Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Traplining foraging behavior in a tropical hummingbird species phaethornis superciliosus Garrison, Jennifer Susan Eileen 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1995-0600.pdf [ 3.78MB ]
Metadata
JSON: 831-1.0099029.json
JSON-LD: 831-1.0099029-ld.json
RDF/XML (Pretty): 831-1.0099029-rdf.xml
RDF/JSON: 831-1.0099029-rdf.json
Turtle: 831-1.0099029-turtle.txt
N-Triples: 831-1.0099029-rdf-ntriples.txt
Original Record: 831-1.0099029-source.json
Full Text
831-1.0099029-fulltext.txt
Citation
831-1.0099029.ris

Full Text

T R A P L I N I N G F O R A G I N G B E H A V I O R I N A T R O P I C A L H U M M I N G B I R D SPECIES PHAETHORNIS SUPERCILIOSUS by JENNIFER S U S A N E I L E E N G A R R I S O N BS, Hons., University of California at Davis, 1992 A THESIS SUBMITTED I N P A R T I A L F U L F I L M E N T OF T H E R E Q U I R E M E N T S FOR T H E D E G R E E O F M A S T E R OF SCIENCE in T H E F A C U L T Y OF G R A D U A T E STUDIES D E P A R T M E N T OF Z O O L O G Y We accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH C O L U M B I A October, 1995 © Jennifer Susan Eileen Garrison, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date OJr n DE-6 (2/88) A B S T R A C T Traplining nectarivores are those that visit widely dispersed and often nectar-rich flowers that cannot be directly defended and are thought to follow similar routes on different foraging bouts. Phaethornis superciliosus is a large (6 g) hummingbird found in the lowland tropical forests of Central and South America, and is considered a preeminent example of a traplining hummingbird. Though P. superciliosus is considered a trapliner, no detailed studies of its movements have confirmed this. Through field observations and enclosure experiments on P. superciliosus, I examined whether they followed similar routes on different foraging bouts, and considered some factors which could affect their visitation rates to patches of flowers. In the field, birds' arrival and departure angles to/from a patch were quite similar over one to several days, and different birds used different arrival and departure directions from the same patch. This supports the idea that they were following routes, and that these routes were based on the locations of patches of flowers, rather than on open flyways through the forest. Both my field and enclosure studies suggest that P. superciliosus can detect and respond to changes in nectar production rates at individual patches along their traplines. Birds in the enclosure increased or decreased relative use of feeders when I manipulated their nectar production rates. Because traplining birds do not defend their flowers from competitors, they should respond to competition by returning sooner to flowers visited by other nectarivores (exploitative defense). In my field study, I observed a positive relationship between how long a bird waited between visits to a patch and the number of competitive visits by other birds to the patch. There was a positive relationship between the amount of nectar removed from the feeder and birds' relative use of the feeder, indicating that birds do respond to competition. i i The currently accepted model of constant net energy intake by territorial hummingbirds does not accurately reflect feeding behavior of traplining birds. P. superciliosus has a gross nectar intake that decreases through the day, mirroring nectar production rates in its food-flowers. I present a simulation model in which trapliners have decreasing rather than constant net energy intake rates. Mode l birds wi th decreasing net intake rates can meet their energetic needs wi th fewer flowers than model birds with constant net intake. i i i T A B L E OF C O N T E N T S Abstract i i Table of Contents i v List of Tables v List of Figures v i Acknowledgments ix Chapter 1: General Introduction • 1 Chapter 2: A Field Study of Traplining in Phaethornis superciliosus 5 Methods 7 Results 11 Discussion . 22 Chapter 3: A n Enclosure Study of Traplining Behavior 29 Methods 32 Results 38 Discussion 45 Chapter 4: H o w Trapliners Regulate Energy Intake 49 Methods 50 Results and Discussion 51 Chapter 5: General Conclusions 66 Traplines 66 Affects of Nectar Availabil i ty 67 Compet i t ion 68 Energy Regulation 70 Literature Cited 72 Appendix I: Mode l Parameters 79 Appendix II: Nectar Production Rates 80 Appendix III: Fat Accumulation Rate Model 81 iv LIST OF T A B L E S Table 1. Ar r iva l and departure angles and Rao's Spacing Test results for birds at patches F l ( a) and L I ( b). Date: numbers indicate which days of late M a y or early June observations took place. Dir= direction (arrival or departure). N = number of arrival or departure angles recorded. <|>= mean angle of arrival or departure, r = mean vector length. P= significance level for Rao's spacing test. ¥= the small r values for these data sets are due to a bimodal distribution of the arrival and departure angles. The Rao's test allows for multi-modal distributions, while calculation of r does not. *= Rao's test not performed because sample size was too small. Table 2. Act iv i ty /hour means for all birds visiting patch F l . Time= time of day (h). N A I I = number of observation periods (days) for all birds. A l l = mean + s.e. for al l birds combined. N B Y B = the number of observation periods for BYB alone. BYB= mean ± s.e. for BYB alone. Individuals = mean number of individual marked (or unmarked) birds visiting the patch for each time period. Visits to patch= total number of times birds visited the patch. Total Probes= total number of times flowers were probed. Time Elapsed= time to next visit (min) ±S.E. Table 3. Activi ty totals for Patch F l over days. Date= date of observation. Al l= total number observed for each day for all birds. BYB= total number observed for each day for BYB. Individual visitors^ total number of indiv idual birds visiting the patch on each day. Visits to patch= total number of visits by all birds to the patch for each day. Probes= number of times flowers were probed for each day. v LIST OF FIGURES Figure 1: a) Ar r iva l and b) Departure Angles for BPB at Patch L I , June 1-3. Dots indicate individual arrival or departure angles. See Table 1 for description of symbols. Arrows point to mean angles, and the length of the arrow is dictated by r. Figure 2: Ar r iva l Angles for BYB at Patch F l . a) M a y 19-20, b) M a y 21. Symbols as in Fig. 1. Figure 3: Mean time between visits (min) by all birds at Patch F l vs time of day (h). (r2= 0.218, N= 53 ,F 4= 2.98, P= 0.029). Error bars = 1 S.E. Figure 4: Number of competitive visits by other birds vs time elapsed between visit to the patch for individual birds. Time Between Visits (min)= the amount of time an individual spent between visits to the patch. Competitive Visits = total number of visits by other birds between that individual 's last and current visit. The relationship is still significant if the largest value is removed. N= 41, r 2 = 0.313, F= 17.81, P< 0.001, using ranks of competitive visits. This relationship is still significant without the two highest values: r 2= 0.203, F 4 = 9.45, P= 0.004 Figure 5: Top view of enclosure. Enclosure is 3.5 x 7.2 x 2.6 m. Symbols are defined as follows: = blind, T = perch, O = feeder Figure 6: Estimated nectar production rates for feeders in enclosure (dashed line) and Hel iconia pogonantha flowers in the w i l d (solid line; from Stiles 1975. This estimate may be low if nectar removal by birds increases total nectar output by the flower; G i l l 1988b). Production rates are in u l /h r . Figure 7: Vis i ts /hour to feeders over time of day. Mean for 20 bird-days (Control only). Error bars indicate 1 S.E. Figure 8: Relative use of feeder for Control, Decrease, and Increase days (mean for 2 birds for each treatment). C O N - D= Control prior to Decrease treatment, D E C = Decrease, C O N -1= Control prior to Increase treatment, INC= increase. Error bars= 1 S.E. Figure 9: Total visits per day to all feeders for Control and Satiation days. C O N = Control days. SAT= Satiation days. Mean for 4 birds. Error bars = 1 S.E. Figure 10: Relative use of feeder (with 1 hour time lag) vs total p i removed from vi feeder. Relative uses of feeders are taken from 1 hour later than the amount of nectar removed to account for the time lag in responses to the removal of nectar. Figure 11: Relative use of feeder on Competition day (solid line), and total amount of nectar (in ul) that I was able to remove from feeder (dashed line). This graph shows the response of a bird to the removal of a large quantity of nectar. Figure 12: Gross Energy Intake (kj) vs Time of Day. Error bars = 1 S.E. Spearman rank correlation of mean (for all birds) nectar consumed and time of day: r s= -0.74 Figure 13: Mean visits per hour (for all birds) to feeder. Error bars= ± 1 SE. Figure 14: The effect of % time flying and % of flying time hovering on net intake of energy for P. superciliosus at an ad libitum feeder, a) Observed gross intake and assumed energy expenditures ( i n kj) for 5 levels of % time flying, b) net intake (gross intake - assumed expenditures) in kj, c) cumulative net intake in kj. % time flying is set at 0600 to one of five levels and then decreases throughout the day in proportion to nectar consumption. H= min imum net intake for bird to survive the night (Hainsworth 1978). Figure 15: The effect of % time flying and % of flying time spent hovering on energy accumulation for a 12 hour day. Contours connect points of equal energy accumulation (in kj). 0 energy accumulation means that bird exactly matches expenditures with intakes, but has no reserves for the night. The line labelled 14 represents Hainsworth's (1978) estimated energy accumulation required for the bird to survive the night and achieve a 24 hour neutral energy budget. Figure 16: Whole day values for Constant and Decrease birds over different numbers of flowers, a) Number of flower visits per day/20000 (solid), Number of bouts per day/500 (large dash), and % time spent flying (small dash), b) Energy taken in (solid), Energy expended (large dash), and Net energy intake (small dash), all in kj. c) Change in mass (starting mass - mass at end of day) in g-Figure 17: Values for Constant and Decrease birds for traplines wi th 30 flowers, a) Number of flowers visited per bout/50 (large dash), Mean volume per flower in u l / 40 (solid), and % time spent flying (small dash), b) Duration (s) of perching/ 120 min (large dash), foraging bout/20 min (small dash) and whole cycle (perching + flying)/120 min (solid), c) Energy taken in/3.6 (solid), Energy v i i expended/3.6 (large dash), and Net energy intake/3.6 (small dash), all in kj. d) Relative mass (Current Mass/Starting Mass - 0.5) in g (small dash; A bird that stays the same weight all day would generate a horizontal line at O.5.), Cumulative fat gain/20 in kj (large dash) and Fat accumulation rate in watts (solid). v i i i A C K N O W L E D G E M E N T S M a n y people contributed time and ideas to help me complete this thesis. I am especially grateful to my advisor, Lee Gass, for unending encouragement and support, and for the many brainstorming sessions about traplining. He also taught me not to give up when things don't work out as planned. M y research committee members, Jamie Smith and Tony Sinclair, provided much valuable advice on how to focus my research. Thanks, Tony, for sitting out in the woods wi th me at La Selva and encouraging me to keep at it. Thanks to Tony L u m and Don Brandys who built my data logger, and to the students in the Gass Lab who originally developed the hummingbird lab. I am deeply indebted to Lance 'Zaphod' Bailey and Alistair 'Big A ' Blachford for spending hours helping me develop computer programs to analyze my data, and to Wesley 'Hoocher' Hochachka for answering all of my questions about statistics, all the while with a smile on his face. Thanks also to Marc Mangel , N e i l Dryden, and Rob Ritchie, who provided mathematical advice on the model. Several people made my research in the tropics a little easier: thanks to Mar ian Sanchez and Marybel Soto of OTS for arranging my stay, and for arranging for permits and various construction materials, and to Dav id and Deborah Clark for being wonderful station directors. Rick Ree, who assisted me in the field, made my life a lot easier and a lot more fun. I hope he didn't mind the bot flies, and that some day he w i l l master dancing Merengue! I also want to thank the many friends I made at La Selva and in Vancouver, for making my stay in both places an adventure. Thanks especially to Devon Graham and Paul Keck for saving my life. Most of all , I wish to thank my parents, John and Hillegonda, for both emotional and financial support, and for always encouraging me to follow my dreams. This research was funded by a N S E R C operating grant to Lee Gass, and by a Frank M . Chapman Memorial Grant from the American Museum of Natural History. Personal funding was supplied by a University Graduate Fellowship and a Teaching Assistantship from the University of British Columbia, and by my parents. ix C H A P T E R 1 G E N E R A L I N T R O D U C T I O N Optimal foraging theory predicts that the foraging strategies and tactics of animals w i l l vary with changes in the habitat, the competitive environment, and the renewal rate and distribution of resources (Miller 1967; Linhart 1973; Waser 1981; Stephens and Charnov 1982; Yamamura and Tsuji 1987; Feinsinger et al. 1988). For example, nectar production rates can substantially affect nectarivore activity budgets and movement patterns (Janzen 1971; Hainsworth & Wol f 1972; Feinsinger & Colwel l 1978; Schemske 1980; Gass & Sutherland 1985). In nectarivorous birds, foraging strategies form a continuum between territoriality and traplining depending on the spacing of flowers and the quantity of nectar produced (Feinsinger & Colwel l 1978; Stiles & Wolf 1979; Stiles 1981). Territoriality is common when nectar dispersion, quantity, and quality allow birds to gain more energy from defending and using territories than they spend (Brown 1964; Stiles and Wolf 1970; Linhart 1973; Carpenter & M a c M i l l e n 1976). Nectarivore territories usually consist of large numbers of low-yield flowers clumped in meadows, or bushes or trees with many inflorescences (Wolf & Hainsworth 1971 ; Carpenter & MacMil len 1976; Kodric-Brown & Brown 1978), although hummingbirds also defend artificial feeders (Ewald & Orians 1983) and sapsucker drillings (Southwick & Southwick 1980; Sutherland et al. 1982). Nectarivores that do not utilize or are excluded from rich resource clumps should use any dispersed resources available (Feinsinger 1976; Gass 1978b). There are a number of foraging strategies used by birds who exploit dispersed resources, including haphazard foraging and traplining (Feinsinger 1976). By definition, 1 General Introduction traplining nectarivores exploit dispersed resources by following regular feeding routes between flowers on scattered trees, shrubs, vines, and herbs (Feinsinger and Chapl in 1975; Feinsinger 1987; G i l l 1988a). Routes are not static: trapliners explore patches of flowers outside their traplines and incorporate profitable patches into their traplines and drop unprofitable patches (Thomson el al. 1982, 1987). Traplines, or foraging circuits, enable animals to harvest reliably renewing food such as nectar from isolated sites in a regular fashion in order to minimize the time and energy spent moving between clumps (Schoener 1971; Wolf et al. 1975; Stiles & Wolf 1979). A s wel l , if traplining nectarivores revisit only profitable clumps, they can obtain substantially more nectar per feeding bout than individuals who forage haphazardly (Gil l & Wol f 1977; Kami l 1978; Armstrong et al. 1987). Visi t ing feeding sites regularly may also deter competition by keeping resources at a low level (Charnov et al. 1976; Janzen 1971; Thomson & Plowright 1980; Davies & Houston 1981; G i l l 1988a). Whi le territorial nectarivores have been studied in great detail, there is much to be learned about the foraging behavior of trapliners. For example, there is very little evidence in the literature that "traplining" nectarivores actually follow routes. As wel l , only a few studies have examined how trapliners pattern visits to patches of flowers along their traplines (Thomson et al. 1982, 1989), or how they keep track of changing profitabilites of different patches due to declining nectar production rates over the day (Stiles 1975, Stiles & Wolf 1979), competition by other nectarivores (Gil l 1988a), and number of flowers open in patches. This study examines traplining behavior of Phaethornis superciliosus, a long-bil led hermit hummingbird species which uses widely-dispersed, highly productive flowers, such as Heliconia pogonantha in the lowland tropical rainforests of Central America. The amount of nectar produced at small- to medium-sized patches of H . pogonantha and the wide spacing of these patches predict that the 2 General Introduction birds that utilize these plants w i l l be trapliners rather than territorial, because while individual patches provide large amounts of nectar, the amount is not great enough to support an individual for an entire day (Stiles 1975; Stiles & Wolf 1979). P. superciliosus rarely (if ever) defend patches of Heliconia. Patches that produce enough nectar to support a bird for a whole day are usually defended by more aggressive species of non-hermit hummingbirds (Stiles & Wolf 1979). Al though the non-hermits are less specialized for use of H . pogonantha (no long, curved bills) they can utilize large patches that produce copious amounts of nectar, because nectar overflows from the main nectar chamber into the corolla of the flower (Gi l l 1989). By observing P. superciliosus feeding at patches of H . pogonantha and through laboratory studies of their foraging behavior, I hoped to determine whether they follow routes and what factors affect how they pattern their visits over time. In chapter 2,1 discuss results from a field study of the foraging behavior of P. superciliosus at patches of Heliconia pogonantha. I address whether P. superciliosus follow repeated routes, patterns of total activity at patches of flowers, and how competition affects visitation rate. In chapter 3 I discuss an enclosure study of P. superciliosus on how changes in nectar production rates affect use of feeders, including decreasing nectar production rate over the day and changes in nectar production at individual feeders as wel l as in total nectar availability in the enclosure. As well , I address how birds respond to simulated competition at feeders. In Chapter 4,1 discuss possible methods of energy regulation in traplining birds. The currently accepted model of energy regulation in territorial hummingbirds through constant net intake over the day (Hainsworth 1978) does not apply to traplining birds. I present 3 models to help understand how traplining birds regulate energy and discuss the implications of Hainsworth's (1978) assumption that hummingbirds w i l l increase foraging effort with decreasing nectar availability in order to maintain constant net intake rates throughout the day. In chapter 5 I bring 3 General Introduction together the results from chapters 2-4, and discuss how they relate to current ideas about traplining foraging behavior. 4 CHAPTER 2 A FIELD STUDY OF TRAPLINING IN NECTARIVOROUS BIRDS The term traplining was first coined by Janzen (1971) to describe the foraging behavior of female euglossine bees, which flew long distances between flowers and seemed to visit plants along foraging routes that were consistent from bout to bout. Traplining has since been observed in other nectarivores, including bumble-bees (Thomson 1988; Thomson et al. 1982, 1987,1989), sunbirds (Stiles 1981), and two groups of hummingbirds: low-reward trapliners: small, short-billed species excluded from territories by more aggressive nectarivores (Feinsinger & Chapl in 1975; Feinsinger 1976), and high-reward trapliners: larger, long-billed species that exploit widely dispersed nectar-rich flowers (Linhart 1973; Feinsinger & Colwe l l 1978; G i l l 1988a; Stiles 1978; 1981). It is, however, unknown whether the nectarivores that have been called trapliners actually follow the same routes repeatedly (Stiles & Wolf 1979; Schemske 1980). Chlorosti lbon canivetii is considered to be a low-reward traplining hummingbird and it revisits clumps of flowers in nearly the same sequence throughout the day (Feinsinger 1976). Phaethornis superciliosus has been described as a high-reward traplining hummingbird, and is thought to use travel routes between patches which remain fairly consistent over many separate visits (Stiles and Wol f 1979). However, neither Feinsinger nor Stiles and Wolf presented much evidence to support the claim that the visits were repeated in the same sequence on different bouts or that the arrival and departure directions were consistent, i.e. that the birds were actually traplining. Bombus spp. bees visit sets of flowering shoots in a regular sequence and repeat this sequence several times in a single trip (Thomson 1988; Thomson et al. 1982, 1987,1989). They also sample shoots outside of their 5 traplines, add or drop flowers from their routes, and gradually shift their foraging routes to include more rewarding plants (Thomson et al. 1982, 1987). Bees' foraging patterns were similar over successive days, which suggests that they remembered foraging routes from day to day. Thomson et al. (1982) provided maps of the bees' traplines, but d id not statistically test the routes for consistency. The bees d id seem to be loosely following routes, although there was a lot of "noise" around them. It is l ikely that the routes these bees follow w i l l not be entirely similar over different bouts, simply because they forage in dense fields of flowers and have many flowers to choose from on any bout. In the tropics, many flowers utilized by trapliners have decreasing nectar production rates over the day (Stiles 1975; Feinsinger 1976; Schemske 1980; G i l l 1988b). If nectarivores need constant net intake rates (Hainsworth 1978) and they have a l imited number of flowers to forage from, they should visit patches less often as nectar production rates decrease (to avoid spending more energy than they take in). This is seen in several species of tropical nectar feeding birds in the laboratory (Tiebout 1992: territorial Amaz i l i a saucerottei and traplining C . canivetii), and at flowers in the w i l d (Feinsinger 1976: C. canivetii; Frost & Frost 1980: sunbirds; Stiles and Wolf 1979: P. superciliosus). Another way of maintaining constant intake rates is to increase foraging effort (i.e. visit more flowers) with decreasing standing crop (with the assumption that visiting more flowers w i l l increase intake rates). Many territorial hummingbirds (such as Selasphorus rufus) increase foraging effort with decreasing returns, both in the laboratory (Gass 1978a) and in the field (Gass 1978b; Gass & Montgomerie 1981; Sutherland et al. 1982; Gass & Sutherland 1985). However, territorial hummingbirds generally do not have to fly long distances to feed from flowers in their territory, which may keep the costs of their increased activity lower than for traplining birds. As well , territorial birds 'control' the standing crop of their 6 A Field Study of Traplining territories while trapliners do not, making it more likely that increased foraging effort by territorial birds w i l l lead to increased nectar intake. Trapliners whose visitation rates do not mirror nectar production rates w i l l most l ikely lose much nectar to competitors. It is likely that traplining birds have evolved different foraging strategies to deal with the dispersion and abundance of nectar they encounter i n their environment. Because trapliners can "defend" the flowers in their traplines only by keeping nectar levels depleted, they should respond to competition by returning more often to patches with competition in order to keep nectar levels low and discourage the other visitors (Frost & Frost 1980; Feinsinger 1987; G i l l 1988a). P. superciliosus detected and responded to competition at high-reward feeders in the field by returning sooner after losing nectar to competitors than when not faced by competition (Gi l l 1988a). Whether they do the same at natural flowers is unknown. This chapter examines traplining behavior of Phaethornis superciliosus, a long-billed hermit hummingbird species at patches of their main food flower, Hel iconia pogonantha, in a lowland tropical rain-forest of Central America. METHODS I observed patches of H . pogonantha for 99 hours at La Selva Research Station in Sarapiqui, Costa Rica from May 13 to June 3 1993. I chose small patches (less than 10 inflorescences per patch) of H . pogonantha to study, because larger ones were often included in non-hermit hummingbird territories, as most hermit species tend not to be territorial (Stiles 1975; Stiles & Wolf 1979). H . pogonantha are found in light gaps and along forest edges. The flowers produce copious quantities of rich nectar (0.92 M) : on a daily basis an individual flower can produce 10 times that of flowers used by territorial birds (Gass et al. 1976; Feinsinger 1978; Feinsinger 1987; Wolf & Hainsworth 1991; Stiles & Freeman 1993). 7 A Field Study of Traplining Nectar production is highest in the morning and decreases rapidly throughout the day (Stiles 1975). H . pogonantha patches normally have 2-6 inflorescences and up to 53 bracts/inflorescence (Stiles 1975), although the plants in my study usually had less than 20 bracts/inflorescence. Each bract can have one open flower at a time, but consecutive flowers in a bract open at 5 day intervals. Inflorescences can have up to 7 open flowers, but normally have only 1-3 at a time (Stiles 1975; personal observation). Flowers last only one day, so nectar production rates at patches can change daily if the number of bracts with open flowers changes. To identify individual birds who visited the patch, I color marked them by attaching small strips of flagging tape to their backs wi th super-glue (Hixon et al. 1983). These markings usually remained on the birds' backs for at least a week (personal observation). Individuals are referred to in the text by the colors of their flagging tape (B= blue, Bl= black, P= pink, W= white, Y= Yellow). A few visiting birds were not captured but only 5% of all flower-probes in this study were made by unmarked individuals . I began observations at each patch the day after color-marking, and watched patches for 1 or 3 days depending on the number of birds visiting the first day (1 day for no visits, 3 days for >0 visits). I observed from 0700 to 1130 h, sitting 2-4 m from the patch of flowers. I named patches of flowers according to their location on the research station grounds (F for forest and L for lab-clearing), followed by a number that I assigned to the patch (example: F l ) . I recorded the identity of each bird visiting the patch, and its arrival and departure directions, time of arrival, number of flowers probed, and any other noteworthy behavior such as interactions wi th other birds. For each visit I calculated the time to the next visit, and the number of visits by other birds (competition) between the last and current visit. I estimated angles of arrival and departure (to the nearest 5 degrees) with a Silva sighting compass, and noted times (to ± 1 min) with a digital wrist watch. I was aided in detecting birds 8 A Field Study of Traplining arriving at a patch by their flight calls as they approached the patch (Stiles & Wolf 1979). A r r i v a l and Departure Angles Bi rd arrival and departure angles were graphed on circles and analyzed for non-random spacing using Rao's Spacing Test with an a level of 0.05 (Batschelet 1981). Mean angle ((j), rounded to nearest degree) and the length of the mean vector (r) were also calculated for each bird-day. r represents the variance around the mean angle; r= 1.0 indicates that a bird arrived from (or departed in) exactly the same direction every time, and values near zero indicate that the arrival or departure directions were widely scattered around the mean (())). If sample sizes for a bird were small and mean angles were similar between days (if the 95% confidence intervals (Batschelet 1981) overlapped between days for both arrival angles and departure angles), I pooled days. Sample sizes were too small in most of these cases to statistically test for differences between days. I pooled data i n the cases of BYB (May 19-20), PB1 (May 20 - 21), and BPB (May 31 - June 3). Patch Act iv i ty I tallied several variables for each hour of each day: 1) the number of individuals visiting the patch (each marked bird was counted only once and unmarked birds were considered to be one individual and counted only once), 2) total number of bird-visits (number of times birds visited the patch), and 3) total probes (the total number flowers investigated by all birds). The effects of time on the above variables were tested using A N C O V A s . T iming of Visits Time to next visit was was log transformed to normalize its distribution, and tested for normality wi th the Kolmogorov-Smirnoff one sample test against a 9 A Field Study of Traplining standard normal distribution. The effects of date, time, number of flowers probed, number of competitive visits by other birds, and all two-way interactions on time to next visit were tested using an A N C O V A . This was done in two steps: first, all independent variables and possible two-way interactions were placed in the model. For the second step, all insignificant two-way interactions were removed from the model. Independent variables were not removed. Three way and four way interactions were not used due to small sample sizes. A l l F, P, and r 2 values presented below are from the second, reduced, version of the model. Patch Descriptions A t most patches, the visitation rate was extremely low. I had enough records of bird visits to analyze visitation rates and flight angles for only two (F l and LI ) of the 12 patches observed (comprising 27 of the 99 hours of observation). Patch F l had one unmarked and four marked birds visiting it. Of these, three (BYB, PB1, and BWP) visited often enough to use in the analysis of angles. Two other birds (BB1 and unmarked) were included in the visitation analysis but not in the analysis of angles because of missing data. I had 3 continuous mornings of observation for BYB and PB1 (May 19-21), two (May 19 and 20) for BB1, and one (May 21) for BWP. BYB accounted for 50% of all probes (by 5 birds) to flowers for the three days, while the other individuals each made only 5-17%. Since BYB was the primary visitor to the patch, I examined its activity patterns more closely and separately from others. Another patch (F2) was ~2 m from F l at 270°, although its flowers were not easily visible from my observation point. The other patch (LI) wi th enough bi rd visits to do analysis of angles had only one visitor, BPB, on June 1-3. Another patch (L2) was about 20 m away from L I , at 265°. The area surrounding L I was less densely vegetated than around patch F l and visibility was high, making it easier to observe 10 A Field Study of Traplining BPB's arrival and departure angles than those of the birds at F l . RESULTS A r r i v a l and Departure Angles If individual birds follow foraging routes they should approach and depart patches from a small set of directions. Except for BYB's departure angles on M a y 21 (which were randomly distributed), and PBl's departure angles on M a y 20-21 (which had too small a sample size to analyze), all birds' flight patterns were significantly non-randomly distributed (clumped) (Table 1). In most cases birds approached from one direction and departed in a different direction. BYB appeared to have bimodal sets of arrival and departure angles, so that it arrived from 2 directions and departed in 2 directions on different foraging bouts. Each bird at patch F l used different directions for travel than the others, although all birds arrived from or departed toward the general direction of patch F2 (270°) for at least one observation period (Table 1). BPB flew from another patch (L2: 265°) directly to patch L I on every visit (Fig. 1). Some birds were consistent in their mean arrival and departure angles over the three day periods (Table 1, Fig. 1) while others changed between days (Table 1, Fig. 2). This switch in direction was rapid: there were few points of overlap between days in the arrival (or departure) angles for birds that changed direction. In most cases, it was the arrival angle that changed, while departure angles remained similar between days. Patch Act iv i ty If birds regulate their foraging to maintain the profitability of visiting patches, the number of birds visiting a patch and the number of times birds visit per hour should decrease during the day because the profit of visiting the patch after a given duration decreases with decreasing nectar production. A t patch F l , wi th several 11 A Field Study of Traplining visitors, the number of individual birds visiting decreased significantly (r 2= 0.765, F 3 = 5.5, P= 0.037, Table 2). The total number of bird visits also seemed to decrease as the day progressed, but the differences were not statistically significant (r2= 0.489, F 3 = 1.626, P= 0.280, Table 2). It is not known whether P. superciliosus visit other "fall back" patches during periods of low nectar production, or if they simply decrease their foraging effort wi th decreasing nectar production. Whi le visitation rates may decrease with decreasing nectar production rates, there is no reason to expect the number of flowers probed on any visit to change much over time (since there are so few flowers). Indeed, the total number of flowers that birds probed per hour fluctuated as the day progressed, but differences between hours were not significant (r2= 0.596, F 3 = 2.74, P= 0.136, Table 2). Whi le activity levels should decrease during the day, the levels should be similar between days as long as similar numbers of flowers remain open at the patch between days, the flowers produce nectar at similar rates, and similar numbers of individuals are active. A t patch F l , the total number of individuals visit ing the patch was similar on all days (r2= 0.765, F 2 = 1.5, P= 0.296, Table 3). The total number of bird-visits to the patch over the three days remained fairly consistent (r2= 0.489, F 2 = 0.429, P= 0.670, Table 3), as d id the total number of times birds probed flowers (r2= 0.596, F 2 = 0.307, P= 0.746,Table 3). Small patches with few open flowers produce less nectar and should receive fewer visitors than more productive patches (such as F l , which had 4-5 flowers open on any given day). Patch L I had only 1-2 flowers open, and this is reflected in activity at the patch: BPB was the sole visitor. The total number of bird-visits to patch L I (1-4 visits per observation period) was much lower than to F l (16-20 visits per observation period). 12 A Field Study of Traplining Table 1: Ar r iva l and departure angles and Rao's Spacing Test results for birds at patches F l ( a) and L I ( b). B i r d Date Dir N 4> r P BYB a 19-20 Arr 15 293 0.70 0.01 Dep 15 322 0.66 0.01 ¥ 21 Arr 10 112 0.43 0.01 ¥ Dep 11 326 0.37 0.10 ¥ PB1 a 19 Arr 4 89 0.96 0.01 Dep 5 255 0.91 0.01 20-21 Arr 3 0 0.99 * Dep 5 289 0.99 0.01 B W P a 21 Arr 4 278 0.92 0.05 Dep 8 266 0.99 0.01 BPB b 1-3 Ar r 8 267 0.99 0.01 Dep 8 101 0.98 0.01 Date: numbers indicate which days of late May or early June observations took place. Dir= direction (arrival or departure) N= number of arrival or departure angles recorded. <|>= mean angle of arrival or departure, r = mean vector length. P= significance level for Rao's spacing test. ¥= the small r values for these data sets are due to a bimodal distribution of the arrival and departure angles. The Rao's test allows for multi-modal distributions, while calculation of r does not. *= Rao's test not performed because sample size was too small. 13 A Field Study of Traplining Figure 1: a) Ar r iva l and b) Departure Angles for BPB at Patch L I , June 1-3. Dots indicate individual arrival or departure angles. See Table 1 for description of symbols. Arrows point to mean angles, and the length of the arrow is dictated by r. 14 A Field Study of Traplining a) 90 b) 270 Figure 2: Ar r iva l Angles for BYB at Patch F l . a) M a y 19-20, b) M a y 21. Symbols as in Fig. 1. 15 A Field Study of Traplining Table 2: Act iv i ty /hour means for all birds visiting patch F l . Activi ty Time N A U A l l N B Y B BYB Individuals 7 3 3.0 ± 0.00 8 3 2.3 ± 0 .33 9 3 1.7 ± 0 . 3 3 1 0 3 1.7 ± 0 . 3 3 Visits to Patch 7 3 5.0 ± 0.58 2 2 .5 ± 0 . 5 0 8 3 5.3 ± 0 .33 2 4 .0 ± 1 . 0 0 9 3 3.3 ± 1.45 3 1.7 ± 0 . 6 7 1 0 3 3.0 ± 0.58 3 2 .0 ± 0 . 5 8 Total Probes 7 3 8.0 ± 2 .00 2 5.0 ± 2 . 0 8 3 11.3 ± 2 . 0 3 .2 10.5 ± 2.5 9 3 4.3 ± 0.88 3 4 .0 ± 1 . 0 1 0 3 10.3 ± 1 . 7 0 3 8.0 ± 2 . 3 Time Elapsed 7 1 2 24.1 ± 5.54 5 18.8 ± 9 . 1 5 8 1 3 28.4 ± 6.83 8 20.1 ± 5 .80 9 9 39.8 ± 13.3 5 37 .0 ± 8 .23 1 0 6 31.0 ± 8 . 3 6 4 25 .5 ± 5 .27 11 1 23.0 1 23 .0 Time= time of day (h). NAII= number of observation periods (days) for all birds. Al l= mean ± S.E. for all birds combined. N B Y B = the number of observation periods for BYB alone. BYB= mean ± S.E. for BYB alone. Individuals = mean number of individual marked (or unmarked) birds visiting the patch for each time period. Visits to patch= total number of times birds visited the patch. Total Probes= total number of times flowers were probed. Time Elapsed= time to next visit (min) ± S.E. 16 A Field Study of Traplining Table 3: Activi ty totals for Patch F l over days. Activity Date A l l BYB Individuals May 19 4 May 20 4 May 21 3 Visits to Patch May 19 18 5 May 20 16 9 May 21 20 11 Probes May 19 38 24 May 20 37 27 May 21 36 22 Date= date of observation. Al l= total number observed for each day for all birds. BYB= total number observed for each day for BYB. Individual visitors= total number of individual birds visiting the patch on each day. Visits to patch= total number of visits by all birds to the patch for each day. Probes= number of times flowers were probed for each day. 17 A Field Study of Traplining Timing of visits T iming of visits to the patch should depend on the amount of nectar available and on the competitive environment. However, there was no obvious temporal patterning of visits to patch F l : birds allowed variable amounts of time to pass between visits. Timing of visits by individual birds seemed to vary independently of date, time of day, number of flowers probed, and competition (r2= 0.173, N = 32, all F-values < 0.9, all P > 0.4). This indicates that birds were not waiting longer between visits as nectar production declined throughout the day as I had originally expected. The times between visits to the two patches (Fl and LI ) were quite different: BYB visited patch F l on average every -25 min., while BPB visited patch L I every ~85 min . Time elapsed between visits by birds to a patch w i l l affect the amount of nectar available at the patch. If birds arrive closely together in time, very little nectar w i l l accumulate from one visitor to the next. Average time between visits by any bird to patch F l was 22.4 ± 3.4 minutes. Time between visits ranged from 0 to 97 min and was heavily skewed towards zero. Median time elapsed between visits to the patch was 11.5 min. Time between visits increased with time of day (r2= 0.218, F 4 = 2.98, P= 0.029, Fig. 3) but d id not change between days (F2= 0.766, P= 0.471). Compet i t ion Wait ing longer before visiting a patch should expose birds to increased competition by increasing the chance of other birds removing nectar (assuming the competitors do not also wait longer). In fact, the longer a bird waited between visits to patch F l , the more competition (number of visits by other birds) it faced there (r2= 0.313, F 4 = 17.8, P= 0.001, using rank transformed competition, Fig. 4. This relationship is still significant without the two highest values: r 2 = 0.203, F 4 = 9.45, P= 0.004). 18 A Field Study of Traplining If birds respond to competition at patches by returning to them more often (Gil l 1988a), the number of competitive visits and the time elapsed between two consecutive visits by an individual bird should be negatively related. Birds as a whole returned no sooner to patch F l after more competitive visits than after few ( A N C O V A mentioned above in Timing of Visits: F 1 = 0.430, P = 0.518). If birds respond only to the presence or absence of competition, and not to the number of competitive visits, then they would wait longer between visits when not faced by competition than when faced by it. Since all birds at F l except BYB always had at least one competitive visit between their own visits, I analyzed BYB's return rates alone to determine if it waited longer between visits when not faced by competition than when it was. BYB spent more time between visits when it faced no competition since the last visit (27.7 ± 5.3 min) than when it d id (17.7 ± 3.3 min). However, the difference is not significant (Mann-Whitney U Test: N = 22, U= 77, P= 0.262). It appears that P. superciliosus does not respond to competition as strongly as expected at patches of natural flowers based on its response to competition at highly profitable feeders (Gill 1988a), at least by varying time between visits to the patch in response to level of competition. It is possible that birds respond only to presence or absence of competition, and that a change in the presence or absence of competition at a patch would evoke a response, such as is seen at Gi l l ' s feeders. 19 A Field Study of Traplining 70 n 'c 60 -E 0 H—'—i—•—i—•—i—•—i—i 1—i 1 6 7 8 9 10 11 12 TIME OF DAY (h) Figure 3: Mean time between visits (min) by all birds at Patch F l vs time of day (h). (r2= 0.218, N= 53 ,F 4= 2.98, P= 0.029). Error bars = 1 S.E. 20 A Field Study of Traplining 20 n <2 15 H > LU > H h-UJ Q_ o o f- 10 50 100 150 TIME BETWEEN VISITS (min) 200 Figure 4: Number of competitive visits by other birds vs time elapsed between visit to the patch for individual birds. Time Between Visits (min)= the amount of time an individual spent between visits to the patch. Competitive Visits = total number of visits by other birds between that individual 's last and current visit. The relationship is still significant if the largest value is removed. N= 41, r 2 = 0.313, F= 17.81, P< 0.001, using ranks of competitive visits. This relationship is still significant without the two highest values: r 2= 0.203, F 4 = 9.45, P= 0.004 21 A Field Study of Traplining DISCUSSION As a result of the small sample sizes, the generality of my findings must remain in question pending further data. However, these data do provide some insight into the foraging behavior of P. superciliosus, and suggest what avenues future research might take to better understand traplining. in general. A r r i v a l and Departure Angles A s predicted by the traplining hypothesis, most birds approached and departed patches of flowers from a small set of directions (Table 1, Figs. 1 and 2). This suggests that they either visited a similar sequence of patches on successive foraging bouts, or were arriving and departing from similar directions on each bout because they used flyways in the forest (Westcott 1994) or were foraging from a central place. Different birds used different arrival (and departure) angles, which argues against use of flyways (Table 1). I know that birds d id not use the most obvious fly way (a trail passing close by) to approach and depart patch F l . A s wel l , arrival angles changed for some birds over the 3 day period, which indicates that they may have been fol lowing routes rather than arriving from some central place. Another argument against central place foraging is that birds' arrival angles were not the same as their departure angles for most days, suggesting that they flew to another patch rather than returning to a central place. I interpret these results to indicate that the birds I observed followed routes repeatedly. However, the only way to determine without a doubt that these birds were using the same routes on different foraging bouts wou ld be to place observers at many patches of flowers used by P. superciliosus, and map out the traplines used by individual birds over several days. It is l ikely that traplines change with productivity of patches or changing profitability due to competition with other birds. Traplining bees often add or drop flowers from their routes as nectar production rates change (Thomson et al. 1982, 1987), and the fact that some birds changed their angles of arrival or departure over 22 A Field Study of Traplining the three days of observation suggest that P. superciliosus do the same (Table 1, Fig. 2). Individual birds should react to the same patch in different ways depending on the profitability of the rest of their trapline, the level of competition at the patch, or other such factors. P. superciliosus changed the timing of visits to high reward feeders after a sudden change in the amount of nectar available, but it is unknown how this affected the rest of the trapline or timing of visits to individual patches (Gil l 1988a). In this study, both BB1 and PB1 visited patch F l less often as days progressed, while BYB visited more often, suggesting that the traplines of all three individuals were shifting over time. Patch Act ivi ty There is a l imit to how many birds can profitably visit a patch. When there are too many visitors, the nectar reward per visitor is likely to be quite low, which w o u l d in turn cause individuals who do not control the nectar harvesting schedule of a patch (Gil l 1988a) to drop out as it becomes unprofitable for them to visit it (as nectar production rates drop). This was the case: the number of individuals visiting patch F l decreased with time of day (Table 2). G i l l (1988a) also observed that if competition was intense at one of his artificial feeders, one or more of the birds visiting that feeder would decrease its use of it over time. Whi le patch productivity limits the number of individuals who can use it successfully, there should be similar use of the patch between days as long as patch productivity remains stable. Even though the usage patterns of indiv idual birds visit ing patch F l shifted over days, it received the same amount of attention from potential pollinators: total number of individuals, visits, and flower-probes were similar between days (Table 3). The size of a patch of a given species is likely the main determining factor in 23 A Field Study of Traplining how many visitors it w i l l receive. Small patches with few open flowers have zero to one visitors (personal observation). A bird visiting small patches may encounter less competition than one that visits large patches, but large patches produce more nectar. Which patch is more profitable to visit depends on the difference in the number of flowers, and the timing of visits by competitors. The more frequently birds visit a large patch the more flowers that patch must have in order to be equal in profitability to a small patch visited infrequently by one bird. It wou ld be interesting to discover whether there are individuals who specialize in small patches in order to avoid competition, and how this affects their overall net energy intake. T iming of Visits G i l l (1988a) observed that P. superciliosus revisited feeding sites from three to seven times per morning at intervals from a few minutes to two hours. I also found that birds spent variable amounts of time between visits. Contrary to my expectations, individual birds d id not increase the time elapsed between visits with decreasing nectar production rates, at least for the morning hours. However, the time elapsed between visits to the patch by any individual increased wi th time of day, and the total number of individuals visiting the patch decreased wi th time, suggesting that birds responded to decreasing nectar production. Stiles and Wolf (1979) found that fewer P. superciliosus visited patches of H . pogonantha in the afternoon when nectar production rates fall close to zero. It remains unknown what individuals do in the afternoon, but they likely spend more time foraging for insects (Stiles and Wolf 1979) and they may also assess possible foraging sites for the next morning (Feinsinger 1987). The high variability in return rates to patches of flowers is l ikely due to the fact that these are lekking birds, and activity on the lek may affect when a bird leaves for 24 A Yield Study of Traplining a foraging bout. Nesting activity w i l l have similar effects. As wel l , G i l l (1988a) observed that patches closer to the lek were visited more frequently than those farther away. Central place foraging theory predicts that hummingbirds w i l l profitably take smaller nectar loads from close sources than from far sources (Tamm 1989). Very short return times may indicate that the bird remained at or near the patch instead of moving on or returning to the lek because it could not consume all the nectar available in one visit. Stiles and Wolf (1979) observed that individual P. superciliosus perched near large (undefended) patches of H . pogonantha for 15-30 min after initial feeding, and then returned to the patch to feed again. It is unclear from this study whether variable return rates indicate variable amounts of time spent between foraging bouts, or variable numbers of patches visited on a bout (so that traplines wou ld be longer or shorter depending on nectar production and competition). The only way to determine which is the case wou ld be to follow birds on their routes or place observers at flowers along the routes, and see how many flowers birds visited on each bout. Compet i t ion It has often been assumed that trapliners, like territorial nectarivores, have almost exclusive use of the flowers along their routes (Feinsinger & Chapl in 1975; Feinsinger 1976; G i l l & Wolf 1977; Stiles & Wolf 1979). However, it is clear from Gil l ' s (1988a) and my studies that a traplining bird does not have complete control over the flowers it uses. G i l l found that visits by lek males to feeding sites far from the lek constituted 17-50% of all visits recorded at these sites. While I was unable to distinguish lek males from females or non-lek males, my results showed similar division of the resources. Patch F l had one primary visitor (BYB) and several others who came less often but still regularly. In Gil l ' s study the level of competition at flowers was rather low. He found that 16% of visits by 2 lek males visiting different 25 A Field, Study of Traplining patches were preceded by competitive visits, while I found that almost 50% of BYB's visits to patch F l were. It seems likely that levels of competition w i l l vary greatly wi th patch size and location, as well as with number of flowering patches in the environment as a whole. If competition decreases the total amount of nectar a bird can take from a patch, it is expected that P. superciliosus wi l l detect and respond to competitive visits by other birds. One variable that birds can control is how long they spend between visits to a patch. Would returning sooner decrease the competition a b i rd faces at a patch? I found that the longer birds waited to return to Patch F l , the more competitive visits from other birds they faced, indicating that they could "beat the competition" by returning sooner to the patch. Surprisingly individuals d id not clearly respond to the level of competition at natural flowers by returning sooner after greater number of competitive visits by other birds. Constraints on timing of visits and on timing of bouts such as activity the lek or nest may prevent birds from "fine tuning" their visitation rates enough to beat the competition. G i l l (1988a) found that P. superciliosus d id return more often to high reward feeders when he removed nectar to simulate the presence of a competitor. The differences in birds' responses to competition in Gil l ' s and my studies may be partly due to the difference in reward levels. The amounts of nectar in Gi l l ' s feeders were an order of magnitude higher than birds encountered at flowers in patch F l . In fact, one of his feeders could supply a bird with up to 50% of its daily energy requirements (King 1974, Gass & Montgomerie 1981), and emptying the feeder wou ld deprive the bird of a significant amount of food. This provided sufficient incentive for the birds to visit the feeder more often. As wel l , they most l ikely decreased visitation to other patches, making it easier for them to respond to the competit ion. It seems likely that time elapsed between visits to a patch by individual birds w i l l be bounded on one side by nectar renewal rates and by the probability of 26 A Yield Study of Traplining competition on the other. Returning often may discourage competitors but rewards may be too small to make it worth while. A l lowing large amounts of time to pass between visits may encourage competitors, which would also reduce nectar rewards. Trapliners w i l l likely return more frequently to patches wi th competitors than to those without, and their visitation rates w i l l probably rely on a combination of number of competitors and number of open flowers, as wel l as the distance of the patch from the lek/nest and activity on the lek/nest for reproductive birds. In the absence of competition, it is energetically more profitable for traplining birds to allow nectar resources to build up to higher levels by waiting longer before returning to each flower. The birds in Gill 's (1988a) study d id just that when they faced no competition at feeders. In natural flowers, this not only increases the nectar available per bout, but also frees time for other activities. Only one bird (BPB) visited patch L I , and return times for this bird (^85 minutes) were much longer than those of BYB, the primary user of the more frequently visited patch F l (~25 minutes). This difference in return rates to the two patches suggests that competition plays an important role in structuring bird visitation patterns to patches of flowers, although differences in the number of flowers (and thus total nectar availability) between the two patches makes this difficult to tease out. Comparisons of similarly-sized patches with different numbers of visitors should help clarify the role of competition in individual return rates. Conclusions P. superciliosus seem to follow flexible foraging routes that can change over time, most l ikely in response to nectar production and competition. The timing of visits to patches is not simply based on nectar production rates or level of competition alone, although the number of individuals visiting a patch decreases with decreasing nectar production during the day. Timing of visits may be a 27 A Field Study of Traplining compromise between activity at the lek/nest, nectar production rates, and level of competition. However, my conclusions are based on only two patches: more data and manipulative experiments are needed to determine whether birds follow the same route within and between days, and to further understand what factors influence traplining hummingbirds' behavior. N o one to date has studied how traplining birds structure their traplines spatially or how this spatial structure is affected by factors such as competition and decreasing or increasing nectar availability at individual patches. Controlled field and laboratory studies w i l l help clarify the roles of nectar production and competition in the timing of visits to patches by individual birds. 28 CHAPTER 3 AN ENCLOSURE STUDY OF TRAPLINING FORAGING BEHAVIOR IN PHAETHORNIS SUPERCILIOSUS A variety of nectarivore foraging strategies deal with the spacing of flowers in the environment, the quantity of nectar produced by each flower, and competition for the nectar from other nectarivores (Feinsinger 1976; Feinsinger & Co lwe l l 1978; Stiles 1981; Feinsinger et al. 1988). Nectarivores that visit undefended widely spaced flowers on repeated routes are called "trapliners" (Janzen 1971; Feinsinger and Chapl in 1975; Feinsinger 1976; Feinsinger 1987; G i l l 1988a). Traplining is economical only when few flowers are visited, the probability of visits to these flowers by competitors is low (Gill & Wolf 1977), and the nectar rewards are rich enough to compensate for energy spent travelling between flowers and still provide energy to be set aside for overnight survival (Hainsworth 1981). Traplines, or foraging circuits, enable animals to harvest regularly renewing food such as nectar from isolated sites in a regular fashion, and may deter competition by keeping resources at a low level (Schoener 1971; Wol f et al. 1975; Paton & Carpenter 1984; G i l l 1988a, Possingham 1989). Traplining has been described in many studies of nectarivorous animals (Janzen 1971, Feinsinger 1976 & 1987, Feinsinger & Chaplin 1975, Feinsinger & Colwel l 1978, Stiles 1975, Stiles & Wolf 1979, G i l l 1988a, Thomson 1988, Thomson et al. 1982, 1987,1989), yet few experimental studies have been conducted to determine what factors affect traplining nectarivores' behavior (Gill 1988a, Thomson et al. 1989; Tiebout 1993). 29 An Enclosure Study of Traplining For instance, how do they respond to changes in the profitability of individual patches of flowers along their routes due to competition, number of flowers open in the patch, or plant quality? Previous studies have suggested that nectarivores respond to changes in profitability of patches of flowers by adjusting visitation rate to increase reward size (Wolf & Hainsworth 1983; Gass & Sutherland 1985; G i l l 1988a; Thomson et al. 1989) Do traplining birds adjust visitation rates in response to changes in individual patches along their traplines, and if so how do they accomplish this? They could change the trapline spatially, temporally, or both. This paper describes an enclosure study on the foraging behavior of the Long-tailed hermit hummingbird (Phaethornis superciliosus). Hermits are found in the lowland tropical rain forests throughout Central America (Slud 1964) and are considered to be model trapliners (Gill 1988a). They are large hummingbirds (~6 g) that utilize highly productive flowers such as Heliconia pogonantha in light gaps and along forest edges (Stiles & Wolf 1979). This study considers how P. superciliosus adjusts its rate of visiting feeders with respect to changes in nectar production rates and to simulated competition. Changes in nectar production Nectar production rates of many flowers utilized by traplining hummingbirds (including H . pogonantha) are high in the early morning and decrease rapidly during the day. To maximize net energy intake P. superciliosus should pattern their visits wi th respect to declining nectar production throughout the day (Stiles 1975; Feinsinger 1976; Frost & Frost 1980; Stiles & Wolf 1979). Individuals in the enclosure should visit feeders less often as the day progresses to allow for nectar accumulation. In many cases, hummingbirds w i l l increase foraging effort when 30 An Enclosure Study of Traplining standing crop is low (Gass 1978a, b; Gass & Montgomerie 1981; Sutherland et al. 1982; Gass & Sutherland 1985). However, because P. superciliosus has evolved wi th a food-flower that produces little nectar in the afternoon and because they cannot defend nectar sources from competitors, I hypothesize that they take nectar when it is available rather than attempting to maintain constant net intake (Hainsworth et al. 1981). This means that they should take more nectar in the morning and less in the afternoon, as is seen in the laboratory under ad l ibitum conditions (Chapter 4). In order to maximize profit, P. superciliosus should respond to an increase or decrease in nectar available at individual feeders by visiting the changed feeders more or less often than unchanged feeders, as do traplining bees and territorial hummingbirds (Thomson 1982, 1987; Gass & Sutherland 1985). They may also decrease/increase use of unchanged feeders if changing feeders greatly enhances or reduces total energy intake (Gass & Sutherland 1985). G i l l (1988a) suggested that satiation might have affected the foraging behavior of P. superciliosus after feeding at feeders with large amounts of nectar, causing them to wait longer between visits than they would have at feeders wi th smaller amounts of nectar. Total nectar availability in hummingbirds' traplines or habitats should affect how many patches they visit and thus how far they travel in order to meet their energetic needs. Doubling the amount of nectar available to birds in the enclosure should allow them to lower their total visitation rate to feeders, and to stop using some of the feeders in the enclosure. Compet i t ion Trapliners do not defend their flowers aggressively as territorial nectarivores do. One way for trapliners to minimize the effects of competition is to utilize the 31 An Enclosure Study of Traplining strategy called defense by depletion, or exploitative defense (Davies & Houston 1981; Waser 1981; Paton & Carpenter 1984; Feinsinger 1987; Possingham 1989). Defense by depletion requires that birds respond to empty flowers at a patch that they previously found profitable by returning more often to that patch than before (Gil l 1988a). Thus the hypothesis predicts that competition decreases time elapsed between visits to patches of flowers along the trapline. Since nectar production rates should also influence return rates (Frost & Frost 1980; G i l l 1988a), visit frequencies should be a compromise between maximizing nectar accumulation and minimiz ing loss to other individuals through exploitative competition (Gil l and Wol f 1977). If birds can detect simulated competition in the enclosure, they should respond by visit ing the feeder wi th competition more frequently than unchanged feeders, in order to keep nectar levels low and discourage competitors from returning (Gil l 1988a). METHODS P. superciliosus (of unknown sex) were captured with mist nets located either near a lek or at a patch of Heliconia pogonantha at La Selva Research Station in the Sarapiqui region of Costa Rica in May-August 1994. Birds were marked wi th a spot of white paint on the back, which usually remains until the feathers are molted (Stiles & Wolf 1979). The paint markings were used to identify recaptured birds since I desired to use only naive birds in my experiments. After marking, the bird was placed in a small (60 X 60 X 60 cm) cage where it was trained for 1 to 2 day(s) to use a feeder. I ran my study in 2 enclosures, each housing 1 bird. Enclosures were 3.5 x 7.2 x 2.6 m (Fig. 5) and were made of shade cloth which reduced light levels to those of 32 An Enclosure Study of Traplining forest wi th light canopy cover. Ambient temperatures ranged from 22-33° C. Each enclosure had 7 feeders and 1 perch (Fig. 5). Since visibility is relatively poor in the rain forest, feeders were hidden from view from each other by cloth blinds hanging from the ceiling; I hoped that birds would not be led to the next feeder simply by seeing it. Birds were introduced to the enclosure after being trained to use the feeder, and were given one day to get used to the set up. Individuals were kept in the enclosure for 4-7 days. Each feeder in the enclosure had a photocell which was activated by the bird's b i l l when it visited. Times of arrival and departure from all feeders were recorded automatically by a custom built data logger. From this I could calculate time and sequence of consecutive feeder visits and return time to each feeder. I had intended to monitor the perch as well , but that part of my system malfunctioned and I do not report those data. Because perch monitoring failed, I could not identify "bouts" of foraging except by long gaps with no visits to feeders. The data logger ran solenoid valves (General Valve Corporation, Series 3) which delivered sugar water of concentration similar to H . pogonantha nectar to the feeders (0.98 M for enclosure, 0.92 M for H . pogonantha; Stiles & Freeman 1993). Hummingbirds of 6 g require at least 36 kj of energy intake a day to survive (King 1974) and captive P. superciliosus consumed 39.7-58.2 k j /day from an unlimited feeder in the enclosure (Chapter 4). The total energy available to birds in the enclosure ranged from 47 to 101 k j / day, depending on the treatment (51 kj were available on Control Days). Solenoid valves delivered sugar water in discrete quantities: 1 pi per "squirt". Interval between squirts was determined by desired output: for a desired output of 60 u l / h , the solenoid valve wou ld be activated once every minute. Thus if a bird emptied a feeder and returned again before the next 33 An Enclosure Study of Traplining squirt, it wou ld receive nothing. A t each feeder, the solenoid valves delivered sugar water at rates comparable to about 10 H . pogonantha flowers (Fig. 6). I had to make each feeder worth many flowers in order to meet the birds' energy needs because data logger memory and space in the enclosure limited the number of feeders I could use. Because each feeder contained so much nectar, they are equivalent to patches rather than individual flowers. Each morning, I put 200 ul of sugar water into the feeders to simulate the amount of nectar available in Heliconia flowers when they open at dawn. I started the pump program at 6 am and birds were then allowed access to the feeders. For the first one to three days, all feeders received the same amount of sugar-water at the same declining rate. Birds were allowed to forage at w i l l , wi th no interference on my part. I w i l l refer to these days as "Control". These days were used to determine how often and in what order birds used the feeders. The next few days were treatment days, in which the output of one or all of the feeders changed. Each bi rd experienced the control period and one of the following treatments: Increase/Decrease: To determine how an increase or decrease in nectar production at one feeder would affect P. superciliosus feeding behavior, I programmed the computer to double/halve the output rate of one feeder. A l l other feeders remained the same. I used 2 birds for Increase and 2 for Decrease treatments. Satiation: To determine how an increase in total nectar availability affected foraging behavior, I doubled the output rate of all feeders in the enclosure, so that birds received roughly twice the amount of nectar as in the control treatment. I used 4 birds in this treatment. 34 An Enclosure Study of Traplining Figure 5: Top view of enclosure. Enclosure is 3.5 x 7.2 x 2.6 m. Symbols are defined as follows: = blind, T = perch, O = feeder 35 An Enclosure Study of Traplining Figure 6: Estimated nectar production rates for feeders in enclosure (dashed line) and Hel iconia pogonantha flowers in the w i l d (solid line; from Stiles 1975. This estimate may be low if nectar removal by birds increases total nectar output by the flower; G i l l 1988b). Production rates are in ul /hr . 36 An Enclosure Study of Traplining Competi t ion: To test how competition affected foraging behavior, I entered the enclosure every 20 minutes from 0700-1200 h and removed any nectar found in one specific feeder wi th a glass microcapillary tube and recorded the amount of nectar removed. Total amount of nectar removed varied between birds (depending on their rate of visitation to the feeder), and ranged from 7 to 114 ul . The bi rd experienced competition at only one of its entire suite of feeders, and it was the same feeder every time. Nectar output was similar at all feeders in this treatment. I used 4 birds in this treatment. Analysis : To determine if nectar production rates influenced foraging, I analyzed visitation rates to feeders over time on Control days wi th a Kruskal-Wall is one way A N O V A . I used two way contingency tables to test the effects of the various treatments on the number of visits to a particular feeder relative to visits to all other feeders for whole days on the "Control" and "Treatment Days". In the Satiation treatment, I compared the total number of visits to a l l feeders for Control and Satiation days using a paired t-test. To determine if the birds' behavior affected how much nectar I was able to remove as a competitor, I used an A N O V A to compare the amount of nectar removed and the mean time elapsed between visits to that feeder. The volumes removed were log transformed before analysis. Us ing a Spearman rank correlation, I analyzed birds' responses to competition (nectar removal) by comparing the amount of nectar I removed from the feeder in one hour to the bird's relative use of that feeder (with respect to the other feeders). 37 An Enclosure Study of Traplining RESULTS Changes in Nectar Production The number of feeder visits per hour differed significantly at different times of day (Kruskal-Wallis A N O V A : K W = 50.145, df= 9, P < 0.001). Birds visited feeders less in early morning and late afternoon than mid-morning, with a peak at 0900 h (Fig. 7). Birds visited feeders with higher nectar production rates twice as often as they visited the same feeders on Control days (Bird 1: X 2 = 8.08, df= 1, P= 0.005; Bird 2: X 2 = 32.25, df= 1, P<0.001; Fig. 8). There was no difference in use of the feeder on the two control days before the increase treatment (P> 0.5 for each bird), which suggests that the change in the feeder caused the birds to increase use of it, rather than random fluctuations in feeder use. Birds visited the feeder only 75% as often when its output was decreased compared to Control days (Bird 1: X 2 = 10.04, df= 1, P= 0.002; Bird 2: X 2 = 13.75, df= 1, P<0.001; Fig. 8). Feeder use d id not change between Control days preceding the Decrease treatment (P> 0.5 for each bird). Birds visited feeders only 5% as much on Satiation days as on Control days (Paired T-test: t= 8.33, P= 0.004, df= 3; Fig. 9). The use of individual feeders changed dramatically between Control and Satiation days, with some feeders left almost completely unvisited during Satiation days. Compet i t ion Individual birds differed in their response to the competition study. One bird increased its relative use of the experimental feeder (Bird 1: X 2 = 6.5, df= 1, P= 0.01), 2 birds decreased their use of the feeder (Bird 2: X 2 = 4.64, df= 1, P= 0.03; Bird 3: X 2 = 38 An Enclosure Study of Traplining Figure 7: Visi ts /hour to feeders over time of day. Mean for 20 bird-days (Control only). Error bars indicate 1 S.E. 39 An Enclosure Study of Traplining cc m Q LU LU LL LL O UJ CO ID LU > LU DC 0.20 -i 0.15 H 0 . 1 0 H 0.05 H 0.00 CON - D D E C CON -1 INC Figure 8:. Relative use of feeder for Control, Decrease, and Increase days (mean for 2 birds for each treatment). C O N - D= Control prior to Decrease treatment, D E C = Decrease, C O N -1= Control prior to Increase treatment, INC= increase. Error bars= 1 S.E. 40 An Enclosure Study of Traplining 19.75, df= 1, P< 0.001) and the fourth d id not respond (Bird 4: X 2 = 0.013, df= 1, P= 0.91). The differences in response to the competition treatment can be accounted for by the amount of nectar that I, as the competitor, was able to remove from the feeder. The amount I removed depended on how long the bird had waited between visits to the feeder. Relative use of the feeder increased wi th the amount of nectar removed, wi th a time lag of one hour (Fig. 10; Spearman rank correlation: r s = 0.32, N= 45, a (two tailed) = 0.05). After birds increased visitation rates to the competition feeder, I could no longer remove any nectar (they effectively out-competed me) and they then reduced their visitation rates to original levels (Fig. 11). Thus, the fact that most birds did not respond (on average, over whole days) as expected due to competition can be explained by the fact that I was not a successful competitor. When I was successful in removing large amounts of nectar, birds responded as expected by increasing visitation rate to the feeder. 41 An Enclosure Study of Traplining to oc LU Q UJ UJ CO CO > _ l < H O 1200 1000 800 600 400 200 C O N S A T Figure 9: Total visits per day to all feeders for Control and Satiation days. C O N = Control days. SAT= Satiation days. Mean for 4 birds. Error bars = 1 S.E. 42 An Enclosure Study of Traplining CC LU Q LU LU O LU 00 LU > h-< _l LU DC 1 . 0 1 0.8 H 0.6 H 0.4 H 0.2 0.0 —r~ 20 40 0   60 80 100 TOTAL VOLUME REMOVED (ul) Figure 10: Relative use of feeder (with 1 hour time lag) vs total fil removed from feeder. Relative uses of feeders are taken from 1 hour later than the amount of nectar removed to account for the time lag in responses to the removal of nectar. 43 An Enclosure Study of Traplining r 100 DC UJ Q LU UJ U_ LU CO z> LU > LU DC Q LU > o LU DC UJ O > _1 o TIME OF DAY (h) Figure 11: Relative use of feeder on Competition day (solid line), and total amount of nectar (in pi) that I was able to remove from feeder (dashed line). This graph shows the response of a bird to the removal of a large quantity of nectar. 44 An Enclosure Study of Traplining DISCUSSION Changes in Nectar Production In the enclosure birds visited fewer feeders in the afternoon when nectar production rates were low (Fig. 7). The low visitation rate in the early morning probably reflects satiation; nectar production rates were high and visiting only a few feeders early in the morning could easily fil l a bird's crop; at 0600 h the feeders contained 200 ul , and visiting 3-4 feeders would fil l the birds' ~780 ul crops (Hainsworth & Wolf, 1972). Stiles and Wolf (1979) observed fewer P. superciliosus visits to H . pogonantha patches in the afternoon than in the morning and noted that total activity levels including activity at the lek decreased in the afternoon. This suggests birds are conserving energy because nectar production levels are quite low at this time of day (Stiles 1975; Stiles & Wolf 1979). P. superciliosus responded to changes in nectar production at individual feeders by visiting them more or less often, depending on the direction of the change (Fig. 8). Traplining nectarivores in the w i ld drop patches that become unprofitable, and incorporate more profitable ones into their traplines (Thomson et al. 1982, 1987). Both bumble-bees and territorial rufous hummingbirds (Selasphorus rufus) increase visitation rates to more rewarding plants (Thomson et al. 1989; Gass & Sutherland 1985) as P. superciliosus did in this study. Unfortunately, I was not able to determine how changes in nectar production rates affect the entire trapline (such as changes in routes taken). A bird could rearrange the trapline so that one patch is visited more/less often than others, or could visit different subsets of its total trapline on different bouts. This problem merits further investigation. If traplining birds are satiated after taking nectar from just a few natural 45 An Enclosure Study of Traplining flowers, this could strongly affect the entire trapline In the enclosures, P. superciliosus visited feeders less often on Satiation days than on Control days (Fig. 9). For territorial hummingbirds, both territory area and foraging time varies inversely wi th flower density and nectar availability (Gass 1979; H i x o n et al. 1983; Armstrong 1992). S. rufus adjust territory size on a daily basis to maintain or improve territory quality (Gass 1979). To keep energy expenditures down, traplining birds should drop feeders or flowers from their trapline and/or decrease visitation frequency if they no longer require the nectar they produce. The size and shape of a trapline w i l l depend on nectar production rates in the environment as a whole, as wel l as on how much nectar each patch provides. Thus plants can affect how many potential pollinators (and potential mates) they have by the amount of nectar they produce (Stiles 1981). If there is a shortage of plants wi th respect to pollinators, increasing nectar rewards w i l l increase number of birds visiting. If there is a shortage of pollinators (i.e. if the number of visitors remains similar), the number of patches visited by each bird would decrease with increasing nectar production, lowering pollination potential for plants. Compet i t ion In the enclosure, birds that waited longer before returning to the feeder lost more nectar to the "competitor". In the wi ld , the number of competitive visits by other birds is a direct function of how long a bird waits before returning to the patch of flowers (Chapter 2). Therefore, birds returning more often than based on nectar production rates alone would decrease the number of competitive visits and therefore decrease the nectar depletion they face at a patch. 46 An Enclosure Study of Traplining M y enclosure study indicates birds' responses to competition w i l l depend on how successful the competitors are (i.e. how much nectar they remove). Birds that faced high levels of competition in the enclosure returned more often to that feeder than to the unchanged feeders (Fig. 10), and P. superciliosus in the w i l d respond similarly to competition at high reward feeders (Gill 1988a). However, at natural patches of H . pogonantha, birds did not respond to the number of competitive visits between their consecutive visits (Chapter 2). Wi th the relatively low nectar production rates of flowers in the w i ld , there may not be a close relationship between the number of competitive visitors and the amount of nectar removed. One competitor could remove as much nectar as several, depending on how far apart the competitors spaced their visits. It is more likely that birds were reacting to competition at a patch on a presence or absence basis. Comparisons of visitation rates by birds wi th competition at patches and those without indicate that birds with competitors returned sooner than those without them (Chapter 2). However , sample sizes were too small in Chapter 2 to draw any firm conclusions about responses to competition. Further Studies Phaethornis superciliosus exhibit much of the behavior predicted for nectarivores by Janzen (1971), Stiles (1975), Feinsinger (1976), and G i l l (1988a). P. superciliosus increased/decreased the relative use of individual patches when their profitability increased/decreased for whole days (Figs. 9 & 10). O n a shorter time scale, birds responded to sudden large changes in nectar availability at a feeder (due to simulated competition) by increasing their visitation rate (Figs. 12 & 13). If and how they track variation in individual patches of flowers along their traplines 47 An Enclosure Study of Traplining remains unknown. What level of change in profitability is necessary to provoke changes in foraging, and how does this affect the rest of the trapline? Since patches are widely spaced, it is costly for the bird to revisit a patch often in order to beat competitors. While birds respond appropriately to changes at a single site, it is unknown how they would respond to changes at multiple sites. Management of multiple sites requires that the bird respond to multiple changes accordingly, while keeping expenditures low (which includes flying the shortest possible distance between patches). It is unknown whether P. superciliosus are truly able to respond to changes in profitability of all the sites on their traplines. I suspect that changes in nectar availability at patches would have to be relatively large for these birds to respond by altering their traplines. 48 CHAPTER 4 ENERGY REGULATION IN TRAPLINING BIRDS Hummingbirds are good models for energy regulation because their small body size and high metabolic rates require that they maintain positive energy balances on relatively short time scales (Gass & Roberts 1992). Hainsworth (1978, 1981) considered the relationship between short term regulation of feeding behavior and longer term regulation of body mass and internal energy reserves. He proposed an energy accumulation-depletion model of short term food consumption in which feeding behavior is regulated to maintain constant rate of net energy intake throughout the day. This rate is set once at the beginning of each day based on energy reserves at that time relative to some standard. Individuals who fail to reach this goal the previous day or experience higher than expected energy expenditures during the night would have lower than expected reserves in the morning, set a higher required net intake rate, and accumulate energy faster during the day. In principle this model accounts for both short term (hour to hour) and long term (day to day) regulation of energy reserves, and allows individuals to respond to seasonal contingencies such as high rates of fat gain during migratory stop-overs (Carpenter et al. 1983). Hainsworth's model is supported by the observation that hummingbird energy accumulation rates tend to be constant over a day, and also by three experimental tests. After Hainsworth and Wolf (1983) deprived Eugenes fulgens and Lampornis clemenciae of food for several hours each day for several days, the birds accumulated energy at a faster rate than before the perturbation. Similarly, after nocturnal temperatures were experimentally decreased, imposing an energy deficit at dawn, diurnal energy accumulation rates increased (Hainsworth 1978). Tooze and 49 Energy Regulation in Traplining Birds Gass (1985) tested the assumption of constant energy accumulation during the day by imposing fasts. As predicted by the model, intake rates after the fast were no higher than before, resulting in energy deficits at nightfall. Most individuals recovered energy balance by dawn through nocturnal torpor; those who had not completely recovered accumulated energy at a higher rate on the next day. The accumulation-depletion model with its assumption of constant required energy accumulation predicts constant food intake rate under constant temperature and ad l ibitum food availability, and hummingbirds tend to do this (Gass 1978a and unpublished observations; Wolf & Hainsworth 1977; Hainsworth 1978; Hainsworth and Wolf 1979; Gass & Montgomerie 1981; Tooze & Gass 1985). These tests include 5 hummingbird species, all territorial species with high wing disc loading (Feinsinger & Chapl in 1975; Feinsinger & Colwel l 1978). Does the model apply to all hummingbirds? Here I show that a hermit hummingbird (Phaethornis superciliosus) does not feed at a constant rate, either in the field or under ad libitum laboratory conditions. In the field, where its natural food flowers (Heliconia pogonantha) produce nectar at a rate that decreases to near zero by late afternoon, P. superciliosus feeds more in the morning than afternoon (Stiles & Wol f 1979). Its total intake was also higher in the morning than in the afternoon under laboratory ad l ibi tum conditions. I show that although net energy accumulation rates are unlikely to be constant in either of these situations, a short term accumulation-depletion model can still account for observed patterns of feeding behavior and predicts changes in other variables such as body mass. METHODS Phaethornis superciliosus is a large (~6 g) species of traplining hummingbird that is common in lowland secondary forests of Central America. Its foraging behavior has been studied in the field by Stiles & Wolf (1979) and G i l l (1988a) but little work has been done with its energy requirements and the patterning of its 50 Energy Regulation in Traplining Birds nectar consumption. As part of a larger study, I measured temporal patterning of nectar consumption by captive wild-caught P. superciliosus throughout the day from a commercial hummingbird feeder (Perky Pet #214) wi th a modified corolla (4 cm of 8 m m diameter tygon tubing extending from the opening of the feeder). Five P. superciliosus were kept (one at at time) in a 3.5 x 7.2 x 2.6 m enclosure at La Selva research station, in the Sarapiqui Atlantic lowland tropical region of Costa Rica, between June 28 and August 2 1993. Because they were held in captivity at different times, they experienced different weather patterns during the experiments. Each bird was kept for 4 to 6 days, depending on the experiments it participated in. Nectar consumption data were taken on either the second or third day in the shade-house for each bird. On these days, birds were given unrestricted access from 0600-1800 h to a feeder containing 1.06 M sugar solution; similar to the nectar concentrations (0.92 M ) of Heliconia pogonantha, a main nectar source for P. superciliosus (Stiles 1975; Stiles & Freeman 1993). Feeders were weighed every hour to determine how much solution (± 0.1 g) had been consumed. O n occasion feeders wou ld drip, but I estimated that drips accounted for only 5% of daily nectar removal from the feeders (by counting drips per observation period, estimating volume of drips, and extrapolating to entire days). For 3 hours, between 0700-0800 h, 1100-1200 h, and 1500-1600 h, I recorded from behind a bl ind the time, duration, and number of probes during visits to the feeders. From this I could calculate time to next visit, and total number of probes and visits during the observation period, The effects of time of day, bird identity, and number of probes on nectar consumed and total number of visits during an observation period were analyzed using A N C O V A s . RESULTS AND DISCUSSION Under ad l ibi tum conditions, gross nectar consumption rate decreased significantly over the day (Spearman rank correlation of mean (for all birds) nectar consumed and time of day: r s= -0.74). Birds consumed the most solution in the first 51 Energy Regulation in Traplining Birds 3 hours, during which the consumption rate dropped rapidly. Nectar consumption rates were steady between 0900 -1600 h, and dropped off rapidly towards zero between 1600-1800 h (Fig. 12). Visitation frequency also declined in the same pattern, suggesting that intake per visit was fairly constant (although differences were not significant: r 2= 0.238 ,N= 15, F u = 1.878, P= 0.195; Fig. 13). There is a positive relationship between kj consumed and total visits in an hour ( r2= 0.262, N= 15, Fi= 4.62, P= 0.05), indicating that the birds are decreasing visitation rate to the feeder as the day progresses and that this decrease accounts for decreased nectar consumption. C o u l d birds wi th this pattern of intake meet the required storage rate through the day hypothesized by Hainsworth (1978)? I explored this by estimating expenditure rates for a wide range of hypothesized activity budgets: from 0 to 100% of total flying time and from 0 to 100% of flying time spent hovering. There are no published activity budgets for hermits, but various investigators have suggested that they fly more than territorial hummingbirds (Feinsinger & Chapl in 1975; Feinsinger 1976, 1987; G i l l 1985, 1988a). Territorial hummingbirds fly 10 to 40% of the time, depending on the quality of their territories (Sutherland et al. 1982). One estimate of % flying time spent hovering comes from my observation that P. superciliosus spent a mean of 15.4 s (and up to 1 min.) hovering to visit several flowers in a patch of H . pogonantha. A t 50% of flight time spent hovering, this wou ld mean 15.4 s flight time between patches, which at 11 m/ s (Gill 1985) wou ld carry them 169 m (or 660 m for 1 min.). This distance is probably farther than many inter-patch distances in nature (Stiles 1975), so I conclude that P. superciliosus spend > 50% of their flight time hovering. I calculated net intake rates in the enclosure by subtracting estimated expenditure rates from observed gross intake. I assumed a constant min imum of 2.4% feeding time, all hovering —the mean for 5 birds at the ad l ibi tum feeder. In a laboratory situation, Chlorostilbon canivetii (a low-reward trapliner) decreased feeding effort and flying time in the afternoon (Tiebout 1991, 1992), and in this 52 Energy Regulation in Traplining Birds 7 l 1 i o H—•—i— 1—i—•—i—•—i—•—-i—•—i 6 8 10 12 14 16 18 TIME OF DAY (h) Figure 12: Gross Energy Intake (kj) vs Time of Day. Error bars = 1 S.E. Spearman rank correlation of mean (for all birds) nectar consumed and time of day: r s= -0.74 53 Energy Regulation in Traplining Birds 14 -i 12 W 10 Q w UH 8 o E-i CD H £ 41 6 i 2H 11 15 TIME OF D A Y (h) Figure 13: Mean visits per hour (for all birds) to feeder. Error bars= ± 1 SE. 54 Energy Regulation in Traplining Birds model I assumed that non-foraging flying activity began at a set level and changed in proportion to food consumption (Fig. 12): as nectar consumption decreased throughout the day so d id non-foraging flight and total expenditures. I assumed forward flight was at Vmr , although G i l l (1985) observed that P. superciliosus often flies faster than Vmr . Al though there is evidence that the weight of a meal increases expenditures (DeBenedictis et al. 1978; Tamm 1989), I assume no such dependency. In all cases, net intake rate decreased throughout the day (Fig. 14). For pure forward flight (no extra hovering over the 2.4% required for feeding) net intake was higher than required even for a starting value of 100% of time spent flying. For 50% of flight time spent hovering, net intake was higher than required for all activity levels but 100% of time spent flying, for which net intake fell slightly below required in the late afternoon. For the most expensive mode (100% of flight time spent hovering), net intake was higher than required to survive the night for all but the two highest activity levels (75 and 100% of time spent flying; Fig. 14). In all of the above analyses, total flight costs and therefore total energy expenditure decreased throughout the day in proportion to decreasing nectar consumption. Cou ld birds meet energy requirements if their expenditures d id not decrease throughout the day, but instead remained constant at 0600 h values? I modeled total net intake for an entire day over a range of % of time spent flying and % of flying time spent hovering assuming constant activity budget over the day (Fig. 15), using observed gross intake and estimated perching and flight costs. Unl ike the above model (Fig. 14), this one assumes no min imum % of time spent hovering. M o d e l birds maintained positive diurnal budgets for all combinations except the highest activity levels and the highest proportions of flying time spent hovering. Of course, a neutral energy budget for the day leaves nothing for the night. Birds met nocturnal requirements below 32% flying time even if all their f lying was hovering. 55 Energy Regulation in Traplining Birds TIME OF DAY (h) TIME OF DAY (h) TIME OF DAY (h) Figure 14: The effect of % time flying and % of flying time hovering on net intake of energy for P. superciliosus at an ad libitum feeder, a) Observed gross intake and assumed energy expenditures ( i n kj) for 5 levels of % time flying, b) net intake (gross intake - assumed expenditures) in kj, c) cumulative net intake in kj. % time flying is set at 0600 to one of five levels and then decreases throughout the day in proportion to nectar consumption. H= min imum net intake for bird to survive the night (Hainsworth 1978). 56 Energy Regulation in Traplining Birds % TIME F L Y I N G Figure 15: The effect of % time flying and % of flying time spent hovering on energy accumulation for a 12 hour day. Contours connect points of equal energy accumulation (in kj). 0 energy accumulation means that bird exactly matches expenditures with intakes, but has no reserves for the night. The line labelled 14 represents Hainsworth's (1978) estimated energy accumulation required for the bird to survive the night and achieve a 24 hour neutral energy budget. 57 Energy Regulation in Traplining Birds They also met nocturnal requirements for 100% flying time if they hovered for less than 5% of flight time and they barely met requirements under the more realistic assumption of 50% flying time and 50% flying time spent hovering. However, in the w i l d , birds may take in more energy than they d id in the laboratory, making it easier to meet needs with higher flying times. Assuming that the rate of storage of excess energy is flexible and can keep up wi th the high net intake rates in the morning, these birds should be able to maintain a positive 24 hour energy balance. They would not be able to meet energy demands wi th a constant storage rate for two reasons. First, it is not possible to maintain a constant energy storage rate with no net intake in the afternoon. A n d if storage rates were lower than net intake rates in the morning, the birds wou ld not be able to use all the food they ate then. This implies that Hainsworth's assumption of constant net intake does not apply to traplining birds. The strategy of harvesting and storing food as it is available may be the only option for traplining hermits because they do not defend resources aggressively as territorialists do. They can respond to competition only exploitatively (i.e. if they don't eat it some other bird w i l l ; G i l l 1988a). This strategy implies a diurnal pattern of changes in body weight very different than what has been observed for territorial hummingbirds with high wing disc loading, which tend to gain weight at a constant rate during the day (Wolf & Hainsworth 1977; Carpenter et al. 1983; Gass & Tooze 1985; but see Calder 1990). The strategy of harvesting food as it is available predicts that hermits w i l l gain weight fast in the morning and lose some of it i n the afternoon. Because both flight and perching costs depend on body mass, early morning weight gain should be uneconomical (Wolf & Hainsworth 1977; Hainsworth et al. 1981). DeBenedictis et al. (1978) assumed that the mass of a meal added enough flight and perching costs to make large meals uneconomical. Tamm (1989) made the similar assumption that the cost of carrying water made low 58 Energy Regulation in Traplining Birds concentration food less economical than high concentration food at long distances from a central place. In both cases the hummingbirds behaved as predicted. Tiebout (1991) noted that trapliners consume more energy per g body weight than territorialists do, and suggested that trapliners' low wing disc loading may help compensate for the costs of carrying the larger meals. This study extends that idea to the costs of carrying the fat gained early in the day. Early weight gain may be inefficient, but it may also be the only choice available to trapliners. The energetic consequences of this foraging strategy deserves closer scrutiny. To determine whether a traplining hummingbird wi th a variable net intake rate could do 'better' than one with a constant net intake rate in a system wi th decreasing nectar production rates, I parameterized for my system a numerical simulation model developed earlier for territorial hummingbirds (Armstrong et al. 1987; C L . Gass, unpublished). The model incorporates measured or assumed values for many components of the system, and includes sub-models for metabolism, feeding behavior, and environmental conditions (Appendix I). Environmental conditions include altitude (air density), maximum and m i n i m u m air temperature and pattern of change, number and spacing of flowers, nectar concentration, rate and pattern of change in nectar production, and day length. The metabolism sub-model converts activities into energy expenditures using allometric cost equations reviewed by Montgomerie (1979), given assumed wing morphology and current body mass (which changes during the day). Based on Stiles' (1975) field measurements, flowers begin the day with some amount of nectar and produce nectar at an exponentially decreasing rate during the day (Appendix II). After Wolf & Hainsworth (1977) and Hainsworth (1978, 1981), this is an accumulation-depletion model. Birds fly on foraging bouts, visit flowers in some sequence and alternate between flying and hovering metabolic rates, then perch, get hungry at a rate determined by their fat storage and perching metabolic rates, and forage again (the model includes no direct or indirect interactions wi th competitors). 59 Energy Regulation in Traplining Birds The model is also event-based, not time based. The duration of any perching event is computed from net intake from foraging, required fat storage rate, air temperature for the time of day, and current body mass. Similarly, transit times between flowers are computed from assumed distances and velocities. The model computes the nectar volume accumulated since the last visit to any flower, transfers the nectar to the bird, and resets the flower to zero. Handl ing time at flowers is a linear function of nectar volume (Gass and Montgomerie 1981) and is based on Wolf et al.'s (1972) estimates for P. superciliosus at H . tortuosa. Transit time between flowers is assumed to be at 11 m / s (Gil l 1985) and to cost an amount calculated for V m r at current body mass, from Montgomerie (1979). After a foraging bout, a portion of net energy intake from the bout sufficient to meet the required accumulation rate for one cycle of activity (the foraging bout and the current bout of perching) is converted to fat, then the remainder is depleted at the perching metabolic rate and the next cycle of activity begins. A n important assumption of this model at its present stage of development is that fat storage is instantaneous, i.e. that crop clearance, digestion, and absorption take no time. This model, when parameterized for rufous hummingbirds (Selasphorus rufus), generates realistic time and energy budgets for entire days under a variety of conditions. M y primary objective here was to compare Hainsworth's (1978, 1981) assumed constant required net intake rate with an alternative in which P. superciliosus accumulated net energy (fat) at a decreasing rate during the day (Figs. 12 and 14). In both cases I assumed that required total energy accumulation for the day is the product of nocturnal metabolic rate and duration of the night (Hainsworth 1978); i.e. I assumed a neutral 24-h energy budget. In the former case, required net intake rate was simply the total required intake divided by day length. The latter case was more complex. I assumed that required net accumulation declines exponentially between dawn and dusk from some starting value such that the area under the accumulation rate curve is the total required accumulation for the day (Appendix III). In practice, I 60 Energy Regulation in Traplining Birds modeled constant net intake rate by setting the rate of change of accumulation rate (m in Appendix III) to zero. I refer to birds with exponentially decreasing net intake rates as "Decrease birds" and birds with constant net intake rates as "Constant birds". Decrease birds gain more of the required fat for overnight survival than Constant birds at intermediate size traplines of 25-40 flowers: they lose more weight than Constant birds with fewer flowers (< 25) and are not different from Constant birds at very high numbers of flowers (> 40; Fig. 16). Decrease birds do worse than constant birds at low numbers of flowers because they have high intake requirements in the morning, and there are not enough flowers to support the required net intake. Because Constant birds are restrained in their intake, they do not drain the flowers first thing in the morning, postponing the time when their flowers are depleted. Once standing crops are reduced to near zero, the required foraging effort increases to 100%, although even this effort is insufficient to make Hainsworth's (1978) net intake requirements (Fig. 17). Both Constant birds and Decrease birds do the same at very high numbers of flowers because at any given time of day there are enough full flowers in the trapline to exceed intake requirements for either model. Mea l size can range from zero to crop capacity, which increases allometrically wi th body mass (Hainsworth & Wolf 1972). Crop volume for P. superciliosus estimated from this relationship is 780 pi. Most territorial hummingbirds take meals of roughly half crop capacity, but any hummingbird whose transit time is more than 15 s on a bout should fill its crop (DeBenedictis et al. 1978). I explored a range of meal sizes from 100 to 780 pi. Smaller meal sizes increase the total number of flowers that birds require to meet energy needs (because, energy expenditure increases wi th visitation rates). For any meal size, Decrease birds satisfy energy needs (Hainsworth 1978) with fewer flowers than Constant birds (Fig. 16). This is because net intake rates of Decrease birds follow nectar production rates, and in 61 Energy Regulation in Traplining Birds CONSTANT DECREASE 22 24 26 28 30 32 34 36 38 22 24 26 28 30 32 34 36 38 40 NUMBER OF FLOWERS Figure 16: Whole day values for Constant and Decrease birds over different numbers of flowers, a) Number of flower visits per day/20000 (solid); Number of bouts per day/500 (large dash), and % time spent flying (small dash), b) Energy taken in (solid), Energy expended (large dash), and Net energy intake (small dash), all in kj. c) Change in mass (starting mass - mass at end of day) in g. 62 Energy Regulation in Traplining Birds CONSTANT DECREASE 0.8 - 0 . 5 - 1.0 d) 0.8 TIME OF DAY (h) Figure 17: Values for Constant and Decrease birds for traplines with 30 flowers, a) Number of flowers visited per bout/50 (large dash), Mean volume per flower in u l / 40 (solid), and % time spent flying (small dash), b) Duration (s) of perching/ 120 min (large dash), foraging bout/20 min (small dash) and whole cycle (perching + flying)/120 min (solid), c) Energy taken in/3.6 (solid), Energy expended/3.6 (large dash), and Net energy intake/3.6 (small dash), all in kj. d) Relative mass (Current Mass/Starting Mass - 0.5) in g (small dash; A bird that stays the same weight all day would generate a horizontal line at O.5.), Cumulative fat gain/20 in kj (large dash) and Fat accumulation rate in watts (solid). 63 Energy Regulation in Traplining Birds contrast to Constant birds, they do not need to struggle to meet energy needs in the afternoon that their traplines cannot provide (Fig. 17). O n the contrary, their foraging activity is lowest while nectar standing crop is lowest. Constant birds, on the other hand, do not take full advantage of high nectar production rates in the morning and must work harder to make required intake rates in the afternoon (Fig. 17). Vis i t ing flowers more when nectar production rates are higher provides Decrease birds wi th greater gross nectar rewards from the same number of flowers, because emptying flowers stimulates nectar production (Appendix II): if birds drain flowers early in the morning while production rates are high, these flowers w i l l produce more nectar than flowers that are first visited later in the day. This relationship between increased nectar production and visitation rate is seen in the w i l d in H . imbricata, and probably holds true for many flowers visited by traplining birds (Gil l 1988b). Because Constant birds do not visit enough in the morning to stimulate much extra nectar production (even on traplines that w i l l not satisfy their nocturnal energy requirements), the flowers in their traplines have lower standing crops in the afternoon than decrease birds (Fig. 17). The model also shows that in the absence of competition there is no advantage for traplining birds to visit more flowers than necessary to meet overnight requirements (Fig. 16). A d d i n g competition to the system would increase the number of flowers a bird must visit to meet requirements, and wou ld increase the importance of having intake rates that parallel nectar production rather than being constant. Competition affects visitation rates to flowers (Gil l 1988a, Chapter 3); birds faced by competition may have to forage more frequently than based on nectar replacement rates alone, to prevent excessive loss of nectar to competitors. 64 Energy Regulation in Traplining Birds Further Studies M y models indicate that model birds with decreasing net intake rates can meet energy requirements necessary to survive the night, and in most cases are more successful than model birds with constant net intake at ut i l izing flowers wi th decreasing nectar production rates over the day. More detailed energetic studies of P. superciliosus should be done in order to understand how traplining hummingbirds regulate energy over the day. For instance, it should be determined whether traplining birds change their fat storage rate during the day and if limits on rate of digestion affect hummingbirds' ability to consume large quantities of nectar in the morning. As well , it would be useful to know how traplining birds cope with decreased net energy intake in the afternoon. Tiebout (1991,1992, 1993) suggests hummingbirds can reduce perching metabolic rates to help compensate for low food availability. Do traplining birds drastically decrease energy expenditures as nectar production decreases in the afternoon on a daily basis in the wild? Comparative studies wi th other species could determine if trapliners are the only hummingbirds that utilize flowers with decreasing nectar production which decrease net intake over the day. Non-nectarivorous birds such as sparrows have variable intake rates (high in the morning, and decreasing throughout the day; Kendeigh et al. 1969), but this may be due to the fact that their food requires longer processing times (J .M.N. Smith, personal communication). 65 C H A P T E R 5 G E N E R A L D I S C U S S I O N TRAPLINES It has always been assumed that traplining birds follow similar routes between patches on different bouts (Feinsinger 1976; G i l l 1988a) , but until now no one has statistically tested whether these routes were indeed similar. M y results in chapter 2 suggest that P. superciliosus do follow similar routes on different foraging bouts. Birds d id not forage randomly in their environment: they approached and departed patches of flowers from a small set of directions (Table 1, Figs. 1 and 2). I interpret these results to indicate that the birds I observed followed routes repeatedly. However, it is impossible to rule out simpler alternative explanations like basic central place foraging without observing movements between several patches of flowers, and I was not able to do that in this study. Future studies are needed to determine that these birds use the same routes on different foraging bouts. This could be done by placing observers at many patches of flowers used by P. superciliosus, and mapping the routes flown by individual birds over several days. By following the same routes between patches on different foraging bouts, traplining birds can learn which patches are profitable, and w i l l be able to react when there is a sudden change in nectar availability at a patch (e.g. when a competitor removes a large amount of nectar). As well , regular harvesting of the nectar in the patches may deter competition by keeping resources at a low level (Schoener 1971; Wol f et al. 1975; Paton & Carpenter 1984; G i l l 1988a, Possingham 1989). 66 General Conclusions EFFECTS OF NECTAR AVAILABILITY M y field study (Chapter 2) suggests that patches of different sizes w i l l have different numbers of visitors depending on number of flowers open and nectar production rates. As well , factors such as distance from the lek and activity levels on the lek w i l l affect how many visitors a patch w i l l receive at any given time. There is a l imit to how many birds can profitably visit a patch. When there are too many visitors, the nectar reward per visitor is likely to be quite low, which wou ld in turn cause individuals who do not control the nectar harvesting schedule of a patch (Gil l 1988a) to drop out as it becomes unprofitable for them to visit it (as nectar production rates drop). Individual birds should react to the same patch in different ways depending on the profitability of the patch with respect to other patches in their traplines, the level of competition at the patches, and distance from the lek/nest. The results from my enclosure experiments indicate that Long-tailed hermit hummingbirds, like many other nectarivores, are capable or detecting and responding to changes in nectar availability at single feeders. P. superciliosus responded to changes i n nectar production at individual feeders by visiting them more or less often, depending on the direction of the change (Fig. 8). Traplining nectarivores in the w i l d drop patches that become unprofitable, and incorporate more profitable ones into their traplines (Thomson et al. 1982, 1987). Both bumble-bees and territorial rufous hummingbirds (Selasphorus rufus) increase visitation rates to more rewarding plants (Thomson et al. 1989; Gass & Sutherland 1985) as P. superciliosus d id in this study. Nectar availability in the environment as a whole has a strong affect on the behavior of all nectarivores. For territorial nectarivorous birds, both territory area and foraging time varies inversely with flower density and nectar availability (Gass 1979; Armstrong 1992; Hixon et al. 1983). S. rufus adjust territory size on a daily 67 General Conclusions basis to maintain or improve territory quality (Gass 1979). M y studies indicate that an increase in total nectar availability w i l l affect traplining birds traplines in two ways: first birds w i l l visit fewer patches and secondly they w i l l decrease visitation rates to individual patches. In the enclosures, P. superciliosus visited feeders less often on Satiation days than on Control days (Fig. 9). This relationship is important to plants pollinated by trapliners, as they can affect how many potential pollinators (and potential mates) they have by the amount of nectar they produce (Stiles 1981). If there is a shortage of plants with respect to pollinators, increasing nectar rewards w i l l increase number of birds visiting. If there is a shortage of pollinators (i.e. if the number of visitors remains similar), the number of patches visited by each bird w o u l d decrease wi th increasing nectar production, lowering pollination potential for plants. COMPETITION It has often been assumed that trapliners, like territorial nectarivores, have almost exclusive use of the flowers along their routes (Feinsinger & Chapl in 1975; Feinsinger 1976; G i l l & Wolf 1977; Stiles & Wolf 1979). It seems clear from Gi l l ' s (1988a) and my studies that in many cases a traplining bird does not have complete control over the flowers it uses. G i l l found that visits by lek males to feeding sites up to 500 meters from the lek constituted 17-50% of all visits recorded at these sites. While I was unable to distinguish lek males from females or non-lek males, my results showed similar division of the resources. If competition decreases the total amount of nectar a bird can take from a patch, it is expected that P. superciliosus w i l l detect and respond to competitive visits by other birds. One variable that birds can control is how long they spend between visits to a patch. In my field study, I found that the time to next visit and number of competitive visits between current and next visit were positively correlated. The 68 General Conclusions longer birds waited to return to a flower, the more competitive visits from other birds they faced, indicating that birds should respond to competition by returning sooner to a patch and "beating the competition" (Chapter 2). These results suggest that birds could respond to competition by returning more to a patches wi th competitive visits than to those without. M y enclosure study (Chapter 3) indicates that birds' responses to competition w i l l depend on how successful the competitors are (i.e. how much nectar they remove). Birds that faced high levels of competition in the enclosure returned more often to that feeder than when the level of competition was low (Fig. 10). P. superciliosus in the wi ld respond to competition at high reward feeders (Gil l 1988a). However, at natural patches of H . pogonantha, birds d id not respond to the number of competitive visits between their consecutive visits (Chapter 2). Wi th the relatively low nectar production rates of flowers in the w i l d , there may not be a close relationship between the number of competitive visitors and the amount of nectar removed: one competitor could remove as much nectar as several depending on how far apart the competitors spaced their visits. If a bird were to encounter a sudden change at a patch that it had previously found profitable, it wou ld most l ikely react as birds d id in the enclosure study and in Gi l l ' s feeder study, by returning more often to the patch after the first indication of competition. Because competitors were always present at the patch I observed in the w i l d , it is more likely that birds were reacting to competition on a presence or absence basis. Comparisons of visitation rates by birds with competition at patches and those without indicate that birds that were faced by competition returned sooner than those who were not (Chapter 2). However, sample sizes here were too small to draw firm conclusions: further studies should investigate the relationship between competition and return rates. 69 General Conclusions ENERGY REGULATION Researchers of energy regulation by territorial hummingbirds suggest that the energy regulation of these birds can be described by an accumulation-depletion model of short term food consumption in which feeding behavior is regulated to maintain constant rate of net energy accumulation throughout the day (Hainsworth 1978, 1981; Wolf & Hainsworth 1977). This rate is set once at the beginning of each day based on energy reserves at that time relative to some standard. In principle this model accounts for both short term (hour to hour) and long term (day to day) regulation of energy reserves, and allows individuals to respond to seasonal contingencies such as high rates of fat gain during migratory stop-overs (Carpenter et al. 1983). This accumulation-depletion model with its assumption of constant required energy accumulation predicts constant food intake rate under constant temperature and ad l ibitum food availability, and territorial hummingbirds tend to do this (Gass 1978a and unpublished observations; Wolf & Hainsworth 1977; Hainsworth 1978; Hainsworth and Wolf 1979; Gass & Montgomerie 1981; Tooze & Gass 1985). In my enclosure study I discovered that P. superciliosus does not have constant net intake rates as predicted by Hainsworth et al. These birds have decreasing net intake over the day, mirroring the decreasing nectar production rates of their food-flowers. In chapter 4 I suggest that traplining birds do not have constant net intake rates as territorial birds do, but rather have intake rates that change over the day, so that they forage the most in the morning when nectar production rates are highest. M y model shows that birds with decreasing net intake rates over the day are able to reach their energy goals (store enough fat for overnight survival without entering torpor) with fewer flowers than are birds with constant net intake rates. This is important for traplining birds, who must travel long distances between patches of flowers in their trapline. As well , hummingbirds who forage more when 70 General Conclusions more nectar is available w i l l decrease the chance of nectar loss to competitors. When hummingbirds can only defend flowers by depleting them, it may be their only option to take in nectar as it becomes available, rather than spread out nectar consumption over the day. 71 L I T E R A T U R E C I T E D Armstrong, D. 1992. Correlation between nectar supply and aggression in territorial honeyeaters: causation or coincidence? Behav. Ecol. Sociobiol. 30: 95-102. Armstrong, D. P., Gass, C. L . , Sutherland, G. D. 1987. Should foragers remember where they've been? Explorations of a simulation model based on the behavior and energetics of territorial hummingbirds. In: Foraging Behavior (Ed by A . C. Kami l , J. R. Krebs, and H . R. Pulliam), pp 563-586. N e w York: Plenum Press. Batschelet, E. 1981. Circular statistics in biology. Academic Press, Toronto. 371 p. Brown, J .H. 1964. The evolution of diversity in avian territorial systems. Wilson Bul l . 76: 160-169. Calder, W . A . , Calder, L . L . , Fraizer, T.D. 1990. The hummingbird's restraint: a natural model for weight control. Experientia 46: 999-1002. Carpenter, F.L. & MacMi l len , R.E. 1976. Threshold model of feeding territoriality and test wi th a hawaiian honeycreeper. Science 194: 639-642. Carpenter, F. L . , Paton, D. C , Hixon, M . A . 1983. Weight gain and adjustment of feeding territory size in migrant hummingbirds. Proc. Nat l . Sci. U S A 80: 7259-7263 Charnov, E.L. , Orians, G . H . , Hyatt, K. 1976. Ecological implications of resource depression. A m . Nat. 110: 247-259. Davies N . B . and Houston, A.I . 1981. Owners and satellites: the economics of territory defense in the Pied Wagtail, Motacilla alba. J. A n i m . Ecol. 50: 157-180. DeBenedictis, P .A. , G i l l , F.B., Hainsworth, F.R., Pyke, G . H . , and Wolf, L . L . 1978. Opt imal meal size in hummingbirds. A m . Nat. 112: 301-316. Ewald , P.W. and Orians, G . H . 1983. Effects of resource depression on use of inexpensive and escalated aggressive behavior: experimental tests using Anna Hummingbirds. Behav. Ecol. Sociobiol. 12: 95-101. 72 Feinsinger, P. 1976. Organization of a tropical guild of nectarivorous birds. Ecol. Monogr. 46: 257-291. Feinsinger, P. 1978. Ecological interactions between plants and hummingbirds in a successional tropical community. Ecol. Monogr. 48: 269-287. Feinsinger, P. 1987. Approaches to nectarivore-plant interactions in the N e w Wor ld . Revista Chilena de Historia Natural 60: 285-319. Feinsinger, P., Busby, W., Murray, K. G. , Beach, J. H . , Pounds, W . Z . and Linhart, Y. B. 1988. Mixed support for spatial heterogeneity in species interactions: Hummingbirds in a tropical disturbance mosaic. A m . Nat. 131: 33-57. Feinsinger, P. and Chaplin, S.B. 1975. On the relationship between wing disc loading and foraging strategy in hummingbirds. A m . Nat. 109: 217-224. Feinsinger, P., Colwel l , R.K. 1978. Community organization among Neotropical nectar-feeding birds. Amer. Zool. 18: 779-795. Frost, A . K . and Frost, P.J.H. 1980. Territoriality and changes in resource use by sunbirds at Leonotis leonurus (Labiatae). Oecologia 45: 109-116. Gass, C L . 1978a. Experimental studies of foraging in complex laboratory environments. Amer. Zool . 18: 617-626. Gass, C L . 1978b. Rufous hummingbird feeding territoriality in a suboptimal habitat. Can. J. Zool . 56:1535-1539. Gass, C L . 1979. Territory regulation, tenure, and migration in rufous hummingbirds. Can. J. Zool. 57: 914-923. Gass, C .L . , Angehr, G. & Centa, J. 1976. Regulation of food supply by feeding territoriality in the rufous hummingbird. Can. J. Zool . 54: 2046-2054. Gass, C L . and Montgomerie, R.D. 1981. Hummingbird foraging behavior: Decision-making and energy regulation. In: Foraging behavior: ecological, ethological, and psychological approaches (Ed. by A . C Kami l and T.D. Sargent), pp 157-194. N e w York: Garland S T P M Press. 73 Gass, C .L . and Sutherland, G.D. 1985. Specialization by territorial hummingbirds on experimentally enriched patches of flowers: energetic profitability and learning. Can. J. Zool. 63: 2125-2133. G i l l , F.B. 1978. Proximate costs of competition for nectar. Amer. Zool . 18: 753-763. G i l l , F.B. 1985. Hummingbird flight speeds. A u k 102: 97-101. G i l l , F.B. 1988a. Trapline foraging by hermit hummingbirds: Competit ion for an undefended, renewable resource. Ecology 69: 1933-1942. G i l l , F.B. 1988b. Effects of nectar removal on nectar accumulation in flowers of Heliconia imbricata (Heliconiaceae). Biotropica 20(2): 169-171. G i l l , F.B. 1989. Ecological fitting: use of floral nectar in Heliconia stilesii Daniels by three species of hermit hummingbirds. Condor 89: 779-787. G i l l , F.B. & Wolf, L . L . 1977. Nonrandom foraging by sunbirds in a patchy environment. Ecology 58:1284-1296. Hainsworth, F.R. 1978. Feeding: Models of costs and benefits in energy regulation. Amer. Zool . 18: 701-714. Hainsworth, F.R. 1981. Energy regulation in hummingbirds. Amer . Sci. 69: 420-429. Hainsworth, F.R., Tardiff, M . , Wolf, L . L . 1981. Proportional control for daily energy regulation in hummingbirds. Physiol. Zool . 54: 452-462. Hainsworth, F.R. and Wolf, L . L . 1972. Crop volume, nectar concentration and hummingbird energetics. Comp. Biochem. Physiol. 42A: 359-366. Hainsworth, F.R. and Wolf, L . L . 1979. Feeding: A n ecological approach. In: Advances in the study of behavior (Ed. by J.S. Rosenblatt, R . A . Hinde, C. Beer, and M . C . Busnel). pp 53-96. N e w York: Academic Press Hainsworth, F.R. and Wolf, L .L . 1983. Model and evidence for feeding control of energy. Amer. Zool . 23: 261-272. 74 Hixon , M . A . , Carpenter, F.L. , Paton, D.C. 1983. Territory area, flower density, and time budgeting in hummingbirds: an experimental and theoretical analysis. A m . Nat. 122: 366-391. Janzen, D . H . 1971. Euglossine bees as long-distance pollinators of tropical plants. Science 171: 203-205. K a m i l , A . C . 1978. Systematic foraging by a nectar-feeding bird, the Amak ih i (Loxops virens). J. Comp. Physiol. Psychol. 98(3): 388-396. Kendeigh, S.C., Kontogiannis, J.E., Malzac, A . , and Roth, R.R. 1969. Environmental regulation of food intake by birds. Comp. Biochem. Physiol. 31: 941-957. K i n g , J.R. 1974. Seasonal allocation of time and energy resources in birds. In: A v i a n energetics (Ed. by R.A. Paynter, Jr.), pp 4-79. Cambridge, Mass.: Publ . Nuttal l Ornithol. Club No. 15. Kodric-Brown, A . & Brown, J.H. 1978. Influence of economics, interspecific competition, and sexual dimorphism on territoriality of migrant rufous hummingbirds. Ecology 59(2): 285-296. Linhart, Y. B. 1973. Ecological and behavioral determinants of pollen dispersal in hummingbird-pollinated Heliconia. A m . Nat. 107: (956) 511-523. Mi l le r , R.S. 1967. Pattern and process in competition. A d v . Ecol. Res. 4: 1-74. Montgomerie, R.D. 1979. The energetics of foraging and competition in some Mexican hummingbirds. PhD Thesis, M c G i l l University, Montreal , Quebec. Paton, C C . & Carpenter, F.L. 1984. Peripheral foraging by territorial rufous hummingbirds: defense by exploitation. Ecology 65: 1808-1819. Possingham, H.P . 1989. The distribution and abundance of resources encountered by a forager. A m . Nat. 133: 42-60. Schemske, D.W. 1980. Floral ecology and hummingbird pollination of Combretum farinosum in Costa Rica. Biotropica 12(3): 169-181. 75 Schoener, T.W. 1971. Theory of feeding strategies. A n n u . Rev. Ecol. Syst. 2: 369-404. Slud, P. 1964. The birds of Costa Rica. Bull . Amer. Mus. Nat. Hist. 138: 1-430 Stephens, D.W. and Charnov, E.L. 1982. Optimal foraging: some simple stochastic models. Behav. Ecol. Sociobiol. 10: 251-263. Stiles, F .G. 1975. Ecology, flowering phenology, and hummingbird pollination of some Costa Rican Heliconia species. Ecology 56: 285-301. Stiles, F .G. 1978. Ecological and evolutionary implications of bird pollination. Amer. Zool . 18: 715-727. Stiles, F .G. 1981. Geographical aspects of bird-flower coevolution, wi th particular reference to Central America. Ann . Missouri Bot. Gard. 68: 323-351. Stiles, F .G. and Freeman, C.E. 1993. Patterns in floral nectar characteristics of some bird-visited plant species from Costa Rica. Biotropica 25(2): 191-205. Stiles, F .G. and Wolf, L . L . 1970. Hummingbird territoriality at a tropical flowering tree. A u k 87:467-491. Stiles, F .G. and Wolf, L . L . 1979. Ecology and evolution of lek mating behavior in the long-tailed hermit hummingbird. Ornithological Monographs 27: 1-78. Sutherland, G.D. , Gass, C .L . , Thompson, P.A. , Lertzman, K.P . 1982. Feeding territoriality in migrant rufous hummingbirds: defense of yellow-bellied sapsucker (Sphyrapicus varius) feeding sites. Can. J. Zool . 60: 2046-2050. Southwick, E.E. and Southwick, A . K . 1980. Energetics of feeding on tree sap by ruby-throated hummingbirds in Michigan. A m . M i d i . Nat. 104: 328-334. Tamm, S. 1989. Importance of energy costs in central place foraging by hummingbirds. Ecology 70(1): 195-205. Thomson, J.D. 1988. Effects of variation in inflorescence size and floral rewards on 76 the visitation rates of traplining pollinators of Ara l i a hispida. Evol . Ecol. 2: 65-76. Thomson, J.D., Maddison, W.P. & Plowright, R.C. 1982. Behavior of bumble bee pollinators of Ara l ia hispida Vent. (Araliaceae). Oecologia 54: 326-336. Thomson, J.D., McKenna, M . A . & Cruzan, M . B . 1989. Temporal patterns of nectar and pollen production in Ara l ia hispida: implications for reproductive success. Ecology 70(4): 1061-1068. Thomson, J.D., Peterson, S.C. & Harder, L .D. 1987. Response of traplining bumble bees to competition experiments: shifts in feeding location and efficiency. Oecologia 71: 295-300. Thomson, J.D., Plowright, R.C. 1980. Pollen carry-over, nectar rewards, and pollinator behavior with special reference to Diervi l la lonicera. Oecologia (Berlin) 46: 68-74. Tiebout, H . M . 1991. Daytime energy management by tropical hummingbirds: responses to foraging constraint. Ecology 72(3): 839-851. Tiebout, H . M . 1992. Comparative energetics of divergent foraging modes: a doubly labelled water experiment on hummingbird competition. A n i m . Behav. 44: 895-906. Tiebout, H . M . 1993. Mechanisms of competition in tropical hummingbirds: Metabolic costs for losers and winners. Ecology 74(2): 405-418. Tooze, Z . and Gass, C L . 1985. Responses of rufous hummingbirds to midday fasts. Can. J. Zool . 63: 2244-2253. Waser, P . M . 1981. Sociality or territorial defense? The influence of resource renewal. Behav. Ecol. & Sociobiol. 8: 231-237. Westcott, D. 1994. Leks of leks: A role for hotspots in lek evolution? Proc. R. Soc. Lond. B 258: 281-286. Wolf, L . L . and Hainsworth, F.R. 1971. Time and energy budgets of territorial 77 hummingbirds. Ecology 52(6): 980-988. Wolf, L . L . and Hainsworth, F.R. 1977. Temporal patterns of hummingbird feeding. A n i m . Behav. 25: 976-989. Wolf, L . L . and Hainsworth, F.R. 1983. Economics of foraging strategies in sunbirds and hummingbirds. In: Behavioral energetics: the cost of survival in vertebrates. (Ed. by E. P. Aspey and S. I. Lustic), pp 223-264. Columbus, Ohio: Ohio State University Press. Wolf, L . L . and Hainsworth, F.R. 1991. Hummingbird foraging patterns: visits to clumps of Ipomopsis aggregata inflorescences. A n i m . Behav. 41: 803-812. Wolf, L . L . , Hainsworth, F. R., G i l l , F. B. 1975. Foraging efficiencies and time budgets in nectar-feeding birds. Ecology 156:117-128. Wolf, L . L . , Hainsworth, F.R., and Stiles, F.G. 1972. Energetics of foraging: rates and efficiency of nectar extraction by hummingbirds. Science 176: 1351-1352. Yamamura, N . and Tsuji, N . 1987. Optimal patch time under exploitative competition. A m . Nat. 129(4): 553-567. 78 Appendices Appendix I: Mode l parameters. Flower parameters are based on Stiles (1975) and on personal observations. Asymptotic Amax and Aslope determine the production rate of flowers, starting at 100 p l / h at 0600 h (Appendix II). Flowers in trapline ranges from 20 to 50. Perch to patch distances and spacing of flowers are estimates based on Stiles 1975, G i l l 1988, and personal observation. Bird weight, wing length, and flight speeds are from G i l l 1985. Extraction intercept and 1/Extraction rate are from Wolf et al. 1972 for P. superciliosus at H . tortuosa. Meal size ranged for the model from 100 to 780 pi. Target fat gain is based on Hainsworth 1978. Parameter Value Al t i tude Day Length (h) M i n . A i r Temp (C) Max. A i r Temp. (C) Starting Nectar V o l (pi) Asymptotic Amax Asymptotic Aslope Nectar Concentration (%) Flowers in Trapline Perch to Patch Distance (m) Flower Spacing (m) Mass of Bird (g) W i n g Length of Bird (cm) Flight Speed (m/s) Extraction intercept (s) 1/Extraction Rate (s/ul) Meal Size (pi) Target Fat Gain for Day (kj) 0 12 20 30 40 0.285 0.022 33 80 100 20 6.1 6.2 11 1.26 0.39 600 13.8 79 Appendices Appendix II: Nectar production rates for our model, a) Nectar volume (ul) accumulated over time since flower was emptied. Numbers on lines represent time of day when flower is drained. Thus, the nectar production rate decreases with time of day, and nectar replacement rates are lower when the flower is drained later in the day. b) A n example of the pattern of nectar availability in a flower over time of day due to foraging by a bird. In this example, the flower is drained every 3 hours by the bird, and then nectar accumulates at a rate determined from above (a). More nectar accumulates after the flower is drained early in the day than later in the day. 100 0 3 6 9 12 TIME SINCE EMPTIED (h) 75 TIME OF DAY (h) 80 Appendices Appendix III: Fat Accumulation Rate Model . Instantaneous and cumulative net intake rates based on Hainsworth (1978) assumption that a bird of 6 g in a 12 hour day at 25° C would need to store 13.8 kj excess energy to survive the night without entering torpor, m = rate of change of accumulation rate in k j / h 2 , where h= time of day. For this graph, m = -0.5 k j / h 2 . H 0.6 2 2 0.4 » 0.2 0 ' 1 ' 1 — H 1 — > — 1 1 I TIME (h) 81 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0099029/manifest

Comment

Related Items