M

DDa

b

c

a

ARRAA

KMVDDSP

1

leSsarPSar

F

h0

Acta Tropica 136 (2014) 58–67

Contents lists available at ScienceDirect

Acta Tropica

jo u r n al homep age: www.elsev ier .com/ locate /ac ta t ropica

easurement of landing mosquito density on humans

onald R. Barnarda,∗, Catherine Z. Dickersona, Kadarkarai Muruganb, Rui-De Xuec,aniel L. Klinea, Ulrich R. Berniera

Center for Medical, Agricultural, and Veterinary Entomology, USDA-ARS, Gainesville, FL 32608, USADepartment of Zoology, School of Life Sciences, Bharathiar University, Coimbatore 641-046, IndiaAnastasia Mosquito Control District, 500 Old Beach Road, St. Augustine, FL 32085, USA

r t i c l e i n f o

rticle history:eceived 27 January 2014eceived in revised form 8 April 2014ccepted 13 April 2014vailable online 22 April 2014

eywords:osquito

ectorensityetectionurveillanceeriodicity

a b s t r a c t

In traditional vector surveillance systems, adult mosquito density and the rate of mosquito-human hostcontact are estimated from the mosquito numbers captured in mechanical traps. But the design of thetraps, their placement in the habitat and operating time, microclimate, and other environmental factorsbias mosquito responses such that trapped mosquito numbers may be at variance with the numbersactually making human contact. As an alternative to mechanical traps, direct measurement of landingmosquito density enables real-time estimation of the mosquito–human-host-contact parameter. Basedon this paradigm, we studied methods to measure mosquito landing responses to a human host. Ourresults showed: (a) an 18% difference (P < 0.0001) in the mean number of female Aedes albopictus (Skuse)making initial contact with the skin (9.11 ± 0.74 min–1) compared with the number remaining on the skinfor 5 s (7.42 ± 0.69 min–1); (b) an increase (P < 0.05) in the mean per minute (min−1) landing responses ofCulex nigripalpus Theobald and Cx. quinquefasciatus Say with increased sampling time; (c) no difference(P > 0.55) in the average number of Ae. albopictus landing on the arm (8.6 ± 1.6 min–1) compared with theleg (9.2 ± 2.5 min–1) of the same human subject; (d) differences among day-to-day landing patterns for

the mosquito species we studied but measurable periodicity (P < 0.05) in each case when daily patternswere averaged for four or more diel periods; and (e) an effect on landing mosquito density from airtemperature (P < 0.0001) for Ae. albopictus and Cx. nigripalpus and dew point temperature (P < 0.0001) forCx. quinquefasciatus. Results from this study were used to develop a procedure for safely and accuratelymeasuring mosquito landing density on a human subject.

Published by Elsevier B.V.

. Introduction

Surveillance systems for adult mosquitoes rely on relative samp-ing methods (usually mechanical traps) for vector detection and tostimate density and other population parameters (Service, 1993;ilver, 2008; Southwood and Henderson, 2000). But the relation-hip between data from mechanical traps and the actual presence,bundance, and species composition of adult mosquitoes or theate of mosquito–host contact is unknown (Barnard et al., 2011).ast failures to detect mosquito presence (Peyton et al., 1999;prenger and Wuithiranyagool, 1986) or mosquito-borne disease

ctivity (Nash et al., 2001) thus may be a consequence of traditionaleliance on mechanical trapping methods for vector surveillance.

∗ Corresponding author at: USDA-ARS-CMAVE, 1600 SW 23rd Drive, Gainesville,L 32608, USA. Tel.: +1 352 374 5930; fax: +1 352 374 5870.

E-mail address: don.barnard@ars.usda.gov (D.R. Barnard).

ttp://dx.doi.org/10.1016/j.actatropica.2014.04.019001-706X/Published by Elsevier B.V.

In a chronology of adult mosquito sampling methods that relyon human bait, Silver (2008) writes that human bait catches orhuman landing catches are the single most useful method for col-lecting anthrophagic [mosquito] species. The technique is simpleto perform, requires no expensive equipment, and when properlyconducted precludes mosquito biting and disease agent transmis-sion (Carroll, 2008; Schmidt, 1989; Service, 1969, 1971; Ulloa et al.,1997).

Direct catches of landing mosquitoes have been used to esti-mate the human biting mosquito population density (Corbet andSmith, 1974; Dia et al., 2005; Ho et al., 1973; Malaithong et al., 2010;Moreno et al., 2007; Roberts et al., 2002; Silver, 2008;). The directcatch method eliminates bias associated with mosquito samplingtechniques that combine mechanical traps, semiochemicals, andvisual stimuli to capture adult mosquitoes (Barnard et al., 2011;

Davis et al., 1995; Sexton et al., 1986; Silver, 2008; Ulloa et al.,1997) or that intercept mosquitoes flying to a host (Achee et al.,2006; Costantini et al., 1998; Dia et al., 2005; Hiwat et al., 2011;Kweka and Mahande, 2009).

a Trop

toiJaftdnabmt

cmobsaptTwmt

ehqstCftsma

itnodtip

2

2

sqwiemisbfo

D.R. Barnard et al. / Act

Elements of the experimental design in common to studieshat use the direct catch method include (a) teams comprisingne human subject (exposed as the bait) and a second (protected)ndividual who collects the mosquitoes landing on the bait (Garrett-ones, 1964a,b; Haddow, 1954; Woke, 1962) and (b) a systematizedpproach to the selection of field sites for mosquito sampling andor scheduling observations (reviewed by Silver, 2008). Amonghe details lacking in these studies (and elsewhere) is a preciseefinition for the mosquito landing response and a constant tech-ique for its measurement. Thus, for purposes of sampling thedult mosquito population, it is unclear what constitutes landingy a mosquito, where on the subject’s body observation for landedosquitoes should be made, the duration of such observations, and

he area of skin to be examined.Categorizing the response of female mosquitoes that land and

ommence immediately to probe/penetrate host skin with theirouthparts is easy to do. However, female mosquitoes may alight

n the skin, touch it with the proboscis, take flight, then repeat thisehavior; land and move haphazardly across the skin; or remaintationary on the skin without initiating blood feeding (Grossmannd Pappas, 1991; Service, 1971). The latter behaviors may com-rise latent components of the mosquito landing response duringhe exploratory (foraging) phase of blood feeding (Clements, 1992).heir consideration as part of the landing response is importanthen seeking to estimate the rate of contact between humans andosquitoes, particularly when landing mosquito density is substi-

uted for the mosquito biting rate.The main objective of this study was to devise a technique that

nables reliable measurement of landing mosquito density on auman host. Our goal was the capacity to correctly classify anduickly count each female mosquito that lands on exposed hostkin. Given achievement of this objective, we used the techniqueo study temporal patterns of landing for Aedes albopictus (Skuse),ulex nigripalpus Theobald, and Cx. quinquefasciatus Say and otheracets of the mosquito landing response, including differences inhe numbers of females that make coincidental contact with thekin compared with those that land and remain, differences inosquito responses to the arm and the leg of a human subject,

nd the periodicity of landing activity.A second objective was to devise a procedure for compar-

ng the density of landing mosquitoes on a human subject withhe mosquito capture rate in mechanical traps. This procedure iseeded to evaluate estimates of mosquito population density basedn data from mechanical traps for comparability with data fromirect observation of landing mosquito density and to determinehe usefulness of capture data from mechanical traps for depict-ng rates of mosquito-human host contact in a vector surveillancerogram.

. Material and methods

.1. Mosquito rearing and test venue

Landing responses were observed for Ae. albopictus (Gainesvilletrain [1992]), Cx. nigripalpus (Vero Beach strain [1999]), and Cx.uinquefasciatus (Gainesville strain [1995]). Cohorts of each speciesere reared from the egg stage (Gerberg et al., 1994) outdoors

n two 6.4 m L × 4.3 m W × 2.8 m H (77.7 m3) aluminum-framednclosures each with fiberglass window screen sides, a sheet alu-inum roof, and a single door for entry and exit. Adults emerged

nto 0.125 m3 screened holding cages placed inside each enclo-

ure. Twenty-four hours before mosquito landing observationsegan, we released 1000–5000 4–9 day-old nulliparous femalesrom the holding cages into the enclosure (the number and agesf mosquitoes varied with time of year/environmental conditions).

ica 136 (2014) 58–67 59

Adult mosquitoes in the holding cages and the enclosure had con-tinuous access to 10% sucrose/water solution (via cotton wick).

2.2. Landing response defined: the exploratory phase of bloodfeeding

The classification of exploratory activity as part of the landingresponse is problematic. It requires a technique for differentiat-ing mosquitoes that make coincidental contact with the skin frommosquitoes that land to blood feed. To develop this technique, wetested the hypothesis that the female mosquito population initiallytouching human skin in a defined period of time is not significantlydifferent from the population remaining on the skin after initialcontact. The test comprised a series of 1 min-long observationsof mosquito landing in which the number of female mosquitoesthat made any contact with exposed skin was compared with thenumber of females doing so that made mouthpart-skin contact ormaintained continuous tarsal-skin contact for 5 s. Ae. albopictus wasselected for these observations. This species lands readily and inhigh numbers on human hosts during daytime when exact countsof landing mosquitoes can be made without supplementary light.We used a 5 s time limit to differentiate incidental contact fromexploratory (foraging) contact and classified the latter as the Landresponse (henceforth termed: Land, land, landed, or landing). The5 s time limit reflects a 1 s increase from the average 4 s exploratoryphase recorded for Aedes aegypti (L.) feeding on a human host withskin temperature of 36 ◦C (Grossman and Pappas, 1991).

To differentiate incidental contact from Land responses, welimited mosquito access to 75 cm2 of skin on the left forearm.The skin was exposed through a 5 cm W × 15 cm L window cutinto a flexible clear vinyl sleeve tailored for fit to the forearm ofthe test subject. The sleeve was placed on the left arm with thewindow centered in the medial forearm area and held in positionby fastening the ends together (in the lateral forearm area) withVelcroTM. During each 1 min observation session, the forearm wasextended forward from the body (with the window in an uprightposition) and the exposed skin observed for landing mosquitoes.Females that made incidental contact with the skin were countedand recorded. Landing females were captured using a mechanicalaspirator (Hausherr’s Machine Works, Toms River, NJ, USA) andtheir number recorded at the end of the test. None was allowedto bite. Observations were replicated 150 times each time usingthe same human subject (DRB).

2.3. Landing response defined: collection of mosquitoes fromdifferent body areas

In this study, we compared mosquito landing responses to thearm and the leg of the same human subject. A matched-pairs designwas used. Each test required two outdoor screened enclosures, eachwith approximately the same number of female mosquitoes. Tomeasure landing responses to the arm, we captured mosquitoesfrom 415 cm2 of exposed skin for 15 min using a mechanicalaspirator. The area of exposed skin was between two lines of cir-cumference on the left forearm, one at 3 cm below the elbow, theother at 3 cm above the wrist, with the boundary in each caseformed by protective clothing held in place with masking tape. Toobserve for landing mosquitoes, the forearm was extended in frontof the seated test subject (approximately 45 cm above ground level)and rotated slowly in a counterclockwise then clockwise fashion.

The same method was used to measure landing responses to theleg, in which case the area of exposed skin formed a 415 cm2 cylin-

der around the lower leg. Proximal and distal limits of the exposedskin were, respectively, 7 cm and 18 cm below the horizontalmidline of the left kneecap. The circumference of each exposedskin area was bounded by protective clothing held in place with

6 ta Trop

msa

elwi3otTosmfpu

2

qem2ltptstiloooeopw

c56aaCtp

0 D.R. Barnard et al. / Ac

asking tape. To observe for Land mosquitoes, the test subjecttood in the center of the enclosure with the feet shoulder-widthpart while inspecting the exposed leg skin for landed mosquitoes.

In the first half of each matched-pairs test (in one of the twonclosures) we collected landing mosquitoes from the arm (oreg). None was allowed to bite. Collection vials on the aspirator

ere changed every 5 min so the number of mosquitoes landingn 1–5 min, 6–10 min, and 11–15 min (collection intervals 1, 2, and, respectively) could be recorded separately. For the second halff each test (in the remaining enclosure), we collected mosquitoeshat landed on the leg (or arm) in collection intervals 1, 2, and 3.he order in which each enclosure was used and the limb (armr leg) that was observed for mosquito landing in the first enclo-ure were selected by coin toss before the start of each set ofatched-pair comparisons. Ae. albopictus was used in these studies

or the reasons stated in Section 2.2. Thirty-two sets of matched-air observations were made. The same human subject (DRB) wassed in all tests.

.4. Measuring daily patterns of mosquito landing

Landing responses of Ae. albopictus, Cx. nigripalpus, and Cx. quin-uefasciatus were observed in individual 24 h-long tests. Withinach test, we made eight separate 15 minute-long observations ofosquito landing. The astronomical time (U.S. Naval Observatory,

005, 2006, 2007, 2008, 2009) for each observation (translated toocal clock time) on the day of a test was determined by dividinghe 24 h (diel) day into eight Periods (Fig. 1). Periods 1–2–3–4 com-rised the time between sunset and sunrise and Periods 5–6–7–8,he time between sunrise and sunset. Period 1 embodied the tran-ition time from day to night (sunset) and Period 5 the transitionime from night to day (sunrise). Observations of mosquito land-ng in Period 1 were made at sunset and in Period 5 at sunrise. Theength of the day and night components of Periods 1 and 5 consistedf, respectively, one-eighth of all minutes in the photophase andne-eighth of all minutes in the scotophase at W 82◦20′, N 29◦40′

n the day of the test. The length of Periods 2–3–4 (night) wasach one-fourth of all minutes in the scotophase and the lengthf Periods 6–7–8 (day) was each one-fourth of all minutes in thehotophase. Mosquito landing responses in night and day Periodsere observed at the midpoint of the Period.

Females that landed and probed the skin or maintained tarsalontact with the exposed skin (415 cm2) on the left forearm for

s (Land) were collected in intervals 1 (minutes 1–5), 2 (minutes–10) and 3 (minutes 11–15) in each Period using a mechanicalspirator. None was allowed to bite. For nighttime observations,

6000 candle power VisorLIGHTTM (Model LT06, Donegan Opticalompany, Lenexa, KS) attached to the top of the aspirator and fit-ed with a red acetate lens cover (Grimstad and DeFoliart, 1974)rovided on-demand illumination.

Fig. 1. Schematic of mosquito collection t

ica 136 (2014) 58–67

Tests were replicated six times for Ae. albopictus, five times forCx. nigripalpus, and four times for Cx. quinquefasciatus. To the extentpossible (as limited by mosquito rearing failures), the order of the15 tests approximated a completely randomized design. The samehuman subject (DRB) was used in all tests.

2.5. Environmental parameters

We monitored air temperature (Ta) and dew point temper-ature (Td) coincidentally with observations of mosquito landing(Section 2.4). A Kestrel 4000 Pocket Weather Tracker (Nielsen-Kellerman, Boothwyn, PA, USA) suspended from the enclosureceiling 45 cm above ground level and 3 m from the test subject wasprogrammed to record and store the readings at 1 min intervals.Downloaded data were averaged separately according to collectioninterval in each Period in each test. Saturation deficit (S) was cal-culated according to Perret et al. (2000). The relationship betweenlanding responses and environmental parameters was evaluatedin regression models using Ta, Td, (Ta − Td), or S as the indepen-dent variable (SAS, 2003). A curvilinear response to each factorwas evaluated by addition of a quadratic coefficient to the linearmodel.

2.6. Data analysis

Statistical analyses of landing response data were made usingtabulation, Student’s t, analysis of variance, and linear regressionprocedures (SAS, 2003). Pre-planned comparisons of means weremade using the least significant difference (LSD) test at the 5% levelof significance. Prior to analysis, the total number (n) of landingmosquitoes captured in each 5 min collection interval (i), the totalnumber captured in each Period (np), and the total number cap-tured in a 24 h replicate were transformed to log10(n + 1) (Williams,1937).

We used a catch-per-unit-effort model (Southwood andHenderson, 2000) to evaluate landing mosquito density. Catch wasthe number of Land females collected by aspiration. Additional(implicit) inputs for the catch-per-unit parameter include (1) theduration of collection time (usually 1 min) (needed to calculatelanding rate) and (2) the area of skin exposed to landing mosquitoes(415 cm2 in most cases). Responses are expressed as the numberof landing mosquitoes per minute (min−1) per cm2 of skin area.Depending on data analysis requirements, we summed the numberof Land females observed in 5 min (for a single collection inter-val), in 15 min (for the three collection intervals in a Period), or in120 min (for all eight Periods in a 24 h replicate).

−1

The mean landing rate min for a 5 min collection interval wascalculated as:

log10

(n + 1

5

).

imes in a 24 h test (TWL = twilight).

D.R. Barnard et al. / Acta Tropica 136 (2014) 58–67 61

Table 1Mean number (±SE) of female Aedes albopictus landing on the forearm and on thelower leg of a human subject min−1.

Collection interval (min) Forearm Lower leg

1 (1–5) 3.8 (1.2) a 6.0 (2.9) a2 (6–10) 11.0 (3.3) a 9.3 (3.7) a

3 (11–15) 11.0 (2.4) a 12.3 (6.4) a1–3 (1–15) 8.6 (1.6) a 9.2 (2.5) a

DaL

i

l

aa

l

3

3

ntf

3d

actt

Table 3Correlation of the mosquito landing rate in a replicate and the number of Periods(in the same replicate) in which landing mosquito density is ≤1 female min−1 (lowdensity) or ≥10 females min−1 (high density).

Species Periods with a mean landing rate of

≤1 female min−1 ≥10 females min−1

r P r P

Aedes albopictus −0.874 0.023 +0.986 0.001Culex nigripalpus +0.363 0.547 −0.161 0.796

TR

L

ata transformed (log10(n + 1)) prior to statistical analysis. Row means (i.e., within collection interval) followed by same letter are not significantly different (P = 0.05,SD test).

The mean landing rate min−1 for all three (5 min) collectionntervals in one Period was calculated as:

og10

{∑3i=1 (n + 1)

5

}

nd for a 24 h replicate, the mean landing rate min−1 was calculateds:

og10

{∑8p=1

(np + 1

)120

}

. Results

.1. Landing response defined: exploratory phase of blood feeding

There was an 18% difference (t = 11.4, P < 0.0001) in the averageumber min-1 of Ae. albopictus females making coincidental con-act with the skin (9.11 ± 0.74) compared with the number of Landemales (7.42 ± 0.69).

.2. Landing response defined: collection of mosquitoes fromifferent body areas

Differences in the mean landing rate of Ae. albopictus on the fore-

rm and the leg of a human subject were not significant (Table 1). Inollection interval 1, the mean response to the leg was 57% higherhan to the arm. The pattern reversed in interval 2 (arm 18% higherhan leg) and again in interval 3 (leg 18% higher than arm). After

able 2ank order of Periods according to the percent of each (in brackets) as a component of th

Replicate Period

1(Sunset) 2 3 4

Aedes albopictus6 3 [20.6] 5 [8.6] 6 [4.1] 8 [1.2]8 3 [11.7] 6.5 [0] 6.5 [0] 6.5

10 2 [24.0] 5 [5.6] 6 [0.9] 7.5

12 4 [15.2] 5 [2.5] 6 [1.2] 8 [0.2]13 4 [14.5] 7 [4.9] 5 [9.5] 8 [3.4]14 4 [11.2] 5 [9.0] 6 [6.0] 7 [4.5]

Culex nigripalpus3 2 [37.9] 4 [4.5] 6 [2.3] 6 [2.3]4 1 [38.9] 5 [5.9] 7 [1.5] 6 [3.4]5 1 [44.0] 3 [10.0] 6 [4.8] 4 [8.9]11 1 [45.6] 3 [13.9] 7 [1.3] 8 [0.6]15 4 [6.7] 5 [3.0] 6 [0.8] 7 [0.5]

Culex quinquefasciatus1 1 [31.5] 5 [10.0] 4 [10.9] 6 [3.7]2 2 [13.9] 4 [12.9] 5 [5.1] 6 [3.7]7 1 [23.4] 2 [21.0] 3 [16.4] 7 [7.4]9 1 [32.2] 3 [12.8] 4 [12.2] 6 [9.1]

anding response data transformed (log10(n + 1)) before the calculation of percentages; 1

Culex quinquefasciatus −0.923 0.076 +0.987 0.012

Data transformed (log10(n + 1)) before statistical analysis.

15 min, the mean number of females collected from the leg and thearm differed by <0.7%.

3.3. Daily patterns of mosquito landing

Temporal trends of mosquito landing are difficult to general-ize because diel patterns of activity vary from one replicate to thenext (Fig. 2). To analyze for specific differences in the patterns foreach species, we converted the total number of mosquitoes cap-tured in each of Periods 1–8 (in a replicate) to a percentage of thetotal number of mosquitoes captured in the (same) replicate. Thepercentages were ranked in descending order within the replicateand the rank order identified for each Period compared among allreplicates. Results of the analysis are shown in Table 2.

Further inspection of daily landing pattern data suggested acorrelation of landing activity with mean landing rate and an asso-ciation between the density of landing mosquitoes in a replicateand the number of landing females in specific Periods. For exam-ple, daily landing patterns in replicates 8, 10, and 12 (Fig. 2) indicatethat mosquitoes landed in some Periods but not others. In con-trast, in replicates 13, 14, and 9, landing females were capturedin every Period. A test for correlation between the landing ratein a replicate and the number of Periods in the (same) replicatewith ≤1 female landings min−1 (i.e., low density) was significantand negative for Ae. albopictus (Table 3). The same test indicatedpositive and significant correlations for Ae. albopictus and Cx. quin-

quefasciatus when using Periods with ≥10 landing females min−1

(i.e., high density). These results indicate that when the overall den-sity of landing Ae. albopictus is low, landing activity takes place atcertain times within the diel period, but not others, whereas high

e total number of landed females collected in a replicate.

5(Sunrise) 6 7 8

7 [1.4] 4 [10.4] 1 [27.4] 2 [26.3][0] 6.5 [0] 4 [7.5] 2 [32.9] 1 [47.9][0.2] 7.5 [0.2] 3 [19.9] 1 [33.1] 4 [16.1]

7 [0.9] 3 [16.9] 1 [33.8] 2 [29.3] 6 [5.9] 2 [19.0] 1 [26.2] 3 [16.6] 8 [4.3] 2 [17.9] 1 [29.8] 3 [17.3]

8 [1.1] 6 [2.3] 3 [11.3] 1 [38.3] 4 [9.6] 8 [1.2] 3 [14.5] 2 [25.0] 7 [3.9] 8 [0.4] 2 [20.0] 5 [8.0] 5 [3.8] 4 [10.1] 6 [2.5] 2 [22.2] 8 [0] 3 [16.9] 2 [28.4] 1 [43.7]

8 [0] 7 [2.8] 3 [11.8] 2 [29.3] 8 [1.4] 7 [3.4] 3 [13.8] 1 [45.8] 4 [13.5] 8 [2.3] 5 [8.3] 6 [7.7] 7 [6.3] 8 [2.9] 5 [10.0] 2 [14.5]

= highest rank, 8 = lowest rank, same number = tied ranks.

62 D.R. Barnard et al. / Acta Tropica 136 (2014) 58–67

Fig. 2. Daily landing patterns of Aedes albopictus, Culex nigripalpus, and Culex quinquefasciatus according to collection interval ((�) collection interval 1 [0–5 min]; (�) collectioninterval 2 [6–10 min]; (�) collection interval 3 [11–15 min]).

a Trop

dcp

3

attgbltw8iPmroCi

npwaspC[re

3

aa(c−fi

TM

D(

D.R. Barnard et al. / Act

ensities of landing Ae. albopictus and Cx. quinquefasciatus are asso-iated with continuous (albeit phasic) activity throughout the dieleriod.

.4. Diel periodicity of mosquito landing

Each mosquito species manifested a recurring cycle of landingctivity (periodicity) when responses from multiple replicates (Sec-ion 3.3) were combined by Period and averaged (Table 4). Becausehe resulting diel patterns of landing activity were generally con-ruent among collection intervals, we compared the landing ratesy Period (and without regard to collection interval) to quantify

anding periodicity(s) for each mosquito species. For Ae. albopictus,he maximum (P = 0.05, LSD test) landing rate (F7,40 = 4.12, P = 0.002)as observed in Period 7 (29.5% of the total) followed by Period

(20% of the total) and the minimum rate (P = 0.05, LSD test)n Periods 4 (3.2%) and 5 (3.5%). For Cx. nigripalpus (F7,39 = 5.52,

= 0.001) and Cx. quinquefasciatus (F7,31 = 2.97, P = 0.022), the maxi-um landing rate was at sunset (Period 1 [34% and 28.2% of the total

esponse, respectively]), although 24.8% of Cx. nigripalpus landingccurred in Period 8, before sunset. The minimum landing rate forx. quinquefasciatus was in Period 6 (2.8%), following sunrise, but

n Periods 3–6 (15.6% of the total) for Cx. nigripalpus.The effect of collection interval on the mean landing rate was

ot significant in any Period (P > 0.05) for Ae. albopictus or Cx. nigri-alpus. In contrast, significantly (P = 0.05) fewer Cx. quinquefasciatusere captured in interval 1 in Periods 7 and 8 than in intervals 2

nd 3. Regression of the mean landing rate on collection intervalhowed effects of the latter factor to be significant for Cx. nigri-alpus (F1,118 = 5.28, P = 0.023; slope [±SE]): 0.0176 [0.0076]) andx. quinquefasciatus (F1,94 = 14.69, P = 0.0002; slope [±SE]: 0.03800.0099]) but not Ae. albopictus (P = 0.326). This means that landingates for Culex spp. increase in proportion to the cumulative timexpended collecting landed females.

.5. Environmental effects on mosquito landing

Three functional relationships between mosquito landing ratend environmental variables were identified. Landing rates for Ae.lbopictus and Cx. nigripalpus were influenced by air temperature

Ta). For the first species (F1,121 = 159.49, P < 0.001), the fitted modeloefficient for slope (±SE) was 0.0646 (0.0051) with intercept:0.7575 (0.1347). For Cx. nigripalpus (F2,107 = 10.24, P < 0.0001),tted linear (Ta) and quadratic (T2

a ) coefficients for slope were,

able 4ean number (±SE) of landing female mosquitoes min−1 on 415 cm2 of exposed left fore

Collection interval (min) Period

1(Sunset) 2 3 4

Aedes albopictus1 (1–5) 11.5 (5.0) abc 7.3 (5.0) bcd 4.0 (2.8) dc 2.9 (2 (6–10) 14.2 (4.2) ab 7.1 (4.8) bc 4.5 (3.1) c 3.0 (3 (11–15) 11.6 (3.9) abc 4.9 (2.1) bcd 5.9 (3.2) dc 2.5 (1–3 (1–15) 12.4 (4.3) ab 6.4 (3.9) bc 4.8 (2.7) bc 2.8 (

Culex nigripalpus1 (1–5) 6.2 (2.1) a 0.7 (0.2) cd 0.1 (0.1) d 0.2 (2 (6–10) 7.7 (2.4) a 1.7 (0.7) bc 0.6 (0.3) c 0.8 (3 (11–15) 7.0 (3.2) a 2.1 (0.7) abc 0.8 (0.5) c 1.6 (1–3 (1–15) 7.0 (2.4) a 1.5 (0.5) bc 0.5 (0.3) c 0.8 (

Culex quinquefasciatus1 (1–5) 7.9 (5.7) a 4.5 (2.3) a 3.0 (2.0) a 2.4 (2 (6–10) 18.5 (9.7) a 8.1 (3.3) ab 8.5 (3.6) ab 4.0 (3 (11–15) 18.9 (7.8) a 11.1 (3.) ab 8.4 (3.4) abc 5.7 (1–3 (1–15) 15.1 (7.7) a 7.9 (3.1) abc 6.6 (3.0) abc 4.0 (

ata transformed (log10(n + 1)) before statistical analysis. Row means (within species aP = 0.05, LSD test).

ica 136 (2014) 58–67 63

respectively, 0.1454 (0.5515) and −0.00232 (0.0011) with inter-cept of −1.7332 (0.7034). Landing rates for Cx. quinquefasciatus(F2,121 = 57.97, P < 0.0001) were influenced by dew point, with lin-ear (Td) and quadratic (T2

d) coefficients for slope of −0.0519 (0.0515)

and 0.0037 (0.0015), respectively, and intercept = 0.6336 (0.3941).Based on the fitted models, estimated air temperature thresh-

olds for the commencement of landing activity by Ae. albopictus andCx. nigripalpus were 11.7 ◦C and 16.0 ◦C, respectively. The estimateddew point temperature threshold for Cx. quinquefasciatus landingactivity was 7.1 ◦C.

4. Discussion

The components of the landing response that we studied firstwere those defined subjectively in the scientific literature andwhich we could duplicate only poorly or not at all in practice. Initialobservations were made to devise a simple yet explicit definitionfor the landing response.

4.1. The landing response defined

Not every landed female mosquito readily probes the skin andnot all females doing so ultimately engage in blood feeding (Khanet al., 1965). The absence of a measurable biological or behavioralsignal that foraging is about to end thus creates a decision-makingdilemma, particularly when seeking to classify a landed femaleas a potential vector. Even when the boundary between forag-ing and blood feeding is clearly defined, as with physical contactbetween mosquito mouthparts and the skin of a host, foragingtimes for Aedes species range between 3 and 11 s (Gillett, 1967;Grossman and Pappas, 1991; Service, 1971) and for Anophelesplumbeus (Stephens) and Mansonia [Coquillettidia] richardii Ficolbiare 16 and 31 s, respectively (Service, 1971).

Our selection of a 5 s time limit for the classification of Landmosquitoes was based on two requirements. The first was foran unequivocal decision-making procedure for classifying landedmosquitoes that would be easy to communicate to technical sup-port personnel. Using a 5 s time limit, the attentions of the collector(to each landed female mosquito) are obligated for a brief period

wherein either of two discrete mosquito behaviors leads to theLand classification: (1) continuous tarsal contact with the skin or(2) contact with skin by the female mosquito mouthparts. Beyondthis point, the ability to correctly classify a Land response requires

arm skin on a human subject.

5(Sunrise) 6 7 8

2.3) d 2.7 (1.6) d 11.1 (6.3) abcd 20.2 (9.3) a 12.9 (5.5) ab2.2) c 3.4 (2.3) c 15.3 (7.2) ab 27.5 (12.0) a 18.1 (6.8) a1.7) d 3.1 (2.1) d 18.3 (9.3) ab 30.0 (15.5) a 21.7 (8.2) a2.0) c 3.1 (1.9) c 14.9 (7.6) ab 26.0 (12.1) a 17.6 (6.7) a

0.1) d 0.3 (0.1) d 0.6 (0.4) cd 2.1 (0.8) bc 3.3 (1.0) ab0.4) c 1.0 (0.5) c 1.0 (0.6) c 3.8 (1.4) ab 6.1 (1.8) a1.3) c 1.0 (0.5) c 1.9 (1.2) bc 5.4 (2.3) ab 6.1 (1.9) a0.6) c 0.8 (0.4) c 1.1 (0.7) c 3.7 (1.4) ab 5.1 (1.5) a

1.8) a 3.6 (2.2) a 0.9 (0.5) a 1.4 (0.4) a 2.2 (0.3) a2.2) ab 3.4 (2.0) b 1.9 (1.0) b 5.8 (2.7) ab 9.7 (3.0) ab2.7) abc 4.2 (2.4) bc 1.6 (0.4) c 6.4 (2.7) abc 16.6 (5.2) a2.2) abc 3.7 (2.0) bc 1.5 (0.6) c 5.1 (2.1) abc 9.5 (2.7) ab

nd collection interval) followed by the same letter are not significantly different

6 ta Trop

ns

rolhfismLW(ssdwf

4

iacncal2

ofoslm(tscatwanAas

w(l(aat

aaooal

4 D.R. Barnard et al. / Ac

ominal training and practice by the collector and no secondarykills other than operation of the aspirator.

The second requirement was to eliminate experimental errorelated to the presence together (in time and space) of twor more human subjects (i.e., bait + collector). We believe thatanding female mosquitoes should be collected only by theuman host to which they are attracted. Doing so obviates con-

ounding of multiple-human-host-presence effects with variationsn the mosquito landing rate caused by location and time-of-ampling factors, changes in microclimate, and the periodicity ofosquito flight and host seeking activity (Garrett-Jones, 1964a;

indsay et al., 1993; Moore et al., 2007a,b; Shidrawi et al., 1974;oke, 1962). Ancillary benefits of the single collector technique

compared with the team method) include (1) reliable mea-urement of the mosquito landing rate on a specific humanubject and (2) the option to increase the number of indepen-ent measurements of landing rate (without additional resources)ith benefits to statistical analyses from increased degrees of

reedom.

.2. Which body area to sample?

Selection of the body area to sample for landing mosquitoess problematic. More Ae. albopictus landed on the leg than on therm in the first 5 min of sampling (interval 1) but this relationshiphanged in successive collection intervals and ultimately favoredeither limb. Continuous sampling of the same body area thus mayontradict the initial rationale used to select that body area. Thisppears to be the case for Cx. quinquefasciatus, the females of whichanded on the forearm at a lower rate in interval 1 than in intervals

and 3 (and only in Periods 7 and 8).Shirai et al. (2002) showed that body position affects the choice

f landing site by Ae. albopictus with preference given to the feet,ace, and hands (in that order) but with landing rates reducedn the feet when test subjects (wearing only shorts) were in aupine rather than upright position. Biting site selection by Anophe-es atroparvus (van Thiel) and An. gambiae s.s. (Giles) on a naked

otionless human host is for the head and feet, respectivelyDeJong and Knols, 1995), with correlations in both cases betweenhe biting site and (combinations of) skin temperature and eccrineweat gland density. For some species, such as Eretmapoditeshrysogaster Graham, female mosquitoes land in relation to heightbove ground level (6–18 in.) rather than in response to some fea-ure of the external anatomy (Haddow, 1954). Cultural customs,hich dictate clothing styles and bodily accouterments, may also

ffect where mosquitoes land on the body. This relationship wasoted by Garrett-Jones (1964b) who compared “sleepers in tropicalfrica [that] are typically muffled to the eyes, to prevent annoy-nce from mosquitoes” with “children [in Bengal] of various agesleeping naked in the warm season”.

Clothing can affect the landing rate of biting flies in yet otherays. Kettle (1969) showed that landing rates by biting midges

Ceratopogonidae) were lower on lightly clad test subjects (arms,egs, and head exposed) in Jamaica compared with heavily clad onesarms and legs exposed, head with balaclava). Differences werettributed to a combination of factors that included the locationnd amount of exposed skin area and radiant energy signatures ofhe various test subjects.

Experimental evidence favoring selection of one body area overnother for observation of landing mosquitoes is equivocal. Studiesre needed on a species-by-species basis to determine the effects

f geographic location, duration of sampling time, mosquito peri-dicity (and other behaviors), body position, clothing and bodilyccouterments, culture, and other lifestyle factors on the mosquitoanding response. Lacking this knowledge, we believe that

ica 136 (2014) 58–67

consistent use of the same body area for observation of landingmosquitoes is a technically defensible sampling strategy.

4.3. Landing mosquito density

The number of mosquito landing responses per unit oftime defines the landing rate but not the density (D) of land-ing mosquitoes. Estimation of D requires measurement of themosquitoes (n) landing on a known area of exposed skin (a) ina unit of time and is calculated as D = (n/a)/time (Southwood andHenderson, 2000). Most reports of mosquito landing in the lit-erature do not specify a, although this parameter varies amonghumans by as much as 45% depending on body mass index andthe anatomical landmarks used for measurement (Haycock et al.,1978; Livingstone and Lee, 2000). To facilitate consistent and reli-able estimation of landing mosquito density and a practical contextfor the comparison of landing response data, selection of the sameanatomical feature and amount of exposed skin area observed forlanding mosquitoes should be standard practice. In cases wherethe measurement of a may not be stipulated (e.g., mosquito repel-lent field studies), it is important to verify that the area of exposedskin (and thus the dose of repellent applied) is the same for all testsubjects.

4.4. Temporal patterns of mosquito landing and environmentalinfluences

Mosquito landing on the skin is the ultimate response to acomplex of stimuli that includes host and environmental effectsand endogenous cycles in the mosquito population. It is typi-cally depicted as a functional response to time (over a 24 h [diel]period), even though the numeric process used for this purpose mayobscure effects due to the frequency and timing of sampling andfor interaction(s) between environmental variables and cycles ofadult mosquito activity. Some of these sampling concerns were dis-cussed by Haddow (1954) and Garrett-Jones (1964a). Our study wasan attempt to add to their findings and to devise methodologicalimprovements in the sampling process, such as compartmental-ization of the diel period into discrete yet generalize-able temporalunits (Periods). This procedure enables consistent and reliable mea-surement of landing rate, landing mosquito density, daily patternsof landing activity, and diel landing periodicities and the com-parison of landing response data obtained from sympatric andallopatric mosquito populations (of the same or mixed species) atdifferent points in time.

Many abiotic factors in the environment (e.g., light, tempera-ture, humidity) evidence 24 h periodicity. Some of these factorscue circadian activity in mosquitoes that manifests as cycles oflanding and biting (Haddow, 1954). The usefulness of these datafor local forecasts of mosquito landing activity depends on whichparameters are measured. In addition, the high variances that oftenaccompany landing response data can mask the effects of micro-climate and other environmental co-variables that influence flightactivity (e.g., wind speed, moon phase, precipitation, and change inambient light level) (Bidlingmayer, 1971, 1974; Bidlingmayer andHem, 1979). Given this context, a practical approach for using envi-ronmental data to anticipate mosquito landing activity is to identifythe conditions that restrain activity but, when no longer present,release mosquitoes to fly, land, and bite (Haddow, 1954). One suchrestraining condition is wind. Speeds above 0.25 m s−1 reduce flightactivity by 75% for many mosquito species (Bidlingmayer et al.,1985). Air temperature (Ta) is a second restraining condition. In our

study, temperatures below 15 ◦C were not suited to the collectionof landing Ae. albopictus and Cx. nigripalpus (Fig. 3). And while themean landing rate for Ae. albopictus increases linearly with increas-ing air temperature (through 35 ◦C), a curvilinear response to this

a Trop

flboselaao

5

bibrpm

(

(

(

(

(

uomtd

D.R. Barnard et al. / Act

actor by Cx. nigripalpus indicates that deceleration (slowing) of theanding rate commences beyond 30 ◦C. Dew point temperature maye a third restraining condition. The curvilinear effect of dew pointn the landing rate of Cx. quinquefasciatus indicates this species isensitive to water vapor level in the air and that landing rate accel-rates (increases) as dew point temperatures increase through ateast 25 ◦C (note: low dew point temperatures correlate with lowir temperatures, whereas a dew point of 27 ◦C at 32 ◦C air temper-ture [65% relative humidity] is considered severely high in termsf human comfort [Lawrence, 2005]).

. Conclusions

Direct counts of landed mosquitoes on a human subject cane used to determine landing adult mosquito density; however,

mportant details for measuring mosquito landing activity cannote found in the scientific literature. In this study, we identified andesolved five landing response-related issues that limit the com-arability of mosquito landing rate data and estimates of landingosquito density:

1) We defined landing to mean (a) continuous contact betweenone or more mosquito tarsi and the host skin for 5 s and/or(b) landed females that immediately probe the skin with theirmouthparts.

2) We defined the mosquito landing rate as the number ofmosquito landings per unit of time and the density of landingmosquitoes as the landing rate per unit of exposed skin area. Therelation between the duration of sampling time and mosquitolanding rate and the density of the landing mosquito populationmay be species-dependent.

3) In studies of preference for specific body areas by landingmosquitoes, we found no measurable difference in the land-ing rates of Ae. albopictus on the arm or leg of a human subject.The result contraindicates selection of either limb as the onepreferred for landing by Ae. albopictus mosquitoes. It also under-scores the significance of body area as a classification variablein landing rate studies and the requirement for knowledge-based species-relevant criteria when pre-selecting a body areato observe for landing mosquitoes.

4) We found temporal trends of mosquito landing difficult togeneralize because daily patterns of landing activity vary inresponse to the sampling venue and the overall density oflanding mosquitoes. However, when daily landing patterns areaveraged, each mosquito species manifests a recurring cycle oflanding activity (periodicity) that allows measurement of thetemporal relation between landing responses and the differen-tiation of times of maximum and minimum landing within thediel period.

5) The functional response of landing mosquitoes to environmen-tal conditions ranged from linear to curvilinear. Landing Ae.albopictus were not observed below 15 ◦C but landing ratesfor this species increased linearly in response to increasing airtemperature through 35 ◦C. In contrast, a curvilinear responseto air temperature for Cx. nigripalpus indicated deceleration ofthe landing rate beyond 30 ◦C. Increasing dew point temper-atures correlated with acceleration of the landing rate by Cx.quinquefasciatus (through 25 ◦C).

The protocol we devised to sample the landing mosquito pop-lation is based on division of the diel period into discrete units

f astronomical time. These units (Periods) enable observation ofosquito landing rates using local clock times indexed to pho-

operiod. The result is a practical framework for the analysis ofaily landing patterns, landing periodicities, and other aspects of

ica 136 (2014) 58–67 65

the mosquito landing response. The protocol is simple and easy tounderstand and can be generalized to the use of more than onehuman subject at the same time in different locations or at differ-ent times in different locations. Data obtained using this samplingprotocol enable comparison of landing mosquito densities forallopatric mosquito populations as well as the infra-seasonal com-parison of landing mosquito densities for multivoltine mosquitospecies from mixed- or single-species populations.

Acknowledgement

We thank Gregory J. Knue for technical support during thesestudies.

Appendix.

A procedure for measuring the density of landing mosquitoeson a human subject.

The goal of our study was a method that could be used by anindividual to safely, efficiently, and accurately self-measure thedensity of adult mosquitoes landing on a human host. The pro-cedure devised for this purpose has three phases: Preparation,Protection, and Collection.

A.1. Preparation

The experimental protocol should designate the date and phys-ical location(s) selected for sampling, the number of samples to becollected, the astronomical start and stop times for each sample(translated to local time), a unique and easily deciphered monikerfor each sample, and a field-reference data sheet that contains alldetails of the sampling protocol.

A storage system for samples that prevents escape of or damageto captured specimens before and during transport to the labora-tory and that precludes misidentification or mixing of samples atany time is critical. In practice, after a sample is collected, we sealthe pre-labeled and capped collection vial inside a 0.94 l clear plas-tic freezer bag (Hefty® One-zip), which is placed in a portable coolerchilled with BlueIce®. These samples are stored in the laboratoryat −20◦ C until needed.

Supplies and equipment acquired in advance of field workshould include replacement aspirators and aspirator parts, col-lection vials and vial caps; a headlamp, flashlights, and a digitalelectronic clock/timer (with on-demand illumination for night-time use); batteries, plastic storage bags, nylon ties, and rubberbands; marking pens, masking and plastic tape, writing instru-ments, and replacement head nets and gloves. Proper functioningof all mechanical equipment should be verified before traveling tothe field.

A.2. Protection

Direct access to skin on the arms, legs, and trunk areas of thecollector by landing mosquitoes is prevented by clothing (Fig. A.1).A white cotton (100%) coverall (Red Cap®) sized large enough to beworn loosely over street clothing is the main garment used for thispurpose. The coverall features a concealed stainless front buttonclosure, breast, hip, and back pockets, and a one piece collar. The hippockets are sewn closed to prevent mosquito entry and an elasticband is sewn to the end of each coverall arm to ensure a tight fitto the wrist. Feet are protected by ankle-high (25 cm) leather boots

and each coverall leg is either taped to or tucked inside the top of theboot. Ambidextrous heavy weight cotton/polyester jersey gloveswith polyvinylchloride (PVC) grip dots on both sides (for dexterity)protect the hands. On each arm, the sleeve of the coverall overlaps

6 ta Tropica 136 (2014) 58–67

thwpeachn

A

semcwe

F

Table A.1Mean density of landing female mosquitoes captured min−1 using three collectionmethods in a Florida floodplain swamp.

Mosquito species Collection methoda

LMD LT ST

Anopheles quadrimaculatus Say 0.13 a 0.03 b 0.04 bCulex nigripalpus Theobald 0.13 a 0.05 b 0.08 aCulex salinarius Coquillett 0.52 a 0.08 b 0.05 bCulex erraticus (Dyar and Knab) 0.19 a 0.05 b 0.04 bCuliseta melanura (Coquillett) 0a 0.02 b 0.02 bMansonia dyari (Belkib, Heinemann and Page) 0.04 a 0.019 a 0bOchlerotatus infirmatus (Dyar and Knab) 1.58 a 2.11 b 0.34 cPsorophora ferox (von Humboldt) 0.54 a 0.20 b 0.09 b

a On 400 cm2 of exposed skin; LMD = landing mosquito density; LT = light trap

6 D.R. Barnard et al. / Ac

he proximal end of the glove by 2–3 cm with both clothing itemseld in position by 5 cm-wide masking tape wrapped around therist. The head and neck are protected from mosquito bite by a netlaced over a straw hat. The netting is held away from the head,ars, and neck by the hat brim (10 cm-wide) and can be drapedround the shoulders or tucked loosely inside the coverall at theollar. The coverall collar can be buttoned and turned up (or theead net-collar junction closed with 5 cm-wide masking tape) ifeeded for additional protection from landing mosquitoes.

.3. Collection

We collect landing mosquitoes from a 200 cm2 area of exposedkin on each thigh. These areas are normally protected by cov-rall material but can be uncovered as needed for exposure to

osquitoes. The cover is an 11 cm wide × 21 cm-long rectangle of

overall fabric placed over a 10 cm wide × 20 cm long (200 cm2)indow located in the mid-center of the thigh area on the front of

ach coverall leg (Fig. A.1). The hinge side of the cover is sewn to the

ig. A.1. Protective clothing used to observe mosquito landing responses in the field.

(miniature [CDC] with CO2); ST = suction trap. Data transformed (log10(n + 1)) beforestatistical analysis. Row means followed by the same letter are not significantlydifferent (P = 0.05, LSD test).

lateral edge of the window on each coverall leg, whereas the upper,lower, and medial edges of the cover are attached to the coverallleg using Velcro®. Pulling the cover to detach it from the coverallleg exposes the skin beneath.

We collect landing mosquitoes from a seated human subjectposition. A portable fabric camp chair (with cup holders in the armrests) works well for this purpose because it enables observationand collection of mosquitoes from exposed thigh skin with minimalmovement by the collector and without disturbance of the landedmosquitoes when using the aspirator. Cup holders can be used forstorage of items needed immediately because of equipment failureor other problems. An electronic timer (suspended from the armrest of the chair) with chime function is used to signal the stoptime for sampling.

A.4. Field testing of the collection procedure

Initial field tests of the collection procedure were made in2009–2010. Our objective was to measure the landing mosquitodensity min-1 (LMD) on 400 cm2 of exposed skin for comparisonwith the mosquito capture rate using a miniature light trap (LT)(with CO2) and a suction trap (ST). Observations were made in a(Sumter County) Florida floodplain swamp using three study plots50–75 m apart and each invisible to the next. Adult mosquitoeswere collected simultaneously and continuously in each plot for30 min before and 30 min after sunset. Separate tests were madeon six dates each time with LMD, LT, or ST assigned randomly to astudy plot. Results of these tests are shown in Table A.1.

References

Achee, N.L., Grieco, J.P., Rejmankova, E., Andre, R.G., Vanzie, E., Polanco, J., Briceno,I., King, R., Roberts, D.R., 2006. Biting patterns and seasonal densities of Anophe-les mosquitoes in the Cayo District, Belize, Central America with emphasis onAnopheles darlingi. J. Vector Ecol. 31, 45–57.

Barnard, D.R., Knue, G.J., Dickerson, C.Z., Bernier, U.R., Kline, D.L., 2011. Relation-ship between mosquito (Diptera: Culicidae) landing rates on a human subjectand numbers captured using CO2-baited light traps. Bull. Entomol. Res. 101,277–285.

Bidlingmayer, W.L., 1971. Mosquito flight paths in relation to the environment. 1.Illumination levels, orientation, and resting areas. Ann. Entomol. Soc. Am. 64,1121–1131.

Bidlingmayer, W.L., 1974. The influence of environmental factors and physiologicalstage on flight patterns of mosquitoes taken in the vehicle aspirator and truck,suction, bait and New Jersey light traps. J. Med. Entomol. 11, 119–146.

Bidlingmayer, W.L., Hem, D.G., 1979. Mosquito (Diptera: Culicidae) flight behaviornear conspicuous objects. Bull. Entomol. Res. 69, 691–700.

Bidlingmayer, W.L., Evans, D.G., Hansen, C.H., 1985. Preliminary study of the effectsof wind velocities and wind shadows upon suction trap catches of mosquitoes

(Diptera: Culicidae). J. Med. Entomol. 22, 295–302.

Carroll, S.P., 2008. Prolonged efficacy of IR3535 repellents against mosquitoes andblacklegged ticks in North America. J. Med. Entomol. 45, 706–714.

Clements, A.N., 1992. The Biology of Mosquitoes, vol. 1: Development, Nutrition, andReproduction. Chapman & Hall, London.

a Trop

C

C

D

D

D

G

G

G

G

G

G

H

H

H

H

K

K

K

L

L

L

M

M

D.R. Barnard et al. / Act

orbet, P.S., Smith, S.M., 1974. Diel periodicities of landing of nulliparous and parousAedes aegypti (L.) at Dar es Salaam, Tanzania (Diptera: Culicidae). Bull. Entomol.Res. 64, 111–121.

ostantini, C., Sagnon, N., Della Torre, A., Diallo, M., Brady, J., Gibson, G., Coluzzi,M., 1998. Odor-mediated host preferences of West African mosquitoes withparticular reference to malaria vectors. Am. J. Trop. Med. Hyg. 58, 56–63.

avis, J.R., Hall, T., Chee, E.M., Majala, A., Minjas, J., Shiff, C.J., 1995. Comparisonof sampling anopheline mosquitoes by light-trap and human-bait collectionsindoors at Bagamoyo, Tanzania. Med. Vet. Entomol. 9, 249–255.

eJong, R., Knols, B.G., 1995. Selection of biting sites on man by two malaria mosquitospecies. Experentia 51, 80–84.

ia, I., Diallo, D., Duchemin, J., Ba, Y., Konate, L., Costantini, C., Diallo, M., 2005. Com-parisons of human-landing catches and odor-baited entry traps for samplingmalaria vectors in Senegal. J. Med. Entomol. 42, 104–109.

arrett-Jones, C., 1964a. A method for estimating the man-biting rate. Unpublisheddocument WHO/Mal/450.

arrett-Jones, C., 1964b. Prognosis for interruption of malaria transmissionthrough assessment of the mosquito’s vectorial capacity. Nature 204,1173–1175.

erberg, E.J., Barnard, D.R., Ward, R.A., 1994. Manual for Mosquito Rearing andExperimental Techniques. American Mosquito Control Association Bulletin No.5, revised. Allen Press, Lawrence.

illett, J.D., 1967. Natural selection and feeding speed in a blood-sucking insect. Proc.R. Soc. Lond. (B) 167, 316–329.

rimstad, P.R., DeFoliart, G.R., 1974. Nectar sources of Wisconsin mosquitoes. J. Med.Entomol. 11, 331–334.

rossman, G.L., Pappas, L.G., 1991. Human skin temperature and mosquito (Diptera:Culicidae) blood feeding rate. J. Med. Entomol. 28, 456–460.

addow, A.J., 1954. Studies of the biting-habits of African mosquitoes. An appraisalof methods employed, with special reference to the twenty-four-hour catch.Bull. Entomol. Res. 45, 199–242.

aycock, G.B., Chir, B., Schwartz, G.J., Wisotsky, D.H., 1978. Geometric method formeasuring body surface area: a height–weight formula validated in infants,children, and adults. J. Pediatr. 93, 62–66.

iwat, H., De Rijk, M., Andriessen, R., Koenraadt, C.J.M., Takken, W., 2011. Evaluationof methods for sampling the malaria vector Anopheles darlingi (Diptera: Culici-dae) in Suriname and the relation with its biting behavior. J. Med. Entomol. 48,1039–1046.

o, B.C., Chan, Y.C., Chan, K.L., 1973. Field and laboratory observations on landingand biting periodicities of Aedes albopictus (Skuse). Southeast Asian J. Trop. Med.Public Health 4, 238–244.

ettle, D.S., 1969. The biting habits of Culicoides furens (Poey) and C. barbosai With& Blanton. I. The 24-h cycle, with a note on differences between collectors. Bull.Entomol. Res. 59, 21–31.

han, A.A., Maibach, H.I., Howard, I., Strauss, W.G., Fenley, W.R., 1965. Screeninghumans for degrees of attractiveness to mosquitoes. J. Econ. Entomol. 58,694–697.

weka, E.J., Mahande, A.M., 2009. Comparative evaluation of four mosquitoes samp-ling methods in rice irrigation schemes of lower Moshi, northern Tanzania. Mal.J. 8, 149–153.

awrence, M.G., 2005. The relationship between relative humidity and the dewpoint temperature in moist air: a simple conversion and applications. Bull. Am.Meteorol. Soc. 86, 225–233.

indsay, S.W., Adiamah, J.H., Miller, J.E., Pleass, R.J., Armstrong, J.R.M., 1993. Variationin attractiveness of human subjects to malaria mosquitoes (Diptera: Culicidae)in The Gambia. J. Med. Entomol. 30, 368–373.

ivingstone, E.H., Lee, S., 2000. Percentage of burned body surface area determina-tion in obese and nonobese patients. J. Surg. Res. 91, 106–110.

alaithong, N., Polsomboon, S., Poolprasert, P., Parbaripai, A., Bangs, M.J.,Suwonkerd, W., Pothikasikorn, J., Akratanakul, P., Chareonviriyaphap, T., 2010.Human-landing patterns of Anopheles dirus sensu lato (Diptera: Culicidae) in

experimental huts treated with DDT or deltamethrin. J. Med. Entomol. 47,823–832.

oore, S.J., Davies, C.R., Hill, N., Cameron, M.M., 2007a. Are mosquitoes diverted fromrepellent-using individuals to non-users? Results of a field study in Bolivia. Trop.Med. Int. Health 12, 532–539.

ica 136 (2014) 58–67 67

Moore, S.J., Darling, S.T., Sihuincha, M., Padilla, N., Devine, G.J., 2007b. A low-costrepellent for malaria vectors in the Americas; results of two field trials inGuatemala and Peru. Mal. J. 6, 101.

Moreno, J.E., Rubio-Palis, Y., Páez, E., Pérez, E., Sánchez, V., 2007. Abundance, bitingbehavior, and parous rate of anopheline mosquito species in relation to malariaincidence in gold-mining areas of southern Venezuela. Med. Vet. Entomol. 21,339–349.

Nash, D., Mostashari, F., Fine, A., Miller, J., O’Leary, D., Murray, K., Huang, A., Rosen-berg, A., Greenberg, A., Sherman, M., Wong, S., Campbell, G.L., Roehrig, J.T.,Gubler, D.J., Shieh, W., Zaki, S., Smith, P., Layton, M., 2001. The outbreak of WestNile Virus infection in the New York City area in 1999. N. Engl. J. Med. 344,1807–1814.

Perret, J.L., Guigoz, E., Rais, O., Gern, L., 2000. Influence of saturation deficit andtemperature on Ixodes ricinus tick questing activity in a Lyme borreliosis-endemicarea (Switzerland). Parasitol. Res. 86, 554–557.

Peyton, E.L., Campbell, S.R., Candeletti, T.M., Romanowski, M., Crans, W.J., 1999. Aedes(Finlaya) japonicus japonicus (Theobald), a new introduction into the UnitedStates. J. Am. Mosq. Control Assoc. 15, 238–241.

Roberts, D.R., Manguin, S., Rejmankova, E., Andre, R., Harbach, R.E., Vanzie, E., Hakre,S., Polanco, J., 2002. Spatial distribution of adult Anopheles darlingi and Anophelesalbimanus in relation to riparian habitats in Belize, Central America. J. VectorEcol. 27, 21–30.

SAS, 2003. Statistical Analysis System 9.1 for Windows. SAS Institute, Inc., Cary.Schmidt, R.F., 1989. Landing rates and bite counts for nuisance evaluation. Proc. N.J.

Mosq. Control Assoc. 76, 34–38.Service, M.W., 1969. Observations of the ecology of some British mosquitoes. Bull.

Entomol. Res. 59, 161–194.Service, M.W., 1971. Feeding behavior and host preferences of British mosquitoes.

Bull. Entomol. Res. 60, 653–661.Service, M.W., 1993. Mosquito Ecology. Field Sampling Methods, second ed. Elsevier

Applied Science, London.Sexton, J.D., Hobbs, J.H., St Jean, Y., Jacques, J.R., 1986. Comparison of an experimen-

tal updraft ultraviolet light trap with the CDC miniature light trap and bitingcollections in sampling for Anopheles albimanus in Haiti. J. Am. Mosq. ControlAssoc. 2, 168–173.

Shidrawi, G.R., Boulzaguet, J.R., Ashkar, T.S., Brøgger, S., 1974. Night bait collection:the variation between persons used as collector-baits. WHO Technical Note 17,MPD/TN/74.1, pp 9–18 (unpublished).

Shirai, Y., Funada, H., Kamimura, K., Seki, T., Morohashi, M., 2002. Landing sites onthe human body preferred by Aedes albopictus. J. Am. Mosq. Control Assoc. 18,97–99.

Silver, J.B., 2008. Mosquito Ecology. Field Sampling Methods, third ed. Springer,Dordrecht.

Southwood, T.R.E., Henderson, P.A., 2000. Ecological Methods, third ed. BlackwellScience, Oxford.

Sprenger, D., Wuithiranyagool, T., 1986. The discovery and distribution of Aedesalbopictus in Harris County, Texas. J. Am. Mosq. Control Assoc. 2, 217–219.

Ulloa, A., Rodríguez, M.H., Rodríguez, A.D., Roberts, D.R., 1997. A comparison of twocollection methods for estimating abundance and parity of Anopheles albimanusin breeding sites and villages of southern Mexico. J. Am. Mosq. Control Assoc.13, 238–244.

U.S. Naval Observatory, 2005. The Nautical Almanac. U.S. Government PrintingOffice, Washington, DC.

U.S. Naval Observatory, 2006. The Nautical Almanac. U.S. Government PrintingOffice, Washington, DC.

U.S. Naval Observatory, 2007. The Nautical Almanac. U.S. Government PrintingOffice, Washington, DC.

U.S. Naval Observatory, 2008. The Nautical Almanac. U.S. Government PrintingOffice, Washington, DC.

U.S. Naval Observatory, 2009. The Nautical Almanac. U.S. Government PrintingOffice, Washington, DC.

Williams, C.B., 1937. The use of logarithms in the interpretation of certain entomo-logical problems. Ann. Appl. Biol. 24, 404–414.

Woke, P.A., 1962. Observations on differential attraction among humans for Man-sonia nigricans and Anopheles albimanus (Diptera: Culicidae). Proc. N.J. Mosq.Exterm. Assoc. 49, 173–181.