Behavioral Ecology Advance Access originally published online on June 12, 2007
Behavioral Ecology 2007 18(4):750-757; doi:10.1093/beheco/arm043
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Avian defensive behavior and blood-feeding success of the West Nile vector mosquito, Culex pipiens
Department of Entomology, 3138 Comstock Hall, Cornell University, Ithaca NY 14850, USA
Address correspondence to L.C. Harrington. E-mail: lch27{at}cornell.edu.
Received 10 February 2007; revised 26 March 2007; accepted 1 April 2007.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Avian defensive behavior against host-seeking arthropods influences transmission of vector-borne pathogens by affecting mosquito biting rate, either by preventing vector–host contact or by increasing the rate of multiple host feeding. We exposed house sparrows (Passer domesticus L.) and chickens (Gallus gallus domesticus L.) to Culex pipiens pipiens L. overnight in a large observation cage and measured avian defensive behavior rates and mosquito blood-feeding success. Both bird species exhibited a range of defensive behaviors, 90% of which were foot stomps, head movements, and wing shakes. Total behavior rates increased proportionately with mosquito density in both species, increased after the first hour of mosquito exposure, and decreased as individual birds were exposed to mosquitoes multiple times. Mosquito blood-feeding success on house sparrows was high overall (82 ± 5%) and independent of behavior rates. Blood-feeding success on chicks was lower (58 ± 5%) and negatively correlated with defensive behavior rate after the first hour of mosquito exposure. Results revealed a higher percentage of partial blood meals on chicks (18 ± 3% of all blood meals on chicks) than on house sparrows (4.9 ± 3%). Birds of both species ate an average of 9.4 ± 1.2% of mosquitoes, and this percentage was positively correlated with defensive behavior. High mosquito feeding success on house sparrows supports its role as a potential amplifying host of West Nile virus.
Key words: chicken, Culex p. pipiens, defensive behavior, house sparrow, mosquito, West Nile virus.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Host defensive behavior against mosquitoes may either facilitate or hinder West Nile virus (WNV) transmission by infected vectors. Virus transmission is facilitated if an infected mosquito has to feed on multiple hosts in order to get a blood meal and hindered if a mosquito fails to feed because hosts are too defensive. Previous studies have shown that mosquito feeding success is reduced on more defensive vertebrates, including pigeons (Columba livia) (Blackmore and Dow 1958
), ciconiiform birds (Edman and Kale 1971; Webber and Edman 1972), house sparrows (Passer domesticus), Carolina wrens (Thryothorus ludovicianus), eastern towhees (Papilo erythrophthalmus), young opossums (Didelphis marsupialis) (Edman et al. 1974), rabbits (Oryctolagus cuniculus) (Klowden and Lea 1978; Waage and Nondo 1982), eastern gray squirrels (Sciurus carolinensis) (Edman et al. 1974; Walker and Edman 1986), eastern chipmunks (Tamias striatus) (Cully et al. 1991), cotton mice (Peromyscus gossypinus), wood rats (Neotoma spp.), cotton rats (Sigmodon hispidus) (Edman et al. 1974), and Homo sapiens (Walker and Edman 1985).
In some of the above studies, the ability of vertebrates to defend themselves was experimentally modified by a number of methods, such as physical restraint (Edman and Kale 1971; Edman et al. 1974; Waage and Nondo 1982; Walker and Edman 1986) or anesthesia (Cully et al. 1991). Other studies compared different host species with different natural extremes of defensive behavior (Kale et al. 1972) or a single host species during active and inactive periods of the day (Day and Edman 1984). Variation between "naturally behaving" individuals has been analyzed with rabbits (Waage and Nondo 1982) and Japanese quail (Coturnix japonica) (Anderson and Brust 1996), but not with passerine birds. Furthermore, although rabbit defensive behavior trials lasted 4 h, avian defensive behavior has been observed no longer than 1 or 2 h, even though avidity of host-seeking mosquitoes can persist for longer periods of time.
WNV can be readily transmitted by Culex mosquitoes (Goddard et al. 2002; Turell et al. 2005). Two primary Culex West Nile mosquito vectors in the northeastern United States, Culex pipiens pipiens and Culex restuans, feed largely on passerine birds (Apperson et al. 2002, 2004; Molaei et al. 2006), and susceptible passerine birds are thought to serve as amplifying hosts for WNV (Komar et al. 2003; Godsey et al. 2005; Reisen et al. 2005; Kilpatrick et al. 2006). This system may represent the primary enzootic cycle of WNV, yet little is understood about the ecology of mosquito–bird interactions in nature and how these dynamics influence WNV transmission.
We evaluated anti-mosquito defensive behavior of house sparrows (P. domesticus), a major amplifying host of WNV (Komar et al. 2003) and chickens (Gallus domesticus), which are used as sentinel animals for WNV surveillance, against a primary enzootic WNV vector mosquito (C. pipiens pipiens L.) to test the hypothesis that mosquitoes have lower feeding success on more defensive birds. After exposing individual birds to host-seeking mosquitoes overnight, we analyzed video-recorded defensive behavior up to 6 h after initial exposure to evaluate changes in avian defensive behavior rates over time. We also investigated the role of intra- and interbird variation on defensive behavior in explaining variation in mosquito blood-feeding success. We compared the relationship between mosquito density and host anti-mosquito behavior rates at hourly intervals. In addition, we categorized specific behaviors and evaluated their roles in reducing mosquito feeding success. These studies are the first to explore the behavioral impact of house sparrow reservoirs on vector-borne infections such as WNV and the efficacy of chickens as WNV sentinels. We also provide biting rate estimates for vectoral capacity models (Garrett-Jones 1964).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Mosquitoes
We hatched C. pipiens pipiens larvae, from a colony established with wild-captured mosquitoes in New York State in 2003 and raised the larvae at 25 °C in plastic trays with approximately 200 larvae per liter of water containing 60 ml of diet slurry (1:2:1 ratio of ground fish food [TetraMin Blacksburg, VA]: rabbit pellets [L/M Classic Blend Rabbit Food, Secaucus, NJ]: liver powder [ICN Biomedicals, Inc., Aurora, OH]). For each trial, we placed pupae into a 7.6 l bucket with a cloth mesh cover to synchronize adult eclosion within a 24-h period. Each bucket contained males and females to ensure mating. No autogeny was observed in this colony.
Adults were fed only 20% sucrose solution during the first 5 days of the holding period. On days 6 and 7, mosquitoes were given only water. Six days after emergence, we anesthetized mosquitoes in each bucket with cold and placed 100 females into a clean, empty bucket cage. On day 7, we confirmed the number of live females in the bucket and performed the experiment.
Chickens
We reared K-strain leghorn chicks as described previously (Darbro and Harrington 2006). Edman et al. (1974) reported that anti-mosquito behavior in chickens can vary with age. We accounted for this by measuring the age and weight of each bird (K-strain leghorn chicks or sparrows) before the beginning of each trial. To determine the effect of age, we performed 2 trials with 6 chicks (aged 3–6 weeks old) and 4 trials with 2 individual chickens at 5, 9, 14, and 17 weeks of age.
House sparrows
We captured 6 after-hatch-year house sparrows (P. domesticus) and 1 hatch-year sparrow from barns in Tompkins County, NY, in June and July 2004, and we identified their ages according to Pyle (1997). Four sparrows were used in experiments 3 times, and 3 sparrows were used twice. We maintained sparrows in captivity in a free flight room with food (Flyers Choice Premium Mix) and water provided ad libitum.
Researchers complied with Cornell University Institutional Animal Care and Use Committee approved protocols (04-45 and 04-56) concerning housing, maintenance, and handling for all birds.
Experimental setup
We constructed an observation cage (2.4 x 1.2 x 1.2 m) with a wooden frame covered with a fiberglass and aluminum screen (New York Wire, Mount Wolf, PA). Entry to the cage was gained through a 1.2 x 0.6–m door, within which 2 sheets of overlapping muslin (cotton) prevented mosquito escape. Within the observation cage, we placed a bird cage (30 x 30 x 60 cm) either on the floor or supported 1.2 m off the floor. We utilized a Sony 1/3'' Super HAD CCD Sensor Color Camera (Liberty Hill, TX) with built-in infrared illuminator to record all avian behaviors. Natural light from a nearby window illuminated all trials. A data logger (HOBO, Onset Corporation, Bourne, MA) recorded hourly temperature and humidity. Room temperature and humidity varied with time of year (17.1–25.5 °C and 23.4–32.2% relative humidity (RH), respectively). Within a single experiment, the temperature and humidity varied by no more than 2 °C or 3% RH.
Experimental procedure
Each trial began approximately 90 min before sundown, with the placement of a chicken or sparrow in the bird cage within the observation cage. Once the bird was in place, we set a 7.6 l bucket with a fiberglass cover, containing 100 7-day-old female mosquitoes, on top of the cage. The first trial began in October 2004, and the last trial concluded in July 2005.
Placement of the bird and mosquitoes in the observation cage ended 30 min before sundown. To record avian behavior in the absence of mosquitoes, we began videotaping 30 min before mosquito release. We released mosquitoes by pulling a string tied to the lid from outside the cage. Mosquito release occurred after sundown, between 2030 and 2130 h in all trials.
Data collection
Mosquitoes were allowed free access to the bird overnight and were aspirated from inside the cage at 0700–0800 the next morning. We determined mosquito mortality by observing whether or not the collected mosquitoes could still fly. Dead mosquitoes that remained in the release bucket were subtracted from the total released. We categorized all mosquitoes visually as non-blood fed (NBF), partially blood fed (PBF), or fully blood fed (FBF). NBF had no visible blood in the abdomen; FBF had more than 50% of the abdomen filled with blood. Mosquitoes with lesser amounts of ingested blood than FBF were classified as PBF. These classifications were based on findings made by Edman et al. (1975) that female Culex nigripalpus with no more than a 50% blood meal usually attempted to re-feed within 30 s of interruption. We considered any mosquitoes not recovered from the cage as eaten by the bird.
We recorded all avian behavior for 30 min before mosquito release and 6 h afterwards. While viewing the entire 30-min prerelease footage, we counted and categorized any observed behaviors. After mosquito release, we viewed 15 min of each hour for the first 6 h.
While under mosquito attack, both bird species exhibited a group of behaviors that were only rarely observed during the control periods. We considered these to be "defensive behaviors." The birds rarely reacted (e.g., by taking flight) to the drop of the plastic lid or the release of the mosquitoes, and birds that did react became calm within 5 min. We observed similar categories of defensive behaviors as those described for ciconiiform birds by Webber and Edman (1972); consequently we used similar names and descriptions (Table 1).
|
Mosquito density was estimated by visually counting the maximum number of mosquitoes seen in proximity (<15 cm) of the bird for each 15-min scoring period. For the purposes of this study, we defined "mosquito density" by these visual counts, not the number of mosquitoes placed into the observation cage per night.
Data analysis
Our primary goals were to evaluate differences in avian anti-mosquito defensive behavior rates between chicks and sparrows and to evaluate the effect of avian defensive behavior and bird species on mosquito blood-feeding success. Our interest extended to adult chickens because they are used as sentinel animals for arbovirus surveillance. To evaluate the effect of chick age, we investigated the differences in defensive behavior and mosquito blood-feeding success on 5-, 9-, 14-, and 17-week-old chickens. Finally, we investigated the effect of bird species on mosquito density.
To confirm that our sampling method of recording behaviors for the first 15 min of each hour was representative of the entire period, we counted 6 continuous hours of total behaviors for 6 trials with 3 different chicks and 3 different house sparrows. We compared the continuous counts with counts obtained by only counting the first 15 min of each hour using a general linear model (PROC MIXED, SAS 9.1, SAS Institute, Cary, NC). The dependent variable was total behavior rates. We entered sampling method ("census" or "sample"), bird species, and time (hours) as main effects. We also entered into the initial model interactions between sampling method and bird species, as well as between sampling method and time. Individual bird ID nested within bird species was entered as a random effect.
We constructed mixed models (PROC MIXED) for analyses using bird ID nested within bird species as a random effect in order to control for interbird variation. Depending on the analysis, the dependent variables included total defensive behaviors per minute, individual defensive behaviors per minute, the proportion of mosquitoes per trial that successfully blood fed (full or partial blood meal), the proportion of mosquitoes that obtained partial blood meals, or the proportion of mosquitoes eaten by the bird. We added temperature, experiment replicate number, relative humidity, bird age, bird weight, and bird location (e.g., on the perch, standing on the floor, or moving back and forth between the 2) to the initial model as fixed effects along with all 2-way interactions.
The exact age of the wild-caught house sparrows could not be determined, beyond hatch-year and after-hatch-year, because they were captured wild. Therefore, we entered sparrow ages as "0" (hatch-year) or "1" (after-hatch-year), and we entered chick age by their age in days. Bird age represented an interaction between bird species and age in the model. Bird weight was handled in a similar way because the chicks always weighed more than the sparrows. We tested age and weight in different models because we found a high degree of correlation between chick age and chick weight. We also recorded and entered the resting position of the bird (on the perch versus on the floor of the cage) into models as "bird location," but we did not manipulate this factor experimentally.
We calculated degrees of freedom (df) using the Kenwood–Rogers method (Kenwood and Roger 1997). The final model included all variables of a significant (P value 
0.05) interaction even when a main effect was not significant (Neter et al. 1996). We excluded all other interactions found not significant (P value > 0.05).
We assessed bird behavior counts for homoscedasticity using the nonparametric Levene's test (Levene 1960) and for normality using a normal probability plot. If data did not satisfy these assumptions for general linear analysis, it was square root transformed or natural log transformed prior to analysis, depending on which best fit statistical assumptions. Similarly, we assessed proportions of mosquito blood-feeding success for normality and homoscedasticity using an arcsine square root transformation when necessary.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sampling method
All means are given as "mean ± standard error (SE)." Results using our sampling method were not statistically different from those obtained by counting all the behaviors for 6 trials (least squared mean [LSM] difference = –0.103 ± 0.006, df = 157, t = –1.04, P = 0.299). None of the other variables entered into the model (bird species, time, bird species x sampling method, and bird species x time) returned significant values. Intrabird variation was not a significant component of random variance (Z = 1.52, P = 0.064).
Variance of mosquito density
Overall, we observed more mosquitoes flying around house sparrows than around chickens (t = 8.28, df = 1, P < 0.0001), although this difference varied over time (Figure 1). Mosquito counts around chickens did not vary over time.
|
Defensive behavior
Observed defensive behaviors and their relative frequency for both bird species are presented in Table 2. Over 90% of observed defensive behaviors fell into 1 of 3 categories: foot stomp, head movement, or wing movement. Foot stomps were the most frequent defensive behavior in sparrows (56%) and chickens (47%).
|
House sparrows exhibited 11.7 ± 1.1 total defensive behaviors per minute, and 3- to 6-week-old chicks exhibited 6.2 ± 1.2 total behaviors per minute. The results of a mixed linear model revealed that significant predictors of total defensive behavior rates (natural log transformed) included bird species, mosquito density, temperature, replicate number, and bird location (Table 2). Chicks displayed lower behavior rates than house sparrows (LSM difference = –0.490 ± 0.183, df = 26.2, t = –2.68, adjusted P = 0.013). Total behavior rates increased proportionately with mosquito count (b = 0.177 ± 0.047, df = 124, t = 3.78, P = 0.0002) and temperature (b = 0.054 ± 0.023, df = 64.5, t = 2.34, P = 0.023). Birds tended to display lower behavior rates with each subsequent replicate. Birds displayed significantly higher behavior rates during the first replicate than the third (difference = 0.404 ± 0.149, df = 123, t = 2.71, adjusted P = 0.021). Other pairwise differences were not significant. Results revealed humidity, bird age, and bird weight as not significant fixed effects. Interbird variance was not a significant component of the residual variance (Z = 1.17, P = 0.1200). The only significant 2-way interaction was mosquito density and time: the effect of mosquito density on total behavior rates was significantly reduced during the first hour of exposure than during other hours (Table 2).
An analysis of 2 chicks tested 4 times over 12 weeks revealed that age was the only significant variable (Figure 2). Nine-week-old chicks displayed higher behavior rates than chicks of other ages (difference = 2.117 ± 0.482; F1,43 = 19.27, P < 0.0001). Mosquito density, bird location, and 2-way interactions were not significant predictors, and we detected no significant interbird variation.
|
Individual defensive behaviors: foot stomps
We examined the 3 most common behaviors for both bird species in greater detail. House sparrows exhibited 6.4 ± 9.5 foot stomps per minute, and 3- to 6-week-old chicks exhibited 2.9 ± 0.2 foot stomps per minute. A mixed model revealed that foot stomps (natural log transformed) positively correlated with mosquito numbers (b = 0.177 ± 0.017, df = 197, t = 10.41, P < 0.0001). We observed lower foot stomp rates during the first hour of exposure than in the following hours (estimate = –0.867 ± 0.102; contrast: F1,187 = 71.77, P < 0.0001). Birds on the cage floor stomped their feet more than perching birds (difference = 0.283 ± 0.131, df = 44.8, t = 2.16, adjusted P = 0.036). We found bird species, replicate number, temperature, humidity, bird age, and bird weight to not be significant main effects, and we detected no significant interactions. Intrabird variation within bird species remained a significant component of random variance (estimate = 0.054 ± 0.030, Z = 1.78, P = 0.038).
In chicks aged 5–17 weeks, the mean foot stomp rate was 2.3 ± 0.2 foot stomps per minute. None of the measured variables (time, age, weight, and mosquitoes) were found to be significant.
Head movements
House sparrows exhibited 2.4 ± 0.2 head movements per minute, and 3- to 6-week-old chicks exhibited 2.0 ± 0.2 head movements per minute. Rates of head movements (natural log transformed) positively correlated with mosquitoes (b = 0.102 ± 0.016, df = 142, t = 6.39, P < 0.0001) and negatively correlated with temperature (b = –0.054 ± 0.027, df = 135, t = –1.98, P = 0.050). Both species exhibited lower movement rates during the second trial (b = –0.393 ± 0.139, df = 139, t = –2.83, P = 0.005) than during the first or third trials (difference = –0.310 ± 0.093; contrast: F1,137 = 11.16, P = 0.001). We detected no significant values or interactions when we entered time, bird species, humidity, bird location, bird age, and bird weight. Intrabird variance within each bird species was a significant component of random variance (estimate = 0.145, SE = 0.068, Z = 2.13, P = 0.0148).
The only significant factor of head movement rates (square root transformed) in chicks 5–17 weeks old was the number of mosquitoes (b = 0.126 ± 0.047, df = 45, t = 2.70, P = 0.010), which was positively correlated. Chicks 5–17 weeks old exhibited 1.4 ± 0.3 head movements per minute.
Wing movements
House sparrows exhibited 35.7 ± 6.6 wing movements per minute, and 3- to 5-week-old chicks exhibited 14.0 ± 1.2 wing movements per minute. Wing movement rates (natural log transformed) positively correlated with mosquitoes (b = 0.104 ± 0.020, df = 140, t = 5.24, P < 0.0001) and temperature (b = 0.106 ± 0.031, df = 133, t = 3.45, P = 0.001). Birds on the cage floor had lower wing movement rates during the first hour than during subsequent hours (b = –0.450 ± 0.156, df = 125, t = –2.86, P = 0.005). We detected no significant interactions or significant predictors for bird species, replicate number, humidity, bird location, bird age, or bird weight. Intrabird variance within bird species was a significant component of random variance (estimate = 0.215 ± 0.110, Z = 1.96, P = 0.025).
Older (5–17 weeks) chicks exhibited 8.0 ± 1.8 wing movements per minute. Wing movement rates (natural log transformed) positively correlated with mosquitoes in older chicks (b = 0.316 ± 0.145, df = 37, t = 2.17, P = 0.036). Chicks exhibited significantly lower wing movement rates during the first hour than during the subsequent hours (b = –1.259 ± 0.404, df = 37, t = –3.11, P = 0.004). Wing movement rates decreased with bird age: 5- and 9-week-old chicks displayed significantly higher wing movement rates than 13- and 17-week-old birds (difference = 0.755 ± 0.170; F1,37 = 19.62, P < 0.0001).
Mosquito blood-feeding success
Overall, the proportion of mosquitoes that successfully fed on sparrows (out of all mosquitoes released into the cage) was 0.822 ± 0.0472 (69.2 ± 4.2 blood meals per sparrow), and the proportion for chicks was 0.579 ± 0.0447 (42.7 ± 6.2 blood meals per chick). After controlling for other variables in the mixed model (transformation of proportion data was not necessary), mosquitoes had significantly higher blood-feeding success on sparrows (difference = 0.243 ± 0.0651, df = 34, t = 3.73, P = 0.0007).
To predict blood-feeding success, we entered the mean total defensive behavior rates per trial into a mixed model. We omitted the first hour of total behaviors before computing the mean because we observed significantly lower behavior rates during the first hour of the trial compared with the rest of the night (Table 2). Defensive behavior as a main effect was not a significant predictor of blood-feeding success (t = 0.66, P = 0.5193), but there was a significant interaction between bird species and defensive behavior. Chick behavior had a negative correlation to mosquito blood-feeding success on chicks (b = –0.00358 ± 0.00099, df = 28.3, t = –3.63, P = 0.0011). Mosquitoes tended to feed more on birds on the cage floor rather than perching birds, but this trend was not significant (difference = 0.137 ± 0.070, t = 1.97, P = 0.0596). Intrabird variation within each bird species composed a significant portion of the random variance (estimate = 0.0167 ± 0.00902, Z = 1.85, P = 0.0319). Nonsignificant factors of blood-feeding success included replicate number, temperature, humidity, bird age, and bird weight.
We performed an additional analysis to evaluate factors affecting the proportion of partial blood meals taken by mosquitoes. The proportion of partial blood meals on chicks was 0.179 ± 0.030, and the proportion for house sparrows was 0.0491 ± 0.0315. We observed more partial blood meals from chicks (t = 8.85, df = 11.5, adjusted P = 0.0004). Additionally, mosquitoes tended to take partial blood meals from perching birds rather than from birds standing on the floor (difference = 0.233 ± 0.0652, df = 22.7, t = 3.57, P = 0.0016). Two-way interactions among several factors yielded nonsignificant values including replicate number, temperature, humidity, bird age, bird weight, and defensive behavior. We found intrabird variance not to be a significant component of random variance (Z = 0.12, P = 0.4532).
The proportion of mosquito blood-feeding success was 0.792 ± 0.0189 on 5- to 17-week-old chicks (64.3 ± 2.5 blood meals per chick), and the proportion of partial blood meals was 0.110 ± 0.0242. None of the variables tested (bird age, bird weight, and defensive behavior) emerged as significant factors of blood meal success or partial blood meals.
Ingested mosquitoes
We assumed the birds consumed any missing mosquitoes not recovered the next morning. The proportion of ingested mosquitoes was 0.094 ± 0.0119. This proportion was arcsine square root transformed to satisfy assumptions of homoscedasticity and normality. Total defensive behaviors positively correlated with the proportion of eaten mosquitoes (b = 0.00157 ± 0.000565, df = 2.78, P = 0.0091). None of the other variables entered into the initial model proved significant, including bird species, replicate number, bird weight, bird age, temperature, or 2-way interactions. Intrabird variance was a significant portion of random variance (estimate = 0.0136 ± 0.00626, Z = 2.17, P = 0.0152).
The proportion of mosquitoes eaten by 5- to 17-week-old chicks was 0.0858 ± 0.0259. None of the variables tested (bird age, bird weight, and defensive behavior) contributed significantly to the rate of mosquito consumption.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study represents the most detailed analysis of anti-mosquito behavior within a single avian species. Previous studies assessed behavior for 1 (Kale et al. 1972
; Edman et al. 1974) or 2 h (Day and Edman 1984; Scott et al. 1988), possibly being limited by nightfall. We chose to record for 6 h because most birds experienced little mosquito attack after that time. A longer observation period allowed us to observe variation and time-dependent decline in the actual number of mosquitoes attacking a host.
We measured 82% feeding success of C. pipiens pipiens on house sparrows, which is higher than that for other mosquito species reported elsewhere. Edman et al. (1974) reported 24% blood-feeding success of 100 C. nigripalpus on house sparrows in Florida, and Scott et al. (1988) reported 10–30% blood-feeding success of 10 Culex tarsalis and 10 Culex p. quinquefasciatus on adult house sparrows. Differences in blood-feeding success may be due to variation in biting persistence on birds of each mosquito species, tolerance threshold variation in different populations of house sparrows, differences in the initial number of mosquitoes used in the experiment, or different-sized experimental cages. Also, C. pipiens pipiens is known to specialize on avian hosts to a greater degree than C. p. quinquefasciatus, C. Nigripalpus, and C. tarsalis, so C. pipiens pipiens may be better adapted to feeding on birds.
We did not detect a significant relationship between house sparrow defensive behavior and C. pipiens pipiens blood-feeding success. Mosquito density around house sparrows declined over time throughout the night, most likely because mosquitoes cease host seeking immediately after engorging on blood. The house sparrow's permissiveness to biting mosquitoes, in combination with its peridomestic lifestyle and high virus titer (Komar et al. 2003), make it an ideal amplifying host for WNV.
The defensive behavior of chickens is of interest because they are used as sentinel animals to detect arboviruses such as WNV or St Louis encephalitis virus. In this study, we report 57% blood-feeding success rate of C. pipiens pipiens on chicks. If mosquitoes preferentially feed on passerines such as house sparrows in nature, sentinel chickens may have limited sensitivity in areas where C. pipiens pipiens is a primary vector of WNV, although a recent field study detected no preference in C. pipiens pipiens between mosquito traps baited with chicks or house sparrows (Darbro and Harrington 2006). As the number of attacking mosquitoes increased, both the sparrows and the chicks exhibited an increase in their defensive behaviors. Furthermore, mosquitoes fed less frequently on chicks that exhibited more defensive behaviors.
We measured defensive behaviors and mosquito blood-feeding success repeatedly for 2 chicks from 5 to 17 weeks of age. Sentinel chickens have been evaluated in the laboratory at 17 weeks (Langevin et al. 2001). Six-week-old chicks used for arbovirus surveillance are exposed to local mosquito populations at the beginning of the season and continue to be used until they seroconvert (Moore et al. 1993). Even with a relatively small sample size, we found that older chickens seem to be more sensitive sentinels than younger chicks. This may have implications for the sensitivity of surveillance programs during the spring, using young sentinel chicks.
It is interesting to note that chicks exhibited fewer defensive behaviors than sparrows while being fed upon 24% less by mosquitoes. Furthermore, it was much more likely that a blood meal taken from a chick would be a partial blood meal. We suspect that factors apart from defensive behavior may play a role. Because chicks displayed similar individual behavior types as the house sparrows, it is possible that chicks had additional protection (compared with house sparrows) in the form of thicker plumage. Extra plumage may reduce blood-feeding success either alone or in the presence of defensive behaviors. Comparisons of feathered versus unfeathered chicks, not conducted in this study, may answer this question. There was no detectable relationship between defensive behavior and blood-feeding success among the older chicks (6–17 weeks), even though they exhibited fewer behaviors and higher mosquito loads than younger chicks. Another explanation for the differences we observed is that house sparrows are more attractive to C. pipiens pipiens than chicks, although our recent field study (Darbro and Harrington 2006) did not detect any difference in mosquito capture rates between chicken- and sparrow-baited traps. In the current study, house sparrows experienced more mosquito attacks than chicks, although the difference decreased over the course of a night. Choice assays using an olfactometer may illuminate differences in attraction between sparrows and chicks.
Whole-body behaviors (i.e., full-body feather fluffs and flights within the cage), which would be expected to be more energetically costly than localized behaviors (e.g., foot stomps), were rarely observed (0.3% and 1.3% of total sparrow behaviors, respectively), suggesting that birds favor low-energy behaviors to high-energy behaviors. Within the low-energy behaviors, we observed foot stomps most frequently, which corresponds to mosquitoes' observed tendency to attack the feet. Temperature correlated negatively with head movements and positively with wing movements, so these behaviors may have had a role in thermoregulation in addition to anti-mosquito defense.
Employing defensive behavior against attacking mosquitoes presents a trade-off for birds. The disadvantages of being bitten by mosquitoes include blood loss and infection with pathogens. On the other hand, frequent defensive activity at night is energetically costly, especially for a relatively small bird. Based on the high feeding success of mosquitoes on house sparrows in this study, it may be that the energetic costs of "effective" defensive behavior in sparrows may be prohibitively high compared with the costs of being bitten by mosquitoes. The tendency of both house sparrows and chicks to react to subsequent mosquito exposures with fewer behaviors may also be adaptive.
Birds displayed significantly more total behaviors and wing movements with higher temperatures and more head movements at lower temperatures. One explanation for this result is that mosquitoes may be more active at higher temperatures and chicks may be behaving more vigorously in response. An alternative explanation is that behaviors may have had a role in thermoregulation in addition to anti-mosquito defense. A third possibility is that relative costs are temperature dependent, and certain behaviors are less costly at higher temperatures.
We detected some significant intraspecific variation in defensive behavior within our study populations of house sparrows and chicks. Although individual birds did not vary significantly in total defensive behaviors, we observed significant variation in head and wing movements. We also noted significant intrabird variation in mosquito blood-feeding success. Because this study evaluated only 7 house sparrows and 8 chickens, natural populations may vary even more.
We did not find a significant replicate effect on blood-feeding success, even though we found that individual birds tended to display progressively fewer total behaviors with each trial. Lower blood-feeding success in subsequent replicates of an individual bird would suggest that the bird is becoming more effective at defending itself against mosquitoes. For example, Waage and Nondo (1982) found the blood-feeding success of Ae. aegypti on individual rabbits decreased as the rabbits experienced multiple exposures to mosquitoes during the day. Anderson and Brust (1996) found similar results with Japanese quail. Edman and Scott (1987) found that C. tarsalis had lower feeding success on house sparrows that had experienced biting mosquitoes once than on naive sparrows. Our results suggest that house sparrows do not become more effective at defending themselves, at least within 3 exposures to mosquitoes.
If avian defensive behavior decreases arbovirus transmission by preventing vector–host contact, we would expect to see lower blood-feeding success on more defensive birds. House sparrows exhibited more defensive behaviors while being fed upon more frequently by attacking mosquitoes. House sparrows received an average of 69.7 bites per night per bird. Young chicks (3–6 weeks) that exhibited more defensive behavior tended to be fed upon less. The most defensive chick (10.0 total behaviors per minute) was fed on by 32% of mosquitoes (n = 23), and the least defensive chick (4.2 total behaviors per minute) was fed on by 83.5% of mosquitoes (n = 71). Although natural biting rates would be expected to vary by mosquito and host density, these differences in experimental rates could have significant impacts on arbovirus transmission dynamics.
For birds of both species, the rate of eaten mosquitoes increased with defensive behaviors. We could not determine if eaten mosquitoes had blood fed or not, but eating mosquitoes would reduce vector–host contact, as mosquitoes would certainly not be able to bite subsequent hosts after being eaten. The risk of being eaten also serves as a source of mortality for host-seeking mosquitoes. In this case, C. pipiens pipiens attempting to feed on house sparrows or chicks experience an average of 9% mortality. This can affect the probability of an infected vector surviving long enough to transmit an arbovirus such as West Nile.
If avian defensive behavior increases arbovirus transmission by increasing vector–host contact, then we would expect to see more partial blood meals on more defensive bird species. Blood meals taken from chicks contained a higher proportion of partial blood meals than those from sparrows, suggesting that mosquitoes are more likely to be interrupted while feeding on chicks, even though chicks displayed fewer defensive behaviors. This study may have underestimated the rate of partial blood meals because we could not determine how many times each fully engorged mosquito had to bite a bird in order to obtain a full blood meal. If mosquitoes require multiple feedings to obtain a blood meal in nature, where there are also multiple hosts, then defensive behavior would increase the amount of vector–host contact.
One potential limitation of our study was reliance on visual counts to estimate mosquito density. Although we consistently detected positive relationships between mosquito density and defensive behaviors, mosquito counts could have been underestimated if mosquitoes were visually obstructed by the bird, resting against a dark background or flying in a group.
Because passerines such as house sparrows serve as reservoirs for WNV in North America, knowledge of vector–host interactions, such as avian defensive behavior, may be useful to make predictions of arbovirus transmission. According to ArboNET (http://www.cdc.gov/ncidod/dvbid/westnile/birdspecies.htm, as of 24 January 2007), WNV has been detected in 284 bird species, so other bird species could serve as amplifying hosts. House sparrows have been found infected with WNV during periods of transmission in New York City (Komar, Burns, et al. 2001; Komar, Panella, et al. 2001) and Louisiana (Komar et al. 2005). Kilpatrick et al. (2006) and Molaei et al. (2006) reported that American robins (Turdus migratorius) had disproportionately high prevalence of WNV infection in Washington DC/Maryland and Connecticut, respectively. These studies suggest that WNV amplifying hosts vary geographically, and the results of studies of defensive behavior of other passerine hosts would be informative.
Average blood-feeding success on house sparrows in our study was uniformly high, suggesting that the unrestrained defensive behavior of a single house sparrow is insufficient to prevent mosquito blood feeding. Defensive behavior of individuals within groups is more likely to have an effect in nature (i.e., mosquitoes are more likely to feed on a less defensive host than a nearby host who is more defensive). Anderson and Brust (1996) exposed Japanese quail to mosquitoes for 55 min, randomly sampling 10 min of behavior for analysis. In our studies, we found that mosquito attack rate and total defensive behavior rate change over time, so comparative studies in this system would be more appropriately observed for longer periods of time. Such comparative studies of anti-mosquito behavior in passerines are currently underway.
|
ACKNOWLEDGEMENTS |
|---|
The authors would like to thank J. Edman, D. Rutz, and A. Dhondt for helpful advice and for reviewing earlier versions of this manuscript, E. Swarthout for assisting with trapping of wild house sparrows, T. Van Deusen for assistance with sparrow housing and maintenance, P. Kaufman for advice on cage construction, and F. Vermeylen for invaluable statistical advice. This project was funded by Centers for Disease Control Grant USO/CCU 220512 and Hatch Project NYC-139410.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Anderson RA, Brust RA. Blood feeding success of Aedes aegypti and Culex nigripalpus (Diptera: Culicidae) in relation to defensive behavior of Japanese quail (Coturnix japonica) in the laboratory. J Vector Ecol (1996) 21(1):94–104.[Web of Science]
Apperson CP, Harrison BA, Unnasch TR, Hassan HK, Irby WS, Savage HM, Aspen SE, Watson DW, Rueda LM, Engber BR, et al. Host-feeding habits of Culex and other mosquitoes (Diptera: Culicidae) in the borough of Queens in New York City, with characters and techniques for identification of Culex mosquitoes. J Med Entomol (2002) 39(5):777–785.[Web of Science][Medline]
Apperson CS, Hassan HK, Harrison BA, Savage HM, Aspen SE, Farajollahi A, Crans W, Daniels TJ, Falco RC, Benedict M, et al. Host feeding patterns of established and potential mosquito vectors of West Nile virus in the Eastern United States. Vector Borne Zoonotic Dis (2004) 4(1):71–82.[CrossRef][Web of Science][Medline]
Blackmore JS, Dow RP. Differential feeding of Culex tarsalis on nestling and adult birds. Mosq News (1958) 18(1):15–17.
Cully JF Jr, Heard PB, Wesson DM, Craig GB Jr. Antibodies to La Crosse virus in eastern chipmunks in Indiana near an Aedes albopictus population. J Am Mosq Control Assoc (1991) 7:651–653.[Web of Science][Medline]
Darbro JM, Harrington LC. Bird-baited traps for surveillance of West Nile mosquito vectors: effect of bird species, trap height, and mosquito escape rates. J Med Entomol (2006) 43(1):83–92.[CrossRef][Web of Science][Medline]
Day JF, Edman JD. Mosquito engorgement on normally defensive hosts depends on host activity patterns. J Med Entomol (1984) 21(6):732–740.[Web of Science][Medline]
Edman JD, Cody E, Lynn H. Blood-feeding activity of partially engorged Culex nigripalpus (Diptera: Culicidae). Entomol Exp Appl (1975) 18:261–268.[CrossRef]
Edman JD, Kale HW, 2nd. Host behavior: its influence on the feeding success of mosquitoes. Ann Entomol Soc Am (1971) 64(2):513–516.[Web of Science]
Edman JD, Scott TW. Host defensive behaviour and the feeding success of mosquitoes. Insect Sci Appl (1987) 8:617–622.
Edman JD, Webber LA, Schmid AA. Effect of host defenses on the feeding pattern of Culex nigripalpus when offered a choice of blood sources. J Parasitol (1974) 60(5):874–883.[CrossRef][Medline]
Garrett-Jones C. The human blood index of malaria vectors in relation to epidemiological assessment. Bull World Health Org (1964) 30:241–261.[Medline]
Goddard LB, Roth AE, Reisen WK, Scott TW. Vector competence of California mosquitoes of West Nile virus. Emerg Infect Dis (2002) 8(12):1385–1391.[Web of Science][Medline]
Godsey MS Jr, Blackmore MS, Panella NA, Burkhalter K, Gottfried K, Halsey LA, Rutledge R, Langevin SA, Gates R, Lamonte KM, et al. West Nile virus epizootiology in the southeastern United States, 2001. Vector Borne Zoonotic Dis (2005) 5(1):82–89.[CrossRef][Web of Science][Medline]
Kale HW, Edman JD, Webber LA. Effect of behavior and age of individual ciconiiform birds on mosquito feeding success. Mosq News (1972) 32(3):343–350.[Web of Science]
Kenwood MG, Roger JH. Small sample inference for fixed effects from restricted maximum likelihood. Biometrics (1997) 53:938–997.
Kilpatrick AM, Daszak P, Jones MJ, Marra PP, Kramer LD. Host heterogeneity dominates West Nile virus transmission. Proc R Soc Lond B Biol Sci (2006) 273:2327–2333.[Medline]
Klowden MJ, Lea AO. Blood meal size as a factor affecting continued host-seeking in Aedes aegypti (L.). Am J Trop Med Hyg (1978) 27:827–831.
Komar N, Burns J, Dean C, Panella DA, Dusza S, Cherry B. Serological evidence for infection of West Nile virus in birds in Staten Island, after an outbreak in 2000. Vector Borne Zoonotic Dis (2001) 1(3):191–196.[CrossRef][Medline]
Komar N, Langevin S, Hinten S, Nemeth N, Edwards E, Hettler D, Davis B, Bowen R, Bunning M. Experimental infection of North American birds with the New York 1999 strain of West Nile virus. Emerg Infect Dis (2003) 9(3):311–322.[Web of Science][Medline]
Komar N, Panella NA, Burns JE, Dusza SW, Mascarenhas TM, Talbot TO. Serological evidence for West Nile infection in birds in the New York city vicinity during an outbreak in 1999. Emerg Infect Dis (2001) 7(4):621–625.[Web of Science][Medline]
Komar N, Panella NA, Langevin SA, Brault AC, Amador M, Edwards E, Owen JC. Avian hosts for West Nile virus in St. Tammany Parish, Louisiana, 2002. Am J Trop Med Hyg (2005) 73(6):1031–1037.
Langevin SA, Bunning M, Davis B, Komar N. Experimental infection of chickens as candidate sentinels for West Nile virus. Emerg Infect Dis (2001) 7(4):726–729.[Web of Science][Medline]
Levene H. Robust tests for equality of variance. In: Contributions to probability and statistics—Olkin I, ed. (1960) Palo Alto (CA): Stanford University Press. 278–292.
Molaei G, Andreadis T, Armstrong P, Anderson J, Vossbrinck C. Host feeding patterns of Culex mosquitoes and West Nile virus transmission, northeastern United States. Emerg Infect Dis (2006) 12:468–474.[Web of Science][Medline]
Moore CG, McLean RG, Mitchell CJ, Nasci RS, Tsai TF, Calisher CH, Marfin AA, Moore PS, Gubler DJ. Guidelines for arbovirus surveillance in the United States. (1993) Fort Collins (CO): United States Department of Health and Human Services/Public Health Service/Centers for Disease Control/National Center for Infectious Diseases/Division of Vector-borne and Infectious Diseases.
Neter J, Kutner MH, Nachtsheim CJ, Wassserman W. Two-factor analysis of variance—equal sample sizes: meaning of ANOVA model elements. In: Applied linear statistical models (1996) 4th ed. Boston (MA): WCB/McGraw-Hill. 811–812.
Pyle P. The identification guide to North American birds: part 1. (1997) Bolinas (CA): Slate Creek Press.
Reisen WK, Fang Y, Martinez VM. Avian host and mosquito (Diptera: Culicidae) vector competence determine the efficiency of West Nile and St. Louis encephalitis virus transmission. J Med Entomol (2005) 42(3):367–375.[CrossRef][Web of Science][Medline]
Scott TW, Edman JD, Lorenz LH, Hubbard JL. Effects of disease on vertebrates' ability to repel host-seeking mosquitoes. Misc Publ Entomol Soc Am (1988) 68:9–17.
Turell MJ, Dohn DJ, Sardelis MR, O'Guinn ML, Andreadis TG, Blow JA. An update on the potential of North American mosquitoes (Diptera: Culicidae) to transmit West Nile virus. J Med Entomol (2005) 42(1):57–62.[CrossRef][Web of Science][Medline]
Waage JK, Nondo J. Host behaviour and mosquito feeding success: an experimental study. Trans R Soc Trop Med Hyg (1982) 76(1):119–122.[CrossRef][Web of Science][Medline]
Walker ED, Edman JD. The influence of host defensive behavior on mosquitoes (Diptera: Culicidae) biting persistence. J Med Entomol (1985) 22(4):370–372.[Web of Science][Medline]
Walker ED, Edman JD. Influence of defensive behavior of Eastern Chipmunks and Gray Squirrels (Rodentia: Sciuridae) on feeding success of Aedes triseriatus (Diptera: Culicidae). J Med Entomol (1986) 23(1):1–10.[Web of Science][Medline]
Webber LA, Edman JD. Anti-mosquito behaviour of ciconiiform birds. Anim Behav (1972) 20:228–232.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. C. Jansen, C. E. Webb, G. C. Graham, S. B. Craig, P. Zborowski, S. A. Ritchie, R. C. Russell, and A. F. van den Hurk Blood Sources of Mosquitoes Collected from Urban and Peri-Urban Environments in Eastern Australia with Species-Specific Molecular Analysis of Avian Blood Meals Am J Trop Med Hyg, November 1, 2009; 81(5): 849 - 857. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Darbro, A. A. Dhondt, F. M. Vermeylen, and L. C. Harrington Mycoplasma gallisepticum Infection in House Finches (Carpodacus mexicanus) Affects Mosquito Blood Feeding Patterns Am J Trop Med Hyg, September 1, 2007; 77(3): 488 - 494. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||