Behavioral Ecology Vol. 10 No. 4: 401-408
© 1999 International Society for Behavioral Ecology
Seasonal variation in the sex allocation of a neotropical solitary bee
a Departmento Biologia Geral, ICB-UFMG, Cx. Postal 486, 30161-900, Belo Horizonte, MG, Brazil b Departmento Zoologia, ICB-UFMG, Cx. Postal 486, 30161-900, Belo Horizonte, MG, Brazil c Institute of Cell, Animal and Population Biology, University of Edinburgh King's Buildings, Edinburgh EH9 3JT, UK
Address correspondence to S. A. West. E-mail: Stu.West{at}ed.ac.uk
Received 21 July 1998; revised 13 November 1998; accepted 21 December 1998.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We carried out a field study on the life history and sex allocation of the ground-nesting solitary bee Diadasina distincta (Hymenoptera: Anthophoridae). This species is multivoltine, undergoing five generations a year between February and September. The numerical sex ratio of this species was female biased overall (approximately 38% males) and showed a strong and consistent seasonal pattern. The numerical sex ratio was extremely female biased (approximately 20% males) from February until May, and then slightly male biased (approximately 60% males) from June until September. Females were 3.26 times the size of males, and so the overall investment ratio was female biased throughout the year. The overall female bias and seasonal variation in sex allocation is unlikely to be explained by models that invoke overlapping generations or competition between brothers for mates (local mate competition). We suggest that a possible explanation for the female bias in the early part of the season is local resource enhancement (LRE): nesting near larger numbers of sisters reduces parasitism. LRE is likely to decrease in importance in the later part of the season, when the biased numerical and investment ratios may be explained by models in which male and female offspring gain different fitness returns from resources invested.
Key words: bees, Diadasina distincta, eusociality, local resource enhancement, sex ratio, variance.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
The study of sex allocation has been one of the most successful areas of behavioral ecology (Charnov, 1982
; Herre et al.,
1999; Leigh et al.,
1985). Frequency-dependent selection provides greater mating
success for the rarer sex and so favors equal allocation of resources to sons
and daughters (Fisher, 1930).
If sons and daughters are equally costly to produce, then this equal
investment also translates into equal numbers of males and females. This
argument, however, assumes that the fitness returns from the production of
sons and daughters are linear, or if not linear then identical
(Frank, 1987b,
1990). Many different
biological mechanisms break this assumption in which case-biased sex
allocation is predicted. In these situations there is a rich experimental
literature exploring the predictions of many theoretical models
(Charnov, 1982;
Godfray, 1994;
Herre et al., 1997;
Wrensch and Ebbert, 1993).
In this paper we are concerned with the predictions of sex allocation
theory for solitary bees and wasps. Like other haplodiploid species, bees and
wasps are able to control the sex of their offspring by whether or not they
fertilize the egg: males develop from unfertilized eggs and females from
fertilized eggs (Cook, 1993).
Understanding the reasons for biased sex allocation in solitary bees and wasps
is particularly important because they may have facilitated the evolution of
eusociality (Grafen, 1986;
Stubblefield and Charnov,
1986). A number of theories have been developed to predict and
explain patterns of sex allocation in these species, and the available data
are generally consistent with theoretical predictions
(Brockman and Grafen, 1992;
Cronin and Schwarz, 1997;
Frank, 1987b,
1995;
Helms, 1994;
Rosenheim et al., 1996;
Schwarz, 1988,
1994;
Seger, 1983). However, support
for these models is generally only qualitative, and detailed studies of
individual species are required to test the assumptions as well as the
predictions of these models (Brockman and
Grafen, 1992). To address this problem we have carried out a field
study on the life history and sex allocation of Diadasina distincta,
a solitary bee that nests in holes that females dig in the ground. Before
describing our work we examine the various assumptions and predictions of the
relevant theoretical models.
Werren and Charnov (1978;
Seger, 1983) constructed a
model to explain seasonal sex allocation biases in species with overlapping
generations (partially bivoltine). The general prediction is that in any
generation, selection favors an excess investment in the sex that has
relatively greater overlap with future generations. For example, many species
in the wasp family Sphecidae have two generations a year
(Seger, 1983). Males and
females diapause as late-instar larvae during the autumn, emerge in the
spring, mate, and produce the offspring that will be the summer generation.
The summer generation produce the offspring that diapause during the autumn.
Some of the males that emerge in the spring may survive to mate females of the
summer generation, as well as those of their own. In this case a male bias is
favored in the offspring that diapause over winter, and a female bias in the
other generation. These models provide a possible explanation for the seasonal
sex allocation patterns that have been observed in partially bivoltine wasp
and bee species (Brockman and Grafen,
1992; Seger,
1983). Seasonal sex allocation patterns may also evolve in
response to seasonal perturbations (variation in recruitment or mortality
rates) of the population (Werren and
Charnov, 1978; Werren and
Taylor, 1984; West and
Godfray, 1997).
Frank (1987b,
1995; see also
Charnov et al., 1981;
Trivers and Willard, 1973)
considered the situation when male and female offspring gain different fitness
returns from the resources invested for offspring provisioning, such as hosts
or pollen. For example, suppose that females gain more from an increase in
size than males. In this case theory predicts (1) more resources should be
provided to females, who will subsequently be larger; (2) the investment sex
ratio (defined as the proportion of resources allocated to males) should be
female biased; (3) the numerical sex ratio (defined as the proportion of
males) should be male biased; and (4) seasonal variation in either the
distribution of resources or the relationships between fitness and size should
lead to a change in the relative size of the sexes, the overall investment sex
ratio, and the numerical sex ratio.
More recently, Rosenheim et al.
(1996) have extended this type
of model to the case when egg supply may also limit reproduction. In this case
the factor limiting a female's reproduction may vary across a continuum from
purely egg limited to purely resource limited, with intermediate states where
a female is partially limited by both. There is observational and experimental
evidence that female parasitoid wasps facultatively adjust their behavior in
response to such variation (Hunter and
Godfray, 1995; West et al.,
1999). Rosenheim et al.'s model makes two additional predictions.
As the availability of resources for provisioning increases and females become
more egg limited they should increase provision masses per offspring and
produce a less male biased or even an equal numerical sex ratio and a more
female biased investment sex ratio.
Another factor that has been suggested to lead to female-biased sex
allocation in solitary bee species is local resource enhancement (LRE)
(Schwarz, 1988). Local
resource enhancement can occur when the presence of individuals of one sex
increases the fitness of other related individuals
(Greeff, 1999;
Seger and Charnov, 1988;
Taylor, 1981;
Toro, 1982). This has been
shown to potentially be important in explaining the female-biased sex
allocation of some bee species where related females cooperate to build and
use nests (Cronin and Schwarz,
1997; Schwarz,
1988, 1994;
Stark, 1992). An important
point is that LRE between siblings can result in two outcomes, depending on
whose fitness is increased (Greeff,
1999; Seger and Charnov,
1988; Toro, 1982).
If both sexes benefit, then an overall bias in sex allocation is favored, with
all females investing in the same way. However, if only the sex that provides
the benefit gains a fitness increase, then, although the overall sex
allocation may still be biased, there is no single investment ratio that all
females should produce. Instead, some individuals are predicted to produce
purely males and the remainder to produce purely females, leading to greater
than binomial (random) variance in the numerical offspring sex ratios.
In this paper we examine the neotropical solitary bee, D.
distincta. This species has several overlapping generations per year
(Martins and Antonini, 1994).
Consequently, any of the above sex ratio models could theoretically be
applicable, and so it is an excellent species on which to carry out a detailed
study. Our first aim was to document the numerical and over-all investment sex
ratio of the species and any seasonal variation. Our second aim was to test
some of the assumptions and predictions of the various theoretical models
described above. We determined the population dynamics of the species so that
we could evaluate the importance of models which invoke overlapping
generations. To examine the role of models where males and females gain
different fitness returns from resources, we determined (1) the availability
of the resource (pollen) used to provision offspring; (2) whether there was
any seasonal variation in the number of nests produced; (3) the amount of
pollen used to provision nests containing males and females, and (4) the
relative size of males and females. Finally, we assessed the importance of LRE
by documenting the prevalence of parasitism and the variance in numerical
offspring sex ratios produced by individual females.
Background biology
D. distincta is a neotropical ground-nesting bee
(Martins and Antonini, 1994).
We have studied populations in a 3 x 50 m study site at the Ecological
Station of the Universidade Federal de Minais Gerais (EE-UFMG) in Belo
Horizonte, Minas Gerais, Brazil (see
Martins and Antonini, 1994).
The adult bees emerge at the end of the rainy season during February, and
nesting activities occur until October, through the drier and colder months of
the year. Females nest in aggregations, with the size of aggregations varying
seasonally (Martins and Figueira,
1992). In the beginning of the reproductive season aggregations
are small (20-50 nests), but they then grow in size until August (50-2000
nests), after which they decrease in size until the end of the reproductive
season. Each nest is a burrow (single cell) in the ground approximately 4 cm
deep and 1 cm in diameter. A single female constructs each nest, and old nests
are not reused. Females collect pollen, predominantly (>90%) from
Ludwigia laruotteana (Onagraceae;
Martins and Borges, in press),
and mold it into a ball in at the bottom of each nest. They then lay a
solitary egg under this ball and close the nest entrance with moist soil.
Excavating, provisioning, and closing a nest takes up to 5 days. The egg-adult
development time observed in the field is approximately 27 days. Adult females
sleep in nests, and males sleep in the surrounding vegetation on which they
also rest between patrolling periods. On sunny days the males search for mates
in aggregations. Males strongly compete for females and pursue both newly
emerged females and females provisioning nests. Females mate repeatedly
throughout their lives.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Population dynamics and parasitism
We estimated the bee population size (number of male and female bees in our study site) every 15 days from March to September during 1994 and 1995. The distance between aggregations in our study site averaged approximately 1.5 m. We used sweep nets to capture all the individuals flying over nest aggregations at the peak of daily activity. Sweeping was done over relatively large aggregations (30-50 nests at the beginning of the reproductive season): 6 in 1994 and 4 in 1995. The captured individuals were marked with a fast-drying acrylic paint and released. Twenty four hours later we captured all flying individuals using the same technique. We then estimated the population size by the Petersen method (Bailey, 1952
). We also
estimated pollen availability every 15 days over the same periods by counting
the number of receptive flowers on 20 marked individuals of the plant L.
laruotteana. Flowers of L. laruotteana are receptive for 1 day.
This species is the main (>90%) source of pollen for female nests
(Martins and Borges, in
press).
We examined the demography of the population in further detail by marking
and following individual bees and their offspring. Females from the first
generation in 1994 and 1995 were marked as above and followed. Plastic cups
were placed over any nests that they built and observed daily. Emerging adults
were sexed and females marked. Emerging parasites were recorded. Cups were
placed over all nests produced by the marked females. This was repeated with
the subsequent generations until the end of the reproductive season. By
September, the population was composed entirely of dormant larvae
(Martins and Antonini, 1994),
and we observed the nests containing these larvae until the emergence of
adults the following year. Ten days after the emergence of the last adult in a
generation we excavated the remaining nests. In these cases the nest contents
were generally moldy pollen balls or, rarely, dead adults. We estimated the
longevity of females by recording the number of days that marked individuals
from the first generation of 1994 were observed at the nest aggregation after
their emergence.
Sex allocation
We estimated the numerical sex ratio by recording the sex of adults caught
in emergence traps placed over six nest aggregations. These six aggregations
were sampled every month during 1994 and 1995. The nests examined covered a
range of sizes, varying between 30 and 50 nests at the beginning of the
reproductive season. These traps caught all the bees that emerged and so could
not be biased by different behaviors of the two sexes.
We also recorded the numerical offspring sex ratios produced by individual females. This was done by recording the offspring sex ratios produced by the females that we had individually marked. Plastic cups were placed over all the nests constructed by marked females, and the sex of the emerging offspring recorded. An important caveat with this data is that it was only collected on individuals that were in small aggregations (<50 nests throughout the reproductive season). Marked individuals could not be followed in large aggregations. This will have biased our results if the sex ratio differed with aggregation size.
Sexual dimorphism
We measured the width of the thorax in males and females. This measurement
was taken as the maximum distance between the external margins of the tegulae.
We cubed the thorax width, as this has been shown to be highly correlated with
dry weight (Silveira et al.,
1993). The females measured were collected during April and August
for both 1995 and 1996. The males measured were collected during April 1995,
April 1996, and August 1996. We used these measurements in conjunction with
the observed numerical sex ratios to estimate the overall investment sex ratio
(see Boomsma, 1989, for why
this may underestimate the overall allocation to females). We also measured
the length and diameter of cells from which males and females had emerged.
These cells had been brought into the laboratory during May 1993.
Statistical analysis
All analyses were carried out with the GLIM statistical package
(Crawley, 1993). Proportion
data such as sex ratio (or proportion parasitized) usually have non-normally
distributed error variance and unequal sample sizes. To avoid these problems,
we analyzed all proportion data with a general linear model analysis of
deviance, assuming binomial errors, and a logit link function. The number of
males in a sample was used as the response variable and the total number of
males and females as the binomial denominator. Initiall, a full model was
fitted to the data, including all explanatory variables and their
interactions. Terms were then removed from the full model by stepwise deletion
(Crawley, 1993). Whether the
removal of a term caused a significant increase in deviance was assessed with
a chi-square test.
The variance in the numerical offspring sex ratios produced by individuals
were analyzed with the two methods used and described in detail by West and
Herre (1998). The rationale
behind both methods was to compare the variance observed in the sex ratio data
with that expected given a binomial (random) distribution. One of the methods
is based on least squares regression (GV;
Green et al., 1982;
West and Herre, 1998) and the
other on maximum likelihood (HF; West and
Herre, 1998). The values of GV and HF represent the observed
variance divided by that expected given a binomial distribution. Values of GV
or HF <1 indicate less than binomial variance, and values of GV or HF >1
show greater than binomial variance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Population dynamics and parasitism
Population size showed a strong seasonal pattern, with a similar pattern observed in 1994 and 1995. The number of bees increased until June, after which time it declined (Figure 1a). The population dynamics of the bee strongly followed that of its host plant, L. laruotteana. The number of receptive flowers on 20 marked plants increased until May (1994) or June (1995), after which time the number declined (Figure 1b).
|
The number of generations in a year was determined by marking and following individual bees and their offspring. There were five generations in both 1994 and 1995. All of the offspring of the fifth generation entered diapause. No offspring from earlier generations entered diapause. The periods between which the first and last female emerged each generation are shown in Figure 2. There was some overlap in emergence time between the first and second generations (both years) and between the second and third generations (1995 only).
|
The marked females produced between two and five nests during their
lifetime. The average number of nests produced by a female was 2.90 (95% CI,
0.09, n = 270), and this did not differ significantly between
generations (F4,265 = 0.91, ns). The proportion of a
female's nests that were parasitized in 1994 varied significantly between
generations (
(4)2 = 15.19, p <.005,
Table 1), and showed no
significant relationship with the number of nests that the female produced
((1)2 = 0.99, p >.1). Within
generations, neither male nor female offspring emerged consistently before the
other. The longevity of the marked females in the first generation of 1994
ranged from 6 to 12 days and averaged 9.00 (95% CI, 1.20).
|
Sex allocation
In both 1995 and 1996 the numerical sex ratio of emerging bees was strongly
female biased (approximately 20% males) from March until June (corresponding
to the first three or four generations), and then slightly male biased
(approximately 60% males) from July until September (corresponding to the last
one or two generations; Table 2
and Figure 3a). Summing over
the relevant months, these numerical sex ratios were significantly different
from 50% males for both March until June (1995:
(1)2 = 148.70, p <.001; 1996:
(1)2 = 121.20, p <.001) and July until
September (1995: (1)2 = 7.71, p <.01;
1996: (1)2 = 17.90, p <.001). The
numerical sex ratio differed significantly between months
((1)2 = 220.90, p <.0001), but not
between years ((1)2 = 1.88, p >.1).
Summing over both years, the overall numerical sex ratio was significantly
female biased ((1)2 = 80.50, p <.0001;
38% males).
|
|
The pollen used for stocking nests, or the time required to locate and
collect the pollen, may be factors limiting the sex allocation of females. Two
possible indices of pollen availability in a month are the number of receptive
flowers and the number of receptive flowers divided by the average number of
foraging female bees (the estimated population size). The numerical sex ratio
of emerging bees showed no correlation with either the number of receptive
flowers ((1)2 = 0.12, p >.1) or the
number of receptive flowers divided by the average number of foraging female
bees ((1)2 = 2.52, p >.1). There is a
large discontinuity in the numerical sex ratios between June and July, so we
also tested for correlations separately in these two parts of the year. In
both cases there was no significant correlation with either the number of
receptive flowers (March-June: (1)2 = 3.73,
p >.05; July-September: (1)2 = 0.35,
p >.1) or the number of receptive flowers divided by the average
number of foraging female bees (March-June: (1)2 =
3.81, p >.05; July-September: (1)2 =
0.17, p >.1), although it should be noted that the same
close-to-significance positive correlation from March to June was observed in
both years.
We also collected data on the numerical offspring sex ratios produced by
individual marked females during 1994
(Table 3). The numerical sex
ratio was highly female biased (approximately 15% males) in the first two
generations, less female biased (approximately 25% males) in the third and
fourth generations, and slightly female biased (approximately 40% males) in
the fifth generation. This difference between the generations was significant
((4)2 = 24.80, p <.001), and was
significantly positively correlated with the number of nests that a female
produced ((1)2 = 28.32, p <.001). The
slope of this relationship (the interaction term) did not differ significantly
between generations ((4)2 = 3.26, p
>.1). The numerical sex ratio was significantly different from the expected
50% males in all five generations. Overall, the numerical offspring sex ratio
produced by the marked females was significantly more female biased than that
caught in emergence traps placed over the aggregations (1995:
(1)2 = 4.52, p <.05; 1996:
(1)2 = 9.49, p <.005).
|
Overall, the numerical sex ratios produced by individuals showed significantly less variation than that expected with a binomial distribution (GV = 0.82, p <.001; HF = 0.79). Examining the numerical sex ratios from each generation separately also showed less than binomial variation in each generation, although this difference was only significant for generation five (Table 3). However, sample sizes and thus the power of the test were lower for the other generations.
Size dimorphism
Females were significantly larger than males (F1,74 =
161.74, p <.01, Table
4). Female size did not differ significantly between the beginning
(April) and end (August) of the reproductive season (F1,44
= 1.25, ns), or between 1995 and 1996 (F1,43 = 0.01, ns).
The interaction between these two factors (stage of season and year) was also
not significant (F1,42 = 0.09, ns). Male size did not
differ significantly between the beginning and end of the reproductive season
(F1,27 = 0.52, ns). Males were significantly smaller in
1996 than in 1995 (F1,28 = 161.71, p <.01).
Females were 2.62 and 3.70 times the size of males for 1995 and 1996,
respectively; these estimates were used to calculate the overall investment
sex ratios which are given in Table
2.
|
Females developed in significantly larger nests (cells) than males (length: F1,38 = 45.57, p <.01; diameter: F1,38 = 40.45, p <.01). The average length and diameter of female nests were 1.14 mm (95% CI, 0.03, n = 25) and 0.85 mm (95% CI, 0.03, n = 25), respectively. The average length and diameter of male nests were 0.97 mm (95% CI, 0.04, n = 15) and 0.71 mm (95% CI, 0.04, n = 25), respectively.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We have collected life-history and sex ratio data on the neotropical solitary bee, D. distincta. This species is multivoltine, undergoing five generations a year between March and September (Figure 2). The numerical sex ratio of emerging bees in this species showed a strong and consistent seasonal pattern (Figure 3a), and this is the first time this pattern has been observed in a neotropical bee species. The numerical overall sex ratio was extremely female biased (approximately 20% males) from February until May (corresponding to the first three to four generations), and then slightly male biased (approximately 60% males) from June until September (corresponding to the last one or two generations). Summing over the year, the overall numerical sex ratio was significantly female biased (38% males). Estimates from 1995 and 1996 suggest that females were, on average, 3.25 times larger than males, and so the overall investment ratio was female biased throughout the year (Figure 3b).
Variable sex allocation may evolve in response to a number of factors in
species with overlapping generations, such as variation between the
generations in diapause, population recruitment, or mortality rates
(Seger, 1983;
Werren and Charnov, 1978;
Werren and Taylor, 1984;
West and Godfray, 1997). The
general prediction of these models is that selection favors an excess of the
sex that will experience lower reproductive competition
(West and Godfray, 1997).
Variable diapause between the generations is unlikely to be important in
D. distincta because only individuals of the fifth generation
diapause, and all of these individuals emerge in the first generation the
following year. Adults do not survive over winter. Likewise, variation in
either population recruitment or mortality rates is unlikely to be important
because the overlap between generations is not of a form predicted to lead to
sex allocation biases: female longevity (approximately 9 days) is considerably
less than the egg-to-adult development time (approximately 27 days), and so
individuals are unlikely to compete with their own offspring. We have not
measured male longevity, but even if it weree enormously greater, it could not
explain the female bias in earlier generations—it would actually favor a
bias in the opposite direction toward males.
Frank (1987b,
1990,
1995) predicted the
consequences of differing relationships between the resources invested for
offspring provisioning and fitness differ between the sexes. In D.
distincta, more resources were invested in female offspring, who were
consequently larger. This suggests that females gain more from increased
resources than males, a pattern which is believed to hold across many
hymenopteran species (Charnov et al.,
1981; Godfray,
1994). In this case Frank's models predict that the overall
numerical sex ratio will be male biased, and the investment sex ratio female
biased. Frank's models therefore predict the observed patterns in both the
numerical and investment sex ratios during the latter part of the season,
corresponding to the last one or two generations. However, these models cannot
predict the extreme female bias in both the numerical and investment sex ratio
during the earlier part of the season.
Rosenheim et al. (1996)
extended the type of model considered by Frank to include egg limitation.
Rosenheim et al. predict that as the total resources available decreases,
fewer resources will be provided for each offspring and that, when males are
the smaller sex, the numerical sex ratio should become more male biased.
However, we found no significant correlation between the numerical sex ratio
and possible correlates of the amount of resources (pollen) available for
provisioning offspring: the number of receptive flowers and the number of
receptive flowers per female bee.
We have concentrated on discussing the overall sex allocation of emerging bees (Figure 3). However, we also examined the numerical offspring sex ratios produced by individual females (Table 3). An important caveat with this data is that it was only collected on individuals that were in small aggregations; marked individuals could not be followed in large aggregations. This would have biased our results if the sex ratio differed with aggregation size. The offspring sex ratios produced by these females became less female biased in the later generations but did not become male biased like the data collected by emergence traps over the aggregations. The data collected on individual sex ratios showed that the largest sex ratio shift was between the last two generations of the year. This is consistent with the male bias in the emergence-trap numerical sex ratios during the latter part of the season being due primarily to the fifth generation.
The numerical offspring sex ratio produced by individual females was
significantly positively correlated with the number of nests that they
produced. This correlation would be predicted by Frank's
(1987b,
1995) models if females
producing more offspring also provided fewer resources to each individual
offspring. Future work could test this possibility.
A commonly mentioned explanation for female-biased sex ratios in
hymenopteran species is competition for mates between brothers and inbreeding
in subdivided populations [termed local mate competition (LMC);
Hamilton, 1967]. The biology
of the species makes LMC seem unlikely: each nest contains only a single egg,
males search for mates in aggregations, and females mate repeatedly throughout
their lives. A role for LMC seems particularly unlikely, as either extreme
competition between brothers or large (>50%) numbers of matings between
siblings would be required to explain the extreme female bias observed. In
addition, we observed a positive correlation between sex ratio and brood size,
which is the opposite direction to that predicted and observed under
conditions of LMC (Flanagan et al.,
1998; Griffiths and Godfray,
1988; Stubblefield and Seger,
1990; Werren,
1980; West et al.,
1997). More generally, there is no evidence for LMC in solitary
wasps and bees (Frank,
1995).
We suggest that a possible explanation for the extreme female bias in both
the numerical and investment sex ratios during the early generations is LRE.
Parasitism would provide the driving force given the following scenario.
Parasitism is lower in large aggregations (Antonini Y, Martins RP, in
preparation), and so the fitness of females is likely to be greater in larger
aggregations. Female-biased offspring sex ratios have at least two
consequences that would lead to fitness benefits: (1) females stay in the same
area and nest near their sisters (Antonini Y, Martins RP, in preparation), and
so having more sisters leads to being in a larger aggregation and (2) because
the population is growing in the first generations of the year
(Figure 1), having more sisters
would lead to future generations (children, grandchildren, etc.) nesting in
exponentially larger aggregations (see also
Frank, 1987a).
A testable hypothesis arising from this scenario is that the female bias in both the numerical and investment sex ratios would be greater in smaller aggregations. Interestingly, the numerical offspring sex ratios produced by the individual marked females, who were in relatively small aggregations, were significantly more female biased than the numerical sex ratio of bees caught in emergence traps placed over aggregations. However, further work on how males disperse is required to back up this prediction. For example, if males dispersed to other aggregations, then females nesting in small aggregations might be favored to produce males, who could go to large aggregations to mate with females whose offspring would suffer from relatively low parasitism. Theoretical work would be required to understand the relative consequences of these different factors. Predictions may also be complicated by the fact that the optimal strategies for individuals to pursue (e.g., sex allocation and dispersal) and the population dynamics will both have consequences for and depend upon each other.
To fully explain our data, this hypothesis would require that LRE becomes less important in the later generations, especially the fifth. Local resource enhancement would be expected to be most important (and therefore predict a greater female bias) in the early part of the season: early in the year aggregations are small (20-50 nests) and parasitism is high (Table 1), whereas later in the year aggregations are large (50-2000 nests) and parasitism is low (Table 1). Future work will be required to determine whether this is able to explain the extreme jump, in both the emergence trap and individual numerical sex ratio data, between the last two generations of the year.
Local resource enhancement theory predicts that if only females benefit
from having extra sisters, then some individuals should produce only
daughters, and the rest of the population only sons (Greeff, unpublished
manuscript: Seger and Charnov,
1988; Toro, 1982).
This would lead to an overdispersed pattern of individual sex allocation and
much greater variance than that expected given a binomial distribution.
However, we observed the opposite pattern, significantly less than binomial
variation. This would be predicted under LRE if both sexes gained a fitness
advantage from extra sisters. A possible explanation for this is that the
benefit through an increased growth rate and larger aggregations in future
aggregations would benefit both sexes. Another possibility is that some
important factor remains to be documented. However, it should also be pointed
out that our analyses were based on the numerical sex ratios of individual
females in relatively small aggregations. Thus we do not know the sex ratio
variance among females in larger aggregations (where females may produce more
sons), nor in the total population.
The observed positive correlation between numerical sex ratio and brood
size was in the same direction as that previously noted in allodapine bee
species where the female-biased sex allocation is most likely due to LRE
(Cronin and Schwarz, 1997;
Schwarz, 1994). The
theoretical basis of such a correlation is unclear
(Greeff, 1999). An important
difference with our study is that the allodapine species studied are
primitively (facultatively) eusocial, with related females cooperating to
build, use, and defend nests. Related females of D. distincta merely
nest near each other and do not actively cooperate. Nonetheless, the
association between related females nesting near each other and female-biased
sex ratios could certainly help facilitate the evolution of eusociality and
could represent a primitive step on the way toward cooperation.
To conclude, our data on D. distincta support the importance of
some areas of sex ratio theory and allow others to be rejected. Both female
offspring gaining relatively more than males from increased resources and LRE
appear to be strong possibilities, and we suggest that their relative
importance varies seasonally. In contrast, any effect due to overlapping
generations or LMC seems unlikely. Further work is required and should be
centered around the models that our current study suggest are likely to be
important. Empirical studies on this and related species are urgently required
to determine the relationship between size and fitness for both sexes under
field conditions (e.g., Visser,
1994; West et al.,
1996), examine the relationships between resources invested and
adult body size for the two sexes, and test how the fitness of an individual
female and the offspring sex ratio is affected by the size of an aggregation.
Theoretical work is required to provide sex allocation models that make
quantitative predictions from LRE. More generally, this is also true for a
range of insect and vertebrate examples (e.g.,
Komdeur et al., 1997;
Lambin, 1994;
Schwarz, 1988;
Stark, 1992) where within-
and/or cross-generational interactions can lead to LRE.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We thank James Cook, Jaco Greeff, Ashleigh Griffin, Jack Werren, and Doug Yanega for useful discussion and comments on the manuscript; Arturo Roig-Alsina, N. Evenhuis, and Z. Boucek for identification of insects; the Biotechnology and Biological Sciences Research Council, Brazilian National Research Council, Minas Gerais State Research Funding Agency, and the U.S. Fish and Wildlife Service for funding. This is a contribution of the program in ecology and wildlife management of the Federal University of Minas Gerais.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Bailey NTJ, 1952. Improvements in the interpretation of recapture data. J Anim Ecol 21:120-127.
Boomsma JJ, 1989. Sex-investment ratios in ants: has female bias been systematically overestimated? Am Nat 133:517-532.[Web of Science]
Brockman HJ, Grafen A, 1992. Sex ratios and life-history patterns of a solitary wasp, Trypoxylon (Trypargilum) politum (Hymenoptera: Sphecidae). Behav Ecol Sociobiol 30:7-27.
Charnov EL, 1982. The theory of sex allocation. Princeton, New Jersey: Princeton University Press.
Charnov EL, Los-den Hartogh RL, Jones WT, van den Assem J,1981 . Sex ratio evolution in a variable environment.Nature 289:27-33.[Medline]
Cook JM, 1993. Sex determination in the Hymenoptera: a review of models and evidence. Heredity 71:421-435.[Web of Science]
Crawley M, 1993. GLIM for ecologists. Oxford: Blackwell Scientific.
Cronin AL, Schwarz MP, 1997. Sex ratios, local fitness enhancement and eusociality in the allodapine bee Exoneura richardsoni. Evol Ecol 11:567-577.
Fisher RA, 1930. The genetical theory of natural selection. Oxford: Clarendon Press.
Flanagan KE, West SA, Godfray HCJ, 1998. Local mate competition, variable fecundity, and information utilization in a parasitoid.Anim Behav 56:191-198.[Web of Science][Medline]
Frank SA, 1987a. Demography and sex ratio in social spiders. Evolution 41:1267-1281.
Frank SA, 1987b. Individual and population sex allocation patterns. Theor Popul Biol 31:47-74.[Web of Science][Medline]
Frank SA, 1990. Sex allocation theory for birds and mammals. Annu Rev Ecol Syst 21:13-55.[Web of Science]
Frank SA, 1995. Sex allocation in solitary bees and wasps. Am Nat 146:316-323.
Godfray HJC, 1994. Parasitoids. Behavioural and evolutionary ecology. Princeton, New Jersey: Princeton University Press.
Grafen A, 1986. Split sex ratios and the evolutionary origins of eusociality. J Theor Biol 122:95-121.
Greeff JM, 1999. Cooperative breeding, offspring
packaging and biased sex ratios in allodapine bees. Behav Ecol
10:141-148.
Green RE, Gordh G, Hawkins B, 1982. Precise sex ratios in highly inbred parasitic wasps. Am Nat 120:653-665.[Web of Science]
Griffiths NT, Godfray HCJ, 1988. Local mate competition, sex ratio and clutch size in bethylid wasps. Behav Ecol Sociobiol 22:211-217.
Hamilton WD, 1967. Extraordinary sex ratios.Science
156:477-488.
Helms KR, 1994. Sexual size dimorphism and sex ratios in bees and wasps. Am Nat 143:418-434.
Herre EA, Machado CA, West SA, 1999. Selective regime and fig wasp sex ratios: towards sorting rigor from pseudo-rigor in tests of adaptation. In: Adaptation and optimality (Orzack SH, Sober E, eds). Cambridge: Cambridge University Press.
Herre EA, West SA, Cook JM, Compton SG, Kjellberg F,1997 . Fig wasp mating systems: pollinators and parasites, sex ratio adjustment and male polymorphism, population structure and its consequences. In: Social competition and cooperation in insects and arachnids, vol. I. The evolution of mating systems (Choe J, Crespi B, eds). Princeton, New Jersey: Princeton University Press; 226-239.
Hunter MS, Godfray HCJ, 1995. Ecological determinants of sex allocation in an autoparasitoid wasp. J Anim Ecol 64:95-106.
Komdeur J, Daan S, Tinbergen J, Mateman C, 1997. Extreme adaptive modifications in sex ratio of the Seychelles warbler's eggs.Nature 385:522-525.
Lambin X, 1994. Sex ratio variation in relation to female philopatry in Townsend's voles. J Anim Ecol 63:945-953.
Leigh EG Jr, Herre EA, Fischer EA, 1985. Sex allocation in animals. Experientia 41:1265-1276.[Web of Science][Medline]
Martins RP, Antonini Y, 1994. The biology of Diadasina distincta (Holmberg, 1903) (Hymenoptera: Anthophoridae).Proc Entomol Soc Wash 96:553-560.
Martins RP, Borges JC, in press. Pollen use by a specialist bee, Diadasina distincta (Hymenoptera: Apidae), at a nesting site in southeastern Brazil. Biotropica.
Martins RP, Figueira JEC, 1992. Spatial distributions of nests in Diadasina distincta (Hymenoptera: Anthoohoridae).J Insect Behav 5:527-529.
Rosenheim JA, Nonacs P, Mangel M, 1996. Sex ratios and multifaceted parental investment. Am Nat 148:501-535.
Schwarz MP, 1988. Local resource enhancement and sex ratios in a primitively social bee. Nature 331:346-348.
Schwarz MP, 1994. Female biased sex ratios in a facultatively social bee and their implications for social evolution.Evolution 48:1684-1697.
Seger J, 1983. Partial bivoltinism may cause alternating sex-ratio biases that favour eusociality. Nature 301:59-62.
Seger J, Charnov EL, 1988. Benevolent sisterhood.Nature 331:303.
Silveira FA, da Rocha LB, Cure JR, de Oliveira MJF,1993 . Abelhas silvestres (Hymenoptera: Apoidea) da zona da mata de Minas Gerais.II. Diversidade, abundância e fontes de alimento em uma pastagem abandonada de Ponte Nova. Rev Brasil Entomol 37:595-610.
Stark RE, 1992. Sex ratio and maternal investment in the multivoltine large carpenter bee Xylocopa sulcatipes. Ecol Entomol 17:160-166.
Stubblefield JW, Charnov EL, 1986. Some conceptual issues in the origin of eusociality. Heredity 57:181-187.
Stubblefield JW, Seger J, 1990. Local mate competition
with variable fecundity: dependence of offspring sex ratios on information
utilization and mode of male production. Behav Ecol
1:68-80.
Taylor PD, 1981. Intra-sex and inter-sex sibling interactions as sex determinants. Nature 291:64-66.
Toro MA, 1982. Altruism and sex ratio. J Theor Biol 95:305-311.[Web of Science][Medline]
Trivers RL, Willard DE, 1973. Natural selection of the
parental ability to vary the sex ratio of offspring. Science
179:90-92.
Visser ME, 1994. The importance of being large—the relationship between size and fitness in females of the parasitoid Aphaereta minuta (Hymnenoptera, Braconidae). J Anim Ecol 63:963-978.
Werren JH, 1980. Sex ratio adaptations to local mate
competition in a parasitic wasp. Science
208:1157-1159.
Werren JH, Charnov EL, 1978. Facultative sex ratios and population dynamics. Nature 272:349-350.[Medline]
Werren JH, Taylor PD, 1984. The effects of population recruitment on sex ratio selection. Am Nat 124:143-148.
West SA, Flanagan KE, Godfray HCJ, 1996. The relationship between parasitoid size and fitness in the field, a study of Achrysocharoides zwoelferi (Hymenoptera, Eulophidae). J Anim Ecol 65:631-639.
West SA, Flanagan KE, Godfray HCJ, 1999. Sex allocation and clutch size in parasitoid wasps that produce single sex broods.Anim Behav 57:265-275.[Web of Science][Medline]
West SA, Godfray HCJ, 1997. Sex ratio strategies after perturbation of the stable age distribution. J Theor Biol 186:213-221.
West SA, Herre EA, 1998. Stabilizing selection and variance in fig wasp sex ratios. Evolution 52:475-485.[Web of Science]
West SA, Herre EA, Compton SG, Godfray HCJ, Cook JM,1997 . A comparative study of virginity in fig wasps. Anim Behav 54:437-450.[Web of Science][Medline]
Wrensch DL, Ebbert MA, 1993. Evolution and diversity of sex ratio in insects and mites. New York: Chapman and Hall.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||