Abstract Littorina planaxis, an intertidal snail, is abundant on a variety of substrata. Iwo rocks, granite and basalt, were examined to determine if the rock type or the roughness of the rock influenced the movement pattern or population density of the snail. A diffusion model was used to quantify the spread of the snail across the rock in rough, intermediate and smooth areas on each rock type. Additionally, the fractal dimension of the rock was measured on each site in an attempt to quantify the topography of the rock and to provide a less subjective estimate of roughness or smoothness. The roughness of the rock was not found to have an effect on the diffusion of the snail. However, there was an almost significant difference (p=.052) between the movements of the snails on the granite and the basalt, with snails spreading faster on the granite. Neither the rock type nor the roughness of the rock was found to influence the population density of the snails. While the diffusion model provided interesting information about the movement patterns of the snail, the fractal dimension analysis was not found to be helpful because no differences could be found in the fractal dimension of the rock. Introduction Littorina planaxis is a an intertidal snail found abundantly on the west coast of North America from Charleston, Oregon to Baja California. It grazes on the microalgae which grow upon the rocks, living high in the intertidal zone (Morris, et al., 1980). The rocks upon which the snails live are highly variable in topography and composition, providing a range of habitats. The substrate is likely to play a large role in the ecology of many intertidal animals. Cracks and indentations in the rock provide the snails with protection from wave exposure and desiccation. The microalgae they eat may grow differently from rock to rock. Smoother rock may allow the snail to crawl farther in a given amount of time, rather than navigating the jagged bumps in rocks with a more complex topography. Studies on other intertidal animals have shown that the substrate may affect their population turnover. density and movement patterns (Underwood and Chapman, 1985; Raffaelli and Hughes, 1978). In this study, two aspects of the substrate were examined for their affects on Littorina planaxis. The topography of two types of rock, granite and basalt, was examined to find whether the roughness of a rock played a role in the snail’s movement patterns or population density. In examining the movement of a population of animals, it can be difficult to describe which factors play a role in the movement of the group Measurements can be made of the speed, direction or distance of an individual, but often cannot be extended to describe the more complex movements of the population. In most cases, the movement pattern of the population as a whole is more relevant than that of a given individual. For example, genetic mixing of a population may be affected by the distance a group of snails spread in a breeding period. In order to more closely examine the behavior of a population, the pattern of movements of a group of animals should be measured rather than those of single animals. A diffusion model is one way to measure the movement of a population. Diffusion is usually used to describe the random motion of gas molecules due to thermal energy However, it can also be used to describe other systems (Berg, 1993). In this case, à diffusion model is used by analogy to describe the random motion of a group of snails across a rock face. Using this model, the spread of the snails across a substrate over a given period of time can be measured. Finding the rate of spread of the population can be relevant information in learning about immigration, emigration, and the gene flow within or among populations in a given area. Quantifying the differences in the substrate is an obstacle to objectively examining its effect on intertidal animals. It is easy to categorize different rocks, but how can the topography of the rock be adequately described? Often the best attempt when studying the effect of substrate on animals is simply to use a subjective measure, such as rough or smooth. This is usually less than ideal because one person's perception of the differences in a rock do not alloy a comparison to other rocks on an absolute scale. Additionally, as humans we cannot know how a rock is perceived by an animal on a totally different size scale. In an attempt to circumvent this problem, the fractal dimension o the rock was used to describe the complexity of the topography of the rock. Fractal geometry provides a useful way to describe the irregularities of the surfaces often found in nature because it takes into account not only the complexity of the surface but also the scale of its measurement. Materials and Methods The experiment was conducted at a point in Carmel Bay where basalt and granite can be found in close proximity. Three sites were chosen on each rock type that were visually classified as having rough, intermediate and smooth topography. The sites were chosen at equivalent tidal heights in the mid-littorina zone for and checked for similar wave exposure at high tide in order to control as much as possible for these factors. Three trials were then run on each of the six sites. On day 0 of each trial, a 0.25m2 quadrat was cleared of Littorina planaxis, and the number of snails recorded. Thirty of the snails taken from the site were then chosen randomly, marked with a dot of enamel paint and placed at a point in the center of the quadrat marked as the origin. Each day for the following three days, the positions of the snails were recorded to the closest centimeter as Cartesian coordinates from the origin. A minimum of 20 snails were recovered at any time. Three replicate diffusion trials were performed on each of the rock sites. To find the diffusion of the snails, the spatial variance was calculated at each time period as: L(X-X)? n-1 here X = mean position of the snails x = position of one snail n = total f of snails recovered. Because the diffusion model assumes random movement of the snails, the diffusion of the snails in both the x and y direction were calculated. Using the Student's t-test, movement in both directions could not be shown to be different. Therefore, the radial distance of the snails from the mean position of the group was considered to be an acceptable measure of diffusion. Because the day to day measurements of the spread of the snails was so variable, likely due to changing environmental factors such as weather, temperature or tidal height, diffusion was measured as the net spread of the snails over the entire three day period. This tends to average the spread over the three days and decreases the noise from environmental factors. The diffusion of the snails in the six sites were then tested for the effect of rock type and roughness using a nested ANOVA. Because the variances of among sites were heterogeneous (Cochran’s test), data were log-transformed prior to analysis. A second ANOVA was performed on the population counts taken from each of the sites to test for differences in population density between rock types and topographies. The population data were analyzed in their untransformed state. In an attempt to quantify the fractal dimension of the rock, a mold was taken of each rock site using liquid latex in four cases and Instamold in two cases. The molds were then cut lengthwise and the profile of the rock traced so it could be digitized by a computer. A computer program was then used to analyze the fractal dimension of the rock in the following way (Morse, Lawton, Dodson and Williamson, 1985). A box was drawn around each digitized rock profile. The box was then divided into four, and four into sixteen, up to 1024 boxes. For each level of subdivision, the computer counted the number of boxes in which the line was found. The slope of the regression of the natural logarithm of the number of boxes in which the line was found to the natural logarithm of the size box is an estimate of the fractal dimension of the rock. Then the digitized line was rotated a random amount and the same count done. The line was rotated four times, giving four estimates of the fractal dimension of the rock, which were averaged. A dimension of 1 indicates a perfectly smooth surface, with increasing values as the complexity of the surface increases. Results The fractal dimensions of the rock profiles were found to lie between 1.08 and 1.12 for all the sites, and were not statistically different among sites. The mean spatial variance after three days ranged from 246.5 cm'on the rough basalt site to 777.3 cm’ in the intermediate granite site. In all three granite sites, the means of diffusion were greater than the means of diffusion on the three basalt sites. However, the standard error was large because of the variance in the diffusion rates from day to day and trial to trial (fig. 1). The pooled mean for the spatial variance of granite sites was 543.4 cm’ while the pooled mean of the basalt sites was 286.1 cm? The ANÖVA performed on the diffusion of the snails showed the roughness of the rock not to have a statistically significant effect on the diffusion of the snails. However, the diffusion on the granite was found to be significantly different than the diffusion on the basalt to a level of P =,052 (table 1), with granite having the larger diffusion. An ANOVA was also done to compare the population densities of the snails on the six rock sites (table 2). Although a graph shows a trend of decreasing population density with decreasing rugosity in both rock types, and a higher mean density in the basalt than the granite, these trends were not statistically significant (fig. 2). Discussion The ANOVA performed on the diffusion data showed no statistical difference of the movement of Littorina planaxis within the rock types. This would indicate that the roughness of the rock did not substantially affect the diffusion of the snails. Similar studies on related intertidal snails, Littoring unifasciata showed an increased distance traveled on less complex surfaces related to directional snail movement (Underwood and Chapman, 1989). However, by using diffusion as a model, the distance traveled by the snails in this study was only indirectly measured. Rather, the spread of a group of snails was calculated. Implicit in the calculation of diffusion is a correction for a changing mean position of the group. The group of snails could travel a great distance in the same direction, but if the spread of their positions relative to each other was not great, a small diffusion coefficient would be calculated. From this data no conclusions can be made about the absolute distance traveled by the snail, or how it may or may not relate to the roughness of the substrate. The ANOVA of diffusion as a function of rock type, although not quite statistically significant at p =.052, suggests that the diffusion of the snails on granite may be higher than that on the basalt. Although this experiment does not meet the traditionally accepted significance level of p-.05, it is close enough to warrant a closer look at rock type as a factor in snail movement. Repeated trials with an increased sample size are necessary before these results can be considered conclusive. Other rock types should also be examined for an influence on the movement of the snails. If it were found that the effect of rock type is real. interesting questions are raised. If it is not the roughness of the different rock types that influences the movement of the snails, what other aspects of the rock might? One possibility could be a difference in growth of microalgae on the two rock types. BecauseLittorina planaxis grazes on the microalgae, differential abundance or types which grow on the granite and the basalt might be examined as a factor in the movement of the snails. The population density of Littorina planaxis was not found to be dependent on the complexity or type of the substratum, and data on related species do not show a consistent correlation of topography and population density. Crevices and pits have been shown to affect population density of Littorina rudis (Emson and Faller-Fritsch, 1976; Raffaelli and Hughes, 1978) while Littorina unifasciata did not show any relationship between topography and population density (Underwood and Chapman, 1985). Intuitively, it might seem that surface complexities would influence the population density of snails. In rougher areas, the snails have more cracks and depressions for protection from wave exposure. Without protection, the littorines might be more easily washed from the rock, which would tend to decrease the size of the population. Because all density counts were taken between the months of April and June, these results reflect only the springtime population densities. It may be interesting to repeat the comparison of population densities during the winter months when storms may impose greater wave forces. Fractal dimension is a useful tool in describing some aspects of nature. It has been used to describe the complexity of algae (Dau, 1995), the path of clownfish while swimming (Coughling, Strickler and Sanderson, 1992), and the shape of coral reefs (Bradbury and Reichelt, 1983; Mark, 1984). However. In this study, fractal analysis did not prove to be useful in quantifying the different complexities of specific sites on rock. Despite the perceived differences in the complexity of the substrate, the fractal dimension was not found to be significantly different from site to site. Using the diffusion model to describe the movement of a population across a substrate provided interesting information about the snails and questions for further thought and study, such as the average spread of the snails over a given period of time. A diffusion coefficient can be calculated from the temporal rate of change in the variance (Berg, 1993): 05 D-2 9t here D = diffusion coefficient og- change in variance ot = change in time. With the diffusion coefficient, the spread of the snails over a given period of time can be calculated. r(15 - JADt Here «r' (t)2 - rms distance, or average displacement of snails for a given time period diffusion coefficient time Using these equations, the average spread of the snails on the basalt and on the granite can be calculated. The mean variance after three days on basalt was 286 cm' while on granite the mean variance was 543.4 cm’. Plugging in these values gives a diffusion coefficient of 47.5 cm’day" for basalt and 90.6 cm'day“ for granite. Using these diffusion coefficients in the second equation, the average displacement for a snail in one day on basalt is 13.8 cm. and on granite is 19.0 cm. However, according to the diffusion model the displacement of the snails from the mean varies with the square root of time. rather than with time. For one year, the average displacement of a snail from the mean position on basalt would be 186 cm and on granite would be 257 cm. If the snails do conform to the diffusion model, this would mean that the average displacement of the snail in a day may be on the scale of tens of centimeters, but over a year the average displacement is only on the scale of a few meters! Of course, snails do not perfectly mimic the motion of molecules. The diffusion model assumes random motion, with an equal probability of the snail moving in any direction. In reality, the snails are likely to be influenced by the cracks in the rock or by following the trail of öther snails. It would be interesting to track snails over a longer period of time to see if their movements do match the predicted values. Further analysis should be conducted to find how diffusion may be related to the ecology of animals. References 1.Berg, H., 1983. Random Walks in Biology, Princeton University Press 2.Bradbury, R. & Reichelt, R., 1983. Fractal dimension of a coral reef at ecological scales. Mar. Ecol. Prog. Ser. vol 10, pp 169-171 3.Coughlin, D., Strickler, J. & Sanderson, B., 1992. Swimming and search behaviour in clownfish, Amphiprion perideraion, larvae. Anim. Behav, vol 44, pp 427 - 440 4. Dau, B., 1995. The relationship between the fractal dimension of the intertidal alga mastocarpus papillatus and the abundance of grazing snails. Unpublished report, Bio 175H. Hopkins Marine Station 5.Emson, R.H. & Faller - Fritsch, 1976. An experimental investigation into the effect of crevice availability on abundance and size-structure in a population of Littorina rudis. J. Exp. Mar. Biol. Ecol., vol 23 pp 285-297 6.Mark, D., 1984. Fractal dimension of a coral reef at ecological scales: a discussion. Mar. Ecol. Prog. Ser. vol 14, pp 293 - 294 7.Morris, Abbot & Haderlie, 1980. Intertidal Invertebrates of California. Stanford University Press 8.Morse, D.R., Lawton, J.H., Dodson M. & Williamson M., 1985. Fractal dimension of vegetation and the distribution of arthropod body lengths. Nature, vol. 314, pp731 -733 9.Petriatis, P., 1982. Occurrence of random and directional movements in the periwinkle, Littorina Littorea.J. Exp. Mar. Biol. Ecol., vol 59, pp 207 -217 10.Raffaelli, D. & Hughes, R., 1978. The effects of crevice size and availability on populations of Littorina Rudis and Littorina Neritoides. J. Animal Ecol. vol. 47, pp71-83 11.Sugihara, G. & May, R., 1990. Applications of fractals in ecology. Trends in Evol. & Ecology, vol. 5, pp 79 -86 12.Underwood, A. & Chapman, M., 1985. Multifactorial analyses of directions of movements of animals. J. Exp. Mar. Biol. Ecol., vol 91, pp 17 - 43 13.Underwood, A.J. & Chapman M., 1989. Experimental analyses of the influences of topography of the subtratum on the movements and density of an intertidal snail, Littorina unifasciata. J. Exp. Mar. Biol. Ecol., vol 134, pp175 -196 Source Sum of F - Ratio DE Mean¬ quares Square Rock 1.261 1.261 —7.482 0.052 1 Site (Rock) 0.674 0.791 0.169 0.42 ErrOt 4.817 121 0.401 Table 1. Analysis of snail diffusion on granite and basalt. Data are for six sites: rough, intermediate and smooth on each of the two rock types. (Data are log-transformed for homogeneous variances.) Sum-of Mean- F-Ratio Source Squares Square 6.589 10706.722 Rock 0.062 10706.7 Site 6499.556 1642.889 1.896 0.176 12 857.111 10285.333 Error Table 2. Analysis of snail density on granite and basalt. Data are for six sites: rough, intermediate and smooth on each of the two rock types. ig. 1. Mean snail variances on six sites: br = rough basalt, bi = intermediate basalt, bs = smooth basalt, gr = rough granite, gi = intermediate granite, gs - smooth granite. Error bars show standard errors. Fig. 2. Mean snail densities on six sites: br = rough basalt, bi = intermediate basalt, bs = smooth basalt, gr= rough granite, gi = intermediate granite, gs - smooth granite. Error bars show standard errors. Spatial Variance, sq.cm + snails /(.25 sq.m) 1 would like to thank Mark Denny and Jim Watanabe for their patient guidance, Brian Gaylord for his countless hours of help, and Judy Thompson for her understanding with my use of the Hopkins truck.