Synergistic effects of ultraviolet radiation and conditions at low tide on egg masses of limpets (Benhamina obliquata and Siphonaria australis) in New Zealand
- First Online:
- Cite this article as:
- Russell, J. & Phillips, N.E. Mar Biol (2009) 156: 579. doi:10.1007/s00227-008-1109-6
The effects of ultraviolet radiation (UVR), desiccation and conditions in tidal pools on embryonic survival were examined for two common pulmonate limpets that lay intertidal benthic egg masses on rocky shores in New Zealand: Benhamina obliquata and Siphonaria australis. Field surveys and manipulative experiments were conducted between December 2006 and September 2007 in the Wellington region of New Zealand (41°17′S, 174°47′E). Egg mass deposition sites in the field were species-specific: B. obliquata deposited eggs primarily in shaded crevices, whereas S. australis predominantly deposited egg masses in the sun and in tidal pools. For both species, however, embryonic mortality was greater in egg masses that had been in full sun compared to shade. For S. australis, there was also high mortality in egg masses in tidal pools or desiccated compared to those that remained submerged in flowing seawater at low tide. In outdoor experiments, embryonic mortality was also always greatest for egg masses exposed to full sun, and lowest for those in shaded treatments. Mortality was also higher if egg masses were in simulated tidal pools, and for S. australis, if desiccated, compared to those submerged in flowing seawater. Periods of particularly sunny conditions with high temperatures also resulted in higher overall mortality. Finally, egg masses of both species that were initially deposited in the shade had greater mortality in response to subsequent UV exposure compared to egg masses initially deposited in full sun. Results from this study suggest that the egg masses of these two species are highly vulnerable to UVR, as well as other intertidal stressors. Embryos of both of these species may be at risk of high mortality particularly during summer when extreme conditions of UV intensity and high temperature coincide with low tide cycles.
Ultraviolet radiation (UVR, 280–400 nm) can cause a range of deleterious effects on aquatic organisms, with early life stages at particular risk (reviewed by Hader et al. 2007). For example, exposure to UVR inhibits photosynthesis and causes DNA damage to kelp zoospores (Wiencke et al. 2000). In yellow perch, 100% of eggs die when exposed to full solar UV (Huff et al. 2004), and developing sea urchin embryos cells self-destruct in response to UV damage (Lesser et al. 2003). UV-induced mortality in early life stages can have important implications for population dynamics of aquatic species; the global decline of amphibian populations may be due in part to mortality caused by UV exposure during embryonic development (Blaustein and Kiesecker 2002).
Human-induced stratospheric ozone loss has resulted in elevated levels of UVR reaching the Earth’s surface in recent decades, particularly for the mid-high latitudes of the Southern hemisphere which were already exposed to higher levels of UV (McKenzie et al. 1999). In New Zealand and Australia UVR can be twice as high in summer compared to similar latitudes in Europe and North America (Seckmeyer and McKenzie 1992; Gies et al. 2004). Organisms in New Zealand may therefore be particularly vulnerable to the damaging effects of UVR, yet few studies have focused on risks to New Zealand’s marine fauna.
A variety of marine species deposit benthic egg masses throughout the intertidal zone, and recent research has been directed toward understanding what mechanisms allow early life stages to persist in this hazardous environment (reviewed by Przeslawski 2004). Although adults can spawn in relatively stable environments where their offspring’s risk from stressful environmental factors is reduced, many species routinely deposit their egg masses in sites where embryos experience periods of desiccation, elevated UVR, high temperature and osmotic stress (Benkendorff and Davis 2004). These environmental stressors can act synergistically to increase embryonic mortality in benthic egg masses on rocky shores at low tide (Przeslawski 2005a; Przeslawski et al. 2005; Przeslawski and Davis 2007). Tidal pools may be a refuge from desiccation stress on rocky shores, however, a number of other environmental stressors likely occur in this microhabitat, particularly in small volume pools, including extremes in salinity, water temperature and oxygen availability (reviewed by Metaxas and Scheibling 1993).
In the present study, we examined the effects of UVR, desiccation and simulated tidal pool conditions on the embryonic mortality of two species of common pulmonate limpet in New Zealand (Benhamina obliquata and Siphonaria australis). These limpets inhabit and deposit benthic egg masses on exposed and semi-exposed rocky intertidal shores throughout New Zealand (Jenkins 1981). Siphonaria australis is a homing species with a broad vertical distribution across the low, mid and upper zones and overlapping with B. obliquata’s range at the mid to high intertidal level (Borland 1950; Morton and Miller 1968). Both species are hermaphroditic and produce gelatinous benthic egg masses over the austral spring and summer months. Benhamina obliquata deposits relatively large, stiff, spiral egg masses with an estimated 45,000–190,000 eggs per mass and 2–3 egg masses per individual over the breeding season (Borland 1950). By contrast, S. australis deposits numerous small, soft ribbon-shaped egg masses (Knox 1955), and produces approximately 350,000 individually encapsulated embryos per season (Creese 1980). Both B. obliquata and S. australis have a mixed life history where embryos spend 5–8 days developing on the benthos, followed by up to several weeks as planktotrophic veligers (Borland 1950; Knox 1955).
Our study took place at the southern tip of the North Island of New Zealand, in the Wellington region, where the topography of the rocky coast is very heterogeneous, characterized by a great deal of vertical relief, crevices, channels, walls and tidal pools of varying sizes. Year-round, including over summer, low spring tides frequently occur from mid-morning through mid-day, with aerial exposure of the mid-intertidal lasting about 4 h on average (Phillips, personal observation). Both B. obliquata and S. australis occur in this region of New Zealand, and their egg masses are commonly found from mid to late spring through summer (Morton and Miller 1968).
Surveys were conducted to quantify egg mass deposition microhabitats in the field and any patterns of embryonic mortality associated with different microhabitats. We also conducted manipulative experiments to determine the separate and combined effects of these conditions on embryonic mortality of B. obliquata and S. australis. We hypothesised that (1) embryos would suffer higher mortality when exposed to full sun compared to shade, (2) that mortality due to light exposure would be higher than mortality due to desiccation or tidal pool conditions and (3) due to differences in the egg masses, the responses of these two species to light and low tide conditions would be different.
Because UVR intensity varies greatly with time of day, season, and cloud cover, we conducted experiments at different times of year to examine whether seasonal variability in climatic conditions might provide refuge from these stressors. Finally, there was evidence that species depositing egg masses in sunny microhabitats may have innate protective mechanisms and lower susceptibility to UVR, than species exclusively depositing their eggs in shaded microhabitats (Przeslawski et al. 2004). Potential differential mortality due to site of egg mass deposition has not been examined in species that lay egg masses in both sunny and shaded microhabitats. We therefore also examined the effect of initial sun exposure of egg masses on embryonic mortality for both species.
Materials and methods
Surveys were conducted during low spring tides, at six semi-exposed to exposed rocky intertidal sites in the Wellington region of New Zealand between 41°01′–41°20′S and 174°42–174°54′E (Makara Beach, Pukerua Bay, Moa Point, Island Bay, Point Halswell and Point Howard). At each site we established a belt transect in the mid-intertidal zone, 2 m wide and 30 m long, parallel to shore. These transects spanned most of the intertidal zone (i.e., from the relatively high barnacle zone to just below the water line at low tide during spring tides), which is vertically compressed. The tidal excursion is <1.5 m during spring tides in the Wellington region. We recorded the species identity (B. obliquata or S. australis), light exposure (sunny vs. shaded), and the microhabitat of deposition for every egg mass in the transects. Light exposure was recorded at the time of sampling (i.e., late morning to early afternoon) and therefore reflects conditions around mid-day when UV intensity is greatest.
Egg mass microhabitats were assigned to one of three categories: desiccated on rock substrate, submerged, or in a tidal pool. Egg masses in the submerged category were deposited low enough on the shore to remain covered by water at low tide during spring tides. Egg masses were collected from each combination of light exposure and microhabitat to assess embryonic mortality. B. obliquata egg masses occurred at four sites, but only in both sun and shade at three sites. For S. australis ten egg masses were collected from each combination of light exposure and microhabitat, from each of six sites. We brought egg masses to the Victoria University Coastal Ecology Laboratory (VUCEL) and examined each at 100× magnification on a compound microscope. We scored the first 60 embryos in each egg mass as either viable (i.e., alive, healthy and developmentally normal in appearance) or inviable (i.e., malformed and grossly abnormal in appearance, often a mass of unidentifiable tissue or cells). Surveys were conducted in December 2006; the same survey methods were repeated again in February 2007 for S. australis at Island Bay only during a period of particularly warm and sunny weather.
Experiment one—light exposure and low tide conditions
In this experiment we examined the effect of light exposure and simulated low tide conditions on embryonic mortality of both B. obliquata and S. australis egg masses using the same methods but at different times.
Egg masses were collected from shaded microhabitats in the field, and examined to ensure that only those still at early stages of embryonic development were used in the experiment (i.e., gastrula stage or earlier). Six egg masses, separated by substantial distances and assumed to be deposited by different individuals, were collected (except in the September run of S. australis, in which only three were collected). Each egg mass was divided into nine equal-sized sections and each section was placed in its own 100 ml plastic container, to ensure each egg mass was represented in all treatments. We drilled numerous small holes in each container to ensure egg masses were retained, but water was free to circulate. For the light exposure treatment there were three levels: full sun, UV-blocked, and completely shaded. Three segments of each egg mass were randomly allocated to each treatment. Light exposure treatments were established by affixing pieces of plastic with the appropriate UV-absorbing properties to each container as lids and around the outside. The containers themselves were transparent to UV penetration; each absorbing plastic was tested using a UV-1601 Shimadzu spectrometer. These containers were then placed in a shallow seawater aquarium with continuously flowing seawater (approximately 5.0 cm deep) in an outdoor location at VUCEL. Seawater was first filtered through a 10-μm bag filter to remove large particles.
We established three levels of simulated low tide conditions as well: constantly submerged in flowing seawater, desiccated for 3 h to simulate aerial exposure at low tide, and 3 h exposure to non-circulating heated seawater to simulate a tidal pool. From each egg mass, one of each of the three sections in each light exposure treatment was randomly allocated to one of these low tide conditions. At 1100 hours each day the containers exposed to 3 h of desiccation were placed in a dry aquarium immediately adjacent to, and similar in all respects to, the aquarium with flowing seawater. The containers subjected to simulated tidal pool conditions were placed in a third aquarium, again immediately adjacent, to which was added seawater that had been allowed to heat in a black tub exposed to full sun for approximately 3 h prior to the experiment. The remaining containers were left submerged in the aquarium with flowing seawater, to act as a control, but were momentarily lifted and replaced to mimic the movement of the other containers to their treatments. At 1400 hours the desiccation and tidal pool treatments were returned to the aquarium with flowing seawater. All positions of containers in all aquaria were random and changed daily. The daily assignment of aquaria to either the desiccated or tidal pool treatment was also random. Each experiment ran for 5 consecutive days, the point at which hatching was imminent but had not yet occurred.
To analyze the embryonic mortality of each egg mass we followed methods similar to Przeslawski et al. (2004). Each segment of egg mass was placed in a vial with 400 μl of filtered seawater and was agitated by hand for 2 min to suspend the embryos from egg mass. From this ten 5 μl aliquots were analyzed microscopically and embryos were scored as viable or inviable, as defined above except in this case all viable embryos were protoconch veligers. The analysis of embryonic mortality was made on samples fixed in 5% buffered formalin for B. obliquata in experiment one, but on un-fixed embryos in the other experiments to make sure that the veligers were alive. Experiments were conducted on 16–20 December (Run 1) and repeated 11–15 February 2007 (Run 2) for B. obliquata and 6–10 February 2007 (Run 1) and 24–28 September 2007 (Run 2) for S. australis.
Experiment two—initial light conditions and light exposure during development
In this experiment we examined the effect of initial light exposure of egg masses at the site of deposition and light exposure during development on subsequent embryonic mortality of both species in one factorial design. Benhamina obliquata and S. australis egg masses were collected from shaded or sunny microhabitats. Six egg masses at early stages of development were selected for each species from each microhabitat, and each was divided into three sections. We established experimental light exposure treatments and scored embryonic mortality at the end of 5 days using identical methods as experiment one. This experiment was conducted once, 12–16 January 2007.
Temperature and UV data
We obtained air and water temperatures, and UV Index (UVI) data during the field surveys at Island Bay (7 December 2006, 9 February 2007), and during the experiments. The UVI is a scale linearly related to UV intensity, weighted by the McKinlay–Diffey Erythema (sunburn) action spectrum. This scale has been adopted by the World Health Organization to represent the risk of skin damage from UV exposure. Scale values range from low (0–2) to medium (3–5), high (6–7), very high (8–10) and extreme (11+). UV measurements were obtained from Industrial Research Limited (IRL), which maintains a UV radiometer <5 km from VUCEL, which records the UV Index every minute. We calculated the mean maximum daily UVI and mean UVI from 1000 to 1600 hours over 5 day preceding each survey. We report both because it is unclear whether responses to UV are related to the maximum intensity or average intensity experienced.
For the surveys, air temperatures were noted at five times over the course of several hours during the survey, and the water temperatures of tidal pools from which egg masses were sampled were also recorded (n = 10 tidal pools on each date). For each experiment, water temperatures of all treatments, and the air temperature, were recorded at 1400 hours every day.
For the field surveys, the percentage of egg masses in the different light exposures and microhabitats was analysed statistically in SPSS using a non-parametric chi-square test for both B. obliquata and S. australis, combining data from all sites for each species. The percentages of inviable embryos in the different microhabitats were analysed using ANOVA, with the fixed factors light exposure (two levels) and for S. australis only, microhabitat (three levels), and site (three levels for B. obliquata, six levels for S. australis) as a random factor. Because in both cases there was no effect of site, nor any interaction of site with exposure or microhabitat (Supplementary material S1), we omitted site as a factor from the final model. An additional ANOVA was done for S. australis egg masses from Island Bay only, which was surveyed twice, and month (two levels) was added as a fixed factor to the model. We considered month a fixed factor because February was deliberately chosen to conduct surveys in mid-summer following several particularly warm, sunny days, and contrasted this with December (early summer) when weather conditions were less extreme (see Supplementary material S2).
Experimental data for each species were also analyzed using ANOVA. For the first experiment the factors were light exposure treatment (three levels), low tide conditions (three levels), and experimental run (two levels), all fixed factors. Experimental run was considered a fixed factor because we deliberately chose times with different weather conditions for the experiments (see Supplementary material S2). The parent of each egg mass was included as a random factor to account for potential variability in embryonic mortality among parents. In the model for the second experiment the factors were species (two levels), initial light exposures at the site of egg mass deposition (two levels) and light exposures during development (three levels) as well as parent (random factor).
Percentage data were arcsine square root transformed (as recommended by Zar 1998) and transformed data fulfilled assumptions of homogeneity. Significant effects and interactions were further explored with post-hoc Tukey HSD tests.
Percentage of limpet egg masses deposited in different intertidal microhabitats on rocky shores in central New Zealand
By contrast, a significantly greater proportion of S. australis egg masses were found in the sun compared to the shade (Table 1, χ2 = 73.38, P < 0.001). A greater proportion of S. australis egg masses were also found in tidal pools compared to desiccated or submerged (Table 1, χ2 = 70.04, P < 0.001), however, substantial proportions of egg masses (i.e., >20%) were found in all three microhabitats. Although the greatest proportion of S. australis egg masses were deposited in tidal pools, this was the microhabitat where the greatest embryonic mortality occurred, followed by those that were desiccated at low tide (Fig. 1, F2,355 = 34.216, P < 0.001, Tukey HSD P < 0.001). Similar to B. obliquata, regardless of microhabitat, embryonic mortality was significantly greater in egg masses in the sun compared to the shade (Fig. 1, F1,355 = 30.935, P < 0.001), with no interaction between microhabitat and light exposure (F2,355 = 0.620, P = 0.539).
Experiment one—light exposure and low tide conditions
Results of ANOVA on the effect of experimental treatments (light exposure and simulated low tide conditions), experimental run and parent on embryonic mortality of B. obliquata
Source of variation
Light exposure × tidal condition
Run × light exposure
Run × tidal condition
Run × light exposure × tidal condition
Results of ANOVA on the effect of experimental treatments (light exposure and simulated low tide conditions), experimental run and parent on embryonic mortality of S. australis
Source of variation
Light exposure × tidal condition
Run × light exposure
Run × tidal condition
Run × light exposure × tidal condition
Experiment two—initial light conditions and light exposure during early development
Results of ANOVA for the effect of species, initial light exposure at site of egg mass deposition (=Site), experimental light exposure treatment, and parent on embryonic mortality of B. obliquata and S. australis
Source of variation
Species × site
Species × light exposure
Site × light exposure
Species × site × light exposure
Temperature and UV data
For the field surveys at Island Bay, the mean seawater temperature was only slightly different between December 2006 (15.5°C) and February 2007 (17.4°C), but the air temperature (December: 14°C, February: 19°C) and tidal pool temperatures (December: 17.8°C, February: 22.6°C) differed greatly. The mean mid-day UVIs were in the very high to extreme range (December: 8.86, February: 10.85). During the experiments over summer (December 2006–February 2007) air temperature ranged from 20.4 to 27.4°C, peaking in early February. Running seawater temperatures (17.2–19°C) and simulated tidal pool temperatures (23.5–25.2°C) did not differ greatly over the experiments. However, mean temperatures were much lower in September (air: 10.2°C, seawater: 9.92°C, simulated tidal pools: 10.06°C). Mean mid-day UVIs were in the extreme range during the experiments in December and February (10.96–12.48), and were in the medium range in September (5.30). Complete temperature and UVI data can be found in Supplementary material S2.
This study demonstrates that, for two species of intertidal pulmonate limpet, egg mass deposition site is important in determining susceptibility of embryos to different environmental stressors. Benhamina obliquata predominantly deposited their offspring in shaded, desiccated crevices, which corresponded to the highest probability of survival. It is not known what drives choice of egg mass deposition site in this species. One study reports that B. obliquata egg masses are deposited at night at low tide (Borland 1950) when natural cues about exposure to sun are reduced, but the generality of this observation has not been further documented. This microhabitat may also offer protection against biotic factors as crevices high in the intertidal are often inaccessible to mobile predators (Rochette and Dill 2000).
Compared to B. obliquata, S. australis egg masses were not found predominantly in microhabitats that resulted in highest embryonic survival. Egg masses were more likely to be found in sun than shade, even though this resulted in high mortality. Further, the greatest proportion of S. australis egg masses was found in tidal pools, the microhabitat where egg masses suffered the greatest mortality. A possible obstacle to protecting benthic egg masses may be the scarcity of safe deposition sites (Wilson 1986; von Dassow and Strathmann 2005), particularly for homing species such as Siphonaria spp.; favourable sites may be scarce in the vicinity of a home scar and individuals may therefore be limited by availability. However, the heterogeneous topography of Wellington’s rocky intertidal shores creates a high availability of shaded microsites. This is supported by the fact that B. obliquata and S. australis co-occurred at several survey sites, and yet B. obliquata egg masses were more often found in shade and were not found in tidal pools. It is also possible that minimizing exposure of embryos to environmental stressors is not the primary driver of adult choice of egg mass deposition site for either species, but that it is instead more closely tied to predator avoidance, space competition or adult life history (Spight 1977).
The ability of B. obliquata to deposit offspring in favourable microhabitats may reflect the relative energetic investment in the egg mass of this species (which is relatively large, requires a long time to deposit, and only a small number are deposited by each individual each year). Therefore, there would be great selective pressure to develop behaviours that maximise protection of egg masses and thereby increase fitness. In comparison, individual S. australis produce vast numbers of small, energetically inexpensive egg masses over the reproductive season (Knox 1955). Our study shows that environmental stressors that are great in summer can be reduced in early spring, so mortality risk may be seasonal. Nevertheless, there is likely a huge cost to reproductive success due to loss of embryos exposed to the sun in summer, in particular for S. australis.
The field survey results were supported by manipulative experiments where, for both species, the overwhelming factor in mortality was exposure to UVR. Similar results have been observed among benthic egg laying gastropods in Australia (Przeslawski et al. 2004). Our study, however, is the first to demonstrate the potential cost of tidal pool immersion on benthic egg masses. Tidal pools are often considered refuges that protect against desiccation (reviewed by Metaxas and Scheibling 1993), and perhaps they are during non-summer months. It is not known whether embryonic mortality in tidal pools in this study was driven by elevated water temperatures, or other factors (i.e., potentially higher salinity, or oxygen stress), and distinguishing among these will be the focus of future work. This study also shows the importance of temporal variability in environmental stressors in the survival of benthic-developing intertidal organisms. Experiments showed that embryonic mortality was generally greatest in the middle of summer (i.e., early February) when conditions were most stressful and interacted with each other to elevate embryonic mortality.
The apparent lower susceptibility of B. obliquata egg masses to desiccation compared to S. australis (<10% compared to ~60% in February in experiment one) is likely due to species-specific egg mass characteristics. Thickness of the egg mass wall, shape and size can profoundly affect the rate of water loss and therefore the survival of the encapsulated embryos (Rawlings 1995); egg masses like those of B. obliquata with stiffer walls (Feder et al. 1982) and a coiled shape (Chambers and McQuaid 1994) are thought to reduce desiccation.
Results here show that initial exposure to sun in the microhabitat where egg masses are deposited to some degree mitigates effects of further UVR exposure for these two species. Similarly, a study in Australia found that among species, molluscs that deposit eggs exclusively in the shade have greater embryonic mortality when exposed to UVR than those that deposit eggs in the sun (Przeslawski et al. 2004). Przeslawski et al. (2004) showed that the Siphonaria sp. in their study deposited egg masses in the sun and, in contrast to this study, they found little embryonic mortality due to further exposure to UVR. A possible explanation for the differences between the two studies may be the timing of the experiments, which was not specified by Przeslawski et al. (2004), but may not have been conducted in mid-summer. In our September experiment, S. australis also had low mortality with no response to experimental manipulations of UVR.
There has been a lot of interest in the role of chemical sunscreens such as mycosporine-like amino acids (MAAs) in mitigating effects of UVR (Shick and Dunlap 2002). However, preliminary MAA analysis on the egg masses from this study (Russell, unpublished data), as well as analyses from gastropods in Australia (Przeslawski 2005b), have found no significant difference in MAA concentration between egg masses deposited in the sun versus those from the shade. Further, Wraith et al. (2006) explicitly tested embryonic mortality in egg masses with different natural levels of MAAs and found that these compounds were probably not the most important source of protection against UV damage. Alternative protective mechanisms may involve DNA repair mechanisms (Kaufmann and Paules 1996; Epel et al. 1999), and UVR absorption in egg mass walls (Rawlings 1996). These may offer explanations as to species-specific vulnerability to UVR (as in this study B. obliquata was at greater risk than S. australis), however, it is more challenging to speculate about what might drive the intra-specific differences in mortality dependent on initial sun exposure that were found here, and further study is required.
Despite the substantial increases in UVR reaching the Earth’s surface over the last 30 years, particularly in the Southern hemisphere, research on the effects on New Zealand’s marine fauna has been minimal (but see Lamare et al. 2004). Unlike other studies of the detrimental effects on early development using artificial sources and extreme levels of UV, the conclusions from this study are ecologically realistic representations of natural mortality for these two common species. Further, geographical differences in the timing of summer low tides coinciding with the middle of the day, when conditions are most extreme, mean that organisms in some locations are exposed to much higher thermal stress than others (Helmuth et al. 2002). In New Zealand, because summer low tides commonly fall in the middle of the day, organisms are not only subjected to extremes of thermal stress, but also to high levels of UVR. This suggests that intertidal fauna in New Zealand and Australia may be particularly vulnerable to synergistic effects of UVR and other stressors, particularly in light of predicted climate change.
We thank C. Smith for field assistance, and IRL and for the use of their UV data. This work benefited from discussions with K. Ryan and funding in the form of a scholarship to J. Russell from Victoria University of Wellington. The methods of the study reported here comply with the current laws of New Zealand.