STUDY AREA

King George Island is located in the climatic zone of maritime Antarctica. The annual mean temperature is −1.7°C (1977–1996), but from December to April the range of the monthly mean is from 0.9 to 2.3°C, with maximum daily temperatures of 16.7°C in January and minimum daily temperatures of −32.3°C in July (Rakusa-Suszczewski et al., 1993; Kejna, 1999). Yearly precipitation is 510 mm with a nearly homogeneous distribution over the year. In March there is an extreme minimum due to the decreasing temperature at the onset of the antarctic winter. The recent vegetation in the ice-free oasis of King George Island is characterized by mosses, some liver mosses, lichens, and algae (Olech, 1993; Zarzycki, 1993) as well as by the higher plants Deschampsia antarctica and Colobanthus quitensis. Huge areas of landscape are nonvegetated, and the occurrence of vegetation shows a high variation and an extreme patchiness.

The investigation site is located near the permanently occupied Polish Station H. Arctowski at King George Island (58°20′E, 62°10′S), located at the shore of Admiralty Bay (Fig. 1).

The parent material of soil is mainly neoglacial moraine rubble (centuries old and younger) and fluvioglacial sands influenced by eolian deposits and volcanic ash (Blume et al., 2002). Relic penguin rookeries are scattered widely within the ice-free areas (Tatur and Myrcha, 1984; Myrcha and Tatur, 1991), and active colonies are located next to the coastline. Most soils are affected by permafrost between 50 and 200 cm depth (Kuhn, 1997).

TRANSECTS

Sampling transects are described in Table 1. Soil samples were taken along two catenary soil transects perpendicular to the nearby Ecology Glacier (Kuhn, 1997: Fig. 1). They were established for botanical studies (Olech, 2002). Transect A is 10 to 20 m inland from the shore, whereas transect B is located several hundred meters farther inland. The direction of transect A is south-north, starting with site A0, about 2 m away from the glacier, and ending at site A15, 306 m north of the glacier. The direction of transect B is south-southeast to north-northwest, starting with site B1, 16 m from the glacier, and ending at site B12, 237 m from the glacier. The vegetation has been characterized along both transects (Olech, 2002). The sampling sites are located in valleys or depressions between the moraines as well as on feet, slopes, or tops of moraines. On transect A, the first visible vegetation—the grass Deschampsia antarctica and the moss Polytrichum piliferum along with certain other mosses—occurred nearly 100 m away from the glacier, although the surface coverage was still extremely low (Table 1). The lichen Usnea antarctica is found 200 m from the glacier, whereas the alga Prasiola crispa is restricted to sites A13–A15, which are characterized by a visible penguin impact (Table 1: remarks). Mosses are absent at these sites. With higher vegetation cover at distances >200 m from the glacier, D. antarctica becomes the dominant plant. Transect B is characterized by a similar pattern of plant occurrence with respect to the distance from the glacier (Table 1). The vegetation cover, however, is greater beginning 130 to 140 m from the glacier, and D. antarctica is no longer the dominant plant. Here, mosses and, at the last point (B12), U. antarctica have a much greater abundance. Seabird excrements are visible on moraines most distant from the glacier (B9–B12).

Soils are classified according to the International Society of Soil Science (ISSS-FAO, 1998, Table 1). Some of the sites on both transects are enriched with volcanic glass. Hence, they are classified as vitric Andosols (A7, A8, B3, B4, B6, B10) or vitric subunits of Cryosols (A6, A9, B1). Only some soils show active cryoturbation phenomena and are thus classified as Cryosols (A1, A2, A6, A9, B1, B2), whereas the others are gelic subunits of varying soil types.

SOIL SAMPLING

For reasons of environmental protection and in accordance with the Antarctic Treaty, sampling had to be restricted (UNOG, 2000). The sampling design followed the vegetational gradient along both catenary transects (A0–A15, B1–B12). A spade was used to take the soil samples from different horizons. One bulk sample of the topsoil (0–20 cm) and a subsoil sample (30–55 cm, depending on depth of soil) were taken from each site along the transects (Table 1) and stored in plastic bags at −20°C. For determinations of bacterial biomass and bacterial counts, a subsample was taken from the 0–5-cm surface layer of each site along both transects. The samples for enzymatic studies were allowed to thaw at +4°C for 3 d; those for physicochemical analyses air dried at room temperature before being sieved (<2-mm fraction) and analyzed. The total number of soil samples was 58 for microbial and chemical measurements and an additional 29 for bacterial counts and bacterial biomass.

SOIL PHYSICAL AND CHEMICAL ANALYSES

Most chemical and physical properties were determined by methods given by Schlichting et al. (1995). In brief, soil texture was determined by the sieving and sedimentation method after H₂O₂ treatment and (NaPO₃)₆ dispersion. Total organic carbon (TOC) and nitrogen (N_t) were measured by dry combustion at 1200°C and thermal conductivity detection. Kjeldahl digestion and subsequent colorimetric detection of NH₄⁺ were used to determine total nitrogen (N_t). Exchangeable bases (BEC = Ca²⁺ + Mg²⁺ + K⁺ + Na⁺) were extracted by BaCl₂ at pH 8.2 and available phosphates (P_citr) by citrate solution. The pH was measured in CaCl₂ solution (soil-solution ratio 1:2.5), electrical conductivity (EC) in the saturation extract. Soil moisture (%ww⁻¹) was measured gravimetrically after drying at 105 °C for 24 h.

MICROBIOLOGICAL ANALYSES

Total bacterial numbers (TBN), total bacterial biomass (C_bac), and related parameters were determined by epifluorescence microscopy (acridine orange staining of suspensions on polycarbonate membranes, pore size 0.2 µm; Bölter, 1992, 1995). Bacteria were identified by shape and size, and biovolumes were calculated from geometrical characteristics. Biovolumes were further taken as the basis for C_bac by assuming a C content of 10% of total biovolume. Microbial biomass–C (C_mic) was indirectly determined by substrate-induced respiration (SIR) using glucose added to soil and incubation at 25°C for 4 h. The CO₂ evolved was trapped in NaOH and measured by titration (Jäggi, 1976). C_mic (µg C_mic g⁻¹ soil) was calculated as [µg CO₂ g⁻¹ soil] × 20.6 × 0.847 (Jenkinson et al., 1987; Kaiser et al., 1992). Microbial biomass–N (N_mic) was determined by the chloroform-fumigation-extraction (CFE) and ninhydrin-reactive-N NHR-N (Amato and Ladd, 1988). N_mic (µg N_mic g⁻¹ soil) was calculated as [µg NHR-N g⁻¹ soil] × 3.1 according to Amato and Ladd (1988). Microbial respiration is obtained by the titration procedure (Isermeyer, 1952; Jäggi, 1976). The CO₂ evolved was determined as described in the SIR-method. The metabolic quotient (qCO₂) is expressed as mg CO₂-C evolved g⁻¹ C_mic h⁻¹.

N-mineralization was measured under 1-wk water-logged incubation (Keeney, 1982; Kandeler, 1996). Alkaline phosphomonoesterase activity (referred to in the text as phosphatase activity) was assayed by using buffered disodium phenylphosphate (0.2 M borate buffer, pH 10, 20 mM phenylphosphate) as a substrate. The released phenol was estimated colorimetrically at 400 nm (Hoffmann, 1968). Arylsulfatase activity was measured according to Tabatabai and Bremner (1970) using p-nitrophenylsulfate solution as the substrate. These enzyme assays measure p-nitrophenol colorimetrically as the reaction product. Urease activity was assayed as described by Kandeler and Gerber (1988) using urea as the substrate. Xylanase activity was determined as described by Schinner and von Mersi (1990) using 1.2% xylan as the substrate followed by colorimetric determination of the reducing sugars. All enzyme assays were done at 37°C except xylanase, which was done at 50°C.

DATA HANDLING AND STATISTICAL ANALYSIS

C_mic, N_mic, C_bac, respiration, and soil enzymatic activity were calculated on an oven-dry weight (105°C) basis. Ratios of enzyme activity-to-C_mic were calculated as indicators of the hydrolytic potential of the microbial community to break down organic material. For simplicity, we refer to the enzyme activity-to-C_mic ratio as specific enzyme activity. All parameter measurements were tested for normality (Kolmogorov-Smirnov Goodness of Fit test) and homogeneity of variances (Levene's test). Normal distribution of the data was obtained after log-transformation of all data. The classification of the investigated sites based on nine variables (microbial biomass C and N, respiration, N-mineralization, metabolic quotient, enzyme activities) and the subsequent relation of the classification to ornithogenic impact was carried out by cluster analysis using standardized data, Euclidean distance as a measure of dissimilarity, and the centroid method as the agglomeration algorithm. In order to determine whether the ornithogenic impact levels at the sites can be identified by their microbiological properties and what the discriminatory importance of each microbial variable is, discriminant function analysis was applied. The groups were defined according to the ornithogenic impact, transect and soil depth. Multivariate Wilks' Lambda was used for the stepwise selection of the variables. Differences of means of microbiological variables between ornithogenic and mineral soils were tested by univariate analysis of variance using ornithogenic impact as the main factor and P_citr, TOC, and P_citr as covariates, followed by the multiple range test (Student Newman Keuls test). The relationship between TOC, P_citr content, pH, and soil microbial parameters was tested by partial correlation analyses where parameters other than those to be tested were controlling factors.

CHEMICAL AND PHYSICAL CHARACTERISTICS OF THE ANTARCTIC SOIL

Most investigated soils are enriched with gravel and stones, reflecting the special conditions of the parent material (moraines). The dominant soil texture is sandy loam (Table 2) with clay contents up to 20%. Abrupt textural changes between top- and subsoil within one profile occur at A4 and A9 due to the mixing of differently textured sedimentary deposits during glacial transports (Blume et al., 1997; Kuhn, 1997).

The soils (A0–A11, B1–B6) on young moraines (10 to 30 yr) have nearly no vegetation cover, carbonate contents of up to 2.9 g kg⁻¹, and are almost free of organic matter (TOC <2 mg kg⁻¹, Table 2). By contrast, soils (A12–A15, B7–B12) on older moraines (hundreds of years old) are strongly acidified (pH < 5) and enriched with organic matter. These soils show carbonate loss, acidification, and base desaturation together with organic matter accumulation, processes which occur very quickly on soils with high nutrient reserves (due to basaltic origin) and rapidly growing vegetation cover.

Nitrogen compounds are only found in trace amounts (0.10–2.86 g kg⁻¹) in ornithogenically influenced soils (A13–A15, B9, B11–B12). For soils with N concentrations close to detection limits, the C/N ratios are not calculated (A0–A12, B1–8, B10). Along with an increasing occurrence of vegetation and/or the impact of bird excrement, the TOC and N_t contents in the topsoils also greatly increase. At transect A, narrow C/N ratios (4.8–6.7) confirm the penguin impact at site A13–A15. This is in line with data from other areas in Antarctica (see reviews: Campbell and Claridge, 1987; Beyer et al., 1999a), which show such ornithogenic soils to be characterized by C/N ratios of 5 or lower.

The base exchange capacities (BEC), which characterize the released and bound nutrients (Ca, Mg, K, Na), are lowest at the far ends of both transects, which show lowest pH levels (Table 2). However, the constant decline of pH values with increasing distance from the glacier is not reflected by the BEC, probably due to increasing TOC levels, textural and mineral differences of the soil matrix, and the parent materials (Kuhn, 1997).

ORNITHOGENIC SOILS IN ANTARCTICA

In antarctic environments penguins and seabirds play an important role in soil development. Tops of the old moraines are covered by penguin guano (A13–A15) and seabird excrement (B9–B12). The influence of penguins and seabirds has resulted in a strong enrichment of phosphate. Guano and bird excrement enrich soils with organic nutrients (e.g., proteins, urea), which can be mineralized and transformed into nitric and sulfuric acids. This causes extremely low pH at bird nests and sites downslope of the nests (A15). The acidification results in elevated levels of oxalate-extractable iron from basaltic rocks, showing high contents of easily weatherable pyroxenes (Blume et al., 2002). Some soils are characterized by medium to high electrical conductivity due to the influence of guano excrement (A13–A15) (Table 2).

CLASSIFICATION OF ORNITHOGENIC SOILS ACCORDING TO MICROBIAL PROPERTIES

The dendrogram of the cluster analysis, based on microbial biomass and enzyme activity data, displays an unequivocal result: soils influenced by penguin guano or seabird excrement (A13–A15, B9–B12) are clearly distinguished from the mineral soils (A0–A12, B1–B8) (Fig. 2). At the first level of clustering the ornithogenic topsoil B10 is separated from all other sites, showing highest levels of microbial biomass and enzyme activity. This may reflect the low P_citr concentration at B10 compared to the other ornithogenic sites B9, B11, and B12, while all other chemical properties are similar. At the second level of clustering the ornithogenic topsoils are separated from the ornithogenic subsoils and the mineral soils. Only one mineral site (A12) is misclassified into the ornithogenic topsoil cluster. This site belongs to the older moraines (A13–A15) and has a more rapidly growing vegetation cover and soil development than the mineral soils of the younger moraines (A0–A11). It is therefore more similar to the ornithogenic sites A13–A15. While at the third to fifth step of clustering the ornithogenic topsoils are clustered according to their location (transect A and B), the mineral soils and ornithogenic subsoils are not grouped with respect to their ornithogenic impact, location (transect A, B), or soil depth (topsoil, subsoil). We summarize that the effect of penguin guano and bird excrements on soil microbial properties is only evident in the topsoils, whereas the subsoils of ornithogenically affected soils do not differ from the mineral soils.

DISCRIMINANT ANALYSIS

Discriminant function analysis reveals a similarly distinct pattern. Along discriminant axis 1 there is a highly significant discrimination between ornithogenic and mineral soils irrespective of transect and soil depth (Fig. 3). Discriminant function 1 (DF 1) explains 76% of the total variance of the data set and is dominated by N-mineralization (Table 3). Because of the high eigenvalue of DF 1, N-mineralization is the most important variable in discriminating ornithogenic from mineral soils, followed by N_mic and urease activity. DF 2, which is dominated by the metabolic quotient, explains 12% of the variance and is mainly responsible for the discrimination between transect A and transect B of the ornithogenic soils. Within the mineral soils there is neither a significant differentiation between transects nor between soil depths (topsoil, subsoil). DF 3 explains only 7% of the variation, and the highest correlation coefficient between the microbial variable and the canonical discriminant function is found for phosphatase activity. DF 4–7 cover the remaining 5% of the total variance of the data set and are not important for the discrimination of the data (P > 0.05).

Discriminant functions are used to classify the soils into 8 groups (ornithogenic topsoil, ornithogenic subsoil, mineral topsoil, and mineral subsoil of transect A and B, respectively). On the basis of 9 microbiological variables (microbial biomass C and N, respiration, N-mineralization, metabolic quotient, 4 enzyme activities), 100% of the ornithogenic topsoils and only 29% of the ornithogenic subsoils are correctly classified according to ornithogenic impact and transect location. Seventy-one percent of the ornithogenic subsoils are grouped into the mineral soils, and 70–75% of the mineral soils are discriminated according to transect location (A, B) and soil depth (topsoil, subsoil). On that basis soils influenced by penguin guano and seabird excrements are termed ornithogenic soils, while undisturbed sites are termed mineral soils.

ANALYSIS OF VARIANCE

The influence of penguin guano and seabird excrements on each microbial parameter is tested by univariate analysis of variance including TOC, pH, and P_citr as covariates (Table 4). The analysis reveals that all measured microbial variables are significantly favored by ornithogenic excrement input, except qCO₂. Microbial biomass (C_mic, N_mic, C_bac, TBN), respiration, N-mineralization, and enzyme activities of the ornithogenic soils are up to 2 orders of magnitude higher compared to the mineral soils (Figs. 4, 5, 6). Differences of qCO₂ between the two soils are less pronounced, a significant difference being detected only within transect A, where qCO₂ values are 10-fold lower in the ornithogenic soils. Differences between transects are only found for N_mic, N-mineralization and phosphatase activity in the ornithogenic soils. Covariance analysis reveals that the ornithogenic impact as a main factor contributes 99% to explained variance, whereas the importance of the covariates TOC, P_citr, and pH is small. A significant influence of the covariates is detected only for N_mic, qCO₂, N-mineralization, and phosphatase activity (Table 4), albeit explained variance is below 11%.

RELATIONSHIP BETWEEN SOIL CHEMICAL AND MICROBIOLOGICAL PARAMETERS

Partial correlation analysis tests the relationship between microbial data sets and soil pH, P_citr, and TOC distribution (Table 5). A distinct pattern emerges: all microbial parameters, except qCO₂, are significantly correlated with TOC, if P_citr and pH are controlling factors. P_citr is correlated only with qCO₂, while soil pH shows no relationship to microbial variables. The results show that the microbial parameters are related to TOC, if the influence of P_citr and pH is considered. Otherwise, no linear relationship between chemical and microbial variables can be detected, although a quadratic model fits the data (data not shown).

SPECIFIC ENZYME ACTIVITY

Specific enzyme activity (biomass-based enzyme activity) is calculated to determine whether the ratio between enzyme activity and microbial biomass changes according to ornithognenic impact. Specific urease activity and specific xylanase activity are up to 2 orders of magnitude higher at the center of penguin rookeries or bird nest colonies compared to mineral soils, indicating significant ornithogenic impact by excrements (Fig. 6). Specific phosphatase activity shows a similar pattern, although the differences between the two soils are not significant due to the heterogeneity of the data. A significant difference is only detected between top- and subsoils. Specific arylsulfatase activity does not show any pattern, neither between ornithogenic and mineral nor between top- and subsoil (Fig. 6).

MICROBIAL BIOMASS INDICATORS

Measured C_mic values are in the range of data for soils from continental Antarctica (Roser et al., 1993), but about 10-fold lower than levels normally observed in temperate soils (cf., Tabatabai and Bremner, 1969; von Mersi et al., 1992). However, other observations of TBN and C_bac in maritime (Tearle, 1987; Bölter, 1995; Bölter et al., 1997) and continental Antarctica (Ramsay and Stannard, 1986; Bölter, 1992, 1993) show quantities about 2 orders of magnitude larger (10⁸–10¹⁰ n g⁻¹) than those found in the study area. Our results show that the quantity of microbial biomass (C_mic, N_mic, C_bac, TBN) in antarctic soil is primarily controlled by ornithogenic impact. Penguin guano and seabird excrements boost microbial biomass up to 2 orders of magnitude. Cofactors such as TOC, pH, and P_citr were less significant, only N_mic and C_bac being significantly influenced by soil pH. This could indicate a shift in the microbial community from a more bacterial- to a more fungal-dominated community with decreasing pH along the transects. The observed pattern of C_mic is consistent with that in Roser et al. (1993), who record C_mic (SIR) values between 54 µg C_mic g⁻¹ (control sites) and 6700 µg C_mic g⁻¹ (active penguin site) in continental Antarctica. Also, the TBN and C_bac results are similar to those reported in previous studies of continental Antarctica (Roser et al., 1993; Tibbles and Harris, 1996), which shows that ornithogenic soils possess more C_bac and TBN than sites distant from penguin colonies. The 2 to 3 orders of magnitude higher microbial biomass (C_mic, N_mic, C_bac, and TBN) in the ornithogenic soils of both transects can be explained by the positive influence of TOC (from vegetation cover and bird excrements). Accordingly, these parameters are closely correlated, if the influence of P_citr and pH is considered (Table 5) though the beneficial effect of high TOC versus the inhibitory effect of high P_citr and low pH levels does not allow microbial biomass to increase constantly. It seems likely that along transect A C_mic, N_mic, C_bac, and TBN are negatively affected by P_citr levels of >9000 mg kg⁻¹ (A15), compared to sites where P_citr contents are below 3000 mg kg⁻¹ (A13, A14). Within the ornithogenic soils of transect B, P_citr concentrations of >700 mg kg⁻¹ (B9, B11, B12) reduce C_mic, N_mic, C_bac, and TBN. These results contradict those of Ramsay and Stannard (1986), who report higher TBN levels in active versus abandoned penguin colony sites at Cape Bird, Ross Island. Similar observations on microbial biomass have been reported at Windmill Islands in continental Antarctica (Roser et al., 1993). The authors of both studies, however, do not explicitly relate their results to the P_citr content of the soil. The present study suggests that ornithogenic excrements favor microbial biomass by input of substrate and nutrients, until accumulating P_citr becomes inhibitory on microbial growth.

RESPIRATION AND METABOLIC QUOTIENT (qCO₂)

While respiration is generally in the low range, qCO₂ is as much as 10-fold higher than normally reported for temperate soils (Insam and Haselwandter, 1989; Insam and Öhlinger, 1995). Guano and seabird excrements significantly affected the respiratory activity by substrate input, raising respiration up to 5-fold. The favoring effect can be related to the increase of TOC (Table 5), which is mineralized by microorganisms. These observations are in line with Orchard and Corderoy (1983), who report the highest microbial activity in fresh guano samples compared to abandoned rookery sites at Ross Island, continental Antarctica. As the qCO₂ is an indirect measure of a microbial community's energetic efficiency, the reported levels indicate a low metabolic efficiency of microorganisms in using organic substrates. The high qCO₂ at low TOC levels in the antarctic soils indicates a rapid turnover of organic compounds (Bölter, 1992). Differences in qCO₂ between mineral and ornithogenic sites can be explained by bird impact, TOC and P_citr (Table 4). The qCO₂ of the ornithogenic soils is lower, suggesting more favorable conditions for incorporating nutrients into the cell and/or a higher proportion of dormant microbial biomass (Ohtonen et al., 1999). The qCO₂ is negatively related to P_citr (Table 5), indicating a higher energetic efficiency with increasing P_citr values. Our results show that turnover rates of organic matter are high, but efficiency of the microbial turnover is low.

N-MINERALIZATION

The low rates of N-mineralization in the mineral versus temperate soils (Öhlinger, 1993; Kandeler et al., 1999a) indicate a severe deficiency of degradable organic N-compounds, evident in the low TOC levels (<2 mg g⁻¹). Accordingly, a significant correlation between TOC and N-mineralization is detected (Table 5). Within the ornithogenic sites, organic matter input (high in TOC, N, and P) by penguins and seabirds favors N-mineralization up to 25-fold rates, which correspond to levels in temperate soils (Öhlinger, 1993; Kandeler et al., 1999a). Highest N-mineralization rates are found at ornithogenic sites (A14, B10) high in TOC (>7 mg kg⁻¹) but relatively low in P_citr (<700 mg kg⁻¹). However, levels of >9000 mg P_citr kg⁻¹ (transect A) and >1000 mg P_citr kg⁻¹ (transect B) apparently inhibit N-mineralization independently of TOC. N-mineralization turned out to be the main factor for discrimination between ornithogenic and mineral soils (Fig. 3). One explanation is that actual mineralization rates are directly influenced by recent input of organic material.

SOIL ENZYMES

Along both transects, a striking feature is the high enzyme activity near recent penguin rookeries (A13–A15) and seabird colonies (B9–B12). For the specific enzyme activities, however, this pattern is only pronounced for penguin sites along transect A, which have thick guano layers, while in samples from transect B, these activities show weak ornithogenic impact (B9–B12), indicating less input of fecally derived enzymes by seabirds.

Urease activity rates are at least 1 order of magnitude lower compared to temperate soils (Kandeler et al., 1994; Kandeler et al., 1999a). Urease activity increases along both catenary soil transects with increasing TOC levels. Accordingly, a positive correlation with TOC levels is detected (Table 5). The rates in the ornithogenic soils are at least 1 order of magnitude higher than in the mineral soils. These observations are in line with those reported for continental Antarctica (Speir and Ross, 1984), i.e., urease activity is stimulated by urea from penguin excreta at current penguin sites. The extremely high specific urease activity in the penguin rookeries (up to 1000 µg N mg⁻¹ C_mic) indicates that urease may not only be microbially derived, but also derived from penguins (Fig. 6). Specific urease activity along transect B shows a less pronounced response to seabird colonies, but reflects TOC contents of the soil. At penguin sites, urease is probably derived from mircrobes and penguin guano, whereas the activity of urease per unit microbial biomass is not promoted by seabird excreta.

As phosphatases are involved in cycling of organic P-compounds, activity rates are substantially affected by ornithogenic excrements. Alkaline phosphatase exhibits 2 orders of magnitude greater activity in the ornithogenic versus mineral soils. The different activities are explained mainly by ornithogenic impact (e.g., organic P input), and secondly by the soil pH (Table 4). These results are corroborated by Pietr et al. (1983) and Speir and Ross (1984), who report organic P inputs near and on penguin sites. Similar patterns are reported for soils in continental Antarctica (Speir and Ross, 1984). Beyond ornithogenic impact, TOC influences phosphatase activity (Table 5). The mineralization of soil organic phosphorus is therefore intimately associated with the mineralization of organic matter as a whole. Specific phosphatase activity also responds to ornithogenic impact. Highest specific activities are measured in the center of the rookery, followed by the sites 20 cm and 200 cm away. No impact of bird excrement on phosphatase is detected at transect B (Fig. 6). In contrast to urease and xylanase activity, which are relatively high, alkaline phoshatase activity is generally in the low range compared to temperate soils. As phosphatases are inducible enzymes that are produced largely under conditions of low phosphorus availability and strongly inhibited by inorganic phosphates (Speir and Ross, 1978), high P_citr concentration may have limited phosphatase activity in the ornithogenic soils.

Arylsulfatase activity in the antarctic soil is at least 2 orders of magnitude lower than in temperate soils (Fig. 6). This activity is promoted by penguin guano and seabird excrements, showing 2-fold higher values at ornithogenic versus mineral sites. As arylsulfatase is crucial in mineralizing organic matter, it can be related to TOC concentrations (Table 5) (Speir and Ross, 1978). Arylsulfatase activity is detectable in most of the samples, but values are remarkably low. As inorganic phosphorus inhibits sulfatases (Dixon and Webb, 1979), the high P_citr concentrations probably reduced arylsulfatase activity here. Additionally, inorganic S-compounds from excrement may inhibit arylsulfatase activity (Speir and Ross, 1984). Specific arylsulfatase activity shows no response to penguin and seabird impact, indicating no stimulation of arylsulfatase activity by fecal material (Fig. 6).

Xylanases play a role in the biological cycling of carbon and thus respond to organic C addition into the soil (Kiss et al., 1978). At a threshold of 5 mg TOC g⁻¹, soil xylanase activity approaches levels usually reported for temperate soils (Zechmeister-Boltenstern et al., 1991; Kandeler and Eder, 1993). Penguin and seabirds had a substantial effect on xylanase activity. The rates are about 25-fold higher in ornithogenic versus mineral soils, reflecting organic C input by fecal material. A corresponding relationship exists between xylanase activity and TOC (Table 5). The high specific xylanase activity in the center of the penguin rookery at transect A can also be attributed to the input of organic C compounds by excrements. At transect B, specific xylanase activity responds to the TOC distribution (Fig. 6). As xylanase production is widely reported for fungi (Fenice et al., 1997), the low activity levels in the mineral soils may reflect the minor contribution of fungi to the microbial community. In contrast, Rustemeier (2001) reports 2- to 3-fold higher activity rates at comparable sites in the Alps. Bacteria, algae, and cyanobacteria are the most abundant organisms in antarctic soils (Bölter, 1992; Roser et al., 1993). Thus, the low numbers of fungi might be caused by the inhibitory effect of antifungal agents produced by bacteria (Czekanowska and Zabawski, 1988; Pietr, 1995).

Our data suggest that penguin guano and bird excrement enhance the amount of enzymes in the soil. Activities of the mineral soil are 1 (urease, xylanase) and 2 orders of magnitude (phosphatase, arylsulfatase) lower than those normally found in topsoils of temperate regions (Kandeler et al., 1996; Ajwa et al., 1999; Klose et al., 1999; Senwo and Tabatabai, 1999). The input of organic material by vegetation and birds raises enzyme activities (urease, phosphatase and xylanase) to levels of those in temperate soils (Kandeler and Eder, 1993; Beyer et al., 1999b; Kandeler et al., 1999b). The very high biomass-based specific enzyme activities of urease and phosphatase indicate that ornithogenic soils from presently occupied penguin rookeries exhibit high levels of enzyme activities; this is probably not only soil microbially derived, but also from enzymes in faecal material. Pietr et al. (1983) and Speir and Ross (1994) have reported similar patterns for protease and acid phosphatase, strengthening the hypothesis that mineralization of organic compounds in the soil is supported by activities of enzymes derived from penguin intestines.