Ecology and Sexual Selection in the Barking Gecko (Ptenopus
garrulus)garrulus)Toby J Hibbitts
University of the Witwatersrand, School of Animal, Plant, and E i t l S i P i t B 3 Wit 2050 S th Af iEnvironmental Sciences, Private Bag 3, Wits 2050, South Africa
Sexual selectionSexual selection
• 5 mechanisms5 mechanisms• I focus on endurance rivalry, contest
competition and mate choicecompetition, and mate choice
Endurance rivalryEndurance rivalry
The most persistent males are the mostThe most persistent males are the most successful breeders
Favours: males in the best bodycondition, or males which loose condition slowly.
ContestsContests
Fights between males for the best territories or access to femalesg
Mate choiceMate choice
Mates chosen based on certain characteristicsMates chosen based on certain characteristics, i.e. display behaviour, colour, tail length, call.
SignalsSignals
Animals use signals to avoid conflicts and attract matesAnimals use signals to avoid conflicts and attract mates
Signals used in contests are referred to as armaments d h d i h iand those used in mate choice as ornaments
Signals may also be broadcast through different sensory g y g ymodalities (Auditory, olfactory, vision).
Role of EcologyRole of Ecology
• Without ecological information on aWithout ecological information on a species more in depth studies cannot be undertakenundertaken.
K l d f b di ti• Knowledge of breeding season, mating system and habitat preferences are
f t di i l l tinecessary for studies in sexual selection.
Study organismStudy organism
• Why study the barking gecko?Why study the barking gecko?– Multiple multimodal signals
Unique in the degree of calling for a reptile– Unique in the degree of calling for a reptile– Common
Relatively unstudied– Relatively unstudied
Calling geckoCalling gecko
What is known?-Live in self constructed burrows (Haacke 1975)-Call at dusk (Haacke 1969)Call at dusk (Haacke 1969)-Will defend against calling intruder (P. kochi; Polakow 1997)-Lay one egg clutches (Pianka and Huey 1978)E t tl t it b l (Pi k d H 1978)-Eat mostly termites by volume (Pianka and Huey 1978)
P garrulus distributionP. garrulus distribution
Molopo Nat. Res.
Major QuestionsMajor Questions
• Report the ecology of the common barkingReport the ecology of the common barking gecko (Ptenopus garrulus)di t d ti f i d l idiet, reproductive season, foraging mode, sexual size dimorphism
• What is the mating system?• What is the mating system?• Does olfaction plays a role in male refuge
l ti ?selection?
Objectives contObjectives cont.
• How do males respond to playbacks of aHow do males respond to playbacks of a known frequency and how do the throat patch and call relate to responses?patch and call relate to responses?
• Are males that are active the longest have the highest reproductive successthe highest reproductive success (Endurance rivalry)?Wh t l h t i ti l t ith• What male characteristics correlate with reproductive success?
EcologyEcology
• Sexual size dimorphismSexual size dimorphism• Diet
R d ti• Reproductive season• Foraging mode
Hibbitts et al. 2005
MorphologyMorphology
• I measured: svl tail head width headI measured: svl, tail, head width, head depth, and trunk length. All except svl were size-adjustedwere size adjusted.
• I conducted t-tests on all morphological measurementsmeasurements.
• The only significant result was size-dj t d h d idthadjusted head width.
Head width t = 1.98 p = 0.002 (n = 66f and 82m)
13.00
14.00
10.00
11.00
12.00
d W
idth
7.00
8.00
9.00Hea
6.0030.00 35.00 40.00 45.00 50.00 55.00 60.00
SVL
Dietary raw dataDietary raw data
N = 640 stomachs, 525 (82%) had foodN 640 stomachs, 525 (82%) had food items, 115 (18%) empty
3086 prey items identified and 17 gecko shed skins.shed skins.
MethodsMethods
Prolate spheroid for testes and prey volume:Prolate spheroid for testes and prey volume: V=4/3π(length/2)(width/2)
Niche Breadth using Simpson’s diversityNiche Breadth using Simpson s diversity measure:
P garrulus diet by prey numberP. garrulus diet by prey number
1578 Isoptera (51%)1578 Isoptera (51%)1100 Formicidae (36%)191 C l t (6%)191 Coleoptera (6%)All other groups (7%)
Niche Breadth 2 5Niche Breadth 2.5
P garrulus diet by volumeP. garrulus diet by volume
5883 29 Isoptera (60%)5883.29 Isoptera (60%)875.85 Formicidae (8.93%)788 37 C l t (8 04%)788.37 Coleoptera (8.04%)All other prey (23.03%)
Niche Breadth 2 60Niche Breadth 2.60
Male vs Female dietMale vs. Female diet
No difference between sexes in preyNo difference between sexes in prey number eaten (χ2
3 = 3.62; P < 0.5).No correlation between prey size and HWNo correlation between prey size and HW
while accounting for SVL (r = 0.099, P = 0 44)0.44).
No difference in the size of prey eaten by l d f lmales and females
Testes volume and Egg lengthnot significantly correlated with SVL
16
18
40
10
12
14
16
mm
3 )
33
40
9
10
12
14
16
th (m
m)
2
4
6
8
Volu
me
(m
Inactive48
42
1258
15
0
2
4
6
8
Egg
Leng
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
-2
0
00 1 2 3 4 5 6 7 8 9 10 11 12 13
MonthJan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Foraging modeForaging mode
• Conducted 11 ten minute observations ofConducted 11 ten minute observations of adult barking geckos (6 male, 5 female), noted all movements and length of eachnoted all movements and length of each move.
• MPM = 0 4 ± 1 9 (0 2 1)• MPM = 0.4 ± 1.9 (0 - 2.1) • PTM = 4.6 ± 1.9 s (0 - 16.7)
(criterion of PTM < 10 for ambush foragers)
ConclusionsConclusions
• Male barking geckos have wider heads thanMale barking geckos have wider heads than females and the difference was not due to niche divergence
• Barking geckos eat mainly termites and ants by number but volumetrically termites are most important
• Peak reproduction occurs in September and O b i b hOctober in both sexes
• Barking geckos are classic ambush foragers
Mating systemMating system
• What type of mating system?What type of mating system?• Are males active before females
(Protandry)?(Protandry)?
MethodsMethods• Located gecko burrow in a 1.1 ha study area and g y
marked each location and recorded the sex of each occupant. Burrows were monitored daily for a 67 day period from Sept to Nov 2004period from Sept to Nov 2004.
• The following data was collected on all geckos captured: SVL, mass, sex, and throat patch size (if male). An attempt was made to record all calling males in the study area.
Home range sizeHome range size
Home range sizeHome range size
• Average HRS of 51 malesAverage HRS of 51 males– 10.4 ± 1.5 m2 (range = 3.1 to 53.0)
mean number overlapped = 0 12 ± 0 04• mean number overlapped = 0.12 ± 0.04 • mean percent overlap = 3.8 ± 1.7• Correlations
– HRS and SVL (r = 0.31, P = 0.02, n = 51)( )– HRS and freq (r = -0.24, P = 0.20, n = 31)– HRS and TPS (r = -0.13, P = 0.72, n = 49)
ActivityActivity
60
70MalesFemales2 per. Mov. Avg. (Males)2 M A (F l )
30
40
50
er o
f gec
kos
2 per. Mov. Avg. (Females)
10
20
30
Num
be
08/31 9/10 9/20 9/30 10/10 10/20 10/30 11/9 11/19 11/29
Date
Activity contActivity cont
90
60
70
80
90os
30
40
50
60
Num
ber
of G
ecko
MalesFemales
0
10
20
1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 651 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65
Days
ConclusionsConclusions
• Mating system is consistent with PolygynyMating system is consistent with Polygyny– Males live in largely exclusive home ranges– Males aggressively defend their territory gg y y– Males potentially breed with more than one female in
a season• Males are active before females (Protandry)
– Establish territories before females emergeM l d t b d b f f l– Males are ready to breed before female emergence
Sexual selectionSexual selection
• OlfactionOlfaction
Pl b k t i l• Playback trials
• Reproductive success
OlfactionOlfaction
• Do males avoid refuges scented by otherDo males avoid refuges scented by other males?
– A male was given the choice of a scented and unscented tile refugeunscented tile refuge
– During 20 experiment 10 chose scented and 10 chose unscented10 chose unscented
Hibbitts and Whiting 2005
ConclusionsConclusions
Scent signals not present in barking geckosScent signals not present in barking geckos because:
Never observed to tongue flick– Never observed to tongue flick– Calling likely signals presence for efficiently
Loose sandy substrate– Loose sandy substrate
Playback trialsMultiple signals in barking geckos
Playback trialsMultiple signals in barking geckos
Throat patch (visual) functions over short distances and in adequate light conditions
Call (auditory) functions over long distances
Contest - questionsContest questions• How do signals function in male-male interactions?g• Is calling costly?
Endurance rivalry - question• Do males which are more active have better
reproductive success
Endurance rivalry - question
reproductive success
Mate choice - question• What male characteristics best predict breeding success
ate c o ce quest o
Signal correlationsSignal correlationsCall frequency (n = 74) correlated y = -62.829x + 7208.8
r 2 = 0.4952000
5500
Hz)
with body size
(r2 = 0.50, F1,72 = 69.8, P < 0.001)
r 0.4952
3500
4000
4500
5000
inan
t fre
quen
cy (H
3000
3500
35 40 45 50 55
SVL (mm)
Dom
iThroat patch size (n = 149) not correlated with body size
(r2 = 0.06, F1,120 = 7.13, P < 0.009)
Playback trialsPlayback trials
• We conducted playback trials to determineWe conducted playback trials to determine what gecko characteristics predicted behavioural responsesbehavioural responses
Playback trialsPlayback trials
• videovideo
Playback trials (n = 58)Playback trials (n 58)
• 3 main behaviours3 main behaviours– Retreat into burrow– Call back
Charge– Charge
• Multiple logistic regressions (SVL, condition, frequency, relative TPS)frequency, relative TPS)– Retreat – SVL (small) only significant character– Call back – None significant, condition close (0.06)– Charge – frequency (low) relative TPS (small) only significantCharge frequency (low), relative TPS (small) only significant
characters
Playback trialsPlayback trials
• Aggressiveness RankAggressiveness Rank– Ranked observed responses to playback trials
from 0-5from 0 5.– Did multiple regression with rank treated as
continuous and SVL, condition, frequency,continuous and SVL, condition, frequency, and relative TPS as independent variables
– The model found that SVL (P = 0.01) and ( )relative TPS (P = 0.04) were best indicators of aggressiveness
Cost of callingCost of callingSVL vs weight loss 0
0.005
0.01
30 35 40 45 50 55e pe
r day
y = -0.0008x + 0.0303R2 = 0.3274
-0.025
-0.02
-0.015
-0.01
-0.00530 35 40 45 50 55
avg
wei
ght c
hang
e(g
m)
SVL (mm)
No significant difference between geckos grouped asNo significant difference between geckos grouped as frequent and infrequent callers (P > 0.05)
Six males recaptured after greater than 50 days averagedweight loss of 16% (range 9 – 24%) of their body weight
Reproductive successReproductive success
Photo by David Laurencio
Endurance rivalryEndurance rivalry
I tested whether male activity (males callI tested whether male activity (males call when active in the breeding season) was higher in males who bred successfullyhigher in males who bred successfully.
No difference was found between the groups
t2 29 = 1.7, P = 0.342,29 ,
Successful male characteristicsSuccessful male characteristics• Conducted multiple logistic regression with Bred as p g g
categorical dependent and SVL, HW, HD, TPS, and Frequency as independent variables. The morphological variables were corrected for by size by using the y y gresiduals from regressions against SVL.
• SVL was the only significant indicator of breeding success (P = 0.01) with larger g ( ) gmales more likely to be successful breeders.
ConclusionsConclusions
• Call frequency is strongly correlated withCall frequency is strongly correlated with SVL and therefore likely signals body size.
• TPS was not correlated with morphological• TPS was not correlated with morphological character, is a signal of aggressivenessC ll f i l d TPS i ll• Call frequency is lower and TPS is smaller in geckos which charge the playback and SVL i ll i l th t t t fSVL is smaller in males that retreat from playback calls
ConclusionsConclusions
• Aggressive males are large and haveAggressive males are large and have small relative throat patches
• Activity costly!• Activity costly! – not correlated to calling
f f• Activity is not a major factor influencing breeding success
• Successfully breeding males were larger
AcknowledgementsAcknowledgements• We would like to thank the following people for help in the field and
M l N Rsupport at Molopo Nature Reserve:• Rachel Gallagher• Steven Gore (Formerly Park Warden at MNR)• Gerald and Inna Botha• David and Laura Laurencio• Kate Hodges• Ingrid Stirneman• We also thank Devi Stuart-Fox for statistical advice and the rest of
the Whiting Lizard Lab for support• This project was funded by an NRF grant to Martin J. Whiting• Clearance was obtained for the study through the Wits University
Animal Ethics Committee (2002/77/2a)