Sexual selection in mosquito swarms: may the best man lose

Available online at ww

wsciencedirectcom

ANIMAL BEHAVIOUR 2008 76 105e112doi101016janbehav200801014

Sexual selection in mosquito swarms may the best man lose

KIJA R NGrsquoHABIdagger BERNADETTE J HUHOdagger GAMBA NKWENGULILAdagger GERRY F KILLEENDagger

BART G J KNOLSsect amp HEATHER M FERGUSONsect

Public Health Entomology Unit Ifakara Health Research and Development Centre

yDepartments of Zoology and Wildlife Conservation University of Dar es Salaam

zSchool of Biological and Biomedical Sciences Durham University

xLaboratory of Entomology Wageningen University and Research Centre

Division of Immunity amp Infection and Division of Environmental amp Evolutionary Biology

University of Glasgow

(Received 26 July 2007 initial acceptance 28 September 2007

final acceptance 29 January 2008 published online 5 May 2008 MS number 9471R)

One of the greatest paradoxes in evolutionary biology is the continued maintenance of genetic variationfor phenotypic traits that appear to confer strong fitness advantages Of these traits body size is perhapsthe one that has been most consistently linked to increased longevity and reproductive success in malesWe investigated two hypotheses for how events occurring during mating in the mosquito Anophelesgambiae could impede directional selection on male body size (1) whether male copulation probabilityis random with respect to body size in aerial swarms and (2) whether large males are less competitiveduring mating than smaller shorter-lived rivals By manipulation of larval nutritional conditions (lowintermediate and high food allocation) we generated cohorts of male A gambiae mosquitoes that differedin adult body size and energy reserves (body size and energy reserves being positively correlated withlarval nutrition) When competing against one another in aerial swarms males from the intermediatefood treatment were six and two times more successful at acquiring mates than those from the highand low food treatments respectively The median survival of males from this most sexually competitivegroup was approximately 13 lower than that of the larger males with high larval nutrition Weconclude that phenotypic determinants of long-term survival and mating success may not be correlatedin this system and thus that stabilizing selection as well as environmental condition-dependent expres-sion of traits could account for the maintenance of variation in male body size in this species and inother swarming insects

2008 The Association for the Study of Animal Behaviour Published by Elsevier Ltd All rights reserved

Keywords Anopheles gambiae fitness genetically modified

CorrespoCentre (I(email kUniversinia B GUniversiThe Nethopment Ces SalaamInfectionUniversi

0003e3

insect mating competitiveness mosquito sexual selection

ndence K Ngrsquohabi Ifakara Health Research and DevelopmentHRDC) PO Box 53 Off Mlabani Passage Ifakara Tanzaniaijaihrdcortz ndyaiyahoocom) G Nkwengulila is at the

ty of Dar es Salaam PO Box 35064 Dar es Salaam Tanza- J Knols is at the Laboratory of Entomology Wageningen

ty and Research Centre PO Box 8031 6700 EH Wageningenerlands G F Killeen is at Ifakara Health Research and Devel-entre Coordination Office PO Box 78373 Kiko Avenue Dar Tanzania H M Ferguson is at the Division of Immunity amp

and Division of Environmental amp Evolutionary Biologyty of Glasgow Glasgow G12 8TA UK

10547208$34000 2008 The Association for the Stu

In animal taxa ranging from insects to large mammals themale phenotypic trait that has been most consistentlyassociated with enhanced longevity andor reproductivesuccess is body size (Andersson 1982 1994 Thornhill ampAlcock 1983 Blanckenhorn 2000 Jennions et al 2001Blanckenhorn et al 2004 Maan et al 2004 Malo et al2005) Additional phenotypic traits that are correlatedwith body size (eg energy reserves) have also been linkedto enhanced longevity and reproduction (Yuval et al1994 Sparkes et al 1996) On the basis of this near ubiqui-tous demonstration of the fitness advantages of being largeone would predict directional selection on male body size toexist in most animal species The continued existence of

dy of Animal Behaviour Published by Elsevier Ltd All rights reserved

ANIMAL BEHAVIOUR 76 1106

variation in male body size in nature suggests that this doesnot occur Three general classes of hypotheses have beenput forward to explain this apparent paradox the first beingthat the efficiency of selection is impeded by plasticity inthe expression of phenotypic traits caused by environmen-tal variation occurring during development (conditiondependence) A well-known example of this phenomenonis the lsquosilver spoonrsquo hypothesis which proposes that nutri-tional resources provided during early life history are anequal if not greater determinant of adult fitness thangenotype (Metcalfe amp Monaghan 2001) A second but notnecessarily conflicting paradigm is that traits apparentlyunder strong sexual selection such as body size may be neg-atively correlated with other aspects of male fitness (egpredation risk andor energy requirements) which incombination generate stabilizing selection (Blanckenhorn2000) These costs have been demonstrated in some species(Blanckenhorn 2000) although a thorough cross-taxameta-analysis suggested that male traits under sexual selec-tion such as size are often positively correlated with longev-ity (Jennions et al 2001)

An additional factor that may influence the efficiency ofsexual selection on male traits is the predictability ofencounter rate between males and females Where theprobability of a male coming into contact with receptivefemales is determined primarily by randomly occurringenvironmental variation male mating success may beultimately determined by chance events rather than partic-ular phenotypic characteristics One system where this mayoccur is aerial mating swarms of insects Unlike communalcourtship grounds (leks) where groups of males congregateon a fixed habitat feature or area and compete for a territorywithin it (Kruijt amp Hogan 1967) an aerial swarm is a three-dimensional arena whose physical structure is determinedby the flight activity and abundance of males themselves(Downes 1969 Yuval et al 1993 Yuval 2006) A wide varietyof insects including dipterans trichopterans lepidop-terans hymenopterans hemipterans ephemeropteransand odonates mate in aerial swarms which are typicallymaintained in a fixed position above a marker by the con-stant lsquofigure-of-eightrsquo flight of males within them (Downes1969) Females may be attracted to a swarm from a relativelylong distance (Clements 1992) Some of the most detailedobservations of swarming behaviour have been made onmosquitoes in which males have been observed to dartout and grab approaching females before they reach thelsquobodyrsquo of the swarm and quickly mate with them (Reisenet al 1977 Charlwood et al 2002a 2003) Consequentlymales flying on the edge of the swarm when a female arrivesmay have the highest chance of copulation As malesrepeatedly fly through the swarm it may not be possiblefor them to predict their location (centre or edge) whena female arrives andor secure a fixed lsquooptimalrsquo positionon the periphery This being the case male mating successwith respect to phenotypic traits such as body size mayindeed be random a phenomenon that has been proposedfor some swarming insects (Charlwood et al 2002b) butdismissed in other studies where both large (Yuval et al1994) and small size (Marshall 1988 Fyodorova amp Azovsky2003 LeBas et al 2004) have been associated with highersuccess

To assess the impact of phenotypic traits on matingsuccess under the competitive conditions of an aerialswarm we manipulated the larval nutrition of male Anoph-eles gambiae sensu stricto mosquitoes Our principal aimwas to test whether male mating success was enhancedby commonly proposed indexes of male fitness such asbody size and energetic reserves or random as has beenspeculated for aerial swarms We also investigated the asso-ciation between traits that may be under sexual selectionand long-term survival to assess the possibility of trade-offs between different components of male fitness thatcould generate stabilizing rather than directional selectionon body size

Elucidating the strength and direction of sexual selectioncan have strong applied relevance to the implementation ofdisease control programmes that are based on the massrelease of sterilized or genetically modified male insectvectors that can drive infection-blocking genes into naturalpopulations (Ferguson et al 2005) The creation of geneti-cally modified mosquitoes that are refractory to infectionby malaria parasites (Capurro et al 2000 Ito et al 2002)and dengue virus (Franz et al 2006) has fuelled optimismthat these diseases could be controlled by releasing a suffi-ciently large number of transgenic males into the wildthat could drive these genes into natural populations TheA gambiae mosquitoes studied here are the principal vectorsof malaria in sub-Saharan Africa (WHO 1993) and a primarycandidate for disease control programmes based on themass release of transgenic males (Ito et al 2002 Tabachnick2003 Moreira et al 2004 Andreasen amp Curtis 2005) Crucialto the success of this approach is the ability of transgenicmales to compete successfully for females in the presenceof their wild counterparts (Scott et al 2002 Fergusonet al 2005) Information on the existence and phenotypicdeterminants of heterogeneity in male A gambiae matingsuccess could thus provide useful guidance for the designof rearing protocols for the mass production of sexuallycompetitive males whose release could reduce malariatransmission

METHODS

Rearing

We studied A gambiae sensu stricto mosquitoes froma population at the Ifakara Health Research and Develop-ment (IHRDC) Tanzania This population was establishedfrom a wild population near Njage village Tanzania in1996 We randomly assigned first-instar larvae to one ofthree food quantity treatments 01 mg 02 mg or 04 mgof dried fish food (Tetramin Melle Germany) per larva perday (Lyimo et al 1992) Larvae of the same food treatmentwere housed together in groups of 100 in standard rearingtrays (37 14 cm and 13 cm high) Each tray was filledwith 1 litre of river water and supplied with a total dailymass of fish food sufficient to feed all 100 larvae (a total of10 mg 20 mg and 40 mg of Tetramin added to low mediumand high food treatment trays respectively) We used riverwater as it is most representative of the natural aquatic sub-strate in which larvae develop in the wild To synchronize

NGrsquoHABI ET AL SEXUAL SELECTION IN MOSQUITOES 107

pupation we established cohorts of larvae in the lowest foodtreatment 4 days before the other two food treatments (thistime gap was chosen on the basis of pilot studies of pupationrates) Once established we inspected larval trays twicea day for the presence of pupae When detected pupaewere collected and held individually in vials to allow foremergence Batches of males from all three food treatmentsthat emerged on the same day competed against one an-other in mating trials for females from the medium foodtreatment We used females from only one larval back-ground to minimize variation in female body size andenergy reserves and thus standardize the lsquoresourcersquo forwhich males were competing across trials

Marking

From emergence males were pooled according to larvalfood treatment and held in separate cages On the secondday after emergence cohorts of adult males from two ofthe three food treatments were marked with green or pinkfluorescent powder respectively The third group was leftunmarked We alternated marking treatments betweenfood treatments across trials to ensure no systematic biasin performance from dusting Pilot studies and previousinvestigations (Ngrsquohabi et al 2005) indicated that malesfrom different backgrounds could be reliably identifiedin this way with no evidence that dust was transferredbetween males during the swarming process In swarmsmales make contact with females during copulation butnot with one another during flying As copulating pairswere removed from the mating arena before separating(see below) marking dust could not be transferred fromone male to another through a female

Mating Experiment

On the third day after emergence 30 males (10 males fromeach treatment) were put together in one cage (15 15 cmand 10 cm high) The cage was exposed to natural lighta few hours before dusk We placed a black plastic disk onthe bottom of each cage to act as a swarm marker (Ngrsquohabiet al 2005) In nature A gambiae males form mating swarmsapproximately 15e20 min before sunset (Charlwood ampJones 1980 Charlwood et al 2002a 2003) a behaviourthat is also observed in our laboratory population (Ngrsquohabiet al 2005) We began monitoring activity inside each exper-imental cage approximately 10 min before sunset

Once swarming was underway 10 females from themedium food treatment were simultaneously released intothe cage (making a 31 male to female ratio) We observedmating activity within cages under a low-intensity redlight bulb Pairs observed to form copula were immedi-ately aspirated out of the cage and put together intoa holding cup As a consequence of this design femalescould not be mated more than once in a swarming eventAs the frequency of polyandry in Anopheles populations isknown to be very low (lt3 Tripet et al 2003) thisrestriction on multiple mating in females is not inconsis-tent with nature In addition a male and female in copula

normally fly away from the swarm reducing the likeli-hood of returning for multiple mating

On each evening of the experiment we finished obser-vations of mating after all males had stopped swarming(usually 40e45 min after first initiation) The followingmorning a cotton wad soaked with chloroform was puton top of each holding cup and all mosquitoes capturedin copula were anaesthetized and killed Males were thenobserved under a fluorescent lamp to identify from whichlarval food treatment they originated We measured thewing lengths of both male and female partners undera dissecting microscope Males that did not mate werealso killed and their wing lengths measured for comparisonwith those caught in copula We conducted eight replicatesof this competition experiment

Quantification of Energy Reserves

A subsample of newly emerged males from each larval foodtreatment were killed and transferred individually into glasstest-tubes for the quantification of lipids sugars and glyco-gen Once in the tubes the mosquitoes were crushed witha glass rod We added 100 ml of 2 sodium sulphate (whichadsorbs glycogen) and 600 ml of a 12 chloroformemethanolmixture (which dissolves lipids and sugars respectively) toeach tube then incubated them for 24 h at room tempera-ture We analysed 93 males 28 from low 26 from mediumand 39 from high food treatments For each batch of malesanalysed we prepared one blank by adding the samechemicals to a tube that contained no mosquitos Lipidsglycogen and sugars of each male were then quantified byusing a colorimetric technique adapted for mosquito analysis(Van Handel 1988)

Longevity Experiment

In a separate experiment males emerging from eachlarval food treatment were denied access to females butheld in groups of 32e47 in standard holding cages(15 15 cm and 10 cm high) to monitor their longevityAt least two cage replicates of each larval food treatmentwere monitored While in cages adult males receiveda 10 glucose solution ad libitum We observed eachcage daily and removed and counted all dead males foundwithin it

Statistical Analysis

The main aim of our statistical analysis was to test fordifferences in male mating competitiveness within a swarmas a function of the larval food-set adult phenotype Weconsidered whether male larval food treatment wasa predictor of (1) the first male to mate and (2) the totalnumber of males that obtained copulations across all trialsIn both cases chi-square tests were used to evaluate therelation between the dependent variable (proportion ofmatings) and larval food treatment General linear models(GLM) were used to test whether larval food treatmentinfluenced male body size (as indexed by wing size) andor

15

20

s (micro

g)

LipidsGlycogens

Sugars

ANIMAL BEHAVIOUR 76 1108

the abundance of energy reserves (sugars glycogen andlipids) on emergence Finally we used KaplaneMeiersurvival analysis to test whether the survival of adult malesvaried in response to larval food treatment The SPSSstatistical package version 115 for Windows (SPSS IncChicago IL USA) was used for all analyses Unlessotherwise stated means are given 1 SE

0

5

10

Low Medium High

Larval food treatment

Ten

eral

res

erve

Figure 2 The mean amounts of lipid glycogen and sugars in newly

emerged A gambiae males reared in low medium and high larvalfood conditions Vertical lines indicate 1 SE

RESULTS

By varying larval food we succeeded in generating threedistinct adult male phenotypes which differed with respectto each of the male traits we measured First body sizevaried substantially between groups (GLM F2234 frac14 94 128P lt 001 Fig 1) with males from the high food treatmentbeing 18 and 9 bigger than those in the low andmedium food treatments respectively (low food 249 0015 mm medium food 269 0014 mm high food294 001 mm) The average size of females used in theseexperiments was 283 002 mm

The abundance of teneral energy reserves in adult malesalso varied in response to larval food treatment As with bodysize males emerging from the high food treatment had themost abundant lipids glycogen and sugars (GLM lipidsF226frac14 1576 P lt 0001 sugars F224 frac14 2205 P lt 0001glycogen F237 frac14 2705 Plt 0001 Fig 2) Furthermoremale survival was similarly influenced by the abundance offood received during development (log rank2 frac14 13669df frac14 2 N frac14 294 P lt 001 Fig 3) The median survival ofmales from the low food regime (12 days) was 8 and 11days lower than that of males from the medium (20 days)and high (23 days) larval food treatment respectively Thusthe larval food treatment we imposed yielded three distinctphenotypic groups of males that varied substantially in con-dition as indexed by body size teneral reserves and survival

In the mating experiment males originating from themedium food treatment with intermediate body size energyreserves and survival were the first to obtain a mate in sevenof eight trials (chi-square test c2

2 frac14 1850 P lt 001) Consid-ering all mating events and not just the first to occur in an

2

22

24

26

28

3

32

Low Medium High

Larval food treatment

Mea

n w

ing

len

gth

(m

m)

MatedDid not Mate

Figure 1 The mean wing length (an indicator of body size) of Agambiae males that mated and those that did not from each larval

food treatment Vertical lines represent 1 SE

evening we found that copulations were not distributedrandomly among males from different larval treatmentgroups (c2

2 frac14 1359 P lt 001 Fig 4) Males from the inter-mediate larval food treatment dominated mating activitybeing six times more likely to acquire a female than malesfrom the higher larval food regime (odds ratio 95 CI frac14633 205e1952 P lt 001) Males from the intermediategroup tended to accrue more matings than those in thelow larval food treatment although this difference wasnot statistically significant (odds ratio 95 CI frac14 209093e471 P frac14 008)

Within each of the three larval food treatments males thatobtained mates were larger than those that did not (GLMF1234 frac14 2949 P lt 001 Fig 1) This observation contrastswith the between-treatment observation that males from

0

05

1

0 5 10 15 20 25 30

Time (days)

Prop

orti

on s

urv

ivin

g

Figure 3 The survival of adult male A gambiae maintained on high

(N frac14 58) medium (N frac14 44) and low (N frac14 192) food provision dur-

ing larval development B Males from the low food treatment males from the medium food treatment 6 males from the

high food treatment

0

15

30

45

60

Low Medium High

Larval food treatment

Perc

enta

ge o

f to

tal

mat

ings

Figure 4 Percentage of total matings (N frac14 35) going to adult maleA gambiae from low medium and high larval food treatments

pooled over eight experimental replicates Vertical lines indicate

1 SE

NGrsquoHABI ET AL SEXUAL SELECTION IN MOSQUITOES 109

the group having intermediate body size were more success-ful than males from the higher and lower larval food groupsIn our statistical analysis of the total number of matingsinclusion of male body size as an additional explanatoryfactor did not diminish the significance of larval treatment(chi-square test c1

2 frac14 4064 P lt 001) indicating that adultbody size and larval nutrition were important and possiblyindependent predictors of mating competitiveness

DISCUSSION

The mating success of male A gambiae mosquitoes withinswarms in our study was not random but depended on nutri-ent uptake during larval development and its impact on adultbody size Contrary to expectation we observed that the lsquobestmalersquo with respect to frequently observed predictors of maleinsect fitness (longevity body size and energy reserves) wasnot the most competitive in acquiring females within a mat-ing swarm Instead we observed that males most likely to winfemales were those of intermediate body size and energyreserves Despite being an average of 10 larger and havingtwo to four times more energy reserves (lipids glycogensand sugars) males from the high larval food backgroundwere only one-sixth as likely to obtain a copulation in oneswarming event as males from the intermediate treatmentThere was also no evidence of a correlation between malemating success within a swarm and long-term survival Themedian survival of males from high larval food conditionswas 13 greater than that of males from intermediate foodbackgrounds but their probability of obtaining a mate perswarming event was much lower (84 reduction) Wecaution that these results were obtained under artificialconditions in the laboratory in which an equal ratio of malesfrom all three distinct phenotype categories competedagainst one another and may not be representative of thefull range of competitive pressures experienced by A gambiaeswarming in the wild However as variation in male body sizeand energy reserves generated in this experiment falls withinthe range observed in wild populations from the surrounding

Kilombero valley (19e31 mm Huho et al 2007) we arereassured that our results are representative of natural pheno-types and not simply an artefact of aberrant extremesproduced in the laboratory The females we used were alsowithin the range observed in the local population (234e362mm Lyimo amp Takken 1993)

Although this is not the first report of males of interme-diate size having high mating success (eg Delisle amp Hardy1997) it does stand in contrast to most of the evidencefrom other arthropods that indicates that male mating suc-cess increases with body size (Partridge amp Farquhar 1983Thornhill amp Alcock 1983 Alcock 1996 Blanckenhorn2000 2005 Crean et al 2000) What could thus accountfor the strong mating advantage of males of intermediatesize in this study One possibility is that smaller males aremore agile in flight and can make contact with femalesmore quickly a phenomenon that has been observed inother swarming insects (Neems et al 1998 Cromptonet al 2003) The potential fitness advantage that small malesmay have in flight could be eroded by their lower energy re-serves which may compromise their ability to sustainswarming (eg Yuval et al 1994) and thus render maleswith intermediate body size and reserves the most competi-tive Another explanation for the differential mating successof males in this study could be size-assortative mating Posthoc observation confirmed that males from the intermediatelarval food group were closer in body size to available femalesthan males from other treatment groups Size-assortativemating has been found in several arthropod species (Crespi1989) and anecdotal evidence suggests it may occur inAnophelines (Gomulski 1985) This phenomenon could ariseas a by-product of size-dependent competition betweenmales for large females andor higher copulation and insem-ination success in maleefemale pairs with similarly sizedgenitalia (Arnqvist 1997) Our study was not designedto test for size-assortative mating and indeed we tried tominimize size variation in the females offered to males tostandardize their perceived lsquoqualityrsquo across trials Althoughwe are currently logistically limited in our ability to rearthe large numbers of size-varying males and females thatwould be required to confirm this phenomenon we recom-mend it as a valuable line of inquiry for future studies ofmating biology in this species

A final hypothesis for the success of intermediate-sized andenergetically provisioned males is that they may be most inconcordance with natural patterns of sexual size dimorphismin this species Field data indicate that female Anopheles aregenerally larger than males although there is considerableoverlap in body size between the sexes (Charlwood 2002Lehmann et al 2006) Although the body sizes of all malesused in this study were within the natural range of thisspecies not all were smaller than the females on offer Inparticular males from the high larval food group wereslightly larger than females If females follow a rule wherethey prefer large males as long as they are not larger thanthemselves intermediate-sized males would be favouredover large males as observed here This hypothesis howeverwould require that mating is driven by female choice ratherthan maleemale competition which although possibleseems unlikely in a system such as this where males physi-cally grab females in flight to copulate

ANIMAL BEHAVIOUR 76 1110

We initially set out three hypotheses for why traits such asbody size that apparently confer strong fitness benefits tomales in many animal taxa show substantial phenotypicvariation in nature (1) environmental condition-dependentexpression of traits (2) opposing influences of phenotypictraits on different aspects of total lifetime fitness that lead tostabilizing selection and (3) randomness in the encounterrate between males and receptive females While our studywas not designed to test each of these hypotheses exhaus-tively it provides valuable insights into the likely constraintson male body size evolution in swarming mosquitoes Wecan reject the hypothesis that chance plays a strong role indetermining male mating success in this system one groupof males was consistently more successful at acquiringfemales than others and furthermore success within thesegroups was not random as larger males within groups weremost likely to mate Thus even though it may appear thatmosquitoes adopt similar flight behaviour in swarms andcannot guard particular territories they are not equally likelyto obtain copulations This finding contrasts with Charl-wood et alrsquos (2002b) conclusion that male size does not influ-ence mating success in natural A gambiae swarms Theseauthors based their conclusion on the fact that the averagebody size of males in copula was not statistically differentfrom that of other males in the swarm but ambiguityremains as to whether the body size of copulating maleswas less variable than that of all other males If it was lessvariable it would suggest intermediate-sized males mayalso have had an advantage in this study Alternatively thediscrepancy between these studies could arise from variationin additional male traits such as age and previous matingexperience in natural swarms which may make it more diffi-cult to detect clear associations between male phenotype andmating success of the type we found under controlled labora-tory conditions

In addition we have shown strong condition depen-dence in the expression of both adult body size and energyreserves Within our laboratory population we were able togenerate substantial variation in male body size and energyreserves by moderate manipulation of larval nutritionAlthough body size in A gambiae is known to be heritablein nature (Lehmann et al 2006) variation in larval habitatquality may ultimately be a stronger determinant of adultphenotype than genetic background and thus impede thespeed of selection towards an optimum

In addition to these environmental effects our resultsalso suggest the potential for stabilizing selection to act onmale body size We found that longevity increased withbody size but that large males were less successful atobtaining mates than those of intermediate size Althoughthe extended survival of males from the high foodbackground indicates they will have more mating oppor-tunities than those from other groups this advantage maynot sufficiently compensate for their reduced competitive-ness within a swarm to make them the most successful inthe long run Other potential costs of large body size thatwere not measured here include increased vulnerability topredation (Allan et al 1987 McLachlan et al 2003) andhigher energy requirements which increase the risk ofstarvation in resource-poor environments (Blanckenhorn2000) These factors have been proposed to explain why

intermediate body size may be favoured in natural popula-tions of other swarming insects (Neems et al 1998Peckarksy et al 2002) and may operate to maintain geneticvariation for body size in mosquitoes

Finally our findings have relevance for the design andimplementation of malaria control programmes based onthe mass release of laboratory-reared genetically modifiedandor sterilized Anopheles mosquitoes A commonly citedweakness of this approach is that it generally fails toconsider ecological and behavioural aspects of target species(Scott et al 2002) Ignorance of mating biology in particularhas been implicated in the failure of numerous attempts tocontrol disease vectors by the mass release of sterile males intrials during the 1960e1970s and little progress to increaseknowledge has been made since then (Ferguson et al 2005)Our results clearly show that male mating success at least inexperimentally generated swarms is highly dependent onlarval nutrition and associated adult phenotype Failure toconsider the strong influence of phenotype on A gambiaemales in mass-rearing facilities could therefore lead to theproduction of millions of transgenic insects with littlechance of passing on their genes in the wild The rangeand nutritional quality of larval habitats used byA gambiae in the wild varies substantially (Ye-Ebiyo et al2003 Minakawa et al 2005) We emphasize the importanceof identifying the phenotypic and genetic determinants ofmale A gambiae mating success within target malaria-endemic settings before attempting any mass-release trialso that rearing protocols in mass-rearing facilities can beoptimized to produce competitive individuals

Acknowledgments

We express our sincere gratitude to Japhet KihondaNicolas Kasigudi and Hassani Ngonyani from the IHRDCPublic Health Entomology section and thank the IHRDCUniversity of Dar es Salaam and Wageningen Universityfor institutional support This work was supported bya VIDI grant (no 86403004) awarded by the DutchScientific Organization (NWO) to BGJK and a projectgrant from the International Atomic Energy Agency(IAEAURT 13295) and BBSRC David Phillips Fellowshipto HMF

References

Alcock J 1996 The relation between male body size fighting and

mating success in Dawsonrsquos burrowing bee Amergilla dawsoni

(Apidae Apinae Anthopthorini) Journal of Zoology 239 663e674

Allan J D Flecker A S amp McClintock N L 1987 Prey size selec-

tion by carnivorous stoneflies Limnology and Oceanography 32864e872

Andersson M B 1982 Sexual selection natural selection andquality advertisement Biological Journal of the Linnean Society

17 375e393

Andersson M B 1994 Sexual Selection Princeton New Jersey

Princeton University Press

Andreasen M H amp Curtis C F 2005 Optimal life stage for radi-

ation sterilization of Anopheles males and their fitness for release

Medical and Veterinary Entomology 19 238e244

NGrsquoHABI ET AL SEXUAL SELECTION IN MOSQUITOES 111

Arnqvist G 1997 The evolution of animal genitalia distinguishing

between hypotheses by single species studies Biological Journal of

the Linnean Society 60 365e379

Blanckenhorn W U 2000 The evolution of body size what keeps

organisms small Quarterly Review of Biology 75 385e407

Blanckenhorn W U 2005 Behavioral causes and consequences of

sexual size dimorphism Ethology 111 977e1016

Blanckenhorn W U Kraushaar U R S Teuschl Y amp Reim C2004 Sexual selection on morphological and physiological traitsand fluctuating asymmetry in the black scavenger fly Sepsis cynip-

sea Journal of Evolutionary Biology 17 629e641

Capurro M D Coleman J Beerntsen B T Myles K M OlsonK E Rocha E Krettli A U amp James A A 2000 Virus-expressed

recombinant single-chain antibody blocks sporozoite infection of

salivary glands in Plasmodium gallinaceum-infected Aedes aegyptiAmerican Journal of Tropical Medicine and Hygiene 62 427e433

Charlwood J D 2002 May the force be with you measuringmosquito fitness in the field In Ecological Aspects for Application

of Genetically Modified Mosquitoes (Ed by W Takken amp T W

Scott) pp 47e68 Wageningen Kluwer Academic

Charlwood J D amp Jones M D R 1980 Mating in the mosquito

Anopheles gambiae s l II Swarming behaviour Physiological

Entomology 5 315e320

Charlwood J D Pinto J Sousa C A Madsen H Ferreira C ampdo Rosario V E 2002 The swarming and mating behaviour ofAnopheles gambiae ss (Diptera Culicidae) from Sao Tome Island

Journal of Vector Ecology 27 178e183

Charlwood J D Pinto J Sousa C A Ferreira C amp do RosarioV E 2002 Male size does not affect mating success of Anopheles

gambiae in Sao Tome Medical and Veterinary Entomology 16

109e111

Charlwood J D Thompson R amp Madsen H 2003 Observations

on the swarming and mating behaviour of Anopheles funestus fromsouthern Mozambique Malaria Journal 2 2

Clements A 1992 The Biology of Mosquitoes Development Nutri-tion and Reproduction London Chapman amp Hall

Crean C S Dunn D W Day T H amp Gilburn A S 2000 Femalemate choice for large males in several species of seaweed fly

(Diptera Coelopidae) Animal Behaviour 59 121e126

Crespi B 1989 Causes of assortative mating in arthropods Animal

Behaviour 38 980e1000

Crompton B Thomason J C amp McLachlan A 2003 Mating ina viscous universe the race is to the agile not to the swift Proceed-

ings of the Royal Society of London Series B 270 1991e1995

Delisle J amp Hardy M 1997 Male larval nutrition influences the

reproductive success of both sexes of the spruce budworm Chori-

stoneura fumiferana (Lepidoptera Tortricidae) Functional Ecology11 451e463

Downes J A 1969 The swarming and mating flight of the DipteraAnnual Review of Entomology 14 271e297

Ferguson H M John B Ngrsquohabi K amp Knols B G J 2005 Re-

dressing the sex imbalance in knowledge of vector biology Trendsin Ecology amp Evolution 20 202e209

Franz A W E Sanchez-Vargas I Adelman Z N Blair C DBeaty B J James A A amp Olson K E 2006 Engineering RNA

interference-based resistance to dengue virus type 2 in genetically

modified Aedes aegypti Proceedings of the National Academy ofSciences USA 103 4198e4203

Fyodorova M V amp Azovsky A I 2003 Interactions between

swarming Chironomus annularius (Diptera Chironomidae)males role of acoustic behavior Journal of Insect Behavior 16

295e306

Gomulski L M 1985 Larval density adult size and mating compet-

itiveness in the mosquito Anopheles gambiae Transactions of the

Royal Society of Tropical Medicine and Hygiene 79 276e277

Huho B J Ngrsquohabi K R Killeen G F Nkwengulila G KnolsB G J amp Ferguson H M 2007 Nature beats nurture a case

study of the physiological fitness of free-living and laboratory-reared male Anopheles gambiae s l Journal of Experimental Biology

210 2939e2947

Ito J Ghosh A Moreira L A Wimmer E A amp Jacobs-LorenaM 2002 Transgenic Anopheline mosquitoes impaired in trans-

mission of a malaria parasite Nature 417 452e455

Jennions M D Moslashller A P amp Petrie M 2001 Sexually selected

traits and adult survival a meta-analysis Quarterly Review of

Biology 76 3e36

Kruijt J P amp Hogan J A 1967 Social behavior on lek in black

grouse Lyrurus tetrix tetrix (L) Ardea 55 203e207

LeBas N R Hockham L R amp Ritchie M G 2004 Sexual selec-

tion in the gift-giving dance fly Rhamphomyia sulcata favors smallmales carrying small gifts Evolution 58 1763e1772

Lehmann T Dalton R Kim E H Dahl E Diabate A DabireR amp Dujardin J P 2006 Genetic contribution to variation in larval

development time adult size and longevity of starved adults of

Anopheles gambiae Infection Genetics and Evolution 6 410e416

Lyimo E O amp Takken W 1993 Effects of adult body size on fecun-

dity and the pre-gravid rate of Anopheles gambiae females in

Tanzania Medical and Veterinary Entomology 7 328e332

Lyimo E O Takken W amp Koella J C 1992 Effect of rearing

temperature and larval density on larval survival age at pupationand adult size of Anopheles gambiae Entomologia Experimentalis et

Applicata 63 265e271

Maan M E Seehausen O Soderberg L Johnson L Rip-meester E A Mrosso H D Taylor M I van DoorenT J amp van Alphen J J 2004 Intraspecific sexual selection on

a speciation trait male coloration in the Lake Victoria cichlidPundamilia nyererei Proceedings of the Royal Society of London

Series B 271 2445e2452

McLachlan A J Ladle R amp Crompton B 2003 Predator-prey

interactions on the wing aerobatics and body size among dance

flies and midges Animal Behaviour 66 911e915

Malo A F Roldan E R Garde J Soler A J amp Gomendio M2005 Antlers honestly advertise sperm production and quality

Proceedings of the Royal Society of London Series B 272 149e157

Marshall L D 1988 Small male advantage in Parapediasia teterrella

and Agriphila plumbifimriella American Midland Naturalist 119412e419

Metcalfe N B amp Monaghan P 2001 Compensation for a badstart grow now pay later Trends in Ecology amp Evolution 16

254e260

Minakawa N Munga S Atieli F Mushinzimana E Zhou GGitheko A K amp Yan G 2005 Spatial distribution of anopheline

larval habitats in Western Kenyan highlands effects of land cover

types and topography American Journal of Tropical Medicine andHygiene 73 157e165

Moreira L A Wang J Collins F H amp Jacobs-Lorena M 2004Fitness of Anopheline mosquitoes expressing transgenes that

inhibit Plasmodium development Genetics 166 1337e1341

Neems R M Lazarus J amp McLachlan A J 1998 Lifetime repro-

ductive success in a swarming midge trade-offs and stabilizing

selection for male body size Behavioral Ecology 9 279e286

Ngrsquohabi K R John B Nkwengulila G Knols B G J KilleenG F amp Ferguson F M 2005 Effect of larval crowding on mating

competitiveness of Anopheles gambiae mosquitoes MalariaJournal 4 49

Partridge L amp Farquhar M 1983 Lifetime mating success of malefruitflies (Drosophila melanogaster) is related to their size Animal

Behaviour 35 871e877

Peckarksy B L Mclntosh A R Caudill C C amp Dahl J 2002

Swarming and mating behavior of a mayfly Baetis bicaudatus

ANIMAL BEHAVIOUR 76 1112

suggest stabilizing selection on male body size Behavioral Ecology

and Sociobiology 51 530e537

Reisen W K Aslam Y amp Siddiqui T F 1977 Observation on the

swarming and mating of some Pakistan mosquitoes in nature

Annals of the Entomological Society of America 70 988e995

Scott T W Takken W Knols B G amp Boete C 2002 The ecology

of genetically modified mosquitoes Science 298 117e119

Sparkes T C Keogh D P amp Pary R A 1996 Energetic costs of

mate guarding behavior in male stream-dwelling isopods Oecolo-gia 106 166e171

Tabachnick W J 2003 Reflections on the Anopheles gambiaegenome sequence transgenic mosquitoes and the prospect for

controlling malaria and other vector borne diseases Journal of

Medical Entomology 40 597e606

Thornhill R A amp Alcock J 1983 The Evolution of Insect Mating

Systems Cambridge Massachusetts Harvard University Press

Tripet F Toure Y T Dolo G amp Lanzaro G C 2003 Frequency

of multiple inseminations in field-collected Anopheles gambiae

females revealed by DNA analysis of transferred sperm AmericanJournal of Tropical Medicine and Hygiene 68 1e5

Van Handel E 1988 Assay of lipids glycogen and sugars in individ-

ual mosquitoes correlations with wing length in field collected

Aedes vexans Journal of the American Mosquito Control Association4 549e550

WHO 1993 Implementation of the Global Malaria Control StrategyReport of a WHO Study Group on the Implementation of the Global

Plan of Action for Malaria Control 1993e2000 Geneva WHO

Ye-Ebiyo Y Pollack R J Kiszewski A amp Spielman A 2003

Enhancement of development of larval Anopheles arabiensis by

proximity to flowering maize (Zea mays) in turbid water and

when crowded American Journal of Tropical Medicine and Hygiene68 748e752

Yuval B 2006 Mating systems of blood-feeding flies Annual Reviewof Entomology 51 413e440

Yuval B Wekesa J W amp Washino R K 1993 Effect of body sizeon swarming behavior and mating success of male Anopheles free-

borni (Diptera Culicidae) Journal of Insect Behavior 6 333e342

Yuval B Holliday-Hanson M L amp Washino R K 1994 Energy

budget of swarming male mosquitoes Ecological Entomology 19

74e78