A Comparison of Migration Between
Displaced and Control Gastropods
Steve Smith -- Marine Biology: 19/11/96
Abstract: An experiment was done to test the hypothesis that snails that were displaced from their naturally preferred habitat would migrate a further distance over a given time period than control snails within their preferred habitat. Upper shore and a lower shore experimental groups were compared to the control. Austrocochlea obtusa were used from North Stradbroke Island and Nerita plicata from Lady Elliott Island. It was shown that on N.S.I., the upper shore snails moved significantly further than the control, and the lower shore moved the same distance as the control. On L.E.I., the lower shore was the group that moved further than the control, while the upper shore snails migrated a significantly shorter distance than the control. A number of factors contributed to these results.
Introduction
Gastropods are often noted for their beauty and morphological brilliance. Speed and locomotion, however, are not characteristics that most immediately think of when considering snails. Although snails are not fixated to one particular spot, it is difficult to imagine how far these small organisms really migrate. Further, while thinking about migration, another inquiry is about possible changes in migration distances due to specific circumstances. In particular, if a specific species of snails were removed from their preferred habitat, would there be a change in the distance traveled over a given amount of time? With this question in mind, a reasonably interesting experiment can be designed to see whether or not this is true.
Before the presentation of the experiment, a brief background on species and location will give it some clarity. On North Stradbroke Island, Austrocochlea obtusa were found abundantly on the eastern rocky shore. They are a gregarious species with shells containing broad black and white longitudinal bands (Rippingale, 1961). Located from Tasmania to southern Queensland, they are found living exposed amongst intertidal rocks, averaging in length from 1.5 to 2.0 cm.
Because Austrocochlea can't be found as far north as Lady Elliott Island, another species was selected that also live gregariously in intertidal regions, usually amongst rocks. Nerita plicata are found with strongly spiraled cords, sometimes spotted with black (Dance, 1974). They are known to have three or four denticles on their columellar lip and two larger denticles, enclosing five smaller ones, on the outer lip. They grow to about 2.0 to 3.0 cm and are found commonly in South Africa and regions of the Indopacific.
After considering this information and the question that has been formulated, a hypothesis can now be stated. It was hypothesized that snails moved away from the chosen habitat would move a further distance in a given time period than snails left in the habitat. After concluding that this is a reasonable hypothesis to test, we set out to experiment it.
Methods and Materials
A mark-recapture experiment was carried out on both islands. On North Stradbroke Island, the Austrocochlea habitat was found by doing six , 1 m2 quadrat species counts, spread out evenly within the intertidal range. Because the entire HWS program was also participating in this exercise, and were broken up into six teams, there were six such quadrat counts taken, totaling 36 counts. With the data collected by each group, quadrat 3 contained the most Austrocochlea. This was right around the mid-tide level.
The Austrocochlea were chosen at random, without regard to size. A total of 100 were collected, 90 for the experiment and 10 extra in case of any collection mistakes. Fingernail polish was used for marking the snails: 30 for the down shore experimental, 30 for the control, and 30 for the upper shore experimental. Different markings on the shells ensured that there would be no discrepancies on the recaptures.
When returning the Austrocochlea to the shore during low tide, three trials were set up along the shoreline to minimize any bias established due to terrain differences. The trials were run 7.5 meters apart, with 30 snails in each trial. Ten marked Austrocochlea were put at the mid-tide level (the preferred habitat) for each of the trials serving as our control groups. From the three control points, three meters were measured up shore and three meters down shore. At each of these experimental points, 10 snails were placed. A clear marking of the spot where the snails were placed gives points from which measurements can be taken to determine the distance of migration.
The first migration measure took place at the next low tide (after one tidal cycle). With a 50 meter tape, the distance was measured and recorded from the marked point to each marked snail migrating from that point. Measurements were again repeated after the second tidal cycle. It is from this data that all statistical analysis was carried out.
The Lady Elliott Island experiment was carried out in much the same fashion. Nerita plicata were picked at random, again, without regard to size. The eastern shore of the island had a beach rock zone at the mid-tidal level. This is where a majority of the snails were found and this is the region used as our control. The trials, controls, and experimental groups were set up the same. Measurements were again taken after the first and second tidal cycles. This data has also been statistically analyzed.
A brief summary of the statistical test that was chosen to analyze the data of the experiment will be helpful in understanding our results. Because standard deviations were so large in the collected data, analyzing means is not as valuable as considering medians. Medians are not affected as drastically by outliers, therefore maintaining a higher level of resistance in measurements (Mitchell, 1996). A good way to accurately test the data by considering medians is the use of nonparametric statistics. In particular, the nonparametric test that was chosen uses a system of pooling and ranking the data. By ranking (from smallest to largest) all data between samples together, we expect to see relatively equal average ranks between data sets that are not significantly different, and a difference in average rank for those that are.
The Kruskal-Wallis k-Sample Test for unmatched data allows for this type of statistical analysis. It compares independent random samples from k populations by using a test statistic, H, that sums the differences between the average ranks and the expected ranks (Fig. 1). With this H-value, a comparison can be made to a critical value from a chi-squared table.
Figure 1: The Kruskal-Wallis k-Sample Test formula for calculating the test statistic, H. Mitchell (1996) gives this formula with the following variables: N=total number of measurements in the k samples; Ri denotes the sum of the ranks within the ith sample; Grand mean or average rank is N+1/2; and the sample mean rank for the ith sample is Ri / ni. The 12/N(N+1) factor gives the H value some normality.
H = 12/N(N+1) ni (Ri/ ni - N+1/2)
If a significant difference is found between at least one for the data sets, a number of paired comparisons can be performed to determine exactly which data sets differ from each other.
Results
On North Stradbroke Island, 62.2% of the Austrocochlea were found for measurement after the first tidal cycle, and 63.3% were found after the second. Another general stat that was discovered upon collection was that 93.4% of all the snails found moved from sand to rock. On Lady Elliott Island, 55.6% of the Nerita were recovered after the first tidal cycle, and 54.4% after the second. Because one of the low tides were at night, we were able to note the terrain that these snails were found on diurnally versus nocturnally. About 58% of the snails were found on the beach rock at night, while only 30.6% were found on beach rock during the day.
After calculating means and standard deviations for the data collected, extremely high variability was observed (Table 1). Because of this high variability, the means have been somewhat fluctuated, ultimately giving some wrong impressions of the results of this experiment. Therefore, as described above, the Kruskal-Wallis k-Sample Test was used.
Table 1: A summary of means and standard deviations of migration from (a) N.S.I. and (b) L.E.I. The numbers following the means and standard deviation (S.D.) represent the tidal cycles. All values are reported in cm.
a) North Stradbroke Island
| Level | Trial | Mean 1 | S. D. 1 | Mean 2 | S. D. 2 |
| Upper | 1 | 69.8 | 34.0 | 102.1 | 50.8 |
| 2 | 62.3 | 75.4 | 102.7 | 85.8 |
| 3 | 94.4 | 65.9 | 159.1 | 116.8 |
| Control | 1 | 48.7 | 36.7 | 81.0 | 53.7 |
| 2 | 35.4 | 16.5 | 39.5 | 19.2 |
| 3 | 58.2 | 14.7 | 104.2 | 32.3 |
| Lower | 1 | 35.9 | 8.6 | 43.5 | 35.0 |
| 2 | 9.5 | 5.4 | 20.7 | 17.6 |
| 3 | 115.6 | 88.0 | 145.4 | 102.1 |
b) Lady Elliott Island
| Level | Trial | Mean 1 | S. D. 1 | Mean 2 | S. D. 2 |
| Upper | 1 | 0 | 0 | 56.4 | 104.4 |
| 2 | 23.6 | 15.6 | 171.9 | 255.5 |
| 3 | 0 | 0 | 102.2 | 102.7 |
| Control | 1 | 190.3 | 171.7 | 255.0 | 177.4 |
| 2 | 218.0 | 117.1 | 244.5 | 123.3 |
| 3 | 232.8 | 190.0 | 296.7 | 85.1 |
| Lower | 1 | 659.5 | 192.3 | 802.5 | 85.4 |
| 2 | 537.5 | 173.2 | 855.0 | 35.4 |
| 3 | 508.0 | 180.3 | 625.0 | 262.4 |
The H-value for N.S.I. was 12.39. This is larger than the critical value, 5.99 (alpha level of .05), which permits a rejection of the null hypothesis that there was no difference between data sets and a conclusion that there was at least one significant difference. Paired-comparison analysis showed that the upper shore experimental snails moved significantly further than the control (=.01). There was no significant difference in migration between the control and the lower shore experimental. Because of the test parameters, we were able to compare both experimental, showing that the upper shore moved significantly more than the lower shore.
On L. E. I., the H-value was 28.29, which is considerably larger than the critical value of 5.99. This again shows that there is at least one difference between the three data sets. Paired comparison analysis shows that at the .01 alpha level, the control moved further than the upper shore. However, the lower shore moved significantly further than the control. Between the two experimental groups, the lower shore migrated further than the upper shore. Table 2 summarizes the paired comparisons for both locations together to simplify visual analysis of the experiment results.
Table 2: A summary of the paired comparisons from the Kruskal-Wallis k-Sample test. The following abbreviations were used: u=upper shore, c=control, and l=lower shore. Less than or greater than signs were used for significant results and an equal sign is used for non-significant results.
| Location: | u vs. c | l vs. c | u vs. l |
| N.S.I. | u > c | l = c | u > l |
| L.E.I. | u < c | l > c | u < l |
Discussion
Before starting the experiment, a hypothesis was formed that snails moved away from their preferred habitat would migrate further over a given time than would those that were left within their natural environment. With the experiment that has been described in this write-up, we set out to examine the validity of this hypothesis. As the results have shown, however, our hypothesis was not completely supported. At North Stradbroke Island, only the upper shore snails migrated further than the control, and at Lady Elliott, the opposite was true, only the lower moved more than the control, and in fact, the control moved further than the upper shore. Although our hypothesis was not completely supported, like any experiment, failure or success, there is a lot to be learned from it.
General information was learned about the snails that were chosen for this experiment simply by observing their behavioral patterns. For example, because all the snails in the experiment were taken from the mid-tide level, it can be concluded that the species that were chosen prefer to be half marine and half terrestrial animals. Also, 94.4% of the A. obtusa moved from sand to rock on the shores of North Stradbroke, a reasonable thing to consider is that the snails prefer to attach to the rock surface for stability and perhaps facilitating grazing. If the grazing theory is correct, then there are two possibilities. The first is that the rock surface contains more valuable nutrients that the organism can use. Or, perhaps it is an easier surface to migrate than sand or rubble, and this factor that they can cover more territory during grazing periods may be more valuable to the animal.
More support for this grazing theory was seen on Lady Elliott, where 58% of snails found at night (during their grazing period) were found on the beach rock, while during the day, when the nocturnal Nerita are inactive, only 30.6% were found on the rock. The rest were found burrowed in the rubble that surrounded the beach rock. Although 58% is not an extremely high value, it must be taken into consideration that the snails were very hard to find during the night. Perhaps this could be avoided if there is a repeated experiment by using a reflective marker on the snails that would show up when a light was shined on them.
This idea of a better marker to use for night purposes also carries over into the recapture topic as a whole. As reported in the results, 63% of the Austrocochlea were recovered and only about 55% of the Nerita were recovered. These numbers aren't too bad for a mark and recapture experiment, but some snails were found with only a tiny bit of nail polish on them, indicating that some snails probably lost all of their marking. However, some snails that were not found were probably either hidden within the rock crevasses on Straddie, burrowed too deeply in the rubble on Lady Elliott, or even taken by predation--hopefully not due to the large, hot pink X on their shells!
After collecting all the data and compiling the results that are reported in Table 1, it was understood that nonparametric statistics should be used for the statistics tests in order to minimize the high variability factor in the analysis. A big question is, why was there so much variation in the migration distances traveled by our snails? An easy assumption to think of prior to the experiment is that all of the control will migrate roughly the same distances while grazing, and the experimental snails will migrate for grazing and for the sake of returning to the preferred habitat. Hence, the experimental snails should migrate further than the control snails. This was not the case however, and it is important to discuss some possibilities of what actually went on within the experiment.
First, the most probable reason for the major differences in
migration was due to the immediate location that they were placed (i.e. the substrate they were placed on and surrounding features). For example, at North Stradbroke, the third lower shore trial had the highest means (Table 1a). These snails were placed in the sand with only one small rock nearby. The closest medium-sized rocks were over a meter up shore, and that is were they migrated to during the first tidal cycle. The three possibilities listed above may apply here (not sufficient nutrients from sand substrate, needing stability, or exposure to predation). Trial 2 on the lower shore had much different results, just 7.5 meters away from Trial 3. These snails were placed next to a group of medium sized rocks that the snails migrated to and were quite content with their position, only migrating 9.5 cm in the first tidal cycle and a total of 20.7 cm after the second. These major differences between Trials 2 and 3 contribute to the large variability seen in the data, and these are just two examples.
On Lady Elliott Island, the immediate environment also played a key role in migration distances (Table 1b). Both experimental groups of snails were placed on rubble while the control were placed on the beach rock at the mid-tidal level. It would typically be expected that both experimental snails would migrate a great distance to return to the beach rock, but some intermediate factors came into play here. The lower shore snails migrated the furthest probably due to the wave action giving them some help in getting them back to the beach rock. The upper shore snails did not have this natural force to help them and moved very little, if any, after the first tidal cycle. Considering the way snails move using their muscular foot, it is not hard to imagine that it would be difficult to move across rubble.
Finally, there needs to be some discussion about what the experiment was really about, differences between the experimental snails and the control snails (Table 2). At North Stradbroke, only the upper shore Austrocochlea moved further than the control. It is possible that these snails migrated further because the were not able to tolerate living a majority of the time as terrestrial animals. This is supported because 85% of the upper shore snails moved down shore while migrating. The lower shore snails may have been somewhat content with being a predominant marine animal since they moved about the same as the control snails, although, it should be noted that all of these snails were found migrating towards the mid-tide level as well. At Lady Elliott, the opposite occurred, as stated earlier. This was evident due to the lower snails getting a boost from the wave action, which enabled them to move quite a far distance in a relatively short time period, while the control snails were not affected by wave action since they were suctioned to the beach rock. The control snails had a facilitated migration ability due to the stability of being on the beach rock, thus allowing them to move significantly further than the upper shore snails.
In conclusion, the original hypothesis that the experimental snails would migrate further than the control snails was only somewhat supported. In some cases it was true, but more importantly, a lot was learned throughout this experiment. A number of experimental conditions had to be well thought about before making all the conclusions to our study. If this experiment were ever to be repeated, there are a few things that ought to be changed in order to obtain a more controlled experiment. First, there needs to be a better marking device than the rudimentary nail polish that was used. Also, if using the same species of snails were possible, it would enable the experimenter to determine whether or not different locations actually changed things such as migration rates. Finally, it would be recommended that all snails, both control and experimental were placed on the same sort of substrate to ensure an equal opportunity for locomotion. Aside from these few experimental difficulties, I think that our experiment was practical, interesting, and educational.
References
Dance, S. P. 1974. The Collector's Encyclopedia of Shells. Australia and New Zealand Book Company, Sydney.
Mitchell, K. 1996. A Primer of Nonparametric Statistics.
Rippingale, O. H., and D. F. McMichael. 1961. Queensland and Great Barrier Reef Shells. The Jacaranda Press.