Biologia 65/3: 545—551, 2010Section ZoologyDOI: 10.2478/s11756-010-0051-5

Efficacy of mosquito attractants in various habitats of a floodplain

Enrih Merdic, Željka Jeličic, Stjepan Krčmar, Branimir Hackenberger-Kutuzovic,Nataša Turic, Mirta Sudaric Bogojevic & Željko Zahirovic

Josip Juraj Strossmayer University in Osijek, Department of Biology, Trg Ljudevita Gaja 6, 31000 Osijek, Croatia; e-mail:enrih@biologija.unios.hr

Abstract: Efficacy of different mosquito attractants was invesigated at four sites in three plant communities (Galio-Salicetum albae, Populetum nigro-albae, Genisto elatae-Quercetum roboris) of flooded and forest habitats in Kopački ritNature Park, Croatia. The attractants were: dry ice, horse urine, horse urine + acetone, acetone, 1-octen-3-ol and ammoniumhydroxide baited CDC traps. A total of 11,441 mosquito specimens of 12 species were collected. Aedes vexans (91.43%) wasthe most numerous species. A statistically significant difference between the efficacy of dry ice and the other attractants wasshown, whereas there was no difference between the other attractants. A greater number of specimens and species numberwere noted in the flooded plant communities (Populetum nigro-albae). The response of Ae. vexans to dry ice was higher inflooded sites, and it was significantly lower in forest habitat (Genisto elatae-Quercetum roboris) according to fuzzy c-meanscluster analysis. The same analysis shows a higher efficacy of other attractants (horse urine, horse urine + acetone, acetone,octenol and ammonium hydroxide) in forest habitat when compared to the flooded area habitats.

Key words: mosquitoes; attractants; flooded habitat; forest habitat; Kopački rit Nature Park

Introduction

In an effort to find adequate control methods to groundand aerial applications of chemical insecticides formosquito control, several host emanations (carbon diox-ide [CO2], 1-octen-3-ol [octenol], butanone) have beenevaluated against mosquitoes associated with a varietyof habitats including salt marshes, mangrove swamps,phosphate mining pits, rice field, and livestock holdingareas. Mosquitoes associated with these habitats variedgreatly in their responses to different release rates andcombinations of these attractants (Kline 1994).Rudolfs (1922) first reported the value of different

compounds (alanine, ammonia, cholesterol, haemoglo-bin, phenylalanine, vaseline) as attractants for mosqui-toes. Newhouse et al. (1966) showed that, when CO2was used with Centers for Disease Control (CDC)light traps, mosquito collections increased 4-fold andthe number of species collected increased by 20–26%.Besides CO2, numerous other substances have beenshown (butanone, honey extract, octenol, L-lactic acid,phenols, 4-methylphenol, 3-n-propylphenol, ammoniumetc.) to attract mosquitoes (Kline et al. 1990; Bosch etal. 2000; Mihok et al. 2007; Merdic et al. 2007). Ofthese, octenol appears the most promising (Cooper etal. 2004).During the last decade much attention has been

given to different attractants for collecting mosquitoes(Miller et al. 2005; Hurk et al. 2006; Kline 2006). Ruedaet al. (2001) compared attractants and their combina-tions, noticing different relationships in the responses

of mosquitoes collected in different habitats (salt marshand inland creek flood plain). Thus, baited adult trapscould have real value in accruing more target speciesduring mosquito-borne virus surveys (Kline 1994). Theflooded and forest habitat types which are in Kopačkirit Nature Park are ideal for mosquito development.Mosquitoes are insects which live in a world of scents;according to Jordan (1995) efficacy of the attractantsdepend on the species and the relationships betweenthe species and habitat. During this study different at-tractants were used to collect mosquitoes in order toobserve any differences in response or effectiveness be-tween the flooded and forest habitats within Kopačkirit Nature Park and between the different attractants.

Material and methods

Study areaOne of the largest remaining floodplain areas in the wholeof Central Europe lies where the Danube and Drava riversmeet. Including territories belonging to Hungary, Croatiaand Serbia, this boundary area stretches from Szekszard inHungary downstream to Dalj in Croatia and Bogojevo inSerbia. The main features of the Danube floodplain are itsregular floods and dynamic water regime, the consequentsedimentation processes that create the typical floodplainrelief and its outstanding biodiversity.

The Kopački rit Nature Park is situated in the north-east of Croatia (45◦34′ N, 16◦24′ E), just in the cornermade by the confluence of the Drava with the Danube River.Kopački rit spreads over some 100 km2, (Fig. 1) it is an oval8–10 km long and 7–9 km wide. The largest part of Kopačkirit lies between 81.5–82.5 NN. In spite of its monotonous

c©2010 Institute of Zoology, Slovak Academy of Sciences

546 E. Merdic et al.

Fig. 1. The investigated area of Kopački rit Nature Park. Local-ities: 1 – Čonakut (Galio-Salicetum albae); 2 – Hordovanj (Pop-uletum nigro-albae); 3 – Kopačko lake (Galio-Salicetum albae); 4– Tikveš (Genisto elate-Quercetum roboris).

lowlands, the structure of Kopački rit is very complex. Shal-low oval or crescent formed recesses and channels directlyconnect Kopački rit with the Drava and Danube allowingflood waters to penetrate into the marsh. The central partof the marsh is occupied by a large permanent water body,Kopačko Lake (ca 200–250 ha). The banks of the lake grad-ually change into marshy meadows.

Because of specific micro relief, habitats in Kopačkirit directly depend on water regime, length and level offlood. Diversity of plant communities (predominantly for-est, marsh and water) is present here. The lowest floodedterrains are covered with an association of Galio-Salicetumalbae. Here the floods are often, long lasting and necessaryif they don’t last too long. The community is monotypicwith white willow trees only. A bush layer does not exist(Rauš 1976). On slightly higher terrain where the floods areshorter and smaller the association of white willow and blackpoplar (Salici-Populetum nigrae) develops. This forest isrepresented by an subassociation of Salici-Populetum nigraerubetosum caesii, where in the bush layer blueberry (Rubuscaesius) dominate. On the highest flooded terrains an asso-ciation of black and white poplar (Populetum nigro-albae)develops. Areas of Kopački rit out of reach of flood are cov-ered with oak forest association (Genisto elatae-Quercetumroboris). The best indicator of stagnant and undergroundwater is Carpinus betulus. This species grows in terrainswhere stagnant and underground water does not exist andspreads on terrains out of reach of flood. In temporarilyflooded areas, vast sections of the marsh are overgrownwith associations of common reed and high sedges (Scirpo-Phragmitetum), and communities of high sedges (Caricetum

elatae) which give a characteristic wetland look. A total of40 plant communities and more than 300 plant species repre-sent a great vegetation diversity in Kopački rit (Topic 1989).

This area is of international importance, containinglarge areas designated as Ramsar sites (Ramsar Conventionon Wetlands, especially as Waterfowl Habitats) or Impor-tant Bird Areas (IBA) under the terms of the EU Wild BirdDirective. It is important to stress, however, that there arestill areas of floodplain, particularly in Croatia and Serbia,that are not protected or that do not have adequate protec-tion status. Since 1999, it has been nominated to be includedin the UNESCO’s list of areas of National World Heritage.This field investigation was done in four sites with differentplant communities. All sites were situated in Nature Parkof Kopački rit, and three of them were situated within theSpecial Zoological Reserve (Fig. 1):

1. Čonakut (45◦36′34.8′′ N, 18◦48′23.9′′ E) – floodedhabitat; plant community Galio-Salicetum albae

2. Hordovanj (45◦36′35.8′′ N, 18◦50′2′′ E) – floodedhabitat; plant community Populetum nigro-albae

3. Kopačko lake (45◦36′26.7′′ N, 18◦50′18.7′′ E) –flooded habitat; plant community Galio-Salicetum albae,with colony of cormorants (Phalacrocorax carbo) on willowtrees

4. Tikveš (45◦41′44.95′′ N, 18◦49′31.4′′ E) – foresthabitat; plant community Genisto elatae-Quercetum roboris

Trap and attractantsTo collect adult mosquito specimens home made Center forDisease Control (CDC) traps without light (Service 1976)were used in every instance. The traps were baited with 6different attractants: dry ice (9 kg), horse urine (40 ml),a combination of horse urine (10 ml) + acetone (30 ml),acetone (40 ml), octenol (4 ml) and NH4OH (4 ml).

The liquid attractants were placed in small glass con-tainers 3 cm in diameter and 5 cm high. To facilitate thesetting up of liquid attractants, the glass containers wereplaced in plastic 500 ml cups that were tied with a wire totree branches. Cotton strips were placed in the attractantsand hung over the lip of the cup to increase the range of theattractants scent. The horse urine had been collected in theearly morning from the same horses, and left for 14 days inthe sun. Other attractants used were uncontaminated (Ace-tone – 0.791 kg L−1, Kemig d.o.o., NH4OH – 25%, AlkaloidSkopje, 1-octen-3-ol 98%, Aldrich). Because dry ice subli-mates at 400 ml/min (Van Essen et al. 1994; Johansen etal. 2003) we calculated that 9 kg would be enough for 24hours. Octenol was released at a rate of 4.5 mg/h from 5-mlvials attached adjacent to the opening of the trap cylinder(Van Essen et al. 1994).

Field trialsTrapping was carried out 12 times in the field from Septem-ber 2006 to October 2007. At every location six CDC trapswith different attractants were set. The distance betweenthe traps was 50 m (according to Hurk et al. 2006). Trapswere set up at a height of about 1.5 meters above groundand operated for 24 hours. Identification of mosquitoes andnomenclature followed Gutsevich et al. (1970), Schaffner etal. (2001) and Becker et al. (2003).

Statistical AnalysisA one-way ANOVA with post-hoc pair-wise comparisons(Tukey) was used to compare efficacies of the various at-tractants. Differences were considered not significant for P> 0.05.

Efficacy of mosquito attractants 547

Table 1. The number of specimens of each mosquito species captured per attractant.

Attractants / species CO2 U U+A A O NH4OH Σ

Aedes vexans (Meigen,1830) 10240 47 52 45 56 20 10460Culex modestus (Ficalbi, 1889) 645 5 6 3 4 663Anopheles messeae (Falleroni, 1771) 132 3 3 1 9 1 149Ochlerotatus caspius (Pallas, 1771) 44 44Aedes cinereus (Meigen, 1818) 40 2 1 43Anopheles hyrcanus (Pallas,1771) 24 24Culex pipiens k. (L., 1758) 14 1 2 1 18Aedes rossicus (Dolbeskhin, Gorickaja, Mitrofanova, 1930) 16 16Ochlerotatus sticticus (Meigen, 1838) 10 10Cocquillettidia richiardi (Ficalbi, 1889) 7 7Anopheles plumbeus (Stephanus, 1828) 4 1 5Anopheles claviger (Meigen, 1804) 2 3

Total 11178 55 62 53 66 27 11441

Explanations: CO2 – dry ice; U – horse urine; U + A – horse urine + acetone; A – acetone; O-1-octen-3-ol; NH4OH – ammoniumhydroxide.

Traditional clustering algorithms generate partitions ofdata where each data instance belongs to one and only onecluster. Fuzzy clustering on the other hand associates eachdata instance with a membership function for each cluster.The output of such clustering is not a partition of data, butan assignation of cluster memberships to all data instances.Each cluster is a fuzzy set of all data instances (Jain et al.1999). For the classification of the four sites fuzzy cluster-ing statistical analysis was used. Fuzzy clustering analysis(Equihua 1990) was performed, in order to obtain classifi-cation of the sites. Fuzzy technique produces clusters of theassemblages of mosquitoes collected on each sites, accordingto their membership values. The numbers of fuzzy clustersare selected according to the higher partition coefficient.

All statistical tests were performed with R softwareenvironment (R Development Core Team 2007), using the“cluster” software package (Maechler 2008).

Results

During the investigation a total of 11,441 specimenswere caught that belonged to 12 species of the followinggenera: Aedes (3), Anopheles (4), Culex (2), Ochlerota-tus (2), Cocquillettidia (1) (Table 1).CDC traps that used dry ice as an attractant

trapped 12 species; 5 species were trapped with ace-tone and NH4OH as an attractant; the attractant com-bination horse urine + acetone caught 4 species, andonly three species were trapped with horse urine and 1-octen-3-ol (Table 1). Out of the 12 species, Ae. vexans,Cx. modestus and An. messeae made up 98.52% of thecatch, while the remaining species (Ae. cinereus, Oc.caspius, An. plumbeus, Oc. sticticus, An. claviger, Cx.pipiens, Ae. rossicus, Cq. richiardii and An. hyrcanus)made up 1.48% (Fig. 2). Eudominant species during theresearch was the species of Ae. vexans which made up91.43% of all mosquitoes captured.The traps placed in the plant community Popule-

tum nigro-albae (Hordovanj) trapped the largest num-ber of mosquitoes (8,452) as well as the largest va-riety of species (11). In the plant community Galio-Salicetum albae two locations were monitored and a dif-ference in number and species composition were noted.

Ae. vexans91.43%

Cx. modestus5.79%

An. messeae1.30%

Other species1.48%

Fig. 2. Proportion of captured mosquitoes.

Within that community the location Čonakut recordeda greater number of individuals (1,727) and species(10) trapped than did the location Kopačko lake whichyielded 851 individuals and 9 species. At the locationTikveš, within the plant community Genisto elatae-Quercetum roboris, the smallest number of individualswas collected (411), but the variety of species was sur-prisingly high (10) (Fig. 3).The investigation in this study lasted two years

(2006–2007). Two-way ANOVA without replication didnot show a statistically significant difference betweenthese years (P = 0.0024). Since there was no statisticaldifference, further analysis was carried out on the datafrom both years.When comparing similarities between the attrac-

tants the only significant difference found was betweendry ice and the other attractants (Tukey multicom-parison test, P = 0.004). Figure 4 shows the statisti-cal difference between every other attractant and CO2(CO2 �= horse urine, CO2 �= horse urine + acetone,CO2 �= acetone, CO2 �= octenol, CO2 �= NH4OH). Thesame statistical test compared the other attractantswithout CO2 and showed no statistical difference.On every investigation site CO2 had the highest

efficacy. The number of mosquitoes trapped with CO2was greatest at the Hordovanj site, with a plant com-munity of Populetum nigro-albae, and it accounted for98.7% of all mosquitoes captured. Similar values wererecorded in other sites within flooded habitats (Ta-ble 2). The amount of mosquitoes captured with CO2

548 E. Merdic et al.

0

1

2

3

4

5

6

7

8

9

10

11

12

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

Galio-Salicetum albae Populetum nigro-albae

Genisto elatae-Quercetum roboris

Galio-Salicetum albae

Čonakut Hordovanj Tikveš Kopačko lake

Num

ber o

f cap

ture

d sp

ecie

s

Num

ber o

f cap

ture

d sp

ecim

ens

Total specimens Total species

Fig. 3. The number of captured specimens and species in different plant communities.

Table 2. Proportion of mosquitoes caught in different plant communities in CDC traps baited with the different attractants.

Sites/attractants CO2 U U+A A O NH4OH

Čonakut 0.959 0.003 0.007 0.012 0.016 0.003Hordovanj 0.987 0.003 0.004 0.002 0.002 0.001Tikveš 0.864 0.044 0.022 0.024 0.032 0.015Kopačko lake 0.973 0.004 0.004 0.005 0.009 0.006

Explanations: CO2 – dry ice; U – horse urine; U + A – horse urine + acetone; A – acetone; O-1-octen-3-ol; NH4OH – ammoniumhydroxide.

Table 3. Membership factors of each location to assumed clustersaccording to fuzzy c-means cluster analysis (according to resultsfor all types of attractants); within error: 2.452997.

Rank ofSites Cluster I Cluster II similarity to Cl I

Čonakut 0.637 0.363 1Hordovanj 0.932 0.068 3Tikveš 0.011 0.989Kopačko lake 0.881 0.119 2

at Tikveš, with a plant community of Genisto elatae-Quercetum roboris, resulted in 86.4% of all capturedmosquitoes, significantly less than flooded habitats butstill far superior to all other attractants.The analysis of the sites with Fuzzy c-means clus-

ter resulted in two clusters (within error 2.45). Clus-ter I is made of three sites: Čonakut, Hordovanj andKopačko lake because attractant CO2 had a muchhigher efficacy than other attractants. Cluster II ismade of only one site, Tikveš, because the efficacy ofCO2 was less (Table 3).Clusters formed without CO2 data again showed

that Tikveš site made a separate cluster (within er-

Table 4. Membership factors of each location to assumed clustersaccording to fuzzy c-means cluster analysis (according to resultsfor all used attractants except CO2); within error: 2.092386.

Rank ofSites Cluster I Cluster II similarity to Cl I

Čonakut 0.720 0.280 1Hordovanj 0.931 0.069 3Tikveš 0.006 0.994Kopačko lake 0.854 0.146 2

ror 2.09). On that site the other attractants: horseurine, horse urine + acetone, acetone, octenol, NH4OHshowed a higher efficacy than in the other sites (Ta-ble 4).If divided into three clusters with all attractants

the results are: cluster I-Hordovanj, cluster II-Čonakutand Kopačko jezero and finally cluster III-Tikveš(within error 1.06). Results without CO2 data are: clus-ter I-Čonakut and Hordovanj, cluster II-Hordovanj andKopačko lake and Cluster III-Tikveš (within error 0.91).These results indicate that the Tikveš site has a

distinct characteristic compared to the other three in-vestigation sites.

Efficacy of mosquito attractants 549

-20 -10 0 10 20

U+A

-UU-O

U-N

H4O

HU-C

O2

U+A

-AO

-ACO2-

A

95% family-wise confidence level

Differences in mean levels of Treat

Fig. 4. Confidence intervals according to Tukey posthoc test.

Discussion

During previous research in Kopački rit, 19 mosquitospecies were noted (Merdic et al. 2004). Within thefauna of Kopački rit, during this investigation 12mosquito species were found, of which 7 (Ae. vexans,Oc. sticticus, Ae. cinereus, An. messeae, An. hyrcanus,Cx. pipiens, Oc. caspius) have breeding sites withinflooded area of Kopački rit (Merdic et al. 2004) whichin certain situations can produce huge populations. Thereason that the population of Ae. vexans can becomeso large (91.41%) is that this species finds ideal breed-ing conditions in Kopački rit: shallow swamps with orwithout vegetation and enough sunlight on that placesmajor role as fluctuation of the water level which pro-duce new generations. Furthermore, Ae. vexans is anautogenous species (Horsfall et al. 1973) which probablyaccounts for its making up to 95% of the mosquito pop-ulation in the area (Merdic 1993). Two other species,Cx. modestus and An. messeae, also occur in big num-bers, and the remaining species are usually rare. Duringa year of average flooding, the composition of mosquitofauna does not vary (Merdic 1993) but a flood of longduration brings changes in mosquito fauna (Merdic &Sudaric 2003), as do long dry periods. The floods ofthe Danube and Drava rivers, therefore, have a big im-pact on the flora and fauna of Kopački rit as a wetlandhabitat.Studies have shown a variety of responses to dif-

ferent mosquito attractants. Many studies show thatCO2 is one of the best attractants (Dekker et al. 2005;Cooperband & Cardé 2006; Merdic et al. 2007). Resultsin this research support this thesis. An ANOVA testshows a statistically significant difference between at-tractants. The Tukey multicomparison test used proveda statistically significant difference between dry ice andother attractants. Other attractants within this study

do not differ significantly. As expected, CO2 had thehighest efficacy at all sites.A difference in the qualitative and quantitative

mosquito compositions in different habitats is expectedand noted in many papers (Rueda et al. 2001; Cooperet al. 2004; Pope et al. 2005; Dale et al. 2008). How-ever, this study points out the relatively small num-ber of species within the association of Genisto elatae-Quercetum roboris caused by consecutive dry seasons,while previous research had shown a large number ofspecies in this habitat (Merdic & Lovakovic 1999)The Fuzzy c-means cluster analysis was used be-

cause of its reliability and the nature of the data, morespecifically because of probability to classify each el-ement to a cluster (Equihua 1990; Jain et al. 1999).The results show that the sites of Čonakut, Kopačkolake and Hordovanj are very similar, but Tikveš is dif-ferent. Tikveš made a separate cluster because on thissite the efficacy of attractants (dry ice and others) issignificantly different from the other sites.Small differences recorded between the sites of

Čonakut, Kopačko lake and Hordovanj are most likelythe result of a difference of vegetation. All three men-tioned sites lie in flooded areas, and the vegetationis basically characterized as open forests. Rare andrelatively low trees without bushes (Topic 1989) en-able big drafts to circulate through the habitat andodors can easily disperse through the area and outof the system. In this case the odors came onlyfrom trap attractants spreading through the habi-tat.The results of long term investigations (10 years)

in Kopački rit have shown that the Hordovanj site hasthe biggest number of mosquitoes (Merdic & Lovakovic1999). That site is covered by plant community Pop-uletum nigro-albae, which is open, but with a denselow bush layer dominated by blueberry (Rubus caesius).

550 E. Merdic et al.

This bush layer serves mosquitoes as a perfect shelterfrom different climate factors.The site Tikveš with oaks (Genisto elatae-Querce-

tum roboris) represents typical forest habitat. In thatsite the number of mosquitoes depends on migrationsfrom flooded areas of Kopački rit. This is because imme-diately after emergence, adult mosquitoes begin migrat-ing (Bidlingmayer 1985), proceeding in several stepsthat depend upon such factors as wind directions andvelocity, the reserves of the mosquitoes and meteoro-logical conditions. However, the richness of mosquitofauna correlates directly with spring rainfalls. Inves-tigated years within this work were dry, for example,and in such conditions fewer mosquito species devel-oped. Oak forest is characterized as distinctly closedwith high trees and a dense crown of trees, with a densebush layer also. In such closed conditions the wind isweak and there is a greater possibility of other odorspresent. In these conditions response of mosquitoes toodors other than CO2 is higher. The presence of manyanimals (breath) and plants (especially Leguminozae)are a trigger that make mosquito receptors keen to otherodors.Since Ae. vexans is predominant in all sites, the

possibility of other species accounting for any differ-ences observed is excluded. Therefore different attrac-tant efficacy at Tikveš site cannot be explained by thepresence of other mosquito species which might havehad a specific and unique response to investigated at-tractants. On the basis of this investigation it cannot beexplained which factors from the environment causedAe. vexans, to be more sensitive to other attractantsat the Tikveš site, nor what factors influence other pa-rameters (distribution and concentration) relevant forattractant efficacy. But this study does show a differentefficacy of attractants at Tikveš site when compared tothe other sites. Since this site differs in vegetation, wa-ter regime and abiotic factors (microclimate, humidity,wind velocity, etc.) compared to the other three sites,it raises new questions for further investigation aboutwhich factors influenced these results.

Acknowledgements

We would like to thanks Rebecca and Larry Javorsky forimprovements in this manuscript.

References

Becker N., Petric D., Zgomba M., Boase C., Dahl C., Lane J. &Kaiser A. 2003. Mosquitoes and their Control. Kluwer Aca-demic/Plenum Publishers, New York, 528 pp.

Bidlingmayer W.L. 1985. The measurement of adult mosquitopopulation changes-same considerations. J. Am. Mosq. Con-trol Assoc. 3: 328–348.

Bosch O.J., Geier M. & Boeckh J. 2000. Contribution of fattyacids to olfactory host finding of female Aedes aegypti. Chem.Senses 25: 323–330.

Centers for Disease Control and Prevention (CDC) 1999. WestNile-like viral encephalitis – New York (1999). Morb. MortalWkly Rep. (MMWR) 48: 890–892.

Cooper R.D., Frances S.P., Popat S. & Waterson D.G.E. 2004.The effectiveness of light, 1-octen-3-ol, and carbon dioxide asattractants for Anopheline mosquitoes in Madang Province,Papua New Guinea. J. Am. Mosq. Control Assoc. 20: 239–242.

Cooperband M.F. & Cardé R.T. 2006. Comparison of plumestructure of carbon dioxide emitted from different mosquitotraps. Med. Vet. Entomol. 20: 1–10. DOI 10.1111/j.1365-2915.2006.00614.x

Dale P.E.R., Knight J., Kay B.H., Chapman H., Ritchie S.A. &Brown M.D. 2008. Habitat characteristics and eggshell distri-bution of the salt marsh mosquito, Aedes vigilax, in marshesin subtropical eastern Australia. J. Insect Sci. 25: 1–8.

Dekker T., Geier M. & Cardé R.T. 2005. Carbon dioxide instantlysensitizes female yellow fever mosquitoes human skin odors.J. Exp. Biol. 208: 2963–2972. DOI 10.1242/jeb.01736

Equihua M. 1990. FUZZY clustering of ecological data. J. Ecol.78: 519–534.

Gutsevich A.V., Monchadskii A.S. & Shtakelberg A.A. 1970.Mosquitoes. Family Culicidae. Fauna of the USSR. Diptera,Nauka, Leningrad, 3 (4), 384 pp. [English translation, Is-rael Program for Scientific Translations, Jerusalem. 1974. 408pp.].

Horsfall W.R., Fowler H.W., Moretti L.J. & Larsen J.R. (eds)1973. Bionomics and Embryology of the Inland FloodwaterMosquito Aedes vexans. University of Illinois Press, Urbana,IL, 134 pp.

Hurk V.D., Montgomery B.L., Zborowski P., Beebe N.W., CooperR.D. & Ritchie S.A. 2006. Does 1-octen-3-ol enhance trapcollections of Japanese encephalitis virus mosquito vectors innorthern Australia? J. Am. Mosq. Control Assoc. 22: 15–21.

Jain A.K., Murty MN. & Flynn P.J. 1999. Data Clustering: AReview. ACM Computing Surveys 3: 264–323.

Johansen C.A., Montgomery B.L., Mackenzie J.S. & Ritchie S.A.2003. Efficacies of the Mosquito Magnet and counter flowgeometry traps in North Queensland, Australia. J. Am. Mosq.Control Assoc. 19: 265–270.

Jordan A.M. 1995. Control of tsetse flies (Diptera: Glossinidae)with the aid of attractants. J. Am. Mosq. Control Assoc. 11:249–255.

Kline D.L., Takken W., Wood J.R. & Carlson D.A. 1990. Fieldstudies on the potential of butanol, carbon dioxide, honeyextract, 1-octen-3-ol, L-lactic acid, and phenols as attrac-tants for mosquitoes. Med. Vet. Entomol. 4: 383–391. DOI10.1111/j.1365-2915.1990.tb00455.x

Kline D.L. 1994. Olfactory attractants for mosquito surveillanceand control: 1-octen-3-ol J. Am. Mosq. Control Assoc. 10:280–287.

Kline D.L 2006. Traps and trapping techniques for adult mosquitocontrol. J. Am. Mosq. Control Assoc. 3: 490–496.

Maechler M. 2008. The Cluster Package Version 1.11.10. ClusterAnalysis Extended Rousseeuw et al., Anja Struyf and MiaHubert, 73 pp.

Merdic E. 1993. Mosquitoes (Diptera, Culicidae) of the specialzoological reserve Kopački rit (NE Croatia). Natura Croatica1: 47–54.

Merdic E. & Lovakovic T. 1999. Comparison of mosquito faunain Kopački rit in the period 1990–1998. Natura Croatica 4:431–438.

Merdic E., Krčmar S., Sudaric Bogojevic M. & Jeličic Ž. 2007.Response of mosquito to different synthetic and natural ol-factory attractants (Diptera: Culicoidae). Entomol. Gener. 4:253–261.

Merdic E. & Sudaric M. 2003. Effects of prolonged high wa-ter level on the mosquito fauna in Kopački rit Nature Park.Perod. Bil. 2: 198–193.

Merdic E., Sudraric M., Lovakovic T., Boca I. & Merdic S. 2004.Checklist of mosquitoes (Diptera, Culicidea) of Croatia. Eur.Mosq. Bull. 17: 8–13.

Mihaljevic M., Getz D., Tadic Ž., Živanovic B., Gucunski D.,Topic J., Kalinovic I. & Mikuska J. 1999. Kopački rit. Pregledistraživanja i bibliografija. HAZU, Zavod za znanstveni radOsijek, Zagreb-Osijek, 188 pp.

Mihok S., Carlson D.A. & Ndegwa P.N. 2007. Tsetse andother biting fly responses to Nzi traps baited with octenol,

Efficacy of mosquito attractants 551

phenols and acetone. Med. Vet. Entomol. 21: 70–84. DOI10.1111/j.1365-2915.2007.00665.x

Miller R.J., Wing J., Cope S., Davey R.D. & Kline D.L. 2005.Comparison of carbon dioxide-and octenol-baited encephali-tis virus surveillance mosquito traps at the shoal water baytraining area, Queensland, Australia. J. Am. Mosq. ControlAssoc. 4: 497–500.

Newhouse V.F., Chamberlain R.W., Johnstone J.G. & SudiaW.D. 1966. Use of dry ice to increase mosquito catches ofthe CDC miniature light traps. Mosquito News. 26: 30–35.

Pope K., Masuoka P., Rejmankova E., Grieco J., Johnson S. &Roberts D. 2005. Mosquito habitats, land use, and malariarisk in Belize from satellite imagery. Ecol. Soc. Amer. 15:1223–1232.

R Development Core Team 2007. R: A Language and Environ-ment for Statistical Computing, R Foundation for StatisticalComputing, Vienna, Austria, ISBN 3-900051-07–0.

Rauš Đ. 1976 Vegetacija ritskih šuma dijela Podunavlja od Alj-maša do Iloka. Annales pro Experimentis Foresticis 19: 5–75

Rudolfs W. 1922. Chemotropism of mosquitoes. New Jersey Agr.Exp. Sta. Bull. 367: 5–23.

Rueda L.M., Harrison B.A., Brown J.S., Whitt P.B., HarrisonR.L. & Gardner R.C. 2001. Evaluation of 1–octen-3–ol, car-bon dioxide, and light as attractants for mosquitoes associ-ated with two distinct habitants in North Carolina. J. Am.Mosq. Control Assoc. 17: 61–66.

Schaffner F., Angel G., Geoffroy B., Hervy J.P., Rhaiem A. &Brunhes J. 2001. The Mosquitoes of Europe/Les moustiquesd’Europe. An Identification and Training Programme (CD-Rom), Montpellier, France: IRD Éditions & EID Méditer-ranée.

Service M.W. 1976. Mosquito Ecology: Field Publishers SamplingMethods. Wiley, New York, 583 pp.

Topic J. 1989. Vegetation of the special zoological reserve ofKopački rit. Hydrobiology 182: 149–160.

Van Essen P.V.H.A., Kemme J.A., Ritchie S.A. & Kay B.H.1994. Differential responses of Aedes and Culex mosquitoesto octenol or light in combination with carbon dioxide inQueensland, Australia. Med. Vet. Entomol. 8: 63–67.

Received September 29, 2008Accepted March 5, 2010