The Effects of Patch Depth and Cover on Giving-up Density for Passer domesticus (House Sparrow) at UIC

The Effects of Patch Depth and Cover on Giving-up Density for Passer domesticus (House Sparrow) at UIC

Adrian Raygoza (Senior BioS undergraduate major)

The University of Illinois at Chicago

1200 W Harrison St, Chicago, IL 60607

Introduction:

This study took place at the University of Illinois at Chicago. The emphasis of this study was to better understand two different variables that affect the giving-up density of the house sparrow, Passer domesticus; which is a small bird that is commonly found on and around the university campus. The giving-up density (GUD) is a point where a foraging organism decides to stop foraging due to a variety of factors. Usually, these factors can include high handling time, metabolic cost, predation risk, and missed opportunity cost. These factors can vary by situation and by species. Smaller birds, such as the house sparrow, face increased pressures of predation while foraging; as they are a common food source for larger predatory birds (Macleod et al. 2003). Costly and deadly risks are associated with any given variable that prolongs a foraging attempt. For example, when a species ventures too far into the open, away from the safety of cover, and predators easily locate them (Macleod et al. 2003). Sometimes the pressures of these risks can grow too high, depending on the factor involved, causing a species to forgo a meal for their own safety.

In this study, we were interested in two different independent variables that were compared against the GUD of the Passer domesticus. One factor that was tested was the depth of the food (substrate) in a given food patch. In a study investigating mourning doves and cottontail rabbits, by Baker and Brown (2009), deeper substrates led to a higher GUD in those species. We were interested in testing this factor for comparison with the species Passer domesticus. The second factor tested was the effectiveness of cover given by the canopy a focal tree to a given food patch. Oyugi and Brown (2014) conducted a study that suggests the GUD was small when a patch was within cover by a canopy of a focal tree, as opposed to further away in an open field.

Our study combined both of these elements into a two by two factorial design. Our factors included depth of substrate (food) in a patch and the effectiveness of cover. Each of these factors was given two levels, shallow and deep for substrate, and cover and no cover of a food patch. This study took span over three days, testing both variables in comparison to the giving-up density of the house sparrow. We hypothesized that if the sparrow has a shallow depth of food and optimal tree-cover, then the GUD will be minimized.

Materials and Methods:

The site used for this study was the Chicago Circle Memorial Grove, located on the southeast end of the university grounds. This site boasts some large trees that provide a good amount of cover and small, open grassy fields. The area has a light, but constant amount of foot traffic by students and staff. The food patches were created using two different sized treys. Two larger treys were used to create a shallow substrate patch, as the longer length stretched the sand a food across a larger area creating a shallow depth for food access. Two smaller treys were used to create deeper substrate food patches, as the sand and food were compressed into a smaller area pushing the food into deeper depth. Four treys were used total for the study. Each tray was filled with 10-grams of black oil sunflower seeds as the substrate. Each tray was also filled with 1000-grams of sand poured over the seeds. The food patches were weighed, measured, and kept constant for each given day of testing.

The food patches were set near and underneath the canopy of a focal pine tree located on the site. Two of the trays, one large and one small, were placed underneath the focal pine tree canopy, which was used to create the cover. Two of the trays, one large and one small, were placed approximately 20 meters away from the focal tree canopy, in the open field. The trays were placed out in the morning and picked up in the afternoon. They remained outside for about 7-8 hours each day of the study. Once the trays were collected, they treys were emptied out over a sifter that separated the sand from the remaining bird seed. Once fully separated, the remaining seeds were weighed in grams. The weight of remaining seeds, in grams, was used as the measurement of GUD in grams. The data was collected over a span of three days. All data were uploaded into Microsoft Excel. The mean and standard deviation was calculated and used to run a Two-Way Factorial ANOVA test, for independent samples, using the Vassar website.

Results:

Using a visual comparative method, as seen in Fig. 1 and Table 1, the trend in the data showed a dramatic difference in comparison of GUD for food patches in cover and in the open. The food patches in the open had higher GUDs at an average of 9.00 grams with shallow food and an average of 8.78 grams for deep food. The patches underneath cover had a much lower average GUDs at 3.11 grams for shallow food and 2.30 grams for deep food.

In comparison, the two depths of the substrate showed larger differences in cover and open field as seen in Fig. 2. The average GUDs for the deep (smaller) food trays were 2.30 grams for open and 8.78 grams for cover. The average GUDs for shallow (larger) food trays were 3.11 grams in cover and 9.00 grams in the open field. A Two-Way Factorial ANOVA test, for independent samples, was run on the data and revealed that the trey type factor, that was used for the creating depth of the substrate, had a P-Value of 0.0729, with 1 degree of freedom. The ANOVA test also revealed that the cover provided by the focal tree factor, open (no cover) and cover (under the tree), had a P-value of <.0001, with 1 degree of freedom.

Discussion:

Similar to the research work produced by Oyugi and Brown (2014), our data showed a trend toward the effectiveness of cover. In our data, as shown in Fig. 1, there is a large difference between the GUD between the open field food patches and the food patches that were placed under the canopy of the focal tree. On average, GUD remained low for the covered patches. This is about what was expected to be seen. According to the T-test that was run, the P-Value was <0.001, which supports a statistical difference between the location of the patches. We reject the null hypothesis that there was no significant difference in between patch location. The data also supported our hypothesis that the shallow depth of food and optimal tree-cover would minimize the GUD. However, we noted that the lowest overall GUD recorded was not the combination of shallow, but of deeper depth and optimal tree-cover. The findings were slightly askew from what we expected.

Unlike the work of Baker and Brown (2009), our data did not support a higher GUD for deeper substrates. Though there was a higher GUD, in Fig. 2, deeper substrate depth was paired with an open field, there was also a lower GUD when deeper depth was paired with the cover. This was mismatched pairing was very similar for shallow depth as well, as seen in Fig. 2. The T-test that was run using the average GUD of the data collected had a P-Value of 0.0729, which supports that there was no significant difference between food depth. In this case, we fail to reject the null hypothesis that there is no significant difference between shallow and deeper food depth. This finding did come as a surprise, as we expected to see shallow food patches have a significantly lower GUD than deeper patches. This was not the case. The reason for this expectation was that we presumed that a deeper depth would cause time constraint and produce a higher handling time, metabolic cost, predation risk, and missed opportunity cost; especially given that shallow patches were not too far away.

Because the shallow and deep substrate patches showed the similar trends, relevant to the data, and were not statistically different, based on the T-test, we could not conclude that food depth played a significant role in the GUD of Passer domesticus in given food patches. However, with the support of statistical analysis, and the relevant difference in trends in our data, we can conclude that cover by tree canopy does play a significant role in the GUD of Passer domesticus in given food patches. For further studies, we would like to include food type as a variable, as well as the time of day. We would also like to acknowledge that though giving-up density is a common approach to ecological foraging studies regarding food patches, there are shortcomings in studies revolving around GUD; as argued by Perez et al. (2013). Some of the shortcomings are due to the several exclusion factors, such as the behavior of different species, differences in foraging techniques, and multiple species visiting designated food patches (Perez et al. 2013). Though it can be difficult, and nearly impossible in some cases, to stop visits from multiple species, we can modify food type to specific diets of the test species. We can also set up feeders that are placed to the appropriate favorable distances and heights to specific species; as some prefer ground feeding, and some do not. Moving the food patches to a less foot trafficked area, by humans, may also play a role in findings. An overall compilation about the effect of different factors can help lead to accurate and effective data findings in GUD. 

Literature Cited

Baker, M. A. A., and J. S. Brown. 2009. Patch area, substrate depth, and richness affect giving-up densities: a test with mourning doves and cottontail rabbits. Oikos 118:1721–1731.

Bedoya-Perez, M. A., A. J. R. Carthey, V. S. A. Mella, C. Mcarthur, and P. B. Banks. 2013. A practical guide to avoid giving up on giving-up densities. Behavioral Ecology and Sociobiology 67:1541–1553.

Macleod, R., P. Barnett, J. Clark, and W. Cresswell. 2006. Mass-dependent predation risk as a mechanism for house sparrow declines? Biology Letters 2:43–46.

Oyugi, J. O., and J. S. Brown. 2003. Giving-Up Densities And Habitat Preferences Of European Starlings And American Robins. The Condor 105:130.

Appendix

This data was collected by Adrian Raygoza and Shivram Patel from 06/03/2019-06/06/2019. One day of data was lost due to rain.

Table 1 GUD measurements for 2-factor variables Position from focal tree and Substrate Depth 

Table 2 Raw Data Collection

Table 3 Statistical Values from Vassar

Figure 1 Graph of GUD Measurement for Open vs Cover for each depth of substrate

Figure 2 Graph of GUD Measurement for Deep vs Shallow for position from the focal tree

Figure 3 A picture of the two different trey sizes with the same substrate used for the study

Figure 4 A Picture of the food patch under cover of the focal tree canopy

Figure 5 A Picture of the study site where the cover and open field were tested at Chicago Circle Memorial Grove on the UIC Campus