing Burkholderia (for example, B. xenovorans LB400), can contribute to both asymbiotic and plant-associated nitrogen fixation (Chen et al., 2003; Choudhury and Kennedy, 2004).

An exponential correlation between depth-integrated maximum potential CO uptake and a surrogate measure of nitrogen fixation, acetylene reduction, for the PP sites indicates that controls of activity differ for the two processes, irrespective of the extent of contributions by CO oxidizers. Differential controls for expression of the aerobic carbon monoxide dehydrogenase and nitrogenase systems likely include nitrogen and organic matter availability, associative interactions with plant roots and responses to physical–chemical variables, for example, pH and water potential. In addition, populations of strictly heterotrophic nitrogen fixers may proliferate to a much greater extent than nitrogen-fixing CO oxidizers in material at the canopy site, where acetylene reduction rates appear disproportionately high. Of course, variations in acetylene reduction rates among sites may only partially reflect patterns in nitrogen fixation, since ratios of acetylene reduced to nitrogen fixed may also vary among sites.

Plant roots also affect the distribution and activity of CO oxidizers. Maximum potential CO uptake rates for MU and PP sites rise exponentially with root mass to an asymptote in PP-edge and PP-canopy sites (Figure 4). This trend likely reflects an increase in organic matter availability that can support CO oxidizer biomass, but may also reflect an increase in CO availability. In particular, roots of all three plant species examined in this study emit CO at rates comparable to rates previously reported for other plants (Table 3; King and Crosby, 2002). As a result, increases in root biomass across sites represent a gradient of increasing CO within the deposits, which may promote CO oxidizer biomass or activity. In addition, the ability of roots of all three taxa to consume CO (Table 3) indicates that they may support CO oxidizer populations that can affect bulk and rhizosphere activities as well as exchanges with the atmosphere (King and Crosby, 2002; King and Hungria, 2002). Interactions between plants and the diversity of CO-oxidizing populations are being assessed at present.

Figure 4
figure 4

Maximum potential CO uptake rates integrated over 0- to 10-cm depth intervals for PP-bare, -edge and -canopy sites versus respiration rates. All data are means of triplicates ±1 s.e. Curve represents a linear least-squares best fit (r2=0.99).

Full size image

In conclusion, atmospheric CO supports significant metabolic activity on volcanic deposits lacking vascular plant growth. Relationships between maximum potential CO uptake, root mass, organic carbon and estimates of microbial biomass indicate that subsequent to colonization on barren material, CO-oxidizing populations expand in concert with the microbial community as a whole, even though the role of CO in metabolism overall diminishes with increasing plant development. CO oxidizers remain an active component of the microbial community throughout succession and may contribute to nitrogen fixation, rates of which increase through successional development.