Feeding Behaviors of the Eastern Grey Squirrel

Assessing the Feeding Habits of Eastern Grey Squirrels Using Varying Food Source Locations Coleman Drane, Marissa Moran, Noah Silverman, Zoey Lim 12-6-2019 Abstract The giving up density (GUD) is the ratio of food available in a foraging patch along with the measured consumption of the resource by an organism. Giving up densities for the eastern grey squirrels have been observed and measured previously, however, previous researchers have not determined differences in the giving up density between sheltered and unsheltered foraging locations. Here the preferential habits in feeding behavior with the assumption of a variation in perceived predation was analyzed. In this experiment, five two-hour trials were run with two feeding trays that each contained roughly 200 grams of peanuts. The results showed that the eastern grey squirrel foraged more in sheltered locations than unsheltered locations most likely due to a lower perceived risk of predation. Conclusions like this are valuable in conservation efforts based on behavioral ecology. Introduction One of the best animals to observe foraging habits in the wild is the eastern grey squirrel (Sciurus carolinensis) due to their scattering habits and their abundance. The way eastern grey squirrels search for wild resources, also known as their foraging strategy, can play a significant role in their ability to survive and reproduce. As degradation of natural forests is becoming more and more pervasive, fitness levels of squirrels have a greater potential to decrease, and they must adjust their foraging habits in order to obtain steady sources of food. As a result, organisms are confronted with the challenge of either adapting to these challenges or facing certain death. Squirrels living in urban areas that are in close proximity to humans, appear to be more limited by food, they might be more sensitive to predatory risk than those living in more natural areas (Bowers & Breland, 1996). The forests surrounding Ohio State University’s Columbus campus provides a mix of urban and natural settings which was ideal to assess foraging behavior. The campus is a useful indicator of various foraging behaviors because of the great diversity of habitat types that include wetland, forest, urban, and open fields. Eastern grey squirrels are perfect test subjects to study the effects of environmental food source availability. They have been found to adapt and adjust their behavior and foraging strategies when faced with specific stressors such as predation (Dill and Houtman, 1987). Using peanuts, the density of available food for eastern grey squirrels was analyzed to see if location and the level of exposure of the food source had any effect on foraging behavior. The experiment assessed the average percent consumption of food sources at various locations within a specific forested site because the amount of food remaining at a specific location serves as a good indicator of an organism’s Giving Up Density (GUD). This belief is based on prior research that concluded that the average percent of peanuts consumed by grey squirrels was significantly higher in forested locations than in suburban (exposed) areas (Pospiech et al., 2012). The hypothesis for this study was that the GUD of eastern grey squirrels will be lower in areas where sources of food are sheltered by nature than in areas of the forest that are more exposed and possess higher visibility to predators. Thus, it was predicted that an exposed source of food will have a higher perceived risk of predation because the grey squirrels will be more visible to potential predators compared to areas where the sources of food are more sheltered by the natural environment. Methods The purpose of this study was to analyze the differences in the foraging behavior of squirrels between a sheltered and unsheltered source of food. These experiments were conducted in the forested area beside the Carmac 3 bus lot on the west campus of the Ohio State University. This area was bordered by an open clearing without any tall plants or structures. In order to minimize bias and maintain consistency, a total of five experimental trials were run and the duration of each trial was two hours each. For each trial, two black trays were filled with 202 grams of peanuts and 1 kilogram of sand. These materials were collected and measured in the supply room of Jennings Hall. All five of the experiments were conducted between 12 pm to 6 pm, to avoid different foraging times and other confounding variables. The trays were placed within the same forest but the amount of overhead exposure varied. Data was collected by looking at foraging habits in areas that were sheltered versus areas that were open. The first tray was used for the open treatments and was placed on grass patches that were well-lit and located in a fairly accessible area with little to no trees in the general vicinity. The second tray was used for the sheltered treatments so it was placed in an area with thick ground cover and dense foliage to block most of the light and predator exposure. The image shown by figure 1 depicts the locations used for the unsheltered (left) and sheltered (right) treatments. In this experiment, the giving up density (GUD) for eastern grey squirrels was measured at both the sheltered and unsheltered food sources. At the end of each session, the leftover peanuts from each tray were then measured and recorded. Since the GUD was the average percentage of food left behind, it was calculated based on the final weight and the original weight to determine how much food was taken by the squirrels. The independent variable for this study was the location where the tray was placed and the dependent variable was the amount of food that was left after the experimentation period. The data has been presented as a bar graph (Fig. 2) that shows the total amount of food eaten in the sheltered areas versus the open foraging areas. The average % GUD left behind for each treatment was calculated and plotted (Fig. 3) to show the visual significance. Results from the experiment were analyzed with a one-way ANOVA test to determine the statistical significance. The F-value was used to determine the effect the sheltered and unsheltered areas had on the giving up densities. A T-test was then constructed to show if there was a significant difference between the means of the two groups. This allowed for an effective analysis of the hypothesis based on the P-values that were generated. In order to ensure that every member of the project contributed fairly, the fieldwork was divided into equal portions. There were five main tasks performed for each test that was conducted. The first task was to weigh the peanuts that would be used for each tray before and after the experiments. The peanut measurements needed to be precise to ensure consistency and accuracy. The second task was to gather the sand and trays and transport them to and from the testing locations. The third task was to set up the experimental design for both the sheltered and unsheltered treatments. This was done on-site within the same forest. The fourth task was to collect the data for each treatment during each trial. This included recording the date, the start and end times, the location, the starting weights, and the ending weights of the trays. The fifth and final task was to clean up the experimental site by collecting the remaining peanuts, dumping the sand, and ensuring that all traces of the setup were removed. Each group member was assigned to do different tasks during every trial to eliminate random errors. All team members contributed their time and work evenly in the project and the report. Noah worked on the abstract and introduction, Zoey worked on the methods, results and discussion, Marissa worked on the introduction and discussion, and Coleman worked on the results and acknowledgements. -218440247650 Figure 1 - This figure above is a photo reference for how the experiment placement of both the unsheltered (left) and sheltered (right) trays. Both trays are filled with sand and 202g of peanuts. Results All the collected results were recorded in Table 1 in the appendix section. The experiments were carried out between September and October of 2019. Each trial was conducted between 12 pm and 6 pm. Figure 2 displays the GUD of the sheltered and unsheltered results of all 5 experiments. This figure indicates that the final measurement of peanuts for the sheltered site were lower than the final measurements at the unsheltered site. Figure 3 shows the average percentage of food that was left behind, which served as a good indicator of the GUD. This was determined by dividing the total mass (in grams) of peanuts left after the trial by the total initial mass at the beginning. This figure highlights a consistent trend that the average GUD for the sheltered environment was lower than the GUD of the unsheltered environment. A significant percentage difference was found between the sheltered and unsheltered food source (Fig. 3 - sheltered 77.09% and unsheltered 85.79%). On average, the GUD was greater at the location of the unsheltered source of food than at the location for the sheltered source of food. The ANOVA test was used to examine the statistical significance in the F-value of the sheltered and unsheltered areas and to determine the effect it had on the giving up density. The experiment’s independent variable was the location where the tray was placed and the dependent variable was the amount of food that was left after the experimentation period. To analyze the results obtained from the experiments, a one-way ANOVA test was used to explore the impact of the food source’s location on the giving up density. The alpha level was set to 0.05; anything below 0.05 was considered to be statistically significant. The F-value was determined to be 8.669. The p-value was 0.0186, which is a statistically significant value. Therefore, the null hypothesis was rejected. To further reconfirm the results of the correlation between the two sets of data, a Welch two sample t-test was performed. The t-test was used to determine whether the mean of the sheltered environment significantly differed from the mean for the unsheltered environment. Like the ANOVA test, the alpha level was set to 0.05. The t-test calculated a confidence interval of -70.9073 to -4.2127 with the means for the sheltered and unsheltered experiments being 155.74 grams and 193.30 grams, respectively. This was graphically represented by a box plot as shown by figure 4. These means were used to calculate the standard deviation of the two trays. The standard deviation was 24.40 for the sheltered source of food and 7.44 for the unsheltered source of food. The interaction between the means resulted in a p-value of 0.03428 which was once again below the threshold. Since the p-value was statistically significant, the null hypothesis was rejected. Visual Representation of Results Table 1 - The weight of peanuts at the end of each trial as well as the average amounts left, average amount eaten, and the standard deviation of these averages. Initial weight of peanuts - 202g Date sheltered(g) Unsheltered (g) September 18 112.5 179 September 28 181.2 195.5 October 22 177 194 October 23 155 200 October 26 153 198 Mean Amount Left 155.74 193.3 Average Amount Eaten 46.26 8.7 Standard Deviation 24.40 7.44 Table 2 - This shows the results of the ANOVA test performed to determine if there was a significant difference between the two treatments. Degrees of Freedom: Sum Sq: Mean Sq: F Value: P Value: Treatment: 1 3527 3527 8.669 0.0186 Residuals: 8 3255 407 N/A N/A Significance Level: 0.05 0.05 0.05 0.05 0.05 Table 3 - This table shows the results of the t-test used. It shows the significant differences between the means of the two treatment groups and their confidence level. Significance Level: 0.05 Degrees of Freedom: 4.7372 T value: -2.9443 Confidence Interval: (-70.9073, -4.2127) Sheltered Mean: 155.74 Unsheltered Mean: 193.30 P value: 0.03428 Figure 2 - The Giving Up Density graph shows the the visual differences of GUD in sheltered and unsheltered locations in the five trials. Y-axis represents the amount of food left in grams and the X-axis refers to trial numbers. Figure 3 - A bar graph of the average GUD percentage of the sheltered and unsheltered area. The Y-axis is the percent of the food left behind by the squirrels and the X-axis is the trial number and result of each treatment group. Figure 4 - A boxplot that depicts the results of the t-test that was performed. Mean GUD for Sheltered (Covered) is 155.74, and mean for unsheltered (uncovered) is 193.30. Discussion Within the first trial alone, as shown by Table 1, the amount of food eaten from the sheltered tray was more noticeable than the amount eaten from unsheltered tray. This pattern persisted for all of the trials. For the unsheltered trays, there was an average of 155.74 grams left behind. On average, the squirrels ate 46.26 grams of peanuts with a standard deviation of 24.40 grams. For the unsheltered tray, there was an average of 193.3 grams of peanuts left. This meant that on average only 8.7 grams of peanuts were being eaten with a standard deviation of 7.44 grams (Table 1). This trend supports the hypothesis stating that the giving up density (GUD) will be lower in areas where the food source is sheltered by nature than in areas where it would be more exposed and visible to predators or risks related to their environment. The conclusion is consistent with the findings from (Bowers, 1996) where scientists determined that there is a correlation between human urbanization and grey squirrel foraging behavior. Discoveries outlined in their paper describe squirrels that preferred areas that were untouched by humans and sheltered by trees and bushes. For this experiment, it is demonstrated in Figure 3, where the average percentage of GUD left behind in unsheltered areas were about 8.7% higher than the sheltered food source. It also supports the prediction that the exposed food source will have a higher perceived risk of predation than food source that are protected by the natural environment. Additionally, densely forested areas appear to have a lower perceived risk of predation than areas that are less dense and more exposed. The results of the t-test and the ANOVA test support this conclusion because they both had p-values below the alpha level of 0.05. According to the experiment’s one-way anova test, the independent variable of location placed yields a statistical significance p-value of 0.0186. With that, we conclude that the location where the food is placed does affect the GUD of grey squirrels directly. In this case, food source that was placed in a sheltered environment increases the GUD of grey squirrels, and the null hypothesis is rejected. Moving on to the experiment’s t-test result, there is a significant difference between the means of the two groups. The p-value of the t-test is 0.0342, allowing us to conclude that the interaction between the means is statistically significant and reject the null hypothesis. The results indicate that there is a correlation between the coverage of the food source and the amount that was consumed from the tray. Our data is consistent with the study performed by Pospiech et al. (2012) as the GUD was greater in the unsheltered area than in the sheltered area. Although the data does support the hypothesis and predictions, there were some limitations to the study because it was impossible to maintain total consistency throughout the trials. In between the time frame of trials two and three, the majority of trees at the study site had been cut down so the area resembled more of an open clearing than a small forest. Although this change was unforeseen, the location of the sheltered trays was adjusted to better fit the independent variable. This tray was moved to areas that were still sheltered by nature but further in the forest, so that the experiment would still fit the original treatment. Forces outside of the control of the experimental design may have negatively influenced the validity of the results. It is recommended that any and all future studies should be conducted at protected forests or nature preserves so there is less risk of human interference on the natural landscape. In addition to the changes that occurred to the experimental site during the trials, this study was also conducted in a very small period of time during September and October. These months could be considered a prime foraging period, as it is right before the colder winter months. However, it is still unclear if the data collected could actually be indicative of foraging patterns during any other time of the year. Since the variable temperatures were quite atypical for fall weather in Ohio, it is unclear what effect this may have had on the foraging behavior of the squirrels. A more consistent weather pattern could be better for determining a true connection between squirrel foraging behavior and location preferences. In a scientific paper, (Dill, 1989), the research team ran their experiments between June 26 and July 24. This period of time had more consistent temperature readings which helped to remove a lot of environmental confounding variables. Ideally, future studies should employ a longer period of data collection and account the temperature for days of the on-going experiment. This would help determine if the trends that were witnessed were actually consistent year-round or if squirrels exhibit seasonal differences in their foraging behavior as a result of fluctuations in the temperature and climate. Behavioral ecology is a fairly well researched field and has been an increasingly popular field of study in recent years. However, even with its popularity there are still gaps in research within it. This study has helped to fill in a few of those gaps, but with every gap that is filled another question arises. The biggest gap in knowledge that remains after this study is how the ecology of fear and perceived risk of predation would affect the grey squirrel’s choice of foraging location. This study gave a surface level understanding of preferred choice of foraging location, but there are many other factors that could also help determine the GUD of a foraging squirrel. Any future researchers who wish to pursue and improve upon the methods and results of this study should take those two factors into account. A good idea would be to run an experiment where the same two treatments groups are used but to also use another set of variables. These variables could include but are not limited to playing sounds of predators near the sheltered and unsheltered trays or even using models to see if this has an effect on the GUD for each treatment. This would help to further clarify how squirrels go about their foraging habits and would be a great step forward in helping ecologists understand how animals adapt their foraging strategies based on the changes in their environment. Acknowledgements A lot of work went into this project and none of it would have been possible without the help of some very important people. First and foremost, we would like to thank Dr. Curtis and Dr. Hovick for giving us the opportunity to conduct research and providing us with the guidelines for the project. Next, we would like to thank our lab TA Rosa Rodriguez-Pena for an amazing amount of help and guidance on the project. We especially appreciated her help in understanding the statistical software that was used for the data analysis. Additionally, we would like to thank our GTA Whitney King for her feedback on our initial drafts which helped us greatly in terms of focusing and correcting our ideas to write this final draft. Finally, we would like to thank The Ohio State University for allowing us to use their campus to conduct our experiments. Literature Cited: Bowers, M. A., & Breland, B. (1996). Foraging of Gray Squirrels on an Urban-Rural Gradient: Use of the Gud to Assess Anthropogenic Impact. Ecological Society of America, 6(4), 1135–1142. Dill, L. M., & Houtman, R. (1989). The influence of distance to refuge on flight initiation distance in the gray squirrel (Sciurus carolinensis). Canadian Journal of Zoology, 67(1), 233-235. Pospiech, T., Levinson, B., & Wind, H. (2012). Giving-up densities of gray squirrels in a forest and suburban environment and the effect of substrate depth. Eukaryon, 8(1), 94-96. Appendix Trial Sheltered Unsheltered Total 1 112.5 179 291.6 2 181.2 145.5 376.7 3 177 194 371 4 155 200 355 5 153 148 351 Total 778.7 466.5 1745.2 Starting Weight: 202 grams