Journal of the American Mosquito Control Association, 9(3):316-320, 1993Copyright @ 1993 by the American Mosquito Control Association, Inc.

COMPARISON OF SAMPLING EFFICACY OFSWEEPING AND DIPPING FOR

AEDES AEGYPTI LARVAE IN TIRESTIAN-MIN ZHENr.'� nNo BRIAN H. KAY'

ABSTRACT. A sweep net with a flexible plastic frame was compared to a dipper for sampling efficacyfor Aedes aegypli larvae in different-sized tires. When known numbers of mosquito larvae were introducedinto tires, from 2 to 6.8 times more larvae were recovered by sweeping than by dipping. Sediment in thetires significantly reduced the recovery oflarvae by both methods. Sampling efrcacy was affected by waterlevel and tire size. In relation to dipping, sweeping showed less variation in percentage recovery, highersensitivity for detecting larvae and had better correlations with known numbers of larvae in tires. Weconclude that sweeping is superior to dipping for sampling mosquito larvae in tires.

INTRODUCTION

Discarded tires constitute an important breed-ing habitat for mosquitoes, especially Aedes ae-gypti (Linn.) ard, Aedes albopictus (Skuse). InPuerto Rico and Miami, respectively, scrap tiresaccounted for < l5olo and 800/o of containers har-boring Ae. aegypti (Frank l98l) whereas the af-flnity of Ae. albopictu^s for this habitat is wellknown (Hawley 1988).

Halstead (1992) considered that Ae. aegypticontrol had "a glorious past but a dismal pres-ent". This breakdown in control is partly due topolitical, financial and social reasons, and wewould contend that it is also due to overrelianceon outdated and inappropriate surveillance tools.For example, the Breteau index (Breteau 1954)is an indicator of prevalence rather than abun-dance and yet, certain levels are used to infersafety or risk of dengue transmission (WorldHealth Organization 1972). Implicit in the de-velopment of better methods for both surveil-lance and control is consideration of breedingsite productivity.

The dipper is undoubtedly the most common-ly used tool for sampling immature mosquitoes.Service (1976) reviewed different types and theirinherent bias with respect to collection ofdiffer-ent instars or species. The water inside the casingcan be siphoned (Trpis 1972) or shaken out intoa tray or individual larvae picked out. These pro-cedures are often laborious and may not reflectthe relative importance of the container type interms of Aedes production.

The plankton net is a standard zoological toolfor sampling aquatic organisms. We routinelyuse a 200 x 100 mm net of 100-pm mesh fordetecting predacious cyclopids and mosquito im-matures in tanks, drums and other large con-

I Queensland Institute of Medical Research, P. O.Royal Brisbane Hospital, Brisbane, Queensland 4029,Australia.

'�Shandong Institute of Parasitic Diseases, Jining,Shandong 272133, P.R. ofChina.

tainers (Brown et al. 1992). Nets are particularlyuseful for sampling large habitats, especially whenimmature populations are low (Heathcote 1970,Service 1976).

In this paper, we describe a net with a flexibleplastic frame that can accommodate differentsized tires. We compare the practicalities andrecovery rates of this net with a dipper with re-spect to detection of larvae and estimation ofabsolute abundance.

MATERIALS AND METHODS

Sampling devlces.' The lO0-pm-mesh net wasmounted on a flexible polyvinylchloride framewith a circumference of 63 cm and attached toa short handle (Fig. l). The frame is constructedwith 2 interlocking cable ties (Wattmaster P/L,Sydney). The ties, 548 x 12 (ribbed width 8 mm)x 2 mm, were threaded together via the top claspand looped together so that the bottom half ofthe loop was of 2 thicknesses of cable tie. Eachend was then fixed to its partner halfway up itslength with a press stud. The shaft of the loopwas then fitted with either a press stud or a leatherloop that was used to adjust net circumference.The seam ofthe net was kept on the outside sothat mosquito larvae would not be trapped inthe seam or be injured by it. Because ofthe flex-ibility ofthe frame, the volume sampled dependson the dimensions of the tire and also on thevolume of water therein.

A l4-cm-diam metal dipper holding 350 mlof water was fixed to a wooden handle 35 cmIong. The inside of the dipper was white to fa-cilitate detection of larvae.

Sampling methods: Three types of tires wereevaluated: A) light truck, 8-ply non-radial, 70cm diam x 17 cm wide; B) radial automobile,56 x 17.5 cm, and C) tractor, 128 x 70 cm.When standing and filled with water, their ca-pacities were 4, 3 and 28 liters, respectively.

For each tire,20,50, 100, 150 and 200 eachof 2nd- to 4th-instar Ae. aegypti from the

3 1 6

3 r 8 Jour.Ner, or rns Ar,reRrcAN Mosqulro Coxrnol Assocrerrorl Vol. 9, No. 3

SWEEPING (TIRE A) SWEEPING (TIRE B)

Sediment

6 0

20

60

5 0

20

1 01 0Mr4oQf*ltrt{(J

Fzl'1()Ef'lA

5 0 50

40 40

5 0 50

50 1 0 0 1 5 0 2 0 0 1 0 0 1 5 0 2 0 0

5 0 1 0 0 1 5 0 2 0 0 5 0 1 0 0 1 5 0 2 0 0

NUMBER OF FOURTH INSTAR LARVAE INTRODUCEDFig.2. Effect ofsediment on percentage recovery of Aedes aegypti larvae by sweeping and dipping in tires

A Oight truck) and B (automobile).

water level (F : 77.8, df : 2, P < 0.0001), sed-iment (F: 55.3, df : l , P < 0.0001), and instarlevel (F : 14.5, df : 2, P < 0.0002). Togetherthese factors accounted for 29.60/o of the vari-ability in the model (r, : 0.2956, n : 787, P <0.0001).

Sediment: For sweeping, in the presence of4th-instar larvae, sediment in the tires was themost important factor influencing recovery ratesbut it was only 3rd in importance for dipping(Fig. 2). When sediment was present, catches bynet were 2.3 times lower overall (e.g., in tire Awithout sediment, the average recovery was 37.7

+ 9.4o/o but this was reduced to 16.3 + 5.6yo).By dipper, average catch in tire A fell from 9.6+ 4.lo/oto 6.4 + 3.4o/o.

Tire size: For both sweeping and dipping, tiresize was the 2nd most important factor. For the3 tires with 3-28 liters capacity, the average dif-ference in percentage recovery in the initial trialswas greater for dipping (l l3olo increase) than forsweeping (590/o). These trends are consistent whendifferent water levels (Table l) and instars (Table2) are also considered. For tires A and B, therewas no significant difference between percentagerecoveries by sweeping (t : 1.56, df: 543, P >

DIPPING (TIRE A) DIPPING (TIRE B)

SerreMsEn 1993 SweepINc lNo DIpprNc Tnss

Table l. Percentage recovery (+SD) of different instars of Ae. aegypti lawae in different waterlevels in tires without sediment.

Water level

Method TireNo.

samples Vr full

3 1 9

Full 2/s full

Sweeping

Dipping

ABC

9.4to.24 . 1

8 .7 + 4 .19.6 + 4.55.2 + 3.2

ABC

272272240272272240

31.9 + 12.829.0 + 12.223.6 + 7.3

4 5 . 8 + l l . l47.0 + 10.8

l4 . l + 4 .016.3 r 5 .3

43.1 + l l.740.2 ! lo.536 .4 + 11 .5

18 .7 +20.8 +

8 .9 +

0.05) but more larvae were always sampled fromtire B than A when dipping (t:2-55, df : 543,P < 0.05).

Water level: The volume of water inside thetire was the most important factor affecting re-covery rates using a dipper but was the 3rd mostimportant factor afecting sweeping. For sweep-ing and dipping, respectively, recoveries werehighest when the tires were 2/t and Vz full (Tablel), but both ofthese gave higher percentage re'turns compared to full tires.

Number of lamae introduced: In relation tothe models developed for both sweeping (P <0.001) and dipping (P: 0.09), this factor had aminor influence compared to the 3 presentedabove (Fig. 2).

Larval instar: Although differences in recov-ery rates of2nd-, 3rd- and 4th-instarlarvae weresignificant for both methods, this factor was ofminor importance to the regression models. Thepercentage recovery of 2nd-instar larvae (Table2) was lower than for the other instars tested byeither sweeping (F : 85.3, df : 2, P < 0.001) ordipping (F : 45.6, df : 2, P < 0.001).

DISCUSSION

The net described herein is light, easily carriedand inexpensive (U.S. $3) and if necessary, canbe fitted with an aluminum handle for more gen-eral sampling purposes. For sampling water in-side tires, the cable ties extending above the net-ting could be comfortably used for this purpose.The pigskin leather around the flexible frameensures its durability.

In surveying container habitats, the main ob-jective is usually detection of immature stages ofAe. aegypti and other species, in order to definerisk using a Breteau index (Breteau 1954). Withrespect to the 3 tire sizes chosen under a varietyof conditions. we demonstrated that this netwould detect immatures more consistently thana dipper. In our hands, all net samples were pos-itive, whereas dipping returned 20 negative fromI 5 8 ( I 2. 7%) trials with 20 larvae introduced. With

50 larvae introduced, 3 of 166 (1.8o/o) sampleswere negative by dipping. Thorough detection ofbreeding sites is a prerequisite of successful andlasting control, especially in the latter stages ofa campaign when larval densities are low.

For this reason, we chose to introduce andevaluate sampling efficacy with respect to 20-200 larvae. It is also relevant that in northQueensland (Tun-Lin 19923) and in Fiji (Andreet al. 19924), the average number of Ae' aegyptilarvae per tire was 90 (n:68) and 157 (n: 68),respectively, and in Louisiana, Marten (1990)recorded an average of 142 Ae. albopictus pettire. Hence, we believe that the range chosen forour evaluation applies to natural conditions butacknowledge that tires with fewer or greaternumbers of immatures do exist.

With respect to estimation of productivity, thelower coefrcients of variation of sweeping, usu-ally less than the mean recovery rates, indicatedthat it is more reliable than dipping, where co-efficients of variation ranged from 3.4 to ll.9times higher. In view of the importance of thepresence of sediment in the bottom of the tire,tire size, and water volumes, we were surprisedthat there was such a high correlation (r: 0.89,n : 787, P < 0.001) between numbers sampledand numbers introduced. However, we have notincluded the regression formulae because of therelatively small number of variables examined.

Because the net could sample almost all of thewater in the radial automobile and the 8-ply liehttruck tires, its efficiency was clearly higher thanin the larger tractor tire, which was more than 4

rTun-Lin, W. 1992. Studies on the ecology andbiology of Aedes (Stegomyia) aegypti (Linnaeus) (Dip-tera: Culicidae) immatures in Queensland, with specialreference to improved surveillance. Ph.D. thesis. Trop-ical Health Program, University of Queensland.

a Andre, R. G., B. H. Kay, E. Kikau and S. H. Wa-terman. 1992. Assessment of the epidemiology ofdengue and its past, present and future control in Fiji.Vector Biology and Control Project Report No. 8 I 269,Arlington, VA.

320 Jounr.tal or rHE AuenrcAN Mosqurro Covrnor Assocrerron Vor-. 9, No. 3

Table 2. Percentage recovery (+SD) of different instars of Ae. aegyptl in tires full of water andwithout sediment. bv swee and dipping.

No.samples

Larval instar

Method IVIIIIITire

Sweeping

Dipping

ABC

ABC

l 3 ll 3 l120

36.034.r25.6

t l .4r 3 . l4 .4

3 7 . t + 9 . 133 .8 + l 0 . lt8.2 + 5.2

40.1 + 10.635 .0 + 9 .727.2 + 6.3

10.3 +10.9 +6 .7 +

+ 8 . 1+ 8 .8+ 7 . 1+ 2.9+ 4 . 1+ 2.6

4.54.03 .5

l 3 ll 3 r120

8.0 + 3.99 .7 + 3 .94.4 + 3.0

times wider. Overall the net collected 4 timesmore than the dipper. However, sweeping wasmore affected by the presence of bottom sedi-ments because, in contrast to dipping, which pri-marily took a surface sample, the net sampledthe total water column. Recovery rates werehigherwith both methods when water levels werereduced but it appears likely that the sweepingprocedure is hindered somewhat with small vol-umes. The edges ofa dipper are finer than thoseof a net and are more able to scoop up anythingremaining. Although more immatures were al-ways collected by dipper in tire B (nonradial trucktire) than in tire A (radial automobile tire), shapewas also important to recovery, but for sweepingthis was unimportant because the net was ableto accommodate the different shapes.

More 3rd- and 4th-instar larvae were collectedthan 2nd instars, as was found by Tun-Lin ( I 9923)in 220liter drums. It is well established that firstand second instar Ae. aegypti will remain sub-merged for longer periods than any of the otherstages, particularly at cooler temperatures(Christophers I 960, Tun-Lin 19923).

We recommend the use of the flexible net forsurveys of container-breeding Aedes as a meansof reducing effort and increasing precision, es-pecially as we have frequently encountered dif-ficulty in inserting dippers and removing sampleswhere the gap between the tire walls is narrow.As most personnel will only be interested in es-tablishing presence or absence ofimmatures, thisnet saves time. It also has the dual benefit ofbeing suitable for sampling predacious copepods,promising biological control agents for Aedes(Rividre et al. 1987, Marten 1990).

ACKNOWLEDGMENTS

This study was carried out while T.-M. Zhenwas a recipient of a WHO Research TrainingGrant under the UNDP/World BanVWHO Spe-cial Program for Research and Training in Trop-ical Diseases. We thank Jim and Kay Marshalland Alan Stockman for making the net, S. A.

Ritchie and D. McManus for review of themanuscript and J & J Tyre Service, West End,Brisbane, for providing the tires.

REFERENCES CITED

Breteau, H. 1954. La fievre jaune en Afrique occi-dentale francaise. Un aspect de la medicine preven-tive massive. Bull. W.H.O. I l:453-481.

Brown, M. D., P. Mottram, I. D. Fanning and B. H.Kay. 1992. The peridomestic container-breedingmosquito fauna of Darnley Is. (Torres Strait) (Dip-tera: Culicidae), and the potential for its control bypredacious Mesocyclops copepods. J. Aust. Entomol.Soc .3 l :305-310.

Christophers, S. R. 1 960. Aedes aegypti (L.): the yel-low fever mosquito: its life history, bionomics andslructure. Cambridge lJniv. Press, Cambridge.

Frank, J. H. 1981. Recycling of discarded tires forcontrol of Aedes aegypti. J. Fla. Anti-Mosq. ControlAssoc. 52:44-48.

Halstead, S. B. 1992. Introductory remarks, pp. 15-17 . In: S. B. Halstead and H. Gomez-Dantes (eds.).Dengue a worldwide problem, a common strategy.Ministry of Health, Mexico and Rockefeller Foun-dation, New York.

Hawley, W. A. 1988. Thebiology of Aedes albopictus.J. Am. Mosq. Control Assoc. 4 (Suppl. 1): l-40.

Heathcote, O. H. U. 1970. Japanese encephalitis inSarawak: studies on juvenile mosquito populations.Trans. R. Soc. Trop. Med. Hyg. 64:483-488.

Marten, G. G. 1980. Elimination of Aedes albopictusfrom tire piles by introducing Macrocyclops albidus(Copepoda, Cyclopidae). J. Am. Mosq. Control As-soc. 6:689-693.

Rividre, F., B. H. Kay, J.-Ivt. Klein and Y. S6chan.1987 . Mesocyclops aspericornrs (Copepoda) and Ba-cillus thuringiensrs var. israelensis for the biologicalconfiol of Aedes and Culex vectors (Diptera: Culic-idae) breeding in crab holes, tree holes and artificialcontainers. J. Med. Entomol. 24:425430.

Service, M. W. 1976. Mosquito ecology. Field sam-pling methods. Applied Science Publishers, London.

Trpis, M. 1972. Seasonal changes in the larval pop-ulations of Aedes aegypti in two biotopes in Dar esSalaam, Tanzania. Bull. W.H.O. 47:245-255.

World Health Organization. 1972. An internationalsystem for the surveillance of vectors. Wkly. Epi-demiol. Rec. 47:73-80.