Behavioral Ecology Vol. 10 No. 3: 220-226
© 1999 International Society for Behavioral Ecology
Avian prey-dropping behavior. II. American crows and walnuts
a Department of Biology, The College of William and Mary, Williamsburg, VA 23187-8795, USA, b Department of Zoology, Eastern Illinois University, Charleston, IL 61920, USA
Address correspondence to D.A. Cristol. E-mail: dacris{at}facstaff.wm.edu
Received 30 September 1997; accepted 12 September 1998.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONCHARACTERIZING WALNUTSSTUDY SITES AND GENERAL...CHARACTERIZING PREY LOSSTESTING PREDICTIONS OF THE...DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Complex and energetically expensive foraging tasks should be shaped by natural selection to be efficient. Many species of birds open hard-shelled prey by dropping the prey repeatedly onto the ground from considerable heights. Urban-dwelling American crows (Corvus brachyrhynchos) forage in this way on two species of walnuts in central California, USA. As predicted from a theoretical model, crows dropped nuts with harder shells from greater heights and dropped them from greater heights when over softer substrates. The height selected for dropping nuts decreased in the presence of numerous nearby conspecifics, indicating that crows were sensitive to the risk of kleptoparasitism when selecting drop heights. Drop height decreased with repeated drops of the same walnut, suggesting that crows adjusted for the increasing likelihood that a repeatedly-dropped nut would break on subsequent drops. Crows did not alter height of drop in accordance with differences in the mass of the prey. When faced with multiple prey types and dropping substrates, and high rates of attempted kleptoparasitism, crows adjusted the height from which they dropped nuts in ways that decreased the likelihood of kleptoparasitism and increased the energy obtained from each nut.
Key words: American crow, Corvus brachyrhynchos, foraging behavior, Juglans spp., kleptoparasitism, prey dropping, walnuts.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERIZING WALNUTSSTUDY SITES AND GENERAL...CHARACTERIZING PREY LOSSTESTING PREDICTIONS OF THE...DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Birds that break open hard-shelled prey items by dropping them repeatedly onto the ground have attracted scientific attention for at least a century (e.g., Saunders, 1891
). This
foraging behavior is more complex than most because it requires repeated
flights with a prey item, choice of drop sites and drop heights, and retrieval
of prey. Such a demanding task is ideal for studying how closely an animal's
behaviors are adapted to the rigors of its environment. Because avian
prey-dropping occurs in at least three distantly-related orders of birds
(Table 1), this behavior has
presumably evolved a number of times. The occurrence of avian prey-dropping on
nearly every continent, in a wide-range of habitats, and with many different
types of prey suggests the potential for interesting adaptive modifications
(Table 1). Finally, the
occurrence of prey dropping among both highly gregarious and essentially
solitary species raises the question of whether, when, and how this complex
behavior is learned.
|
Much of the early interest in avian prey-dropping behavior arose from
debate over whether some birds (crows) selected efficient hard surfaces for
dropping, whereas others (gulls) were less "intelligent" because
they reportedly failed to discriminate between hard and soft substrates
(Oldham, 1930;
Tinbergen, 1960). More
recently authors have attempted to understand avian prey-dropping in the
context of optimal foraging (e.g., Barash
et al., 1975; Beck,
1980; Maron, 1982;
Siegfried, 1977;
Whiteley et al., 1990;
Zach, 1978,
1979). In our companion paper
(Switzer and Cristol, 1999) we
presented a general theoretical model that identifies important variables that
should shape all avian prey-dropping systems. The model suggests that the
height from which hard-shelled prey will be dropped is influenced by the
hardness of the prey item, the hardness of the substrate onto which it is
dropped, the mass of the prey item, the potential weakening of the shell
during repeated drops, and the likelihood of loss after dropping. The
objective of the present study was to test the qualitative predictions of our
model on a previously undescribed avian preydropping system.
American crows, Corvus brachyrhynchos, which drop two species of
walnuts (English, Juglans regia, and northern California black,
J. hindsii), presented an ideal opportunity to test our model. We
first gathered background data on the heights from which we had to drop each
type of walnut until it was cracked sufficiently to be consumed by a crow and
data on the effects of different substrates, nut mass, and repeated impacts on
the breakability of the walnuts we dropped. We also determined, by observing
foraging crows, whether the probability of a crow losing its walnut was
dependent on the height from which the nut was dropped. These data allowed us
to choose a specific set of predictions from our general theoretical model
that could be applied to the crow—walnut system
(Switzer and Cristol, 1999).
To test these predictions we gathered data on the heights from which crows
dropped walnuts in situations where important factors could be varied. For
example, we observed crows feeding simultaneously on two walnut species that
differed markedly in the ease with which they could be broken, thereby
allowing us to test predictions about the effects of prey hardness on the
height selected for dropping.
A central prediction of our model was that the presence of prey loss should
have dramatic consequences on dropping behavior
(Switzer and Cristol, 1999).
Although prey loss and theft are mentioned in most previous studies of prey
dropping (e.g., Barash et al.,
1975; Siegfried,
1977; Whiteley et al.,
1990; Zach, 1979)
few quantitative data are available. In the present study, crows frequently
stole walnuts dropped by other crows, and dropping occurred under varying
degrees of kleptoparasitism intensity. Thus, we were able to examine, for the
first time, the effects of prey loss on this complex foraging behavior.
|
CHARACTERIZING WALNUTS |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERIZING WALNUTSSTUDY SITES AND GENERAL...CHARACTERIZING PREY LOSSTESTING PREDICTIONS OF THE...DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Methods
Before making specific predictions using the model presented in our companion paper, it was necessary to understand the effects of height, prey type, substrate hardness, repeated drops, and prey mass on the probability of a prey item breaking open on a given drop (Switzer and Cristol, 1999
).
We determined these characteristics for walnuts by dropping them from an
exterior stairway at various heights. To determine the effect of height on the probability of a dropped walnut breaking, we dropped 20 English walnuts onto asphalt at each of the following heights: 1.55, 3.1, 6.2, and 9.3 m. Each nut was dropped repeatedly from the same height until its shell had cracked open sufficiently to be consumed by a crow (throughout the paper "cracked" refers to a hole of at least 1 cm2). The same procedure was repeated with black walnuts from 3.1, 6.2, and 9.3 m. A two-factor (species, height) ANOVA was used to compare the number of drops required to break English and black walnuts dropped from these three heights.
To determine the effect of substrate hardness on the probability of a dropped walnut breaking, English walnuts were dropped from 6.2 m and 9.3 m into a container of hard-packed soil from an agricultural field used by crows for nut dropping. These were compared with English walnuts dropped from the same heights onto asphalt. Nuts that had not broken after 20 drops were assigned a value as if they had broken on the next drop. We used a nonparametric Mann-Whitney U test to compare the breakability of all English walnuts dropped on each of the two substrates because variance was distorted by the fact that all nuts dropped on soil from 6.2 m were scored as having broken after 21 drops.
To determine whether walnuts weakened with repeated drops, 100 English
walnuts weighing from 6-17 g were dropped repeatedly from 3.1 m until cracked.
So that we could compare our results to those of a previous study on dropped
mollusks (Zach, 1979), we
plotted the percentage of nuts broken in each batch of redropped nuts. With
arcsine-transformed data from our walnuts and Zach's mollusks
(1979), we used linear
regression to determine whether either relationship was positive. A positive
slope would indicate weakening with successive drops. This was repeated for
100 black walnuts of similar weights dropped from 10.85 m, which is a
comparable height in terms of the number of drops required to crack these
harder nuts. Black walnuts exhibited a similar pattern to English walnuts
(data not shown), but because we have insufficient data to examine the
behavior of crows successively dropping black walnuts (see below), there was
no reason to include black walnuts in this analysis. The same black and
English walnut data were used to determine the relationship between number of
experimental drops required to break a walnut and its mass.
Results
By dropping nuts experimentally, we found that the number of drops required
to crack a walnut decreased with height of drop, regardless of species
(Fheight = 27.7, df = 2, p <.0001;
Figure 1). English walnuts
broke more easily than black walnuts, regardless of height
(Fspecies = 119.8, df = 1, p <.0001;
Figure 1). There was also a
significant interaction between species and height
(Finteraction = 19.7, df = 2, p <.0001),
reflecting the fact that further increases in height, beyond the height at
which all were broken, could have no effect on the likelihood of an English
walnut breaking. Thus, walnuts are more likely to break when dropped from
higher heights, and black walnuts require more drops to break than do English
walnuts.
|
English walnuts dropped experimentally onto asphalt broke in fewer drops than those dropped onto soil (asphalt: n = 40 walnuts, mean = 1.3 ± 0.5 drops; soil: n = 40 walnuts, mean = 13.9 ± 7.6 drops; z = 7.2, p <.0001; Figure 1), indicating that asphalt is a more effective substrate for breaking nuts than agricultural soil.
English walnuts tended to be more likely to break after repeated drops
(Figure 2). By way of
comparison, the mollusks studied by Zach
(1979) did not become
increasingly likely to break after repeated drops.
|
The relationship between mass and the number of drops required to break English and black walnuts was weakly but significantly negative (English: y = -0.2x - 5.2, r2 =.10, df = 78, F = 8.7, p =.004; black: y = -1.0x + 15.5, r2 =.419, df = 96, F = 22.3, p <.0001). In other words, heavier nuts were slightly easier to break, but large sample sizes were required to detect the effect.
|
STUDY SITES AND GENERAL METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERIZING WALNUTSSTUDY SITES AND GENERAL...CHARACTERIZING PREY LOSSTESTING PREDICTIONS OF THE...DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We observed crows foraging on nuts in English and black walnut trees in urban areas of Davis, Yolo County, California, USA. From 12 September to 1 November, 1995 crows were observed almost daily at our primary location, Birch Lane, a low-traffic residential street with 25 English walnut trees. Occasionally we observed crows dropping English walnuts from Birch Lane onto the dried soil of nearby agricultural fields. We also observed crows dropping black walnuts 3-9 January 1996 at a second site, Russell Boulevard, a high-traffic road located 5 km from Birch Lane with approximately 100 large black walnut trees.
All English walnuts had fallen from the trees and been removed by birds, squirrels, and humans by mid-November, and all black walnuts were gone by mid-January. Therefore, we established a new study site, on the campus of University of California-Davis, where we provisioned free-living crows with walnuts nearly daily from 16 January to 11 March 1996. Our objective at the campus site was to determine experimentally the drop heights of English and black walnuts of known mass (to 0.1 g) dropped at the same site under identical conditions.
We recorded the following data for each instance of a crow dropping an English walnut at Birch Lane: height (estimated by comparison with previously measured trees, utility poles, and other objects); distance each nut rolled; number of times each nut was dropped; number of crows present within 100 m; ground distance from where the crow dropped the nut to the nearest other crow; and response of other crows (kleptoparasitize, attempt to kleptoparasitize, or no response, all within 30 s). In nearby agricultural fields, and at Russell Boulevard, we recorded only the heights from which walnuts were dropped and the number of drops for each nut. At the campus site walnuts were presented singly or in pairs, and each crow taking a provisioned walnut was followed until the nut was either eaten or stolen or until the bird disappeared from view.
All provisioned walnuts were gathered in October 1995 from single trees of
each species, except for English walnuts >17.5 g, which were gathered from
a different tree. At Birch Lane and Russell Boulevard crows found naturally
occurring nuts and removed the soft hulls before dropping, whereas at the
campus site all nuts were provisioned with hulls already removed. Throughout
the study all observations were made during the first 4 h after dawn. Crows
were unmarked, but the rapid turnover of flocks and large number of
individuals present simultaneously at the study sites (100-1000 individuals
daily), the large number of crows roosting in the vicinity (approximately
10,000; Gorenzel P, personal communication), and an analysis of the
variability in dropping behavior of 21 recognizable individuals (Cristol and
Switzer, unpublished data; see Leger and Didrichsons,
1994) indicate that our
conclusions are robust despite the occasional inclusion of some multiple
observations from the same individuals. Except when considering the effects of
repeated drops, all data were from the first-time drops only. Throughout,
unpaired t tests, simple linear regression, and ANOVA were used
unless otherwise noted.
|
CHARACTERIZING PREY LOSS |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERIZING WALNUTSSTUDY SITES AND GENERAL...CHARACTERIZING PREY LOSSTESTING PREDICTIONS OF THE...DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Methods
An essential characteristic of any avian prey-dropping system is the likelihood of losing prey before it can be retrieved, such as through kleptoparasitism. Prey loss, if it occurs, might be constant with regard to drop height, or it might increase when a bird drops prey from greater heights. We examined our data from crows feeding on naturally occurring English walnuts at Birch Lane to determine both the prevalence of walnut loss in this system and whether the probability of loss through kleptoparasitism was dependent on height of drop. To determine whether height of drop affected the likelihood of kleptoparasitism, we used logistic regression (SAS JMP 3.1) with "kleptoparasitized" or "not kleptoparasitized" as the dependent variable, and drop height, distance to nearest crow, number of crows present, and distance the nut rolled as independent variables. Because attempted kleptoparasitism was much more frequent than actual kleptoparasitism, we performed an alternative analysis with "attempted kleptoparasitism" as the dependent variable.
Results
Prey loss occurred almost exclusively through kleptoparasitism (data not
shown), with nearly half of all first-time drops resulting in attempted
kleptoparasitism, and 12% of all dropped nuts being stolen. A significant
portion of the uncertainty in likelihood of kleptoparasitism was explained by
the factors included in our logistic regression model (df = 4, likelihood
ratio 
2 = 20.2, p <.0005). Height from which a
walnut was dropped had the greatest effect on the model (df = 1, Wald
2 = 6.9, p =.009). Our alternative logistic
regression model explained more of the uncertainty in the likelihood of
attempted kleptoparasitism (df = 4, likelihood ratio 2 = 58.9,
p <.0001), and the drop height again had the greatest effect on
the model (df = 1, Wald 2 = 7.3, p =.007). Thus, the
likelihood of prey loss through kleptoparasitism was related to the height of
drop for English walnuts.
|
TESTING PREDICTIONS OF THE MODEL |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERIZING WALNUTSSTUDY SITES AND GENERAL...CHARACTERIZING PREY LOSSTESTING PREDICTIONS OF THE...DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Having determined the relevant prey characteristics and details of kleptoparasitism for this previously undescribed avian prey-dropping system, we tested five key predictions generated by our model (Switzer and Cristol, 1999
).
Prediction 1: prey breakability
Because English walnuts break more easily than black walnuts, crows should
drop English walnuts from less high than they do black walnuts.
Methods. We compared the heights from which crows dropped each walnut species when they were foraging on naturally occurring English walnuts at Birch Lane and black walnuts at Russell Boulevard. We also compared drops of the same two nut species when they were provisioned simultaneously at the campus site where birds dropped them under identical conditions.
Results. The mean height for 311 first-time drops at Birch Lane was 3.41 ± 2.17 m, with a maximum recorded drop of 10.85 m. Crows dropping black walnuts at Russell Boulevard dropped them from significantly greater heights than crows dropping English walnuts at Birch Lane (n = 34, mean = 7.01 ± 3.18, df = 343, t = 5.7, p <.0001). Where both walnut species were dropped together, at the campus site, crows also dropped the black walnuts from significantly greater heights than they did English walnuts (English: n = 153, = 4.66 ± 3.02 m; black: n = 72, = 5.84 ± 3.15 m, df = 223, t = 2.7, p =.0075). Because English walnuts varied in mass more than black walnuts, we also compared drop height for only those English walnuts within the range of weights represented by black walnuts (6.7-13.7 g) and found a similar result (English: n = 95, = 4.83 ± 3.48 m; df = 165, t = 2.3, p =.02). Crows dropped the harder-to-break black walnuts from greater heights than they did English walnuts, supporting this prediction of our model.
Prediction 2: substrate hardness
Because English walnuts break more easily on asphalt than on soil, crows
should drop them from greater heights when they are over soil. As a corollary
to this prediction, crows should select paved surfaces in preference to soil
when both are available.
Methods. We compared heights from which crows dropped naturally occurring English walnuts on paved surfaces at Birch Lane and on agricultural fields nearby. To determine whether crows preferentially dropped English walnuts onto hard surfaces when both paved and softer substrates were available, we observed an area of 771 m2 at Birch Lane and recorded the surface onto which the first 50 walnuts were dropped by crows. The measured site was 41.7% asphalt, 20.0% concrete, and 38.3% lawn. We recorded only the first drop for each nut.
Results. Crows that dropped walnuts onto the soil of agricultural
fields did so from significantly greater heights than those dropping onto the
asphalt or concrete of Birch Lane (n = 36, mean = 8.47 ± 6.72
m, df = 345, t = 9.5, p <.0001). When dropping in an area
where we had measured substrate availability, crows dropped 48 English walnuts
onto asphalt or concrete and only 2 onto lawn during our observations
(2 = 24.8, p =.001, df = 1). As predicted, crows
dropped walnuts from greater heights on the softer substrate and selected
paved surfaces as dropping substrates significantly more often than expected
by chance.
Prediction 3: height of successive drops
Because the probability of a walnut breaking increases with each additional
drop, crows should reduce the height of successive drops of the same nut.
Methods. To determine whether height of drop changed during series
of repeated drops, we compared mean heights of first, second, third, fourth,
and all other drops combined (fifth through seventeenth). Because our
independent variable is not on an interval scale, we used a nonparametric
Kruskal-Wallis test with multiple comparisons
(Siegel and Castellan, 1988).
To determine whether any relationship between drop height and successive drop
number might be affected by a third variable, attempted kleptoparasitism, we
also compared the percentage of first, second, third, fourth, and all other
drops that resulted in attempted kleptoparasitism using the same statistical
test described above.
Results. At Birch Lane 11.9% of 337 English walnuts dropped once were dropped again by the same crow (mean = 3.42 ± 1.87 drops). The maximum number of repeated drops observed at Birch Lane was 17, although at other sites crows dropped English walnuts up to 50 times before cracking them (data not shown). As predicted by our model, drop height decreased significantly with successive drops (Figure 3). Attempted kleptoparasitism did not change with successive drops (df = 4, Kruskal-Wallis H = 4.0, p =.41), indicating that this variable is unlikely to explain the relationship between drop height and successive drop number.
|
Prediction 4: prey mass and breakability
Because heavy walnuts break more easily than light ones, crows should drop
heavier walnuts from lower heights than lighter ones.
Methods. Using data from the provisioned walnuts of known mass dropped at the campus site, we determined whether there was a significant relationship between nut mass and height of drop using linear regression.
Results. Contrary to our prediction, there was no relationship between mass and drop height for provisioned English walnuts (y = 11.38 + 0.28x, r2 <.01, df = 151, F = 1.0, p =.32). Further, heavy black walnuts were dropped from significantly greater heights than light ones, although mass explained only a small percentage of the variation in drop height (y = -0.65 + 1.81x, r2 <.08, df = 70, F = 5.6, p =.02). Thus, crows did not drop heavier walnuts from lower heights, failing to support the prediction of the model.
Prediction 5: prey loss
Because the probability of prey loss through kleptoparasitism by other
crows increases with drop height, crows should decrease drop height when there
is a greater risk of kleptoparasitism.
Methods. To determine whether crows adjusted drop height with
respect to potential kleptoparasitism, we tested for a correlation between
drop height and an index of "kleptoparasite intensity," using the
data from all first drops of English walnuts at the Birch Lane site. Our index
of kleptoparasite intensity was the number of crows present divided by the
distance to the nearest crow. We do not know what measure crows might use in
assessing risk of kleptoparasitism, and complicating factors such as social
status could make assessment difficult. In the absence of such information, we
used a simple index because it combined the two pieces of information that we,
and presumably the crows, could easily measure during the rapid-paced events
of a walnut drop. Our index of kleptoparasite intensity proved to be an
excellent predictor of the probability of kleptoparasitism and/or attempted
kleptoparasitism. To determine this we classified each drop as being one of
eight ordinal categories of increasing kleptoparasite intensity (levels chosen
to equalize sample sizes in each category), and then we performed a
nonparametric test of association between this index and the observed levels
of kleptoparasitism and/or attempted kleptoparasitism (Spearman 
= 0.98,
p =.01).
We used two alternative analyses to determine whether drop height was related to risk of kleptoparasitism: (1) we carried out a nonparametric test of association between the index of kleptoparasite intensity and height of each drop, and (2) we used linear regression after log transforming both the drop height and kleptoparasitism intensity index data so that they closely approximated normal distributions.
Results. As predicted, there was a significant negative relationship between drop height and kleptoparasite intensity, indicating that crows dropped walnuts from greater heights when the potential for kleptoparasitism was lower. Both the nonparametric analysis on the untransformed data (Figure 4a) and linear regression on log-transformed data indicated significant relationships (Figure 4b).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERIZING WALNUTSSTUDY SITES AND GENERAL...CHARACTERIZING PREY LOSSTESTING PREDICTIONS OF THE...DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Prey breakability
Our model predicted that birds should drop harder-to-break prey from greater heights, a prediction best tested with a single species that drops multiple prey types. Crows chose significantly greater heights when dropping black walnuts, which are harder than English walnuts. This is the first demonstration that birds adjust drop height according to hardness of prey and was true for naturally occurring English and black walnuts dropped at separate sites and for experimental data from provisioned walnuts dropped simultaneously at a single site.
Substrate hardness
For the same reason that birds should drop harder prey from greater
heights, our model predicted that they should drop prey from lower heights
when they are over harder substrates. When dropping English walnuts onto
harder, paved surfaces crows dropped prey from lower heights than when
dropping the same walnuts over the softer substrate of nearby agricultural
fields. Thus, as predicted by our model, crows adjusted the height of their
drops according to substrate hardness, consistent with earlier reports of
crows and gulls dropping mollusks on different substrates
(Siegfried, 1977;
Ward, 1991;
Whiteley et al., 1990)
Numerous reports of birds selecting distant or traditional drop sites, such
as ossuaries used by lammeregeiers (Gypaetus barbatus;
Huxley and Nicholson, 1963),
are highly suggestive of substrate selectivity by most prey-dropping birds,
but the issue of substrate selectivity has not been satisfactorily resolved
because researchers have consistently failed to report the available area of
each type of substrate (e.g., Barash et
al., 1975; Beck,
1980). At a site where we had measured relative substrate
availability, American crows selected paved areas as drop sites significantly
more frequently than if they had been dropping without regard to substrate.
Whether there is great variability between species in the ability to select
effective dropping substrates, as suggested by Tinbergen
(1960), remains to be
determined from comparative studies in which substrate availability is
quantified.
Height of successive drops
Our model predicted that a foraging bird should adjust drop height
according to the probability that the prey item will break on that drop. Some
prey items, such as certain mollusk species, have the same chance of breaking
regardless of how many times they have been dropped. Crows dropping those
mollusks have been reported to drop them from constant or slightly increasing
heights on successive drops (Whiteley et
al., 1990; Zach,
1978). By dropping English walnuts repeatedly until they broke, we
demonstrated experimentally that, in contrast to many mollusks, walnuts have
an increasing probability of breaking with repeated impacts. As predicted by
the model, American crows dropped English walnuts from lower heights on
successive drops. Thus, our model successfully predicted a behavioral
difference between two avian prey-dropping systems based on a specific
characteristic of the prey—change in breakability on successive
drops.
Prey mass and breakability
For cases in which prey breakability reliably increases with prey weight,
our model predicted that birds would adjust drop height and drop
harder-to-break prey from greater heights. For both species of walnuts there
was a significant but weak positive relationship between prey mass and
breakability. Because heavier walnuts break more easily crows should drop them
from lower heights than lighter walnuts. Crows apparently adjusted height of
drop according to prey mass when dropping black, but not English, walnuts.
However, heavier black walnuts, which break more easily, were dropped from
greater heights than light ones, opposite of what was predicted by our model.
This lack of support for the model's prediction could be the result of crows
ignoring walnut mass because it is a poor predictor of breakability or crows
failing to accurately assess walnut mass. Another possibility is that crows
chose greater drop heights for heavier black walnuts because the greater
energetic value of the larger nuts outweighed the potential energetic savings
resulting from their slightly greater breakability
(Switzer and Cristol, 1999).
Little is known about the ability of corvids to assess prey mass per se, so
without further study we cannot distinguish between these explanations
(references in Vander Wall,
1990).
Prey loss
Little attention has been paid to the role of prey loss in avian
prey-dropping systems. Given that kleptoparasitism was the main form of prey
loss for American crows dropping walnuts and that its likelihood increased
with drop height, our model predicted that drop height should increase when
potential kleptoparasitism was lower. Using an index that combined the number
and proximity of conspecifics, we showed that crows dropped walnuts from
greater heights as the likelihood of kleptoparasitism decreased, supporting a
key prediction of the model. Although the negative relationship between drop
height and kleptoparasitism risk was significant, the nonparametric analysis
shown in Figure 4a suggests
that there may be a threshold of kleptoparasite intensity above which drop
height is reduced, rather than the linear response suggested by the parametric
analysis (illustrated in Figure
4b). Further work is necessary to uncover the means by which crows
assess risk of kleptoparasitism and to reveal the details of their behavioral
response.
Kleptoparasitism is of varying importance in other avianprey dropping
systems. It is apparently unimportant for territorial northwestern crows
dropping whelks (Zach, 1978),
while other studies mention that theft occurs (e.g., Siegfried,
1977), or even suggest the
possibility of territorial defense of drop sites against kleptoparasites
(Barash et al., 1975). Only one
previous study has presented any data addressing this question. Whiteley et
al. (1990) found that the
average height from which mussels were dropped when there were more than five
crows and/or gulls within 8 m of a dropping crow was 52% lower than the height
used when no potential kleptoparasites were present, but no other data on
kleptoparasitism were reported. There are other data in the literature
suggesting that birds adjust drop height to account for probability of
kleptoparasitism, but they have not been interpreted in this light by the
authors. Gulls tended to drop prey from approximately 60% lower heights when
at a more crowded site, but Siegfried
(1977) attributed this to the
fact that the more crowded site had a harder substrate. Beck
(1980) reported that the mean
drop height for gulls using a rock wall with many other gulls was 34% lower
than that for a parking lot with fewer gulls, but his explanation was that the
narrow wall required better aim. Many studies mention kleptoparasitism and
other forms of prey loss (e.g., lost under water or lost among broken shells),
but ours is the first in which the likelihood of prey loss and its
relationship with drop height has been quantified.
Conclusions
American crows foraging on walnuts presented an opportunity to test our
model of prey-dropping behavior on a previously undescribed system. Because
crows dropped two species of walnuts that differed substantially in hardness,
and they dropped prey onto both paved and unpaved substrates, we had the
opportunity to test several predictions of our model within the same avian
prey-dropping system. Because birds drop many types of hard-shelled prey
(Table 1), the present study
also allowed us to compare the crow—walnut system to other examples of
avian prey-dropping. Walnuts differ from commonly dropped mollusks in two ways
that affect the predictions of our model: walnuts are more likely to break
with each successive drop, and heavier walnuts break more easily than lighter
ones.
Our model predicted that if birds maximize the energy obtained from each dropped food item, they should take into account the intensity of kleptoparasitism, the likelihood of each item breaking given its hardness and previous weakening, and the hardness of the substrate. American crows adjusted the height from which they dropped walnuts as our model predicted, if they were accounting for each of these factors. Crows did not behave as predicted with regard to the mass of a walnut. The mass of each walnut may be harder to assess quickly than a constant or gradually changing characteristic such as substrate hardness or intensity of kleptoparastism. Overall, we were successful in predicting many qualitative aspects of walnut-dropping behavior by American crows. This suggests that these crows have evolved or learned to maximize the energy obtained from each dropped walnut. Thus, American crows feeding on walnuts, and avian prey-dropping in general, provides an excellent system for future studies of the ontogeny of a complex behavior and the process of prey assessment.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERIZING WALNUTSSTUDY SITES AND GENERAL...CHARACTERIZING PREY LOSSTESTING PREDICTIONS OF THE...DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We thank A. Ashton, A. Carmichael, I. Cristol, M. Hershdorfer, K. Johnson, C. Kondzela, K. Radasky, R. Reimers, I. Switzer, J. Switzer, L. Tu, L. Walke, T. Weber, C. Weightman, and especially J. Stamps for their help with this research. Partial support during the development of this research was provided by the Faculty Research Committee at the College of William and Mary and National Science Foundation DIR 9113287 from the University of California-Davis to D.A.C. and by Eastern Illinois University and M. Mangel to P.V.S.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERIZING WALNUTSSTUDY SITES AND GENERAL...CHARACTERIZING PREY LOSSTESTING PREDICTIONS OF THE...DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Barash DP, Donovan P, Myrick R, 1975. Clam dropping behavior of the glaucous-winged gull (Larus glaucescens).Wilson Bull 87:60-64.
Beck BB, 1980. Animal tool behavior: The use and manufacture of tools by animals. New York: Garland STMP;199 -214.
Berrow SD, Kelly TC, Myers AA, 1992. The diet of coastal breeding hooded crows Corvus corone cornix.Ecography 15:337-346.
Bindner CM, Jr, 1968. Bald eagles use tools.Flor Nat 41:169.
Boswall J, 1977. Further notes on tool-using by birds and related behaviour. Avic Mag 84:162-166.
Boswall J, 1983. Tool-using by birds and related behaviour: yet more notes. Avic Mag 89:170-181.
Cadée GC, 1989. Size-selective transport of shells by birds and its paleoecological implications. Palaeontology 32:429-437.
Cavanagh PM, 1992. Application of prey-dropping behaviour to a spiny dogfish squalus acanthias by a herring gull Larus argentatus. Mar Ornithol 20:91-92.
Conner J, 1993. The shell game. Living Bird 12:6-7.
Drennen DJ, 1995. Ring-billed gulls (Larus delawarensis) feeding on mussels. Alab Bird Life 41:7-9.
Förstel A, 1993. Rabenkrähe Corvus corone öffnet Walnuss Juglans regia durch Fallenlassen. Ornithologisc. Anzeiger 32:165.
Gibson C, 1992. Jackdaws feeding on horse chestnuts.Br Birds 85:138.
Goodwin D, 1986. Crows of the world. London: British Museum (Natural History).
Grobecker DB, Pietsch TW, 1978. Crows use automobiles as nutcrackers. Auk 95:760-761.
Harber DD, Johns M, 1947. Great black-backed gull dropping rat. Br Birds 40:317.
Hartley, PHT, 1967. Feeding habits. In: A new dictionary of birds (Thomson AL, ed). New York: McGraw-Hill;287 -288.
Heinrich B, 1989. Ravens in winter. New York: Summit.
Hunt GR, 1996. Manufacture and use of hook-tools by New Calendonian crows. Nature 379:249-251.
Huxley J, Nicholson EM, 1963. Lammergeier Gypaetus barbatus breaking bones. Ibis 105:106-107.
Ingolfsson A, Estrella BT, 1978. The development of shell-cracking behavior in herring gulls. Auk 95:577-579.
James PC, Verbeek NAM, 1984. Temporal and energetic aspects of food storage in northwestern crows. Ardea 72:207-215.
Kent BW, 1981. Prey dropped by herring gulls (Larus argentatus) on soft sediments. Auk 98:350-354.
Leger DW, Didrichsons IA, 1994. An assessment of data pooling and some alternatives. Anim Behav 48:823-832.
Lorek G, Oleksik I, 1992. Hooded Crows (Corvus corone cornix) break nuts by throwing them at the ground. Notaki Ornithol 33:331-332.
Maple T, 1974. Do crows use automobiles as nutcrackers? West Birds 5:97-98.
Maron JL, 1982. Shell-dropping behavior of Western Gulls (Larus occidentalis). Auk 99:565-569.
Meinis HK, 1993. A change in the nut-cracking technique used by hooded crows in Kibbutz Netzer Sereni, Israel.Ornithol Soc Mid East Bull 30:29.
Nihei Y, 1995. Variations of behaviour of carrion crows Corvus corone using automobiles as nutcrackers. Jpn J Ornithol 44:21-35.
Oldham C, 1930. The shell-smashing habit of gulls.Ibis 6:239-243.
Priestley CF, 1947. Rook feeding on mussels. Br Birds 40:239.
Richardson H, Verbeek NAM, 1986. Diet selection and optimization by northwestern crows feeding on Japanese littleneck clams.Ecology 67:1219-1226.
Richardson H, Verbeek NAM, 1987. Diet selection by yearling northwestern crows (Corvus caurinus) feeding on littleneck clams (Venerupis japonica). Auk 104:263-269.
Saunders H, 1891. Notes on birds observed in Switzerland; chiefly in the cantons of Vaud and Neuchâtel. Ibis 6:157-188.
Siegel S, Castellan NJ, Jr, 1988. Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill.
Siegfried WR, 1977. Mussel-dropping behaviour of kelp gulls. S Afr J Sci 73:337-341.
Switzer PV, Cristol DA, 1999. Avian prey-dropping
behavior. I. The effects of prey characteristics and prey loss. Behav
Ecol
10:213-219.
Teichert C, Serventy DL, 1947. Deposits of shells
transported by birds. Am J Sci
245:322-329.
Tinbergen N, 1960. The herring gull's world. New York: Basic Books.
van Impe J, 1978. Mediterranean gull dropping bivalves. Br Birds 71:128-129.
van Lawick-Goodall J, 1970. Tool-using in primates and other vertebrates. In: Advances in the study of behavior, vol. 3 (Lehrman, DS, Hinde, RA, Shaw E, eds). London: Academic Press; 195-249.
Vander Wall SB, 1990. Food hoarding in animals. Chicago: University of Chicago Press.
Ward D, 1991. The size selection of clams by African black oystercatchers and kelp gulls. Ecology 72:513-522.
Wheeler R, 1946. Pacific gulls and mussels.Emu 45:307.
Whiteley JD, Pritchard JS, Slater PJB, 1990. Strategies of mussel dropping by carrion crows Corvus c. corone.Bird Study 37:12-17.
Zach R, 1978. Selection and dropping of whelks by northwestern crows. Behaviour 67:134-138.
Zach R, 1979. Shell dropping: decision-making and optimal foraging in northwestern crows. Behaviour 68:106-117.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. V. Switzer and D. A. Cristol Avian prey-dropping behavior. I. The effects of prey characteristics and prey loss Behav. Ecol., May 1, 1999; 10(3): 213 - 219. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||