Behavioral Ecology
doi:10.1093/beheco/arp044
Advance Access publication 1 June 2009
Nutritional enrichment increases courtship
intensity and improves mating success in
male spiders
Johannes Peter Lomborg and Søren Toft
Department of Biological Sciences, Genetics and Ecology, Aarhus University, Building 1540,
DK-8000 Århus C, Denmark
The development of male sexual ornaments and the intensity of male courtship behavior are often used by females as criteria for
mate choice and by other males to evaluate the strength of a rival. We tested the hypotheses that courtship intensity and mating
success depend on the males’ nutritional status (enriched or deficient) and that courtship intensity predicts mating success in
males of the same nutritional status. We used wolf spiders, Pardosa prativaga, which have an elaborate display of courtship
behaviors, including encircling, palp vibrations, abdomen vibrations, hopping, etc. Viability parameters indicated enhanced
condition of enriched males. Mating success was higher for nutrient-enriched males in direct competition with deficient males.
Enriched males had higher courtship intensity and were also larger (carapace width) but not heavier than deficient males. The
statistical analysis indicated that diet effects on courtship intensity were indirect, through its effect on size. In competition tests
between males of equal mass and the same diet treatment, the previously most active male (high levels of palp vibrating,
abdomen vibrating, and hopping) had the highest mating success, though this result depended on male nutrient status. The
widely used residual condition index (RCI) did not distinguish the treatments. It is suggested that the index is unsuitable in
a situation of nutritional stress caused by nutrient imbalance. The results underscore the importance of nutrient balancing to all
aspects of performance also in predatory animals. Key words: Araneae, condition dependence, courtship, Lycosidae, nutrient
balance, nutrition, Pardosa prativaga, sexual selection, wolf spider. [Behav Ecol 20:700–708 (2009)]
nherent in most theories of sexual selection is that female
choice is influenced by the size of male sexual ornaments
or the intensity of his courtship behavior (Andersson 1994).
Production and maintenance of ornaments as well as the performance of courtship are costly and therefore depend on the
animal’s condition. It has further been recognized that these
costs are greater for an animal in poor condition than for one
in good condition (Grafen 1990; Kotiaho 2000). Animals can
be considered in good condition if they can mobilize stored
physiological reserves in appropriate situations (Rolff and
Joop 2002; Cotton et al. 2004), for example, for production
of more or less permanent morphological structures (like sexual ornaments), for behavioral performance in competitive
situations (confronting females or rival males, cf., Kotiaho
et al. 1999a), or in other stressful situations (e.g., predator
avoidance), and thus improve their fitness. Thus, conditiondependent traits of an animal may provide information about
an individual in both sexual and nonsexual contexts.
It is increasingly realized that nutritional factors influence
the reproductive success of males, even in cases where the
males provide nothing but sperm. Thus, nutrition may influence several morphological and behavioral traits of importance for male competitive ability and female choice (Rowe
and Houle 1996; Cotton et al. 2004; Hunt et al. 2004). This
relates not only to lack of food (e.g., Kotiaho et al. 2001) but
also to the nutrient composition of the diet. For example,
insects reared on protein-rich diets had increased calling
activity (Kaspi et al. 2000; Hunt et al. 2004), and crickets with
I
Address correspondence to S. Toft. E-mail: [email protected].
dk.
Received 14 May 2008; revised 13 November 2008; accepted 4
February 2009.
The Author 2009. Published by Oxford University Press on behalf of
the International Society for Behavioral Ecology. All rights reserved.
For permissions, please e-mail: [email protected]
a high calling rate had higher body loads of phosphorous
than less active callers (Bertram et al. 2006). Among vertebrates, nutritional condition is known to affect male ornament brightness, immune function, and attractiveness to
females in birds and fishes (e.g., Hill and Montgomerie
1994; Ohlsson et al. 2001; Fisher and Rosenthal 2006), and
in some cases, this has been related to specific dietary constituents like protein, carotenoids, vitamin D, or calcium (KodricBrown 1989; Martin and Lopez 2006; Smith et al. 2006;
McGraw 2007).
Nutrient balancing through dietary self-selection enhances
animal viability as evidenced by measurements of several fitness parameters (Raubenheimer and Simpson 1997). This
has been documented extensively as regards herbivores and
omnivores (Simpson et al. 2002; Raubenheimer and Jones
2006) and recently also for predators (Mayntz et al. 2005;
Raubenheimer et al. 2007). Evidence for nutrient limitation,
in particular nitrogen, of predators (Fagan et al. 2002; Fagan
and Denno 2004) further stresses the need to include nutritional quality as a factor. A good nutrient balance may not
only result in the development of morphological traits of importance for sexual selection (e.g., body size or mass and
ornament size or brightness) but also provide the ability to
mobilize surplus resources in an emergency situation. The
ability to select an optimal diet in males is therefore expected
to be associated with well-developed sexual signals and high
attractiveness to the females. As such, dietary self-selection
should itself be a target of sexual selection.
Courtship in many spiders, including wolf spiders, may be
elaborate, consisting of combinations of visual, vibratory,
and auditory signals, at times coupled with ornaments on legs
and palps (Kronestedt 1979; Kotiaho et al. 1999b; Hebets and
Uetz 2000). Species of the genus Pardosa exhibit a wide range
of male courtship behaviors, varying from almost no visible
Lomborg and Toft
•
Nutrient balance and mating success
display in Pardosa pullata to repeated cycles of different movements in Pardosa prativaga (Richter et al. 1971; Den Hollander
et al. 1973). Nutritional effects on wolf spider courtship have
recently been studied by Hebets et al. (2008) and Wilder and
Rypstra (2008). None of these found diet effects on courtship
intensity, but Hebets et al. (2008) demonstrated diet effects
on female mate selectivity.
In the present study, we tested whether the intensity of courtship display in 2 groups of male spiders depends on male nutritional state after a period of differential diet treatment and if
males exhibiting more intense courtship also have a higher
mating success. Previous laboratory studies of courtship display
and condition have mostly used food limitation as a treatment
to affect condition (Mappes et al. 1996; Kotiaho 2000; Kotiaho
et al. 2001; Uetz et al. 2002), arguing that the treatment reflects variation in foraging history, which would subsequently
affect the male’s condition. In contrast, the present study used
prey of differing nutritional value as treatment, offered ad
libitum to both groups. Rather than hunger state, condition
then reflects the spider’s ability to optimize its nutritional
balance, for example, through prey selection for high-quality
prey (Toft 1999; Mayntz et al. 2005) or for a diverse diet (Uetz
et al. 1992).
We examined the significance of nutritional quality of prey
on the condition of male spiders and on the performance of
male courtship display and mating success. First (experiment 1),
we tested if nutrition affected the mating success of males in
a 2-male competitive situation. Then we tested (experiment
2) whether the courtship intensity in the presence of a female
depends on the nutritional balance of the male and whether
the male most active in courtship is also the most successful in
a 2-male competitive situation. Males fed high-quality prey were
expected to display more intensive courtship and have increased mating success.
Condition can be quantified in various ways. Because it
encompasses many aspects of viability, it may be expressed
in terms of life-history parameters like survival, growth rate,
age at maturation, mass, or biometric size. Condition indices
based on mass, size, or mass/size relationships (Jakob et al.
1996; Kotiaho 1999; Rolff and Joop 2002) are often used as
shortcuts, but they are not direct measures of condition and
may sometimes fail to correlate with fitness (Rolff and Joop
2002). Viability indicators are affected only little by short-term
events like single prey captures but instead depend mainly on
effects integrated over longer periods of time. In contrast,
indicators of current energetic reserves only reflect the present state and can vary from day to day as a result of fluctuations in food availability. To estimate the energetic reserves of
an animal, a residual condition index (RCI) of body mass in
relation to body size (Jakob et al. 1996) can be used, but it
may depend on assumptions that cannot be fulfilled (Garcı́aBerthou 2001). To compensate for possible shortcomings of
the various parameters as condition estimators, survival rate,
time to maturation, size (carapace width), mass, and RCI were
used to compare the 2 groups of spiders.
MATERIALS AND METHODS
Rearing
Medium-sized juvenile P. prativaga, 2–3 molts from maturity,
were collected by hand from a grass field at Stjær near Århus,
Denmark, in late April 2002 for experiment 1 and in late
October 2002 for experiment 2. The species matures in May
when mating occurs and reproduces through spring and summer. The spiders were kept singly in small plastic tubes (Ø 2 cm
and height 6 cm) sealed with foam rubber plugs. Each tube
had a 1-cm bottom layer of plaster mixed with charcoal.
701
The bottom layer was moistened regularly to ensure adequate water for the spiders. All tubes were kept at a constant
temperature of 20 C and a 16:8 light:dark cycle. Experiments were carried out during daylight hours at room temperature (20–25 C).
For both experiments, the spiders were divided randomly into 2 groups. Treatment of these groups differed only in the
diet, which was manipulated according to Mayntz and Toft
(2001) to be either nutritionally enriched or deficient. One
group was raised on live nutritionally enriched fruit flies,
Drosophila melanogaster, reared on a medium of 1.8 g Carolina
Formula 4-24 Drosophila medium (Carolina Biological Supply,
Burlington, NC) mixed with 1.0 g dry, grounded pellets of Techni-Cal adult dog food (Martin Pet Foods, Ontario, Canada)
with demineralized water and baker’s yeast added. The
other group was fed live, nutritionally deficient fruit flies
reared on plain Carolina Formula 4-24 medium, baker’s
yeast, and demineralized water. This resulted in spiders of
2 nutritional states, balanced and unbalanced, respectively
(Mayntz and Toft 2001). Among the nutrients affected is
the nitrogen content, which is 8.7% in the enriched flies
against 7.9% in the deficient flies (Mayntz D, personal communication). In the experiments, male spiders of both diets
were used, whereas only females from the enriched diet
were used in the tests.
Spiders on both diets were offered flies and water ad libitum.
The dates of molts and deaths were recorded at least every second day. As P. prativaga may experience a delay of sexual
activity and receptivity for up to 9 (males) and 11 (females)
days after the final molt (Richter et al. 1971), spiders were
kept for at least 11 days from last molt before being used in
the experiments. No change in diet was made during the
various stages.
Courtship and mating
Based on previous descriptions of the courtship movements
and activities in P. prativaga (Den Hollander et al. 1973) and
on preliminary observations from the current study, specific
behaviors were defined according to Table 1. The activities
‘‘palp shaking’’ and ‘‘abdomen shaking’’ were not described
by Den Hollander et al. (1973), being most likely pooled with
‘‘palp vibrating’’ and ‘‘abdomen vibrating,’’ respectively. Shaking movements are distinguished from vibrations by the lack
of rhythm and by consisting of only a few movements at a time,
whereas vibrations always come in a sequence of several repetitions. Non-courtship behavior patterns, that is, behaviors
observed both in the presence and absence of a conspecific
female, namely, ‘‘walking’’ and ‘‘palp cycling,’’ are also included in the analysis to allow comparisons with the noncourtship activity level of the spiders (cf., Cotton et al. 2004).
When staged with a female, males generally initiate courtship
within a minute. Courtship consists of cycles of behaviors occurring in a nearly fixed order (except palp and abdomen shakings,
which occur erratically): first encircling, then palp vibration,
and followed by abdomen vibrating and hopping. This cycle
is repeated several times. However, the male may simply stop
the cycle without apparent cause, especially during his first
attempts at courtship, or the female may interrupt it halfway
through a cycle by moving away. After a time, the male moves
closer to the female and attempts to mount her, which is the behavior ‘‘copulation attempt.’’ At this stage, the female can reject
the male merely by moving away or by threatening. If allowed,
the male will climb on top of the female’s carapace, facing the
opposite direction of her, and initiate copulation by stretching
the pedipalps around the pedicel to her epigynum on the ventral side of the abdomen. All courtships in this study were premating, that is, before palp insertion and sperm transfer.
Behavioral Ecology
702
Table 1
Definitions of behaviors of male spiders during courtship display and normal activity in Pardosa prativaga
Activity
Description
1) Encircling
2) Palp vibrating
3) Abdomen vibrating
Movement in a circle around the female, with the front toward the female.
A sequence of fast and regular movements of the pedipalps in a rhythmic movement back and forth.
A sequence of heavy and rhythmic jerks (up and down) of the abdomen. Often seen subsequent to palp
vibrating and concurrent with hopping.
A jerky, hopping movement toward the female, most often seen subsequent to palp vibrating.
The slow climbing on top of the female by the male. Often subsequent to hopping.
A few fast and sudden movements of the pedipalps. Not in a steady rhythm.
A few short and fast jerks of the abdomen, sideways or up and down. Not in a steady rhythm.
Any movement around the arena in no particular direction relative to the female. Not a courtship activity.
The slow stretching and subsequent withdrawal of a pedipalp. Not a courtship activity.
4)
5)
6)
7)
8)
9)
Hopping
Copulation attempt
Palp shaking
Abdomen shaking
Walking
Palp cycling
Experimental design
Experiment 1
The spiders used in this experiment were kept on their respective diet for at least 30 days before the experiments were
carried out. All spiders were weighed 1 day prior to the experiments (Sartorius MC5, 0.001 mg resolution). Pairs consisting
of 1 male from the diet of enriched flies with 1 from the deficient group were established. Males were paired as to minimize the weight difference between them. In all cases, the
difference was less than 1 mg, with weight of males in the
whole experiment ranging from approximately 12 to 17 mg.
The maximum difference in date of maturation between the
2 males was 1 week. Adult virgin females raised on the enriched diet were chosen randomly for each male pair. The
males were marked on the tibia of a hind leg with nail polish.
One of a pair, chosen randomly, was marked on the left leg and
the other on the right.
In all, 22 tests were made, each male being used only once.
The bottom of a 14-cm diameter Petri dish was covered with
a moist filter paper. The 3 spiders were placed gently in the
Petri dish as far from each other as possible, the female first
and the males after a few minutes. Matings and the successful
males were noted, as was agonistic behavior between males, either physical fights or threat posturing (Foelix 1996). If no
mating occurred within 2 h, the experiment was terminated
and discounted. If a male was allowed to mate within the 2 h,
the experiment was terminated when he dismounted or was
thrown off. Mating success was also measured as the time
before initiation of the copulation and as the duration of
the copulation.
The first male to mount the female and stay there for at least
10 min was defined as the successful (or primary) mate. In
some cases, the second male would try to move over or under
the mating couple to reach the female’s epigynum with a pedipalp. Such a subsequent ‘‘copulation’’ by the second male was
termed a secondary mating, but the male was considered unsuccessful because it was doubtful whether secondary matings
allowed for sperm transfer.
Experiment 2
Spiders for this experiment were caught in the autumn and
thus were preparing to hibernate; they were kept cold
(5 C) for about 2 months to simulate seasonal conditions until
the treatment commenced. Then the temperature was raised
gradually over a week to 20 C. The spiders were kept on their
respective diet for at least 40 days before they were used in the
experiment.
Courtship analysis. All males (44 enriched and 32 deficient)
were staged individually with a female. Four randomly chosen
females were used in these tests. Available males were analyzed
in random order. The female was placed within a glass cylinder
(Ø 6 cm and height 5 cm) on a bottom of moistened filter paper covering a layer of plaster. Allowing 2 min for the female to
leave silk draglines on the substrate and adjust to the enclosure, the male was then gently introduced into the same glass
arena as far from the female as possible. A video camera (Panasonic AG-450 S-VHS) recorded for 10 min after introduction of
the male from directly above. If a male started to mount the
female before the 10 min had elapsed, he was stopped by gently
separating the spiders with a soft brush. The recording period
was then increased by the amount of time it took the male to
resume courtship display after being disturbed. In the later
analysis, any such periods of disturbance were discounted, giving approximately 10 min of courtship observation for each
male. Premating courtship may last from a few minutes to several hours. The 10 min were chosen to secure measurements
over a reasonable period of consistent courtship. Whenever
separation of coupling spiders was impossible, the observation
and both spiders were discarded. Also, injured males were discounted, as were males that died within a day after the courtship analysis.
The analysis of the recordings was carried out with The Observer 3.0 software package (Noldus Information Technology,
Wageningen, The Netherlands). Occurrences and durations of
male behaviors were recorded to gain the total time spent on
each behavior. Intensity of each behavior was calculated as the
time spent on the behavior divided by the total time of the observation period. As individual behaviors were not mutually exclusive, a summed total of the intensities could in principle
exceed 100%, but this was never observed as the spiders were
inactive much of the time.
Still images were digitally grabbed from the video recordings
and enhanced by use of Final Cut Pro 3 (Apple Computers,
Inc., Cupertino, CA) and Adobe Photoshop 7 (Adobe Systems,
Inc., San Jose, CA) software. From these images, the carapace
width between second and third leg pairs was measured using
the picture analysis program Scion Image (Scion Corporation,
Frederick, MD). Each picture was scaled by the diameter of the
arena in the image, resulting in scales of approximately 0.1
millimeters per pixel.
Mating success. The day after a group of males had thus been
screened for courtship display, they were ranked roughly based
on the intensity of their courtship. Males were then paired
within nutritional groups, enriched with enriched (N ¼ 18)
and deficient with deficient (N ¼ 12), to eliminate possible
secondary effects of diet. Also, to eliminate an effect of which
female the males had been tested against, only males tested
the day before against the same female were paired. To ensure
a high degree of variation of intensity within couples, a highly
active male was paired with a less active male while weight
difference within couples was reduced to a minimum. By this
Lomborg and Toft
•
Nutrient balance and mating success
procedure, we intended to expose the differences in individual male condition, whether genetically or environmentally
determined, that stemmed from before they were collected
in the field. The largest weight difference of a pair was 2.6
mg, with males ranging from approximately 10 to 22 mg in
weight. These pairs were used in tests with a female, which
were carried out 1 day after the courtship measurements. No
male or female was used more than once in these tests nor was
any female used that had previously been used in the courtship measurements. The tests were carried out over 3 weeks,
as males matured, resulting in 32 pairings. The outcome of
the tests was recorded as in experiment 1, noting agonistic
behavior, time to acceptance, duration of the mating, and
which males were allowed to mate.
Statistical analysis
We used both morphometric (mass, carapace width, and a condition index) and viability parameters (mortality and development time) as potential indicators of male condition. Mortality
of male spiders from the start of treatment was compared by
log-rank test for right-censored data (Pyke and Thompson
1986) between the 2 diet groups. As courtship and mating
may affect mortality, males used in tests were censored from
the date of the test, and all unused spiders that were still alive
at the termination of the experiment were censored at this
date. Development time was analyzed by means of a 2-way
analysis of variance (ANOVA) including diet and sex as factors. As the interaction between the factors was nonsignificant, it was deleted from the final model. Tests of equal
variances were made and appropriate transformations used
if necessary.
The RCI was calculated as the residuals from a regression of
the ln-transformed weight against the ln-transformed carapace
width of all males (Jakob et al. 1996; Kotiaho 1999). However,
following Garcı́a-Berthou (2001), an analysis of covariance
(ANCOVA) on weight with diet and width as factors was first
carried out to ensure that the diet 3 width interaction was
nonsignificant and that weights did not differ with diet.
Because the 9 behaviors (Table 1) might be correlated, we
extracted 2 principal components (PCs) that accounted for
50.6% and 45.8% of the variation in the original variables,
respectively. We used the varimax-rotated factors for the analyses (Krzanowski 2000). PC1 and PC2 did not depend on
which of the 4 females the males were tested against. Their further analysis therefore included only diet, carapace width, and
RCI. As none of the interaction terms of the 3-way ANOVAs
were significant, they were deleted from the final models. Results for the single behaviors are illustrated (Figure 3) to
help interpretation of the PCs. For the calculations, Statgraphics Plus 4.0 (Statistical Graphics Corporation, Princeton,
NJ) and JMP 6.0 (SAS Institute, Inc., Cary, NC) were used.
RESULTS
Treatment effects on condition
In experiment 1, survival of male spiders showed no significant
difference between enriched (mean 6 standard error [SE]:
27.6 6 0.7 days, N ¼ 137) and deficient (27.1 6 0.6 days,
N ¼ 135) males (log-rank test, v21 ¼ 0.458, P ¼ 0.50). Development time depended on both diet (F1,236 ¼ 7.8, P ¼ 0.0057;
least squares means [LSMs]: enriched 25.1 6 0.3 days, deficient 26.1 6 0.3 days) and sex (F1,236 ¼ 30.8, P , 0.0001;
LSMs: females 26.7 6 0.3 days, males 24.6 6 0.3 days) with
no significant interaction. Finally, there was no significant diet
difference in male adult weight (F1,78 ¼ 1.12, P ¼ 0.29; enriched 14.5 6 0.3 mg, deficient 14.0 6 0.4 mg). Notice, how-
703
ever, that all diet differences are in the direction expected
from the diet quality.
In experiment 2, males differed significantly in survival
(v21 ¼ 11.5, P ¼ 0.0007), enriched males (N ¼ 125, mean 6
SE ¼ 59.6 6 1.2 days) living longer from treatment initiation
than deficient males (N ¼ 100, 52.6 6 1.3 days). Development
time depended on sex, males maturing faster than females (2way ANOVA on Box–Cox transformed data: F1,239 ¼ 13.8, P ¼
0.0002; LSMs: females 41.5 6 0.4 days, males 38.3 6 0.8 days),
but not on diet (LSMs: enriched 39.8 6 0.6 days, deficient
40.0 6 0.6 days) and not on the interaction. Again, all diet
differences were in the expected direction.
Diet did not significantly affect adult male weight
(ANCOVA, F1,73 ¼ 0.246, P ¼ 0.62; enriched 14.6 6 0.3
mg, deficient 14.5 6 0.3 mg; diet 3 carapace width interaction: F1,73 ¼ 3.58, P ¼ 0.063), but weight was significantly
correlated with carapace width (ANCOVA, F1,74 ¼ 6.83, P ¼
0.011, pooled data: ln [weight] ¼ 0.315 3 ln [carapace
width] 1 3.20). Thus, the RCI could be calculated for each
male from a common regression of all males in experiment 2.
The calculated RCI did not significantly depend on the diet
(ANOVA, F1,74 ¼ 0.226, P ¼ 0.64; diet 3 carapace width interaction: F1,73 ¼ 3.26, P ¼ 0.075) but was slightly higher for
deficient than for enriched males (0.12 6 0.35 vs. 20.09 6
0.25 mg). Male carapace width depended on the nutritional
value of the diet (ANOVA, F1,74 ¼ 7.94, P ¼ 0.0062), as enriched males were larger (mean 6 SE ¼ 0.194 6 0.003 cm)
than deficient males (0.180 6 0.003 cm). Thus, although
males of the 2 groups attained similar weights, enriched males
grew larger than deficient males. Overall, though the significant effects varied between experiments 1 and 2, all diet effects except on RCI (see Discussion) were in the expected
direction. We therefore conclude that diet had the intended
effect on spider condition.
Experiment 1
In all, 22 tests were carried out, of which 9 resulted in a mating.
No agonistic behavior between males was observed. Two cases
resulted in both a primary and a secondary mating; in both, the
enriched male was the primary mate. Seven tests had only a primary mating, from which 6 resulted in the enriched male being
the successful one. Discounting the 2 cases of both primary and
secondary matings, the number of replicates is insufficient to
reveal a significant difference in success of the 2 diet groups
(2-tailed binomial test, N ¼ 7, P ¼ 0.12). Including the 2 cases
with the primary male as the successful one, however, enriched males gained the mating significantly more often than
deficient males (2-tailed binomial test, N ¼ 9, P ¼ 0.039). The
mean weights of successful and unsuccessful males were not
significantly different (paired t-test, t8 ¼ 0.602, P ¼ 0.95).
Because only one deficient male was successful, 1-sample
t-tests were applied to the enriched-males winners’ data on
acceptance time and copulation time against the single corresponding values of the deficient-male winner. The results were
in the predicted directions but nonsignificant (acceptance time
[mean 6 SE] 40.5 6 17.1 min against 71 min [t ¼ 1.78, P ¼
0.12]; copulation time 149.5 6 27.4 min against 109 min [t ¼
1.48, P ¼ 0.18]).
Experiment 2
Courtship intensity
Of the 76 tests of 1 male and 1 female, 44 were with enriched
males and 32 with deficient males. The 7 courtship behaviors
all scored positively and the 2 non-courtship behaviors scored
negatively on PC1. PC1 had high positive loadings for abdomen vibrating, palp vibrating, hopping, and copulation
Behavioral Ecology
704
Figure 1
Comparison of male behavioral
activity (a, b: 2 PCs extracted
from 9 male behaviors [see
Table 1]) recorded during 10
min of confrontation between
a female and a male and biometric parameters (c: carapace
width; d: mass; e: RCI) of male
wolf spiders, Pardosa prativaga,
fed either nutrient-enriched or
nutrient-deficient prey. PC1 represents the intensity of courtship
behaviors; PC2 represents mostly
non-courtship activities. RCI was
calculated from an ln (mass) 2
ln (carapace width) regression.
*P , 0.05 (t-tests).
attempts and negative loadings for walking. PC2 had high positive loadings only for walking and lower loadings for palp shaking, palp cycling, and encircling. PC1 and PC2 could therefore
be interpreted to represent courtship and non-courtship
behaviors, respectively. PC1 was higher in enriched than in deficient males (t74 ¼ 2.4, P ¼ 0.018; Figure 1), whereas PC2
showed no difference (t74 ¼ 0.45, P ¼ 0.65). However, when
carapace width and RCI were included as covariates, the diet
effect disappeared, whereas carapace width was a significant
covariate for both variables (Table 2). PC1 was positively and
PC2 was negatively correlated with carapace width (Figure 2).
RCI had only a marginal effect on PC2 and none on PC1
(Table 2).
Table 2
Three-way ANOVA of behavioral performance of male spiders kept
on diets of different nutritional quality (enriched vs. deficient fruit
flies)
PC1—whole model
Diet
Carapace width
RCI
PC2—whole model
Diet
Carapace width
RCI
df
F
P
3,72
1
1
1
3,72
1
1
1
5.30
1.54
2.93
0.83
2.88
0.0045
4.83
3.24
0.0023
0.128
0.0045
0.41
0.042
0.95
0.031
0.076
Two PCs (PC1 and PC2) were extracted from 9 male behaviors (see
Table 1) recorded during 10-min confrontations with a female. PC1
reflects courtship intensity and PC2 mainly non-courtship behaviors.
All interactions were nonsignificant. df, degrees of freedom.
Of the single behaviors that loaded positively on PC1, abdomen vibrating, palp vibrating, hopping, and abdomen shaking
all showed the same pattern as PC1, correlating significantly
with carapace width, and with no diet dependence when carapace width is accounted for (Figure 2). Encircling and walking varied like PC2, correlating negatively with carapace width
though neither this nor the diet effect was significant.
Mating success
Of the 32 pairs staged with a female, in 3 pairs, neither of the
males was allowed to copulate. In 5 pairs, 3 of enriched and 2 of
deficient males, both a primary and a secondary mating were
observed. Twenty-four pairs had only a primary mating, of
which 14 were enriched diet pairs and 10 were deficient pairs.
In 5 pairs, agonistic behavior between the males was observed.
Four of these agonistic cases were actual fights, where the males
grappled for a few seconds, then they disengaged, and no further fighting was observed. In a single case, 1 male dominated
the other by threat posturing, extending its front legs above the
submissive male, which meanwhile remained in a lowered, resting position. The dominant male was subsequently successful
in mating. None of the encounters resulted in any apparent
injuries or fatalities.
Paired t-tests were first conducted on PC1 and PC2 including data from tests resulting in both a primary and a secondary
mating. Both PCs showed significant differences between winner and loser for enriched males but no difference for deficient males (Table 3). Excluding tests with secondary males,
the results were the same except that PC2 for enriched males
was now only marginally significant. Among enriched males,
4 courtship behaviors, namely, abdomen vibrating, palp vibrating, hopping, and copulation attempts, were used more intensively by successful than unsuccessful males (Figure 3A),
Lomborg and Toft
•
Nutrient balance and mating success
705
Figure 2
Intensities (proportion of time used) of 9 male behaviors (c–k) and 2 PCs (PC1 and PC2) extracted from the same behaviors (a, b) recorded
during 10 min of confrontation between a female and 1 male. The males were treated as subadults and adults with 1 of 2 diets (nutrient
enriched: closed circles, solid lines; nutrient deficient: open circles, dotted lines). Results of ANCOVAs with diet treatment as factor and ln (male
carapace width) as covariate are given in each panel. Degrees of freedom ¼ 1,73 for all. None of the interactions were significant.
whereas walking, palp shaking, and encircling were most intense in the unsuccessful males. For deficient spiders, the
differences were in the same direction though much smaller
(Figure 3B). Weight or RCI did not influence which male was
allowed to mate, but the size of the male mattered, as successful males had larger carapace width than unsuccessful males
(Table 3). Similar comparisons for carapace width, weight,
RCI, and the 9 behaviors revealed no significant differences
for deficient males.
Diet, size, or RCI showed no significant effect on the time
before the male was accepted or on copulation duration, nei-
ther was there any correlations between copulation latency or
copulation duration and PC1 or PC2. The overall average time
before copulation (mean 6 SE) was 34 6 9 min, and the duration of matings was 314 6 22 min.
DISCUSSION
The results supported our hypotheses that a balanced nutrition
confers males with an advantage in sexual competition and
that this may have been mediated by a more vigorous courtship
behavior of the nutrient-enriched males. When staging 2 males
Behavioral Ecology
706
Table 3
Results of paired t-tests of differences between successful and unsuccessful males of Pardosa prativaga from confrontation experiments of 2
males and 1 female.
Enriched (mean 6 SE)
Parameter
Winner; loser
PC1 including secondary males
PC2 including secondary males
PC1 excluding secondary males
PC2 excluding secondary males
Carapace width (mm)
Mass (mg)
RCI 3 100
1.23
0.34
1.36
0.35
2.05
15.0
0.15
6
6
6
6
6
6
6
0.18; 0.54
0.09; 0.84
0.19; 0.50
0.10; 0.91
0.03; 1.88
0.5; 14.5
1.31; 0.04
6
6
6
6
6
6
6
0.15
0.21
0.18
0.25
0.06
0.3
0.08
Deficient (mean 6 SE)
jtj
df
P
Winner; loser
2.91
2.19
2.92
1.86
3.16
1.79
0.29
17
17
13
13
13
13
13
0.0097*
0.0424*
0.012*
0.085
0.0076*
0.096
0.78
0.55
0.69
0.46
0.80
1.73
14.6
1.30
6
6
6
6
6
6
6
0.20; 0.36 6 0.24
0.21; 0.82 6 0.29
0.23; 0.25 6 0.27
0.24; 0.97 6 0.33
0.05; 1.80 6 0.06
0.7; 14.8 6 4.6
2.03; 1.25 6 1.94
jtj
df
P
0.77
0.44
0.70
0.45
0.71
0.85
0.18
11
11
9
9
9
9
9
0.45
0.66
0.50
0.65
0.49
0.42
0.86
Pairs were divided into either enriched (18/14 pairs) or deficient (12/10 pairs) males (including/excluding secondary male matings).
Parameters tested were carapace width, RCI, intensity of 9 different behaviors, and 2 PCs extracted from the 9 behaviors. P values significant at the
0.05 level are labeled with *. df, degrees of freedom.
of the same mass but different nutritional history with a female,
the male from the enriched diet treatment was successful at
mating. When staging 2 males of the same mass and nutritional
history but differing in previously measured courtship activity
with a female, we found that the one with the more active courtship behavior was successful in mating, a result also obtained by
Delaney et al. (2007). The differences between males of the
same treatment may have been due to genetic or to environmental factors from before they were captured. As courtship is
costly in terms of both energy use (Watson and Lighton 1994;
Kotiaho et al. 1998) and subsequent survival of the male
(Hoefler 2008) and thus probably also on condition factors
other than energy store, it was expected that a high performance depended not only on a sufficient amount of food but
also on a nutrition of high quality. However, due to a diet effect
on male size but not on body mass, the most active males in our
experiments were also larger than their less active competitors
of similar mass. Due to this unexpected pattern, we cannot
conclude to what extent each of the 2 factors were decisive.
Also, as we staged pairs of males, mate selection might be
due to either female choice or competition between males.
But the low level of direct aggression between the males even
in the limited space of the test containers indicated female
choice as the most likely mechanism. A combined effect might
consist of female choice for courtship intensity and male competitive dominance due to larger size, as argued to be the case
in another wolf spider (Delaney et al. 2007).
In male wolf spiders, success of mating with virgin females is
likely to be a good indicator of total reproductive success. Most
spiders have a first male precedence system of sperm competition (Austad 1984; Elgar 1998). This is at least partly true
Figure 3
Intensities (percentage of time
used) of courtship behaviors by
nutrient-enriched (a) and deficient (b) males of Pardosa prativaga, comparing the winning
(black) and losing (gray) males
of a 1 female–2 males competition setup. Paired males had
the same mass. Courtship
intensity was measured in a previous 1 female–1 male noncompetitive setup.
also for species of the genus Pardosa (Kiss B and Toft S, unpublished data). There is also evidence that males may be able
to sense the mating status of a female (Rypstra et al. 2003). As
every mating reduces the value of a female to later rival males,
there is severe competition between males for early access to
females. The ability to compose a nutritionally balanced diet,
normally considered a product of natural selection through
better survival, reduced developmental time, and larger adult
size (Toft 1999; Mayntz and Toft 2001; Uetz et al. 2002), will
also be enhanced through sexual selection when males in
good condition due to a balanced food choice achieve an
increased mating rate. Apart from an advantage in direct sexual competition, nutritionally balanced males may also benefit
indirectly due to a faster maturation and a longer survival.
Thus, they have a better chance of being present when the
females mature and still are virgins, and they may have a longer reproductive life span.
The mechanisms of obtaining an enhanced nutrient balance
can be pre- or postingestive, that is, through selection of food
type, differential ingestion of components of the selected food,
and through differential utilization of ingested and/or assimilated food (review in Raubenheimer and Simpson 1997). For P.
prativaga, Mayntz et al. (2005) documented an increased ingestion of prey that was rich in macronutrients the spiders were
short of. Thus, at least 1 mechanism of improving nutrient
balance is known from our experimental species. The enriched
flies fed to the spiders had a higher nitrogen content than the
deficient ones; however, as the enrichment was by means of dog
food, many other nutrients could have added to the effect.
Whereas 1 biometric indicator (carapace width) increased
under the enriched feeding treatment, RCI and body mass were
Lomborg and Toft
•
Nutrient balance and mating success
not good indicators of condition, as they failed to demonstrate
any difference between the 2 diet groups and were not related to
mating success. Wilder and Rypstra (2008) obtained the same
result with the same diet treatments as ours. Uetz et al. (2002)
found a diet effect on RCI in male wolf spiders (Schizocosa
ocreata), but their treatment differed in food quantity, not in
quality of prey. We believe that RCI and body mass failed as
condition indicators because we used food quality rather than
quantity as our food stress treatment. Nutritionally imbalanced
ad libitum feeding may lead to compensatory consumption
that may again result in increased weight per unit linear size.
As a matter of fact, we found a higher RCI in deficient than in
enriched males (Figure 1) associated with a reduced viability.
This result supports the advice from Rolff and Joop (2002) to
empirically establish the positive relationship between a condition indicator and some relevant fitness measures because such
relationships cannot be assumed. When used in connection
with a starvation treatment, RCI may correctly reflect the animal’s energy stores available for sexual signaling (Kotiaho
2000; Kotiaho et al. 2001). But fat stores only reflect condition
if they enhance fitness. Mayntz et al. (2003) found reduced
survival of nutrient-deficient ad libitum fed spiders but not of
spiders fed enriched prey in low food rations. Thus, reduced
viability due to nutrient imbalances may not follow from lack of
energy stores but from inability to properly utilize energy resources that may very well be present.
That size (carapace width) and not energy stores (weight and
RCI) was the determining factor of mating success makes
sense considering the information these 2 signals represent.
Although current energy stores may reflect the foraging success
of a male, they do not necessarily offer reliable signals for
females or rival males to make decisions on a male’s quality
or strength. Energy stores may fluctuate within short periods,
simply due to random variation in prey availability. Because it
is fixed at the maturity molt and results from high growth ratios
over several molts, final adult size offers information on the
lifetime foraging success and ability to avoid nutrient deficiency and may therefore be a more reliable indicator of
long-term feeding success. For animals with a visual display,
the ability to obtain surplus resources may be most effectively
communicated through the width of the carapace and possibly
the length and thickness of the legs. These structures are exposed in front of the opponent (female or male) and can be
evaluated even at a distance. This may also be the reason why
males of many sexually dimorphic wolf spiders have thickened
front legs or hair tufts on the front legs (e.g., Hebets and Uetz
2000; Stratton 2005). The cephalothorax and front legs are also
the organs used in direct contests with rivals, so investment in
these structures is likely to enhance fighting power.
In conclusion, manipulation of the nutritional value of the
diet has demonstrated that nutritional balance has consequences for size and the successful performance of courtship
and via these parameters for mating success of males. Our
results underscore the notion that selection for nutrient balancing has been an important evolutionary force not only for
herbivores and omnivores but also for predators (cf., Mayntz
et al. 2005) and that nutrient balance needs to be considered
also in studies of condition-dependent sexual signaling.
They also stress the need to use true viability parameters
rather than morphometric indices as expressions of animal
condition because the latter may not be of sufficient generality to reflect all variation in condition independently of the
causes of this variation.
FUNDING
Danish Research Council Forskningsrådet for Natur og
Univers; grant no. 272-05-0246 to S.T.
707
We thank Jens Hofman Hansen and Allan Nielsen of the Audio-Visual
division of the Department of Information and Media Studies, Aarhus
University, for guidance and access to equipment for image and video
analysis. We also thank David Mayntz, Janne Kotiaho, and 2 anonymous
reviewers for suggestions that improved the manuscript considerably.
REFERENCES
Andersson M. 1994. Sexual selection. Princeton (NJ): Princeton University Press.
Austad SN. 1984. Evolution of sperm priority patterns in spiders. In:
Smith RL, editor. Sperm competition and the evolution of mating
systems. Cambridge (MA): Harvard University Press. p. 223–249.
Bertram SM, Schade JD, Elser JJ. 2006. Signalling and phosphorous:
correlations between mate signalling effort and body elemental
composition in crickets. Anim Behav. 72:899–907.
Cotton S, Fowler K, Pomiankowski A. 2004. Do sexual ornaments
demonstrate heightened condition-dependent expression as predicted by the handicap hypothesis? Proc R Soc Lond B Biol Sci.
271:771–783.
Delaney KJ, Roberts AJ, Uetz GW. 2007. Male signaling behavior and
sexual selection in a wolf spider (Araneae: Lycosidae): a test for
dual functions. Behav Ecol Sociobiol. 62:67–75.
Den Hollander J, Dijkstra H, Alleman H, Vlijm L. 1973. Courtship
behaviour as species barrier in the Pardosa pullata group (Araneae,
Lycosidae). Tijdschr Entomol. 116:1–22.
Elgar MA. 1998. Sperm competition and sexual selection in spiders
and other arachnids. In: Birkhead TR, Møller AP, editors. Sperm
competition and sexual selection. San Diego (CA): Academic Press.
p. 307–339.
Fagan WF, Denno RF. 2004. Stoichiometry of actual vs. potential predatorprey interactions: insights into nitrogen limitation for arthropod predators. Ecol Lett. 7:876–883.
Fagan WF, Seimann E, Mitter C, Denno RF, Huberty AF, Woods HA,
Elser JJ. 2002. Nitrogen in insects: implications for trophic complexity and species diversification. Am Nat. 160:784–802.
Fisher HS, Rosenthal GG. 2006. Female swordtail fish use chemical
cues to select well-fed mates. Anim Behav. 72:721–725.
Foelix RF. 1996. Biology of spiders. 2nd ed. New York: Oxford University Press.
Garcı́a-Berthou E. 2001. On the misuse of residuals in ecology: testing
regression residuals vs. the analysis of covariance. J Anim Ecol.
70:708–711.
Grafen A. 1990. Biological signals as handicaps. J Theor Biol. 144:
517–546.
Hebets EA, Uetz GW. 2000. Leg ornamentation and the efficacy of
courtship display in four species of wolf spider (Araneae: Lycosidae). Behav Ecol Sociobiol. 47:280–286.
Hebets EA, Wesson J, Shamble PS. 2008. Diet influences mate choice
selectivity in adult female wolf spiders. Anim Behav. 76:355–363.
Hill GE, Montgomerie R. 1994. Plumage colour signals nutritional
condition in the house finch. Proc R Soc Lond B Biol Sci. 258:
47–52.
Hoefler CD. 2008. The costs of male courtship and potential benefits
of male choice for large mates in Phidippus clarus. J Arachnol. 36:
210–212.
Hunt J, Brooks R, Jennions MD, Smith MJ, Bentsen CL, Bussière LF.
2004. High-quality male field crickets invest heavily in sexual display
but die young. Nature. 432:1024–1027.
Jakob EM, Marshall SD, Uetz GW. 1996. Estimating fitness: a comparison of body condition indices. Oikos. 77:61–67.
Kaspi R, Taylor PW, Yuval B. 2000. Diet and size influence sexual
advertisement and copulatory success of males in Mediterranean
fruit fly leks. Ecol Entomol. 25:279–284.
Kodric-Brown A. 1989. Dietary carotenoids and male mating success in
the guppy: an environmental component to female choice. Behav
Ecol Sociobiol. 25:393–401.
Kotiaho JS. 1999. Estimating fitness: comparison of body condition
indices revisited. Oikos. 87:399–400.
Kotiaho JS. 2000. Testing the assumptions of conditional handicap
theory: costs and condition dependence of a sexually selected trait.
Behav Ecol Sociobiol. 48:188–194.
Kotiaho JS, Alatalo RV, Mappes J, Nielsen MG, Parri S, Rivero A. 1998.
Energetic costs of size and sexual signaling in a wolf spider. Proc R
Soc Lond B Biol Sci. 265:2203–2209.
708
Kotiaho JS, Alatalo RV, Mappes J, Parri S. 1999a. Honesty of agonistic
signalling and effects of size and motivation asymmetry in contests.
Acta Ethol. 2:13–21.
Kotiaho JS, Alatalo RV, Mappes J, Parri S. 1999b. Sexual signalling and
viability in a wolf spider (Hygrolycosa rubrofasciata): measurements under laboratory and field conditions. Behav Ecol Sociobiol. 46:123–128.
Kotiaho JS, Simmons LW, Tomkins JL. 2001. Towards a resolution of
the lek paradox. Nature. 410:684–686.
Kronestedt T. 1979. Etologiska karaktärer vid taxonomiska studier av
vargspindlar. Entomol Tidskr. 100:194–199.
Krzanowski WJ. 2000. Principles of multivariate analysis: a user’s perspective. Oxford: Oxford University Press.
Mappes J, Alatalo RV, Kotiaho J, Parri S. 1996. Viability costs of condition-dependent sexual male display in a drumming wolf spider.
Proc R Soc Lond B Biol Sci. 263:785–789.
Martin J, Lopez P. 2006. Vitamin D supplementation increases the
attractiveness of males’ scent for female Iberian rock lizards. Proc
R Soc Lond B Biol Sci. 273:2619–2624.
Mayntz D, Raubenheimer D, Salomon M, Toft S, Simpson SJ. 2005.
Nutrient-specific foraging in invertebrate predators. Science. 307:
111–113.
Mayntz D, Toft S. 2001. Nutrient composition of the prey’s diet affects
growth and survivorship of a generalist predator. Oecologia.
127:207–213.
Mayntz D, Toft S, Vollrath F. 2003. Effects of prey quality and availability
on the life history of a trap-building predator. Oikos. 101:631–638.
McGraw KJ. 2007. Dietary mineral content influences the expression
of melanin-based ornamental coloration. Behav Ecol. 18:137–142.
Ohlsson T, Smith HG, Råberg L, Hasselquist D. 2001. Pheasant sexual
ornaments reflect nutritional conditions during early growth. Proc
R Soc Lond B Biol Sci. 269:21–27.
Pyke DA, Thompson JN. 1986. Statistical analysis of survival and removal rate experiments. Ecology. 67:240–245.
Raubenheimer D, Jones SA. 2006. Nutritional imbalance in an extreme generalist omnivore: tolerance and recovery through complementary food selection. Anim Behav. 71:1253–1262.
Raubenheimer D, Mayntz D, Simpson SJ, Toft S. 2007. Nutrientspecific compensation following diapause in a predator: implications for intraguild predation. Ecology. 88:2598–2608.
Behavioral Ecology
Raubenheimer D, Simpson SJ. 1997. Integrative models of nutrient
balancing: application to insects and vertebrates. Nutr Res Rev.
10:151–179.
Richter CJJ, Den Hollander J, Vlijm L. 1971. Differences in breeding
and motility between Pardosa pullata (Clerck) and Pardosa prativaga
(L. Koch), (Lycosidae, Araneae) in relation to habitat. Oecologia.
6:318–327.
Rolff J, Joop G. 2002. Estimating condition: pitfalls of using weight as
a fitness correlate. Evol Ecol Res. 4:931–935.
Rowe L, Houle D. 1996. The lek paradox and the capture of genetic
variance by condition dependent traits. Proc R Soc Lond B Biol Sci.
263:1415–1421.
Rypstra AL, Wieg C, Walker WE, Persons MH. 2003. Mutual mate
assessment in wolf spiders: differences in the cues used by males
and females. Ethology. 109:315–325.
Simpson SJ, Raubenheimer D, Behmer ST, Whitworth A, Wright GA.
2002. A comparison of nutritional regulation in solitarious- and
gregarious-phase nymphs of the desert locust Schistocerca gregaria.
J Exp Biol. 205:121–129.
Smith HG, Råberg L, Ohlsson T, Granbom M, Hasselquist D. 2006.
Carotenoid and protein supplementation have differential effects
on pheasant ornamentation and immunity. J Evol Biol. 20:
310–319.
Stratton GE. 2005. Evolution of ornamentation and courtship behavior in Schizocosa: insights from a phylogeny based on morphology
(Araneae: Lycosidae). J Arachnol. 33:347–376.
Toft S. 1999. Prey choice and spider fitness. J Arachnol. 27:301–307.
Uetz GW, Bischoff J, Raver J. 1992. Survivorship of wolf spiders
(Lycosidae) reared on different diets. J Arachnol. 20:207–211.
Uetz GW, Papke R, Kilinc B. 2002. Influence of feeding regime on
body size, body condition and a male secondary sexual character
in Schizocosa ocreata wolf spiders (Araneae, Lycosidae): conditiondependence in a visual signalling trait. J Arachnol. 30:461–469.
Watson PJ, Lighton JRB. 1994. Sexual selection and the energetics of
copulatory courtship in the sierra dome spider, Linyphia litigiosa.
Anim Behav. 48:615–625.
Wilder SM, Rypstra AL. 2008. Diet quality affects mating behaviour and egg production in a wolf spider. Anim Behav. 76:
439–445.