| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
 |
ABSTRACT |
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We investigated the roosting behavior of two bird species that may be very important in the transmission dynamics of WNV, particularly in the Midwest; American crows (Corvus brachyrhynchos) and northern cardinals (Cardinalus cardinalus). Since WNV was first detected in New York in 1999, crows have been the primary bird species associated with WNV transmission activity and human risk.4 Public heath departments use dead crows as sentinels of WNV activity.4–6 The reason crows are linked with WNV is that they experience extremely high mortality when infected with WNV,7,8 and based on laboratory tests, they are highly competent hosts.9,10 The average viremia lasted for 3.5 days in a study by Komar and others10 and 4–5 days in a study by Weingartl and others.11 Crows typically die within seven days of being infected with WNV.9 Furthermore, when healthy and infected crows have been confined together, bird-to-bird transmission has been documented.9,12 Crows also develop particularly high titers of the virus,10–12 with the highest three days post-infection.11
Understanding the movement of crows provides valuable information on how far individuals travel and how quickly WNV may spread across an area as a result of viremic crows infecting mosquitoes in new geographic areas. Based on an analysis of movements of healthy crows in east-central Illinois, Yaremych and others suggested that crows occupy a diurnal home range of 7.6 km2 and may potentially infect mosquitoes over this area if viremic crows exhibited normal movements.13 However, a better understanding of the role of crows in dispersing the virus may be gained through analysis of crow movement at the time when they are most likely to encounter mosquitoes at nocturnal roosts. Furthermore, analysis of the movement of crows that later die of WNV infection may provide specific data on differences in movement of sick versus healthy crows and implications for transmission and dispersal of the virus. Because WNV causes neurologic problems,14 viremic individuals may behave differently than healthy individuals.
Recently, there has been focus on species other than crows (or corvids) as potentially important reservoirs of WNV (i.e., American robins [Turdus migratorius]).1 In the Midwest, one bird species with the highest seroprevalence is the northern cardinal.15,16 Northern cardinals in Georgia and Florida had the highest prevalence of WNV antibodies among passerine birds,17 and cardinals in Louisiana have been suggested to be the primary amplifying hosts.18 West Nile virus is known to cause some mortality in a wide range of bird species, including northern cardinals (www.cdc.gov).18 Although northern cardinals were not included in an experimental infection study by Komar and others,10 three of the five most competent avian hosts were non-corvids (i.e., common grackle [Quiscalus quiscula], house finch [Carpodacus mexicanus], and house sparrow [Passer domisticus]). Northern cardinals were also among the most frequently fed upon species by several Culex mosquitoes as determined by blood meal analysis.1,19 There is growing body of evidence that non-corvid passerines are involved in the transmission cycle of WNV, primarily based on the observation that in some areas where crow populations have significantly decreased,20 WNV activity is still relatively high. If the roosting behavior of crows is responsible for the amplification and spread of the virus throughout an area, and if the roosting behavior of crows is different from other species, then once the crow population is severely reduced the WNV activity will decrease and/or the transmission dynamics will change.
We documented the roosting behavior of American crows and northern cardinals and recorded how far they moved between roosts from one night to the next. We also determined whether the roosting behavior of viremic crows differed from that of healthy individuals and how far a viremic crow may move and spread the virus. We investigated whether crow and cardinal populations were experiencing declines and discuss how population declines of crows may affect the transmission dynamics of WNV.
|
MATERIALS AND METHODS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Radio-tagged crows were tracked using vehicle-mounted twin Yagi antennas (Primus Inc., Joliet, IL). Roosts were defined as those areas where crows were located between 11:00 PM and 4:00 AM. We located the roosts of radio-tagged crows as often as possible. On average, we obtained roost location on each individual six of every seven days. Once a roosting crow was located, its position was recorded using Global Positioning Systems. Roosts were generally cohesive and their boundaries easily distinguished. Roosts were considered to be separate if the most peripheral individuals in each roost were at least one kilometer from each other. We calculated the distance between roosts as a straight line distance between the roosting locations of a bird in each roost. The average area used by a viremic crow was calculating as the area of the circle around the point where the crow died to the furthest average roosting location.
Every radio-tagged crow that was found dead was submitted to the University of Illinois, College of Veterinary Medicine Veterinary Diagnostic Laboratory for necropsy. This laboratory conducted a gross necropsy and an immunohistochemical evaluation of major organs to determine if the individual was infected with WNV, and then determined the cause of death based on the gross neocropsy and immunohistochemical results. Because all radio-tagged crows were tracked daily, it is likely that we recovered dead crows within 24 hours of their death. We were assuming that WNV progressed in these birds as has been recorded in crows held in captivity.10–12 Therefore, we documented roosting behavior for the five days prior to death, which encompasses the time when crows are most viremic.
Northern cardinals were studied from July to September 2005 in the same area of Champaign-Urbana. Cardinals were captured using mist nets and potter walk-in traps. Once captured, a 1.2-gram transmitter was attached to their tail. These radio-tagged cardinals were tracked using a hand-held receiver. Roosting locations were determined by locating these radio-tagged birds between 11:00 PM and 4:00 AM an average of six of every seven nights.
The population trends of crows and cardinals were determined for the entire state and Cook County, the county with the historically highest WNV activity (www.idph.state.il.us). We did not use Christmas bird count (CBC),20,23 data because many of the crows in Illinois in winter are migrants. Therefore, the CBC is a poor index of the breeding crow population. We know that at least some of our wintering crows are from northern Michigan because we have radio-tracked two crows from Illinois to areas where they were breeding in northern Michigan (Raim A, unpublished data). We also did not use the breeding bird survey (BBS) because this survey does not sample urban areas, e.g., there are no active BBS routes in Cook County. For determining avian population trends in Illinois, we used the spring bird count (SBC) data, which are available upon request from one of the co-authors (M.P.W.). The SBC has been conducted since 1972 in Illinois, and is conducted by volunteers in all 102 counties on a Saturday that falls between May 4 and May 10. These volunteers record all birds and the number of hours spent by each party in the field. In any given year, there are between 1,100 and 1,600 volunteers in the field. Because of the timing of this census, it provides an ideal index of the breeding crow and cardinal populations in Illinois. Crows in Illinois breed from March through June,24 and cardinals breed from March through August.25 This census samples both rural and urban areas, and provides a good index of birds in areas with high WNV activity such as Cook County where they annually have more than 100 volunteers.
The route-regression technique26 was used to determine the population trend of crows and cardinals. We analyzed the population trend from May 2001, a year before the arrival of WNV, to May 2005, four years after the arrival of WNV. Because decreases in the population could be part of natural fluctuations in the population, we compared the trend after the arrival of WNV to the population trend from 1975 to 2000. In the route-regression technique, a separate trend is calculated for each county (i.e., route) by linear regression. Once county-level estimates of trends are obtained, the overall trend is estimated using the weighted mean of the county trends. Weights are proportional to the total number of individual birds observed in each county. Statistically significant trends were assessed using a one-sample t-test comparing the mean slope of the county population trends to zero.
|
RESULTS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The average distance a crow and a cardinal moved between roosts was significantly different (t = 15.84, degrees of freedom [df] = 65, P < 0.01).
|
). Over the last five days when a crow viremia is highest, they may roost within a mean ± SE area of 20.84 ± 0.79 km2 or within 2.58 km of where the crow was found dead. Although no cardinals died of WNV infection, if we assume viremic cardinals behave similarly to healthy individuals, then they would have roosted within a mean ± area of 0.03 ± 0.01 km2.
|
).
|
|
DISCUSSION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
One of the interesting results of our study on roosting behavior is that viremic crows use more roosts than healthy crows. This may seem counter-intuitive because sick birds would generally seem less likely to move around. The data show that infected birds used the same roosts as healthy birds until one or two days before they died. Because individuals often forage several kilometers from where they roost, it is possible they became sufficiently ill during the day and were unable to return to their usual roosting site that evening. Our radio telemetry data indicate that the night before their death they tend to roost at the nearest roost site. In some situations, it was unclear whether a crow had died before roosting, or died during the night.
In contrast to crow behavior, which may be dispersing WNV by one to four kilometers per night, the roosting behavior of cardinals is more consistent with focal WNV activity. Because cardinals do not move very far between roosts and do not roost communally, there is less possibility of bird-to-bird transmission. However, given their abundance in urban habitats and fidelity to roost sites, cardinals or other passerines like house sparrows may be important reservoir hosts in the focal maintenance or amplification of WNV transmission.
The density of dead crows has been suggested to be a means of predicting the area of highest risk for human cases.4 If crows were the sole reservoir host for WNV, then one may expect that the WNV infection rates in mosquitoes and the number of human cases would decrease as crows die out from WNV-induced mortality. Clearly, this is not the case in Illinois, which had a resurgence in WNV in Cook County in 2005 despite a continuing downward trend in crow abundance. It seems more plausible that crows are important dispersers of WNV throughout an area, especially since the average distance between roosts was slightly greater than a kilometer. Therefore, a dramatic decrease in the crow population might not affect the amplification of WNV, but could change the rate of spread to adjoining areas. That is, the transmission dynamics would then change to reflect the roosting behavior of the major reservoir hosts. Assuming WNV is maintained throughout the winter by infected, overwintering mosquitoes,28–31 then in the absence of infected crows dispersing WNV, transmission activity may be expected to be more focal. Although the roosting behavior for many common species is poorly described, it is generally assumed that species such as American robins, house sparrows, and mourning doves (Zenaida macroura) are roosting at or near their breeding territories. However, house sparrows and mourning doves do occasionally roost in groups, but not to the extent to which crows roost communally. Once a crow population is extirpated or severely reduced in an area, you would expect transmission activity to show a slower spread from one area to the next and instead expect more focal hot spots maintained by the breeding species in that area.
In Illinois, this pattern of more focal activity and less dispersal may be starting to occur. Crows have been decreasing throughout the state but their decrease is most precipitous in Cook County (i.e., Chicago area). In 2002, there was a statewide outbreak of WNV with Cook County having the most human cases. However, we can be reasonably certain that WNV transmission occurred throughout northern, central, and southern counties because of the number of horse cases and detection of pools of WNV-positive Culex (www.idph.state.il.us/ and Table 1). After this initial outbreak, which at the time was the largest outbreak of WNV every recorded, fewer counties reported activity in mosquitoes, dead WNV-positive birds, and human cases. Most of the activity was relatively localized in the northeastern part of the state (i.e., Cook County). In 2005, Illinois experienced a resurgence of WNV; however, unlike the outbreak year of 2002, the activity has remained focal within the state with fewer counties detecting WNV and more than half of the cases being in Cook County (Table 1). Even within Cook County, it was evident there were areas of very high and very low activity (www.idph.state.il.us/ and Novak R, unpublished data).
|
Received January 6, 2006. Accepted for publication March 9, 2006.
Acknowledgments: We thank the University of Illinois at Champaign-Urbana for use of research sites, and Tara Danner and Christine LaPointe for assistance in the field.
Financial support: This study was supported by the Illinois Department of Natural Resources Waste Tire Act, The Centers for Disease Control and Prevention (grant PHS U50/CCU 52051), and Animal Protocol no. 04010.
* Address correspondence to Robert J. Novak, Illinois Natural History Survey, 1816 South Oak Street, Champaign, IL 61820. E-mail: rjnovak{at}uiuc.edu 
Authors’ addresses: Michael P. Ward, Arlo Raim, Richard Lampman, and Robert J. Novak, Illinois Natural History Survey, 1816 South Oak Street, Champaign, IL 61820, Telephone: 217-333-0305, Fax: 217-333-2359, E-mails: mpward{at}uiuc.edu, arain{at}uiuc.edu, rlampman{at}inhs.uiuc.edu, and rjnovak{at}uiuc.edu. Sarah Yaremych-Hamer, Department of Fisheries and Wildlife, Michigan State University, East Lansing, MI 48824, Telephone: 517-353-7981, Fax: 517-432-1699, E-mail: yaremych{at}msu.edu.
Reprint requests: Michael P. Ward, Illinois Natural History Survey, 1816 South Oak Street, Champaign, IL 61820.
|
REFERENCES |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. M. NEMETH, S. BECKETT, E. EDWARDS, K. KLENK, and N. KOMAR AVIAN MORTALITY SURVEILLANCE FOR WEST NILE VIRUS IN COLORADO Am J Trop Med Hyg, March 1, 2007; 76(3): 431 - 437. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
