The vertebrate hosts of 672 blood-engorged Culex quinquefasciatus Say, collected in Harris County, Texas, during 2005, were identified by nucleotide sequencing PCR products of the cytochrome b gene. Analysis revealed that 39.1% had acquired blood from birds, 52.5% from mammals, and 8.3% were mixed avian and mammalian blood meals. Most frequent vertebrate hosts were dog (41.0%), mourning dove (18.3%), domestic cat (8.8%), white-winged dove (4.3%), house sparrow (3.2%), house finch (3.0%), gray catbird (3.0%), and American robin (2.5%). Results are interpreted in conjunction with concurrent avian and mosquito West Nile virus (WNV) surveillance activities in Harris County. We conclude that Cx. quinquefasciatus is an opportunistic feeder and principal mosquito vector of WNV in this metropolitan area; however, transmission by other mosquito species or by other modes of infection, such as ingestion, must account for the high WNV infection rates among local blue jays and American crows.

Harris County, Texas, is an endemic area of St. Louis encephalitis (SLE); and an active surveillance program that monitors SLE virus activity in mosquitoes, birds, and humans has been in place there for the past 28 years. In June of 2002, West Nile (WN) virus appeared in Houston and quickly spread throughout the region. This report describes the results of 12 years of SLE surveillance in Harris County and the contrasting pattern of WN virus activity, when it arrived in 2002. Our data indicate that both SLE and WN viruses can coexist, despite their ecologic, antigenic, and genetic similarities, and that both viruses will probably persist in this geographic region.

St. Louis encephalitis virus (SLEV) has been regularly isolated throughout the Americas since 1933. Previous phylogenetic studies involving 62 isolates have defined seven major lineages (I–VII), further divided into 14 clades. In this study, 28 strains isolated in Texas in 1991 and 2001–2003, and three older, previously unsequenced strains from Jamaica and California were sequenced over the envelope protein gene. The inclusion of these new sequences, and others published since 2001, has allowed better delineation of the previously published SLEV lineages, in particular the clades of lineage II. Phylogenetic analysis of 106 isolates identified 13 clades. All 1991 and 2001–2003 isolates from Nueces, Jefferson and Harris Counties (Texas Gulf Coast) group in clade IIB with other isolates from these counties isolated during the 1980s and 1990s. This lack of evidence for introduction of novel strains into the Texas Gulf Coast over a long period of time is consistent with overwintering of SLEV in this region. Two El Paso isolates, both from 2002, group in clade VA with recent Californian isolates from 1998–2001 and some South American strains with a broad temporal range. Overall, these data are consistent with multiple introductions of SLEV from South America into North America, and provide support for the hypothesis that in most situations, SLEV circulates within a locality, with occasional incursions from other areas. Finally, SLEV has much lower nucleotide (10.1 %) and amino acid variation (2.8 %) than other members of the Japanese encephalitis virus complex (maximum variation 24.6 % nucleotide and 11.8 % amino acid).

Associations between Culex quinquefasciatus, Aedes albopictus and West Nile virus (WNV) activity, temperature, and rainfall in Harris County, Texas 2003-06 are discussed. Human cases were highly correlated to Cx. quinquefasciatus (r = 0.87) and Ae. albopictus (r = 0.78) pools, blue jays (r = 0.83), and Ae. albopictus collected (r = 0.71), but not Cx. quinquefasciatus collected (r = 0.45). Human cases were associated with temperature (r = 0.71), not rainfall (r = 0.29), whereas temperature correlated with Ae. albopictus and Cx. quinquefasciatus collections (r = 0.88 and 0.70, respectively) and Cx. quinqueftsciatus pools (r = 0.75), but not Ae. albopictus pools (r = 0.55). Both species (collections and pools) and blue jays were weakly correlated (r 5 0.41) with rainfall, but blue jays were better correlated with Cx. quinquefasciatus pools (r = 0.87), compared with Ae. albopictus pools (r = 0.67), Ae. albopictus collections (r = 0.69), and Cx. quinquefasciatus collections (r = 0.46). Peak minimum infection rate for Cx. quinquefasciatus (4.55), and Ae. albopictus (4.41) was in August with highest human cases (17.87), blue jays (55.58), and temperature (29.01 degrees C). Between both species, blood meal analysis indicated 68.18% of Cx. quinquefasciatus mammalian hosts were dog, while 22.72% were human, whereas Ae. albopictus had higher human (44.44%) but fewer dog hosts (22.22%). Ten bird species were identified as hosts for Cx. quinquefasciatus, with northern cardinal and blue jay representing 26.66% and 20.00%, respectively. No bird feeding activity was observed in Ae. albopictus. The earliest and latest human blood meal occurred in May (Ae. albopictus) and November (Cx. quinquefasciatus); 66.66% of human host identifications between both species occurred in October-November, after the seasonal human case peak. Based upon our data, WNV activity in both mosquito species warrants further investigation of their individual roles in WNV ecology within this region.

West Nile virus (WNV) was detected in 11 dead birds and two mosquito pools collected in east Texas and southern Louisiana during surveillance studies in the winter of 2003 to 2004. These findings suggest that WNV is active throughout the year in this region of the United States.