Potential transmission of West Nile virus in the British Isles: an ecological review of candidate mosquito bridge vectors

Authors

  • J. M. Medlock,

    Corresponding author
    1. Microbial Risk Assessment, Health Protection Agency, Porton Down, Salisbury, Wiltshire and
      §Jolyon Medlock, Microbial Risk Assessment, Health Protection Agency, Centre for Emergency Preparedness and Response, Porton Down, Salisbury, Wiltshire SP4 0JG, U.K. Tel: + 44 (0)1980 619956; Fax: + 44 (0)1980 612731; e-mail: jolyon.medlock@hpa.org.uk
    Search for more papers by this author
  • K. R. Snow,

    1. School of Health and Bioscience, University of East London, London, U.K.
    Search for more papers by this author
  • S. Leach

    1. Microbial Risk Assessment, Health Protection Agency, Porton Down, Salisbury, Wiltshire and
    Search for more papers by this author

  • *

    The British Isles (population 63 million) form a fairly compact archipelago of > 6000 islands (between 49–61°N and 10°W−2°E) spanning ∼600 000 km2 with total land area > 300 000 km2, westward from continental Europe. The two main islands are Great Britain (politically subdivided between the States of England, Scotland and Wales, comprising the United Kingdom) and Ireland (politically divided between the Irish Republic, Eire, and Northern Ireland Province of the United Kingdom). Largest of many subsidiary islands are the Isle of Man, between Britain and Ireland, and the Channel Islands between Britain and France. Until ∼6000 years ago, when seas rose after the last glaciation, Great Britain was joined to continental Europe by a land bridge, now broken by the shallow English Channel ≥ 30 km wide.

§Jolyon Medlock, Microbial Risk Assessment, Health Protection Agency, Centre for Emergency Preparedness and Response, Porton Down, Salisbury, Wiltshire SP4 0JG, U.K. Tel: + 44 (0)1980 619956; Fax: + 44 (0)1980 612731; e-mail: jolyon.medlock@hpa.org.uk

Abstract

Abstract.  West Nile virus (WNV) transmitted by mosquitoes (Diptera: Culicidae) infects various vertebrates, being pathogenic for birds, horses and humans. After its discovery in tropical Africa, sporadic outbreaks of WNV occurred during recent decades in Eurasia, but not the British Isles*. WNV reached New York in 1999 and spread to California by 2003, causing widespread outbreaks of West Nile encephalitis across North America, transmitted by many species of mosquitoes, mainly Culex spp. The periodic reappearance of WNV in parts of continental Europe (from southern France to Romania) gives rise to concern over the possibility of WNV invading the British Isles.

The British Isles have about 30 endemic mosquito species, several with seasonal abundance and other eco-behavioural characteristics predisposing them to serve as potential WNV bridge vectors from birds to humans. These include: the predominantly ornithophilic Culex pipiens L. and its anthropophilic biotype molestus Forskål; tree-hole adapted Anopheles plumbeus Stephens; saltmarsh-adapted Ochlerotatus caspius Pallas, Ocdetritus Haliday and Oc. dorsalis (Meigen); Coquillettidia richiardii Ficalbi, Culiseta annulata Schrank and Cs. morsitans (Theobald) from vegetated freshwater pools; Aedes cinereus Meigen, Oc. cantans Meigen and Ocpunctor Kirby from seasonal woodland pools. Those underlined have been found carrying WNV in other countries (12 species), including the rarer British species Aedes vexans (Meigen), Culex europaeus Ramos et al., Cx. modestus Ficalbi and Oc. sticticus (Meigen) as well as the Anopheles maculipennis Meigen complex (mainly An. atroparvus van Thiel and Anmesseae Falleroni in Britain). Those implicated as key vectors of WNV in Europe are printed bold (four species).

So far there is no proof of any arbovirus transmission by mosquitoes in the British Isles, although antibodies to Sindbis, Tahyna, Usutu and West Nile viruses have been detected in British birds. Neighbouring European countries have enzootic WNV and human infections transmitted by mosquito species that are present in the British Isles. However, except for localized urban infestations of Cxpipiens biotype molestus that can be readily eliminated, there appear to be few situations in the British Isles where humans and livestock are exposed to sustained risks of exposure to potential WNV vectors. Monitoring of mosquitoes and arbovirus surveillance are required to guard the British Isles against WNV outbreaks and introduction of more anthropophilic mosquitoes such as Stegomyia albopicta (Skuse) and Ochlerotatus japonicus (Theobald) that have recently invaded Europe, since they transmit arboviruses elsewhere.

Introduction

Since the 1999 detection of West Nile virus (WNV, Flaviviridae) in the United States, this arbovirus has spread across 47 mainland states (USGS, 2004), resulting in > 16 600 human cases with > 650 deaths recorded in the U.S.A. (by the end of 2004), reaching Canada in 2001 (HC, 2004) and countries of the Caribbean and central America by 2003 (Mexico, 2004; PAHO, 2004). Unlike most other mosquito-borne arboviruses, which tend to be transmitted by few species of mosquitoes, WNV is transmitted by several genera and many species of mosquitoes, having been detected in 60 North American species (CDC, 2005) and at least 75 species in > 10 genera of mosquitoes worldwide (Goodman et al., 2003; Higgs et al., 2004). WNV is primarily enzootic among birds (Malkinson & Banet, 2002) and less frequently mammals, including humans and equines as dead-end hosts (Deubel & Zeller, 2001). Apart from the epizootic of current concern in the Americas (Epstein, 2001), WNV outbreaks occur sporadically in southern Europe, attributed to importation of the virus through migration of infected birds from Africa and involvement of local mosquito populations (Murgue et al., 2002). The importation of arboviruses by migratory birds is not restricted to WNV, with Usutu virus appearing for the first time recently in central Europe (Gratz, 2004a).

The continuing outbreak of WNV in North America has focused attention on the possibility of its emergence in the British Isles. Phylogenetic analysis of the virus from U.S.A. has shown it to be genetically distinct from strains of WNV circulating in parts of Western Europe. This, coupled with the relative abundance of efficient mosquito vectors in North America as compared to the British Isles, suggests that the threat posed by WNV is relatively low in the British Isles (Crook et al., 2002). Some evidence, however, has pointed to the presence of WNV neutralizing antibodies in British resident birds (Buckley et al., 2003) but, as yet, no virus has been isolated from either birds or mosquitoes, and furthermore no human cases of WNV have been reported in the U.K. (Great Britain & Northern Ireland). Even so, a review of viral encephalitis cases in the U.K. by Davison et al. (2003) revealed that 60% of fatalities due to viral encephalitis are of unknown aetiology.

As the British Isles experience climate change (http://www.defra.gov.uk/environment/climatechange/), mosquitoes are likely to flourish better through expected milder winters and wetter summers. Therefore, we need to understand more about the mosquito fauna of the British Isles, monitor the endemic species populations and keep watch for possible importation of exotic mosquitoes and associated diseases. Numerous viruses comprise the West Nile (WN) group in Africa, Eurasia and now America, including the oriental strains of Kunjin (KU) extending to Australasia (Scherret et al., 2001): their propensity for transmission by diverse species of mosquitoes is a cause for concern. Before such a mosquito-borne virus arrives in the British Isles, we should build upon knowledge of the ecology and biology of endemic mosquitoes, assess their relative roles in possible enzootic and human bridge arbovirus cycles, develop a greater understanding of their current distribution and abundance, and assess the possible impacts of British climate change. For these reasons, Higgs et al. (2004) reviewed the mosquito species most likely to become involved in WNV transmission in the U.K. and the present review aims to provide more ecological understanding of endemic mosquitoes of the British Isles and their potential as enzootic and bridge vectors of WNV, to assess the implications of potentially imported exotic species, and to provide an overview of other mosquito-borne arboviruses potentially or apparently occurring in the British Isles.

Mosquitoes of the British Isles

Thirty-three species of mosquitoes have been recorded in the British Isles (Table 1), comprising six species of Anophelinae (genus Anopheles) and 27 species of Culicinae in six genera: Aedes (2), Coquillettidia (1), Culex (4), Culiseta (7), Ochlerotatus (12) and Orthopodomyia (1). Nine of these species have been linked to WNV transmission in other parts of Europe and the U.S.A., namely, Aedes cinereus, Aedes vexans, Anopheles maculipennis, Coquillettidia richiardii, Culex europaeus, Cx. modestus, Cx. pipiens, Ochlerotatus cantans and Oc. caspius (Labuda et al., 1974; Hubalek & Halouzka, 1999; Durand et al., 2002; CDC, 2004), the principal European WNV vectors being Cx. modestus, Cx. pipiens (typical Cx. pipiens sensu  stricto and the biotype molestus, see below) and Cq. richiardii (Higgs et al., 2004). Six more British species of mosquitoes, An. plumbeus, Culiseta annulata, Cs. litorea, Cs. morsitans, Oc. detritus and Oc. punctor, are known to feed on birds and humans, so could be considered as potential bridge vectors. Eleven additional species are known to bite humans but are considered infrequent avian feeders (An. algeriensis, An. atroparvus, An. claviger, An. messeae, Cs. subochrea, Oc.  annulipes, Oc. communis, Oc. dorsalis, Oc. flavescens, Oc. geniculatus and Oc. rusticus) and would therefore pose only a low or unlikely risk of transmitting WNV to humans. Of the others, four species (Cs. longiareolata, Oc. leucomelas, Oc. sticticus, Orthopodomyia pulcripalpis) are so rare as to pose negligible risk in the transmission of WNV. The remaining three species are not known to bite humans: Cx. torrentium feeds exclusively on birds; for Cs. fumipennis no biting on humans or domestic animals has ever been recorded; and for Cs. alaskaensis there is insufficient information on its habits in the British Isles (Marshall, 1938; Cranston et al., 1987; Snow, 1990).

Table 1.  Mosquitoes recorded from the British Isles, and some factors relevant to their vector potential for West Nile virus in Great Britain (England, Scotland, Wales) and Ireland (Eire and Northern Ireland).
SpeciesGreat BritainIrelandOccurrence (Ashe et al., 1991;
Snow et al., 1998)
Implicated in WNV transmission elsewhereOrnithophagic (bird-biting)Anthropophagic (human-biting)Likely bridge vector (readily bites both birds and humans)
  • a

    Nomenclature for Aedini follows Reinert et al. (2004) recognizing 46 genera globally, two of which (Aedes and Ochlerotatus) are represented by species in the British fauna, plus the threat of Stegomyia considered in this study (see Discussion of exotic species).

  • b

    Comprising An. atroparvus Van Thiel 1927, An. daciae(Nicolescu et al., 2004) and An. messeae Falleroni 1926, feeding mainly on livestock (i.e. cattle, horses, pigs, etc.), opportunistically anthropophilic but seldom biting birds; apparently without An. maculipennis Meigen sensu stricto in the British Isles.

  • c

    Mainly typical anautogenous Culex pipiens, plus local pockets of the autogenous biotype molestus.

  • d

    According to Service (1971a) these species were found negative for bird biting.

aAedes (Aedes) cinereus Meigen 1818XXWidespread, patchyXXXX
Aedes (Aedimorphus) vexans(Meigen, 1830)XSporadic reportsXX
Anopheles (Anopheles) algeriensis Theobald 1903XXFew reportsX
Anopheles (Anopheles) claviger Meigen 1804XXWidespreadXd
bAnopheles (Anopheles) maculipennis Meigen 1818 sensu latoXXWidespreadXX
Anopheles (Anopheles) plumbeus Stephens 1828XXWidespreadXXX
Coquillettidia (Coquillettidia) richiardii(Ficalbi 1889)XXWidespreadXXXX
Culex (Barraudius) modestus Ficalbi 1889XFew historical reportsXXXX
cCulex (Culex) pipiens Linnaeus 1758 sensu latoXXWidespread, abundantXXXX
Culex (Culex) pipiens Linnaeus 1758 sensu strictoXXWidespread, abundantXX
Culex (Culex) pipiens biotype molestus Forskål 1775XLocally sporadicXXXX
Culex (Culex) torrentium Martini 1925XAbundant in S. EnglandX 
Culex (Neoculex) europaeus Ramos et al. 2003XWidespread, few recordsXXXX
Culiseta (Allotheobaldia) longiareolata(Macquart 1838)XVery few reports
Culiseta (Culiseta) alaskaensis(Ludlow 1906)XXOccasional reports in north
Culiseta (Culiseta) annulata(Schrank 1776)XXWidespreadXXX
Culiseta (Culicella) fumipennis(Stephens 1825)XWidespread: E, S & W England
Culiseta (Culicella) litorea(Shute 1928)XXWidespread in southXXX
Culiseta (Culicella) morsitans(Theobald 1901)XXWidespread throughoutXXXX
Culiseta (Culiseta) subochrea(Edwards 1921)XXUncommonX
Ochlerotatus (Finlaya) geniculatus(Olivier 1791)XWidespread in EnglandX
Ochlerotatus (Ochlerotatus) annulipes(Meigen 1830)XWidespreadXd
Ochlerotatus (Ochlerotatus) cantans(Meigen 1818)XXWidespreadXXXX
Ochlerotatus (Ochlerotatus) caspius(Pallas 1771)XXLimited in southXXd
Ochlerotatus (Ochlerotatus) communis(DeGeer 1776)XFew recordsX
Ochlerotatus (Ochlerotatus) detritus(Haliday 1833)XXWidespread, patchyXXX
Ochlerotatus (Ochlerotatus) dorsalis(Meigen 1830)XXLocalizedXX
Ochlerotatus (Ochlerotatus) flavescens(Müller 1764)XLocalizedXd
Ochlerotatus (Ochlerotatus) leucomelas(Meigen 1804)XOne report
Ochlerotatus (Ochlerotatus) punctor(Kirby 1837)XXWidespreadXXX
Ochlerotatus (Ochlerotatus) sticticus(Meigen 1838)XFew historical reportsX
Ochlerotatus (Rusticoides) rusticus (Rossi 1790)XXWidespreadX
Orthopodomyia (Orthopodomyia) pulcripalpis Rondani 1872XFew records in S. EnglandX

Arthropods and West Nile virus in other countries

Prior to the outbreak in North America, WNV had been detected in 44 mosquito species from other parts of the world (Higgs et al., 2004). With the virus spreading across North America, so far it has been isolated from at least 75 spp. of mosquitoes, including 60 species in the U.S.A. (CDC, 2005) not all being necessarily involved in transmission. Twelve of these species have been recorded in the British Isles, namely, Aedes cinereus, Ae. vexans, Anopheles maculipennis s.l., Coquillettidia richiardii, Culex europaeus, Cx. modestus, Cx. pipiens, Culiseta morsitans, Ochlerotatus cantans, Oc. caspius, Oc. donsalis and Oc. sticticus (Table 1).

Among other haematophagous arthropods, WNV has been detected in the biting midge Culicoides sonorensis Wirth & Jones (Diptera: Ceratopogonidae) in the U.S.A. (Naugle et al., 2004) and in 10 species of ticks (Higgs et al., 2004) in six genera of Acari: Amblyomma, Argas, Dermacentor, Hyalomma, Ornithodoros, Rhipicephalus, some associated with migratory birds and seabirds in the British Isles (Hoogstraal et al., 1976; Martyn, 1988). Ornithodoros maintains such arboviruses for months, if not years, and can transmit WNV experimentally (Hoogstraal et al., 1976; Lawrie et al., 2004), although the virus does not amplify, so these argasid ticks could serve as reservoirs but not major vectors. Since direct bird-to-bird transmission of WNV should also be considered (Buckley et al., 2003, 2004; Austin et al., 2004; Naowarat & Tang, 2004), the relative roles of representative types of all haematophagous arthropods require further investigation as vectors of WNV, with emphasis on certain Ceratopogonidae and other ornithophilic flies such as some Simuliidae in comparison to mosquitoes.

The Culex pipiens complex in the British Isles

Culex pipiens Linnaeus sensu lato has been repeatedly implicated as an important vector of WNV in continental Europe (Hubalek, 2000) and North America (Apperson et al., 2002; Anderson et al., 2004), habitually biting both humans and birds, so that it serves as a bridge vector of infection from birds to humans. As in continental Europe, British Cx. pipiens populations differ physiologically, being of either the ‘typical’ or the molestus biotype [see footnote, page X], these two forms being morphologically indistinguishable (Marshall, 1938; Cranston et al., 1987; Snow, 1990). Females of the typical form, Culex pipiens L. sensu stricto, are mostly ornithophagic (bird biting) and rarely bite humans in Britain: their adult activity occurs only during warmer months between the equinoxes; females hibernate (diapause) from autumn until spring (mostly in cool humid buildings) whereas males do not overwinter. Females of the typical form are anautogenous (require blood for egg development) and eurygamous (require large open spaces for the mating arena). Aquatic habitats of Cx. pipiens s.s. immature stages are various types of pools and containers of water above the ground outdoors.

By contrast, females of the molestus biotype of Cx. pipiens bite humans avidly (anthropophagy), although molestus females are primarily autogenous (egg production without a first bloodmeal) and stenogamous (capable of mating in confined spaces). Biotype molestus undergoes continuous generations in all months of the year (no hibernation): aquatic habitats of immature stages usually occur underground, such as in flooded basements, from where the emergent females of molestus can readily find humans to bite indoors for production of successive egg batches. Whereas molestus occurs throughout the holarctic range of Cx. pipiens s.s. and in Australia (Drummond, 1951; Vinogradova, 2000), the molestus biotype does not occur among populations of Culex quinquefasciatus Say 1823 (formerly known as Culex fatigans Wiedemann 1828), the tropical counterpart of Cx. pipiens (sometimes ranked as its tropical subspecies Cx. p. quinquefasciatus), that is responsible for urban lymphatic filariasis transmission (White & Nathan, 2002) and a vector of various arboviruses including WNV in Africa (Murgue et al., 2002) and southern States of the U.S.A. (Rutledge et al., 2003). Although Cx. quinquefasciatus has been found alive on board aircraft arriving in Britain from tropical Africa (Curtis & White, 1984), this very anthropophilic species cannot hibernate and would not survive winter in the British Isles. However, Cx. quinquefasciatus and other tropical mosquitoes might import and transmit arboviruses, so disinsection of aircraft coming from endemic countries should be implemented with vigilance against all vector-borne diseases (Gratz et al., 2000).

Fonseca et al. (2004) showed by molecular markers that, in North America, ∼40% of Culex pipiens females have genetical characteristics of hybrids between the two European biotypes (molestus and typical pipiens s.s.): they blood-feed readily on both birds and humans, thereby serving as efficient bridge vectors of WNV. Furthermore, Fonseca et al. (2004) showed that underground molestus in the British Isles bears more genetic resemblance to autogenous populations in southern Europe than to British overground populations of the typical form, thus refuting the possibility that subterranean molestus populations are locally derived from typical pipiens having undergone physiological and behavioural shifts.

Culex pipiens Linnaeus 1758 sensu stricto: typical biotype

Typical Culex pipiens L. sensu stricto populations are ubiquitous throughout the British Isles, breeding in a wide variety of natural and artificial water collections, overwintering as adult females (inseminated but not blood-fed) that hibernate in cool and damp enclosed situations, nourished by body fat developed in early autumn (Marshall, 1938; Onyeka & Boreham, 1987). After breaking diapause in spring, Cx. pipiens females blood-feed and produce eggs, few surviving more than one gonotrophic cycle until the summer generation appears by the end of May (Cranston et al., 1987). Females deposit egg rafts on the water surface of various aquatic habitats such as ponds, marshes, backwaters of streams, hoof prints, pools, ditches and water-containing tree-holes, tanks, butts (Cranston et al., 1987; Snow, 1990) and discarded tyres. The water may be fresh or brackish, clean or foul, with or without emergent or overhanging vegetation. Eggs hatch within a couple of days after oviposition and the larvae reach pupation after 2–8 weeks (Marshall, 1938; Snow, 1990). Rising temperatures promote oviposition and larval development, with immature stages becoming more abundant through summer until August (Cranston et al., 1987). Mating occurs soon after emergence, and both sexes begin feeding on nectar and plant juices, deriving energy from sugars (Snow, 1990), females also seeking bloodmeals from birds, with some records of feeding on frogs, lizards, snakes (Marshall, 1938) and rarely humans (Cranston et al., 1987).

The summer generation of Cx. pipiens adults appears at the end of May, resting on vegetation and walls, indoors as well as outdoors. Adult densities increase rapidly during June, becoming the predominant mosquito species in many parts of the British Isles and remaining abundant until autumn (Cranston et al., 1987). During summer, some females achieve multiple gonotrophic cycles through to September, but these adults do not survive beyond late autumn (Snow, 1990). Adults fly at heights up to 5–6 m according to Service (1971a) with Rothamsted collections in suction traps up to 12 m (Sivell & Harrington, 2004). Flight numbers peak around sunset and, to a lesser degree, sunrise (Service, 1971a). Females emerging after August feed exclusively on plant juices, building up fat reserves ready for hibernation. Inseminated females enter diapause to overwinter and gonotrophic development follows their first bloodmeal in spring. No males survive the winter (Cranston et al., 1987; Snow, 1990).

The onset and persistence of cyclic hibernation is due to photoperiodism, with mosquitoes resisting hibernation in the presence of artificial light (Tate & Vincent, 1936). Females seek suitable hibernation sites (in sheds, stables, animal shelters, hollow trees, lofts, attics and outbuildings) in late summer and early autumn; the numbers plummet during early diapause before stabilizing until January (Sulaiman & Service, 1983; Onyeka & Boreham, 1987). Several factors are important in the decline of adult densities in hibernation sites: the population in a basement studied by Onyeka & Boreham (1987) lost ∼30% due to episodes of warm weather (leading to premature movement and subsequent depletion of fat reserves), and spider predation caused ∼20% further reduction. Pathogenic fungal infections (Entomophthora conglomerate and E. destruens) accounted for 50% reduction of populations in a sealed room free from predation (Sulaiman & Service, 1983). Only ∼10% of mosquitoes entering hibernation survive the winter (Sulaiman & Service, 1983), with re-activation of some in December, peaking around the vernal equinox in March, with the final hibernators departing in early May (Onyeka & Boreham, 1987). An important consideration for Cx. pipiens is its competition with Cx. torrentium (vide infra) where these two sibling species are both abundant and syntopic in most parts of southern England (Gillies & Gubbins, 1982).

Culex pipiens biotype*molestus Forskål 1775

Unlike the typical form of Cx. pipiens, females of which diapause to overwinter in a sheltered hibernation site (vide supra), populations of the biotype molestus do not hibernate: adult males, gonoactive females and aquatic stages occur throughout the year in the British Isles (Cranston et al., 1987). Despite being morphologically identical to the typical form, the biotype molestus differs biologically in many ways, being stenogamous, primarily autogenous and then haematophagic, non-diapausing and with breeding sites usually in flooded underground chambers (Cranston et al., 1987; Snow, 1990). Autogeny is an uncommon trait in mosquitoes, being an inherited character in molestus rather than simply the result of larval nourishment. In natural populations, the proportion of molestus females that lay autogenously may be quite small, but the ability to do so increases markedly in succeeding autogenous generations (Tate & Vincent, 1936), thus permitting molestus to maintain large populations through periods when hosts are not available (Cranston et al., 1987).

The molestus biotype of Cx. pipiens is the only anthropophagic Culex endemic in the British Isles (Snow et al., 1998), apart from a few old records of Cx. modestus Ficalbi (Table 1). Larval sites include cisterns and tanks in buildings, sumps at the bottom of lift shafts (elevators), leakage pools in heated cellars and boiler rooms and in the foundations of buildings (Snow, 1996), where larvae occur perennially (Cranston et al., 1987). In such conditions, molestus breeds continuously, with many generations annually (Snow, 1990). Nearly all U.K. records of molestus are from underground or overbuilt sites, usually in complete or semi-darkness, developmental stages of this mosquito often inhabiting water rich in decaying organic material (e.g. sewage), being capable of tolerating human urine and ammonia (Mattingly, 1951a; Cranston et al., 1987).

Biotype molestus adults are seldom encountered far from their breeding sites, with only limited summer dispersal; the females attack humans voraciously and can even be troublesome biters in winter indoors (White & Chase, 1980; Vinogradova, 2000). Larval populations of molestus are generally not exposed to predation, but larval overcrowding reduces the incidence of autogeny in resultant females (Cranston et al., 1987). Autogenous egg rafts are small (50–80 eggs/raft), whereas blood-fed females produce twice as many eggs (Tate & Vincent, 1936). At 24°C, eggs hatch within 48 h, larvae develop within 12 days, the pupal stage lasts 48 h and autogenous females oviposit 4 days after emergence – so the complete life cycle takes only 3 weeks. Adult females of molestus readily bite birds as well as humans, with avian blood yielding more eggs than human blood (Mattingly, 1951a). Because urban situations provide plentiful habitats for molestus to breed perennially in flooded basements, sewers, subterranean chambers and tunnels, this biotype occurs mainly in British cities, but has also been reported from suburban and rural areas, including Scotland (Fonseca et al., 2004). In central London, where molestus infests the underground railway system (Byrne & Nichols, 1999), these blood-thirsty mosquitoes sometimes reach the streets where they might pose the threat of WNV transmission from birds such as feral pigeons to humans and police horses.

Culex torrentium Martini

Culex torrentium was first recognized in Britain by Mattingly (1951b) and had been present since the early 20th century, as determined from museum specimens (Service, 1968c). Males of Cx. pipiens and Cx. torrentium differ in their terminalia, whereas the females and immature stages of the two species are almost identical morphologically. Intensive studies of these two sympatric species have revealed no striking bionomic differences, except that the hibernation sites of Cx. torrentium remain obscure. Both species are strongly ornithophilic and occupy similar larval habitats, except that Cx. torrentium immatures are more often found in flooded tree-holes (Cranston et al., 1987). Studies on the proportions of larvae and egg rafts belonging to pipiens and torrentium in southern and eastern England revealed that the latter constituted ∼80% of egg rafts (Gillies & Gubbins, 1982), 15–37% of larval counts (Service, 1968c; Gillies & Gubbins, 1982) and 62% of adults reared from larvae and pupae (Jupp, 1979). Until now, Cx. torrentium has not been found in Ireland, although it predominates over Cx. pipiens in peri-domestic habitats of southern and eastern England, and across much of continental Europe to Scandinavia (Becker et al., 2003), but has never been recorded to bite a human.

Other West Nile virus vectors to consider in the British Isles

Because Aedes vexans, Culex europaeus, Cx. modestus and Oc. sticticus are so rare in the British Isles, they can be discounted from eco-epidemiological consideration, despite their WNV vector roles elsewhere (Higgs et al., 2004). For the 1960s outbreak of WNV in the south of France, Cx. modestus was the main bridge vector with Cx. pipiens the enzootic vector (Mouchet et al., 1970; Durand et al., 2002). The only records of Cx. modestus in the British Isles are from 1944 to 1945 along the central south coast (Portsmouth, Gosport and Hayling Island) breeding in fresh and brackish water (Snow et al., 1998). Presumably it was imported inadvertently during wartime, but did not become established. Where Cx. modestus is endemic around the Mediterranean, it bites humans viciously in woodlands and houses, during daytime as well as by night (Rees & Snow, 1992; Becker et al., 2003). In many parts of the world, Aedes vexans is an abundant pest species (Horsfall et al., 1973) implicated in transmission of WNV and several other arboviruses (Becker et al., 2003; Goodman et al., 2003), but it has seldom been found in the British Isles (Rees & Snow, 1995), mostly in south-east England (Mattingly, 1950; Ramsdale & Smith, 1990). Little is known of Ae. vexans ecology and life cycle in the British Isles; elsewhere it is a troublesome biter causing serious annoyance to humans and domestic animals by day and night (Snow, 1990; Becker et al., 2003).

Culex europaeus (Ramos et al. 2003), formerly known as Culex territans Walker 1856 (an American species), is regarded as scarce in the British Isles, although there are reports of this species from scattered localities ranging from the south coast of England to Scotland (Rees & Snow, 1992). Unlike most Culex, which deposit their egg rafts on the water surface, Cx. europaeus oviposits above the waterline, so that larvae enter the water after hatching. Little is known of this species ecology in the British Isles: the main aquatic habitats are probably small, permanent collections of groundwater with emergent vegetation, although container breeding may occur (Snow, 1990). Considering the obscurity of its preferred habitats, Cx. europaeus could be commoner than the few records indicate. Amphibians and reptiles (mainly frogs) appear to be the main food source for Cx. europaeus, plus a few records of it biting humans, other mammals and birds (Cranston et al., 1987).

The remaining five British mosquitoes that are known to transmit WNV elsewhere in Europe warrant specific attention to their biology and ecology in the British Isles. Briefly, Aedes cinereus, Coquillettidia richiardii, Ochlerotatus cantans and Oc. caspius are thought to bite birds as well as mammals (Table 2), although there are remarkably few data on hosts (Table 3): these four species could be potential bridge vectors from birds to humans or livestock. The Anopheles maculipennis complex (comprising mainly An. atroparvus and An. messeae in the British Isles) are not regarded as avian feeders (ornithophagy), so would be unlikely bridge vectors for WNV in the British Isles.

Table 2.  Basic bionomics of candidate mosquito enzootic vectors and human bridge vectors of WNV in the British Isles: vector potential inferred from host-preference(s). A, adults; H, hibernating adults; L, larvae. Months of activity, from January (1) to December (12), in parentheses.
SpeciesAquatic habitatAnnual generationsOverwintering stagePreferred hostsMonthly activityPotential principal enzootic vectorPotential local bridge vector
Aedes cinereusFlooded habitatsUnivoltineEggMainly cattle, also birds and humansL (4–6) A (6–8)X
Anopheles plumbeusTree-holesBivoltine4th instar larvaMammals, inc. humans, also birdsL (1–12) A(4–10)X
Coquillettidia richiardiiPermanent watersUnivoltineLarvaeCattle, humans and birdsL (1–12) A (5–9)X
Culex pipiens typical formPermanent watersMultivoltineAdultsBirdsLA (4–11) H (9–4)X
Culex pipiens biotype molestusUsually undergroundMultivoltineAll stagesHumans, also birdsLA (1–12)X
Culex torrentiumPermanent watersMultivoltineAdultsBirdsLA (4–11) H (9–4)X
Culiseta annulataPermanent watersMultivoltineAll stagesBirds and humansLA (1–12) H (9–4)X
Culiseta litoreaCoastal watersUnivoltine4th instar larvaMainly birds, some humansL (1–12) A (5–9)XX
Culiseta morsitansPermanent watersUnivoltine4th instar larvaMainly birds, some humansL (9–6) A (4–8)XX
Ochlerotatus cantansWoodland poolsUnivoltineEggMainly cattle, also birds and humansL (1–7) A (4–9)X
Ochlerotatus detritusCoastal watersMultivoltine4th instar larvaMainly cattle and humans, some birdsL (1–12) A (3–11)X
Ochlerotatus dorsalisCoastal watersMultivoltine?CattleA (5–9)X
Ochlerotatus punctorWoodland poolsUnivoltineEggs & larvaeMainly cattle, some humans and birdsL (11–8) A (4–10)X
Table 3.  Mosquito hosts in the British Isles (excluding the largely ornithophilic Culex spp.). For two rare species marked *the generalized information comes from continental Europe ( Becker et al., 2003).
Mosquito speciesGenerally accepted hosts
(Cranston et al., 1987; Snow, 1990)
Precipitin test results
(Service, 1969, 1971a)
  • Omitting negative test results: total usually < 100%.

Aedes cinereusMainly cattle, also birds and humansn = 159: bovid 84%, bird 3%, human 3%
*Anopheles algeriensisHumans and other mammals
Anopheles clavigerVarious mammals: readily bites humansn = 20: rabbit 78%, bovid 22%
Anopheles plumbeusHumans and birdsn = 17: bovid 47%, human 29%, bird 18%
Coquillettidia richiardiiMainly large mammals including humans; also birds and amphibiacattle 38%, human 33–77%, bird 8–18%, rabbit 4%
Culiseta annulataHumans, birds, various mammalsn = 27: human 30%, bird 26%, rabbit 11%, pig 4%, other mammals 7%
Culiseta litoreaBirdsbird 85–91%, reptile 3%, human 0–8%
Culiseta morsitansBirdsn = 356: bird 96%, human 1.4%
Culiseta subochreaHumans and domestic animals
Ochlerotatus annulipesHumans and other mammals7 bovid
Ochlerotatus cantansHumans and other mammalsbovid 50%, rabbit 28%, human 4–8%, bird 0.3–3%
Ochlerotatus caspiusHumans and large mammalsn = 21: sheep/goat 71%, bovid 29%
*Ochlerotatus communisHumans and other mammals
Ochlerotatus detritusHumans and large mammalsn = 84: bovid 51%, human 32%, bird 4%
Ochlerotatus dorsalisHumans, various mammalsn = 40: bovid 73%, rabbit 13%, horse 10%, pig 2.5%, human 2.5%
Ochlerotatus flavescensHumans and other large mammalsn = 67: sheep/goat 82%, bovid 18%
Ochlerotatus geniculatusHumans and cattle (presumably other animals in woodlands)bovid 2, human 1
Ochlerotatus punctorHumans and large mammalsn > 200: bovid ∼90%, human 3%, bird 1%
Ochlerotatus rusticusHumans

Anopheles maculipennis Meigen sensu lato

The An. maculipennis complex in Europe comprises at least 10 species (White, 1978; Ramsdale & Snow, 2000a; Nicolescu et al., 2004; Linton et al., 2005b), but An. maculipennis s.s. has not been found in the British Isles (Ramsdale & Snow, 1994b). Two members of the An. maculipennis complex are widespread in Britain, namely An. atroparvus and An. messeae (Rees & Snow, 1990), but their differential distributions are unclear due to the difficulty of separating them morphologically (Snow, 1998). Recently the newly described An. daciae (Nicolescu et al. 2004) has been found in Somerset where it was previously misidentified as An. messeae (Linton et al., 2005a). The aquatic stages of both species prefer relatively clean, permanent, standing or slowly moving waters supporting algal growth or emergent vegetation such as ditches, drains, slow moving rivers, ponds and marshes. Larval stages of these species may occur together, although An. atroparvus tolerates high salinities, exhibiting a coastal and estuarine distribution. Adult females show endophilic and endophagic tendencies, with human habitations preferred historically, but now observed more frequently in animal quarters and seldom feeding on humans (Cranston et al., 1987).

The aquatic stages, males and gonoactive females of An. atroparvus disappear before winter, leaving a generation of nulliparous, inseminated females to hibernate, taking bloodmeals without egg development (gonotrophic dissociation) during winter months, the nutrition serving to maintain energy via fat reserves (Ramsdale & Wilkes, 1985). An. atroparvus prefers the comparative warmth of animal shelters, whereas An. messeae undergoes complete diapause in cool shelters (Marshall, 1938). Surviving females reactivate in March/April, gaining gonotrophic concordance (i.e. eggs develop after each bloodmeal). The spring generation of developmental stages metamorphose by the end of May (Marshall, 1938), giving rise to increasing population densities during June/July with two or three generations per year (Snow, 1990). These sibling species feed on animals, including humans, but apparently not birds to any degree of importance.

Aedes cinereus Meigen

Aedes cinereus has a widespread but patchy distribution across mainland Britain and, where it is locally common, is a troublesome biter of humans (Snow, 1990). The univoltine life table is typical of aedine mosquitoes, with oviposition occurring throughout the summer in areas prone to freshwater flooding. Eggs resist hatching during the autumn, even if immersed, diapausing until the spring, hatching after the 8th-12th soaking in flooded meadows, ponds, ditches and marshes (Service, 1968b; Cranston et al., 1987; Snow, 1990). Larval development is fairly rapid, usually complete by the end of May, with adults on the wing from June until September and the peak of biting during July (Service, 1969). Females blood-feed mainly on large mammals and, although not particularly attracted to humans, readily bites them when available at dusk and nocturnally, and also during the daytime in shaded situations. Host preference studies in England by Service (1971a) found that most bloodmeals of Ae. cinereus were from bovids (134/159), others from humans or birds (Table 3). Baited traps in Sweden showed Ae. cinereus preference for rabbit rather than chicken or frogs (Jaenson, 1990).

Coquillettidia richiardii (Ficalbi)

Coquillettidia richiardii is univoltine (Service, 1977b) and widespread in the British Isles – mainly England and Wales (Snow et al., 1998), but multivoltine in southern Europe (Guille, 1976). Rafts of up to 200 eggs are deposited during May–August, hatching within 2 weeks (Marshall, 1938; Cranston et al., 1987). Larvae develop slowly from one year to the next, surviving ice-bound periods under leaves and mud during winter. The larval respiratory siphon and the pupal trumpets are specialized for piercing aerated stems and roots of aquatic plants (Acorus, Glyceria, Ranunculus, Typha) from which they obtain oxygen. Older larvae and pupae cannot survive without suitable plants, whereas early instars are able to breathe air from the water surface via the siphon (Marshall, 1938), as done by all larval instars of other mosquito genera. First instar larvae occur in June–September, other instars throughout the year. Although some larvae reach 4th instar during late summer and autumn, pupation does not occur until the following spring (Cranston et al., 1987; Snow, 1990). In Britain, Cq. richiardii pupae occur during May–July, staying continuously plugging into plants for air, ascending to the water surface only for adult emergence (Marshall, 1938). Adults are active during May–September, most numerously in July–August (Marshall, 1938; Cranston et al., 1987). Male swarming occurs from mid-June to late August, close to larval habitats, an hour after sunset and just after sunrise (Marshall, 1938; Service, 1969).

Some females of Cq. richiardii are autogenous (Service, 1969; Guille, 1976; Cranston et al., 1987). For bloodmeals they feed mainly on humans and other mammals, sometimes birds (Table 3), indoors and outdoors (Service, 1969, 1971a; Rees & Snow, 1994). Foraging females bite soon after entering a house: bloodmeal ingestion takes about 4 min, followed by resting for a period of up to 18 min until departure from the house (Service, 1971a).

Ochlerotatus cantans (Meigen)

Ochlerotatus cantans is univoltine, widespread and patchily abundant in woods and scrublands across most of the British Isles (Rees & Snow, 1994). Eggs are laid on damp leaf litter of shaded pools as they dry out, during June–September, requiring nearly saturation humidity for embryogenesis, tolerating 90% after 15 days, but needing > 85% relative humidity to survive the winter (Service, 1977a). With good habitat conditions, overwintering egg losses are only 2–8%, plus 1–2% not hatching due to sterility and failure of embryogenesis (Service, 1977a, b). Eggs do not hatch when flooded by autumn rains; they require environmental conditioning before hatching will occur in response to flooding by melted snow or early spring rains (Marshall, 1938; Snow, 1990). If unflooded, some Oc. cantans eggs can remain viable for at least 3 years (Service, 1977a). Factors that determine the termination of diapause are poorly understood, apparently involving gradual decrease of the oxygen concentration in the water caused by microbial activity (Snow, 1990). Egg hatching rates are low when flooded during early weeks of the year: alternate immersion, drying and re-flooding increases the proportion hatching incrementally; by March most eggs hatch after their first immersion (Service, 1977a). From eggs soaked experimentally at the beginning of February, 21% hatched after 1 week, 50% after 2 weeks and 95% by late March. Larval habitats begin drying during May–September, with no eggs hatching after September (Service, 1977a).

By March, the densely shaded, freshwater pools are heavily infested with larvae, which continue to appear over the next four months (Marshall, 1938). Larval development is normally completed by June (Service, 1977a) when pools dry out, with high levels (∼95%) of larval and pupal mortality mostly during the younger instars (Service, 1977a). In addition to the drying out of breeding sites, further mortality is due in part to an iridescent virus, Coelomomyces fungi, various microsporidian infections and arthropod predators (Service & Streett, 1976; Service, 1977b; Cranston et al., 1987). Furthermore, competition for food and space, and toxic chemicals produced during overcrowding further reduce larval populations (Service, 1977a). Larval and pupal development is temperature dependent (140–145 days at 4°C, 65–70 days at 10°C, 20–25 days at 25°C), with most adults emerging during April–May (Service, 1977a), mating occurring within a day or so.

Adult emergence requires 7°C threshold temperature (Birley & Charlwood, 1989), with males emerging before females (by 0–4 days), with an ultimate sex ratio in favour of males (1.2 : 1.0 m : f) – the greater number and earlier emergence allowing the male terminalia to rotate to the functional position and allowing male numbers to build-up in readiness for female emergence (Snow, 1987). Male swarms (usually less than 50 individuals) congregate 1 m above the ground in the early evening awaiting prospective female mates prior to dispersing before sunset in late May to early August (Snow, 1990). Following emergence, both sexes feed on naturally occurring sugar solutions to procure energy for flight and dispersal, with females not taking bloodmeals until they are 3 weeks old (16.8–18 days) in response to levels of lipid reserve (Service, 1977a; Renshaw et al., 1994). Female Oc. cantans can be troublesome biters outdoors (Snow, 1990) with maximum biting occurring in July, 15–20 min after sunset, and remaining on the wing until September (Service, 1969, 1971b, 1977a). Adults rest in deeply shaded vegetation (Cranston et al., 1987) with most flight activity within a few feet of ground level (Service, 1971b). Daily biting density is strongly temperature dependent, with little biting below 12–13°C mean temperatures (Birley & Charlwood, 1989).

According to Service (1971a), the principal hosts of Oc. cantans are cattle (50%) and rabbits (28.1%), plus some feeding on humans (3.8–7.9%), birds (0.3–2.6%) and horses (Renshaw et al., 1994), these proportions reflecting host availability as well as the species preference. In June, 11–15 days elapse between bloodmeals, with female survival rates of 0.963–0.987 daily (Renshaw et al., 1994). Bloodmeal digestion takes 30 days at 4°C, 14 days at 8°C, 5 days at 20°C and 3 days at 30°C, averaging 8–9 days (Service, 1977a). Life expectancy of females averages 66–68 days, some surviving > 100 days (Service, 1977a), despite predation by various dipterans and arachnids (Service, 1973). Females undergo no more than two or three gonotrophic cycles. Oviposition begins in late June or early July and eggs are laid in small groups (mean 32); the rate of embryonic development is temperature-dependent (42 days at 4°C, 22 days at 12°C, 8 days at 20°C, 4 days at 31°C), averaging 14–15 days during June–September at mean temperatures of 13.6–16.1°C (Service, 1977a).

Ochlerotatus caspius (Pallas)

Ochlerotatus caspius has been implicated in human WNV transmission in Israel. It is multivoltine (Marshall, 1938), breeding in saltmarshes and meadows subject to intermittent flooding. Like most temperate aedines, Oc. caspius overwinters in the egg stage, with facultative diapause induced by short photoperiod and cold; in spring the longer days allow inundation to stimulate egg-hatching in April (Cranston et al., 1987). Unlike other European mosquitoes that breed in saltmarshes (e.g. Oc. detritus), no larval stages of Oc. caspius are present between December and March (Marshall, 1938). The aquatic stages are salt-tolerant, commonly found in coastal marshes (Marshall, 1938), indicating a restricted distribution in the British Isles (Snow et al., 1998), although inland records from non-saline waters are not infrequent (Marshall, 1938). Larval growth is rapid (2 weeks), with adults appearing in April. The number of generations in coastal marshes varies according to tidal and rainfall patterns, with adults present in large numbers until October (Marshall, 1938; Cranston et al., 1987). Adults rest in dense, grassy vegetation in the vicinity of larval sites and females bite humans viciously, sometimes entering houses to feed (Rees & Snow, 1996). Along with Oc. detritus they are a cause of widespread annoyance in many low-lying coastal districts of the British Isles, especially on the east (mainly caspius) and south (mainly detritus) coasts (Marshall, 1938; Mattingly, 1950; Rees & Snow, 1996), with some evidence of Oc. flavescens replacing Oc. caspius in the north Kent marshes (Rees & Snow, 1996). Among 21 bloodmeals of Oc. caspius identified by Service (1971a), the majority (15/21) were from ovine hosts, the other females having fed on bovids.

Other mosquitoes of the British Isles that habitually bite both birds and humans

Six more British species of mosquitoes could be considered as potential bridge vectors of WNV because their females blood-feed habitually on birds and humans (Tables 1, 2 and 3). Culiseta annulata is widespread in many habitats and the adults are common in all months of the year (Table 2), the females readily biting humans and birds, both indoors and outdoors; hence this species is a prime candidate bridge vector. Anopheles plumbeus is widespread at low density, biting humans and birds in both urban and rural situations, with endophagic tendencies. Ochlerotatus detritus is a biting nuisance for humans and livestock in many coastal areas, together with Oc. caspius, being predominantly exophagic with limited feeding on birds. Ochlerotatus punctor shares woodland habitats with Oc. cantans and bites both humans and birds. The predominantly avian feeders Cs. morsitans and Cs. litorea (mainly coastal) are prime candidate enzootic vectors that also bite humans and so could also be potential bridge vectors. Summaries of the ecology of these six species follow.

Anopheles plumbeus Stephens

Anopheles plumbeus is almost exclusively a tree-hole mosquito, distributed widely throughout the British Isles, equally prevalent in urban and rural situations wherever there are mature trees with water-filled holes for its aquatic stages (Snow, 1990, 1998); occasionally it develops in containers and groundwater (Marshall, 1938). Eggs are laid just above the water-line in tree-holes, the majority (88%) hatching when first flooded (Service, 1968b). Studies of seasonal activity by Service (1969) suggest that An. plumbeus is bivoltine. Overwintering larvae are able to survive long periods when the water surface is frozen (Cranston et al., 1987). They pupate in early spring and the first generation of adults emerge in April–June (Table 2). These produce a second generation that reaches adulthood by August–September. Anopheles plumbeus females are persistent biters, with peak blood-feeding activity around dusk (Service, 1971b), readily entering houses to bite humans and sometimes transmitting imported malaria (Becker et al., 2003). Host preference studies by Service (1971a) identified An. plumbeus bloodmeals from bovids (8/17), humans (5/17) and birds (3/17).

Culiseta annulata (Schrank)

Culiseta annulata is widespread and abundant in most parts of the British Isles, being one of the commonest mosquitoes in Ireland (Ashe et al., 1991), with the longest biting season of any British mosquito. Larvae of Cs. annulata develop in a wide range of natural and artificial aquatic habitats (ponds, ditches, marshes, cisterns and water butts in gardens) in sunlit and shaded situations; the water may be fresh and clean, polluted or brackish (Snow, 1990). Host preference studies by Service (1969) revealed that Cs. annulata fed as often on birds (7/27) as on humans (8/27), indoors and outdoors, other hosts being rabbits (3/27), pigs (1/27) and other mammals (2/27). Unlike most other mosquito species in the British Isles, gonoactive females of Cs. annulata and all the life stages occur perennially, overwintering without diapause. Adult females pass the winter mainly quiescent, with periods of normal activity and biting if the temperature rises, with no evidence that photoperiod exerts any influence on any stage (Ramsdale & Wilkes, 1985). Reports of mosquito biting during late autumn and early spring in the British Isles are mostly attributable to this species (Snow, 1990).

Culiseta litorea (Shute)

Culiseta litorea is univoltine and essentially coastal across southern England and Ireland. However, its breeding sites are not restricted to saline waters; larvae often occur in freshwater habitats in coastal areas, restricted to open sunlit situations (Snow, 1990; Service, 1994). Studies on Cs. litorea by Service (1994) revealed that females lay ∼110 eggs per batch, 81% of which hatch on first flooding. Fourth instar larvae of Cs. litorea overwinter and pupate in April. During March the pupal duration lasts about 3 weeks, decreasing to ∼7 days in June. Adults appear before mid-May and this species remains on the wing until late September or early October (Service, 1994). Host preference studies revealed that Cs. litorea feeds predominantly on birds: 84.6% on Brownsea Island, Dorset (Service, 1969) and 90.6% at Monks Wood, Hunts (Service, 1994). Bloodmeals from reptiles (2.6%) and humans (8.1%) were found in Cs. litorea from Dorset, but no human feeding by this species was detected at Monks Wood (Service, 1994).

Culiseta morsitans (Theobald)

Culiseta morsitans is widely distributed in England and Wales, and has been recorded in Scotland and Ireland (Cranston et al., 1987). It is univoltine (Shute, 1933), depositing eggs during the summer months in dry hollows or above the level of standing water. Aquatic sites may be fresh or slightly brackish, in ponds, ditches and pools in either open or shaded situations (Snow, 1990). The eggs of Cs. morsitans, laid in batches of about 110, hatch following immersion by autumn and winter rainfall, with the majority (82%) hatching on first flooding (Service, 1994). Fourth instars appear as early as November (Marshall, 1938), able to withstand winter freezing and surviving under ice for long periods (Shute, 1929). In April, pupal development takes 15 days, reducing to 8 days in June, with adults on the wing May–October and peak abundance in June (Service, 1969). Following emergence, the first bloodmeals are sought after 2–3 weeks, with most feeding on birds. Precipitin tests by Service (1969) showed that 96% of bloodmeals are from birds, with only limited feeding on reptiles and mammals including humans (5/356). Service (1994) reported that, during extensive human bait experiments, no Cs. morsitans were caught biting, although precipitin tests gave evidence of human feeding by wild-caught females. Shute (1933) also reported that this species had never been known to attack humans. Furthermore, 70% of females caught in suction traps were at high elevations, conforming with affinity for birds. Culiseta morsitans adults rest in sheltered hollows, cracks of trees and in sheltered parts of derelict buildings but not occupied houses (Service, 1969).

Ochlerotatus detritus (Haliday) and Oc. dorsalis (Meigen)

Ochlerotatus detritus and Oc. dorsalis have wide but patchy distributions in low-lying coastal areas and some inland saline waters of Britain and Ireland (Rees & Snow, 1996), showing preference for salty ground prone to periodic flooding. These species are multivoltine, generally producing a generation following each immersion, provided that sites are not wiped out by the tide. Not all eggs hatch after the first immersion, but most eggs (80%) hatch with two to five soakings (Service, 1968a). Without flooding, eggs may survive for more than a year (Marshall, 1938). Eggs laid in dry areas during summer and autumn may hatch in winter and develop to 4th instar larvae, staying at this stage until March when pupation first occurs (Service, 1968a). Adults appear in March and bite humans persistently throughout the summer until November, mainly attacking outdoors but sometimes entering buildings to feed. In some coastal areas, Oc. detritus causes the greatest biting nuisance of any British mosquito (Snow, 1990). Precipitin tests by Service (1971a) revealed that the majority of Oc. detritus bloodmeals were from bovids (43/84) and humans (27/84), with some from birds (3/84). Results for Oc. dorsalis were similar with more hosts identified: bovid > rabbit > horse > pig > human (Table 3).

Ochlerotatus punctor (Kirby)

Ochlerotatus punctor is one of the commonest mosquitoes in the British Isles, widespread throughout Britain and Ireland and occupying a similar niche to Oc. cantans as a woodland mosquito (Rees & Snow, 1996). Eggs are generally laid in dried-up hollows and ditches or above the level of standing water during the summer months, which later become flooded by autumn and winter rains, with eggs hatching from December onwards in response to deoxygenation of pools by bacterial growth (Fallis & Snow, 1983). Ochlerotatus punctor often breeds in more or less acid waters, in areas of sandy and gravelly soils, such as those lined with dead leaves or Sphagnum or in open heath or woodland where birch and pine predominate (Marshall, 1938). Oc. punctor is univoltine, the first adults emerging in April or May and remaining on the wing until October, with peak numbers recorded in June (Service, 1969) and few recorded after July. Adults spend the day sheltering in dense vegetation (Service, 1971c), biting various mammals including humans (Marshall, 1938) abundantly outdoors during crepuscular hours (Service, 1969), with records of biting in houses (Snow, 1990). Precipitin tests of > 200 bloodmeals fom Oc. punctor revealed that the majority (∼90%) were from bovids; only five had fed on humans and two on birds (Service, 1971a).

Discussion

West Nile virus eco-epidemiology and British mosquitoes

Buckley et al. (2003, 2004) prompted debate about the eco-epidemiology of WNV and other arboviruses in the British Isles, by suggesting that direct bird-to-bird transmission of them is well established enzootically. Whatever becomes of this theory, mosquitoes are potential vectors of WNV in the British Isles. Birds are generally regarded as the key amplifiers of WNV and migratory birds are assumed to be its main natural carriers to receptive areas having efficient vector mosquitoes. Evidence for WNV reservoir hosts, such as birds and other vertebrates or ticks and other invertebrates, is insufficient for discussion but we should not overlook the likelihood of transovarial (vertical) transmission of WNV and other flaviviruses in the vectors (Baqar et al., 1993; Miller et al., 2000; Goddard et al., 2003). That remains speculative without proof of local WNV activity, so this review considers the potential risks of WNV transmission by mosquitoes that actually or might occur in the British Isles, assuming the vector competence of all such mosquitoes.

Among the potential vectors of WNV in the British Isles, four mosquito species (Culiseta litorea and Cs. morsitans, typical Culex pipiens s.s. and Cx. torrentium) are strongly ornithophagic and therefore most likely candidates for enzootic transmission; eight more species of mosquitoes (Table 2) appear sufficiently ornithophilic to be able to transmit WNV among birds where local circumstances are conducive. Also the widespread Culiseta fumipennis may feed on birds and perhaps amphibia, since it is widespread but does not bite humans nor domestic animals (Snow, 1990; Becker et al., 2003).

From a different perspective, by inference from their host-feeding patterns, up to a dozen mosquitoes qualify for consideration as candidate bridge vectors of WNV from birds to humans in the British Isles (Table 1, column 8). Three of these (Coquillettidia richiardii, Culiseta annulata and Culex pipens biotype molestus) feed avidly on both human and avian blood, so they would pose the greatest risk of serving as bridge vectors. The others are either predominantly bird-biters with less tendency to bite humans (Culiseta litorea and Cs. morsitans), predominantly human-biters with less tendency to bite birds (Anopheles plumbeus and Ochlerotatus detritus) or they feed predominantly on other mammals with less tendency to bite birds or humans (Aedes cinereus, Ochlerotatus cantans, Oc. dorsalis and Oc. punctor). For example, Oc. cantans chooses vertebrate hosts opportunistically, usually foraging in meadows and woodlands where the available mammalian hosts are generally not humans, but also being a serious nuisance attacking humans around habitations near woodlands.

Among other mosquitoes in the British Isles, a dozen species are known to attack humans and other mammals but have not been seen to bite birds (Tables 1, 2 and 3), so they are not regarded as potential bridge vectors of WNV, although they might be suited to transmission of other pathogens. For example, the widespread Anopheles atroparvus and An. messeae were vectors of human malaria in the British Isles until the mid-20th century (Marshall, 1938; Cranston et al., 1987). Anopheles algeriensis has been recorded from three localities with calcareous waters: Norfolk, England (Edwards, 1932; Hart, 1954), Anglesey, Wales (Morgan, 1987; Rees & Rees, 1989) and County Clare, Ireland (Ashe et al., 1991). The holarctic Ochlerotatus communis has been reported only twice, from Nottinghamshire in the British midlands (Marshall, 1938) and from Jersey in the British Channel Islands (Lane, 1965). Baited traps in Sweden showed a strong preference of Oc. communis for rabbit rather than chicken or frogs (Jaenson, 1990). The other seven mosquitoes regarded as not ornithophagic (Service, 1971a) comprise Anopheles claviger, Culiseta subochrea, Ochlerotatus annulipes, Oc. dorsalis, Oc. flavescens, Oc. geniculatus and Oc. rusticus (Table 3).

The remaining five endemic mosquitoes are uncommon in the British Isles: Culiseta longiareolata has been recorded on only three occasions (from Epsom, Portsmouth and Brownsea Island: Cranston et al., 1987) and it seems unlikely that it ever became established (Ramsdale & Snow, 1994b); Ochlerotatus leucomelas was found only once in Nottinghamshire in 1919 (Cranston et al., 1987); the anthropophilic Ochlerotatus sticticus has seldom been detected (from the New Forest, Cumbria and central Scotland: Cranston et al., 1987); Orthopodomyia pulcripalpis bites birds and breeds in water-filled tree-holes of mature woodlands in southern England (Cranston et al., 1987; Becker et al., 2003). The snowmelt mosquito Culiseta alaskaensis constitutes a serious pest around the northern holarctic, but is not a nuisance in the British Isles, with only a few records from Scotland and northern England (Cranston et al., 1987).

Exotic mosquitoes

In addition to 33 species of mosquitoes recorded in Great Britain, there are ∼60 more European species that have not been found in the British Isles (Ramsdale & Snow, 1994a, b, 1995; Becker et al., 2003) and there is no particular reason why they would reach these islands. However, three exotic species (Ochlerotatus atropalpus Coquillett, Oc. japonicus Theobald and Stegomyia albopicta Skuse) have recently become established in Western Europe, breeding in tyres and other water containers, raising the prospect that they might be also introduced to the British Isles. These three invasive species have relatively high vector competence for WNV (Turell et al., 2001) and, typically of aedine mosquitoes (Savage & Strickman, 2004), their eggs can remain viable but dormant when desiccated, then hatch when flooded, allowing the species to be transported and introduced to wherever the infested containers are taken. Historically, this capacity also allowed the yellow fever mosquito ‡Stegomyia aegypti (L) to become established almost everywhere that people settled in the tropics and subtropics (Christophers, 1960). Since the mid-20th century, however, St. aegypti has disappeared from southern Europe, apparently because of improved sanitation. Even so, the continued popularity of St. aegypti for laboratory investigations allows it to escape repeatedly in the British Isles where it fails to establish (Cranston et al., 1987).

The Asian tiger mosquito St. albopicta spread around the world during the 20th century (Estrada-Franco & Craig, 1995), mainly through accidental transportation of dormant eggs in tyres (Mitchell, 1995). In its oriental homeland, St. albopicta breeds naturally in flooded tree-holes, leaf axils, cut bamboo stems, coconut shells and the like. Peridomestically, artificial receptacles of water such as buckets, pots and tyres serve as aquatic habitats for immature stages of St. albopicta. Females of St. albopicta have vector competence for many arboviruses, notably dengue, WN and yellow fever (Moore & Mitchell, 1997), but the natural vector status of this species remains enigmatic except where implicated as the only vector, e.g. for dengue outbreaks in Hawaii, the Seychelles and China (Gratz, 2004b). In addition to being a serious nuisance biting pest, the ecology and behaviour of St. albopicta make it apparently suitable as a bridge vector and WNV has been isolated from it occasionally in the U.S.A. but not elsewhere (Goodman et al., 2003).

Stegomyia albopicta was first recorded in Europe in Albania in 1979 (Adhami & Reiter, 1998; Vazeille-Falcoz et al., 1999) and more recently in Italy 1990 (Dalla Pozza & Majori, 1992), France 1999 (Schaffner & Karch, 2000), Belgium 2000 (Schaffner et al., 2004), Montenegro 2001 (Becker et al., 2003), Israel 2001 (Wilamowski & Schnur, 2003), Switzerland 2003 (Flacio et al., 2003) and Spain 2004 (ProMED, 2004). Initial French foci were tyre storage centres in Poitou-Charentes and Normandy where tyres were imported to Europe and St. albopicta became established on the other side of the English Channel (Schaffner & Chouin, 2003). In Italy, all except one of the original foci were linked to a shipment of tyres imported into Genoa from Atlanta, Georgia (Dalla Pozza et al., 1994; Romi, 1995; Romi et al., 1999), where St. albopicta had recently become established in the U.S.A. (Moore & Mitchell, 1997). Already in Italy St. albopicta has joined in transmission of dirofilariasis among dogs, increasing the risks of human infection (Cancrini et al., 2003).

Distribution of St. albopicta is limited by photoperiod, temperature, annual rainfall and humidity (Hawley, 1988). North of 30°N latitude, it is seasonally affected by day-length, tolerating low temperatures by producing eggs that undergo winter diapause. The 10°C cold-month isotherm appears to separate continuously breeding populations from those that overwinter (Mitchell, 1995), with cold-month isotherms between 0°C and 5°C appearing to limit its northerly distribution in Asia, North America and now Europe (Nawrocki & Hawley, 1987; Kobayashi et al., 2002; Toma et al., 2003). Having become established in northern and central Italy (Romi, 1995), the photoperiodic eggs of St. albopicta hatch in early spring (March in Rome) when climatic conditions become more favourable with spring rainfall, 11–12 h of daylight and temperature not falling below 10°C (Toma et al., 2003). Continuous generations occur until the autumn when the reduced photoperiod induces egg diapause (Estrada-Franco & Craig, 1995; Toma et al., 2003). With these parameters in mind, vector biologists suspect that St. albopicta could survive and become established in the British Isles (Burgess, 1995; Ramsdale & Snow 2000b; Snow & Ramsdale, 2002), if it were to be introduced via the tyre trade as happened repeatedly for continental Europe. Whether St. albopicta can expand its range to more northerly latitudes of Europe remains to be seen: this could have public health implications because St. albopicta is a biting nuisance, with vector competence for > 20 arboviruses including dengue, Sindbis, Tahyna and Rift Valley fever (Mitchell, 1995; Gratz, 2004b); this species may contribute to WNV transmission in the U.S.A. (Holick et al., 2002; Kutz et al., 2003; CDC, 2004).

Ochlerotatus (Ochlerotatus) atropalpus is endemic to eastern North America, breeding in freshwater pools of rocky valleys. Females are a persistent pest near breeding sites, having rather limited flight range, and there may be several broods during the year (Carpenter & LaCasse, 1955). In 1996, adults and developing larvae of Oc. atropalpus were discovered in the Veneto province of Italy (Romi et al., 1997) at a depot that regularly imported scrap tyres from the U.S.A. (Minnesota and Texas). Despite insecticidal control measures, larvae were collected a year later from the same site (Romi et al., 1997) and it is therefore considered to be well established, albeit locally. Ochlerotatus atropalpus has long been recognized as a vector of Eastern Equine Encephalitis (EEE) virus. It readily bites humans along with other vertebrates. With a more urban distribution in Italy, or elsewhere in Europe, it could conceivably lead to the transmission of other arboviruses (Snow & Ramsdale, 2002).

Ochlerotatus (Finlaya) japonicus is native to Korea, Japan, Taiwan and southern China (Tanaka, 1979) where the larvae develop in natural and artificial containers. Laird et al. (1994) reported Oc. japonicus importation to New Zealand in tyres from Japan. This species was recently detected in Orne department of northern France (Schaffner et al., 2003), following its 1990s establishment in the U.S.A. (Darsie, 2002), where it now breeds in a diversity of container habitats including rock pools, tyres, buckets, metal cans, bird baths and tree-holes (Andreadis et al., 2001). As a semi-temperate species, Oc. japonicus appears to be well suited for colonization of the British Isles. Ochlerotatus japonicus has been found positive for WNV repeatedly in the U.S.A. (Andreadis et al. 2001; Snow & Ramsdale, 2002) and is a competent carrier of Japanese encephalitis transmitted both vertically and horizontally (Takashima & Rosen, 1989).

Mosquito-borne arboviruses of Northern Europe

In addition to the concerns raised over WNV, several other arboviruses transmitted by mosquitoes have been recorded in parts of Northern Europe, transmitted by species of mosquito endemic to the British Isles. Serological evidence of Usutu virus and Sindbis virus infection of birds (Buckley et al., 2003) and Tahyna virus in small mammals (Chastel et al., 1985) in the U.K. suggests that other arboviruses might be circulating, at least in enzootic cycles. All known arboviruses are listed at http://www.ictvdb.rothamsted.ac.uk

Sindbis virus in humans causes fever, rash and arthritis and is locally known as Ockelbo disease (Sweden), Karellian fever (western Russia) and Pogosta disease (Finland). Although wild and domestic birds are the main vertebrate hosts, human infections are not uncommon. In 1995, this virus caused > 1000 cases involving considerable morbidity in Finland and Sweden; antibodies have been found in human sera from the former Yugoslavia, Italy, Romania and Greece (Lundstrom, 1999). Sindbis is maintained in an enzootic cycle between passerine birds and culicine mosquitoes (Cs. morsitans, Cx. pipiens, Cx. torrentium). Transmission to humans results from the bites of bridge vectors (Ae. cinereus, Oc. cantans, Oc. communis) that feed on both birds and humans when people visit forest areas (Lundstrom, 1999).

Tahyna & Inkoo viruses (both members of the California serogroup) occur in most countries of central, southern (Tahyna) and northern (Inkoo) Europe. Tahyna virus has been isolated from various mosquitoes including Ae. vexans, Ae. cinereus, Cs. annulata, Cx. modestus, Oc. cantans, Oc. caspius and Oc. sticticus (Lundstrom, 1999). Serosurveys of small mammals by Chastel et al. (1985) around Exeter, south-western England, found that Apodemus sylvaticus (wood mouse) and Clethrionomys glareolus (bank vole) gave positive results for Tahyna virus. In France, Inkoo virus antibodies have been found in several mammals, including the red fox, horses and rabbits (Lundstrom, 1999). In Finland, California viruses (probably Inkoo) have been isolated from several mosquitoes: Oc. communis, Oc. punctor and Oc. sticticus (Lundstrom, 1999); seroprevalence rates were found to be high amongst humans, cattle and reindeer in Lapland, northern Finland, with human seroprevalence rates lower in central and southern Finland (Brummer-Korvenkontio & Saikku, 1975). In Scandinavia there is no evidence that California serogroup viruses cause human disease; but in France, the former Czechoslovakia and the former USSR, Tahyna virus is associated with influenza-like symptoms, sometimes accompanied by acute respiratory infection with evidence of CNS infection (Pilaski, 1987).

Mosquito-borne Bunyaviruses circulating in Europe include Batai and Calovo (a strain of Batai). Batai virus has been reported from Austria, the former Czechoslovakia, Finland, Germany, Hungary, Portugal, the former Yugoslavia and the former USSR (Snow, 1991), being transmitted by Ae. cinereus, An. claviger and Oc. communis in northern Europe (Jaenson, 1988, 1990). Calovo occurs in central Europe and the former USSR where the vectors include An. maculipennis s.l., Cq. richiardii and Oc. punctor (Jaenson, 1990). Normally these viruses affect livestock; reports of human infections are rare, although Calovo sometimes causes headache and fever in humans (Snow, 1991). There is no evidence of these viruses occurring in the British Isles.

Other arboviral diseases with human risks include dengue (Mackenzie et al., 2004), another flavivirus, which, within Europe, last occurred in Greece during the 1920s, affecting over one million people. Dengue is the most prevalent arbovirus globally and scores of cases are imported annually to Britain (CDR, 2002) by infected people returning from tropical situations where transmission occurs, mostly due to the vector Stegomyia aegypti (Deubel & Murgue, 2001), which does not survive in the British Isles (vide infra). However, St. albopicta also transmits dengue (Gratz, 2004b) and this might be of concern in future should the climate continue to warm (http://www.defra.gov.uk/environment/climatechange/) and if this vector becomes established in the British Isles (vide supra). Additional arboviruses with multiple vectors, e.g. Semliki Forest and Chikungunya, have been reported in continental Europe at low frequencies, possibly brought from Africa by migratory birds. The recent emergence of Usutu virus in parts of Austria and Hungary illustrates the ability of exotic arboviruses to be imported by migratory birds and become involved in local mosquito transmission cycles (Gratz, 2004a), although there have been no reports of severe disease in humans.

Finally, yellow fever was transmitted in Swansea, Wales, upon the 1865 arrival of a boat from Cuba carrying yellow fever sufferers and presumably the vector (Smith & Gibson, 1986). Over the next few weeks, nearly 30 cases of yellow fever were reported amongst the local residents (Buchanan, 1866). The outbreak spread from the ship into the town during a period of warm summer weather and was probably transmitted by Stegomyia aegypti mosquitoes which, in those days, commonly thrived on-board boats (coming from the tropics) breeding in containers of freshwater. The epidemic was curtailed by the onset of cold weather and putative demise of the exotic mosquitoes (Snow, 1991). Nowadays the introduction of yellow fever and its vector via ships is considered to be extremely unlikely but, conceivably, infected arbovirus vectors such as St. aegypti could be inadvertently imported via aircraft, flying off to transmit locally, as apparently occurred with malaria-infected Anopheles mosquitoes at Gatwick Airport in 1982 (Curtis & White, 1984; Snow, 1991). This justifies appropriate disinsection of aircraft coming from endemic areas (Gratz et al., 2000).

In summary, perhaps 12 species of mosquitoes endemic to the British Isles appear to have the ability, based upon their ecology and biting preferences, to act as potential vectors of arboviral infections to humans. Potential bridge vectors of WNV comprise Ae. cinereus, An. plumbeus, Cq. richiardii, Cs. annulata, Cs. morsitans, Cx. pipiens biotype molestus, Oc. cantans, Oc. caspius, Oc. detritus, Oc. dorsalis, Oc. punctor and Oc. sticticus. In most British ecosystems the mosquito fauna includes more than one of these species, e.g. An. plumbeus and the molestus biotype of Cx. pipiens in urban areas, Ae. cinereus, Oc. cantans and Oc. punctor in woodlands, Oc. caspius and Oc. detritus in coastal marshlands, with Cs. annulata spanning these and other habitats such as freshwater ponds with Cq. richiardii. So far, however, only limited serological evidence of four mosquito-borne arboviruses (Sindbis, Tahyna, Usutu and West Nile viruses) has been reported from the British Isles, whereas neighbouring countries in northern Europe have widespread enzootic transmission and human infections of these and other arboviruses. Most of the mosquito species involved in such transmission are present in the British Isles, so the risks of arbovirus introductions must be recognized and guarded against. Moreover, the increasing spread of exotic mosquito species, coupled with changing climate, further enhance the possibility of arboviral transmission in new locations. With this in mind, the U.K. Health Protection Agency (HPA) in collaboration with the Chartered Institute of Environmental Health (CIEH) have established a ‘Mosquito Watch’ programme (http://www.cieh-npap.org.uk), whereby entomologists and Environmental health officers are encouraged to send mosquito specimens for identification and recording in the ‘Mosquito Watch’ database. This builds on previous recording schemes for British mosquitoes run by the University of East London (http://www.uel.ac.uk/mosquito/ and http://www.dipteristsforum.org.uk) in conjunction with the Biological Records Centre (http://www.brc.ac.uk) based at Monks Wood, Cambridgeshire. Developing greater understanding of mosquito ecology and distributions should assist with predicting the likely risk areas for arbovirus transmission to humans in the British Isles.

Footnotes

  • *

    Harbach et al. (1984) redescribed Culex molestus Forskål 1775 from Egypt and placed it in synonymy under Culex pipiens Linnaeus 1758, without evidence of its reproductive characteristics. Marshall (1938) ascribed the name molestus to the autogenous/stenogamous biotype of Culex pipiens, due to its prevalence around the Mediterranean, giving it priority over Culex pipiens autogenicus (Roubaud, 1935) as the name for this taxon. According to the Rules of Nomenclature (ICZN, 2000), infrasubspecific forms do not require latin names, but molestus remains the usual name for this biotype in the European and Oriental literature (Tanaka, 1979; Cranston et al., 1987; Becker et al., 2003; Vinogradova, 2000), often ranked as subspecies Culex pipiens molestus or Cx. pipiens pipiens var. molestus (e.g. McAbee et al., 2004).

  • Until Reinert et al. (2004) revised the systematics of aedine mosquitoes, Stegomyia albopicta was known as Aedes albopictus, and Stegomyia aegypti was known as Aedes aegypti.

Acknowledgements

This work forms part of Microbial Risk Assessment's Horizon Scanning activities at the Centre for Emergency Preparedness and Response, Health Protection Agency (HPA), Porton Down, into new and (re-)emerging infectious disease issues for the U.K. This work is funded through the HPA Government Grant-in-Aid. The HPA is an executive agency of the Department of Health, and a national organization for England and Wales dedicated to protecting people's health and reducing the impact of infectious diseases, chemical hazards, poisons and radiation hazards.

Ancillary