One of the remarkable features in the intertidal environment is emersion and submersion of seaweeds by tidal cycles. Intertidal seaweeds commonly experience nutrient limitations as soon as they are exposed to air and suffer from desiccation (Brawley and Johnson 1991, Davison and Pearson 1996). Algal propagules are susceptible to stress and their physiological responses to emergence are clearly related to the subsequent distribution and abundance (Brawley and Johnson 1991, Vadas et al. 1992, Davison et al. 1993). Crowded germlings are able to survive under desiccation conditions that would normally kill well-spaced individuals. This phenomenon has been tested on the shore (Ang and De Wreede 1992, Andrew and Viejo 1998) and in simulated tidal regimes in the laboratory (Hruby and Norton 1979). Crowding negatively affects germling performance under favorable growth conditions whether they are in monoculture or mixed with another species (Reed et al. 1991, Creed et al. 1996, Choi and Norton 2005a, 2005b).

It is generally known that upper shore algae, both in the juvenile and adult stage, are better adapted to emersion stresses than the plants inhabiting the lower shore (Chapman 1995, Davison and Pearson 1996). One can reasonably speculates that germlings of the upper shore species may be more tolerant for the stress than those of lower shore algae, but this has not been experimentally tested. It is of particular interest how the physiological responses of germlings differ between species and whether the outcome of competition in mixtures of species is changed by the degree of the stress. There is evidence that the outcome of competition in mixtures of two species in the adult stage has been altered by temperature, irradiance and nutrient availability (Enright 1979, Fujita 1985, Peckol and Rivers 1995). A few interspecific competition studies have been conducted at the germling stages (Choi and Norton 2005a, 2005b).

Stress tolerance of seaweeds to emersion stress is also age-dependant (Brawley and Johnson 1991, Davison et al. 1993). Silvetia compressa[formerly Pelvetia fastigiata (J. Ag.) De Toni] showed better survival at 6-h-old- and 1-week-old-germlings than 12, and 48 h old germlings during natural emersion trials (Brawley and Johnson 1991). The emersion stress for the embryos and germlings of S. compressa resulted in the reduction of photosynthesis and rhizoid development (Davison et al. 1993). Thus, the growth of seaweed germlings is also a good indicator representing the emersion stress degree experienced and predicting the competitive interactions between species at different germling ages.

The vertical distribution pattern of fucoid algae is consistent on many rocky shores in northern Europe: Pelvetia canaliculata (L.) Dcne. Et Thur. at the very top of the shore followed by Fucus spiralis L., F. vesiculosus L. and / or Ascophyllum nodosum (L.) Le Jolis with F. serratus L. at the lowest levels (Lewis 1964). On the shore, mixed stands of upper shore species, P. canaliculata and F. spiralis are occasionally found, even though their reproductive period are overlapped in summer (Knight and Parke 1950). These species are exposed periodically to air and emersion is an important stress, especially for P. canaliculata and F. spiralis. Thus, P. canaliculata and F. spiralis are suitable species to examine the effects of interactions between varying degrees of desiccation stress and the density of the two competitive species. We tested the following hypotheses: 1) P. canaliculata is more tolerant to desiccation than F. spiralis and the former outgrows the latter under desiccation stress. 2) Crowding of germlings protects them from desiccation, irrespective of the species involved.

Twenty fertile plants of both F. spiralis and P. canaliculata were collected separately from Port St. Mary ledges (54^o0' N, 4^o44' W) on August 22, 1999. Receptacles were cut from the male and female plants, washed several times using filtered seawater, dried for 5 h, and submerged in filtered seawater to induce gamete release. After 1 h, the receptacles were removed and clean egg suspensions were prepared as described by Creed et al. (1996).

To create uniform settlement density, 5 mL of zygote suspension (60 zygotes mL^-1) of each species was inoculated into two Petri dishes for each of six treatments (control, 6, 12, 24, 48, and 72 h). Each dish contained eight glass slides cut to 2.5 × 2.0 cm and 30 mL of autoclaved seawater. After 2 days, 12 slides (from the 16 in each treatment) on which germlings were most evenly distributed were chosen; four were used to determine settlement density, and eight were used for the experiment. For the culture experiment, four replicate dishes were used, each containing six slides, one from each treatment, and 50 mL of culture medium (Kain and Jones 1964). Germlings were cultured at 10 ± 1ºC , 120 ㎛ol photons m^-2 s^-1 and a photoperiod of 16 h : 8 h LD. The culture medium was changed every 3 days.

The effects of germling age on their mortality and growth were examined after moving the slides with attached germlings from the culture dishes into a desiccation tank (33 × 22 × 21 cm) with a relative humidity of 70-90%. Germlings were moved to the desiccation tank after culturing germligs for 2- or 7-days in the dishes. After the desiccation period (6, 12, 24, 48, and 72 h), the slides were returned into the culture dishes and cultured for 16 days. Control slides were continuously cultured in culture dishes.

Settlement density was determined by counting the number of settled germlings within four 4 mm² areas on each slide. Settlement density of germlings was ca. 30 germlings cm^-2. Frond lengths of 25 germlings were measured for each species and each slide. Germling mortality was estimated by counting 100 germlings on each slide after 16 days.

Twenty fertile plants of F. spiralis and P. canaliculata were collected separately from Port St. Mary ledges on 8th August 1999. For each species, four different zygote concentrations (10, 100, 600, and 1,500 zygotes mL^-1) of each species were prepared. Each zygote suspension (5, 5, 10, and 20 mL) was inoculated into three Petri dishes for four different concentrations in order to make different settlement densities (Table 1). The Petri dish contained eight slides (2.5 × 2.0 cm) and 30 mL of filtered seawater. Mixtures were made by inoculating half zygote suspension volume for each species into the same Petri dish so that their combined density was similar to that of monoculture.

After 2 days, 20 slides (from the 24 in each density treatment) on which germlings were most evenly distributed were chosen: four were used for determining density and 16 for the experiment. For the culture experiment, four replicate tidal tanks were used, each containing 16 slides, four from each treatment. The four slides were placed on four steps each exposed to air for different periods (none, 6, 8, and 10 h / 12 h tidal cycle) in 4 tidal tanks see below. Thus, each step had twelve glass-slides (4 density × 3 proportion levels) as described in Fig. 1. Culture conditions were 10 ± 1ºC, 16 h : 8 h LD, 120 ㎛ol photons m^-2 s^-1 and the culture medium was not changed for a period of 16 days.

Settlement density of germlings was determined after setting up the experiment and the lengths and mortalities of germlings were measured after 16 days as described above. The mean settlement density of four replicates is shown in Fig. 1. There were significant differences in density (analysis of variance [ANOVA], p ？ 0.05), but not between proportions of each species within a density.

Artificial tidal tanks (25 L) were prepared for periodic submergence experiment (Fig. 2). Four tanks were placed on a shelf in a culture-room and connected with two tanks each containing 125 L of culture medium. An aquarium pump (Aquaclear Power Head 400; Hagen, Germany) was placed in the reservoir tank from which seawater levels of the tidal tanks were gradually increased when the pump operated for 6 h twice per day (6-12 AM and 6-12 PM). An automatic timer was used to control water pump operation. Water level decreased gradually when the pump stopped and water drained out continuously at a rate of

60 mL per minute through outlet pipes (Fig. 2). Thus, two tidal cycles per day were simulated in the tidal tanks.

A perspex “staircase” was placed on the bottom of each tank. The lowest of the five steps was submerged continuously and the duration of air exposure increased by increasing the step height (none, 4, 6, 8, and 10 h / 12 h tidal cycle). Relative humidity of tidal tanks was 70-90%.

Data were analyzed using one-way and two-way ANOVA. Homogeneity of variances was tested by Cochran’s test. Where necessary, data were transformed before analysis to meet the assumptions of parametric tests (Sokal and Rohlf 1995). The significance of the differences between means was tested with the Tukey HSD test.

Sixteen days after settlement, the mortality of F. spiralis was 11-30% for 2-day-old germlings and 11-25% for 7-day-old germlings, respectively (Fig. 3). The mortality of Fucus germlings significantly increased with the emerged period but it was not significantly different between germling ages, even though the mortality of germlings was slightly decreased at 7-day-old germlings (Table 2). In Pelvetia, the mortality of germlings was 15-23% for 2-day-old germlings and was between 15-25% for 7-day-old germligs (Fig. 3). P. canaliculata mortality was minimal for the 12 h exposure treatment in 2-day-old germlings and at the 6 h exposure for 7-day-old germlings. It was not significantly different between germling ages (two-way ANOVA, F_{1, 36} = 1.38, p = 0.25) and between emerged periods (two-way ANOVA, F_{5, 36} = 2.39, p = 0.06). Furthermore, mortality was not significantly different between germlings of the two species (ANOVA, F_{1, 94} = 0.02, p = 0.89).

Zygote diameter was 106.66 ± 11.28 μm (mean ± standard error, n = 60) for P. canaliculata and 77.58 ± 4.35 μm (n = 60) for F. spiralis. The growth of both species was retarded with increasing exposure period and as they were air-exposed at earlier stage (Fig. 4). After 16 days, the length of Pelvetia germlings was 125-140 μm for 2-day-old germlings and 134-140 μm for 7-day-old germlings. The length of Fucus germlings was ranged from 184 to 278 μm for 2-day-old germlings and between 211-277 μm for 7-day-old germlings. These data indicate that Fuscus grew faster than Pelvetia and that the effects of temporary exposure period and germling age emerged on the growth of germlings are more sensitive in F. spiralis than P. canaliculata.

Mean relative growth rate of F. spiralis was 0.070 day^-1 for 2-day-old and 0.072 day^-1 for 7-day-old germlings and that of P. canaliculata was 0.015 day^-1 for 2-day-old and 0.016 day^-1 for 7-day-old germlings. The relative growth rate of F. spiralis was much greater than that of P. canaliculata in all experimental treatments.

The relative growth rate was significantly greater when germlings were exposed to air after 7 days rather than after 2 days in F. spiralis (two-way ANOVA, F1, 36 = 11.76, p ？ 0.01) but not for P. canaliculata (two-way ANOVA, F_{1, 36} = 2.04, p = 0.16). With increasing air exposure, the relative growth rate decreased significantly in both species (F. spiralis: two-way ANOVA, F_{5, 36} = 68.60, p ？ 0.001 and P. canaliculata: two-way ANOVA, F_{5, 36} = 11.60, p ？ 0.001). A significant interaction between germling age and temporary exposure period with respect to the relative growth rate was found for both species (F. spiralis: two-way ANOVA, F_{5, 36} = 3.51, p ？ 0.05 and P. canaliculata: two-way ANOVA, F_{5, 36} = 3.82, p ？ 0.01). For F. spiralis, a Tukey test revealed that there were no significant differences in growth rate between the control and 6 h emergence, the 6 and 12 h, and the 12 and 48 h, but significant differences

were found among the other treatments. Significant differences for P. canaliculata were observed between 72 h and the other treatments but not for the intermediate duration (Tukey test).

Germling mortality was not significantly affected by the density of germlings or by the periodic exposure periods of the two species 16 days after settlement. Mortality was 8.3-15.6% for F. spiralis and 10.1-15.9% for P. canaliculata in the monocultures and mixed cultures.

In monoculture, F. spiralis grew significantly faster than P. canaliculata under all exposure periods and at the four different densities (one-way ANOVA, F_{1, 126} = 298.44, p ？ 0.001) (Table 3).

The growth of both species in monoculture was influenced by germling density and by periodic exposure period. After 16 days in culture, the mean lengths of germlings were 209-273 μm for F. spiralis and 159-227 μm for P. canaliculata. However, the effect of settlement density on germling growth varied with increasing periodic exposure period (Table 3). Settlement density and F. spiralis growth was negative in the control and 6 h treatment but positive at the 8 and 10 h exposure treatments. A negative density effect on P. canaliculata growth appeared at up to 8 h of exposure but growth was higher at a high density when germlings were exposed to air for more than 10 h. Due to such different density effects, germling growth was not significantly different between F. spiralis density levels (two-way ANOVA, F_{3, 48} = 2.64, p = 0.06) but it was significantly higher in the control, and 6 and 8 h treatments than that in the 10 h treatment (two-way ANOVA, F_{3, 48} = 55.36, p ？ 0.001). Interactions between density and emerged period were found (two-way ANOVA, F_{9, 48} = 17.72, p ？ 0.001). The effects of density on P. canaliculata growth were not significantly different among density levels (F_{3, 48} = 35.68, p ？ 0.001) or exposure treatments (F_{3, 48} = 107.68, p ？ 0.001) but interactions were found (F_{9, 48} = 9.20, p ？0.001). Mean germling lengths were significantly different between all exposure treatments and were significantly greater at densities of 10 and 100 than at densities of 1,000 and 7,000 germlings (Tukey test).

In the mixtures, germling lengths were 240-282 μm for F. spiralis and 131-193 μm for P. canaliculata. Germling growth was significantly influenced by the periodic exposure period in F. spiralis (two-way ANOVA, F_{3, 48} = 38.18, p ？ 0.001) and in P. canaliculata (two-way ANOVA, F3, 48 = 5.44, p ？ 0.01). F. spiralis grew better when exposed to air

for 6 or 8 h compared to the control and 10 h treatments, and no significant difference was observed between the 6 and 8 h or between the control and 10 h treatments (Tukey test). Growth was significantly greater when P. canaliculata was grown under exposed conditions than under continuously submerged condition (Tukey test).

The density and growth relationship changed from negative to positive by increasing the duration of emergence (Fig. 5), resulting in no statistically significant influence of density on F. spiralis growth (two-way ANOVA, F_{3, 48} = 1.04, p = 0.38). However, the effect of density on P.

canaliculata growth was significantly different (two-way ANOVA, F_{3, 48} = 22.83, p ？ 0.001). A multiple comparison revealed that P. canaliculata grew faster at lower densities than at a density of 7,000, but no differences were found at densities from 10 to 1,000 germlings cm^-2.

The relative importance of intra- and interspecific competition was compared with respect to the growth of both species under the various emerged periods (Fig. 6). The mean lengths of the four densities within each treatment were pooled for the comparisons. Fucus spiralis grew significantly faster in mixtures than in monocultures (two-way ANOVA, F_{1, 24} = 8.53, p ？ 0.01), whereas P. canaliculata grew better in monoculture (two-way ANOVA, F_{1, 24} = 5.11, p ？ 0.05).

Over the range of density levels, germling growth was also significantly affected by duration of emergence for F. spiralis (two-way ANOVA F_{3, 24} = 6.66, p ？ 0.01) and for P. canaliculata (two-way ANOVA, F_{3, 24} = 6.04, p ？ 0.01). The growth of F. spiralis was significantly retarded at the 10 h exposure more than at any other treatment, and that of P. canaliculata was significantly highest at 6 h exposure in monoculture (Tukey test). Interestingly, P. canaliculata growth was higher in monoculture when the plants were grown in a submerged condition, whereas it was enhanced in mixtures exposed for 10 h.

Pelvetia canaliculata well adapts to desiccation stress better than any other fucoids in the visible stage both on the shore and in the laboratory culture. After severe desiccation, P. canaliculata can survive well (Schonbeck and Norton 1980), take up nutrients (Hurd and Dring 1990), recover photosynthesis (Hurd and Dring 1991) and grow well (Stromgren 1977) better than other fucoids. Furthermore, P. canaliculata has volemitol, a carbohydrate that can be used during desiccation stress (Pfetzing et al. 2000). In the present study, germlings of the high-shore species, P. canaliculata exhibited a high tolerance to desiccation as shown in the adult plants (Schonbeck and Norton 1978, 1979, 1980). Although the superior drought tolerance of P. canaliculata compared to F. spiralis was tested, the mechanism was not examined. P. canaliculata zygotes are surrounded by mesochiton and thick cell walls, which are absent in Fucus spp. (Moss 1974, Hardy and Moss 1979). The mesochiton with amorphous mucilage is broken when rhizoids of zygotes differentiate, and P. canaliculata zygotes probably adapt well under desiccation stress. Whatever the mechanism, the different desiccation tolerance between germlings of the two species may determine their distribution and abundance on the shore.

Emersion stress inhibits photosynthesis and retards embryonic development, and the ability of Silvetia compressa (formerly Pelvetia fastigiata) germlings to withstand emersion stress increases with age (Brawley and Johnson 1991, Davison et al. 1993, Pearson et al. 2000). In the present study, P. canaliculata and F. spiralis grew less well when they were exposed to air after settlement for 2 days rather than 7 days. However, even if the exposure duration is the same, this result confirms that desiccation tolerance of germlings increased with age. These germling growth differences by germling age (2-day or 7-day-old) may result from the recovery period. For example, the ability of Silvetia compressa to recover from desiccation is greater in 7-day-old germlings than that in 8 h zygotes (Davison et al. 1993). The growth and mortality of germlings were influenced significantly by exposure duration even within the same exposure period. Thus, successful recruitment of the two high shore species, P. canaliculata and F. spiralis, may coincide with periods of cool or cloudy weather affecting exposure time of germling and with the spring or neap tides, which determine exposure duration.

Under emerged stress, crowded germlings survive well by mutual protection compared to germlings that are sparsely settled (Hruby and Norton 1979). The present results show that the survival of F. spiralis and P. canaliculata germlings was higher at a high density than at a low density, indicating that crowding is a major means to survive under emersion stress in both species and that such mutual protection may operate irrespective of species. Both species grew faster when exposed to air for 6 h per tidal cycle (12 h) than when they were submerged in water. This result is not surprising because P. canaliculata adults show necrosis after prolonged submergence (Schonbeck and Norton 1979). Thus, F. spiralis and P. canaliculata can inhabit high shore not only because they have ability to withstand emersion stress but also because protect themselves from the stress by crowding.

The outcome of interspecific competition between F. spiralis and P. canaliculata germlings was slightly altered by exposure period, but not to such an extent as to change the outcome. On the shore, there are many additional stressful factors (i.e., high temperature and irradiance). Thus, germling growth and survival are affected by the combined effects of such factors but we tested the effects of individual environmental factors separately. Clearly, this study has shown that the responses of germlings to environmental stresses are different between fucoid species but it would be premature to claim that such responses result in their distribution on the shore.