Calliergon megalophyllum rediscovered in the Netherlands after 50 years: comparison to Swedish habitats

The moss Calliergon megalophyllum is rediscovered in the Netherlands after approximately 50 years of absence, in a location different from before: National Park Weerribben-Wieden. This is a Natura 2000 wetland area, and a Dutch hotspot for rich-fen bryophytes. The species was growing in a fen pool. Plant species composition and water chemistry were compared with Swedish samples collected throughout the country. Water chemistry of C. megalophyllum in Sweden was also compared with four other (semi-)aquatic species: C. giganteum, Scorpidium scorpioides, Sarmentypnum trichophyllum and S. exannulatum. The species is characteristic for poorly buffered habitats, but has nevertheless relatively high pH, which makes it sensitive to acidification, especially when atmospheric deposition is high. In the Dutch locality, buffer capacity is maintained by input of base-rich ditch water through small channels in the fen. The data further suggest that, like other Calliergon species, C. megalophyllum is growing in relatively nutrient-rich habitats, especially with respect to P and K. In the Netherlands, plant nutrient concentrations suggest that P is indeed not limiting, which may enhance survival of the species, as P-poor habitats in this country have become very rare.

21 this paper, we will characterize water chemistry and plant species composition in the new Dutch locality of C. megalophyllum, and compare them to the general habitat conditions of the species in Sweden. To put this characterization in a wider perspective, the Swedish water samples are also compared with four other semi-aquatic to aquatic Calliergonaceae moss species, with which C. megalophyllum has either been confused in the Netherlands, or ones that it sometimes grows together with in Sweden: the closely related C. giganteum, and the species Scorpidium scorpioides (Hedw.) Limpr., Sarmentypnum trichophyllum (Warnst.) Hedenäs, and Sarmentypnum exannulatum (Schimp.) Hedenäs. These species have partly similar, but also slightly different habitats. With this information, we will discuss the perspectives of survival of the species in countries such as the Netherlands, with high environmental pressure.

Study sites and methods
Calliergon megalophyllum was found in October 2014 in a fen pool in National Park Weerribben-Wieden (Fig. 1), a wetland area known for its base-rich and transitional fens belonging to the Natura 2000 habitattype H7140 (van Wirdum 1991, Cusell 2014). The area is also a hotspot for characteristic rich-fen bryophytes (Kooijman 2012). The coordinates of the locality are: 52°42'54.1779''N; 6°7'5.4821''E, and a voucher is deposited in herbarium S (reg. no. B211008). The fen itself consists of a peat layer of approximately 60 cm thickness, formed on top of cover sand from the last Ice age. The fen is mostly covered by Sphagnum spp. and Polytrichum commune Hedw. Calliergon megalophyllum was found in a pool of 10 cm depth in the central part of the fen, probably in remnants of isolated former lakes that terrestrialized only in the 1960s. In the National Park, fens are mainly fed by rain and surface water, as the wetland area is surrounded by lower lying agricultural polders, blocking the influence of groundwater seepage in the nature reserve itself (van Wirdum 1991). In this particular C. megalophyllum fen, the central parts are supplied with base-rich surface water by small ditches which are connected to larger ditches and channels.
After discovery of C. megalophyllum, all other plant species in the habitat were recorded. Vascular plant names are used according to van der Meijden (2005). Samples of surface water and soil pore water were collected for analysis of water chemistry. Soil samples and aboveground vascular plant biomass were collected for their nutrient composition. Depth profiles of the electrical conductivity (EC) were recorded until the sandy bottom was reached with a EC-prikstok, a sampling stick of 2 m which can measure EC and temperature at different depths (van Wirdum 1991). In addition, EC-values of the surface water were recorded with a field meter along a transect from the C. megalophyllum pool, through the small channels, to the ditch where the water originates from. This ditch is located relatively far away from the main channel to reduce nutrient input. EC in the main channel was also measured, and water samples for chemical analysis were collected in both ditch and channel.

22
To compare the Dutch C. megalophyllum habitat with more general habitat characteristics of the species in Sweden, we used a part of the dataset of water chemistry and plant species composition of 481 bryophyte localities in wetlands and rich-poor fens throughout Sweden (L. Hedenäs and A. M. Kooijman unpubl.). This survey was centered on species of the Calliergonaceae and took place in 1991-1997. Even though Sweden has a large population of C. megalophyllum, the species is also relatively uncommon, and only 13 localities could be included (Appendix 1). Thus, to put the data for C. megalophyllum in a wider perspective, the Swedish water samples were also compared with four other semi-aquatic to aquatic Calliergonaceae moss species. The first one is the closely related C. giganteum (n = 21), with which C. megalophyllum has been confused in the Netherlands. The other three are Scorpidium scorpioides (Hedw.) Limpr. (n = 26), Sarmentypnum trichophyllum (Warnst.) Hedenäs (n = 6), and S. exannulatum (Schimp.) Hedenäs (n = 24), which are sometimes growing together with C. megalophyllum, and have slightly different habitats.
The pH and EC of the water samples were measured immediately after collection. The Swedish water samples were divided in two bottles, one acidified with HNO 3 to prevent precipitation of Fe, and the other conserved with chloroform to prevent biological changes. In the Dutch sample, alkalinity (HCO 3 -) was also measured directly after sampling, with titration to pH 4.2, using 0.01 M HCl. For all samples, NH 4 + , NO 3 -, PO 4 3-, Cl and SO 4 2concentrations were measured spectrophotometrically with a Skalar auto-analyzer (Westerman 1990). For the Swedish samples, concentrations of Na, K, Ca, Mg, Fe, Al and HCO 3 were also measured in this way. For the Dutch samples, Na, K, Ca, Mg, Fe and Al were measured with an ICP (Westerman 1990). For the Swedish samples, ionic ratio, a proxy for the relative contribution of groundwater versus rainwater was calculated according van Wirdum (1991), based on Ca-and Cl-concentrations. In the Dutch samples, concentrations of Cl were unfortunately unreliable, as they were 13-35 times lower than Na, while they are normally more or less the same in the Dutch rich fens (I. Mettrop unpubl.). To give at least some indication of the ionic ratio of the Dutch sample, this was based on Na rather than Cl. Soil material and aboveground vascular plant tissue were dried and ground for CNSanalysis (Westerman 1990). P-and K-contents were determined after microwave-digestion with HNO 3 . Nutrient concentrations were measured with ICP.
As there was only one Dutch locality with C. megalophyllum, statistical testing of the habitat conditions was not possible, but values generally fell within the Swedish range. Differences in habitat conditions between the Swedish localities of C. megalophyllum, C. giganteum, S. scorpioides and S. trichophyllum were tested with general linear models, and post-hoc lsmeans tests.

Plant species composition
In the new Dutch locality of Calliergon megalophyllum, accompanying bryophytes were S. scorpioides, Calliergonella cuspidata (Hedw.) Loeske and Sphagnum contortum Schultz, which point to relatively base-rich conditions (Table 1). Vascular plant species included Carex lasiocarpa Ehrh., Menyanthes trifoliata L., Pedicularis palustris L., Juncus subnodulosus Schrank and Ranunculus flammula L., which are also species of relatively base-rich fens. Other vascular plant species of the Dutch locality also occurred in the Swedish localities, such as Carex rostrata Stokes, Comarum palustris (L.) Scop. and Nymphaea alba L., the latter pointing to the former presence of open water. More eutrophic species were also present, such as Lycopus europeus L. and Lythrum salicaria L. The Swedish localities ranged from backwaters of rivers to small (ox-bow) lakes and reed shores. Besides C. megalophyllum, the bryophyte Sarmentypnum exannulatum (Schimp.) Hedenäs was frequently found, which points to slightly lower buffer capacity than in the Netherlands. Among the vascular plants, Equisetum fluviatile L. and Carex aquatilis Wahlenb. frequently occurred. In many Swedish localities, at least some relatively eutrophic species were present, such as Caltha palustris L., Lysimachia vulgaris L., Epilobium palustre L., Carex pseudocyperus L., Phalaris arundinacea L., Typha latifolia L. and Lemna minor L. Typically northern (fen) species were generally absent, even in the north.

pH and buffer capacity
The Dutch locality of C. megalophyllum had a relatively high pH of 6.4 (Table 2). However, buffer capacity was relatively low. Electrical conductivity was only 138 µS cm -1 , and Ca and HCO 3 concentration 289 µmol l -1 (7.6 mg l -1 ) and 560 µmol l -1 respectively. These values are 3-6 times lower than in the base-rich fens dominated by Scorpidium spp. in the area . The pH in the Dutch locality was comparable to the Swedish ones. However, buffer capacity in the Dutch locality, albeit low for Dutch standards, was on the highest part of the Swedish range. In the Swedish localities, mean EC values were 51 (± 32) µS cm -1 , and Ca-concentrations 136 (± 110) µmol l -1 , which is approximately two times lower than in the Dutch locality.
In the Swedish localities, pH values for C. megalophyllum were also relatively high, at least compared to the other moss species (Table 3). The pH values were comparable to those of C. giganteum, but higher than for S. scorpioides, and especially S. trichophyllum and S. exannulatum. Despite high pH, however, buffer capacity for C. megalophyllum was relatively low. Electrical conductivity, ionic ratio,

25
Ca-and HCO 3 --concentrations were significantly higher for C. giganteum and S. scorpioides than for C. megalophyllum. Buffer capacity of the latter species was instead more comparable to S. trichophyllum and especially S. exannulatum. All three species had relatively low values for EC, IR, Ca and HCO 3 -, although lowest values were reached for S. trichophyllum. These results support that the C. megalophyllum habitat has relatively high pH, but low buffer capacity in both Sweden and the Netherlands.
In the Dutch locality, the C. megalophyllum pool was located inside the fen, isolated from the surface water, except for the small ditches which presumably provide base-rich water from the ditch. In the pool, EC-values were below 150 µS cm -1 from the surface to a depth of 50 cm, and EC-values lower down were hardly higher (Fig. 2). In the small ditch system, EC values were around 231 µS cm -1 in the ditch, and gradually decreased along the way to 138 µS cm -1 (Fig. 3). In the ditch, EC values were approximately two times lower than in the main channel, which does not provide water to the fen and pool directly, due to the long pathway before the water reaches the supplying ditch. This was also true for Ca-concentrations, which were 989 and 559 µmol l -1 in the main channel and supplying ditch, respectively. Table 2. Chemical composition of the surface water in the only Calliergon megalophyllum locality in the Netherlands (NL) and 13 localities throughout Sweden (samples 20-459). * = ionic ratio (IR) in the Netherlands was based on Ca and Na, instead of Ca and Cl, because Cl-concentration measurements were unreliable.   NL  20  175  165  195  252  167  459  461  194  150  24  147  332 pH 6.4 6.5 6.7 6.2 6.9 6.6 6.4 6.8 6.9 6.7 6.7 6.6 6.9 7.0  Table 3. Chemical composition of surface water in localities throughout Sweden for Calliergon megalophyllum (n = 13), compared to three other (semi-)aquatic moss species Calliergon giganteum (n = 21), Scorpidium scorpioides (n = 26), Sarmentypnum trichophyllum (n = 6) and Sarmentypnum exannulatum (n = 24). Values given are mean values and standard deviations. * = significant general effect (p < 0.05); different letters indicate significant differences for a parameter between particular species.
C. megalophyllum C. giganteum S. scorpioides S. trichophyllum W. exannulatum pH* 6.7 (0.2) c 6.7 (0.6) c 6.3 (0.6) b 6.0 (0.7) ab 5.6 (0.6) a EC (µS cm -1 )* 51 (32) a 142 (102)  may both affect P-availability, they were measured as well. However, concentrations were low in general, and did not generally differ between moss species. In the Dutch locality with C. megalophyllum, concentrations of N and K in vascular plant tissue showed values of 16.5 and 15.3 mg g -1 (Table 4), which are not elevated or very low. However, P-concentrations were relatively high, and showed values of 1.2 mg g -1 . Also, the N:P ratio of 13.8 was relatively low (Koerselman and Meuleman 1996).

Discussion and conclusions
Based on the available information, the results suggest that the habitat conditions in the new Dutch locality of Calliergon megalophyllum fit well within the range described for Sweden. Also, comparison of Swedish localities for C. megalophyllum and the closely related C. giganteum makes clear that the two species have actually different habitats, especially with respect to buffer capacity. For C. giganteum, buffer capacity was approximately three times higher than for C. megalophyllum. These habitat differences further supports that they are really different species (Jensen 1939, Tuomikoski 1940, Nyholm 1965, Tuomikoski and Koponen 1979, Hedenäs 1993a, in contrast to what was assumed by Rubers (1989).

High pH and low buffer capacity
Calliergon megalophyllum is characteristic for habitats with a high pH of 6.4-7.0, but nevertheless low buffer capacity. This is a difficult combination, as low buffer capacity may be advantageous to photosynthesis under water, but more difficult to combat acidification of the habitat. Low buffer capacity may to some extent help photosynthesis of submerged mosses. Under water, CO 2 supply may be a problem due to low rates of diffusion, but is also chemically limited at high pH, when the main carbon form is HCO 3 -. Many aquatic macrophytes can use bicarbonate as alternative carbon source (Madsen and Sand-Jensen 1991), but there is no evidence that bryophytes can do this, except perhaps Fontinalis antipyretica Hedw. (Glime andVitt 1984, Penuelas 1985). When buffer capacity is low at relatively high pH, submerged mosses may be able to capture at least some CO 2 from the water. This may Figure 2. EC-depth profile (µS cm -1 ) in the Dutch locality with Calliergon megalophyllum, measured in october 2014. Figure 3. EC-surface profile (µS cm -1 ) in the dutch locality with Calliergon megalophyllum, from the ditch where the water comes from, through a series of small ditches inside the fen, to the pool were the species was found. EC-values were measured in October 2014. The mean EC-value in the main canal, which is not directly connected to fen and pool, is also given. be a reason why S. trichophyllum, which has lower pH and slightly lower buffer capacity than C. megalophyllum, is growing completely submerged, while the latter was found at or above the water surface in 50% of the cases. When buffer capacity becomes too high, such as for, e.g. C. giganteum and S. scorpioides, growing above the water may be a better option. While low buffer capacity may be an advantage for photosynthesis in aquatic bryophytes, it is a problem to maintain a relatively high pH, especialy in areas with high atmospheric deposition. In the Netherlands, weakly buffered habitats with S. scorpioides in the Pleistocene, sandy part of the country were still present in the 1960s (Kooijman andWesthoff 1985, Arts 1990). However, the species has disappeared from this region, probably due to high atmospheric deposition. In the Netherlands, approximately 1.5 times higher amounts of buffer components are needed to maintain the pH in the habitat around 6.5 compared to less polluted areas in e.g. Sweden (Kooijman 2012). It is not unlikely that C. megalophyllum has disappeared from part of the former Dutch localities by acidification in a similar way.
In the new Dutch locality, it is not clear whether C. megalophyllum is really new to the area, or whether it has been there for a long time. In any case, the buffer capacity is relatively high compared to Sweden, albeit low compared to real rich-fen habitats in the National Park dominated by Scorpidium spp. ). Buffer capacity is presumably maintained by input of relatively base-rich ditch water into the small channel system during high water levels in the National Park. In other parts of the National Park, similar channels, with higher buffer capacity, have led to an increase of rich-fen bryophytes such as S. scorpioides, S. cossoni (Schimp.) Hedenäs and Hamatocaulis vernicosus (Mitt.) Hedenäs (A. M. Koijman, L. Hedenäs, I. Mettrop and C. Cusell unpubl.). More systematic research is needed, but it is very likely that occasional flooding with base-rich, relatively nutrient-poor water can mitigate at least some of the adverse effects of high atmospheric deposition.

Nutrient availability
Calliergon megalophyllum seems to prefer intermediate to nutrient-rich habitats. In both Sweden and the Netherlands, phosphate concentrations in the water were relatively high, and eutrophic plant species occurred in almost all relevés. In at least some of the old Dutch localities, where C. megalophyllum has disappeared, eutrophic species were also present, such as Glyceria maxima (Hartm.) Holmb. and Lycopus europeus in the Belversven in 1947 (Appendix 1). Also, plant N:P ratios in the present locality suggest that P is not a limiting factor. Other Calliergon species are also growing in relatively nutrient-rich habitats, such as C. cordifolium and to some extent C. giganteum.
This may be an inherent habitat trait of the genus, inherited from ancestors during the evolution of the Calliergonaeceae (Hedenäs and Kooijman 1996).
Growing under relatively nutrient-rich conditions may be an advantage in the Netherlands, where eutrophication in wetlands is a real problem (Lamers et al. 2014). In the National Park, approximately 79% of the annual P-input comes from adjacent agricultural polders, especially in winter (Cusell 2014). Nevertheless, values around 0.3 µmol P l -1 in the water are not really high, and C. megalophyllum would probably not survive hypertrophic conditions. In some of the old Dutch localities, C. megalophyllum may have disappeared due to eutrophication of the water, such as in former ditches. In the National Park Weerribben-Wieden, its present locality, however, nutrient levels have decreased in the National Park over the past decades. Although total P did not decrease, phosphate levels in the main channels of the National Park have clearly dropped to values around 0.1 µmol l -1 (Cusell 2014). In one of the best rich-fen complexes in the area, eutrophic bryophytes such as C. cuspidata have been replaced along the shoreline by characteristic species such as S. scorpioides, while aboveground vascular biomass decreased from approximately 1000 to 250 g m -2 . A further decrease in P-input is much more difficult to achieve, but there is no reason to say that the present conditions are not suitable for C. megalophyllum, especially when buffer capacity in the habitat is maintained the way it is.

Concluding remarks
It is unknown why C. megalophyllum has disappeared from the Netherlands in the past. However, the particular habitat conditions of high pH, but low buffer capacity suggest that acidification has played at least some role, similar to the disappearance of weakly buffered habitats with S. scorpioides (Kooijman and Westhoff 1995). In addition, pollution of surface waters in the 1960s may have led to local extinction, although C. megalophyllum is a relatively eutrophic species and can at least stand slightly elevated nutrient levels, such as in the new Dutch locality. Whether it will survive there probably mainly depends on the buffer capacity, which is maintained by input of baserich surface water through small channels. Whether the species can expand in the Netherlands is more difficult to say, since weakly buffered habitats with relatively high pH have become very rare.