Abstract The barnacle Tetraclita squamosa rubescens exhibits a high degree of variation in its shell morphology. Research on other barnacle species suggests that a variety of factors, particularly whelk predation and temperature stress, may influence the nature of this morphological variation. By sampling T. squamosa rubescens in several sites in Monterey and Pacific Grove, California, we established that barnacle morphology varies significantly with respect to whelk predation and the vertical orientation of the barnacles' substrate. Barnacles are taller on vertical substrate, and barnacles in areas of high whelk predation have more enclosed, recessed opercular plates. By exposing whelks to barnacles with artificially exposed opercular plates in caged experimental treatments, we found that the whelks preferentially prey on barnacles with exposed opercular plates. Our predation field experiment corroborates the field sampling data insofar as the exposure of the opercular plates decreases with increased whelk abundance, and decreased exposure of the opercular plates significantly decreases the risk of whelk predation. Moreover, by measuring the heat flux through barnacles with different shapes, we ascertained a significant difference in the temperature budgets of taller, narrower barnacles versus barnacles with a shorter, broader shape. It appears that barnacles with tall, narrow shells shed heat to the air more rapidly. These observations and experiments support the hypothesis that the shell morphology of the Tetraclita barnacle has significant effects on its susceptibility to predation and temperature change. Introduction Predation and heat stress are two factors in an array of physical and biological influences that determine the optimal morphology of an organism in its environment. Morphological variation within a population may suggest different conditions among microenvironments that favor one shape over another. For organisms in the rocky intertidal zone, temperature stress and predation have been posited to be significant selective pressures in determining variation in morphology. Lively (1986) suggested that predatory whelks influenced the morphology of their barnacle prey, such that chemical cues from the whelks caused Chthamalus to assume a "bent' morphology that protected it from drilling by whelks. The morphological variation was maintained by the fact that, in the absence of whelks, the bent morphology reduced fecundity. Hence, morphology is often a compromise between contradicting environmental pressures. In Monterey Bay, California, the barnacle Tetraclita squamosa rubescens exhibits a high degree of morphological variability. One characteristic that varies significantly is the distance between the barnacle'’s opercular plates and the lip of the shell aperture (hereafter referred to as opercular depth). The whelk Nucella emarginata preys on Tetraclita by sitting on the aperture of a barnacle and drilling through the opercular plates to access the internal tissue. Thus, deeply recessed opercular plates may provide a refuge from whelk predation. In areas of high whelk abundance, Tetraclita may benefit in a manner similar to the bent Chthamalus of Lively (1986) by having a shape that decreases the success of whelk predation. Within Monterey Bay, another highly variable aspect of Tetraclita shell morphology is the height of the barnacle in relation to its basal diameter. Some are tall and narrow while others are short and broad-based. Yip and Shin (1990) observed similar variation among Tetraclita squamosa japonica in the Bay of Hong Kong. They hypothesized that broad Tetraclita were better insulated and therefore more resistant to temperature stress. However, in the absence of excessive temperature stress, they speculated that tall, less insulated barnacles had more internal space for gonad development and were therefore more fecund. In this study we examined the possible adaptive advantages conferred by the variety of morphologies we observed in the field. We did this by quantifying the morphological variation between sites that differed in temperature regimes and predation pressure and by experimentally manipulating barnacle morphology and evaluating subsequent susceptibility to predatory whelks. Materials and Methods We performed all of our field observations and experiments at two study sites in Pacific Grove, California and Monterey, California between April and June of 2001. The first study site was located on the wave-exposed northeast rocky shore of China Point at Hopkins Marine Station, and the second was a short granite outcrop in the wave-protected harbor behind the Coast Guard Breakwater. Both sites were primarily composed of large granite boulders. Tetraclita were common in both sites. Nucella emarginata was abundant at the Hopkins site but entirely absent from the Breakwater site. At each site, we measured spatial variability of barnacle morphology and conducted a field experiment to examine the relationship of morphology and susceptibility to whelk predation. Transects At each site we sampled morphological data from the center of Tetraclita’s vertical range. At Breakwater, this height was approximately 0.7 meters above MLLW and at Hopkins, it was 1.5 meters above MLLW. At each site, we laid out 10 meter transects parallel to the contour of the shore, and at each meter we counted whelks within a square meter quadrat. We randomly placed a 0.25-m' quadrat within each successive meter of transect. In each quadrat we sub-sampled barnacles in one lOcm by lOcm subdivision of the large quadrat. We used calipers to measure the opercular depth, the width of the aperture (= opercular width), the maximum and minimum basal diameter, and the height of each barnacle. Relative Height was then calculated as the ratio of height to average basal diameter and Operuclar Exposure Index as opercular width divided by operuclar depth. We hypothesized that since whelks must access the opercular plates in order to prey on barnacles, deeply recessed opercular plates might protect against whelk predation, while a wider opercular aperture might permit easier whelk access. The opercular exposure index is thus an indicator of susceptibility to predation. The orientation of substratum and the direction of exposure were also recorded for each measured barnacle. This step was repeated until we measured ten barnacles for every meter along the transect. We sampled two transects at Breakwater and three transects at Hopkins. Field Experiment In each of our two study locations, we selected three plots in the center of the Tetraclita zone. The Hopkins plots were 1.2 meters above MLLW, and the Breakwater plots were 0.7 meters above MLLW. The plots were chosen to provide enough flat substrate to accept a square .0625-m“ stainless steel cage. Each plot contained 24 to 30 Tetraclita, and we cleared all other barnacles from the plots. Ten whelks were placed in each cage. Once we selected the experimental sites, we recorded the same morphological data as described above for all Tetraclita present. To test whether a large opercular depth offers any defense against predation by whelks, one-third of the Tetraclita in each cage were selected at random to have their shells filed. Using an iron file, we filed the lip of each shell until the edges were flush with the opercular plates. Filing removed the shell above the operculum, reducing the opercular depth to zero. To ascertain whether the filing process itself acted as an unanticipated signal inviting whelk predation, we filed on the side of one-third of the barnacles without significantly changing the shell shape. The remaining third of the Tetraclita acted as unmanipulated control barnacles. The whelks used in both sites were collected from Hopkins, since there were no whelks at Breakwater. Over the course of four weeks, we removed the cages once a week and recorded the number of experimental barnacles that had been eaten. Heat Flux Experiment To test if heat flux varied with the shape of the barnacle, we selected two barnacles representing extremes in the morphological range. One barnacle had a large relative height (.827; i.e. tall and narrow, see above). The other barnacle had a low relative height (.549; or short and broad). We removed the barnacle tissue from each shell and glued the opercular plates in place. We then filed the bottom of the shells so that they were flat enough to make an effective seal to a flat concrete substrate. The interface of barnacle and rock was coated with KY Jelly to prevent undue heat loss. We calculated the post-filed surface area of the bottom of the shell by tracing the barnacle’s base and cutting out a piece of paper corresponding to the bottom of each shell. These basal areas were then weighed and compared to the mass of a piece of paper of a known area. We measured the volume of each shell by placing the inverted shell on a microbalance, taring to zero, and subsequently filling the barnacle to capacity with water. Äfter determining each barnacle’s internal volume, we drilled a small hole in the side of each shell, inserted a thermocouple into the center of the shell, sealed it with silicone sealant, and measured the distance between the thermocouple and the bottom of the shell. Once the sealant dried, we filled each shell with a solution made from one teaspoon gelatin and one-half cup water. We left the gelatin solution to firm up in a refrigerator for 12 hours. The gelatin solution, being mostly water, approximated the thermal properties of the barnacle's internal structures. The shells were placed within 15 cm of each other on a concrete slab in full sunlight on Hopkins. Each shell was placed directly over two thermocouples, which had been inserted into the concrete block at a known depth. A third thermocouple was placed between the barnacle and the substrate. Finally, we had a fourth thermocouple to measure air temperature at the rock surface. In combination with the thermocouple inside the barnacle, the rock thermocouples allowed us to determine the temperature gradient between the barnacle and the substrate. Using the surface area of contact between the rock and the barnacle, we could then calculate heat flux through each barnacle. Temperatures were recorded every 5 seconds and then averaged over 5-minute increments. The duration of the experiment was 4 hours. After approximately 2 and a half hours, we shaded the barnacles for twenty minutes and then re-exposed them to sun to allow us to see how the barnacles respond to different levels of insolation. Data Analysis Morphological measurements were compared between sites and substrate orientation using a single-factor analysis of variance (ANOVA). Data from all transects were combined for this analysis. Too few barnacles occurred on horizontal substrates at Breakwater to include in the analysis, leaving Horizontal and Vertical at Hopkins and Vertical at Breakwater as the levels of the single factor. The caged predation experiment was analyzed by two-factor ANÖVA, with each cage as a replicate and Filed, Sham Filed, and Control as the treatment levels. Hopkins and Breakwater were the two sites. All data were analyzed using Systat 8.0. Results Field Sampling We found significant differences in the shape of Tetraclita at our two study sites (Table 1A, Fig. 1). At Hopkins, barnacles on vertical substrate had large relative heights (Fig. 1), while barnacles on horizontal substrate had significantly lower relative heights (ANOVA, Pe 0.005). Barnacles on vertical substrate at the Breakwater also had significantly lower relative heights compared to Hopkins vertical (mean = 0.784; SE =.016). There was no significant difference in relative heights between horizontal barnacles at Hopkins and vertical barnacles at the Breakwater. Of the 200 barnacles we measured at the Breakwater, only 12 were found on horizontal substrate. Because of this small sample size, we did not include these barnacles in any of our data analysis. Vertically oriented barnacles at Hopkins had significantly smaller values for the opercular exposure index than either barnacles on horizontal substrate at Hopkins or vertically oriented barnacles at the Breakwater (ANÖVA, P 0.005; Table 1B, Fig. 2). There was no significant difference between horizontal barnacles at Hopkins and vertical barnacles at the Breakwater (ANOVA, PK 0.005; Fig. 2). The frequency distribution of barnacle size (=estimated internal volume) were similar between our two study sites, though the Breakwater had a few larger barnacles than Hopkins (Fig. 3). The similarity of the size distributions at our study sites facilitates direct comparisons of opercular exposure. Since the index increases with increasing barnacle size at both sites (Fig 4), if size distributions were not similar, direct comparisons of the index would not be possible. Field Experiment At Hopkins, Nucella consumed significantly more filed barnacles (two-factor ANOVA and Tukey HSD post hoc comparisons, P « 0.005, Table 1C, Fig. 5).. This trend was also present at the Breakwater, but the differences were not significant, possibly because so few barnacles were drilled there. There was no significant difference between the control or procedural control groups at either Hopkins or Breakwater (Table 1C). Temperature Experiment In our heat flux experiment, the short barnacle heated up at a faster rate than the tall barnacle, and retained a higher internal temperature throughout the course of the experiment (Fig. 6). The average heat flux through the tall barnacle was substantially larger than that through the short barnacle (mean -O.14, SE 0.051; mean =-0.22, SE = 0.038). After 110 minutes, when substrate temperature, air temperature, and internal barnacle temperatures were holding fairly constant (Fig. 6), we measured solar insolation at 925 W m-2, and calculated heat flux from the rock to the barnacle at .167 W for the flat barnacle and .259 W for the tall barnacles. At this constant temperature, we calculated the ratio of convective heat loss for the flat barnacle to the tall barnacle as 0.709. Discussion The purpose of our study was twofold: first, to establish that there is variation in the morphology of Tetraclita, and second, to determine the mechanisms by which the different variations might be adaptive. By sampling barnacles at our two sites, we determined that Tetraclita varies both in its relative height and in its opercular exposure index. There were significant differences both between the two sites and between vertical and horizontal orientations at Hopkins. At Hopkins the barnacles exhibited significantly less exposed opercular plates than at Breakwater. In our predation field experiment, whelks preferentially drilled barnacles with more exposed opercular plates. Barnacles may thus benefit from having less exposed opercular plates in areas of high whelk abundance such as Hopkins. The difference in the opercular exposure index between two locations where whelk abundance varies suggests a potential "whelk effect" on barnacle morphology. Lively (1986) demonstrated how whelks affect the development of barnacle morphology via shape-changing chemical cues. The predation experiment shows that predation pressure may select for less exposed opercular plates. However, opercular exposure differs on a local scale within the Hopkins site, where whelk abundance is variable but consistently high: Tetraclita on vertical substrate have significantly lower opercular exposures than Tetraclita on horizontal substrate. Thus, a whelk effect is unlikely to be the only factor determining barnacle morphology. Otherwise, we would expect all barnacles to exhibit a more recessed morphology. Furthermore, whelk presence is unlikely to be a dominant developmental cue if, as we see within the whelk-abundant Hopkins site, barnacles grow to become more open as they increase in size (Fig. 4). In Tetraclita, size does not appear to provide a refuge that would diminish the importance of having protected opercular plates, as barnacles of all sizes were eaten in the experimental treatments. Heat stress may act as an alternate selective factor in the environment of Tetraclita. The variation we see in the barnacle shape, particularly the variation in relative height, may be a function of Tetraclita's thermal environment. Using relative height as an index of shape for which heat stress might vary, Yip and Shin (1990) concluded that short, broad barnacles were more insulated and had lower heat flux through their shells. They also found that solar insolation per unit area was higher on horizontal substrate and that barnacles on horizontal substrate were shorter. Hence, Yip and Shin (1990) posited that barnacles with low relative heights were better suited for the more thermally stressed horizontal substrates. In our study, we observed similar relative height differences between Tetraclita at different sites as well as on substrates with different orientations. Hopkins barnacles on vertical substrates are taller and narrower than Breakwater barnacles with a vertical orientation. If, as Yip and Shin (1990) suggested, a low relative height indicates a morphology acclimated to a high temperature stress, then we would expect the Breakwater site to be a more thermally stressful environment. Because of higher wave exposure, Hopkins receives more spray and is likely cooler than the Breakwater. Also, Tetraclita inhabited a vertical range which extended higher in the rocky intertidal at Hopkins compared to Breakwater, which correlates with the more abundant spray at Hopkins. Our temperature experiment suggests that the relationship between temperature, morphology, and orientation is unclear. While we found that tall, narrow barnacles exhibit a higher heat flux and maintain a lower temperature on horizontal substrate, we were unable to test if the same were true for vertical substrate. It appeared that conductive heat flux from the substrate provided the majority of the heat acting on the barnacle, so that a more insulated shell would fail to pass this heat effectively to the air. The higher insulation does not limit conductive heat from the rock because the barnacle does not have a basal plate and is uninsulated at its base. Our experimental barnacle with low relative height had a thicker shell than the relatively taller barnacle, and this greater thickness may be responsible for the reduced the heat transfer coefficient. This corroborates the findings of Yip and Shin (1990) who found that a thicker shell insulated the barnacle more. In our experiment, the shorter, thicker barnacle maintained a higher internal temperature when the substrate was warm relative to the ambient air temperature. Because lethal temperatures are most likely to occur where the substrate is hotter than the air temperature (M.W. Denny, personal communication), it appears that the shorter barnacle is not better suited thermally to the horizontal substrate, as Yip and Shin (1990) suggested. Our observations show similar patterns to Yip and Shin (1990), but the conductive heat flux we observed leads us to question their conclusion. Ultimately, a more refined model will be necessary before Tetraclita relative height can be understood as an adaptation to heat stress. At this time we can only conclude that heat flux varies with the size and shape of the barnacle, and the shape varies with substrate orientation. Conclusions regarding the functional relationship between the thermal environment, shell morphology, and substrate orientation will require further study. Beyond any of the influences we have explored, there exist a variety of factors that also could have an effect on optimal Tetraclita shell morphology. The temperature effects may be related more to the size of the barnacle than its shape. Foster (1971) suggests that larger barnacles are more insulated from heat stress. There may also be non-thermal forces at work. Wave forces may favor shorter, broader barnacles, and this selective pressure may counteract, augment, or override completely the predation and temperature effects (Crisp and Bourget, 1985). There may be unobserved tradeoffs for one morphology versus another. For instance, a different morphology may reduce fecundity, as Lively (1986) found with "bent" Chthamalus morphs. Further study is necessary to distinguish the many physical and biological factors from one another. Conclusion Our field sampling data firmly established the existence of morphological variability within Tetraclita living in Monterey Bay, California. Barnacles vary both in regards to their relative height and the exposure of their opercular plates. Our predation experiment demonstrated the importance of opercular exposure to susceptibility to whelk predation. Our temperature experiment demonstrated that variation in a barnacle’s relative height has a significant affect on the heat stress of the barnacle. However, we couldn’t conclude from our temperature data the exact nature of the relation between shell morphology and heat stress. Acknowledgments We thank Eric Sanford and Mark Denny for advice, equipment and field assistance. We especially thank Jim Watanabe for assistance with all facets of this project—from experimental design to statistical analysis—and for the inspiration and hours of tireless work he put into this project. 12. 6 Literature Cited Crisp, D.J. and Bourget, E. 1985. Growth in Tetraclita Barnacles. Adv. Mar. Biol. 22: 199-244 Foster, B.A. 1971. On the determinants of the upper limit of intertidal distribution of barnacles. Journal of Animal Ecology. 40: 33-48. Lively, Curtis M. 1986. Predator-induced shell dimorphism in the acorn barnacle Chthamalus anisopoma" Evolution. 40: 232-242. Yip, May W. and Shin, F.G. 1990. A study of the thermal conductance in relation to the shell geometry of Tetraclita squamosa japonica. Bulletin of Marine Science. 47: 86-93 Tables Table lA. ANÖVA of Relative Height vs Location. Levels of Location were Hopkins Vertical, Hopkins Horizontal, and Breakwater Vertical. Too few barnacles were found on horizontal surfaces at Breakwater to be included in the analysis. Source F-ratic MS 0.97. Location 0.486 11.687 .001 419 Error 17.437 0.042 Table IB. ANOVA of Opercular Exposure vs Location. Opercular Exposure was calculated as the ratio of opercular width to opercular depth (see Methods and Materials) Source MS F-ratio 0.695 6.629 Location 139 .001 43.841 418 0.105 Error Table 1C. ANOVA and Tukey's Test of Predation Field Experiment. Treatment was Filed, Sham filed, and unmanipulated control (see Methods and Materials); Location was either Hopkins or Breakwater. Interaction and Error sums of squares were pooled to test Main Effects. Source MS F-ratic 1389 1.389 640 Location 071* 6.167 3.364 Treatment 12.3 2.39 308 4.8 Site X Treatment 1.3 Error 1.833 *P value for pooled data Post-Hoc Comparisons: Tukey's HSD Matrix of pairwise mean differences. BV HV HH BV HH 0.026 (NS) 1.06* 131* HV * - significant NS = not significant 14. Figure Legends Fig. 1. The relationship between relative height (= height / mean basal width) and location Fig. 2. The relationship between opercular exposure (= opercular width / opercular depth) and location Fig. 3. A histogram showing the size distribution of barnacles according to site Fig. 4. A scatterplot relating the opercular exposure to barnacle volume. Opercular esposure (ÖW / OD) is plotted on the y axis, and called REALÖW/OD Fig. 5. The results from our predation experiment showing a significant whelk preference for filed barnacles Fig. 6. The temperature recordings taken during the four hour run of our temperature experiment Figure 1. 0.95 0.85 E 0.8 0.75 0.7 0.65 Figure 2. 1.2 1.15 1.1 1.05 0.95 0.9 0.85 Relative Height vs Location Breakwater Hopkins Horizontal Hopkins Vertical Vertical Opercular Exposure vs Location Breakwater Vertical Hopkins Horizontal Hopkins Vertical Figure 3 Figure 4 75. 50- 25 2 1 -0.16 LOCATION -0.14 Breakwater 0.12 7 Hopkins 0.10 -0.08 -0.06 -0.04 -0.02 S+0.00 1000 2000 3000 4000 5000 VOLUME Location BreakwaterV X HopkinsH HopkinsV oto 1000 2000 3000 4000 5000 VOLUME 17. Figure 5 Predation Susceptibility Results 4.5 U Hopkins 3.5 5 3 + Breakwater a- 515 0.5 0 - Procedural Filed Contro Control Figure 6. Temperature Recordings 30 e e e e e nenene. sBoog t AAAAA oe 20 15 —Tall 5 10 —— Air Temp —4— Flat —— Substrate Temp o aaa- Time (5 minute increments) 18.