Document Sample
					O R I G I NA L A RT I C L E


Mats Olsson,1,2 Mo Healey,1 Erik Wapstra,3 and Tobias Uller1,4
    School of Biological Sciences, The University of Wollongong, Wollongong, NSW 2522, Australia
         E-mail: molsson@uow.edu.au
    School of Zoology, The University of Tasmania, Hobart 7001, Tasmania, Australia
    Edward Grey Institute, University of Oxford, United Kingdom

     Received February 18, 2008
     Accepted October 13, 2008

     In the Australian painted dragon lizard (Ctenophorus pictus), males occur in two different morphs with respect to gular color,
     with or without a yellow bib. Males without a bib lost within-clutch paternity significantly more often to rivals than bibbed
     males. Thus, it appears that bibs identify some phenotypic advantage linked to competitive ability. To test whether this could be
     related to whole-organism capacity to withstand an increased workload (due to better health and vigor, or evolved differences in
     self-maintenance), we implanted males with a lead pellet (loaded), Styrofoam pellet (controls), or sham-operated males without
     implants (shams), and compared male categories with respect to how they maintained body mass during the mating season.
     Somewhat unexpectedly, bibbed males consistently lost more body weight across all treatments and controls, although we could
     not verify that this translated into higher mortality in this short-lived animal (about 80% survive for one year only). However,
     bibbed males may invest more into “mating success” than nonbibbed males, which agrees with our experimental results and
     paternity data.

     KEY WORDS:        Bib, costly signaling, handicap principle, lizard, multiple paternity.

The handicap theory of sexual selection is based upon the as-                   (Hausken and Hirschleifer 2008; for an excellent review across
sumption that the costs of displaying a trait associated with re-               this field see Searcy and Nowicki 2005).
productive benefits keep the signal honest and reliable, that is,                     Condition-dependent signaling also underpins the important
it cannot be invaded by a mutant strategy that reaps the benefits               link between sexual selection and life-history evolution, namely
without paying the price (Zahavi 1975, 1977). Following much                    the relative investments into mate acquisition, self maintenance,
debate regarding a lack of formalism and the validity of its orig-              and longevity (Kokko et al. 2002; Kokko 2007), in which sce-
inal description (e.g., Maynard Smith 1976), Andersson (1986)                   nario the signaling process can be seen as a component of mate
made an important advance in the formal analysis of the handicap                acquisition either from a perspective of “good-genes” processes
theory by showing theoretically that variation in heritable sexual              or Fisherian selection (Kokko et al. 2002; Getty 2006). Empiri-
ornaments can be based on viability and evolve in the absence                   cal evidence for good-genes models has been considered positive
of Fisherian processes, if the sexually selected trait is condition             viability effects on offspring by attractive males, whereas null or
dependent (Andersson 1986). Since then, a number of models                      negative effects on the mating success of sons have been inter-
of the handicap principle, or components thereof, have demon-                   preted as support for a Fisherian process (Kokko et al. 2002).
strated that it can be robust for both nonheritable quality traits              However, recent theory lends support to the notion that this is
(e.g., Grafen 1990a, b; reviews in Johnstone 1995; Hasson 1997;                 a single process because enhanced mating success and survival
Hausken and Hirschleifer 2004) and for heritable quality traits                 are equally valid genetic benefits (Kokko et al. 2002) where

                       C 2009 The Author(s). Journal compilation   C   2009 The Society for the Study of Evolution
             695       Evolution 63-3: 695–701
M AT S O L S S O N E T A L .

relative fitness is the product of the two. Consequently, the rel-
ative investments into mating success (i.e., relatively more re-
sources invested into mate acquisition and paternity assurance)
versus self-maintenance (i.e., longevity) can be seen as selected
tactics representing different reproductive trade-off strategies.
      Although the theoretical literature on signaling principles
abound, there is still a paucity of empirical work testing their valid-
ity across taxa in natural populations (Andersson 1994; Johnstone
1995). By far the most commonly studied scenario is in the con-
text of female choice, which is lacking in many species (includ-
ing our model species, Healey et al. 2007; Olsson et al. 2007b).
However, signals also form important components of interindi-
vidual communication between sexual and nonsexual rivals, not             Figure 1.   The photographs depict painted dragon (Ctenophorus
only potential partners. For example, badges evolve to aid in-            pictus) males with (A) and without (B) a bib.
dividual recognition and reduce the frequency of costly interac-
tions (Rohwer 1975, 1976), or convey resource-holding power               in nonbibbed males would be consistent with a good-genes sce-
and form integral parts of tactical decision making when to en-           nario with better self-maintenance in bibbed males. We discuss
gage or not in contests for resources. An example of this is badge        our results from a perspective of life-history trade-offs between
“green-ness” in sand lizards (Lacerta agilis; Olsson 1994), and           bib categories.
epaulettes in red-shouldered widowbirds (Euplectes axillaries;
Pryke and Andersson 2003).
      What constitutes the benefit to the female for exercising a         Materials and Methods
costly choice on male signals, or for rivals to consider a signal         FIELD PROTOCOL
in contests, is often far from obvious. Importantly, most exper-          The Australian painted dragon (C. pictus) is a small (adult snout-
imental studies have approached this question by manipulating             vent length (SVL) = 65–95 mm, mass = 8–16 g) diurnal lizard.
the signal to establish its relation to dominance or female mate          Males occur in different head colors (red, orange, yellow; Healey
choice, and ultimately reproductive success, whereas tests of dif-        et al. 2007; Olsson et al. 2007a). Furthermore, regardless of head
ferences in the innate ability of males to withstand the cost of trait    color, males occur in two different phenotypes with respect to
development and associated tactics are rare. This is an unfortunate       the color under their chins, with a bib (yellow, carotenoid-based,
omission because the cost for trait expression has been predicted         C. Isaksson, unpubl. data) or without a bib (white, same as rest
to be higher for males in poorer condition (e.g., Andersson 1994;         of ventral surface; Fig. 1). They typically prefer sandy habitats
Grafen 1990a,b; Getty 2006), and increasing such costs through            and low vegetation in a range covering central and western New
manipulation would be predicted to be higher with a life-history          South Wales to Western Australia. We studied a desert popula-
trade-off that favors relatively lower signaling costs and a longer       tion at Yathong Nature Reserve, New South Wales; 145◦ 35 E;
life span (Kokko et al. 2002; Searcy and Nowicki 2005).                   32◦ 35 S, by monitoring individually marked lizards during the
      In the present article, we exploit a lizard model (the Aus-         mating season (August–January) in 2000 and 2005, and by per-
tralian painted dragon, Ctenophorus pictus) in which males can            forming the field experiment described below in 2007. We caught
be classified into two different phenotypes, with and without a           males by noosing from a slow-moving (and then stopping) ve-
throat bib. We first show that bib expression is positively linked        hicle. Captured lizards were weighed, measured (only relevant
to resource-holding potential in terms of fertilization success of        measures reported), blood sampled (from Vena angularis in the
females on male territories in a natural population. We subse-            corner of the mouth), and scored for head coloration and bib pres-
quently tested whether bib expression is linked to the ability to         ence/absence by eye (following Sinervo and Lively 1996). Blood
maintain body mass under experimental elevation of workload               or toe/nail clips were stored in 96% ethanol until DNA extraction.
in a natural population during the mating season by surgically                  Male paternity was assigned following standard protocol
implanting males in three different categories, with a lead pellet        (Olsson et al. 2007a,b; Schwartz et al. 2007) using microsatel-
(“loaded”), with a Styrofoam pellet (“controls”), and with no im-         lite markers in CERVUS 2.0 (Marshall et al. 1998). DNA was
plant (“shams”). Based on this experimental design we can make            isolated from blood samples (2000) and toe or tail clips from
the following predictions: (1) A greater mass loss in bibbed males        2005 using a Qiagen DNAeasy Tissue Extraction Kits (Qiagen,
would suggest a higher investment into aspects of mate acqui-             Doncaster, Victoria, Australia). DNA samples from 2000 to 2005
sition in bibbed than nonbibbed males. (2) A greater mass loss            were PCR amplified at the six microsatellite loci CP01, CP02,

            696      EVOLUTION MARCH 2009
                                                                                                         S I G NA L I N G I N A L I Z A R D

CP06, CP10, CP11, and CP17 (Schwartz et al. 2007), using flu-         paternity could be successfully assigned at 95% confidence in
orescently labeled primers. The samples from 2005 were am-            80% of cases with the complete exclusion in both years (only
plified at the following eight microsatellite loci CP01, CP02,        80% because we conservatively assumed we had only sampled
CP06, CP10, CP11, CP17, AM41 (Schwartz et al. 2007), and              80% of males in population). Assuming a 0.01 error rate, paternity
C7 (Austin et al. 2006) using fluorescently labeled primers. All      was estimated to be successfully assigned at the 95% confidence
loci were amplified in 7 μl reactions either singleplexed or mul-     level in 62% and 64% of cases in 2000 and 2005, respectively.
tiplexed using 10–20 ng of DNA; 0.22 mM each dNTP; 0.3–0.7            The fathers that were identified after comparing the three types
μM of each primer; 1× PCR buffer (Qiagen) containing Tris-            of analyses never had more than one allele mismatch and always
HCL, KCL(NH 4 ) 2 SO 4 , 15 mM MgCl 2 pH 8.7; additional MgCl 2       fell within the 95% confidence level or the 80% confidence level
(final concentration of MgCl 2 ranged from 1.5 to 3.5 mM, de-         with siblings having been assigned the same father at the 95%
pending on the loci being amplified); and 0.05 U of HotStart Taq      confidence level.
(Qiagen). Cycling conditions included a hot start denaturation at          Data on the role of the bib in antagonistic interactions are
95◦ C for 15 min; 30 cycles of 95◦ C for 30 sec, annealing temper-    lacking. However, using paternity data from natural populations,
ature for 30 sec, 72◦ C for 30 sec; and a final extension at 72◦ C    we were able to assign males to categories that suffered versus
for 30 min. Fluorescent amplifications were electrophoresed on a      those that did not suffer from loss of paternity to rivals (i.e., in
genetic analyzer (ABI3130×l—2000 and ABI3900—2005: Ap-                clutches of females residing in male territories) via some unknown
plied Biosystems Australia, Melbourne, Victoria, Australia) us-       pathway (e.g., decreased success in mate acquisition or sperm
ing the LIZ500 size standard (Applied Biosystems). Alleles were       competition) to assess to what extent bib expression is linked to
scored using GENEMAPPER (Applied Biosystems) and confirmed            competitive ability in a natural setting. Previous work has shown
visually.                                                             that about 80% of painted dragons only survive for one year and
      Parentage analyses were conducted independently for each        that a substantial part of male paternity is set by matings early
year using both the exclusion and likelihood-based approaches         in the season and that offspring are sired posthumously by sperm
in CERVUS 2.0 (Marshall et al. 1998). Genotypes from the adults       stored in females (Olsson et al. 2007b).
were used to calculate allele frequencies, observed and expected
heterozygosites, frequency of null alleles, and the polymorphic       FIELD EXPERIMENT 2007
information content (PIC: a measure of informativeness related to     At the onset of the mating season (early September) noosed lizards
expected heterozygosity) of the loci (Botstein et al. 1980; Hearne    were weighed, measured, and scored for head and bib coloration at
et al. 1992; Schwartz et al. 2007). Mother–offspring genotypes        about 36◦ C, their preferred body temperature (following Sinervo
were compared for the presence of null alleles and loci containing    and Lively 1996), and brought to a laboratory facility at Yathong
null alleles were deleted from the analyses. All males sampled        Nature Reserve. They were then kept overnight, and in the fol-
in the population were assumed to be potential fathers for all        lowing morning participated in an implantation experiment to test
offspring in the year they had been sampled (except for five in-      whether males of different bib status (present/absent) differed in
dividuals in 2005 that had less than four loci genotyped). For        their ability to sustain their body mass in the wild, while engag-
these analyses we assumed that 80% of the males in the popula-        ing in normal mating season activities (e.g., territory and mate
tion were sampled. To determine 95% and 80% confidence levels         acquisition, sperm production, and antagonistic interactions).
for the parentage assignments, two types of simulations were run
on each dataset: the first assumed a zero genotyping error rate       SURGICAL IMPLANTATION
(complete exclusion), and the second assumed a 0.01 genotyp-          Males were haphazardly assigned to either of three treatments:
ing error rate. The datasets were then corrected for any potential    (1) Lead pellet implantation (hereafter “loaded males”). These
null alleles by deleting all homozygous genotypes at loci CP02,       males received a 2.7 g lead pellet, approximately 8 mm in diam-
CP10, CP17, and AM41 (loci most likely to have null alleles)          eter, coated with sterile, surgical silicone prior to implantation.
from the potential fathers and assigning parentage a third time us-   The mass of the pellet was derived from a calculation of body
ing a zero genotyping error rate simulation. Results from all three   condition index (residuals from a mass − SVL regression) based
analyses were cross-checked and assigned fathers were verified        on morphometric data collected from previous years. This was
by eye across mother–offspring pairs and the clutchmates. Any         matched to the males in “best condition,” meaning that the in-
mismatches between father–offspring–mother were checked with          creased mass was within the natural limits of the mass in relation
the raw data.                                                         to skeletal size a male ever carried in the wild. (2) A second
      Based on the loci used, the probability of exclusion when the   group of males received a Styrofoam pellet (hereafter “controls”)
mother’s genotype was known was 0.999 for both years’ datasets.       of the same diameter as the lead pellet, coated with surgical sili-
Simulations estimated that with the known mother’s genotype,          cone, with a mass of 0.001 g. The rationale for this treatment was

                                                                                        EVOLUTION MARCH 2009             697
M AT S O L S S O N E T A L .

that males may suffer weight loss from either carrying an extra         results. We therefore only report the results from the first analy-
load (lead) or from a reduction in ingestion volume (because the        sis here. We started with a full model and backward-eliminated
implant took up some space in the body cavity that would other-         factors at P > 0.25, starting with higher order interactions, and
wise have been available for inflating the stomach during ad-           report the result of the final model. Mass at release did not differ
ditional foraging). (3) A third group was operated on as per            among treatment categories (F 2,35 = 0.30, P = 0.74) or between
above-described procedure but received no implant. The oper-            bib categories (F 1,36 = 0.04, P = 0.84).
ation procedure was conducted as follows: a male was sterilized              Paternity loss was analyzed using likelihood-ratio tests based
ventrally with alcohol and injected subcutaneously with 0.2 mL          on a single clutch per male.
Lidocaine (Lidocaine hydrochloride 0.4%; Sigma), on the right
ventral side. The male was then cooled in a sterile plastic box
on ice (being ectotherms, this immobilizes lizards without caus-        Results
ing any harm). The male was then turned over on its back and            MATING SYSTEM AND BIB STATUS IN 2000
immobilized on a sterile glass surgical table using sterile tape,       AND 2005
thereafter it was wiped ventrally with 70% alcohol a second time.       Males without a bib lost paternity to rivals four times as often
A small incision was made, the sterile implantation took place          as bibbed males did (i.e., at least one young was sired by a male
(not in shams), and the male was closed with surgical superglue         different to the territory-holding male; territory in which a female
(2-octyl cyanoacrylate; Hystoacryl). All males appeared to be           resided; 19% vs. 5% of clutches; likelihood-ratio χ2 = 4.1, df =
in perfect condition at release, that is on the day subsequent to       1, P = 0.040; Fig. 2). There was no difference in the number of
implantation/sham operation, and at the subsequent recapture.           observations made in the wild of bibbed (SE = 5.9 ± 0.9, n =
      All males were then kept in a clean cloth bag for 24 h, checked   39) versus nonbibbed males (SE = 6.6 ± 0.7, n = 53; t-test, t =
for any discomfort, and had individual numbers painted on their         0.62, P = 0.54, df = 90).
backs prior to being released at their places of capture. The date
and position of males at resightings were monitored with global         MASS LOSS: EFFECTS OF HEAD COLOR, BIB,
position system (GPS) coordinates. After four weeks we recap-           AND TREATMENT
tured as many males as possible, resulting in the recapture of 40       Head color did not show a significant effect on male mass loss and
of the 74 males. We describe elsewhere (Olsson et al. 2007b) that       was removed from the final analysis. Treatment, bib, and SVL all
male territories are close to one-dimensional on the embankments        significantly influenced mass loss (Table 1). All male categories
of the fire trails forming our study site, vegetation is too dense      showed mass loss during the treatment period but none of the
on the sides of the embankments, and the open road typically            recaptured or resighted males (all could not be captured) were in
discourages lizard activities (probably due to risk of predation).      poor condition (e.g., all males had excellent muscle tonus and alert
Thus, although males occasionally take refuge from predators            behavior). There was no difference in mass loss between loaded
in the surrounding vegetation, they typically patrol their territo-     and control males, suggesting that the main effect of implanta-
ries only from perch sites along the trail embankments and we           tion was a reduction in ingestion volume, rather than energetic
therefore used the longest axis of their territories as a proxy for     expenditure (mass loss, 0.82 ± 0.22 g, n = 11, vs. 0.84 ± 0.15 g,
territory size. Territory size was successfully normalized by log       n = 12, respectively; t-test, t = 0.06, df = 21, P = 0.95), with
transformation (Wilks’ lambda, W = 0.97, P = 0.40, n = 47).             the main treatment effect being in comparison with sham males
      This study was approved by and conducted in compliance            (mean mass loss, SE = 0.27 ± 0.13 g, n = 15; Table 1). Bibbed
with Wollongong University animal ethics protocols AE 04/03,            males consistently lost more mass than nonbibbed males with an
AE 04, and AE 05. A scientific license was also issued under the        overall average mass loss for bibbed males across categories of
National Parks and Wildlife Act 1974 by the National Parks and          0.82 g (SE = ± 0.15, n = 17) compared to 0.44 g for nonbibbed
Wildlife Service, NSW, Australia.                                       males, respectively ( SE = ± 13, n = 21; t-test, t = 2.8, df = 21,
                                                                        P = 0.015). Finally, there was a significant interaction between
STATISTICAL METHODS                                                     treatment and bib category, showing that bibbed males main-
We analyzed mass loss ([[recapture mass − release mass] − im-           tained their body mass less successfully than nonbibbed males in
plant mass]; response variable) in general linear model with re-        response to the experimental treatment (Table 1; Fig. 3).
spect to three main effects in Proc GLM (SAS Institute); head                To test whether other differences between male categories
color (red, orange, yellow), bib (bib or no bib), treatment (loaded,    could contribute to energetic expenditure, we also looked for
controls, shams), with SVL as a covariate in the model. A model         differences in number of observations (activity) and territory size
with mass at recapture as response variable and mass at release as      (mobility) among treatment categories. However, there was no
a predictor gave qualitatively similar and statistically significant    effect of bib or treatment category on territory size (P > 0.16 for

           698       EVOLUTION MARCH 2009
                                                                                                                                                                          S I G NA L I N G I N A L I Z A R D


                          % of clutches with multiple vs. single paternityfff



                                                                                                                    n = 10

                                                                                                                                                                     n = 43
                                                                                                                                      n = 37

                                                                                         Bibbed                  Non Bibbed           Bibbed                   Non Bibbed

                                                                                                Mult iple Pat ernit y                          S ingle Pat ernit y

Figure 2. Clutches with multiple paternity (i.e., with at least one young sired by a male different to the territory-holding male, territory
in which a female resided) and single paternity in bibbed versus nonbibbed males. Sample sizes are given within the bars in the diagram.
Olsson et al. (2007b) show that densities of perch sites can affect risk of multiple paternity (less risk with more perch sites), but there
was no difference in perch density between bib categories (F 1,53 = 0.19, P = 0.66).

all factors and interactions; mean territory size = 45.7 ± 6.8 m,                                                             and in drumming spiders (Hygrolycosa rubrofasciata), traits im-
n = 47) or on number of observations (P > 0.30 for all factors                                                                portant for mate choice (e.g., duration of drumming) were highly
and interactions; mean number of observations SE = 4.0 ± 0.3,                                                                 repeatable whereas this did not apply to traits unimportant for
n = 47). Likewise, we could not identify any difference among                                                                 male reproductive success (e.g., frequency of drumming; Rivero
treatment categories in the last day of the season, males were                                                                et al. 2000).
observed as a proxy for survival (P > 0.13 for all factors and                                                                      Our proxy of competitive ability (lost paternity to rivals) sug-
interactions; mean Julian date SE = 17438.7 ± 1.53, n = 47).                                                                  gests that bibbed males would suffer less paternity loss to other
                                                                                                                              males, either through better resource-holding power or through
                                                                                                                              greater success in sperm competition. Thus, we would expect
Discussion                                                                                                                    males with stark signals (bibs) to have both better physiological
Most examples in sexual selection biology have demonstrated                                                                   performance and to be more competitive. Instead, we demon-
significant positive links between trait expression, some aspect                                                              strate that nonbibbed males better withstand an extra physiolog-
of condition or performance, and concomitant measures of com-                                                                 ical work load and/or reduced foraging efficiency. What may
ponents of fitness (Andersson 1994). For example, work on sand                                                                explain this? First, in the light of sexual selection from a per-
lizards (L. agilis) showed that male body mass and condition                                                                  spective of life-history trade-offs, our results would make perfect
were significantly correlated with color traits and male reproduc-                                                            sense (i.e., higher investment into “mating success” components
tive success (Olsson 1994). Similar work on Jackson’s widow                                                                   in “signalers” Kokko et al. 2002) if there were accompanying
birds showed that male display rate was significantly correlated                                                              reductions in bibbed male survival. We cannot verify this with
with male tail length and reproductive success (Andersson 1989),                                                              our data—males appeared to live for as long in terms of recorded
                                                                                                                              recapture rates (and dates) regardless of bib and treatment cate-
Table 1. Effects of bib (bib, no bib) and treatment (lead, foam,                                                              gory. Furthermore, in this species > 80% of all males live for only
and sham) on mass loss in males after three weeks exposure time.                                                              a single year and a substantial proportion of offspring sired late
(Model F 6,31 =5.2, P=0.0009, R2 =0.50).                                                                                      in the season are posthumously fathered by sperm stored from
                                                                                                                              much earlier matings (Olsson et al. 2007b, unpubl. ms.). Thus,
 Source               df                                                        Type III SS   MS         F          P
                                                                                                                              sperm storage should contribute to decouple otherwise coevolved
 Treatment            2                                                         3.71          1.85       7.68       0.002     trade-offs between mating success and survivorship as modeled
 Bib                  1                                                         1.56          1.56       6.45       0.016     in Kokko et al. (2002); thus, even though such effects would be
 Treat×bib            2                                                         1.89          0.94       3.91       0.030     expected theoretically, demonstrating them in this species in the
 SVL (covariate)      1                                                         1.00          1.00       4.17       0.049
                                                                                                                              wild would be very hard. Similar systems in which prolonged

                                                                                                                                                  EVOLUTION MARCH 2009                    699
M AT S O L S S O N E T A L .



                         Mass loss (g)aa


                                                 n=6             n=4
                                                                                         n=7            n=5             n=7
                                                 Bibbed     Non bibbed   Bibbed      Non Bibbed        Bibbed       Non Bibbed

                                                          Lead                 Control                          Sham

Figure 3.   Interaction effect between treatment (lead, foam, and sham males), and male bib status (present or absent) with respect to
mass loss over the experimental period (Model F 6,31 = 5.2, P = 0.0009, R2 = 0.50; interaction effect, df = 2, F = 1.89, P = 0.030; Table 1).

sperm storage make up significant parts of male reproductive                      dity). Our results provide experimental evidence that high-quality
success are commonplace in the wild across taxa (e.g., Zamudio                    signalers in the painted dragon lizard are less able to withstand an
and Sinervo 2000), and the importance of this effect needs to                     enforced energetic expenditure, mediated via an increased work-
be assessed theoretically. Second, although our evidence suggests                 load or reduced foraging capacity, suggesting that perhaps their
that the presence of a bib has a positive effect on reproductive suc-             investment into self-maintenance is lower than for low-quality
cess, its effect on overall fitness may not be straightforward. In C.             signalers.
pictus, males may differ in their reproductive tactics in that bibbed
males are more active in territorial defense or other metabolically               ACKNOWLEDGMENTS
costly exercises. We doubt this interpretation, as males of both                  We are grateful to the Australian Research Council (MO) and The
                                                                                  Wenner-Gren Foundations (TU) for financial support, to the Snaith fam-
morphs were observed equally often, have similar territory size,                  ily for logistic support, G. Swan for field assistance, and T. Schwartz for
and there was no apparent difference in their behavior such as the                paternity assignment.
dominance effects we have observed for males of different head
color (Healey et al. 2007). Thus, it seems more likely that males                 LITERATURE CITED
differ in some, as yet, unmeasured innate quality, such as basal                  Andersson, M. 1986. Evolution of condition-dependent sex ornaments and
metabolic rate or parasite resistance, mediated via genetic mecha-                     mating preferences: sexual selection based on viability differences. Evo-
                                                                                       lution 40:804–816.
nisms, or via condition dependence (not known). Furthermore, in                   ———. 1989. Sexual selection and cues for female choice in leks of Jackson’s
this experimental study, we did not find a difference in mass loss                     widowbird Euplectes jacksoni. Behav. Ecol. Sociobiol. 25:403–410.
depending on head color unlike in a previous correlative study,                   ———. 1994. Sexual selection. Princeton Univ. Press, Princeton, NJ.
which identified thinner red males in some years (Olsson et al.                   Austin, J. J., R. J. Rose, and J. Melville. 2006. Polymorphic microsatellite
                                                                                       markers in the painted dragon lizard, Ctenophorus pictus. Mol. Ecol.
2007a). The most likely explanation for this seems to be that the
                                                                                       Notes 6:194–196.
effects of bib status and treatment supersede any effects of head                 Botstein, D., R. L. White, M. Skolnick, and R. W. Davis. 1980. Construc-
color in our experimental study.                                                       tion of a genetic-linkage map in man using restriction fragment length
      Although our evidence for a lower ability to withstand an in-                    polymorphisms. Am. J. Hum. Genet. 32:314–331.
                                                                                  Getty, T. 1998. Handicap signaling: when fecundity and viability do not add
creased workload and/or cope with a reduced foraging efficiency
                                                                                       up. Anim. Behav. 56:127–130.
for bibbed males seems to dispute the “honesty” of the signal, this               ———. 2002. Signalling health versus parasites. Am. Nat. 159:363–371.
may be a simplistic conclusion. As noted by Getty (1998, 2002,                    ———. 2006. Sexually selected signals are not similar to sports handicaps.
2006), Kokko et al. (2002), and others (Hausken and Hirschleifer                       Trends Ecol. Evol. 21:83–88.
2008), signaling theory may sometimes be too casually invoked                     Grafen, A. 1990a. Sexual selection unhandicapped by the Fisher process. J.
                                                                                       Theor. Biol. 144:473–516.
into sexual selection theory because of the multiplicative nature
                                                                                  ———. 1990b. Biological signals as handicaps. J. Theor. Biol. 144:517–546.
of sexual selection (costs paid in reduced maintenance, viability,                Hasson, O. 1997. Towards a general theory of biological signaling. J. Theor.
and survival are multiplied by benefits in reproduction or fecun-                      Biol. 185:139–156.

            700      EVOLUTION MARCH 2009
                                                                                                                           S I G NA L I N G I N A L I Z A R D

Hausken, K., and J. Hirschleifer. 2004. The truthful signaling hypothesis: an    Pryke, S., and S. Andersson. 2003. Carotenoid-based epaulettes reveal male
     explicit general equilibrium model. J. Theor. Biol. 228:497–511.                 competitive ability: experiments with resident and floater red-shouldered
———. 2008. Truthful signaling, the heritability paradox, and the Malthusian           widowbirds. Anim. Behav. 66:217–224.
     equi-marginal principle. Theor. Popul. Biol. 73:11–23.                      Rivero A., R. V. Alatalo, J. S. Kotiaho, J. Mappes, and S. Parri. 2000. Acoustic
Healey, M., T. Uller, and M. Olsson. 2007. Seeing red: morph-specific con-            signaling in a wolf spider: can signal characteristics predict male quality.
     test success, and survival rates, in a colour-polymorphic agamid lizard.         Anim. Behav. 60:187–194.
     Anim. Behav. 74:337–341.                                                    Rohwer, S. 1975. The social significance of avian winter plumage variability.
Hearne, C. M., S. Ghosh, and J. A. Todd. 1992. Microsatellites for linkage            Evolution 29:593–610.
     analysis of genetic-traits. Trends Genet. 8:288–294.                        ———. 1976. The evolution of reliable and unreliable badges of fighting
Johnstone, R. A. 1995. Sexual selection, honest advertisement and the handi-          ability. Am. Zool. 22:531–546.
     cap principle: reviewing the evidence. Biol. Rev. 70:1–65.                  Schwartz, T. S., D. A. Warner, L. B. Beheregaray, and M. Olsson. 2007. Mi-
Kokko, H. 2007. Modelling for field biologists. Cambridge Univ. Press,                crosatellite loci for Australian agamid lizards. Mol. Ecol. Notes. 7:528–
     Cambridge pp. 63–88.                                                             531.
Kokko H., R. Brooks, J. M. McNamara, and A. I. Houston. 2002. The sexual         Searcy, W. A., and S. Nowicki. 2005. The evolution of animal signaling
     selection continuum. Proc. R. Soc. Lond. B. 269:1331–1340.                       systems. Monographs in behavior and ecology. Princeton Univ. Press,
Marshall, T. C., J. Slate, L. E. B. Kruuk, and J. M. Pemberton. 1998. Sta-            Princeton, NJ, pp. 1–223.
     tistical confidence for likelihood-based paternity inference in natural     Sinervo, B., and C. M. Lively. 1996. The rock-paper-scissors game
     populations. Mol. Ecol. 7:639–655.                                               and the evolution of alternative male strategies. Nature 380:241–
Maynard Smith, J. 1976. Sexual selection and the handicap principle. J. Theor.        243.
     Biol. 57:239–242.                                                           Zahavi, A. 1975. Mate selection: a selection for a handicap. J. Theor. Biol.
Olsson, M. 1994. Nuptial coloration in the sand lizard (Lacerta agilis): an           53:205–214.
     intrasexually selected cue to fighting ability. Anim. Behav. 48:607–613.    ———. 1977. The cost of honesty (Further remarks on the handicap principle).
Olsson, M., M. Healey, and L. B. Astheimer. 2007a. Afternoon T: testosterone          J. Theor. Biol. 67:603–605.
     level is higher in red than yellow male polychromatic lizards. Physiol.     Zamudio, K. R., and B. Sinervo. 2000. Polygyny, mate guarding, and posthu-
     Behav. 91:531–534.                                                               mous fertilization as alternative male mating strategies. Proc. Natl. Acad.
Olsson, M., M. Healey, E. Wapstra, T. Schwartz, N. LeBas, and T. Uller. 2007b.        Sci. USA 97:14427–14432.
     Mating system variation and morph fluctuations in a polymorphic lizard.
     Mol. Ecol. 16:5307–5315.                                                                                              Associate Editor: H. Kokko

                                                                                                      EVOLUTION MARCH 2009                    701

Shared By: