Behavioral Ecology Vol. 10 No. 6: 714-725
© 1999 International Society for Behavioral Ecology
Response of a traplining hummingbird to changes in nectar availability
Department of Zoology, University of British Columbia, 6270 University Boulevard, Vancouver, BC V6T 1Z4, Canada
Address correspondence to J. S. E. Garrison, Department of Zoology, University of Hawaii, 2538 The Mall, Honolulu, HI 96822, USA. E-mail: garrison{at}hawaii.edu .
Received 12 February 1998; accepted 7 June 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Theory predicts that nectarivores respond to changes in profitability of patches of flowers or feeders by adjusting visitation rate to increase reward size. We conducted a set of experiments in an outdoor enclosure with seven feeders to determine how Phaethornis longirostris, a traplining hummingbird, adjusts its visitation rates in response to changes in sucrose solution delivery rates. Each experiment tested the response of P. longirostris to the following changes in the timing and volume of sucrose solution delivery: (1) increases in sucrose solution abundance at all feeders (mimicking seasonal increases in numbers of open flowers or nectar output); (2) large changes in sucrose solution availability at one feeder (mimicking increases or decreases of patch profitability); and (3) sudden unexpected decreases in sucrose solution availability at one feeder (mimicking loss of nectar to competitors). We found that P. longirostris (1) decreased visitation rates when the sucrose solution delivery rate was higher at all feeders, (2) increased visitation rates to individual feeders when their profitability increased for whole days but did not significantly decrease visitation rates when feeder output decreased; and (3) responded to sudden food losses at a feeder (due to simulated competition) by increasing use of that feeder for 1 or 2 h after the loss.
Key words: foraging behavior, hummingbirds, nectarivore, Phaethornis longirostris, Phaethornis superciliosus, trapline.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Nectarivores modify their foraging strategies to deal with environmental changes such as the spacing of flowers, the quantity of nectar produced by each flower relative to the harvesting costs, and competition from other nectarivores (Feinsinger, 1976
;
Feinsinger and Colwell, 1978;
Feinsinger et al., 1988;
Stiles, 1981). Nectarivores
that visit but do not defend widely spaced flowers on repeated routes are
called "trapliners" (Feinsinger
and Chaplin, 1975). Traplining is economical only when relatively
few patches are visited, the probability of competition is low, and nectar
rewards are rich enough to compensate for energy spent traveling between
flowers (Gill and Wolf, 1977;
Hainsworth, 1981).
In this paper we describe a study of the foraging behavior of the
long-tailed hermit hummingbird (Phaethornis longirostris; formerly
P. superciliosus; Hinkelmann,
1996). Long-tailed hermits are found in the lowland tropical
rainforests throughout Central America
(Slud, 1964) and are
considered to be model trapliners (Gill,
1988b), although the repeatability of their routes has never been
tested. They are large monomorphic hummingbirds (~6 g) that use highly
productive flowers such as Heliconia pogonantha, which grow in small
patches of one to a few plants in light gaps and along forest edges
(Stiles and Wolf, 1979).
Traplines, or foraging circuits, enable animals to organize their harvest
of regularly renewing food such as nectar from isolated sites, and may deter
competition by keeping resources low
(Gill, 1988b;
Paton and Carpenter, 1984;
Possingham, 1989;
Schoener, 1971;
Wolf et al., 1975). Traplining
has been described in many studies of nectarivorous bird and insects (e.g.,
Feinsinger, 1976;
Gill, 1988b;
Janzen, 1971;
Stiles and Wolf, 1979;
Thomson, 1996;
Thomson et al., 1997), yet
only a few experimental studies have tested behavioral responses to
environmental changes (Gill,
1988b; Thomson,
1988; Thomson et al.,
1987; Tiebout,
1991, 1992,
1993). For instance, how do
trapliners respond to changes in the profitability of individual flower
patches due to competition, number of flowers open in the patch, or nectar
quality and quantity?
Nectarivores often respond to changes in profitability of patches of
flowers or feeders by adjusting visitation rate to increase reward size (e.g.,
Gass and Sutherland, 1985;
Wolf and Hainsworth, 1983).
For example, Gill (1988b)
found that P. longirostris adjusted visitation rates in response to
increases in sucrose solution reward at particular feeders. However, Gill's
study examined only a few feeders in the wild; the total amount of food
available to the hummingbirds was not controlled, and it was not possible to
determine how changes in the feeders affected use of other nectar sources. His
study, and those of Stiles and Wolf
(1979), Thomson
(1988,
1996), Thomson et al.
(1987,
1989,
1997), and Tiebout
(1991,
1992,
1993) produced several
hypotheses about traplining foraging behavior that can be tested with
laboratory experiments (see below).
We conducted a set of experiments to determine how P. longirostris adjusts its visitation rates to feeders with respect to changes in sucrose solution delivery rates. Each experiment tests the response of P. longirostris to the following changes in the timing and volume of sucrose solution availability: (1) increases in total sucrose solution abundance at all feeders (mimicking a large increase in numbers of open flowers or nectar output due to seasonal effects or other factors); (2) large changes in sucrose solution availability at one feeder (mimicking increases or decreases of patch profitability); and (3) sudden unexpected decreases in sucrose solution availability at one feeder (mimicking loss of nectar to competitors).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Laboratory apparatus and protocol
We captured P. longirostris of unknown sex in mist nets located either near a lek or at a patch of Heliconia pogonantha flowers at La Selva Research Station in the Sarapiqui region of Costa Rica from May to August 1994. We placed individual birds in small (60 x 60 x 60 cm) cages where they were trained for 1 day to use an ad libitum feeder. The next day, birds were placed in an experimental enclosure.
For the experiments, we used two outdoor enclosures, each housing one bird. Enclosures were 3.5 x 7.2 x 2.6 m and were made of rain-permeable shade cloth. Each enclosure had seven feeders and one perch (Figure 1). To simulate the poor visibility in a rainforest, we hid the feeders behind cloth blinds. Birds were given one day to learn the feeder arrangement in the enclosure before data collection began. Following this introductory day, individuals were kept in the enclosure for 4-6 days (control + treatment days). The length of stay of an individual bird depended on which experimental treatment the bird received. There was also some variation in length of stay within experiments (1-2 days maximum) due to unpredictable weather conditions; data were not collected on days with heavy rainfall, as birds tended not to forage in the enclosures during periods of heavy rain.
|
Each feeder was equipped with a photocell that was activated by the bird's
bill when it probed the feeder (Brown and
Gass, 1993). A custom-built data logger automatically recorded
times of arrival and departure from all feeders. From this record we
calculated time and sequence of consecutive feeder visits and time between
visits (seconds) to each feeder. The perch monitors malfunctioned, and we do
not report perching times. For this reason we could not accurately identify
individual bouts of foraging.
The data logger was connected to solenoid valves (General Valve
Corporation, series 3; Brown and Gass,
1993) that delivered a 0.98 M sucrose solution to the feeders
(H. pogonantha nectar is 0.92 M;
Stiles and Freeman, 1993).
Hummingbirds of 6 g require at least 36 kJ of energy intake per day to survive
(King, 1974), and captive
P. longirostris consumed 39.7-58.2 kJ/day from an unlimited feeder in
the enclosure during a companion study
(Gass and Garrison, 1999). In
this study, the total energy available to birds in the enclosure ranged from
47 to 101 kJ/day, depending on the treatment (51 kJ on control days). Solenoid
valves delivered sucrose solution in discrete 1-µl squirts. Feeder output
was adjusted by altering the interval between squirts: for example, 1
squirt/min = 60 µl/h. Delivery was discrete; if a bird emptied a feeder and
returned again before the next scheduled squirt, it received nothing. Feeders
began delivering sucrose solution at 0600 h and stopped at 1700h.
On average, each feeder delivered sucrose solution at approximately 10 times the rate a single H. pogonantha flower would produce (Figure 2). In this sense each feeder modeled a patch of 10 flowers, and the set modeled a trapline of 7 patches (or 70 flowers). Although the small size of the enclosure and high spatial concentration of sucrose solution did not mimic natural conditions, it did allow us to closely examine birds' behavioral responses to changes in single feeders and in overall sucrose solution availability.
|
Experimental design and predictions
Control days
We placed 200 µl of sucrose solution into the feeders just before
starting the sucrose solution delivery schedule at 0600 h to simulate the
large amount of nectar available in Heliconia flowers when they open
at dawn (Stiles, 1975). The
control period lasted for 1-3 days, during which time all feeders received the
same amount of sucrose solution at the same declining rate
(Figure 2). Each bird
experienced the control period and one of the experimental treatments
described in the following sections.
We used a declining rate of sucrose solution delivery throughout this study
because nectar production rates of many flowers used by traplining
hummingbirds (including H. pogonantha) are high in the early morning
and decrease rapidly during the day (Figure
2; Stiles, 1975;
Stiles and Freeman, 1993).
Territorial hummingbirds often increase foraging effort when standing crop is
low (Gass, 1978a,
b;
Gass and Montgomerie, 1981;
Gass et al., 1999;
Gass and Sutherland, 1985;
Sutherland et al., 1982).
However, P. longirostris has coevolved with flowers that produce
little nectar in the afternoon, and the birds cannot defend them from
competitors. To maximize net energy intake, P. longirostris was
predicted to decrease its feeding frequency during the day as nectar
production declined (Gass and Garrison,
1999) rather than attempting to maintain constant net intake
(sensu Feinsinger, 1976;
Frost and Frost, 1980;
Hainsworth et al., 1981;
Stiles and Wolf, 1979).
Experiments
Changes in total nectar availability. Total nectar availability in
hummingbirds' traplines or habitats should affect how many patches they visit
and thus how far they travel to meet their energetic needs. Gill
(1988b) suggested that due to
satiation, P. longirostris waited longer between bouts of foraging
after visiting feeders with large amounts of sucrose solution than they did
after visiting feeders with small amounts of nectar. If satiation affects
P. longirostris visitation rates and patterns, increasing the amount
of sucrose solution available in the enclosure should cause them to lower
their total visitation rate to feeders and perhaps stop using some of the
feeders altogether (Gill,
1988b). Lower visitation rates could be due to physical
limitations (such as a full crop), or time constraints (time saved by visiting
less frequently can be used for activities such as digestion, resting, or
singing).
To determine how an increase in total nectar availability affects P. longirostris foraging behavior, we increased the output rate of all feeders in the enclosure so that birds received sucrose solution at roughly 1.5 times the rate as on the control days (Figure 2). We also doubled the amount of initial sucrose solution available at 0600 h to 400 µl per feeder. We used four birds in this satiation treatment, which lasted 3 days for each bird.
Changes in patch quality. In the wild, patches of flowers change
in quality (i.e., amount of nectar available) due to the number of flowers
open, the number of competitors visiting them, the age of the flowers, and
flowering phenology of the individual plant. To maintain high net energy
intake, P. longirostris should respond to an increase or decrease in
the amount of nectar available at individual patches by visiting the patches
more or less often, respectively, than before the change
(Gill, 1988b). Traplining bees
(Thomson et al., 1982,
1987) and territorial
hummingbirds (Gass and Sutherland,
1985) both do this. A change in food availability at one patch may
also affect use of unchanged patches if the response to changed patches
greatly enhances or reduces total energy intake
(Gass and Sutherland,
1985).
To determine how changes in patch quality (sucrose solution delivery rates at one feeder) affect P. longirostris foraging behavior, we programmed the data logger to halve or double the output rate of one feeder (1 and 9; Figures 1 and 2) relative to the others, starting at 0600 h. The change occurred before the bird started foraging in the morning (with 100 µl placed in the decrease feeder and 400 µl placed in the increase feeder at 0600 h). The relative difference persisted throughout the day (Figure 2), so it did not represent sudden changes that might accompany competition. We refer to these treatments as decrease or increase. Decrease and increase treatments were run on separate days, with two different birds for each treatment. Treatments lasted 1-2 days depending on weather conditions.
Nectar loss to competition. Trapliners do not and cannot defend
their flowers as aggressively as territorial nectarivores. One way for
trapliners to minimize the effects of competition is through the strategy of
"defense by depletion," or exploitative defense
(Davies and Houston, 1981;
Feinsinger, 1987;
Paton and Carpenter, 1984;
Possingham, 1989;
Waser, 1981). Defense by
depletion requires that foragers return more frequently than dictated by
replacement rates alone to empty patches that were previously profitable.
P. longirostris did this at high-reward feeders in the wild
(Gill, 1988b). Thus, the
hypothesis predicts that competition increases frequency of visits to some or
all patches of flowers along the trapline, even at the expense of smaller
harvests and lower (or even negative) profitability (at least temporarily;
Gill, 1988b). Because nectar
production rate should also influence return rate
(Frost and Frost, 1980;
Gill, 1988b), visit
frequencies should be a compromise between maximizing nectar accumulation and
minimizing loss to other individuals through exploitative competition
(Gill and Wolf, 1977).
To keep nectar levels low and discourage competitors from returning, P.
longirostris should visit a feeder with competition more frequently than
on control days and more frequently than the other unmanipulated feeders (on
competition days; Gill,
1988b). To test how sudden sucrose solution loss due to
competition affects P. longirostris foraging behavior, we entered the
enclosure every 15 min from 0700 to 1200 h, removed any sucrose solution found
in feeders 3 and 11 (Figure 1)
with a glass microcapillary tube, and recorded the amount of sucrose solution
removed. Total amount of sucrose solution removed varied between birds
(depending on their rate of visitation to the feeder) and ranged from 0 to 92
µl/h (for a total of 4-115 µl for the half day). The sucrose solution
delivery rate was similar at all feeders and was the same as during control
days (Figure 2). We used four
birds in competition treatments, which lasted 2 days for each bird.
Analysis
We analyzed two dependent variables: visits per hour and mean return time.
The number of visits per hour was a tally of the number of probes to a feeder
that were more than 1 s apart in time during an 1-h period. Return time was
the number of seconds that elapsed between consecutive visits to the same
feeder. Mean return time was the average of these return times for one feeder
over an hour. Although number of visits and mean return time were closely
correlated (r2 = 79.6%, df = 2305, p 
.0001),
mean return time was more prone to high variation at low numbers of visits per
hour than at high numbers of visits per hour. Because of this
heteroscedasticity, we used number of visits per hour to feeders rather than
mean return time to a feeder. However, analyses of both variables gave
qualitatively similar results.
Control day
To determine if the different experiments were comparable, and if bird
identity, time of day, or feeder identity had any significant effect on hourly
visitation rates to feeders, we first analyzed the control day data for all
birds using a general linear model. For this model, bird identity was nested
within experiment (satiation, increase, decrease, competition) because no bird
experienced more than one experimental treatment.
Experiments
Because not all experiments had similar methodology (e.g., differing
numbers of feeders manipulated), we analyzed each experiment separately. We
used four general linear models with visits per hour as the dependent
variable, and treatment, time of day, and bird identity as categorical
independent variables. We included all two-way interactions between the
factors in the initial analyses but the final models included only significant
higher order interactions (Stewart-Oaten,
1995). Removal of insignificant higher order interactions did not
change the qualitative results.
Changes in total nectar availability. For the satiation experiment, all feeders changed on the experimental day, so treatment was simply coded as control or satiation. In this model, feeder identity was also included as an independent variable.
Changes in patch quality. Treatment for the increase and decrease experiments was coded as control (decrease/increase), control (unmanipulated), decrease/increase feeder (the feeder which experienced the change on the experimental day), or unmanipulated feeders (those feeders which were not changed on the experimental day). For these experiments, individual feeder identities were not used in the models. For both days of the decrease experiment, the data logger malfunctioned in the afternoon (1400 h onward), and so for this experiment we analyzed data from 0600 to 1300 h.
Nectar loss to competition. Data for the competition experiment
were coded the same as in increase and decrease experiments. In the
competition experiment, some birds removed the sucrose solution before we
could (i.e., they "outcompeted" us by having high visitation rates
before our attempts to remove sucrose solution). These birds did not
experience the treatment (removal of food). Therefore, for the general linear
models we only analyzed data for competition days where birds lost at least 20
µl in any hour (
14% of the amount produced in the competition feeder in
an hour). We removed 35-115 µl per day for bird-days included in the linear
model (n = 5 bird-days), compared to 4-27 µl removed for bird-days
excluded from the analysis (n = 3 bird-days). For this experiment, we
analyzed data from 0600 to 1200 h, when competition ended.
Relative use. During preliminary analyses, we found that feeder visitation rates varied over time of day and between days and that birds behaved differently from each other (some were more active foragers than others). Thus we decided to also analyze the data using a dependent variable that was standardized for bird activity levels. We chose "relative use of feeder," which is the total number of visits to a feeder for a given hour, divided by total number of visits to all feeders for that hour. A relative use index makes it possible to adjust for some of the differences in bird activity levels, as well as differences within birds over time and days. This standardized variable is useful when there are factors that affect behavior but cannot be controlled for (such as time of day or weather). For these analyses, we used similar general linear models as in the analyses discussed above (with manipulation and bird as independent variables), except that time was removed because it was no longer a significant influence on relative use of feeders.
All statistical analyses were performed with Data Desk v. 6.1.1 for
Macintosh (Velleman,
1997).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Control days
There was no significant effect of experimental block on control day visitation rates (Table 1). This is consistent with the assumption that despite the fact we used different birds in each of the four experiments, results are comparable.
|
On control days, birds visited feeders less in early morning and late
afternoon than in mid-morning, with a peak at 0900 h
(Table 1 and
Figure 3a). Individual birds
varied in their visitation rates over whole days (some birds were more active
and visited feeders often, while others visited less frequently). This pattern
also varied for some birds over time of day, possibly due to factors such as
rainfall or other unavoidable disturbances
(Table 1, time x bird).
Birds differed somewhat in their use of particular feeders, but on average
they visited the central feeders more often than the rest, and visited the
feeder at the far end of the room the least (with the exception of feeder 15,
which was near a perch but not visited more frequently;
Figure 3b; Scheffe Post-hoc
test: feeder 4 > feeder 1, p .0009; feeders 11 and 12 >
feeder 9, p =.05 (feeder 11-feeder 9) and p =.0001 (feeder
12-feeder 9); feeder 12 > feeder 15, p .0004).
|
Relative use of the feeders did not differ between control days in the experiments, but did differ between birds and feeders (Table 1), with individual birds using some feeders more frequently than others as described above.
Changes in total nectar availability
On satiation days, birds visited feeders only 11% (range: 4-18%) as often
as on control days. The number of visits on satiation days showed a similar
pattern as on control days (a peak in the morning and another in the
afternoon), except that the peaks were not nearly as large as on control days
(Figure 4a,b). On satiation
days birds visited most feeders infrequently for all hours of the day
(Table 2 and
Figure 4b). Individual birds
favored different feeders, and their visitation rates to those feeders varied
with time (Table 2, feeder
x bird and time x bird interactions;
Figure 4c). The use of
individual feeders changed dramatically between control and satiation days,
with some feeders left almost completely unvisited (and dripping with sucrose
solution) during satiation days (Table
2 and Figure
4c).
|
|
Relative use of the feeders was not significantly affected by treatment but was affected by bird and feeder (Table 2). Birds used feeders differently on satiation days, as above, and different birds preffered different feeders. The effects of the satiation experiment can be clearly seen in the higher order interaction between treatment and feeder (Table 2).
Decrease in patch quality
We chose feeders 1 and 9 (Figure
1) as the decrease feeders for this experiment. As mentioned
above, these feeders were visited much less frequently on the control days
than most of the other feeders (bird 1: Scheffe Post-hoc test, p
=.001; bird 2: Scheffe Post-hoc test, p =.05,
Figure 5a,b). The birds also
visited the decrease feeder significantly less than the unmanipulated feeders
on the decrease days (bird 1: Scheffe Post-hoc test, p =.03; bird 2:
Scheffe Post-hoc test, p =.00001;
Figure 5a,b). The two birds
responded somewhat differently to the decrease experiment
(Table 3). Although both
slightly decreased visitation rates to the decrease feeder for at least part
of the morning on the treatment day, the differences were not significant
(bird 1: Scheffe Post-hoc test, p =.42; bird 2: Scheffe Post-hoc
test, p =.92; Figure
5a,b). One bird also slightly decreased overall visitation rates
to the unmanipulated feeders (Figure
5a; Scheffe Post-hoc test, p =.08), while the other bird
slightly increased visitation rates to the unmanipulated feeders on the
treatment day (Figure 5b,
Scheffe Post-hoc test, p =.07). None of these differences are large
enough to be significant, given the amount of variation in visitation rates
over time.
|
|
Relative use
The two birds did not differ from each other in their relative use of the
feeders (Table 3). On the
treatment day, birds used the decrease feeder slightly (but not significantly)
less than on the control day (Figure
5c; Scheffe Post-hoc test, p .23). Relative use of
the unmanipulated feeders on the decrease day was the same as on control days
(Scheffe Post-hoc test, p =.93). Thus it appears that the hermits did
not react strongly to the decrease in feeder profitability by changing
visitation rates or relative use of the feeder.
Increase in patch quality
We used feeder 9 (the farthest feeder from the perch;
Figure 1) in the increase
experiment. As in the decrease experiment, both birds used the control
(increase) feeder less than the control (unmanipulated) feeders
(Table 4; Scheffe Post-hoc
test, p =.007). Birds differed slightly in their visitation rates on
control days (Table 4, time
x bird). One bird had constant visitation rates in the morning on the
control day (for all feeders), while the other had the more usual peak of
activity at 0900 h (Figure
6a,b).
|
|
Birds responded strongly to the higher sucrose solution delivery rate at the increase feeder on the treatment days by visiting it more frequently than on control days (Table 4; Scheffe Post-hoc test, p =.042) and slightly more frequently than the unmanipulated feeders on the treatment days (Figure 6a,b; Scheffe Post-hoc test, p =.079). This insignificance is due to the similarity in visitation rates after noon (Figure 6a,b); there is a clear increase in visitation rates to the increase feeder in the early morning on the treatment days (Figure 6a,b). The birds also visited unmanipulated feeders less frequently on the treatment day than on control days (Scheffe post-hoc test, p =.0009). The differences between visitation rates to the control (increase), control (unmanipulated), increase, and unmanipulated feeders were greatest when overall visitation rates were highest (0900 h; Figure 6a,b), and the differences decreased over time until feeder usage was similar for all feeders in the late morning and afternoon.
Relative use
One bird responded more strongly to the increase treatment than the other
(Figure 6c and
Table 4, manipulation x
bird). Even so, both birds showed the same pattern of increasing use of the
more profitable feeder on the treatment day. The birds' use of the increase
feeder switched from lower than average on the control days to higher than
average on the increase days (Figure
6c). Bird 1's relative use of the increase feeder was
significantly higher than both the unmanipulated and control (increase)
feeders (Table 4 and
Figure 6c; Scheffe Post-hoc
test, p .00001 for both cases). Bird 2 significantly increased
use of the increase feeder between control and treatment days
(Figure 6c; Scheffe Post-hoc
test, p .01). However, bird 2 did not differ in relative use of
the increase and unmanipulated feeders on the treatment day (Scheffe Post-hoc
test, p =.78). For bird 1, relative use of the unmanipulated feeders
was slightly (but not significantly) lower on the treatment day than on the
control day (Figure 6c; Scheffe
Post-hoc test, p =.07). Bird 2 did not differ in relative use of the
unmanipulated feeders between control and treatment days (Scheffe Post-hoc
test, p =.55).
Nectar loss to competition
Because of the methodology used, the competition experiment ended up
treating individual birds differently depending on their activity levels and
visitation rates to the feeder. See Table
5 for amounts of sucrose solution removed from the competition
feeder. One bird lost relatively large amounts of sucrose solution in an hour
(75 and 92 µl on separate days), another lost relatively constant amounts
over a day (~20 µl per hour on three separate hours), one lost 40 µl
at 1100 h, and the others experienced little or no loss. Birds that visited
the competition feeder frequently throughout the halfday were able to prevent
sucrose solution loss. Birds that experienced competition (defined as > 20
µl lost in any hour) had significantly lower average visitation rates (for
the half-day) than those that lost no sucrose solution (birds with
competition: visits per hour = 18.97 ± 2.27 SE; birds without
competition: visits per hour = 38.14 ± 4.46 SE; t = 4.25, df =
13; p .0005).
|
Visits per hour
There was large variation in visitation rates between different birds both
over time and days, and between the same bird on the control and treatment
days, and within treatment days (Table
6, time x bird, manipulation x time). We chose feeders
3 and 11 for the treatment feeders in this experiment, and these were visited
more frequently than some of the other feeders on control days by two of the
four birds. Two of the four birds were also more active on the control days
than on the treatment days (Table
6, manipulation x bird).
|
Most birds increased their visitation rates to the competition feeder after sucrose solution had been removed from it. However, because of the differences in activity levels between control and competition days, and because results are similar but easier to follow with relative use, we focus on relative use of the feeders rather than absolute visitation rates.
Relative use
Birds that faced competition had a higher relative use of the competition
feeder than those that did not (competition, relative use of competition
feeder = 0.21 ± 0.024 SE; no competition, relative use of competition
feeder = 0.149 ± 0.005; t = 2.59, df = 13; p
.02). Individual birds varied in the magnitude of their response to
sucrose solution loss (Table 6,
manipulation x bird). In general, birds' relative use of the competition
feeder in the hour after sucrose solution removal increased with increasing
amounts of sucrose solution removed (Spearman rank correlation,
rs =.34, n = 32,.01 < p <.05).
The bird that experienced significant sucrose solution loss (76 µl and 92 µl in 1 h on two separate days; bird 1 in Table 5) greatly increased relative use of the competition feeder on both days, within 1-2 h after the removal (Figure 7a,b). This bird's relative use of the competition feeder was also elevated in the second morning of the competition treatment (perhaps indicating an expectation of competition at that feeder). The bird that lost a total of 86 µl spread out evenly over the morning (bird 2 in Table 5) also increased relative use of the feeder after sucrose solution loss, but relative use was only higher for the competition feeder on the treatment day than on the control day after the first sucrose solution loss and not after the other two losses (Figure 7c). The other birds that experienced only slight sucrose solution loss or loss later in the day did not increase relative use of the competition feeder (e.g., see Figure 7d).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Control day behavior
Hermit hummingbirds should visit patches or feeders less often and consume less nectar as the day progresses. We found this to be true in a companion study under ad libitum conditions, an observation that challenges established notions of how hummingbirds regulate energy intake during the day (Gass and Garrison, 1999
). In
this study, feeder visitation rates mirrored decreasing sucrose solution
delivery rates for most of the day as predicted. However, the early morning
hours (0600-0800 h) had lower-than-expected visitation rates. This low
visitation rate may reflect satiation due to an abundance of food
(Gill, 1988b). Sucrose
solution delivery rates and standing crops were high at 0600 h: visiting only
four 200-µl feeders would fill the birds' crops, which have a volume of
approximately 780 µl (Hainsworth and
Wolf, 1972). This early morning surplus of food could affect
visitation rates in two ways: birds would be full after visiting a few feeders
and would have to wait for their crops to empty to start new bouts, and they
would not have to visit many feeders to meet their energy requirements.
After the morning surplus is depleted, foraging effort should increase
because birds must visit more feeders to take in the same amounts of nectar.
However, as nectar production rates continue to decrease and standing crop in
the feeders is diminished, hermits should decrease feeding efforts to save
energy. This was the case in the enclosure: visitation rates peaked and then
began to decline after 0900 h along with sucrose solution delivery rates, as
we observed previously at ad libitum feeders
(Gass and Garrison, 1999).
Stiles and Wolf (1979)
observed fewer P. longirostris visits to H. pogonantha
patches in the afternoon than in the morning and noted that total activity
(including that at the lek) decreased in the afternoon. Stiles
(1995) also noted that total
observed foraging activity (foraging for arthropods and nectar) was lower in
the afternoon than in the morning for several species of hermit and nonhermit
hummingbirds at La Selva. In addition, both Chlorostilbon canivetii,
a low-reward trapliner, and Amazilia saucerottei, a territorial
hummingbird, had peak feeder visitation rates in mid-morning and decreased
visitation rates over the rest of the day in an enclosure study in Monteverde,
Costa Rica (Tiebout, 1992).
These tropical species also feed primarily on flowers that have a decreasing
nectar production rate over the day
(Feinsinger, 1976), although
the decline is not as dramatic as in H. pogonantha. Feinsinger
(1976) noted that A.
saucerottei was only territorial at patches of these flowers until
midmorning, after which point they did not defend them. This suggests that
unlike in temperate systems, where foraging effort is bimodal with peaks in
morning and afternoon visitation rates
(Gass and Montgomerie, 1981),
tropical hummingbirds decrease foraging efforts after mid-morning to conserve
energy because nectar production and standing crop are quite low in the
afternoon (Stiles, 1975;
Stiles and Wolf, 1979). In the
enclosure, P. longirostris did show a small peak in feeder visitation
rates in the afternoon (~ 1400 h) but this peak was considerably smaller
than the midmorning peak (Figure
3a). It would be worthwhile to determine if other tropical
nectarivorous species show this same trend.
It is interesting to note that in the small space of the enclosure, birds
visited some feeders more often than others, even though sucrose solution
delivery rates were similar for all feeders on control days. The highest
numbers of feeder visits in our study were to those feeders in the middle of
the room, which birds would pass on their way to visiting feeders at the far
end of the room. This same pattern is seen in P. longirostris in the
wild: flowers near leks (where males spend a great deal of time perching) were
visited more frequently than those far from leks
(Gill, 1988b). Tiebout
(1991) noted in an enclosure
study that both C. canivetii and A. saucerottei visited
closer feeders more frequently than feeders farther away (although he did not
offer both types of feeders at the same time). Thus it seems that hummingbirds
will stop to check nearby flowers or feeders even if they have visited them
recently. In the wild, this behavior of checking recently visited flowers (on
the way to the rest of the trapline, or while displaying in the lek) would
keep nectar levels low in flowers near the leks, where competition for nectar
is highest.
Changes in total nectar availability
For territorial hummingbirds, both territory area and foraging time vary
inversely with flower density and nectar availability
(Gass, 1979;
Hixon et al., 1983). If
traplining birds are satiated after taking nectar from just a few patches of
flowers, this could strongly affect the size of the trapline. To keep energy
expenditures down, traplining hummingbirds should drop feeders or flowers from
their trapline and/or decrease visitation frequency if they no longer require
the nectar produced there. Thus, the size of a trapline should depend on
nectar production rates in the environment as a whole, as well as on how much
nectar each patch provides. Gill
(1988b) noted that P.
longirostris visited feeders (in the field) less frequently when the
amount of sucrose solution available increased (as long as there were no
competitors visiting the feeders). Tiebout
(1991) noted that C.
canivetii and A. saucerottei both visited ad libitum feeders 6.5
times less frequently than feeders with limited amounts of sucrose solution.
Our study supports these observations. On satiation days, P.
longirostris visited feeders only 4-18% as often as on control days. Some
feeders were left completely unvisited for one to several hours of the day.
The hermits also switched between feeders over time so that they were visiting
feeders they had not previously visited often and which contained large
amounts of sucrose solution (Figure
4c). Thus, it appears that (at least without competition) P.
longirostris adjusts both visitation rates to individual patches
(feeders) and trapline size (number of feeders visited) in response to overall
changes in nectar availability across all patches.
Visitation rates to feeders were low on satiation days and did not vary
temporally as much as on control days. It seems likely that, because feeders
were producing more sucrose solution and were visited less often, food was
abundant in the enclosure at all hours of the day during the satiation
treatment—a bird would not have had to visit many feeders to fill its
crop and meet its energetic needs. Despite this, a similar pattern of total
visits to those on control days can still be seen on the satiation days,
although the pattern has been greatly dampened due to high food availability.
In a companion study, we found that even at ad libitum feeders, P.
longirostris tended to consume more sucrose solution early in the day
than later in the day and that visitation rates mirrored this trend
(Gass and Garrison, 1999).
While P. longirostris responds to the levels of sucrose solution in
the environment (by increasing or decreasing its overall foraging effort),
there seem to be periods of time where it forages more actively than others
(Gass and Garrison, 1999;
Stiles, 1995;
Stiles and Wolf, 1979), even
in an environment with unlimited resources. It would be interesting to
discover whether this behavior is inflexible or if it would change over time
in an environment with constant or increasing levels of nectar availability
over the day.
Changes in patch quality
Gill (1988b) hypothesized
that traplining hummingbirds should adjust return rates to individual patches
to increase reward size. Both traplining bumblebees and territorial rufous
hummingbirds (Selasphorus rufus) increase visitation rates to more
rewarding plants (Gass and Sutherland,
1985; Thomson et al.,
1989), as P. longirostris did in this study. In the
enclosure, P. longirostris responded to steady, whole-day increases
in sucrose solution delivery rates at individual feeders by visiting the
profitable increase feeder more frequently than the unmanipulated feeders. The
increase feeder went from being the least visited feeder on control days to
the most visited feeder on the treatment days. The use of the unmanipulated
feeders decreased between the control and treatment days. This raises the
question of whether there is an upper limit to the number of visits a bird can
perform in an hour, or if the number of visits is simply a reflection of
caloric needs. The mean total number of visits per hour on the control and
increase days were similar. The results from the increase experiment indicates
that P. longirostris directs its foraging effort toward particularly
rewarding patches and that the rest of the trapline is affected by this
increasing visitation rate (with unchanged feeders visited less).
Traplining nectarivores in the wild are thought to cease visiting patches
that become unprofitable (Thomson,
1988; Thomson et al.,
1982, 1987,
1989), so that the traplines
shift in size and shape over time (days). In the decrease experiment, birds
visited the unprofitable decrease feeder slightly (but not significantly) less
often than on the control days. Because the treatment feeder was already the
least visited feeder, there may be a lower boundary on the number of visits
that the birds will pay to a feeder, even when its output is decreased. It
seems likely that the birds would have responded more strongly to a decrease
in the output of a more frequently visited feeder. It is also possible that
the birds continued to visit the decrease feeder because they needed the food
available there, regardless of the fact that it was no longer as profitable as
the other feeders.
Unfortunately, we could not determine how changes in sucrose solution delivery rates affect the entire trapline (such as changes in routes taken) because we could not distinguish bouts (and thus routes). In addition, hummingbirds are unlikely to completely stop visiting feeders in an enclosure study with closely spaced feeders. In nature, where interpatch flight distances are much greater, birds could respond to changes in patch profitability by rearranging the trapline so that one patch is visited more or less often than others, or they could visit different subsets of their total trapline on different bouts. This problem merits further investigation in the field (although it is logistically difficult to do because of the wide spacing of the patches).
Nectar loss to competition
Gill (1988b) found that in
the wild, P. longirostris responded to competition at high reward
feeders by returning more often to a feeder when there was a sudden, large
drop in reward at that feeder. Tiebout
(1993) found that C.
canivetii increased visitation rates to a feeder when faced with a
competitor by decreasing interbout time by 30%. In our enclosure study
individual birds faced differing levels of competition, depending on their
visitation rates to the feeders. Those that visited often did not lose much
sucrose solution to "competitors," whereas those that visited less
often did.
The maximum loss in 1 h faced by birds in our competition experiment was
about half that of Gill's
(1988b) competition experiment
at feeders in the wild. Even the largest losses were never more than 11% of
the total amount of sucrose solution available in the enclosure in a given
hour. Given that, it is not surprising that the birds that did not lose much
sucrose solution in a day did not respond to the treatment. Those birds that
did lose a larger amount of sucrose solution (24-77% of the total amount
available in the competition feeder in one hour) responded by increasing
relative use of the competition feeder in the following 1-2 h. After that
time, their use of the feeders returned to previous levels. It is logical that
hermits would only keep up the high visitation rates to the competition feeder
for a short time (because of high expenditures and little reward) until feeder
rewards returned to expected values. This means that hermits are flexible in
the return rates to individual patches on their trapline, at least for patches
that are close to the lek or central perch. Given the positive relationship
between relative use and amount of sucrose solution removed, it is likely that
the birds' response would have been stronger had larger amounts of food been
removed. Our methodology prevented this, as we were often
"outcompeted" by the birds. However, even though the amount of
sucrose solution we removed was not great, it is quite likely that it was
similar to (or greater than) the amount of nectar that a hummingbird would
lose in the wild to a competitor at a single patch of flowers.
In the wild, the number of competitive visits a trapliner sustains at a
patch is a direct function of how long it waits before returning to the patch
(Garrison, 1995). However, in
the wild, P. longirostris do not respond to the number of competitive
visits between their consecutive visits at patches of H. pogonantha
by increasing visitation rate with increasing number of competitive visits
(Garrison, 1995). It seems
likely that with the relatively low nectar production rates of flowers in the
wild, the relationship between the number of competitive visitors and the
amount of nectar removed may be weak. One competitor could remove as much
nectar as several, depending on how often it visited a patch. Trapliners may
simply react to competition at patches on a presence or absence basis. Because
trapliners may have no way of seeing competitors visiting patches on their
trapline, the loss of expected nectar would be the only way to detect
competition. Thus it is likely that trapliners simply respond to any sudden
large loss of nectar at a patch by returning to that patch more frequently
than they did before the loss.
Comparing the results of the decrease and competition experiments sheds some light on this subject. Birds did not respond to a decrease in sucrose solution availability at one feeder by visiting it more frequently than the others when the decrease was present from the beginning of the day. Trapliners may base their expectations of what each patch should be like on their first few foraging bouts. If the patch was profitable from the beginning of the day, the trapliner should respond to a sudden nectar loss by visiting more frequently. Regardless of the number of actual competitors, trapliners can reduce nectar loss to competitors if they return to a patch of flowers more frequently than is dictated by nectar production rates alone.
If the patch is unprofitable from the beginning (i.e., it is being visited regularly by a competitor), then the traplining forager should visit the patch less often. It may not drop the patch completely, either because some food is available at low levels, or because it is checking the patch to see if the profitability changes. A patch may suddenly become more profitable if competitors stop using it or if more flowers open. If patches are not superabundant in the environment, it would be beneficial for trapliners to sample a patch infrequently to determine if it has changed in quality. Future research could help determine the minimum profitability necessary for a patch to be included in a trapline.
Implications for natural systems
P. longirostris exhibits much of the behavior predicted for
traplining nectarivores by Feinsinger
(1976), Stiles and Wolf
(1979), Gill
(1988b), and Tiebout
(1991,
1992,
1993). Tiebout
(1991) suggested that because
individual hummingbirds often face large fluctuations in food availability,
they should be flexible in managing their energy budgets. In our study, we
found that the size of the trapline (number of feeders visited) was related to
the abundance of food in the environment. In addition, P.
longirostris increased visitation rates to individual feeders when their
profitability increased for whole days. On a shorter time scale, they
responded to sudden large losses at a feeder (due to simulated competition) by
increasing their visitation rate to the competition feeder. If and how they
track variation in individual patches of flowers along their traplines in the
wild remains unknown. How does change in patch profitability quantitatively
affect visitation rates to particular patches, and how does this affect the
rest of the trapline? Because patches are widely spaced, it would be costly to
revisit distant patches often to beat competitors. Although long-tailed hermit
hummingbirds respond as predicted to large changes at high-reward feeders both
in the laboratory (this study) and in the field
(Gill, 1988b), it is not known
how they would respond to smaller changes at multiple sites in the field.
Management of multiple sites requires that the bird respond to multiple
changes accordingly, while keeping expenditures low. It is unknown whether
P. longirostris can detect and keep track of small changes in
profitability of the multiple sites on their traplines. The changes in sucrose
solution availability to which we and Gill
(1988b) subjected long-tailed
hermit hummingbirds were most likely larger than would be encountered in
nature. We suspect that changes in nectar availability at patches would have
to be relatively large and represent a big gain or loss of nectar for these
hummingbirds to respond by altering their traplines.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
This research was funded by a Natural Sciences and Engineering Research Council research grant to C.L.G. and by a Frank M. Chapman Memorial Grant from the American Museum of Natural History and a University Graduate Fellowship from the University of British Columbia to J.S.E.G. We thank Tony Lum and Don Brandys, who built the data logger, and the students in the Gass Lab, who originally developed the prototype hummingbird lab. Lance Bailey and Alistair Blachford spent many hours helping us develop computer programs to analyze the data. Andrew Bohonak and Wesley Hochachka kindly provided statistical advice. Andrew Bohonak, Marc Mangel, and two anonymous reviewers provided helpful discussion and comments on the manuscript. Several people made our research in the tropics a little easier: Marian Sanchez and Marybel Soto arranged for our stay, permits, and various construction materials, and Pedro and Mitzi Leon were wonderful hosts to C.L.G. in San José.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Brown GS, Gass CL, 1993. Spatial association learning by hummingbirds. Anim Behav 46:487-497.
Davies NB, Houston AI, 1981. Owners and satellites: the economics of territory defense in the Pied Wagtail, Motacilla alba. J Anim Ecol 50:157-180.
Feinsinger P, 1976. Organization of a tropical guild of nectarivorous birds. Ecol Monogr 46:257-291.
Feinsinger P, 1987. Approaches to nectarivore-plant interactions in the New World. Revista Chile Hist Nat 60:285-319.
Feinsinger P, Busby W, Murray KG, Beach JH, Pounds WZ, Linhart YB,1988 . Mixed support for spatial heterogeneity in species interactions: hummingbirds in a tropical disturbance mosaic. Am Nat 131:33-57.[ISI]
Feinsinger P, Chaplin SB, 1975. On the relationship between wing disc loading and foraging strategy in hummingbirds. Am Nat 109:217-224.
Feinsinger P, Colwell RK, 1978. Community organization among Neotropical nectar-feeding birds. Am Zool 18:779-795.
Frost AK, Frost PJH, 1980. Territoriality and changes in resource use by sunbirds at Leonotis leonurus (Labiatae).Oecologia 45:109-116.
Garrison JSE, 1995. Trapline foraging behavior in a tropical hummingbird species Phaethornis superciliosus (MSc dissertation). Vancouver: University of British Columbia.
Gass CL, 1978a. Experimental studies of foraging in complex laboratory environments. Am Zool 18:617-626.
Gass CL, 1978b. Rufous hummingbird feeding territoriality in a suboptimal habitat. Can J Zool 56:1535-1539.
Gass CL, 1979. Territory regulation, tenure, and migration in rufous hummingbirds. Can J Zool 57:914-923.
Gass CL, Garrison JSE, 1999. Energy regulation by traplining hummingbirds. Funct Ecol 13:483-492.
Gass CL, Montgomerie RD, 1981. Hummingbird foraging behavior: decision-making and energy regulation. In: Foraging behavior: ecological, ethological, and psychological approaches (Kamil AC, Sargent TD, eds). New York: Garland STPM Press;157 -194.
Gass CL, Romich M, Suarez RK, 1999. Energetics of hummingbird foraging at low ambient temperature. Can J Zool 77:314-320.
Gass CL, Sutherland GD, 1985. Specialization by territorial hummingbirds on experimentally enriched patches of flowers: energetic profitability and learning. Can J Zool 63:2125-2133.
Gill FB, 1988a. Effects of nectar removal on nectar accumulation in flowers of Heliconia imbricata (Heliconiaceae).Biotropica 20(2):169-171.
Gill FB, 1988b. Trapline foraging by hermit hummingbirds: competition for an undefended, renewable resource.Ecology 69:1933-1942.
Gill FB, Wolf LL, 1977. Nonrandom foraging by sunbirds in a patchy environment. Ecology 58:1284-1296.
Hainsworth FR, 1981. Energy regulation in hummingbirds. Amer Sci 69:420-429.
Hainsworth FR, Tardiff M, Wolf LL, 1981. Proportional control for daily energy regulation in hummingbirds. Physiol Zool 54:452-462.
Hainsworth FR, Wolf LL, 1972. Crop volume, nectar concentration and hummingbird energetics. Comp Biochem Physiol 42A:359-366.
Hinkelmann C, 1996. Systematics and geographic variation in long-tailed hermit hummingbirds, the Phaethornis superciliosus-malarislongirostris species group (Trochilidae), with notes on their biogeography. Ornithol Neotrop 7(2):119-148.
Hixon MA, Carpenter FL, Paton DC, 1983. Territory area, flower density, and time budgeting in hummingbirds: an experimental and theoretical analysis. Am Nat 122:366-391.
Janzen DH, 1971. Euglossine bees as long-distance
pollinators of tropical plants. Science
171:203-205.
King JR, 1974. Seasonal allocation of time and energy resources in birds. In: Avian energetics, no. 15 (Paynter RA Jr, ed). Cambridge: Nuttall Ornithological Club;4 -79.
Paton CC, Carpenter FL, 1984. Peripheral foraging by territorial rufous hummingbirds: defense by exploitation.Ecology 65:1808-1819.
Possingham HP, 1989. The distribution and abundance of resources encountered by a forager. Am Nat 133:42-60.
Schoener TW, 1971. Theory of feeding strategies.Annu Rev Ecol Syst 2:369-404.
Slud P, 1964. The birds of Costa Rica. Bull Am Mus Nat Hist 138:1-430.
Stewart-Oaten A, 1995. Rules and judgments in statistics: three examples. Ecology 76:2001-2009.
Stiles FG, 1975. Ecology, flowering phenology, and hummingbird pollination of some Costa Rican Heliconia species.Ecology 56:285-301.[ISI]
Stiles FG, 1981. Geographical aspects of bird-flower coevolution, with particular reference to Central America. Ann Missouri Bot Gard 68:323-351.
Stiles FG, 1995. Behavioral, ecological and morphological correlates of foraging for arthropods by the hummingbirds of a tropical wet forest. Condor 97:853-878.
Stiles FG, Freeman CE, 1993. Patterns in floral nectar characteristics of some bird-visited plant species from Costa Rica.Biotropica 25:191-205.
Stiles FG, Wolf LL, 1979. Ecology and evolution of lek mating behavior in the long-tailed hermit hummingbird. Ornithol Monogr 27:1-78.
Sutherland GD, Gass CL, Thompson PA, Lertzman KP,1982 . Feeding territoriality in migrant rufous hummingbirds: defense of yellow-bellied sapsucker (Sphyrapicus varius) feeding sites. Can J Zool 60:2046-2050.
Thomson JD, 1988. Effects of variation in inflorescence size and floral rewards on the visitation rates of traplining pollinators of Aralia hispida. Evol Ecol 2:65-76.
Thomson JD, 1996. Trapline foraging by bumblebees: I.
Persistence of flight-path geometry. Behav Ecol
7:158-164.
Thomson JD, Maddison WP, Plowright RC, 1982. Behavior of bumble bee pollinators of Aralia hispida Vent. (Araliaceae).Oecologia 54:326-336.
Thomson JD, McKenna MA, Cruzan MB, 1989. Temporal patterns of nectar and pollen production in Aralia hispida: implications for reproductive success. Ecology 70:1061-1068.
Thomson JD, Peterson SC, Harder LD, 1987. Response of traplining bumble bees to competition experiments: shifts in feeding location and efficiency. Oecologia 71:295-300.
Thomson JD, Slatkin M, Thomson BA, 1997. Trapline
foraging by bumble bees: II. Definition and detection from sequence data.Behav Ecol
8:199-210.
Tiebout HM, 1991. Daytime energy management by tropical hummingbirds: responses to foraging constraint.Ecology 72:839-851.
Tiebout HM, 1992. Comparative energetics of divergent foraging modes: a doubly labeled water experiment on hummingbird competition.Anim Behav 44:895-906.
Tiebout HM, 1993. Mechanisms of competition in tropical hummingbirds: metabolic costs f