| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Behavioral Ecology Vol. 13 No. 1: 125-133
© 2002 International Society for Behavioral Ecology
Antipredator defense as a limited resource: unequal predation risk in broods of an insect with maternal care
Neurobiology and Behavior, Cornell University, Ithaca, NY 14853, USA
Address correspondence to R.B. Cocroft, who is now at the Division of Biological Sciences, 105 Tucker Hall, University of Missouri, Columbia, MO 65211, USA. E-mail: cocroftr{at}missouri.edu .
Received 31 July 2000; revised 30 March 2001; accepted 10 May 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The allocation of parental investment is a potential source of conflict within broods whenever offspring are able obtain differential access to the parental resource. Unlike the provisioning of food, parental antipredator behavior is usually considered a resource that benefits all offspring simultaneously. In the thornbug treehopper (Umbonia crassicornis), offspring form aggregations in exposed positions on host-plant stems. They are subject to intense predation, and maternal defense is their primary means of protection. I examined the distribution of risk within these offspring groups, using natural variation in the outcome of more than 500 predation attempts (324 recorded on videotape) by vespid wasps (Pseudopolybia compressa) on 18 U. crassicornis aggregations. I found three influences on an individual offspring's risk of predation. The first was the presence of a defending female: as expected, offspring were much more likely to survive contact with a wasp if the female was present than if the female had disappeared. The second influence was position relative to other offspring: when wasps were successful in removing an individual, they almost always removed it from the edge of the group. The third influence was distance from the female: the closer an offspring was to the female at the time it was contacted by a wasp, the higher its likelihood of survival. The distribution of risk is determined largely by the behavior of defending females and the prey-searching behavior of wasps. The nature of risk within these aggregations sets the stage for two forms of sibling rivalry: selfish herd behavior and competition for access to maternal defense. It also raises the question of how a parent should allocate defense among offspring when it is unable to defend them all simultaneously.
Key words: parental care, predation, Pseudopolybia compressa, sibling rivalry, Umbonia crassicornis.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The allocation of parental investment within broods of offspring is a source of potential conflict, expressed when siblings compete for a disproportionate share of limited resources (Mock and Parker, 1997
;
Trivers, 1974). The extent of
conflict will be influenced by, among other things, the degree to which
individual offspring can monopolize the resource. For example, if a parent
bird provisions its brood with a succession of small food items, each of which
can be consumed by a single chick, there is the potential, through sibling
interactions or parental choice, for some offspring to obtain a
disproportionate share (Mock and Parker,
1997). In contrast, if the parental resource benefits an entire
brood at once, there may be no opportunity for within-brood competition.
Parental vigilance and defense of the brood against predators are often
considered to be a form of parental investment that benefits all offspring
equally and simultaneously (Lazarus and
Inglis, 1986; Ruusila and
Poysa, 1998). If this is true, we might expect cooperative rather
than competitive behavior among siblings when obtaining access to parental
defense against predators.
What is the evidence that parental antipredator behavior benefits all
offspring equally? In contrast to the provisioning of food, the allocation of
parental defense within broods has received little empirical study. It seems
clear that all offspring should benefit when a parent produces an alarm signal
(Klump and Shalter, 1984) or
intercepts a predator and prevents it from reaching the brood
(Owings et al., 1986).
However, if parents are not always successful in defending against predators,
there may be two ways in which risk is unequally distributed. First, when
broods are relatively large, there may be a selfish herd effect, with
individuals on the margins at greater risk of predation
(Hamilton, 1971;
Krause, 1994;
Mooring and Hart, 1992;
Rayor and Uetz, 1990). If so,
interactions among siblings might determine which individuals are marginal and
thus at greater risk (see Kacelnik et al.,
1995; McRae et al.,
1993). Second, a parent may be unable to defend all offspring at
once, depending on the spatial configuration of the brood
(Windsor, 1987). Distribution
of risk may thus depend both on sibling interactions (i.e., competition for
the most protected or defensible locations) and on parental behavior if
parents preferentially defend some offspring over others. Parents may also
influence the position of offspring within broods, as shown by studies of
catfish defense of mixed-species broods
(McKaye and Oliver, 1980;
McKaye et al., 1992). Adult
catfish actively excluded cichlid young from the center of the school, leaving
their own young in safer, central positions.
In this study I examined the distribution of predation risk within broods
of an insect with extended maternal care. In the thornbug treehopper
(Hemiptera: Membracidae: Umbonia crassicornis), females defend their
aggregated offspring from predators (Wood,
1976,
1983). The offspring (or
nymphs) are vulnerable to predation by a range of invertebrates
(Cocroft, 1996;
Dowell and Johnson, 1986;
Masters, 1997;
McKamey and Deitz, 1996; Wood,
1976,
1983), and loss of the female
is associated with greatly increased predation
(Dowell and Johnson, 1986;
Wood, 1975). Maternal defense
is thus an important resource for developing nymphs, as it is in other insects
with care of offspring after hatching
(Tallamy and Schaefer,
1997).
The protection afforded by female U. crassicornis is solicited by
means of offspring signals (Cocroft,
1996,
1999a,b),
analogous to the food-soliciting signals of altricial birds. However, unlike
nestling birds, U. crassicornis nymphs produce signals in synchronous
bursts that exhibit some elements of cooperation (Cocroft,
1996,
1999b). Cooperation would be
expected if all individuals benefit simultaneously from maternal defense;
however, if maternal defense does not benefit all offspring at once, this
coordinated signaling may also contain an element of competition.
Understanding the balance of cooperation and noncooperation among offspring in
soliciting defense requires clarifying the nature of selection exerted by
predators on aggregated offspring.
I investigated three potential influences on predation risk within offspring aggregations, using natural variation in the success rate of predatory wasps (Hymenoptera: Vespidae: Pseudopolybia compressa). First, to evaluate the importance of maternal defense against wasp predation, I compared predator success rates between broods attended by a female and those from which the female had disappeared. Second, to assess the potential for spatial inequalities in risk, I asked if an individual's risk of predation is influenced by whether it is on the margin or in the center of the group. Third, to assess a female's effectiveness in defending all offspring at once, I examined the correlation between an individual's risk of predation, once contacted by a predator, and its distance from the female at the time of the attack. I also observed the interactions of females and predators that underlie the observed patterns of risk.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Study species
Umbonia crassicornis occurs from Mexico through northern South America, with introduced populations in southern Florida, USA, and the Caribbean (McKamey and Deitz, 1996
). Adults and immatures feed on the sap (presumably phloem) of
various trees and shrubs in the mimosoid legumes
(McKamey and Deitz, 1996).
Females are semelparous, depositing a single clutch of 100 or more eggs into a
host-plant stem near the growing tip
(Wood, 1974). Females
typically spend the rest of their lives with these offspring, often surviving
until the offspring reach maturity 6-8 weeks later. Females face the group,
usually from below, and periodically cut slits into the bark to facilitate
offspring feeding (Wood,
1983). When the nymphs first hatch, they form a group only a few
maternal body lengths long (Wood,
1976; adult females are 1-1.5 cm long). However, as the nymphs
grow to adulthood during the next 3 or 4 weeks, they form a cylindrical
aggregation around the stem, often extending up to 10 or more body lengths
away from the mother (Cocroft RB, unpublished data). Nymphs typically complete
their development to reproductive maturity on the same stem on which the eggs
are deposited.
Aggregations of U. crassicornis nymphs often consist of the
offspring of one singly mated female (Wood
and Dowell, 1984). However, females sometimes mate twice (Masters,
1998). Females also may deposit their egg clutch on the same stem where
another female is already guarding eggs; after hatching, the two broods will
form a single aggregation attended by both females (Cocroft RB, unpublished
data). Furthermore, nymphs occasionally move between aggregations on the same
tree (Cocroft RB, personal observation). Group composition can thus range from
full siblings to a mix of full siblings, half-siblings, and unrelated
individuals. The aggregations used as units of analysis in this study were
operationally defined as a cluster of nymphs on a stem, with or without an
attending adult female.
The study was conducted near San Luis on the Pacific slope of the Cordillera de Tilaran, Puntarenas, Costa Rica, at an elevation of approximately 700 m. Host plants at the study site were saplings of Enterolobium cyclocarpum and Acacia centralis (Leguminosae: Mimosoideae) at the edges of clearings and along riverbanks. Aggregations were found 10-20 cm back from the apical meristem of branches 1-3 m above the ground. Umbonia crassicornis is common at the study site during the May—December rainy season (Masters K, personal communication). This study was conducted between 18 July and 13 August, 1996.
The most common predator of U. crassicornis nymphs at the study
site is a social wasp, Pseudopolybia compressa
(Cocroft, 1998).
Figure 1 shows a female
defending an offspring aggregation from a hovering P. compressa.
Wasps approach from the air, then land on or near the aggregation and attempt
to remove a nymph by biting it and pulling it from the branch. Once the wasp
returns to its nest, nest mates may cooperate in reducing the nymph to small
pieces that will be fed to larvae (Cocroft, personal observation). However,
these wasps do not hunt socially.
|
Behavioral observations
I videotaped aggregations of U. crassicornis to provide a record
of predator—prey interactions. All aggregations located during the study
were filmed, except those that occurred in windy locations where movement of
the stem made observations difficult and filming impossible (those
aggregations were not included in the study). The aggregations observed
contained second- to fourth-instar nymphs (nymphs at San Luis apparently
undergo the final molt to their adult morphology after the fourth stadium,
rather than after the fifth as in other localities). To film an aggregation, I
positioned a camcorder (Canon ES 2000 Hi-8 with a 20X optical zoom lens) on a
tripod 1.5-2 m from the group, with the long axis of the stem in the focal
plane of the camera. Observations on a focal aggregation consisted of
continuous monitoring for 2-6 h. I began filming when I observed a wasp within
1-2 m of the aggregation, and continued filming for approximately 1 min after
the wasp departed. Due to technical limitations, filming was limited to
daylight hours (including dusk) that were free of rain or heavy wind. These
records were supplemented with field notes on predator encounters that were
not captured on video. I analyzed the behavior of predators and prey using a
JVC editing deck and a 12-inch monitor. I measured distance relative to female
body length using a clear plastic ruler over the image on the monitor screen.
This method can introduce some error due to parallax; however, I assumed the
error introduced would be small and unbiased with respect to the analysis.
Presence of the mother and the predator's success rate
To assess the effectiveness of females in defending against P.
compressa, I compared the per-encounter success rate of wasps between
aggregations tended by a female and those from which the female had
disappeared. During an encounter, a wasp arrived at the aggregation, hovered
around the aggregation and/or landed on the stem, then left (with or without a
nymph). Operationally, I considered an encounter to begin when a wasp
approached to within 20 cm of the aggregation (as judged from videotape), and
to end when the wasp moved more than 20 cm away, if it then stayed away for 30
s or more.
Center versus edge positions
Are individuals on the edges of groups more vulnerable to predation than
those in the center? If an individual's position influences its predation
risk, then the proportion of individuals contacted by predators in center and
edge positions should differ from the proportion of individuals available in
those positions. For each contact between a wasp and a nymph recorded on film,
I scored whether the individual contacted was in the center or on the edge of
the group. I then counted the number of individuals in both positions. Because
offspring encircle the host plant stem (see
Figure 1), only about half of
each group is visible on videotape. However, because groups were roughly
cylindrical, there should be no systematic bias between the near and far sides
of the groups filmed. An individual was scored as in a center position if
there was at least one other individual on every side of it and in an edge
position it if was exposed on at least one side. Groups were assumed to be
symmetrical, so individuals not clearly on an edge, but for which one side was
not visible on film, were also scored as center individuals. Sensitivity of
the results to potential violations of this symmetry assumption were assessed
by conducting two additional analyses, one with these individuals scored as
edge and one with them excluded. Field notes clarified the position of these
individuals when more of the stem could be seen than was visible from the
camera angle. I conducted two separate analyses: one for contacts that
resulted in predation, and one for those that did not (the latter may be a
better indicator of predator behavior if female behavior influences which
individuals are actually preyed on once contacted).
The null hypothesis for this analysis was that predation risk is
independent of position. To test this hypothesis, I calculated for each
aggregation the probability of observing as many or more edge predations as
observed, based on the number of individuals in each position. For
illustration, suppose that a total of two predation events were observed for a
family group and that both individuals removed were on the edge of the group.
If at the time of the first event there were 10 nymphs on the edge and 8 in
the center, the probability of an edge predation, given the hypothesis of no
effect of position, would have been 10/(10 + 8), or 0.55. If at the second
event there were 9 nymphs on the edge and 8 in the center, the probability
that an edge nymph was taken would have been 9/(9 + 8), or 0.53. The joint
probability of the two edge predations is obtained from the product of the
probabilities of the individual events: 0.55 x 0.53 = 0.29. In other
words, if predation risk is independent of position, the probability that both
individuals removed from that family would have been on the edge is 0.29. If
the individuals removed from another group were 3 edge and 1 center, the
probability of observing results as or more extreme would be equal to the
joint probability of 3 edge and 1 center predation, plus the probability of 4
edge predations. Results from individual family groups were combined to yield
an overall statistic using Fisher's method of combining p values
(Sokal and Rohlf, 1995).
Improbable results (p <.05) would allow rejection of the
hypothesis that predation risk is independent of position.
Distance from the female at the time of contact
Given that a predator has contacted a nymph, is the nymph's likelihood of
capture related to its distance from the female? Distance from the female was
measured between the nearest points on the bodies of the mother and the nymph
at the time of contact, using a clear plastic ruler placed on the monitor
screen. I first measured the female's body length (from the front to the back
of the pronotum, the hard shell covering the body; see
Figure 1), then made two
measurements used in calculating female—nymph distance: distance along
the long axis of the stem and distance across the stem. From these I
calculated the shortest distance the female would have to travel along the
surface of the stem to reach the nymph (expressed in female body lengths and
corrected for the stem's cylindrical shape).
Aggregations have an inherent polarity along their long axis, with one end closer to the base of the stem and one closer to the tip. It is conceivable that this could be reflected in spatial structuring within the aggregation if, for example, larger individuals cluster toward one end. To control for the possibility of such structure, I divided each aggregation into thirds on a per-event basis by placing a clear plastic ruler over the image on the monitor screen. The position of each individual contacted by a predator was scored as in the proximal, medial, or distal third of the group, where "proximal" indicates the edge closest to the base of the branch. An individual was considered to be within a given third based on the location of its thorax; if its thorax was equally in two areas, the nymph was scored as being in the third containing the head.
I assessed the relationship between a nymph's predation risk, once it was
contacted by a wasp, and its distance from the mother using a multiple
logistic regression analysis (Hosmer and
Lemeshow, 1989). The dependent variable was the outcome of the
contact: captured (1) or not captured (0). In addition to the
female—nymph distance, two categorical dummy variables were included to
control for possible sources of variation: aggregation, which controlled for
differences among family groups (such as variation in maternal behavior or
group location) and position, which controlled for the nymph's position within
the aggregation. The model was as follows: outcome (captured or not) =
constant + aggregation + position (proximal, medial, distal) +
mother—offspring distance. The regression analysis was done using SPSS
version 10. The model was constructed following the recommendations of Hosmer
and Lemeshow
(1989)—starting with a
basic set of likely explanatory variables and adding or deleting variables
based on results of a likelihood ratio test.
Female and wasp behavior
If there is a relationship between mother—offspring distance and
predation risk, the nature of this relationship will be influenced by the
behavior of predators and defending females. I measured two aspects of the
timing of predator—female interactions: the speed at which females moved
across the group, and the time required for a wasp to remove a nymph once
contact was initiated. Females may not be able to defend all nymphs
simultaneously if their travel time across the aggregation is longer than the
time required for wasps to remove a nymph. Furthermore, patterns of
vulnerability within the aggregation will depend on how far the female has to
move to reach a particular location, so I also recorded the females'
positions, both during regular censuses and at the times offspring were
contacted by wasps. I measured the rate of movement of females across the
aggregation in female body lengths per second (body length and distance
traveled were measured as above). I also measured the time required for
females to turn around (180° rotation of heading), which they often do
while defending. In each case I obtained the time interval by counting the
number of full frames (each frame represents 1/30 s). I measured the time
required for wasps to remove a nymph from the stem for events in which the
entire sequence was clearly visible. When removal time was short (<5 s), I
measured its duration to the nearest 1/30 s by counting the number of full
video frames. For longer removal times (> 5 s), I measured duration to the
nearest 1/4 s using a hand-held stopwatch.
In most field studies of predation risk, the logistics of the study do not
allow determination of the number of individual predators involved in the
observed mortality (e.g., Clutton-Brock et
al., 1999; Haskell,
1994; McKaye et al.,
1992; Rayor and Uetz,
1990; Stamp and Bowers,
1988; Wisenden and
Keenleyside, 1995). This knowledge would be useful for an
assessment of individual variation among predators. When possible, I marked
wasps on the body with a dot of white enamel paint to allow for individual
recognition. Wasps were marked when they landed near a group, without
capturing them, and were subsequently identified through the location of the
paint dot.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
I observed 20 aggregations of U. crassicornis and 568 encounters with predatory wasps (Pseudopolybia compressa). The number of wasp encounters observed per aggregation ranged from 0 (in two aggregations) to 94 (mean ± SD = 27.0 ± 26.2). Videotape records were obtained for all or part of 324 of these encounters (the number of encounters used in the following analyses therefore varies depending on whether video analysis was required or whether field notes provided sufficient information). The total observation time was 188 h, with a mean of 8.9 ± 6.6 h per aggregation. All groups consisted of one female and a cluster of immatures, presumably reflecting single-female clutches. One host sapling contained 3 aggregations, 3 contained 2 aggregations each, and 11 contained a single aggregation each. The number of nymphs at the start of observations for each aggregation ranged from 7 to 131 (50.1 ± 36.6); this number dropped as nymphs were lost to predators.
The presence of a defending female was associated with a fourfold reduction in the wasps' per-encounter success rate. In 15 groups attended by a female throughout the study, only 23% of wasp encounters, on average, resulted in capture of a nymph (Figure 2). In three groups observed after the female had disappeared, 94% of wasp encounters resulted in capture of a nymph (Mann-Whitney U test, n1 = 15, n2 = 3, U = 0, p <.01). The only individuals that survived wasp encounters in the absence of the female were two nymphs that dispersed to hidden locations on leaves, where wasps searching along the stem failed to locate them.
|
Videotaped predation events (N = 80; this number excludes
uninformative events for which there were no nymphs in central positions)
provided clear evidence that individuals at the edge were taken
preferentially. Out of 80 nymphs removed by a wasp, 77 were removed from edge
positions. This proportion is significantly edge-biased compared to the
relative numbers of nymphs available in each position
(Table 1). The overall
proportion of nymphs available in each position (obtained by calculating a
proportion for each event, then averaging means from each aggregation) was
0.523 edge to 0.477 center, and the overall proportion of nymphs taken from
each position was 0.963 edge to 0.037 center. These proportions can be used to
estimate an odds ratio (Sokal and Rohlf,
1995). When wasps in this study removed a nymph from an
aggregation, they were 23.4 times more likely to remove it from the edge of
the aggregation than from the center (95% CI = 7.2-76.1).
|
The proportion of wasp-nymph contacts that did not result in predation was also highly edge biased (Table 1). The odds of being contacted were similar to those of being preyed upon: nymphs that were contacted but not removed by wasps were 25.6 times more likely be on the edge of the aggregation (95% CI = 10.9-60.0).
I conducted two additional analyses to assess the sensitivity of these
results to violations of the assumption that groups were symmetrical (and thus
that all of the individuals assigned to center positions were correctly
assigned). The results were not affected by this assumption. If individuals
for which one side was not visible on the video screen were excluded rather
than scored as center, the observed predation was still edge biased
(
2 = 58.76, df = 18, p <.001; one aggregation was
left out because exclusion of these individuals left an uninformative group of
all edge individuals). If the individuals in question were assigned as edge
rather than center, predation also remained edge biased (2 =
37.79, df = 18, p <.001). The results of these additional analyses
are the same for contacts that did not result in predation: results remained
highly edge biased if the individuals in question were excluded
(2 = 148.5, p <.001) and if they were scored as
edge individuals (2 = 96.6, p <.001).
The total number of wasps accounting for the more than 500 predation
attempts observed in the study is unknown. However, I was able to individually
mark five wasps, and these accounted for 33% (189/568) of the attempts.
Repeated observation of these individuals allowed an assessment of individual
variation in behavior and a determination of whether pooling data among
predators is likely to be misleading. Four marked wasps that successfully
attacked nymphs did not differ in the proportion taken on the edge versus the
center of aggregations (predation: heterogeneity 2 = 0.24, df
= 3, p >.9). There is thus no evidence of individual differences
in the likelihood of removing nymphs from the edge of the aggregation (in
fact, there was almost no variation in the data). The number of additional,
unmarked individuals is unknown, but observations indicate that there were a
minimum of two. The encounters observed during the study thus involved seven
or more wasps.
Logistic regression analysis revealed that a nymph's risk of capture, once
contacted by a wasp, depended on its distance from the mother.
Table 2 shows the effect of
each variable in the model after the effects of all of the other variables
were controlled for. Calculation of an odds ratio
(Hosmer and Lemeshow, 1989)
revealed that a nymph's chances of being removed increased 3.7-fold (95% CI =
2.2-6.3) for every additional female body length of distance away from the
female (after controlling for variation among aggregations and locations
within the aggregation). There also was a significant spatial component of the
risk of being removed from the stem. Once contacted by a wasp, nymphs in the
distal third of the aggregation were three times more likely to be removed
than were nymphs in the proximal and medial thirds (after controlling for
variation among aggregations and female—nymph distance). Incorporation
of an interaction between the nymph's location and its distance from the
female did not significantly affect the likelihood of the model; in other
words, the relationship between distance and risk was independent of location
within the aggregation (likelihood ratio test, 2 = 0.964, df =
1, ns). Furthermore, there was no significant improvement in the model from
adding in nymph age (likelihood ratio test, 2 =.637, df = 1,
ns) or the number of nymphs in the aggregation (likelihood ratio test,
2 =.235, df = 1, ns).
|
Female and wasp behavior
Females moved slowly as they walked along the stem or across their
aggregated offspring. Females took 2.6 ± 1.3 s (mean ± SD) to
move one body length (N = 11 females, 32 measurements) in a straight
line. Females also took an average of 2.5 ± 0.5 s to turn around,
something they often did while defending (N = 14 females, 56
measurements).
Females were most likely to be below the nymphal aggregation (i.e.,
proximal with respect to the plant) during regular censuses
(Figure 3a; replicated
chi-square test, total c 2 = 58.05, df = 22, p
<.01; heterogeneity 2 = 31.41, df = 20, ns, indicating that
females were similar in their tendency to be below the group). During
encounters, females walked into the aggregation, and at the times when wasps
contacted nymphs, females were more likely to be in the medial portion of the
aggregation than on one end (Figure
3b; replicated G test, total 2 = 83.9, df
= 24, p <.01; pooled 2 = 26.1, df = 2, p
<.01; heterogeneity 2 = 57.8, df = 22, p <.01,
indicating that not all females were alike).
|
Wasps began to remove a nymph by biting it on the thorax or abdomen. While holding a nymph in its mandibles (and sometimes its front legs as well), a wasp pushed against the branch with its hind legs. Wasps sometimes pulled one end of the nymph away from the branch, then bit through one or more of the nymph's legs. The wasp then rolled the nymph over by lifting up the side where the legs were severed and quickly pulled the nymph away from the branch. The time involved in removing a nymph from a stem ranged from 0.3 to 33.0 s (mean ± SD = 6.1 ± 6.35 s; N = 114 nymphs). Nymphs typically were stationary between and during attacks. However, after a wasp briefly contacted a nymph, that nymph and its neighbors sometimes shifted position. For six aggregations, I observed attacks on both stationary and walking individuals. The amount of time required by a wasp to remove nymphs that were walking was shorter (2.5 ± 1.5 s) than that required for nymphs that were stationary (8.8 ± 3.9 s; Wilcoxon matched-pairs, signed-ranks test, n = 6, T = 0, two-tailed p <.05).
Wasps made their final approach to nymphs in one of two ways. In 29%
(37/129) of the contacts for which the wasp's prior behavior was determined,
the wasp landed on the branch above or below the aggregation, then walked
along the branch until it reached the aggregation. In these cases the wasp
contacted and sometimes removed the first individual encountered on the margin
of the aggregation. In the other 71% (92/129) of cases, the wasp simply
touched down or landed directly on a nymph. Wasps concentrated their contacts
with nymphs at the two ends of aggregations, but they appeared about equally
likely to contact nymphs at each end
(Figure 3c; replicated
G test, total 2 = 160.7, df = 22, p < 1;
pooled 2 = 109.1, df = 2, p <.01; heterogeneity
2 = 51.8, df = 20, p <.01, indicating that there
was significant variation among aggregations in the distribution of
contacts).
In contrast to the slow-moving U. crassicornis females, attacking wasps were highly mobile. Flying wasps typically circled the aggregation, moving from one end to the other and contacting up to eight nymphs in the course of a single encounter. While many encounters with female-defended aggregations lasted less than 20 s, some lasted for more than 5 min (mean = 51.6 ± 74.2 s). Encounters with aggregations without a defending female lasted an average of 13.0 ± 10.7 s. I observed two marked wasps at both female-tended and untended aggregations. Encounters were longer when the female was present, by a factor of 4 in one case (Mann-Whitney U test, n1 = 22, n2 = 3, U = 56, p <.05, one-tailed) and by a factor of 11 in the other (Mann-Whitney U test, n1 = 33, n2 = 11, U = 322.5, p <.01, one-tailed).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
I found three major influences on the predation risk of offspring. The first was the presence or absence of the mother: defending females substantially reduced the effectiveness of wasps in removing offspring. In this study, swaps were only successful in about one-fourth (23%) of attacks on female-defended aggregations. In contrast, wasps were successful in removing a nymph in 94% of attacks on aggregations without a defending female. These proportions are similar to those found by Dowell and Johnson (1986
) in a study conducted
with a Florida, USA, population of U. crassicornis. Only 19% of
attempts by the vespid wasp Polistes major to remove nymphs were
successful when a defending female was present, as compared to 90% of attempts
on nymphs not defended by a female. Maternal defense is important in reducing
the success of foraging wasps.
A second important influence on an individual offspring's predation risk
was its position relative to other individuals in the group. Almost all of the
predation by wasps was on individuals on the margins of groups. This by itself
is not necessarily evidence of edge-biased predation: as pointed out by Krause
(1994), one limitation of many
studies of spatial variation in predation risk is that no information is given
on the relative numbers of individuals in center and edge positions. In this
study, I estimated the probability of the observed number of edge predation
events based on the number of individuals available in each position for each
event. Taking into account the relative numbers of individuals in center and
edge positions, wasps were approximately 23 times more likely to take an
individual from the edge. The same bias was revealed in contacts that did not
result in predation. This edge bias occurred in part because wasps often
contacted or removed the first nymph encountered as they walked up to the
aggregation along the plant stem. There is thus a clear spatial component of
predation risk.
The third major influence on offspring predation risk was distance from the mother at the moment of contact by a wasp. Once a nymph was contacted by a predator, its odds of being removed increased by a factor of 3.7 for every female body length away from the female. What accounts for this relationship between survivorship and the female's proximity? Successful defense of nymphs that were being removed by a wasp usually involved approach by the female. Females moved relatively slowly, however, taking 7.7 s, on average, to travel 3 body lengths (one body length is 1-1.5 cm). Females also took an average of 2.5 s to turn around. In contrast, wasps quickly circled the aggregation and moved from one end to the other. Once a wasp contacted a nymph, it required an average of 6.2 s to remove it from the plant stem. A female moving at an average speed toward a wasp taking an average amount of time to remove a nymph would only arrive in time if she were less than 2.4 body lengths away at the time of contact. Because aggregations often extend along the branch for 2-10 female body lengths (Cocroft, unpublished data), a female on one end of the aggregation may not have sufficient time to reach a contacted nymph on the opposite end. The location of the mother is thus an important and dynamic influence on the spatial distribution of risk within the brood.
In species in which parents actively defend a brood, one would expect
defense to be less effective when broods are larger. Indeed, constraints on
the effectiveness of defense have been proposed to impose an upper limit on
brood size (Lack, 1954;
Safriel, 1975;
Wisenden and Keenleyside,
1995). However, in this study I detected no influence of brood
size on the effectiveness of maternal defense. This rather surprising result
may have come about because of the extent to which brood sizes were reduced by
the time nymphs grew large enough for aggregations to extend long distances
along the stem. That is, females may well be less effective at defending long
aggregations, but no aggregations spanned sufficient distances in this study
to reveal this effect. There also was no effect of offspring age, as reflected
by nymphal stage, on predator success rates. Although wasps did not attempt to
prey on nymphs in the first posthatching stage, they did prey on individuals
at all subsequent stages until they attained the adult morphology.
Results of this study suggest that maternal defense does not benefit all
offspring equally and simultaneously. In general, parental resources can be
viewed as a continuum with respect to the proportion of current offspring that
benefit from an act of parental investment, ranging from provision of a food
item to a single offspring to alarm calls that warn the entire brood. Defense
by female U. crassicornis falls between the two: it does reduce the
overall success rate of wasp predators, but at any one time females can only
effectively defend offspring in their immediate vicinity. Maternal defense in
this species is thus a limited resource. Defense may be similarly limited in
other species in which parents protect offspring, as when cichlids defend
their fry from predatory fish (Keenleyside and Wisenden, 1995) or ducks defend
their brood against gulls (Mendenhall and
Milne, 1985).
How should a parent allocate defense within a brood when it cannot protect
all offspring at the same time? One possibility is that the parent will always
attempt to assess the point of greatest offspring need—for example, by
continuously assessing the predator's current location and approaching it.
Another possibility is that the parent will favor some offspring by remaining
closer to them. These alternatives cannot be distinguished in this study of
natural variation in predator success because predator position, offspring
signaling behavior (see Cocroft,
1999b), and maternal position and orientation were changing
continuously.
What are the consequences for sibling rivalry of this unequal distribution
of predation risk? Hamilton
(1971) predicted cover-seeking
behavior among members of groups vulnerable to a predator that appears within
the group and takes the closest individual. This selfish herd model was
extended by Vine (1971) to the
case of a predator that approaches from outside the group. Studies in a range
of taxa have supported the assumption of increased predation on marginal
individuals (reviewed in Krause,
1994), and in some cases have shown competitive behavior that
gives some individuals access to safer, central positions
(Rayor and Uetz, 1993). Other
things being equal, competition for central locations might also be expected
among individuals within aggregations of U. crassicornis. Even in the
case of an aggregation of full siblings, competition for secure, central
positions should occur. Such competition seems even more likely in broods with
lower average relatedness, as when females mate more than once or when broods
of more than one female merge into a single aggregation. I have not observed
any obvious examples of competitive behavior (e.g., shoving) within nymphal
aggregations, and a nymph probably is unable to displace another individual
that is securely attached to the stem. Any rivalry is thus more likely to be a
form of exploitation competition rather than interference competition; if all
central positions are occupied, an individual seeking a position will have no
option but to leave the group or take an edge position.
Offspring in these groups might also be expected to compete for access to
the mother's defense (Lazarus and Inglis,
1986). Nymphs use substrate-borne vibrational signals to elicit
maternal protection, and these signals are therefore a potential means of
competition for the female's proximity. One unusual feature of signaling in
U. crassicornis offspring groups is that it is coordinated: a wave of
signals begins at one point in the group and quickly spreads, with individual
signals superimposing into a characteristic group display (Cocroft,
1996,
1999a,
b). Furthermore, if multiple
individuals signal, they must coordinate their signals to evoke the female's
response (Cocroft, 1999a). This
coordinated signaling thus contains an element of cooperation, in that
maternal defense both benefits the group and requires collective action (see
Dugatkin, 1997). However, there
may be scope for competition if nymphs can, either by signaling or by
withholding signals, influence the female's position within the aggregation.
Field observations show that nymphs signal as long as a predator is present,
suggesting that signals function to influence the mother's behavior throughout
a predator encounter (Cocroft,
1999b).
The characteristics of maternal responses to signals will constrain the
ways in which offspring can influence parental behavior. It is thus necessary
to understand how female behavior is influenced by variation in signaling
behavior within offspring aggregations. It is also necessary to characterize
how signaling behavior in aggregations varies in response to predation risk.
For example, is the direction of propagation of signals across the aggregation
a reliable indicator of the predator's location? Can the mother detect this
direction? Previous playback studies with nymphal aggregations have shown that
nymphs will signal not only in response to perception of a predator, but also
in response to signals from other individuals, even if those signals occur in
the absence of a predator (Cocroft,
1999a). Are individuals in more vulnerable positions more likely
to signal? Do offspring signal (or fail to signal) in ways that influence the
mother's position within the aggregation? It is at least possible that this
communication system is completely cooperative, in spite of the unequal
distribution of risk. This could occur if predation pressure is sufficiently
high and if cooperative communication within the group is important for
reducing predator success. If so, selection among groups might oppose
selection for selfish behavior within groups, especially where relatedness
within groups is high (Sober and Wilson,
1998).
There was a suggestion of additional influences on risk within groups. The
first was whether an offspring was stationary or moving. Because defending
females move relatively slowly, a nymph's survival depends on its ability to
hold onto the branch until the female arrives. Nymphs sometimes changed
position during or after a visit by a predatory wasp. Changing position after
an attack might be important because wasps can return to the same aggregation
(Cocroft, 1998), and thus
experience of one attack may indicate the likelihood of subsequent attacks.
Changing position entails costs, however, because walking nymphs contacted by
wasps were removed in one-third the time it took to remove stationary nymphs.
A walking individual may not be as securely anchored to the stem as a
stationary individual. A second component of vulnerability to predation was an
individual's position along the long axis of the aggregation. Nymphs in the
distal third of the aggregation (farthest from the base of the plant) were
more likely to be removed, once contacted by a wasp, than nymphs in the rest
of the aggregation. The reasons for this are unclear; it was not due to
greater distance from the mother because this result was independent of
mother—offspring distance. It is possible that larger or more
competitive individuals were clustered in the areas of the aggregation nearer
to the female's usual resting position, although no such pattern was obvious
within aggregations.
How does defense by female U. crassicornis reduce the
effectiveness of foraging wasps? Unlike some other subsocial insects that have
biting mouthparts or raptorial forelegs
(Bequaert, 1935;
Kudo et al., 1992), U.
crassicornis females appear unable to injure an attacking wasp. Instead,
females respond to the approach of a wasp by buzzing their wings and by
approaching and kicking a wasp that lands on the stem. Given the limitations
of female defense, why should a wasp ever end a visit without removing a
nymph? In this study, wasps ended about half of their visits without even
contacting a nymph. It may be that females can increase the handling time of
wasps beyond the point at which it is profitable to continue. For two wasps,
attacks on tended aggregations were 4 and 11 times longer than attacks on
untended aggregations. During visits to female-tended aggregations, wasps
spend most of their time in hovering flight. If this activity is sufficiently
energetically demanding, it may sometimes pay wasps to leave an aggregation,
especially if untended or less well-defended aggregations may be found
elsewhere. In areas with a large number of wasps, repeated visits to the same
aggregations will still result in a substantial loss of brood. However, if the
female can reduce the rate of loss, the extra time may be sufficient to allow
a portion of the brood to reach an age at which they are no longer vulnerable.
The wasps observed by Dowell and Johnson
(1986) were apparently unable
to prey on nymphs once they had reached the fourth or fifth stadium; in this
study, wasp predation attempts on newly eclosed or mature adults were
invariably unsuccessful (Cocroft, personal observation).
This study describes the behavior of one predator species at one locality.
How relevant are these results likely to be for other locations and predator
species? In addition to wasps, nymphs of U. crassicornis are preyed
on by predatory Hemiptera, coccinellid beetles, spiders, neuropteran larvae,
and syrphid fly larvae (Cocroft, unpublished data;
McKamey and Deitz, 1996; Wood,
1974,
1976). Although the
effectiveness of parental defense will vary with the species of predator
(Kudo, 1996;
Kudo and Ishibashi, 1996),
defense by U. crassicornis females is likely to be important in
reducing the effectiveness of many or most of these predators. For example,
Wood (1976) observed predation
by coccinellid beetles and predatory Hemiptera on eight first-instar
aggregations where females were experimentally removed, and mortality in these
aggregations was 100%. Although the role of predation was not documented in
that mortality, female-tended aggregations had fewer predators in their
vicinity and higher nymphal survivorship. Wood
(1976) also observed maternal
defense against a syrphid fly larva; however, at my Costa Rican study site,
larvae of the syrphid Ocyptamus phaeoptera are active at night, when
defending females are unable to respond effectively (Cocroft, unpublished
data).
Increased risk of marginal individuals has not been examined with other predators of U. crassicornis nymphs. Predators that approach the aggregation by moving along the stem (which includes all those listed above, except perhaps wasps) are likely to encounter nymphs on the edges of aggregations. Edge-biased predation may thus be a common pattern for a range of predators. Furthermore, given that the mother must approach the predator to drive it away, the relationship between parent—offspring distance and predation risk may be general feature of maternal defense in this species.
Predation risk is unlikely to be the only important factor influencing
choice of position within aggregations. A position in the group is also a
feeding position, and it is currently unknown what, if any, differential
benefits there may be to feeding in central versus edge positions or in
positions closer to the base versus the tip of the stem. Females continue to
make feeding slits in the bark as offspring develop
(Wood, 1974), with newer slits
being made below the group (i.e., toward the base of the branch). The
nutritional quality of these sites may vary depending on the time course of
the plant's wound response and of scarring of older slits. There might be
trade-offs between foraging and safety, as in some other animal groups (e.g.,
Petit and Bildstein, 1987;
Rayor and Uetz, 1990;
Romey, 1995). Information on
feeding-site selection is thus needed before predictions can be made about how
individuals can optimize their position within a group.
In summary, this study shows that, as in other animal social groups
(Krause, 1994), the benefits
of group living are unequally distributed among offspring in the maternally
defended broods of U. crassicornis. Maternal defense in this species
lies on a continuum between shared and unshared parental investment
(Lazarus and Inglis, 1986). As
a result, there is likely to be a complex interplay of cooperation and
competition within broods, especially when offspring signal to elicit maternal
protection. It also raises the question of how a parent should allocate
defense when it cannot defend an entire brood at once.
|
ACKNOWLEDGEMENTS |
|---|
I thank C. Clark, C. Gilbert, R. Hoy, T. Seeley, T. Wood, and two anonymous reviewers for valuable comments on the manuscript and C. McCulloch, D. Winkler, and especially P. Wrege for statistical advice. I am grateful to K. Masters and family for hospitality, generously sharing access to her field site, and stimulating discussions. I thank the Pastre Campos family for hospitality and assistance during my fieldwork. I also thank M. Nelson for producing the drawing used in Figure 1. Financial support was provided by the Animal Behavior Society, the Olin Foundation, Sigma Xi, and a National Science Foundation Dissertation Improvement grant. J. Carpenter identified the wasps. C. Godoy at Inbio, and the staff of the Organization for Tropical Studies, provided logistical help. I thank the government of Costa Rica for permission to conduct the research.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bequaert J, 1935. Presocial behavior among the Hemiptera. Bull Brooklyn Entomol Soc 30: 177-191.
Clutton-Brock TH, Gaynor D, McIlrath GM, Maccoll ADC, Kansky R, Chadwick P, Manser M, Skinner JD, Brotherton PNM, 1999. Predation, group size and mortality in a cooperative mongoose, Suricata suricatta. J Anim Ecol 68: 672-683.
Cocroft RB, 1996. Insect vibrational defence signals. Nature 382: 679-680.
Cocroft RB, 1998. Offspring-parent communication in a subsocial insect (PhD dissertation). Ithaca, New York: Cornell University.
Cocroft RB, 1999a. Offspring-parent communication in a subsocial treehopper (Hemiptera: Membracidae: Umbonia crassicornis). Behaviour 136: 1-21.
Cocroft RB, 1999b. Parent-offspring communication in response to predators in a subsocial treehopper (Hemiptera; Membracidae: Umbonia crassicornis). Ethology 105: 553-568.
Dowell RV, Johnson M, 1986. Polistes major (Hymenoptera: Vespidae) predation of the treehopper, Umbonia crassicornis (Homoptera: Membracidae). Pan-Pacific Entomol 62: 150-152.
Dugatkin LA, 1997. Cooperation among animals: an evolutionary perspective. New York: Oxford University Press.
Hamilton WD, 1971. Geometry for the selfish herd. J Theor Biol 31: 295-311.[ISI][Medline]
Haskell D, 1994. Experimental evidence that nestling begging behavior incurs a cost due to nest predation. Proc R Soc Lond B 257: 161-164.
Hosmer DW, Lemeshow S, 1989. Applied logistic regression. New York: Wiley.
Kacelnik A, Cotton PA, Stirling L, Wright J, 1995. Food allocation among nestling starlings: sibling competition and the scope of parental choice. Proc R Soc Lond B 259: 259-263.
Klump GM, Shalter MD, 1984. Acoustic behavior of birds and mammals in the predator context. I. Factors affecting the structure of alarm signals. II. The functional significance and evolution of alarm signals. Z Tierpsychol 66: 189-226
Krause J, 1994. Differential fitness returns in relation to spatial position in groups. Biol Rev 69: 187-206.[Medline]
Kudo S, 1996. Ineffective maternal care of a subsocial bug against a nymphal parasitoid: a possible consequence of specialization to predators. Ethology 102: 227-235.
Kudo S, Ishibashi E, 1996. Maternal defence of a leaf beetle is not effective against parasitoids but is against pedestrian predators. Ethology 102: 560-567.
Kudo S, Maeto K, Ozaki K 1992. Maternal care in the red-headed spruce web-spinning sawfly, Cephalcia isshikii (Hymenoptera: Pamphiliidae). J Insect Behav 5: 783-795.
Lack DL, 1954. The natural regulation of animal numbers. Oxford: Oxford University Press.
Lazarus J, Inglis IR, 1986. Shared and unshared parental investment, parent-offspring conflict and brood size. Anim Behav 34: 1791-1804.
Masters KL, 1997. Behavioral and ecoogical aspects of inbreeding in natural animal populations: inferences from Umbonia treehoppers (Homoptera: Membracidae). (PhD dissertation). Princeton, New Jersey: Princeton University.
McKamey SH, Deitz L, 1996. Generic revision of the New World tribe Hoplophorionini (Hemiptera: Membracidae: Membracinae). Syst Entomol 21: 295-342.
McKaye KR, Mughogho DE, Lovullo TJ, 1992. Formation of the selfish school. Environ Biol Fishes 35: 213-218.
McKaye KR, Oliver MK, 1980. Geometry of a selfish school: defense of cichlid young by bagrid catfish in Lake Malawi, Africa. Anim Behav 28: 1287.
McRae SB, Weatherhead PJ, Montgomerie R, 1993. American robin nestlings compete by jockeying for position. Behav Ecol Sociobiol 33: 101-106.
Mendenhall VM, Milne H, 1985. Factors affecting duckling survival of eiders Somateria mollissima in northeast Scotland. Ibis 127: 148-158.
Mock DW, Parker GA, 1997. The evolution of sibling rivalry. Oxford: Oxford Univ. Press.
Mooring MS, Hart BL, 1992. Animal grouping for protection from parasites: selfish herd and encounter-dilution effects. Behaviour 123: 173-193.
Owings DH, Henessy DF, Leger DW, Gladney AB, 1986. Different functions of alarm' calling for different time scales: a preliminary report on ground squirrels. Behaviour 99: 101-106.
Petit DR, Bildstein KL, 1987. Effect of group size and location within the group on the foraging behavior of white ibises. Condor 89: 602-609.
Rayor LS, Uetz GW, 1990. Trade-offs in foraging success and predation risk with spatial position in colonial spiders. Behav Ecol Sociobiol 27: 77-85.
Rayor LS, Uetz GW, 1993. Ontogenetic shifts within the selfish herd: predation risk and foraging trade-offs change with age in colonial web-building spiders. Oecologia 95: 1-8.
Romey WL, 1995. Position preferences within groups: do whirligigs select positions which balance feeding opportunities with predator avoidance? Behav Ecol Sociobiol 37: 195-200.
Ruusila V, Poysa H, 1998. Shared and unshared parental investment in the precocial goldeneye (Avel: Anatidae). Anim Behav 55: 307-312.[ISI][Medline]
Safriel UN, 1975. On the significance of clutch size in nidifugous birds. Ecology 56: 703-708.
Sober E, Wilson DS, 1998. Unto others: the evolution and psychology of unselfish behavior. Cambridge: Harvard University Press.
Sokal RR, Rohlf FJ, 1995. Biometry, 3rd ed. New York: Freeman.
Stamp NE, Bowers MD, 1988. Direct and indirect effects of predatory wasps (Polistes sp.: Vespidae) on gregarious caterpillars (Hemileuca lucina: Saturniidae). Oecologia 75: 619-624.
Tallamy DW, Schaefer C, 1997. Maternal care in the Hemiptera: ancestry, alternatives, and current adaptive value. In: The evolution of social behavior in insects and arachnids (Choe JC, Crespi BJ, eds). Cambridge: Cambridge University Press; 94-115.
Trivers RL, 1974. Parent-offspring conflict. Am Zool 14: 249-264.
Vine I, 1971. Risk of visual detection and pursuit by a predator and the selective advantage of flocking behaviour. J Theor Biol 30: 405-422.[ISI][Medline]
Windsor DM, 1987. Natural history of a subsocial tortoise beetle, Acromis sparsa Boheman (Chrysomelidae, Cassidinae) in Panama. Psyche 94: 127-150.
Wisenden BD, Keenleyside MHA, 1995. Brood size and the economy of brood defence: examining Lack's hypothesis in a biparental cichlid fish. Environ Biol Fishes 43: 145-151.
Wood TK, 1974. Aggregating behavior of Umbonia crassicornis (Homoptera: Membracidae). Can Entomol 106: 169-173.
Wood TK, 1975. Defense in two pre-social membracids (Homoptera: Membracidae). Can Entomol 107: 1227-1231.
Wood TK, 1976. Alarm behavior of brooding female Umbonia crassicornis (Homoptera: Membracidae). Ann Entomol Soc Am 69: 340-344.
Wood TK, 1983. Brooding and aggregating behavior of the treehopper, Umbonia crassicornis. Nat Geogr Soc Res Rep 15: 753-58.
Wood TK, Dowell R, 1984. Sex ratio in Umbonia crassicornis Amyot and Serville (Homoptera: Membracidae). Am Midl Nat 112: 58-66.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. G. Zink Quantifying the costs and benefits of parental care in female treehoppers Behav. Ecol., September 1, 2003; 14(5): 687 - 693. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||