There are few data on the effect of long-term manipulations on soil protozoa, and almost nothing is known about soil protozoa in Alaska. I studied the response of testate amoebae to nitrogen and phosphorus addition in an Arctic fen, at Toolik Lake Long-Term Experimental Research (LTER) Station, Alaska. Testate amoebae were extracted from Sphagnum mosses in control and fertilized plots. Of the 35 testate amoebae taxa recorded, 7 are first observations for the Arctic (excluding Russia) and 14 for Alaska. The total density and biomass of testate amoebae were significantly reduced, by 77% and 84%, in the fertilized plots. The structure of testate amoebae communities was also modified in those plots, although for most taxa the changes were not significant. Four taxa (Amphitrema flavum, Assulina muscorum, Placocista spinosa ssp. hyalina, and Hyalosphenia papilio) accounted for over half of the population in the control plots but only for 11% in the N and P plots. The densities of A. muscorum and Difflugia oviformis were significantly lower in the N and P–treated plots. The relative abundance and contribution to biomass of Centropyxis aerophila, Phryganella acropodia, and Tracheleuglypha dentata increased in the fertilized plots, while that of D. oviformis decreased. These effects suggest that testate amoebae respond to nutrient manipulations in the Arctic.
Introduction
The use of microorganisms, and testate amoebae in particular, as bioindicators is often promoted because they have short generation times, high densities, and diversity, and they are likely to respond to changes more rapidly than other components of the ecosystem. The response of soil protozoa communities to environmental gradients has been established in a broad range of ecosystems, from mineral and organic soils to aquatic systems. These relationships have also been used to infer past environmental conditions based on testate amoebae tests (shells) preserved in lake and peat sediments (Charman, 2001).
Soil protozoa, and protists in general, have also been used in a number of studies to assess the biotic effect of a range of environmental perturbation on different ecosystems. Heavy metals (Kandeler et al., 1992) and air pollution (Balik, 1991) had negative effects on testate amoebae abundance and diversity. Several species of soil algae were able to grow in highly acidic soil polluted by heavy metals, but diatoms were absent from the most polluted sites (Shubert et al., 2001). The addition of organically enriched magnesite fertilizers reduced the abundance but not the diversity of soil ciliates (Aescht and Foissner, 1993). The abundance or biomass of testate amoebae was increased by solar ultraviolet-B radiation (Searles et al., 1999), elevated CO2 (Lussenhop et al., 1998), the addition of nutrients to mineral soils as pig slurry (Griffiths et al., 1998), L-lysine (Hodge et al., 1999), and nitrogen (Maraun et al., 2001). However, elevated atmospheric CO2 also had contrasting effects on soil protozoa in mineral soils. The density of flagellates tended to increase, whereas the density of amoebae significantly declined (Treonis and Lussenhop, 1997). In a Sphagnum peatland, N addition increased the relative importance of cyanobacteria, euglenophyceae, diatoms, and ciliates and decreased the relative importance of heterotrophic bacteria, other microalgae, and testate amoebae (Gilbert et al., 1998b). In the same site, the supply of nutrients (PKCa and NPKCa) resulted in increases of the relative biomasses of heterotrophic bacteria, diatoms, and ciliates and a decrease in the relative proportion of testate amoeba and other micro-algae (Gilbert et al., 1998a). Soil acidification due to atmospheric pollution did not affect protozoan biomass (Couteaux et al., 1998).
Taken together, these studies suggest that protozoa, or protists in general, have potential as biomonitors in most cases. But this approach is currently limited by (1) taxonomic uncertainties (Griffiths et al., 2001), (2) an understanding of the importance of phenotypic plasticity in response to environmental conditions (Wanner, 1999), (3) the debate over global versus local distribution of protozoa (Wilkinson, 1994; Finlay et al., 2001; Wilkinson, 2001), and (4) the scarcity of experimental studies that include protists. Most of the existing studies using protists as bioindicators focused either on existing gradients of environmental perturbations such as long-term pollution in highly industrialized regions, or on short-term experiments, often under artificial conditions. Long-term manipulative experiments in natural ecosystems are comparatively rare in ecology, and even more so in the Arctic. Furthermore, the soil biota is usually less intensively studied than the aboveground component of ecosystems.
There are almost no data on the effect of long-term controlled manipulations on soil protozoa, and almost nothing is known about testate amoebae in Alaska. The objectives of this study were therefore (1) to assess how an experimental manipulation of N and P might affect the abundance and diversity of testate amoebae in the Arctic, and (2) to initiate a list of testate amoebae from Alaska. In view of the marked effect of N and P on the vegetation and carbon dynamics (Shaver et al., 1998) a clear response of testate amoebae was hypothesized.
Material and Methods
STUDY SITE AND FERTILIZATION EXPERIMENT
Toolik field station, the Arctic Long Term Experimental Research (LTER) site, is located in the northern foothills of the Brooks Range, Alaska (68°38′N, 149°43′W, elevation 760 m a.s.l.). This area has continuous permafrost and no trees. The dominant vegetation type is tussock tundra vegetation of sedges and grasses mixed with dwarf birch (Betula nana, B. glandulosa) and low willows (Salix spp.). The climate at the site is typical of arctic regions, with a mean annual air temperature of about −10°C and low precipitation (45% of the 20–40 cm of precipitation falls as snow). During the summer the daily average air temperature is 7–12°C with the sun continuously above the horizon from mid-May to late July. The snow-free season lasts from late May to mid-September, with below-freezing temperatures possible at any time.
Several long-term fertilization experiments are being carried out at the Toolik LTER site. This study focuses on a nitrogen and phosphorus enrichment experiment that was initiated in 1989 in two wet sedge tundra sites (wet, nutrient poor, fen community) 2 km apart. One site was located near the main inlet to Toolik Lake, and the other was located near the outlet of the lake. The sites were almost flat, with up to 5 cm of standing water and occasional moss hummocks. The vegetation at the experimental sites was dominated by sedges, Eriophorum angustifolium, Carex rotundata, and C. cordorrhiza, with scattered hummocks of bryophytes, including Sphagnum. Nitrogen was applied every year as granular ammonium nitrate at a rate of 10 gN m−2 yr−1. Phosphorus was applied the first year as granular superphosphate at a rate of 10 gP m−2 yr−1 and in subsequent years at a rate of 5 gP m−2 yr−1 (Shaver et al., 1998).
SAMPLING AND ANALYSIS
Five to 10 Sphagnum samples (top 5 cm) were taken on 28 July 2000 in 3 control and 3 experimental plots. For each treatment 2 plots were located at the outlet site and 1 at the inlet site. In one of the fertilized plots of the outlet site duplicate samples were taken, and the average of these 2 samples was used in the data analyses for that plot. The moss samples were first dried at 65°C and then weighed. Testate amoebae were extracted from the mosses using a sieving and back-sieving method that retained all particles with a size between 10 μm and 300μm (Hendon and Charman, 1997). Exotic Lycopodium spores tablets were added to estimate the density of testate amoebae (Stockmarr, 1971). Wet mounts were analyzed under the microscope at 200× and 400× magnifications. Testate amoebae were identified and counted with a goal of reaching 150 individuals. One rotifer species, Habrotrochoa angusticollis, which is frequently found in Sphagnum, was also observed and counted. The results for this species are presented, but it was not included in the numerical analyses of testate amoebae communities. Biovolumes of testate amoebae and the rotifer Habrotrochoa angusticollis were estimated by assuming geometrical shapes using the following formulae where L is the length (in most cases the longest dimension from the aperture to the tip of the shell), w is the width (the longest dimension perpendicular to L), h is the height (perpendicular to both L and w), r is the radius, and d is the diameter in the cases where w would be equal to h:
Biovolumes were converted to carbon using the following conversion factors: testate amoebae, 1μm3 = 1.1 × 10−7 μgC (Weisse et al., 1990); rotifers 1μm3 = 1.25 × 10−7 μgC (Gilbert et al., 1998a). The effect of the treatment on testate amoebae densities, diversity, biomass carbon, relative abundance (% of total testate amoebae count), and relative contribution to biomass (% of total testate amoebae estimated biomass) were assessed using t-tests. The data was transformed to homogenize variances using the logarithm [x′ = ln(x + 1)], or the square root when needed. If neither of these transformations homogenized the variance, the P-values for nonhomogeneous variances of the t-test on untransformed data were used.
Results
ABUNDANCE, DIVERSITY, AND BIOMASS
A total of 35 testate amoebae taxa were recorded in the 7 samples. The species richness of individual samples varied between 13 and 19 species (Table 1). The density of testate amoebae species varied between 0 and 34,654 individuals per g dry weight of Sphagnum, and the estimated biomass of species varied between 0 and 399 μgC g−1. Total testate amoebae densities varied among the samples from 5723 to 103,403 individuals g−1. Total estimated biomass varied from 45 to 1345 μgC g−1. The density of the rotifer Habrotrochoa angusticollis varied between 0 and 2728 individuals g−1, and its biomass varied between 0 and 103.7 μgC g−1.
DIVERSITY, DENSITY, AND COMMUNITY STRUCTURE IN THE CONTROL AND FERTILIZED PLOTS
After 12 yr of N and P addition, the total density of testate amoebae was 77% lower in the fertilized plots (74,530 individuals g−1 in the control plots and 16,838 individuals g−1 in the N and P plots; P = 0.052). The biomass carbon was 84% lower in the fertilized plots (988 μgC g−1 in the control plots and 159 μgC g−1 in the N and P plots; P = 0.036). However, no significant effect was observed for the diversity, as measured with the Shannon's entropy test (data not presented), or the species richness (Table 1).
The density and biomass C of almost all testate amoebae taxa differed between the control and fertilized plots, but in most cases these trends were not significant owing to high variability (Table 1). Although the percentage treatment effect was very important for many taxa, some of which were absent from either the control or the fertilized plots, it was statistically significant only for 2 species, Assulina muscorum and Difflugia oviformis. The density of both species was lower in the N and P–fertilized plots (−85%, P = 0.024, and −99%, P = 0.008).
The structure of testate amoebae communities, measured as the species' relative frequencies and relative contribution to the total testate amoebae biomass C, was also different between the control and fertilized plots, but as for density and biomass, in most cases these trends were not significant owing to high variability (Table 2). In the control plots, 4 taxa (Amphitrema flavum, Assulina muscorum, Placocista spinosa hyalina, and Hyalosphenia papilio) accounted for over half (51%) of the population. These same 4 taxa accounted for only 11.1% of the population in the N and P–fertilized plots. Two species (Centropyxis aerophila and Phryganella acropodia) accounted for nearly half (47%) of the population in the fertilized plots but only 18% of the population in the control plots. However, the percentage treatment effect on relative frequencies was different than the effect on densities. As the total density was significantly reduced in the fertilized plots, the relative frequency of several species that did not show any response in terms of density was clearly affected, and in some cases this effect was significant. The relative proportion of Arcella discoides, Centropyxis aerophila, Phryganella acropodia, and Tracheleuglypha dentata more than doubled, and their relative contribution to the total testate amoebae biomass C increased by 6- to 14-fold. This effect was significant or marginally significant for the last 3 species. Among the many taxa that decreased in relative frequency and relative contribution to the total testate amoebae biomass C in the fertilized plots, the treatment effect was significant only for one, Difflugia oviformis (−97%, P = 0.015).
Discussion
In this study I aimed at assessing the long-term (12 yr) effects of N and P addition on the testate amoebae of an arctic fen. To my knowledge this is only the second published data set on testate amoebae from the Alaskan Arctic (Beyens and Chardez, 1995) and the first report of the response of testate amoebae to experimental environmental manipulation in the Arctic.
A total of 35 taxa were recorded in the control and fertilized wet sedge tundra plots of Toolik Lake LTER Station. To my knowledge, 7 of these had previously never been recorded in the Arctic (between 27°E and 168°W), and 7 more (14 in total) had not been recorded in Alaska before this date (Beyens and Chardez, 1995; Van Kerckvoorde et al., 2000) (Table 1). These first observations for the arctic are: Amphitrema flavum, Amphitrema wrightianum, Difflugia oviformis, Nebela griseola, Nebela tincta var. major, Placocista spinosa ssp. hyalina, and Quadrulella symmetrica. In addition, the following 7 taxa are first observations for Alaska: Centropyxis aerophila var. sphagnicola, Cyclopyxis kahli, Difflugia globulosa, Euglypha ciliata, Heleopera rosea, Phryganella paradoxa, and Tracheleuglypha dentata. These findings indicate only that very few studies of testate amoebae have been done in Alaska and only a few more in other parts of the Arctic. In addition, some of the taxa listed here, especially A. flavum and A. wrightianum, are found almost exclusively on Sphagnum mosses, which have not been extensively sampled in the published arctic studies. Clearly, many more data are needed before any conclusions can be drawn regarding possible broad-scale distribution patterns of testate amoebae in Alaska or the Arctic.
The experimental addition of N and P to wet sedge tundra had profound effects on the vegetation and ecosystem functioning after 6 yr (Shaver et al., 1998). The total plant biomass increased nearly 3-fold. This effect was due mostly to the increase in vascular plants, although a significant increase in moss biomass was observed in the inlet site. The vegetation composition was also affected, with an increase of Carex cordorrhiza, a decrease of Carex rotundata, and a disappearance of Trichophorum caespitosum. The mass of N and P in the living plant biomass strongly increased with N and P addition. The gross ecosystem production, net ecosystem production, and ecosystem respiration all increased with N and P addition.
The testate amoebae results suggest an apparent response to the experimental addition of N and P, which was significant for 2 species and for the total testate amoebae density and biomass. The strong significant reduction in total testate amoebae density (−77%) and biomass C (−84%) may be due to a deterioration of living condition or availability of prey species for testate amoebae. Alternatively, it may partly be an artifact related to the treatment effect on the bryophytes communities. I aimed at collecting Sphagnum mosses only, but in 2 of the fertilized plots the samples contained either a mix of Sphagnum and other mosses or only non-Sphagnum mosses. The bulk density of Sphagnum mosses is much lower and the water-holding capacity is higher than that of other mosses. Thus, differences in bulk densities may at least in part explain the effect on total density and biomass.
The experimental setup did not allow for more replicate samples. Therefore, the data set was limited, and the clear trends observed for many taxa did not emerge as statistically significant. For example, density of Phryganella acropodia nearly tripled with N and P addition, and several species (Assulina seminulum, Centropyxis aculeata, Hyalosphenia elegans, and Nebela griseola) were absent from the fertilized plots. Other taxa showed a clear reduced density in the fertilized plots. Furthermore, if a conservative significance threshold were applied, such as a Bonferroni correction (0.05 per number of taxa), none of the taxa's responses would be significant. This remark does not apply to the tests on the oval abundance and biomass, as these can be considered as single tests. Conversely, several taxa were present only in the fertilized plots. With small data sets such as this one, the likelihood of a type 2 statistical error is high. Therefore, the statistical significance of the results presented here (or lack thereof) should not be overemphasized.
From a bioindication standpoint, the response of several species is in agreement with their known ecology. Two examples may be highlighted, although neither was statistically significant. (1) Phryganella acropodia is not characteristic for Sphagnum but is also found in forest litter and bryophytes other than Sphagnum. The higher density of this species in the fertilized plots suggests that conditions in Sphagnum fens are not optimal for this species and that the perturbation caused by the fertilization made the environment more favorable for this species. (2) By contrast, A. flavum is often cited as a good indicator species for bog pools (wet depressions in ombrotrophic Sphagnum peatlands). The dramatic decrease of this species in the fertilized plots could be due to physical or chemical changes in the microhabitat colonized by amoebae, or by changes in food sources. Given the sensitivity of testate amoebae to microenvironmental conditions at a very fine scale (Mitchell et al., 2000a) and the existence of vertical trophic gradients within the surface of Sphagnum peatlands (Mitchell et al., 2000b), a Sphagnum hummock in a relatively nutrient-rich peatland (the site was a minerotrophic fen as opposed to an ombrotrophic bog) may represent an environment similar to a bog pool.
At this site the vegetation also changed as a result of the treatment, with a clear reduction in Sphagnum cover. Therefore, an indirect effect of N and P fertilization on testate amoebae through shifts in bryophyte community cannot be excluded and is indeed most likely given the sensitivity of these organisms to microenvironmental gradients. In this study the effect of N and P on the ecosystem was clear before looking at testate amoebae, and their hypothesized response was therefore highly likely. But the identity of the species and the magnitude of their response were unknown. The clear effect of N and P on total testate amoebae density and biomass, the significant effect on 2 species for density and C biomass and 4 species for relative frequency and relative contribution to total testate amoebae biomass, and the clear but nonsignificant response of most other species suggest that testate amoebae are strongly responsive to nutrient manipulations in the Arctic.
Acknowledgments
I wish to thank Dr. M. Sommerkorn (The Ecosystems Center, MBL, Woods Hole, MA, U.S.A., now at MLURI, Aberdeen, U.K.) for assistance during sampling, information on the experiment, and critical comments on the manuscript; Dr. G. Shaver (The Ecosystems Center, MBL, Woods Hole, MA, U.S.A.) for allowing me to visit the Toolik LTER station and take the samples; and Dr. F. S. Chapin III and Dr. J. Braddock (UAF, AK, U.S.A.) for access to facilities at the University of Alaska Fairbanks during my postdoctoral stay and for critical comments on the manuscript. Financial support for this study came from the Swiss National Science Foundation (Postdoctoral Fellowship #81NE-62574). The Arctic LTER project (NSF grant #9810222) is also supported by several other grants. I also thank Dr. Ralph Meisterfeld (Institute for Biology II, Aachen, Germany) for taxonomic clarifications and Dr. David Wilkinson and Dr. Louis Beyens for reviewing the manuscript.
References Cited
Table 1
Density and biomass carbon of testate amoebae species in Sphagnum samples from control and N- and P-fertilized wet sedge tundra plots in Toolik LTER, Alaska
Table 2
Relative abundance and relative contribution to the total testate amoebae biomass C of testate amoebae species in Sphagnum samples from control and N- and P-fertilized wet sedge tundra plots in Toolik LTER, Alaska
