THE FEEDING PREFERENCES FOR COLOR AND TIME OF DAY OF RUBY-THROATED HUMMINGBIRDS IN GARRETT COUNTY, MARYLAND

THE FEEDING PREFERENCES FOR COLOR AND TIME OF DAY OF RUBY-THROATED HUMMINGBIRDS IN GARRETT COUNTY, MARYLAND July 27, 2012

ABSTRACT Ruby-Throated Hummingbirds (Archilochus colubris, or RTHU) are nectarivorous, feed on ornithophilous plants, and are able to digest sucrose nectar. Honey bees are also able to digest sucrose nectar, meaning hummingbirds could possibly pollinate the same plants that honey bees do. This similarity is important because since the mid-20 th century, there has been a sharp decline in the honey bee population. If the honey bee population continues to fall, agriculture could suffer because honey bees are major pollinators. We need to find a substitute for the honey bees until their population can recover. Hence, we are trying to discover the feeding preferences for color and time of day of RTHU. We hypothesized (H 0 ) that there will be no significant difference in average number of times the RTHU feed from one specific color or in the average number of feedings in a given time interval over the length of the experiment. We used two hummingbird feeders filled with solution in a 4:1 water to sugar ratio to see what times hummingbirds fed and from which color they fed. We recorded data of the hummingbirds feeding for four days from 6:45 a.m. to 8:45 p.m. We used a t-test assuming unequal variance to determine if there was a significant difference in the data. We rejected our null hypotheses (H 0 ) for color because our t-tests indicated a significant difference between the average number of times RTHU fed from blue and green in favor of red and yellow. We rejected our null hypothesis (H 0 ) for time because our t-tests indicated a significant difference in the average number of feedings between morning and afternoon and morning and evening in favor of the evening. Our results suggest that if we want to attract hummingbirds to pollinate the plants that honey bees currently do, we need to make sure that nectar from a red or yellow source is available; more studies are necessary to determine the feeding time preference of RTHU.

INTRODUCTION Pollination is the process by which flowers are fertilized. The male reproductive cells, also known as pollen, are transmitted to the female reproductive parts through the use of pollinators. This process is crucial to the continuance of flower and plant species around the world. The main pollinators of the world vary from small mammals, like mice and bats, to insects and birds, like honey bees and hummingbirds. Pollinators are important parts of our ecosystem because without them, the crops that we grow would not be pollinated (Poslethwalt & Hopson, 2006). For example, 1/3 of all of the crops in the United States are pollinated solely by the Western honey bee. Without important pollinators like honey bees, the crops that rely on honey bee pollination would not fertilize effectively, therefore causing a failed crop and possibly leading to food shortages and higher agricultural prices (Berenbaum, 2007). Since late 1947, researchers have discovered a severe decline (40%) in honey bee populations that is affecting the delicate relationship between the bees and the flowers and plants they pollinate. There have been many studies conducted on the reason for the decline. For example, National Agricultural Statistics Service (NASS) has conducted numerous assessments on the decline but the cause of this condition is still unknown. For lack of a cause and name, researchers have officially called this condition Colony Collapse Disorder (CCD). CCD not only directly affects the honey bee population, but also agriculture because the more the bee population declines, the less pollination occurs, and thus the less healthy the crops will be when it comes time to harvest them. Less honey bees means less pollination; this decline can lead to a demand for pollination services, an increase in the demand for crops from other countries, and an increase of cost of crops (Berenbaum, 2007). Hummingbirds are a possible solution to these problems because honey bees and hummingbirds are closely related. 1

Honey bees and hummingbirds are two unique kinds of pollinators. The hummingbirds have a very similar diet to honey bees because both are nectarivorous (Why Do Hummingbirds Hum?, 2000; Nicholson & Fleming, 2003; University of Maine, 2002; U.S. Dept. of Agriculture Wildlife Management). Both pollinators have the ability to digest a type of complex sugar called sucrose, which is found in the nectar (the sugar water solution that is produced by plants to attract pollinators) of plants. Honey bees tend to frequent short tubed flowers because they can easily land on them and obtain nectar (Nicolson, & Fleming, 2003); however, it has been found that hummingbirds can also feed off of the same plants as honey bees. This is significant because if hummingbirds can feed off of the same plants, they will in the process pollinate the same plants as honey bees. The hummingbirds are thus a possible substitute for the lack of honey bees. However, while hummingbirds can pollinate honey bee flowers, they do not typically pollinate those flowers because they prefer a different type of plant. Hummingbirds pollinate a special plant type, ornithophilous plants, which are plants that are pollinated by birds (Nicolson & Fleming, 2003). Ornithophilous flowers are diurnal, which means that they bloom primarily during the day; scientists believe that this characteristic is one of the many adaptations to hummingbirds (Nicolson & Fleming, 2003). According to Nicolson & Fleming (2003), most ornithophilous flowers produce more nectar in the morning right before the hummingbirds come out of torpor, which is a hibernation-like state that hummingbirds go into when energy levels are low (Poslethwalt & Hopson, 2006). Other adaptations include some of their shapes and their dilute nectars. These flowers tend to be bigger and have a long tubular structure, which helps the hummingbird feed more easily because of its long slender beak. Another adaptation is the nectar that the flowers produce. Ornithophilous flowers generally have a 20-25% dilute nectar, which is the preferred mixture of sucrose and water that is needed for 2

hummingbirds so that they have enough energy to support their fast flying speed and high metabolism (Miller & Miller, 1971). All of these adaptations could be why hummingbirds seem to prefer certain characteristics of ornithophilous flowers. Throughout the years, there have been many studies done to find out if hummingbirds have any color preference for the flower they feed on (Miller & Miller, 1971; Hauser, 2008; Grant & Grant, 1968; Why Do Hummingbirds Hum?, 2000). An experimental study in California found that hummingbirds seemed to prefer red colored flowers, but in California the majority of the ornithophilous flowers are, in fact, red. In a different experimental study, the results seemed to show that the position of the flower mattered more than the color (Miller & Miller, 1971). Since these studies contradict each other, more studies are needed to determine if hummingbirds have a color preference. There have also been studies to see if hummingbirds prefer a specific feeding time (Why Do Hummingbirds Hum?, 2000). In one experimental study, the earliest recorded visit in Canada was 30 minutes after sunrise, and the latest recorded visit was an hour and twenty minutes after sunset, but they concluded that the hummingbirds fed equally throughout the day (Miller & Miller, 1971). As a result, more studies are needed because there is not enough evidence to determine if hummingbirds have any type of feeding time preference. Therefore, the focus of our study is to find out whether or not Ruby-Throated Hummingbirds (Archilochus colubris) have a color preference of the flowers they feed on and if they have a feeding time preference by setting out feeders all day. The first purpose of our study is to see if the Ruby-Throated Hummingbirds (RTHU) have a feeding preference for a particular color. If we can find out what color the RTHU prefer, then we can use this information to try to attract more birds; in doing so, we could enhance 3

pollination on crops that need it because of the lack of other pollinators, more specifically, the Western honey bee. The second purpose of our study is to determine if the RTHU have a specific time period during the day they feed, thus pollinate, the most. The honey bees pollinate equally throughout the day, but we do not know if the hummingbirds would do the same. We want to compare the data we collect on the RTHU with the information we already know about the honey bees so that we can determine if the hummingbirds would be an effective substitute for pollination needs. We used this specific species of hummingbird, RTHU, because they are the only nesting hummingbird on the east coast, in the general region where our experiment was done. As shown in Fig.1, the RTHU s typical nesting grounds are on the east coast of the United States and along the outer edge of Canada. We chose our study site, Garrett County, Maryland, because there were frequent sightings of RTHU the previous summer, meaning that we would have a reliable population to sample this summer because the birds would most likely visit the same feeders they knew were there. We chose the time of year, around mid-june, because that is when the birds were going to be in our general area according to their migration pattern. The birds migrate down to Central and South America during the winter time (around September) and come back to the breeding grounds during the spring (around late April). June was also an ideal month to conduct our study because mating season ends around late April, which means that there would be a higher concentration of juvenile RTHU in the area, leading to a larger sample size. We chose our time of day to test during based off of what we know about the RTHU and the plants that they feed on. The ornithophilous plants that they feed on are diurnal, which means that we would not have to worry about collecting data during the night hours because the flowers are closed and the birds are in a state of torpor. We also based our selection off of previous studies 4

that showed that the hummingbirds feed evenly throughout the day when provided with feeders (Miller & Miller, 1971). Fig. 1. - Average spring migration route of the Ruby Throated Hummingbird. In addition, artificial feeders were chosen instead of observing the hummingbirds in their natural environment with flowers because it was easier to manipulate the colors on the feeder than it would be to manipulate the colors of the flowers. Another reason for choosing feeders was that the nectar would not be available in the flowers all day like it is in the feeders because of natural factors such as evaporation from the sun. If the nectar is available all day for the birds, we want to see if the RTHU would continue feeding equally throughout the day in order to determine how they would respond to the crops that need to be pollinated all throughout the day. We chose the color of the disks (red, yellow, blue, and green) that we placed on the feeders based on different reasons. For example, we chose the red color because previous studies showed that the hummingbirds actually did prefer the red color over the other colors that they used 5

(Grant & Grant, 1968). We chose the yellow because it is a common ornithophilous plant color and blue because the hummingbirds are less sensitive to color on that end of the spectrum (Grant & Grant, 1968; Why Do Hummingbirds Hum?, 2000). We chose green because it is not a common flower color at our study site, so it was used as a control to make sure that the birds had a color preference rather than just a position preference. There were two sets of hypotheses developed for this experimental study. The first null hypothesis was given a feeder set out all day for four days, there will be no significant difference in the average number of times the RTHU feed from one specific color. The alternative hypothesis was given a feeder set out all day for four days, there will be a significant difference in the average number of times the RTHU feed from one specific color. The second null hypothesis was given a feeder set out all day for four days, there will be no significant difference in the average number of feedings in a given time interval over the length of the experiment. The alternative hypothesis was given a feeder set out all day for four days, there will be a significant difference in the average number of feedings in a given time interval over the length of the experiment. 6

METHODS The study site was located at 8167 Westernport Rd., Garrett County, Maryland (Fig.2). Our experiment was conducted on June 12, 14, 15, and 16, 2012. Hummingbird feeders were placed in the yard of the residence. This yard was surrounded by dense forest on all sides, which gives the birds a place where they can perch between feedings. The site we used was ideal because it is in an area where RTHU have been seen in previous years and in late spring. The site has many ornithopilous flowers; one of these flowers is the columbine. The columbine was in bloom during the time the hummingbirds first arrived, but it was not in bloom during the time of the experiment. Some of the other flowers that were in bloom during the time of the experiment were evening primrose, black-eyed susans, and daisies. Fig. 2. - Our study site, with a red dot showing where the feeders were located. To set up for this experiment we used two feeders. These feeders were built out of 20 oz. water bottles that were attached to standard deli containers. At the top of the deli containers, we created four holes from which the birds could drink. Around these holes were different color disks which simulated the color of flowers (Fig. 3). To create the sucrose solution, we used a 4:1 7

water to granulated sugar ratio. We boiled the sugar solution, which helped get rid of any germs that could be in the water, it also helped dissolve the sugar completely. We then let the solution cool down before putting it into the feeders. Fig. 3. - One of the feeders we used. After filling the feeders with our sucrose solution, we hung them outside. The feeders were placed approximately a foot away from the house on a pole and were hung about 44 inches off the ground. In order to watch the feeders without disturbing the birds, we placed a camera inside the house at a window facing the feeders. The camera recorded video from about 6:30a.m.to about 9:00 p.m. The feeders were kept full at all times and they were rotated 180 degrees once after two days. The video for each day was divided into two hour increments and downloaded onto CDs. Since the videos were not started at the same time each day, we decided to begin our data collection at 6:45 a.m. We decided to look at the videos in 15 minute intervals because according to University of Maine (2002), the hummingbirds have to eat every ten minutes to sustain their extreme energy levels. Since we started at 6:45a.m., 15 minutes was a better increment to follow 8

than 10 minutes. For each individual 15 minute interval, we counted every time the hummingbirds stuck their beak into a disk as one feeding. We then combined all of the feedings for each color to create our raw color data. We also combined all feedings for each time increment to create our raw time data. In order to analyze our raw color and time data, we calculated the averages of each set and constructed bar graphs to display the results. The bar graphs have 95% confidence intervals, which were used to show a range of sample means that would represent the population. We used the confidence intervals to predict if a difference was present by seeing if the confidence intervals did not overlap. Finally, we conducted a statistical analysis, using a t-test, to determine if there was a statistically significant difference between the individual color data. We then did the same procedure for the time data. In order to know which t-test to use, we conducted an F-test on the raw data from both the color and time. The F-test helped us determine which t-test to use, either to use a t-test for data with unequal variances, or equal variances. After running the t-test we looked at the p value to see if there was a statistically significant difference. If the p-value was less than 0.05, then there was a statistically significant difference. 9

Average Number of Feedings RESULTS As Fig.4 shows, the blue color disk had the lowest average number of feedings out of all the colors. The confidence intervals suggest that there is a difference between the averages for blue and all other colors, as well as between the averages for green and red, and green and yellow. A t-test using the p-values seen in Table 1 later confirmed that there was a significant difference. The confidence intervals suggest that there is no difference between red and yellow, and a statistical t-test was used to confirm this suggestion. 4000 3500 3000 2500 2000 1500 1000 500 Average Number of Feedings Per Color 0-500 Blue Green Red Yellow Disk Color Fig. 4. - The average number of feeding per color. We divided our time data up into four periods: morning (6:45a.m.-10:15a.m.), midday (10:15a.m.-1:45p.m.), afternoon (1:45p.m.-5:15p.m.), and evening (5:15p.m.-8:45p.m.). As Fig.5 shows, the morning time period had the lowest average number of feedings out of all our time periods. Even though the confidence intervals overlap, t-tests using the p-values seen in Table 1 indicated that there was a significant difference between the morning and afternoon intervals and the morning and evening intervals. 10

Average Number of Feedings 2500 Average Number of Feedings Per Time Period 2000 1500 1000 500 0 6:45AM-10:15AM (Morning) 10:15AM-1:45PM (Midday) Time Period 1:45PM-5:15PM (Afternoon) 5:15PM-8:45PM (Evening) Fig. 5. - Average number of feedings per time periods. T-test Results (p-value) Color Time Blue/Green 0.00 Morning/Mid-Day 0.08 Blue/Red 0.01 Morning/Afternoon 0.01 Blue/Yellow 0.00 Morning/Evening 0.01 Green/Red 0.02 Mid-Day/Afternoon 0.32 Green/Yellow 0.00 Mid-Day/Evening 0.29 Red/Yellow 0.17 Afternoon/Evening 0.85 Table 1. - The p-values from our t-tests. 11

CONCLUSIONS & DISCUSSION The null hypothesis for the color variable was given a feeder set out all day for four days, there will be no significant difference in the average number of times the RTHU feed from one specific color. This hypothesis was used to determine if the RTHU actually had a significant preference for a certain color disk on the feeders and in doing so, determine what colors are best suited to attract the RTHU. We reject the null hypothesis for the color variable in favor of the alternative because, based on a t-test conducted, there was a significant difference in the color preference of the birds; overall the RTHU seemed to prefer the red and yellow colors more than the blue and green. The null hypothesis for our time variable was given a feeder set out all day for four days, there will be no significant difference in the average number of feedings in a given time interval over the length of the experiment. This hypothesis was used to determine if the RTHU had a preference in feeding time and in doing so, determine if the RTHU would be a suitable substitute for the honey bee pollinators, who pollinate all throughout the day. We reject our null hypothesis for the time variable in favor of the alternative because, based on the t-test, there was a significant difference between the time intervals; overall the RTHU seemed to feed less in the morning and more in the afternoon and evening. For color, there were many factors we looked at when analyzing our results. We concluded that the number of feedings from the blue disks seems to be the lowest, possibly because the RTHU may be less sensitive to the blue and violet end of the color spectrum (Grant and Grant, 1968). Grant (1968) also found that hummingbirds are more sensitive to red than humans, and equally sensitive to yellow, which may explain why red and yellow were picked more often and are not statistically significantly different from each other. There were a number 12

of possible reasons that the green disk was chosen more than the blue disk, even though the green was supposed to be our control. One is that the two most common places that the RTHU migrates to are Mexico and Nicaragua, which both have native green ornithophilous flowers such as Mistletoe and Hellebores. The presence of the green flowers along the RTHU migration route may indicate that the birds have learned that green flowers do produce nectar and therefore, that the green disk would have a nectar reward. Another reason for the frequency of visits to the green disk may be that according to a color analysis we performed, the color green that we used for the disk had a high percentage of yellow in it, and yellow is a common ornithophilous flower color. Finally, we are confident that the data we collected does indeed reflect color preference of the RTHU and not position preference because we rotated both feeders, each 180 degrees, so that if the birds followed the color, which they did, we would know it was a color preference. For our time variable, we decided to divide our initial average data up into four specific time periods: morning, midday, afternoon, and evening. We came to this conclusion by observing a time when the hummingbird feedings were low in the average data. We then referred to a previous nectar production study that stated that ornithophilous plants produce the most nectar in the morning right after sunrise before the hummingbirds become active, and that the sun evaporates the nectar as the day goes on (Hauser, 2008). This study raised questions about the times the birds fed. We wanted to see if the birds really did have a time preference for feeding so as to conclude if they were a possible substitute for the honey bee population that feed and pollinate equally throughout the day. However, we found that the birds did not feed more in the morning as we had previously thought, which led us to look at how the birds were feeding rather than how many times they actually visited the feeder. A review of the video footage showed that the birds in the morning 13

actually perched and drank for longer periods of time than the birds in the afternoon. This observation means that because of the way we collected data, (i.e., recording each time a bird stuck its beak in a disk), there were less feedings but the birds actually fed for longer periods of time. The fact that the birds consumed more nectar could be because they had just woken from their state of torpor so they needed more energy right away, or because they had learned that since there was more nectar in the morning (in flowers in nature), they should drink the most during those times. The larger consumption of nectar might also explain why the feedings in the afternoon and the evening seemed more frequent than the morning, possibly because the birds were not drinking as much nectar at each feeding, and therefore had to come back more times to sustain their high energy levels. Our study had a variety of limitations. One of our limitations was the weather at the site during our four testing days. According to Dr. Latta (2012), there was slight precipitation on the first day of the study, which could have impacted how many birds came out on the 12 th to feed from the feeders. Another limitation was all the activity that went on around the feeder that could have impacted the hummingbirds feeding activity. The feeders were placed right outside of a residence, which meant that there could have been various disturbances at the site, including people, vehicles, sounds from the residence, and other animals. There were also several wasp sightings on the feeders that may have scared some of the hummingbirds away from the feeder, or impacted what color they chose (for example, if the wasp was on the red color, the bird may have deviated from that color and went to another color instead). There is furthermore the issue of territoriality. Some of the hummingbirds became very aggressive over the feeders and chased other birds away, possibly affecting how many birds had the opportunity to feed at the feeders. The feeder placement could have also affected how many birds saw the feeders; for instance, if 14

they were flying over the opposite side of the house, they might not have seen the feeders on the other side. There were also technical limitations to our study. The camera that was used for data recording had some technical problems that lead to missing time in our data. The times that were missing ranged from around 7:00 p.m. - 8:00p.m. each day with about 5-20 minutes of data missing overall. Finally, there were observational errors while we collected the data. These errors varied from having too many birds feed at once so that we could not count all feedings to not being able to see the birds feed because the water bottles were cloudy, therefore obscuring two of the color disks. Suggestions for future studies include using better camera angles and more study sites, using additional ornithophilous flowers, choosing different disk colors, substituting flowers for feeders, and recording nearer to sunrise. In our study, we only used one video camera and therefore could only see one side of the feeder, which made it harder to collect data on birds that went to the side of the feeder that was obscured; so, additional camera angles would have aided in data collection. More study sites are a suggestion because if more study sites were included, there would be a larger sample to analyze, thus making the results more accurate. Likewise, if there were more ornithophilous flowers throughout a study site, more hummingbirds could be attracted to the feeders, resulting in a larger sample size and consequently, more accurate data. Variation in disk colors is also a suggestion because if black or clear were used as controls, the data would be more accurate because there are no true black or clear flowers in nature. Using flowers instead of feeders was suggested because the study would be more relevant as to what actually occurs in the wild. Finally, recording nearer to sunrise could allow collection of data to begin as soon as RTHU begin to feed. 15

During our study, a few questions were raised that require further research. The first question was: how did the RTHU know to go to the feeder to feed? Throughout our study site, specifically around the feeder, there were ornithophilous plants, but our video recordings only show the RTHU visiting our feeders. If this question was answered, we may be able to train RTHU to pollinate wherever nectar is available. Our second question was: do RTHU return to the same site where they were born to nest? In previous years, it seemed that about the same amount of RTHU returned every year to our study site, but we cannot be certain due to the fact that the RTHU were not tagged. If this question was answered we could know how many hummingbirds to expect at any given area, and thus we could be able to predict how consistent pollination will be. The last question we had was: why do RTHU come to the feeders in clusters? When watching the video recordings, there were moments when RTHU would come together and feed at the feeders, then leave around the same time, so it led the group to wonder whether or not that was the way RTHU preferred to feed. Overall, our results suggest that if we want to attract hummingbirds to pollinate the plants that honey bees currently do, we need to make sure that nectar from a red or yellow source is available; more studies are necessary to determine the feeding time preference of RTHU. If these questions are researched further, it could possibly help others to better understand the feeding preferences for color and time of day for RTHU, which can help local farmers determine whether or not the RTHU are a possible substitute for the lack of honey bees and pollination. 16

REFERENCES CITED Berenbaum, M.R. (2007, March). Colony Collapse Disorder and Pollinator Debate. Retrieved from http://www7.nationalacademies.org/ocga/testimony/colony_collapse_disorder_and_poll inator_decline.asp. Grant, K.A & Grant, V. (1968). Hummingbirds & Their Flowers. New York, NY: Columbia University Press. Hauser, M.A. (2008). Nectar Accumulation Rates of Etilingera elatior and the Possible Effects on Foraging Patterns of Some Costa Rican Hummingbird Species. Willimantic, Connecticut: Department of Biology, Eastern Connecticut State University. Latta, G. (2012). Weather Information Page. Retrieved from http://faculty.frostburg.edu/phys/latta/weather/. Miller, R.M. & Miller, R.E. (1971). Feeding Activity and Color Preference of Ruby-Throated Hummingbirds. The Condor, 73, 309-313. Nicolson, S.W. & Fleming, P.A. (2003). Nectar as Food for Birds: The Physiological Consequences of Drinking Dilute Sugar Solutions. Plant Systematics and Evolution, 238, 139-153. 10.1007/s00606-003-0276-710.1007/s00606-003-0276-7. Poslethwalt, J.H & Hopson, J.L. (2006). Modern Biology. Boston: Holt Reinhart & Winston. University Of Maine Cooperative Extension Publications. (2002). Understanding Ruby-Throated Hummingbirds and Enhancing Their Habitat in Maine. Retrieved from ftp://ftpfc.sc.egov.usda.gov/whmi/web/pdf/rubythroated1.pdf. U.S. Dept. of Agriculture Wildlife Management Institute. (1999, December). Ruby-Throated 17

Hummingbird (Archilochus colubris): Fish and Wildlife Habitat Management Leaflet. Retrieved from ftp://ftp-fc.sc.egov.usda.gov/whmi/web/pdf/rubythroated1.pdf. Why do our Hummingbirds Hum? (2000, Summer). Retrieved from http://www.botgard.ucla.edu/html/membgnewsletter/volume3number3/hummingbirds.html. 18