Comparison of Nest Predation in Two Riparian Habitats Alexandra Kaye Department of Biology, Middlebury College, Middlebury, VT 05753, USA ABSTRACT Clearing of riparian habitat in Monteverde, Costa Rica is prohibited up to 50 meters from the bank of t he waterway. Unfortunately, this regulation is often disregarded, and forests bordering streams are cleared for a variety of purposes. Deforestation creates edge effects where there once was continuous forest, and it has been shown that rates of avian nest predation are higher on forest edges. In this study, I examined the proportions of eggs removed or damaged from artificial nests in four riparian habitats. Two of the chosen sites were forested, and two had been cleared for human use. The artificial nests contained two quail eggs and four clay eggs. The study lasted nine nest nights, and after the morning of each I returned to the site to census the nests and collect any bitten clay eggs for predator identification. I found no significant differences among predator types in the four sites. Contrary to my expectations, though, I found that the site with the greatest proportion of eggs removed or damaged was a forested site. The second forested site and one of the deforested sites were comparable in terms of nest predation proportions. The results of my study did not support my predictions that, in deforested sites, both nest predation and the incidence of bites by predators that thrive in human modified habitats â€“ coatis and opossums among them â€“ would be sig nificantly higher. Nonetheless, both the use of artificial nests and predatorsâ€™ search behavior modification in response to high frequency of occupied nests may have affected the results of this study. Future studies should concentrate on these two factors before reaching any conclusions on the impact of riparian deforestation on avian nest predation. INTRODUCTION Riparian habitats in Costa Rica are formally protected under the Ley Forestal up to 50 meters if in steep terrain (Ley Forestal 7575, Artcul o 33). However, these regulations are frequently flouted, with the result that it is fairly simple to find degraded riparian habitat. In rural parts of Monteverde, where terrain is steep and therefore the banks of streams should be protected up to 50 met ers, riparian habitats are often cleared for agriculture or to provide pastureland. This clearing necessarily modifies the habitat, disrupting delicate ecological interactions; previous studies have found higher rates of avian nest predation along the habi tat edges that are created when forest is cleared (Batary 2004). The effect of clearing riparian habitats on nest predation is interesting to examine because it gives insight into the species composition of these areas. Moreover, as a primary cause of n esting mortality for numerous bird species, nest predation can act as a lens into the overall health of an avian population; if rates of predation are too high, certain bird populations could become severely threatened (Martin 1987). In this study, I soug ht to evaluate the efficacy of riparian protection regulations by comparing proportions of disturbed nests in forested versus deforested habitats, as well as
2 the incidence of bites by different species of predators. Through these experiments, I hoped to ma ke inferences about the health and species composition of both forested and deforested riparian habitats. I expected to find higher rates of nest predation in the deforested sites, and lesser incidence of bites made by species like squirrels and mice that prefer habitats with high tree density. METHODS AND MATERIALS Study Sites I included four riparian sites in my study, two of which were forested, and two of which were deforested. The first of the deforested sites had been cleared for pastureland. The banks of the stream had tall grasses growing on them, which were periodically cut down to prevent them spreading into the pasture. The first of the forested sites was located along the same waterway, but on the opposite side of a roadway, where the land ha d not been cleared. The second of the deforested sites occupied a cleared hillside bordering one side of a stream. The second forested site was located along the same waterway, and was directly adjacent to the second deforested site. Comparison of Nest Predation Rates I built nine nests for each site, with each nest consisting of four artificial clay eggs colored and shaped like quail eggs, and two actual quail eggs. Quail eggs have been used in similar experiments assessing nest predation in the Neotropics because they are both small and commercially available in large quantities (DeGraaf and Maier 1996). I distributed these nests along the ground of the four sites, at various elevations and proximity to the waterway. I returned the morning following setting up the nests to census the number of eggs â€“ both quail and artificial â€“ removed from the site, as well as the number of artificial eggs displaying bite marks, sign of a predator having bitten the artificial egg and rejected it upon realizing it wa s inedible. I repeated this QuickTime and a decompressor are needed to see this picture. QuickTime and a decompressor are needed to see this picture. QuickTime and a decompressor are needed to see this picture. QuickTime and a decompressor are needed to see this picture. QuickTime and a decompressor are needed to see this picture. Figure 1. Study sites. Clockwise from top left: Forested 1, Deforested 1, Forested 2, Deforested 1 Figure 2 . Artificial nest containing both clay and quail eggs
3 methodology for nine â€œnest nights,â€ or cycles of placement and census of nests. To compare proportion of eggs removed or damaged from the sites, I established the average proportions across three categories: quail eggs removed, clay eggs removed or damaged, and total eggs removed or damaged. I then used a Wilcoxen Rank Sum test to compare these three categories of proportions in each of the four sites. Each morning that I returned to the study sites to census the nests, I also collected any artificial eggs displaying bite marks. I later identified the predator that had made each bite, grouping the data for Forested Sites 1 & 2 and Deforested Sites 1 & 2. I evaluated the relationship between bite incidence and site using a Chi Sq uared test. Additional Observations After taking a census of the nests from the fourth nest night, I began to suspect that some predators, particularly in Forested Site 2, had modified their behavior in response to high density of nests. In several nest s where clay eggs had been bitten for the first two nights, now the quail eggs were removed but the clay eggs were completely untouched. To avoid a bias in the proportions of eggs removed or damaged, I began to place disturbed nests in new sites starting o n the fifth night. I took this change into account by calculating proportions in each of the three categories for the first four nights and then the final five nights. I then compared these proportions for each of the sites using an F test. RESULTS Both the average proportion of quail eggs removed and the composite proportion of eggs removed/damaged in Forested Site 2 were significantly higher than those of the three other sites. The proportion of clay eggs removed/damaged in Forested Site 2, however, did not differ significantly with that of Deforested Site 1. In all three categories, the differences in the proportions of eggs removed/damaged from Deforested Site 2 and Forested Site 1 were statistically negligible. The proportion of eggs removed/damaged i n Deforested Site 1 was consistently higher than that of Deforested Site 2; the same was true of Forested Site 2 in comparison to Forested Site 1 (Table 1). Comparison of Incidence of Bites A Chi Square test yielded no statistically significant differen ces in observed versus expected incidence of bites for each predator species in forested and deforested sites (Figure 3). Additional Observations I used an F test to look for variance in the proportions of eggs removed or damaged from the four sites dur ing nights one through four and that of nights five through nine. I found no significant variance in the average proportion of quail eggs removed, artificial eggs removed/damaged, or total eggs removed/damaged among any of the sites (df = 1, critical F val ue = 647.8).
4 DISCUSSION The fact that the highest proportion of quail eggs removed and composite damage occurred in a forested site contradicts my predictions for this study. I expected to find the greatest proportion of eggs removed or damaged in the d eforested sites, based on previous research on the impacts of edge effects and habitat fragmentation on avian nest predation (Batary 2004). However, Forested Site 2 was adjacent to a cleared site, Deforested Site 2; previous studies have found that nest pr edation may also be affected in forested areas by type of land use on bordering fragments (in the case of this study, the cleared riparian site) (Small and Hunter 1988). In considering what factors may have contributed to my results differing from my expectations, it is important to take into account both the limitations of artificial nest usage in examining nest predation and to consider the possibility that predator search behavior adapted to my study sites. The artificial nests were less camouflaged than real nests, and so they may have been more attractive to visual predators (Martin 1987). In addition, predators may become visually accustomed to a certain speciesâ€™ nests, developing a â€œsearch imageâ€ and increasing the overall efficiency with which they forage. Since my nests were all identical in composition and very similar in appearance, I may inadvertently have created a situation in which predators were able to learn quickly precisely what sorts of nests to look for (Martin 1988). This effect may have contributed to the pattern of egg removal in Forested Site 2. Predator search behavior is both density and reward frequency dependent. Predators may modify their behavior, increasing their foraging intensity, if the frequency of nests containing eggs i s high; this finding is supported by data from experiments using both real and artificial nests (Martin 1988). All my nests were occupied, and their density was relatively high. This could have intensified searching behavior in predators, because it facilitated a high reward rate. Unnaturally high predation rate, therefore, may have biased the results of this study (Martin 1988). This studyâ€™s scope was not species specific; nest predation rates were generalized, and the effects of riparian deforestation on particular avian species were not examined. This would be a good next step in comparing the health of the study sites. There is concern that conversion of forest to pastureland reduces the availability of nesting habitat for birds that prefer the forest i nterior for nesting, threatening certain avian species (Lindell 2003). This threat would not necessarily be shown in a study that looks at overall nest predation without distinguishing between the nests of particular species. Although quail eggs and clay or plasticene eggs are commonly used in nest predation experiments, differences in predation rates between real and artificial nests do exist. Of particular concern to this study is that a previous study found that the quail eggs in nests with plasticene e ggs marked by mice were in fact, upon video recording, found to have been removed by raccoons (Thomas and Burhans 2004). This finding opens the door to the possibility that the numerous mice markings were not necessarily indicative of predation by mice spe cies. Only one species of raccoon is found in the Monteverde area, and since no raccoon bites were recorded throughout the study, it is unlikely that raccoons were the predators of these micemarked nests. However, the findings of Thomas and Burhans might suggest that, in this study, the mice bites were misleading because the actual predation of quail eggs was carried out by other mammals, such as coatis or squirrels.
5 The incidence of mouse bites in forested sites was not statistically greater than that in deforested sites, but, outside of statistical analysis, the incidence of bites in forested sites was relatively high. Regardless of whether or not it was the mice who actually removed the quail eggs, the relatively high incidence of mouse bites in the forested sites may speak to an overall greater abundance of mice in those habitats. Of the four species of pocket mice found in Costa Rica, all but one prefer well forested habitats; only Liomys salvini is found in brush or weedy fields (Reid 2009). The exp ected incidence of opossum bites was about half that of the observed incidence in deforested sites. Didelphis marsupialis , which is abundant in Costa Rica, favors disturbed areas and rural garbage dumps, two indicators of human presence. Moreover, the D. marsupialis is frequently found by stream banks (Reid 2009). These three factors could explain the unexpectedly high incidence of opossum bites in the deforested sites. Five species of squirrel are found in Costa Rica, but one is endemic to the Guanacaste Mountains. These four species is are diurnal and spend most of their time in trees, descending to the ground mainly for feeding. Because they need trees, they may be more readily seen in forested areas, and two of the species found in Monteverde â€“ the Mont ane Squirrel and Red tailed Squirrel â€“ are especially associated with secondary growth, which could explain the (albeit not statistically significant) greater incidence of squirrel bites in the forested sites in this study (Reid et al. 2010). The results o f this study, while not supporting my hypothesis that nest predation would be higher in deforested sites, instead suggest a new possibility altogether: that riparian sites, be they forested or cleared, function as forest fragments. Riparian sites, if they exist in areas modified by humans for agricultural purposes, are not likely to represent healthy forests, even if they remain forested, simply because the surrounding areas are cleared. For this reason, it may be best to consider the potential of these rip arian sites to function are corridors, linking larger, healthy forests that are not interrupted by agriculture. Forested riparian habitats in otherwise deforested areas could have great value as refuges for meso predators, which in turn serve as prey for l arger mammals at higher trophic levels. In this manner, riparian habitats could have tremendous conservation value, and should be protected accordingly. ACKNOWLEDGMENTS I would like to thank Karen Masters, my advisor in this project, for her help with structuring my study, dealing with problems that came up along the way, and showing me the appropriate statistical tests and how do use them. I would like to thank Moncho Caldern for identifying which predators had made the bites on the artificial eggs a nd helping me find sources. I would like to thank Raquel Martinez for her insight into my project and for helping with statistical analysis, as well as helping me find sources. And thank you also to Lizzy Schiller, who was a huge help in setting up the study sites! LITERATURE CITED Batary, P. and Baldi, Andras. 2004. Evidence of an Edge Effect on Avian Nest Success. Conservation Biology. 18:389400. DeGraaf, R. M. and Maier, T. J., 1996. Effect of Egg Size on Predation by White Footed Mice. The Wilson Bu lletin, 108: 535 539.
6 Ley Forestal, Artculo 33. (1996) La Gaceta, Imprenta Nacional, La Uruca, San Jos, Costa Rica. Lindell, C. and Smith, M., 2003. Nesting bird species in sun coffee, pasture, and understory forest in southern Costa Rica. Biodiversity a nd Conservation, 12: 423 440. Martin, T. E., 1987. Artificial Nest Experiments: Effects of Nest Appearance and Type of Predator. The Condor, 89: 925928. Martin, T. E., 1988. On the advantage of being different: Nest predation and the coexistence of bird s pecies. Proc. Natl. Acad. Sci. USA, 85: 2196 2199. Martin, T. E., 1988. Processes organizing opennesting bird assemblages: competition or nest predation? Evolutionary Ecology, 2: 3750. Reid, F. A., 2009. A Field Guide to the Mammals of Central America & Southeast Mexico. Oxford University Press, New York. Reid, F. A., Leenders, T., Zook, J., and Dean, R., 2010. The Wildlife of Costa Rica: A Field Guide. Comstock Publishing Associates, New York. Small, M. F. and Hunter, M. L., 1988. Forest fragmentation a nd avian nest predation in forested landscapes. Oecologia, 76: 6264. Thomas, F. R. and Burhans, D. E., 2004. Differences in Predators of Artificial and Real Songbird Nests: Evidence of Bias in Artificial Nest Studies. Conservation Biology, 18: 373 380.
7 (a) Proportion of Quails Eggs Removed Comparison Rank Difference Standard Error q value D1 vs. D2 99.5 31.6 3.15 D1 vs. F1 60 31.6 1.90 D1 vs. F2 107.5 31.6 3.40 D2 vs. F1 39.5 31.6 1.25 D2 vs. F2 207 31.6 6.55 F 1 vs. F2 167.5 31.6 5.30 (b) Proportion of Clay Eggs Removed/Damaged Comparison Rank Difference Standard Error q value D1 vs. D2 136 31.6 4.30 D1 vs. F1 69.5 31.6 2.20 D1 vs. F2 49.5 31.6 1.57 D2 vs. F1 66.5 31.6 2.10 D2 vs. F2 185.5 31.6 5.87 F1 v s. F2 119 31.6 3.77 (c) Composite Proportion of Eggs Removed/Damaged Comparison Rank Difference Standard Error q value D1 vs. D2 116 31.6 3.67 D1 vs. F1 67 31.6 2.12 D1 vs. F2 89 31.6 2.82 D2 vs. F1 49 31.6 1.55 D2 vs. F2 205 31.6 6.49 F1 vs. F2 15 6 31.6 4.94 Table 1. Pairwise comparisons of the eggs from deforested sites (D1 and D2) with forested sites (F1 and F2). Crit ical q value = 2.78; p < 0.05. Composite proportion combines proportions of quail eggs removed and clay eggs removed/damaged.
8 Figure 3. Incidence of bites . Chi square d value = 7.33; df = 4; critical = 9.49. 0 5 10 15 20 25 30 35 40 45Incidence of Bites Deforested, observed Forested, expected Forested, observed Deforested, expected COATI OPOSSUM MOUSE RAT SQUIRREL
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xlink http:www.w3.org1999xlink xmlns:xsi http:www.w3.org2001XMLSchema-instance
leader 00000nas 2200000Ka 4500
controlfield tag 008 000000c19749999pautr p s 0 0eng d
datafield ind1 8 ind2 024
subfield code a M39-00262
Comparativa de depredacin de nidos en dos hbitats ribereos
Comparison of nest predation in two riparian habitats
Clearing of riparian habitat in Monteverde, Costa Rica is prohibited up to 50 meters from the bank of the waterway. Unfortunately, this regulation is often disregarded, and forests bordering streams are cleared for a variety of purposes. Deforestation creates edge effects where there once was continuous forest, and it has been shown that rates of avian nest predation are higher on forest edges. In this study, I examined the proportions of eggs removed or damaged from artificial nests in four riparian habitats. Two of the chosen sites were forested, and two had been cleared for human use. The artificial nests contained two quail eggs and four clay eggs. The study lasted nine nest nights, and after the morning of each I returned to the site to census the nests and collect any bitten clay eggs for predator identification. I found no significant differences among predator types in the four sites. Contrary to my expectations, though, I found that the site with the greatest proportion of eggs removed or damaged was a forested site. The second forested site and one of the deforested sites were comparable in terms of nest predation proportions. The results of my study did not support my predictions that, in deforested sites, both nest predation and the incidence of bites by predators that thrive in human-modified habitats coatis and opossums among them would be significantly higher. Nonetheless, both the use of artificial nests and predators search behavior modification in response to high frequency of occupied nests may have affected the results of this study. Future studies should concentrate on these two factors before reaching any conclusions on the impact of riparian deforestation on avian nest predation.
La tala del hbitat ribereo en Monteverde, Costa Rica est prohibida hasta 50 metros desde la orilla de la va acutica. Por desgracia esta regulacin es a menudo ignorada, y los bosques que bordean las Quebradas son talados para una variedad de propsitos. La deforestacin crea efectos de borde donde una vez hubo bosques continuos, y se ha demostrado que las tasas de depredacin de nidos de aves son ms altas en los bordes del bosque. En este estudio, examine la proporcin de huevos eliminados o daados de los nidos artificiales en cuatro hbitats ribereos. Dos de los sitios elegidos fueron bosques boscosos y dos haban sido talados para uso humano. Los nidos artificiales contenan dos huevos de codorniz y cuatro huevos pintados. El estudio duro nueve noches, y volv todas las maanas al sitio para hacer un censo de los nidos y recoger los huevos mordidos para identificar al depredador. No encontr diferencias significativas entre los tipos de depredadores en los cuatro sitios. Contrariamente a mis expectativas, sin embargo, me di cuenta de que el sitio con la mayor proporcin de huevos retirados o daados era un sitio boscoso. El Segundo sitio boscoso y uno de los sitios deforestados fueron comparables en trminos de proporciones de depredacin de nidos. Los resultados de mi estudio no apoyaron mis predicciones de que, en los sitios deforestados, tanto la depredacin de nidos y la incidencia de mordeduras de depredadores que se alimentan en hbitats modificados por el hombre pizotes y zarigeyas, entre ellos- seria considerablemente mayor. No obstante, tanto el uso de nidos artificiales y bsqueda de modificacin del comportamiento predador en respuesta a la alta frecuencia de nidos ocupados puede haber afectado los resultados de este estudio. Estudios futuros deberan centrarse en estos dos factores antes de llegar a ninguna conclusin sobre el impacto de la tala de rboles de ribera en la depredacin de nidos de aves.
Text in English.
Costa Rica--Puntarenas--Monteverde Zone
Costa Rica--Puntarenas--Zona de Monteverde
Tropical Ecology Summer 2010
Ecologa Tropical Verano 2010
t Monteverde Institute : Tropical Ecology