Changes in Hummingbird Species Richness and Abundance in a Forest Fragment and Agricultural Ecotones Stephen Smith Dept. of Biology, Occidental College, Los Angeles, CA 90041 Abstract Hummingbirds are important pollinators in tropical forests, and their behavior is vitally important to the plants they pollinate. As the tropics become more and more fragmented, hummingbirds will play the increasingly important role of long distance pollination of many plants between isolated tracts of forest. This study loo ked at the different responses of hummingbirds to a forest fragment and the surrounding agricultural land. Hummingbirds preferred edge to open areas and forest, with three feeders out of 18 getting 126 of the 482 total visits, compared to 91 total visits o f the second most visited feeder. Further, the birds preferred an open field and a coffee plantation to a banana plantation, the number of visits being 191, 171, and 120, respectively. Most interesting was the partitioning by two species of hummingbird, on e preferring the forest and the other preferring open areas. The differing behavior of these two species could lead to very different pollination patterns in the open areas and the forest fragment, and could serve to isolate forest fragments in terms of ge ne flow for hummingbird pollinated plants. Resumen Los colibrÃes son muy importantes en los bosques tropicales porque ellos polinizan las flores. Su conducta es tambiÃ©n muy importante para las plantas que ellos polinizan. Porque los bosques se han cortand o mucho, los colibrÃes serÃ¡n mas importantes porque ellos pueden transportar polen de un fragmento a otro. Este estudio observo como responden los colibrÃes a un fragmento y la zona agrÃcola circundante. Los colibrÃes prefieren el margen y no el bosque o l a zona agrÃcola , porque el margen recibio 126 de 482 colibrÃes total. TambiÃ©n , los colibrÃes prefieren la tierra abierta y no la plantaciÃ³n de banano, con 191 colibrÃes en el pasto , 171 en el cafetal, y 120 en la plantaciÃ³n de banano. TambiÃ©n , dos especie s de colibrÃes se segregaron entre el bosque y la zona abierta. Esta conducta diferente entre los dos tipos puede resultar en un arreglo de polinizaciÃ³n muy diferente entre el fragmento y la zona abierta, y puede separar los fragmentos porque polen no va a ir entre los fragmentos. Introduction Human caused disturbance affects tropical forests worldwide, and agricultural clearing is a major reason for the rampant destruction of forests Terborgh 1992, Vitousek et al. 1997. When clearing forest for agriculture, farmers often leave small patches o f forest intermixed with farmed land. These patches act as refugee for the original species present
in the area, and are being viewed as increasingly important as source pools if and when the forest is allowed to re generate Chazdon 1998. However, fragm entation also affects the ecosystem within the relict patches of forest. Edge effects, conditions caused by proximity to open land, have been well documented and include drying, increased temperature, lowered reproductive success, and lowered species richn ess in fragments compared to continuous forest Didham and Lawton 1999; Debinski and Holt 2000. These effects can be found up to 100 meters into a forest Didham and Lawton 1999 and because patches are often small, edge effects permeate the entire patch. The biological impact of these edge effects is far reaching. Bruna 1999 found that seed germination in patches is far less than in forest. Laurance et al. 1997 showed that forest patches lost up to 14% of biomass following fragmentation. Aldrich and H amrick 1998 demonstrated a genetic bottleneck in a tree species in a forest patch; a few Â€superadultsÂ produced over 50% of total seedlings in a patch. The species richness, abundance of plants and animals, and the genetic diversity is often not equal to pre fragmentation levels, resulting in a source pool that may not reflect the original composition of the forest. Because forest patches are often tightly bounded by agricultural land, they are ideal places to study species compositional changes along ec otones. Natural ecotones, a change from one type of environment to another, are often very large, spanning hundreds of miles Smith et al., 1997. The differing selective pressures along ecotones influence species composition. The selective pressures can b e so different that Smith et al. 1997 demonstrated possible speciation of birds occurring along an ecotone in Africa. However, because of the large size and unclear boundaries of natural ecotones, they are hard to study. Ecotones along forest patches are easier to study because in the span of 100 meters, the environment can change drastically, along with selective pressures and species composition. Hummingbirds are vitally important to the ecological health of an ecosystem; they pollinate almost four per cent of flowering plants in Costa Rican lowland rainforest Bawa, 1990. Their importance increases in fragments, where they can increase genetic flow between patches by their ability to carry pollen long distances Bawa 1990. While many bird species disa ppear from fragments Rappole and Morton 1985, hummingbird abundance may increase in fragmented areas, due to additional open area species moving in to cleared areas Feinsinger 1988 or due to their plasticity in habitat requirements allowing them to adj ust to disturbance Bierregaard and Stouffer, 1997. Behavioral changes in hummingbirds living in fragments, such as increased intraspecific competition Feinsinger 1988 and changes in foraging behavior Aldrich and Hamrick 1998 have been documented. Shi fts in pollinator behavior can cause drastic effects on the plants they pollinate, such as inbreeding and genetic bottlenecking Aldrich and Hamrick 1998. Understanding the changes in vital ecosystem interactions caused by fragmentation will become an inc reasingly important part of conservation biology as clearing and fragmentation becomes more prevalent. This study recorded the changes in species composition and abundance of hummingbirds in a small patch of forest and surrounding agricultural land. The f orest
abruptly yielded to coffee and banana plantations as well as grazing pasture. Thus, I present a comparative study of hummingbird responses to different ecotones in an area small enough to negate any distance dependant or elevation dependent differenc es. The different species present along different ecotones and their differing behaviors in agricultural lands influence the plant pollinator interactions, and thus the overall species composition, of the forest patch. Materials & Methods This study was c onducted in a forest patch in CaÃ±itas, just outside of Monteverde, Puntarenas, Costa Rica, at approximately 1400 meters elevation. The patch was surrounded by a 2.5 ha. Coffee plantation, a one hectare c ow grazing pasture , and a one hectare banana plantati on. The forest patch was approximately 1.5 hectares, triangular, and bounded on one side by a stream. See Figure 1 The patch was created approximately 40 to 50 years ago, and is largely undisturbed by humans and livestock. Eighteen identical hummingbird feeders were placed along three transect lines. The feeders were red plastic with white flowers around four feeding holes, and provided four perches. Each line ran from the forest patch into either the coffee, bananas, or pasture land. Two feeders were pl aced in the forest, one on the edge, and three in the agricultural land. Each transect line held six feeders, each 20m apart, for a total of 120 m. Feeders were filled with sugar solution four parts water to one part sugar, boiled which was changed every four days to avoid contamination, and were never allowed to run dry for longer than twelve hours. Along each transect line; feeders were numbered one through six; one being in the agricultural land, four on the edge, and six in the forest. Feeder number one in the field was hung on a solitary tree four meters high, two and three were hung from sticks stuck in the ground. Four was hung on a Piper auratum over the stream on the edge. Five and six were suspended from small trees in the forest. In the coff ee transect, feeders one through three were suspended from coffee bushes along the same row. Feeders four through six were suspended from small understory trees, with four on the edge and five and six in the forest. In the banana transect, feeders five and six were hung from trees in the forest, with five being relatively close to the stream. Four was hung on a small tree on the edge, and one through three were hung on banana trees. All feeders were between one and 1.5 meters above the ground.
Figure 1: The Study Site. Feeders were placed along three transect lines, and numbered 1 to 6. Feeder Number one was 60 m into the field, number four was on the edge, and number six was 40 m into the forest patch. I watched each feeder for twenty minutes at a time, moving along each transect line. I staggered times morning [8 10am] or afternoon [3 5pm] and observation order to eliminate any temporal bias caused by the daily activity patterns of the birds. I sat quietly approximately ten meters away from each feeder and recorded number of visits and species. I identified individuals using a field guide Stiles and Skutch 1989 and I confirmed them to be living in the area using a local checklist Fogden, 1993. I used the Friedman test to find significant diff erences among feeders in different positions along the same transect line and among feeders in the same position along different transect lines. I also used Chi square tests to identify species preferences for different transects or feeder positions. Resul ts I watched feeders a total of 56 hours, during which I recorded 482 visits. I observed a total of seven species foraging at the feeders See Appendix 1, but only three species were observed with great enough frequency more than ten visits and over a l arge enough range of feeders to provide patterns. These were the Violet Sabrewing Campylopterus hemileucurus the Striped tailed Hummingbird Eupherusa exima and the Rufous tailed Hummingbird Amazalia tzacatl .
By visits All of the within forest sites were relatively similar according to the Friedman tests. In a comparison of forested sites six and site five across the three ecotone transects p = .1354 and .358, respectively, the test showed no significant difference in the number of visits Table 1. Feeders five and six combined recorded 47 visits in the pasture, 61 in the bananas and 51 in the coffee. The edge stations four and the first agricultural station three were also similar p = .100 and .4832, respectively. Combined, the feeders record ed 86 visits in the pasture, 52 in the bananas and 76 in the coffee. Although the field edge stations showed considerably more visits than the others 58, compared to 31, and 37 visits the differences were not significant. However, the two stations farthe st from the forest two and one were significantly different, p= .015 and .002, respectively. These feeders combined for 57 visits in the pasture, 44 visits in the coffee, and only seven in the bananas both banana plantation feeders five and six were significantly different from the others. The total number of visits to the banana 120 was significantly less than the number of visits to either the coffee 171 or the banana 191 xÂ² = 16.69, d.f. = 2, p < .05. There also seemed to be a general tr end for hummingbird visits to increase along the edge of the forest. In the field transect, feeder two 21 visits was significantly different from the edge feeder, four 58 visits, P = .0004. Similarly, in the banana transect, feeders one zero visits a nd two seven visits were significantly different from feeder four 31 visits, p < .0001. Finally, in the coffee transect, both feeders three 39 visits and four 37 visits were significantly different from feeder two 18 visits, p=.0257. In all three transects, the edge was significantly different from at least one other feeder. Further, the total number of visits 126 to the three edge feeders was significantly greater than the total visits to any other group of feeders xÂ² = 49.06, d. f. = 5, p < .05, and within each transect was either the most visited feeder or nearly so Table 1. Table 1: Total visits per Feeder. Each Â€feederÂ is the total number of visits in all feeders of the same number across the three transects. Note th e edge, feeder four, attracted more hummingbirds than any other feeder group. Pasture Banana Coffee Total Feeder 1 36 0 26 62 Feeder 2 21 7 18 46 Feeder 3 29 21 39 89 Feeder 4 58 31 37 126 Feeder 5 20 20 28 68 Feeder 6 27 41 23 91 Total 191 120 171 482
By Species The Species richness did not vary much from site to site, and showed no trends. The Shannon Weiner diversity index for individual sites showed no patterns, but when the equivalent stations were combined, there was a slight trend towards greater diversity in the forest Table 2. The Violet Sabrewing showed an overall preference for the banana plantation, with 71 visits compared to 38 and 43 for the field and coffee, respectively xÂ² = 83.4, d.f. = 4, p < .05. It also avoided open areas xÂ² = 121.8, d.f. = 4, p < .05, and was notably absent from the field and coffee plantations, with no visits to stations one or two in these transects see Appendix 1. Conversely, the Rufous tailed hummingbird avoided the banana plantation xÂ² = 83.4, d. f. = 4, p < .05, with only 16 total visits, compared to 79 in the field and 90 in the coffee. It also preferred open areas, with only 15 total forest visits compared to 170 non forest visits xÂ² = 121.8, d.f. = 4, p < .05, see Table 3. The stripe taile d hummingbird seems to show a trend toward higher number of visits in the forest and edge, averaging 27 visits per feeder in forest, 29 visits per feeder on the edge, and 11 visits per feeder in the agricultural lands, but observed abundance did differ gre atly from the expected Chi square abundance. Species composition was quite predictable and similar in forested sites five and six, xÂ² = 4.68, d.f. = 4, p < .05 and xÂ³ = 9.37, d.f. = 4, p < .05, respectively with totals for both feeders of 15 Rufous tail ed, 82 Violet Sabrewings, and 54 Striped tailed hummingbirds. However, the birds were affected by the different agricultural types, and the compositions for edge four and agricultural land three through one was significantly different than expected xÂ² = 25.16, d.f . = 4, p < .05; xÂ² = 24.9, d.f. = 4, p < .05, x Â² = 36.5, d.f. = 4, p < .05, and xÂ² = 10.8, d.f. = 1, p < .05, respectively. The edge feeders recorded 42 Rufous tailed, 51 Violet Sabrewings, and 29 striped tailed hummingbirds, while the combin ed field stations recorded 128 Rufous tailed, 19 Violet Sabrewings, and 33 Striped tailed hummingbirds See Appendix. Table 2: Shannon Weiner Diversity Index and Species Richness Among Feeder Groups. The species richness showed no patterns over the six feeders. The diversity index was higher in the forest fragment than in the agricultural land, but the values were very similar throughout. Species Richness Shannon Weiner HÂ‚ Feeder 1 4 0.9011209 Feeder 2 5 1.03155433 Feeder 3 6 1.11097719 Feeder 4 4 1.11172248 Feeder 5 5 1.05340915
Table 3: Visits per feeder for the Rufous tailed hummingbird and the Violet Sabrewing. Note the low number of visits of the Rufous tailed from forested sites five and six. Also note the absence of Violet Sabrewing from open sites one, two and three. When the banana data is removed, because the environment is more understory like than open, the trend becomes more pronounced. The Stripe tailed hummingbird seems to prefer forested areas, but also frequents open area s. Striped Rufous Violet Violet without Banana Feeder 1 14 41 0 0 Feeder 2 6 32 5 0 Feeder 3 13 55 14 5 Feeder 4 29 42 51 29 Feeder 5 21 3 40 29 Feeder 6 33 12 42 18 Discussion The species richness did not show any recognizable trend, and the Shannon Wiener diversity index, while showing a slight trend towards higher diversity in the forest, was very similar in all six sites. This is due to the low number of species in the system. Because so few species were present, if a rare species happened to visit during an observation period, it altered the diversity index at that site in a way that may not reflect the true in site diversity. Because exhaustive sampling of the individuals present was impossible, and the community was dynamic, this problem was unavoidable. Species composition changed between forested and open areas, although the Stripe tailed hummingbird remained relatively constant. In the forested areas, stations five and six, Rufous tailed hummingbirds were largely absent, while Violet S abrewings abounded. The opposite was true for open areas, with Rufous tailed being more abundant while the Sabrewings were reluctant to leave the forest. This observed segregation could be a response to competition. Both bird species use such plants as Hel iconia, Cephaelis, Stiles and Skutch, 1989 and Erythrina Neill 1987. The hummingbird community is strongly affected by competition. Proctor et al . 1996, and disturbance increases the already high level of competition because the birds increase their niche breadth, and thus increase niche overlap between species Rapole and Morton, 1985. The Violet Sabrewing, however, has been described as a traplining species; a species which visits small groups of widely separated flowers over foraging circuits Nei ll 1987. The Rufous tailed hummingbird is quite the opposite, preferring to guard a small territory of nectar rich flowers Stiles and Skutch, 1989. That these two species with such different behaviors would be in competition with each other seems stran ge. However, often the Violet Sabrewing was observed territorially guarding feeders, suggesting that either in the fragment, of because of the attractiveness of the feeders, it changes its behavior. Hummingbirds are known to subtly change their behavior in fragmented forest by changing flower preference or foraging strategies Neill,
1987. The two birds, while sharing a common food source, may avoid competition through differing behaviors in intact forest. However, as behaviors change in response to fragme ntation, the birds may be driven into competition as the Violet Sabrewing adopts the Rufous tailed hummingbirdÂ‚s behavior, forcing the birds to separate their ranges to avoid competition. Violet Sabrewing lives and reproduces in forest Stiles and Skutch 1 989, so as the forests continue to disappear, it will become less common. The Rufous tailed hummingbird, because of its ability to thrive in cleared areas, will replace the Violet Sabrewing in cleared areas. Because the Violet Sabrewing is reluctant to v enture into open areas, the plants that it pollinates will have their pollen flow restricted to within the fragment. The Rufous tailed hummingbird, while it may travel between fragments, does not enter them, and thus will not contribute to between patch po llen flow. Thus, the job of transporting pollen between patches will fall to Stripe tailed hummingbird, the only bird that will cross open areas and enter the forests. As fragmentation becomes evermore prevalent, the Stripe tailed will become more importan t as a vector of pollen movement between patches. The number of visits was greater at the edge, for a number of reasons Figure 2. The edge was frequented by both the forest dwelling species Violet Sabrewing and the field dwelling species Rufous taile d hummingbird. Further, the edge offered a place to perch; hummingbirds were often observed perching at the edge, flying to a feeder, and then flying back to their edge perch. Plant productivity is higher than the forest at the edge, due to the higher inp ut of light Laurence et al. 1997. Nectar richness is thus greater in edge habitat because plants have more energy to allot to nectar production Bierregaard and Stouffer, 1997, and therefore hummingbirds are often found there. Further, there were more h ummingbird plants flowering on the edge than in agricultural land personal observation. Hummingbirds may have occurred with greater frequency on the edge because the feeding conditions there tend to be more favorable. Beyond the edge, numbers of visits were similar for the coffee plantations and the field, which is not surprising considering they are both relatively open areas. Although coffee bushes offer a place to perch and some limited shelter, the microclimate and non coffee vegetation is more simil ar to the field than either the banana plantation or the forest. Feeding patterns were similar in both habitats as the birds flew from the edge to the feeder and then back to the edge. The difference in visitation can be accounted for by the lack of Stripe tailed hummingbirds in the coffee. There is a lack of food among the coffee bushes, as coffee is wind and self pollinated Boucher 1983 and other species are removed by the farmer, resulting in a lack of hummingbird plants. The Rufous tailed hummingbirds are known to occupy coffee plantations Stiles and Skutch, 1989, while the Stripe tailed preferred the fieldÂ‚s Asclepia and other small flowering plants personal observation. However, overall, the two ecotones were quite similar. Visitation to the ban ana ecotone was very different. The banana trees provided an understory environment, and therefore the species composition reflected that of the forest more than the other two sites. The Violet Sabrewing, a known inhabitant of banana plantations Stiles an d Skutch, 1989 was the common species. However, the most striking difference between the banana plantation and the other ecotones was the lack of
visits to feeders one and paucity of visits to feeder two. Although Violet Sabrewings are known to forage in banana plantations, and were once observed near the feeders foraging at the bananas, they did not visit the feeders. The hummingbirds may have been altogether absent from the plantation due to the lack of everything but bananas to forage on, but because th ey occur in such close proximity and were seen in the plantation, this seems unlikely. An alternate explanation is that the banana flowers offered a richer nectar reward than the feeders, so the hummingbirds ignored the feeders. Different species of hummi ngbirds showed a preference for different edges. In the banana plantation, Violet Sabrewings and Stripe tailed hummingbirds have to pollinate the plants, because the Rufous tailed is absent. Although the Rufous tailed is more abundant at the field edge tha n the coffee edge, all three species of hummingbird occur at both edges. Therefore, plants growing on the banana edge have a different pattern of pollen flow. Because the Rufous tailed hummingbird, the dominant bird on the other edges, ignores the banana e dge, plants growing there may not be able to spread their pollen as effectively as plants growing on the other edges. However, this could reflect the fact that the banana plantation is an understory environment and the edge is less prevalent, and therefore the pollination patterns more closely reflect the patterns in the fragment. Experiments with the hummingbird pollinated tree Symphonia globulifera Gittiferae have shown that trees in pastures, because they are exposed to more sunlight and less competit ion, produce a higher flower set, and that they are often self pollinated Aldrich and Hamrick 1998. This led to a genetic bottlenecking of the population, a few Â€superadultsÂ producing the majority of recruits, and the high selfing rate increased this ef fect. The authors implicated a shift in pollinator behavior from forest traplining to pasture territoriality. A territorial bird will sit at one tree, defend it as a territory, and continually visit flowers on the same tree, self pollinating the tree. A tr aplining bird will cross pollinate the tree by visiting many trees. Aldrich and Hamrick 1998 A change in species composition from Violet Sabrewings in the patch to Rufous tailed hummingbirds in the field would have the same effect. The territorial behavi or of the Rufous tailed hummingbird would result in a higher rate of selfing in pasture trees. Further, if the Violet Sabrewing is also exhibiting territorial behavior, the self pollination trend would continue in the patch. The resulting skewed genetic di versity would greatly affect the quality of the patch as a source pool for re forestation efforts. Further research into the fragment behavior of Violet Sabrewings could reveal much valuable information about the pollination patterns of forest fragments, as could genetic studies on the plants themselves. As human beings continue to convert the worldÂ‚s forests into small isolated patches, it is becoming increasingly important to understand the changes taking place in these patches. Vital ecosystem interacti ons such as pollination must be studied in order to understand the eventual fate of the tree species which have been isolated in patches. This study demonstrated that subtle details, such as the type of crop grown around the patch, can affect species compo sition, abundance and even behavior of hummingbirds. This could affect the quality of the patch as a source bank for genetic diversity, and as a refuge for plants and animals following deforestation.
Only through a detailed knowledge of what we are doing t o biodiversity can we hope to one day restore it. Figure 2: Number of Visits at each Feeder. Feeder number one and two represent agricultural land, numbers five and six forest, and number four edge. Note the higher rate of visitation at the edge. Acknowledgements I would like to thank Alan Masters for his help throughout this project, from conception to completion. I wo uld also like to thank DoÃ±a Ingr acia and Victor Torres for allowing me to study on their farms. I would also like to thank Eric and Fabian Villalobos for helping to reduce the monotony of watching hummingbird feeders. Thanks also to Ben, for his thoughtful critique and f inally, thanks to all the hummingbirds for coming. Literature Cited Aldrich, P. R. and J. L. Hamrick. 1998. Reprod uctive Dominance of Pasture trees in a Fragmented Tropical Forest Mosaic. Science 281: 103 105 Bawa, K. S. 1990. Plant Pollinator Interactions in Tropical Rain Forests. Ann. Rev. Ecol. Syst. 21: 399 422. Bierregaard, R. O. and P. C. Stouffer. 1997. U nderstory Birds and Dynamic Habitat Mosiacs in Amazonian Rainforests. In W. F. Laurance and R. O. Bierregaard, ed. Tropical Forest Remnants. University of Chicago Press, Chicago. Boucher, D. H. 1983. Coffee. In: Costa Rican natural History, D. H. Ja nsen, ed. The University of Chicago Press, Chicago, IL. Pp. 86 88 Bruna, E. M. 1999. Seed Germination in Rainforest Fragments. Nature 402: 139. Chazdon, R. 1998. Tropical Forests: Log Âƒfm or Leave Âƒem. Science 281: 1295.
Debinski, D. M. and R. D. Holt. 2000. A Survey and Overview of Habitat Fragmentation Experiments. Conservation Biology 14: 342 355. Didham, R. K. and J. H. Lawton. 1999. Edge Structure Determines the Magnituse of Changes in Microclimate and Vegetation Structure in Tropical Fore st Fragments. Biotropica 311: 17 30. Fiensinger , P., W. H. Bubsy, K. G. urray, J. H. Beach, W. Z. Pounds, and Y. B. Linhart. 1988. Mixed Support for Spatial Heterogeneity in Species Interactions: Hummingbirds in a Tropical Disturbance Mosiac. The American Naturalist. 131: 33 57. Fogden, M. 1993. An Annotated Checklist to the Birds of Monteverde and PeÃ±as Blancas. Michael Fogden, Monteverde, Costa Rica. Laurance, W. F., S. F. Laurance, L. V. Ferreira, J. M. R. Merona, C. Gascon, T. E. Lovejoy. 19 97. Biomass Collapse in Amazonian forest Fragments. Science 78: 1117 1118. Neill, D. 1987. Trapliners in Trees: Hummingbird Pollination of Erythrina Sect. Erythrina. Ann. Missouri Bot. Gard. 74:27 41 Proctor, M., P. Yeo and A. Lack. 1996. The Natural History of Pollination. Timber Press, Portland, Oregon. Rappole, J. H. and E. S. Morton. 1985. Effects of Habitat Alteration on a Tropical Avain Forest Community. Pages 1013 1019 in M. Foster, ed. Prnithological Monographs No. 36. The American Orn ithologistsÂ‚ Union, Washington, D.C. Smith, T. B., R. K. Wayne, D. J. Girman and M.W. Bruford. 1997. A Role for Ecotones in Generating Rainforest Biodiversity. Science 276: 1855 1857. Stiles, F. G. and A. F. Skutck. 1989. A Guide to the Birds of Costa Rica. Cornell University Press, Ithaca, New York. Terborgh, J. 1992. Diversity and the Tro pical Rainforest, Scientific Am erican Library, New York. Vitousek, P. M., H. A. Mooney, J. Lubchenco, J. M. Melillo. 1997. Human Domination of EarthÂ‚s Ecosyste ms. Science 278: 494 499.
Appendix 1 Rufous Violet Striped Ruby Steely Fork Coppery Field 1 18 0 13 3 0 0 0 Field 2 14 0 6 2 0 0 0 Field 3 16 5 7 3 0 0 0 Field 4 27 15 15 0 0 0 0 Field 5 2 11 8 0 0 0 0 Field 6 2 7 15 0 2 0 0 Banana 1 0 0 0 0 0 0 0 Banana 2 2 5 0 0 0 0 0 Banana 3 7 9 3 1 1 0 0 Banana 4 0 22 8 0 0 1 0 Banana 5 0 11 8 0 0 1 0 Banana 6 7 24 10 0 0 0 0 Coffee 1 23 0 1 0 0 0 3 Coffee 2 16 0 0 0 0 0 2 Coffee 3 32 0 3 0 0 0 1 Coffee 4 15 14 6 0 0 0 0 Coffee 5 1 18 5 0 0 1 3 Coffee 6 3 11 8 0 0 2 0 Totals Field 79 38 64 8 2 0 0 Banana 16 71 29 1 1 2 0 Coffee 90 43 23 0 0 3 9 Grand Total 185 152 116 9 3 5 9 Totals per Position Feeder 1 41 0 14 3 0 0 3 Feeder 2 32 5 6 2 0 0 2 Feeder 3 55 14 13 4 1 0 1 Feeder 4 42 51 29 0 0 1 0 Feeder 5 3 40 21 0 0 2 3 Feeder 6 12 42 33 0 2 2 0 Visits to each feeder, as well as totals per feeder, per ecotype, and the overall totals, sorted by species. Note the absence of Violet Sabrewings in the open areas feeders one through 3 and the absence of Rufous tailed hummingbirds in the forested area s five and six and all banana stations. Abbreviations are as follows: Rufous, Rufous tailed hummingbird, Amazalia tzacatl , Violet, Violet Sabrewing, Campylopterus hemileucurus ; Striped, Striped tailed hummingbird, Eupherusa eximia , Ruby, Ruby throated h ummingbird, Archilochus colubris ; Steely, Steely vented hummingbird, Amazalia saucerrottei ; Fork, Fork tailed emerald, Chlorostilbon canivetii , Coppery, coppery headed emerald, Elvira cuprei .