In the literature on arbuscular mycorrhiza, available soil P levels are considered high when they exceed 200 mg kg−1 soil (Sylvia & Schenck, 1983). Published examples of total P concentrations clearly inhibiting root colonization by arbuscular mycorrhizal fungi range from 228 mg kg−1 soil (Graham et al., 1981) to 600 mg kg−1 soil (Menge et al., 1978). In our study, available soil P contents were orders of magnitudes higher, ranging from 6332 mg kg−1 at plot 2 to 12 249 mg kg−1 at plot 1 (Table 1).
Phosphorus concentrations in roots (0.47, 0.49 and 0.77% at plots 1–3, respectively), stems (0.27, 0.32 and 0.27%) and leaves (0.46, 0.38 and 0.32%) were also high enough to suggest that no, or only slight, colonization by AMF should have occurred. In other studies, root and shoot P concentrations of ≈0.3% were sufficient to suppress more than sporadic mycorrhization (Graham et al., 1981; De Miranda et al., 1989).
Thus, if P had been the only determining factor for mycorrhization, plants at our field site should have been, at most, slightly mycorrhizal. Nevertheless, A. vulgaris was found to have high mean mycorrhization rates of 80 and 85% (internal mycelium), 61 and 67% (arbuscules), and 43 and 55% (vesicles) at plots 1 and 2, respectively. At plot 3 colonization was significantly lower: mean rates were 17% (internal mycelium), 11% (arbuscules), and 1% (vesicles).
The most prominent difference between soils of the three plots was the N level. Whereas plots 1 and 2 had low total N concentrations (0.16 and 0.18%, respectively), N was significantly higher at plot 3 (0.34%), at a greater distance from the former factory. Reports about N-fertilization experiments to define soil N thresholds inhibiting mycorrhizal colonization are scarce, as N has rarely been regarded as an important factor determining root colonization by AMF.
The trend in plant N concentration measured at the three plots paralleled that in soil N. Plants at plots 1 and 2 were characterized by a significantly lower N concentration in leaves (1.76 and 2.03%, respectively), stems (0.46 and 0.49%), and roots (0.61 and 0.65%), compared with plants at plot 3 (leaf N 3.23%, stem N 0.83%, root N 1.31%). Crops are known to be N-limited if their shoot N concentration is below 1.4% (Verhoeven et al., 1996), which applies to plots 1 and 2 when considering mean values of leaves and stems. Azcón et al. (1982) found a reduction of mycorrhizal colonization at a shoot N >1.9%, which would apply to plot 3. Other clear data on tissue N concentrations at which a distinct reduction of mycorrhization occurs are not available in the literature, but there are a number of studies concerning mycorrhizal colonization dealing with the proportion of P and N.
Mosse & Phillips (1971) found that, at high P concentrations, roots were colonized only when the medium lacked N. Nevertheless, they did not consider that AMF may be needed for N nutrition. A later connection between P and N availability was established by Sylvia & Neal (1990), who showed that under N-limiting conditions, P addition had no effect on mycorrhizal colonization whereas, when a sufficient amount of N was supplied, colonization was suppressed by the addition of P. Also Bååth & Spokes (1989) found that only a combination of high P and N concentrations led to reduced mycorrhization.
Whether a plant profits from being mycorrhizal is a question of carbon cost vs benefit through improvement of nutrient uptake by AMF (Corkidiki et al., 2002). If plants are nutrient-limited, they profit from allocating carbohydrates to the fungi because, in turn, the mycobionts provide them with soil nutrients. If all required elements are available without mycorrhiza, the carbon cost exceeds the benefit and plants stop allocating C to the symbionts, which become C-limited (Treseder & Allen, 2002).
Sylvia & Neal (1990) found no relationship between tissue N : P ratios and root colonization rates, but attributed mycorrhization at high P levels to N deficiency in general. In contrast, at least at intermediate N levels, Liu et al. (2000) obtained a negative correlation between shoot N : P and percentage colonization by AMF.
Like the soil and plant N concentrations, plant N : P ratios found in our study were significantly (marginally in the case of root N : P) lower in plants growing at plots 1 and 2, which had respective ratios of 4.29 and 5.46 in leaves, 2.04 and 1.58 in stems, and 1.45 and 1.34 in roots, compared with plants at plot 3 (10.4 in leaves, 3.33 in stems, 1.99 in roots). Koerselman & Meuleman (1996) assessed the suitability of vegetation N : P ratio as an indicator for nutrient limitation, and found that a shoot N : P ratio <14 at a community level indicates N limitation. Unfortunately their study, like most others dealing with this subject, was restricted to wetland ecosystems. However, a recent literature survey by Tessier & Raynal (2003) revealed thresholds for N limitation from different ecosystems, ranging from N : P <6.7 to 16, and the authors suggested that these ratios may as well be used to indicate nutrient limitation in individual species. Aerts & Chapin (2000) assumed that a leaf N : P ratio of ≈10 indicates optimal conditions for plant growth, and Güsewell (2004) concluded that a vegetation N : P ratio <10 indicates N limitation. In reference to all these values, A. vulgaris appears to be N-limited at plots 1 and 2 of our field site; accordingly, the larger N : P ratio of plants at plot 3 indicates a minor, or no, N limitation.
With the exception of root N : P, both N and N : P in different plant parts were significantly negatively correlated with mycorrhization rates. Spearman's rank correlation coefficients (rs) for leaves and stems did not differ much between N and N : P, whereas in the case of roots rs was distinctly higher for the correlation of percentage colonization with N than with N : P. Given the high P availability at our field site, leaf and stem N concentration and N : P ratio both were good indicators of the mycorrhizal status of a plant, whereas in roots N concentration was a better indicator than N : P ratio. These results suggest that N deficiency, as well as N : P imbalances in plants, have similar effects to P deficiency in stimulating root colonization by AMF.
In some cases the relationship between the parameters appeared to be more-or-less linear, for example for internal mycelium vs N concentration of leaves (Fig. 4a); internal mycelium vs N concentration of roots (Fig. 4g); and internal mycelium vs N : P ratios of leaves (Fig. 5a). In other cases, an N or N : P threshold appeared to limit the development of mycorrhizal structures. This is most evident when looking at arbuscules and particularly vesicles, for example in the event of arbuscules vs stem N (Fig. 4e); vesicles vs stem N (Fig. 4f); vesicles vs root N (Fig. 4i); arbuscules vs stem N : P (Fig. 5e); and vesicles vs stem N : P (Fig. 5f), and may well indicate specific reactions of these different fungal structures to changed N levels or N : P ratios in host plants. According to this hypothesis, internal mycelium would directly react proportionally to changes in plant N or N : P level, whereas the production of arbuscules, and especially of vesicles, would depend rather on the crossing of threshold levels.
To our knowledge there are no previous studies specifically relating the development of different AMF structures to changes in nutrient availability, with the exception of internal vs external mycelium. For example, Abbott et al. (1984) and Thomson et al. (1986) found that P fertilization first reduced structures inside the root, and Johnson et al. (2003) concluded that external structures are more responsive to changes in N availability. In our study we did not determine the length of external mycelium because, under field conditions, it is almost impossible to distinguish AMF hyphae from those of other soil fungi.
Soil and plant element concentrations other than P and N were never related to both each other and mycorrhization rate at the same time, and for this reason were not further considered.
The results of the N-fertilization experiment confirm the findings of the previous experiment. Because of the proximity to plot 1 of experiment 1, plants at the control plots were exposed to very similar nutrient conditions, i.e. they were well provided with P but presumably deficient in N.
Above-ground biomass data of A. vulgaris support this assumption. Although the results just bordered on significance, plant dry weight per square metre was several times higher at N-fertilized than at control plots (Table 4).
Like the plants at plots 1 and 2 of experiment 1, A. vulgaris at control plots had high mean mycorrhization rates (71% internal mycelium, 46% arbuscules, 36% vesicles), which probably are a consequence of N limitation. In accordance with plot 3 of experiment 1, plants at N-fertilized plots had significantly lower mycorrhization rates (34% internal mycelium, 22% arbuscules, 10% vesicles). These values are higher than those found at plot 3, which may be caused by the N limitation only being partially relieved at the applied N-fertilization levels. Nevertheless, the N-fertilization experiment clearly shows that increasing the N provision of A. vulgaris decreases mycorrhization rates.