1Many threatened bryophytes are restricted to patchy and temporary substrates such as dead wood and tree stems. Their persistence depends on successful colonizations of new patches. Spore germination may then be limited by substrate quality and wetness.
2In vitro experiments were used to test the effects of pH and moisture on the establishment of spores of the moss species Neckera pennata Hedw. and Buxbaumia viridis (DC) Moug. & Nestl.
3Low pH and water potential prolonged the lag phase preceding germination and reduced final germination. The interaction between pH and moisture suggests that high water availability facilitates germination at suboptimal pH, and vice versa.
4The results reflect the species’ habitats: the wood-inhabiting B. viridis had higher capacity to germinate at low pH, while spores of the epiphyte N. pennata showed earlier germination at low water potential and survived longer in a dry state. This supports the notion that bryophytes are most strongly affected by substrate quality during establishment.
5We suggest that a trade-off exists among moss spores between the ability to colonize substrates with low moisture-holding capacity and low pH, and that the positive effect of high pH is largely that it speeds up germination thereby enabling the spores to exploit short, moist periods.
Dispersal and establishment are of fundamental importance for the persistence of metapopulations, especially for species restricted to patchy substrates of limited duration (Herben et al. 1991; Hanski & Hammond 1995). Many bryophytes are examples of species restricted to patchy, transient habitats such as dead wood or deciduous tree stems. They form ‘patch-tracking’ metapopulations (Thomas 1994; Snäll et al. 2003): if new patches are not colonized before the occupied patches disappear, the species will be lost from the site. Bryophytes establish new colonies either from spores or from asexual propagules or fragments. Establishment is usually more effective from vegetative diaspores than from spores (Keever 1957; Mishler & Newton 1988; Kimmerer 1991). However, spores are small and can be dispersed over long distances by wind (van Zanten & Pócs 1981). Decaying wood and trunks of old deciduous trees host many threatened bryophytes (ECCB 1995; Berg et al. 2002) and their rarity may be caused by the lack of substrate patches or by dispersal limitation. Rare species could have a narrower fundamental niche than do common species (Gaston & Kunin 1997) and, for bryophytes, substrate quality might be most important during the establishment stage (Brown 1982; Bates & Bakken 1998). We need to increase our knowledge about establishment probability under different environmental conditions to improve understanding of the distribution and rarity of species.
Moisture is vital for spore germination; phosphorus availability and pH have also been shown to be important (Thomas, Proctor & Maltby 1994; Sundberg & Rydin 2002; Wiklund 2003). Establishment may also depend on temperature and light. There may also be interactions among those factors so that the response to pH may depend on the availability of moisture. The challenge of laboratory studies is to mimic the range of moisture conditions the species may encounter in the field. However, the use of natural substrates poses a problem because they are often inhabited by fungi or algae, but sterilization can alter substrate chemistry. Culture media offer a practical alternative. Polyethylene glycol (PEG) is often used to maintain culture media at fixed water potentials. PEG is a long-chain inert organic polymer that alters the matric potential of the solution (Steuter, Mozafar & Goodin 1981). PEG makes it easy to perform factorial growth experiments with moisture and pH.
We studied two bryophyte species from contrasting habitats, which are patchy in space and time. Buxbaumia viridis (DC) Moug. & Nestl. is an epixylic moss growing on dead wood in late stages of decay; Neckera pennata Hedw. is an epiphytic moss inhabiting stems of deciduous trees. Both species are red-listed in Sweden (Hallingbäck 1998) and in Europe (ECCB 1995). Wood in late stages of decay remains wet after rainfall because of its large water-holding capacity and because of the production of water as a by-product of cellulose metabolism during decomposition (Rayner & Boddy 1988). Bark, however, is prone to desiccation, and wet periods may offer brief windows of opportunity for germination. Phillips (1951) suggested that, in bark-inhabiting bryophyte communities, an abundance of moisture can compensate for low pH, for example. Another study (Wiklund 2003) indicated a possible interaction between moisture, pH and phosphorus. Our aim was to investigate the combined effect of pH and moisture on germination and protonemal growth of the two species. An important question is how long a period of wetness is required for germination success, and if the required length of the period depends on the levels of moisture and pH. We hypothesized that the effect of pH on germination depends on moisture, and that habitat differences are reflected in properties of spore germination. Spores of N. pennata are dispersed during dry weather conditions, while spores of B. viridis are dispersed during rain (personal observations; Patterson 1953; Schofield 1985). In nature, spores often land on a dry substrate surface and have to wait for rain before they can germinate. We made a study to test the survival time of such dry-stored spores
Materials and methods
Four experiments were carried out to determine the effects of pH and water potential on spore germination and protonemal growth. Sporophytes of N. pennata and B. viridis were collected in eastern central Sweden (59°44–47′ N, 17°32–35′ E). A mixture of spores from several capsules with a green and vital spore mass was used in the experiments. Capsules with the lid still attached were surface sterilized for 1·5 min in 1·5% (v/v) sodium hypochlorite solution and washed in distilled water. A spore suspension in distilled water was homogenized with an ultrasonic processor, and 50 µl of the spore suspension was added to Ehrlenmeyer flasks containing 10 ml full-strength sterilized nutrient solution (N 50, P 10, K 43, Ca 3, Mg 4, S 4, Fe 0·35, Mn 0·2, Zn 0·03, Mo 0·0008, Cu 0·015, B 0·1 mg l−1); the medium was adjusted to pH 3, 4, 4·5, 5, 6 and 7 with HCl or NaOH. Water potentials were adjusted with PEG 6000 to −0·5, −1, −1·5 and −2 MPa. Treatments with no addition of PEG are labelled water potential 0, although they are slightly negative as a result of the nutrient addition. Water potential was measured using a thermocoupled dew-point hygrometer with a Wescor C-52 sample chamber connected to a microprocessor-controlled water potential data system (HP-115 Wescor Inc., Logan, Utah, USA). The final spore concentration in the flasks was 4000–6000 spores ml−1 as calculated from information about spore numbers per capsule (Hagström 1998; Wiklund 2002). The flasks were sealed with parafilm and randomly positioned in a growth chamber with light/dark periods of 16/8 h at 15/7·5 °C. Irradiance (PAR) varied between 67 and 209 µmol m−2 s−1 at different positions in the growth chambers, as measured using a light meter (Li-250, Li-Cor, Lincoln, Nebraska, USA). Each of the experiments consisted of three or four replicates, but all combinations of species, pH and water potential were not exploited in all experiments. The replicates were moved within each shelf every second day. The proportion of germinated spores was assessed with a light microscope at suitable intervals to capture cumulative germination. A subsample of 50 spores was examined and classified as non-vital; ungerminated but vital; or germinated. Undeformed spores with chloroplasts were classified as vital. A spore was classified as germinated if the spore wall was broken by the germ tube. In germinated spores, the number of developed cells in the protonema was counted. For protonemata with more than five cells, the cells were counted in only 15 protonemata.
dry storage of spores
Survival under dry conditions was investigated by adding spores of N. pennata and B. viridis to dry filter paper kept in Petri dishes. Two replicates of each species were placed in a growth chamber under the same conditions as for the germination experiments. After 12, 25 and 48 days, spores were taken from the filter paper and added to a nutrient solution at pH 6. After 9–13 days in nutrient solution, three samples were taken from each replicate, and 50–100 spores were counted under a microscope and classified as germinated; not germinated but vital; or not vital.
Cumulative germination curves for each replicate were modeled using the Gompertz growth model (Draper & Smith 1998):
where ω is cumulative germination at a given time and t is time in days from the start of cultivation. The parameters α, β and k were estimated by non-linear regression (PROC. NLIN with the Marquardt iterative method; SAS 2001) where α represents the final cumulative germination. The parameters β and k were used to calculate the time to germination of 2% of the spores (t2) and to 50% of the final cumulative germination (t50). The value of t2 was interpreted as the lag phase before the start of spore germination; t50 is the time taken for the average spore to germinate, and the difference between t50 and t2 (Δt) was used to evaluate the rate of germination. The Gompertz model was used because it is not symmetrical about its point of inflection, and the cumulative germination rate was faster in the first 50% fraction of the spores than in the second.
All experiments were analysed collectively and, to correct for the different proportions of viable spores from the start of each experiments, all α values were multiplied by a correction factor calculated from germination at pH 6 and water potential 0. To predict final cumulative germination (α) at different pH and water potential values logistic models were constructed (PROC. GENMOD with Pearson scaled deviance; SAS 2001). Similarly, linear models of log-transformed values of t2 and Δt were made (PROC. GLM; SAS 2001). In the analyses of Δt, replicates with less than 10% germination were excluded. All significant (P < 0·05) variables, including quadratic terms and interactions, were entered in the models. The statistical analyses were performed with the sequential sums of squares because of lack of orthogonality, as recommended by Grafen & Hails (2002). Protonemal growth rate was determined as an average for each replicate from the number of cells in the 10% fraction of spores with the fastest germination rate, and was calculated as:
where time1 was the first observation day when at least 10% of the spores had germinated, and time2 was the day when growth rate started to slow down. We thus used the period with linear growth rate (6–45 days in different experiments). To predict the effect of pH and water potential on growth rate, linear models using square root-transformed values of the dependent variable were constructed (PROC. GLM; SAS 2001).
Low water potential and initial pH increased the time it took for spores to start to germinate (t2) and reduced the proportion of spores that finally germinated (α) in both species. Further, the time from the start of germination to 50% of the final germination (Δt) increased (Fig. 1). There was a strong interaction between pH and water potential in the prediction of the final cumulative germination: the moss spores reacted positively to one factor only when the other factor was in a favourable range. The number of days needed for germination to start (t2) was strongly affected by a combination of pH and water potential, and varied between 2 and 50 days under the different treatments in which spore germination occurred. The time to start of germination and time from t2 to t50 (Δt) were significantly correlated (r = 0·56, P < 0·0001, n = 51 for B. viridis; r = 0·25, P = 0·04, n = 68 for N. pennata) with early (t2 < 5 days) and homogeneous (Δt < 5 days) germination at unlimited water supply in combination with a pH of 4·5 and higher. At water potentials below −1 MPa, germination started late (t2 > 15 days) and was heterogeneous (Δt > 15 days). The final proportions of germinating spores in favourable conditions (water potential 0, pH 6) were 63 and 81% in B. viridis in the two experiments that included this species, and varied between 86 and 96% in four experiments that included N. pennata. Germination decreased rapidly at water potentials below −1 MPa in B. viridis, and at pH below 4·5 in N. pennata. At pH 4·5, germination was inhibited when water potential was decreased to at least −0·5 MPa. In B. viridis final cumulative germination was significantly reduced at pH 3 and 4 compared with higher pH, but the effect, especially at pH 3, was less apparent than for t2 and Δt.
The responses varied among the experiments, and a comparison between the species can best be made from one experiment where the two species were grown simultaneously at pH 4·5 and 6 (Fig. 2). Comparison between the species at pH 6 showed that t2 was significantly extended in B. viridis compared with N. pennata at water potentials below −0·5 MPa. Final germination revealed no difference between the species after correction for different initial spore viability. The germination time was not significantly different between the species.
survival of spores
Spores of N. pennata survived dry storage significantly better than did spores of B. viridis (Fig. 3). Only 20% of spores in B. viridis germinated after 12 days’ dry storage, compared with 60% in N. pennata. Virtually all spores of both species died after 48 days’ dry storage.
Protonemal growth was generally affected by pH and water potential in a similar way to germination (Fig. 4). However, low pH had a stronger effect on protonemal growth than on final cumulative germination, severely reducing growth at pH <5 in both species. In B. viridis protonemal growth showed an optimum at pH 5 while growth in N. pennata was greatest at pH 6·5–7.
Although bryophytes are generally desiccation-tolerant (Proctor & Tuba 2002), spore germination is a critical stage which requires water. Our study shows that moss spores have the capacity to germinate at water potential of −2 MPa, a value at which most seeds fail to germinate (Battaglia 1997; Roman et al. 1999; Sy, Grouzis & Danthu 2001), but only if pH >5. The interaction between pH and water potential effects on germination suggests that high moisture facilitates germination at suboptimal pH, or vice versa. Further, our study demonstrates the effect of pH and water potential on the length of the lag phase preceding germination, and on the germination rate. This time effect is ecologically important as slow spore germination increases the risk of desiccation or disappearance of spores through wind or predation.
Our results reflect the habitat of the species. Buxbaumia viridis appears to be adapted to germinate at low pH, and the faster onset of germination in N. pennata at reduced water availability would enable it to establish on bark. These results agree with those of other studies (Cameron & Wyatt 1989; Bosley, Petersen & Rebbeck 1998). Although the two species normally inhabit contrasting microhabitats and responded differently to the treatments, they showed a common pattern: the three-dimensional plots of final germination percentage vs pH and water availability (Fig. 1c,f) clearly show a split between a tenable and an untenable region of the response surface. We suggest a general trade-off between the ability of moss spores to colonize substrates with low moisture-holding capacity and low pH, with the effect that substrates prone to fast desiccation (such as the bark of living trees) can be colonized only if they have a fairly high pH. Substrates with a high water-holding capacity, such as wood in late stages of decay, or peat, can be colonized despite low pH. Interactions between environmental factors, preventing seemingly suitable substrates from being colonized, are probably common in nature. Our study emphasizes the interaction between moisture and pH, but other interactions are possible. For instance an interaction between light and CO2 was found to affect growth in an amphibious liverwort (Andersen & Pedersen 2002). Temperature is another factor to consider, influencing both the length of the lag phase and final cumulative germination (Newton 1972; Furness & Hall 1981).
Establishment constraints have implications for the understanding of metapopulation dynamics. Both B. viridis and N. pennata occupy a limited fraction of seemingly suitable substrate patches (Hagström 1998; Kuusinen & Penttinen 1999; Wiklund 2002). Possibly, many of those patches are of a lower quality, with lower probability of being colonized.
The better survival of dry-stored spores in N. pennata, compared with B. viridis, supports the hypothesis that xerophytic species have more drought-resistant spores (van Zanten & Pócs 1981). Consequently, spores of N. pennata have a higher chance of surviving a prolonged period of drought after dispersal, thereby increasing the probability of establishment. However, spores of both species retained vitality for a short period compared with several other bryophytes (Malta 1922; van Zanten & Pócs 1981; Dalen & Söderström 1999). The results are, however, in agreement with a study of fern spores, where survival of green spores was 48 days, whereas yellow-brown spores remained viable for 3 years as a mean (Lloyd & Klekowski 1970), and both B. viridis and N. pennata have green spore mass.
Buxbaumia viridis and N. pennata are classified as vulnerable in Europe (ECCB 1995) as a consequence of a reduction of decaying wood and host trees in forests. Spores of both species, in particular N. pennata, germinated at a lower pH in the experiments than those at which they normally occur in nature (Wiklund 2003; unpublished data). One explanation could be the interaction with moisture. In the lower pH range, germination was possible only if water availability was unlimited, which would apply only for short periods in the field. Another explanation could be the stronger effect of decreased pH on protonemal growth than on germination. In N. pennata protonemal growth was significantly (P < 0·05) reduced at pH 5 compared with pH 6 and 7. A stronger pH effect on protonemal growth than on germination has also been observed in three Splachnum species (Cameron & Wyatt 1989). Spore germination is of fundamental importance for B. viridis, which is a short-lived moss and entirely dependent on establishment from spores, but is probably also significant for N. pennata. In central Sweden spores of B. viridis are dispersed in June, and prolonged periods without precipitation could be a problem because of short-lived survival of spores. Although B. viridis can tolerate low pH, the probability of establishment increases with increasing pH. According to the models, B. viridis requires at least 25 days to germinate and develop a five-cell protonema at pH 3, but only 11 days at pH 5, provided water is unlimited. Throughfall and litter from deciduous trees increase pH and nutrient availability (Nihlgård 1970; Nordén 1991) for species growing on logs and stumps beneath the trees. Hence favourable germination conditions could be secured by increasing the density of deciduous trees in coniferous forests (see also Wiklund 2003). Neckera pennata, like most epiphytic mosses with frequent spore formation, depends on short, wet periods (temporary windows of opportunity) to establish new colonies. If a spore of N. pennata reaches a tree with a surface at pH 7, according to the models it requires 3 days to form the first germ tube and another 7 days to form a five-cell protonema – in all, a period of 10 days with unlimited moisture availability. At pH 5 or 4 the wet period required increases to 15 and 50 days, respectively. Probability of establishment of N. pennata benefits from a high density of host trees to increase spore load and overcome dispersal limitation (Snäll, Rydin & Ehrlén, in press) and from the high frequency of deciduous trees with a high bark pH, preferably in moist areas, to overcome establishment limitation.
In conclusion, many plant species are dependent on recurrent colonizations of newly formed habitat patches. Metapopulation models focus on interpatch distances to explain colonization events, and our study adds that interactions of habitat factors can strongly affect the probability of establishment for a diaspore once it has reached the patch. For the conservation of threatened species with such patch-tracking metapopulations, both the spatial arrangement and quality of the patches must be considered.
We thank Robin Kimmerer and Staffan Karlsson for useful comments on the manuscript. Financial support was received from Formas.