Background

Host feeding patterns describe an important component of vector capacity, i.e., the probability of a vector–host contact [1]. This interaction is essential to understanding pathogen transmission cycles, e.g., to identify potential vector species [2]. Host feeding patterns of mosquitoes are characterized by intrinsic (genetic) and extrinsic (environmental) factors [3,4,5]. Intrinsic factors are considered the main drivers of host preference for mosquito species with a narrow range of host species, e.g., high preference of Culex territans or Uranotaenia unguiculata for amphibians [6, 7], while extrinsic factors are expected to be relevant for species with a broad range of host species, e.g., host availability for Cx. pipiens [8].

It is proposed that specialists evolve when there is a fitness gain achieved by consuming one optimal host compared with feeding on a range of suboptimal hosts [9]. In contrast, generalists are expected to occur in environments with a low probability of host encounter, and the advantage of waiting for an optimal host is weighed against the risk of death prior to blood feeding and reproduction [1]. To understand the transmission cycle of mosquito-borne pathogens, it is important to accurately describe species-specific differences in host feeding patterns, as it enables the classification of mosquito species as enzootic vectors or bridge vectors of a given pathogen, e.g., Cx. torrentium is considered the enzootic vector (bird–mosquito–bird) of Sindbis virus, while Aedes cinereus the bridge vector (bird–mosquito–human) [10].

Misconceptions about mosquito host feeding patterns are deeply rooted in the literature. One prominent example is Cx. pipiens s.s./Cx. torrentium, including the taxa Cx. pipiens biotype pipiens (Cx. pipiens pipiens), Cx. pipiens biotype molestus (Cx. pipiens molestus), the hybrid between both biotypes Cx. pipiens biotype pipiens × molestus (Cx. pipiens pipiens × molestus), and Cx. torrentium. The females cannot be identified by classic morphology [11], but the taxa differ considerably in their ecology [12,13,14,15]. In the literature, Cx. pipiens pipiens and Cx. torrentium are commonly described as ornithophilic/-phagic [13, 16,17,18], while there is no unified definition for this terminology other than feeding “often” or preferring to feed on the respective host group compared with other host groups without a defined threshold [19]. In contrast, Cx. pipiens molestus is predominantly considered mammalophilic/-phagic [20]. The hybrid between both biotypes with an intermediate host feeding pattern is considered to function as bridge vectors for zoonotic diseases in Northern America [21]. In contrast, recent studies from Europe and Asia show opportunistic host feeding patterns for Cx. pipiens s.s./Cx. torrentium with a considerable proportion of mammals, including humans. There might be no taxa-specific association with one host group and the taxa have to be considered both potential enzootic and bridge vectors [22,23,24,25].

Culex pipiens s.s./Cx. torrentium are potential vectors of different mosquito-borne pathogens with a high relevance for veterinary and public health. This also applies to Germany, Moldova, and Iran, which are examined in more detail in the present study. Culex pipiens s.s./Cx. torrentium is widespread in each of the three countries [22, 23], and field-collected specimens are regularly found to be positive for arboviruses as well as their vector competence was confirmed in the laboratory, for example, Usutu virus or West Nile virus [26,27,28,29,30,31,32]. This is also reflected in the published information on the host feeding patterns for the countries, which showed that Cx. pipiens s.s./Cx. torrentium have to be considered potential bridge vectors feeding on birds and mammals, including humans [22, 23]. Nevertheless, although there are several other studies analyzing the host feeding patterns of Cx. pipiens s.l. with more than 20,000 identified blood meals all over the world, many studies did not differentiate between the members of the species complex [33].

Therefore, the aim of this study was to provide comprehensive insight into the host feeding patterns of Cx. pipiens pipiens, Cx. pipiens molestus, Cx. pipiens pipiens × molestus, and Cx. torrentium by (1) analyzing the host attraction of Cx. pipiens pipiens and Cx. torrentium in a host-choice experiment, (2) summarizing the published and unpublished host feeding patterns for specimens collected in field studies over the last decade analyzed with the same laboratory protocols, allowing for a comparability of the results between Germany, Moldova, and Iran, and (3) finally comparing our results on the host feeding patterns of these taxa with those previously described in the globally available literature.

Methods

Experiment on the host attraction of Cx. pipiens pipiens and Cx. torrentium

Culex pipiens s.s./Cx. torrentium were reared from egg rafts collected in Weinheim, Germany (49.54° N, 8.66° E) between May and August 2020 using gravid-trap bins baited with a yeast hay infusion. About 1–5 egg rafts were placed in larval rearing trays (22 cm × 15 cm × 7 cm) containing 1 L of tap water. Larvae were fed daily with a small amount of crushed flake fish food (TetraMin Flakes, Tetra GmbH, Melle, Germany). Larval rearing was conducted at 22–26 °C and 40–60% relative humidity. Emerging adults were maintained in 32.5 cm × 32.5 cm × 32.5 cm screened cages under the same temperature and relative humidity conditions and were daily provided with 10% sucrose solution ad libitum. Females used in the host selection trials emerged 4 days prior and deprived of sucrose solution 12 h prior to testing.

The trials were conducted with two animals: one grey canary (Serinus canaria form domestica) and one house mouse (Mus musculus). In addition, as an attractant that mimics human skin scents, a packet of BG-Sweetscent (Biogents, Regensburg, Germany) was used with 25 ml CO2/min, which is similar to the amount of CO2 emitted by the mouse. The CO2 emission of the canary (9.22 ml CO2/min (SD 1.09) and the mouse (24.82 ml CO2/min (SD 1.64) was previously measured with a CO2 monitor (AIRCO2NTROL 5000, TFA Dostmann, Wertheim-Reicholzheim, Germany). For this purpose, the individual animals were placed in a box (32 × 25 × 37 cm) and the CO2 content was measured before adding the animal and after 10 min. The experiment was repeated three times. A 1.5 m × 1.5 m mesh enclosure was placed inside the laboratory and two lard can traps (25 × 25 × 80 cm) were hung side-by-side separated by one meter [34] (Fig. 1). The lard can traps were constructed from a large tube (⌀ 25 cm) covered at both ends with removable sampling devices with mesh funnels that allowed mosquitoes to enter but prevented them from escaping the tube. A cage with the attractant was placed inside the tube. Trials were performed with the following combinations inserted within the lard can traps: bird–bird, bird–lure, bird–mouse, lure–lure, mouse–lure, and mouse–mouse. The animal or attractant was randomly assigned to one of the lard can traps. Each trial was repeated five times.

Fig. 1
figure 1

© clipart-library, B lard can traps included in the mesh enclosure (Fig. 1); bird pictogram taken from © clipart-library

A Mesh enclosure with two lard can traps each equipped with an animal or attractant, mosquito pictogram taken from

Full size image

Culex pipiens s.s./Cx. torrentium females entered the trap through one of two removable funnels on either end of the trap. The funnels contained a mosquito-proof mesh that prevented direct contact between the animals and mosquitoes. The trials were conducted from 6 pm to 8 am with an average of 122 females (between 43 and 212 females) for each trial, depending on the availability of 4-day-old females. Mosquitoes in the lard can traps and the remaining mosquitoes in the mesh enclosure were removed with a manual mouth sucking aspirator, stored separately in tubes at −20 °C. All specimens were identified as Cx. pipiens pipiens, Cx. pipiens molestus, Cx. pipiens pipiens × molestus, or Cx. torrentium using a molecular DNA typing assay [12].

Host attraction was analyzed using individual binomial generalized linear models (GLM) per combination of hosts and mosquito species. The proportions of host-seeking female mosquitoes per lard can trap (from now on “attraction”) was used as response variable (N = 10 per GLM) and animal/attractant as two-factorial explanatory variable, e.g., “bird” and “mouse.” Mosquitoes that did not enter one of the lard can traps were not considered as host-seeking and were excluded from the statistical analysis.

Analysis of the host feeding patterns of Cx. pipiens s.s./Cx. torrentium collected in Germany, Moldova, and Iran

Our field data on the host feeding patterns of Cx. pipiens s.s./Cx. torrentium combine previously collected data by us during field studies conducted in Germany [22] and Iran [23] and new, unpublished data collected in different sampling campaigns between 2012 and 2022 in Germany and Moldova. All specimens from the already published studies, as well as the newly collected specimens, were analyzed with the same laboratory workflow [22, 23]. This allows for a better comparability between the results from the three countries, for example, polymerase chain reaction (PCR) primers have been shown to have different specificity [35], potentially influencing the sensitivity for different host taxa between different studies. Sampling sites in all of the three countries covered different dominant land-use categories from urban over rural to natural in each of the countries [22, 23], although an analysis of the differences in host feeding patterns between different land-use categories were not in focus of this study, as it was shown to have no statistically significant impact in our previous studies in Germany [22] and Iran [23]. Mosquitoes were collected with pop-up garden bags as artificial resting sites using a hand-held aspirator [36] or within a nationwide mosquito and pathogen surveillance program using CO2-baited Heavy Duty Encephalitis Vector Survey traps (BioQuip Products, Rancho Dominguez, California, USA), Centers for Disease Control miniature light trap (BioQuip Products, Rancho Dominguez, California, USA), and Biogents Sentinel or BG-Pro traps (Biogents, Regensburg, Germany). The collected mosquitoes were left in the trap bags and stored at −20 °C prior to analysis. Each specimen was morphologically identified under permanent cooling [37].

Whole blood-engorged, morphologically identified Cx. pipiens s.s./Cx. torrentium specimens were placed individually into 2 ml tubes and about 20 pieces of 2.0 mm zirconia beads (BioSpec Products, Bartlesville, USA) as well as 1 ml of cell culture medium (high-glucose Dulbecco’s modified Eagle’s medium; Sigma-Aldrich, St. Louis, MO, USA) were added. The homogenization was performed with a TissueLyser or TissueLyser II (Qiagen, Hilden, Germany) for 2 min at 50 oscillations/s. After clarifying by centrifugation for 1 min at 8000 rpm and 4 °C, the suspension was transferred to a new safe-lock tube. DNA was extracted from 200 μl of the homogenate using the KingFisher™ Flex Magnetic Particle Processor with the MagMAX™ Pathogen ribonucleic acid/DNA Kit (both Thermo Fisher Scientific, Waltham, MA USA).

Two primer sets targeting the cytochrome b or 16S rRNA gens were used [38, 39] following the previously published protocol [22, 23]. All amplicons were further processed with Sanger sequencing (LGC Genomics, Berlin, Germany), sequences pre-processed with Geneious® 7.1.9 [40], and finally compared with GenBank sequences (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Host species were determined using a 95% threshold for percentage identity. Using the same template, all morphologically identified Cx. pipiens s.s./Cx. torrentium specimens were identified as Cx. pipiens pipiens, Cx. pipiens molestus, Cx. pipiens pipiens × molestus, or Cx. torrentium using a molecular DNA typing assay [12].

Differences in the proportion for the avian, human, and nonhuman mammalian host feeding groups were evaluated among the three countries by the test of equal or given proportions (prop.test) in R (Version: 4.2.2) [41].

Global literature review on the host feeding patterns of Cx. pipiens s.s./Cx. torrentium

Data on host feeding patterns were extracted for Cx. pipiens pipiens, Cx. pipiens molestus, Cx. pipiens pipiens × molestus, or Cx. torrentium from publications identified in a systematic search on 17 June 2024 using the PubMed database with the following strategy: '(Mosquito*[Title] OR Culici*[Title] OR Aedes[Title] OR Culex[Title] OR Anoph*[Title] OR "west nile virus"[Title]) AND (Blood*[Title] OR meal*[Title] OR feed*[Title] OR host*[Title] OR preference*[Title] OR pattern*[Title] OR forage*[Title])'. The methods were described in detail by Wehmeyer et al. [33]. In short, two researchers independently screened the publications for suitability on the basis of following inclusion criteria: (1) the study was conducted in the field, (2) studies using vertebrate baits were included only if mosquitoes had no direct contact with the host or were collected before biting, and (3) ingested blood was analyzed using serological or molecular methods. Studies that were only based on behavior observation, laboratory-reared mosquitoes, or laboratory-based feeding experiments were excluded. For this publication, studies were included where Cx. pipiens s.s./Cx. torrentium were identified as Cx. pipiens pipiens, Cx. pipiens molestus, Cx. pipiens pipiens × molestus, or Cx. torrentium using a molecular DNA typing assay. All possible information given on mosquito, detected host taxa, and country were collected and merged into a single database. Blood meal hosts were further categorized into the host groups avian, amphibian or reptilian, reptilian, amphibian, mammalian, human, and nonhuman mammalian.

Data analysis

All computational analysis was performed in R (Version: 4.2.2) using the R-Studio IDE (Version: 2022.12.0) [41]. Additionally, functions from the following packages were used for data preparation and visualization: dplyr [42], ggplot2 [43], tidyverse [44], readxl [45], stringr [46], plyr [47], and magrittr [48].

Results

Experiment on the host attraction of Cx. pipiens pipiens and Cx. torrentium

A total of 268 Cx. pipiens pipiens and 350 Cx. torrentium females were used in the experimental trials comparing the proportional attraction for bird versus lure, bird versus mouse, and mouse versus lure. Both species showed a higher mean attraction for birds compared with lure with a mean of 60.3% [95% confidence interval (95% CI) 30.9–89.8%] against 39.7% (95% CI 10.2–69.1%) for Cx. pipiens pipiens and 58.9% (95% CI 38.4–99.4%) against 38.9% (95% CI 7.1–70.8%) for Cx. torrentium. For the trial bird against mouse it was the other way around with a higher mean attraction for mouse against bird with a mean of 53.3% (95% CI 0.7–100.0%) against 77.3% (95% CI 49.1–100.0%) for Cx. pipiens pipiens and 41.7% (95% CI 14.2–69.1%) against 58.3% (95% CI 30.9–85.8%) for Cx. torrentium. No clear pattern regarding the mean values was observed for the trial lure versus mouse. The 95% confidence intervals of mean attraction for the different trials were highly overlapping (Fig. 2) and neither species showed any statistically significant difference for a host or attractant (binomial GLMs, P > 0.05). In addition, no statistical pattern was observed for the same host/attractant in both lard can traps (Additional file 1: Fig. S1).

Fig. 2
figure 2

Mean attraction with 95% confidence interval for host/attractant for Culex pipiens pipiens and Culex torrentium. Numbers on the bottom indicate the total number of specimens collected in the specific lard can trap over five replicates

Full size image

Analysis of the host feeding patterns of Cx. pipiens s.s./Cx. torrentium collected in Germany, Moldova, and Iran

The host species were identified for a total of 931 blood-fed Cx. pipiens pipiens, 29 Cx. torrentium, 18 Cx. pipiens pipiens × molestus, and 14 Cx. pipiens molestus collected in Iran, Moldova, and Germany (Fig. 3). For Cx. pipiens pipiens, blood meals from human (371, 39.8%) and avian hosts (363, 39.0%) were detected in the highest numbers, followed by non-mammalian hosts detected with 191 blood meals (20.5%) and 4 amphibian blood meals (0.4%). Blood meals of Cx. torrentium were dominated by birds (14, 48.3%) and humans (12, 41.4%), while only 3 blood meals (10.3%) were observed from nonhuman mammalian taxa. Culex pipiens pipiens × molestus fed on humans (8, 44.4%) and showed equal proportions of avian and non-human mammalian blood meals (5, 27.8%). Finally, for Cx. pipiens molestus, blood meals from human (5, 35.7%) and non-human mammals (5, 35.7%) were equally frequently detected, shortly followed by avian hosts (4, 28.6%).

Fig. 3
figure 3

Proportion of host groups detected for Cx. pipiens molestus, Cx. pipiens pipiens, Cx. pipiens pipiens × molestus, and Cx. torrentium. Data collected in our studies (left), data from literature (middle), and both datasets merged (right). Numbers in the bar indicate the number of blood meals per taxon and dataset. The host group “mammalian” is used if studies do not identify the mammalian species

Full size image

As demonstrated above, a high prevalence of humans is evident for all four studied Culex taxa (> 35%, Fig. 4). Focusing exclusively on Cx. pipiens pipiens with a sufficient sample size, further frequent host taxa were Bos taurus (122 blood meals, 13.1% of all blood meals for this taxon), Columba palumbus (68, 7.3%), Anas spp. (62, 6.7%), Turdus merula (54, 5.8%), and Gallus gallus (44, 4.7%). The other blood meals (210, 22.6%) were distributed over many less frequent hosts dominated by different bird species and domestic animals (e.g., Canis lupus, Felis catus). Comparing the host feeding patterns for the three countries in comparison with the remaining two, a significant lower proportion of nonhuman mammals was observed for Germany (Germany versus Iran: χ2 = 33.1, df = 1, P < 0.001; Germany versus Moldova: χ2 = 6.3, df = 1, P < 0.012; Iran versus Moldova: χ2 = 0.27, df = 1, P = 0.6), while we found lower proportions of humans in Moldova (Germany versus Iran: χ2 = 2.7, df = 1, P = 0.09; Germany versus Moldova: χ2 = 13.2, df = 1, P < 0.001; Iran versus Moldova: χ2 = 18.8, df = 1, P < 0.001) and lower proportions of birds in Iran (Germany versus Iran: χ2 = 42.1, df = 1, P < 0.001; Germany versus Moldova: χ2 = 2.8, df = 1, P < 0.09; Iran versus Moldova: χ2 = 29.7, df = 1, P < 0.001).

Fig. 4
figure 4

Number of blood meals per host taxon detected in our studies for Cx. pipiens molestus, Cx. pipiens pipiens, Cx. pipiens pipiens × molestus, and Cx. torrentium

Full size image

For 41 Cx. pipiens pipiens specimens (4.4%), two different hosts were detected: 35 mixed blood meals with human and avian blood, 3 with avian and nonhuman mammalian blood, 2 specimens fed on a human and a nonhuman mammal, and 1 specimen contained blood of a bird and an amphibian. One Cx. torrentium specimen (3.4%) contained blood from Homo sapiens and Sus scrofa.

Global literature review on the host feeding patterns of Cx. pipiens s.s./Cx. torrentium

We found a total of 23 publications on host feeding patterns that used molecular assays to differentiate Cx. pipiens pipiens, Cx. pipiens molestus, Cx. pipiens pipiens × molestus, and Cx. torrentium (5 × USA [49,50,51,52,53]; 4 × Japan [54,55,56,57] [50,51,52, 54,55,56]; 3 × Spain [25, 58, 59]; 2 × each for Australia, Portugal, and UK [17, 60,61,62,63,64]; and 1 × each for Argentina, Iran, the Netherlands, Romania, and Russia [18, 65,66,67,68]). When this dataset was merged with our dataset, 1872 identified blood meals were available for Cx. pipiens pipiens, 460 for Cx. pipiens molestus, and 130 for Cx. pipiens pipiens × molestus (Fig. 3). No additional data from the literature were available for Cx. torrentium. Compared with the new data presented in this study for Germany, Iran, and Moldova with blood meals from birds < 50%, the three Cx. pipiens taxa in the merged dataset had more than 50% blood meals from birds, while human and mammalian species each had less than 30%.

Results from the different countries were heterogeneous. Studies from Romania, the USA, and Portugal showed that Cx. pipiens pipiens predominantly fed on birds, with up to 95.5% (Fig. 5). In contrast, higher proportions of mammalian taxa were observed for the newly collected data from Moldova and Germany (42.7% and 35.5%, respectively), and even reached 64.9% and 75.8% in the Netherlands and Iran, respectively. Similarly, low proportions of mammalian hosts were observed for Cx. pipiens molestus in the USA, Spain, Japan, and Portugal (< 25%); around half of the feeds in Germany, Australia, and Romania; and a high proportion of 68% in Argentina. The few specimens from Iran and Moldova did not contain any avian blood. For Cx. pipiens pipiens × molestus, a dominance of mammals was found for Germany, Iran, the Netherlands, and Romania (> 50%); less than 50% for Portugal and Spain; and only blood meals from birds in the USA.

Fig. 5
figure 5

Proportion of host groups for Cx. pipiens molestus, Cx. pipiens pipiens, Cx. pipiens pipiens × molestus, and Cx. torrentium per country. Data combined blood meals collected by us (Germany, Iran, Moldova) and data from the literature. Numbers in the bar indicate the number of blood meals per taxon and country. The host group “mammalian” is used if studies do not identify the mammalian species

Full size image

Discussion

Due to their wide distribution, abundance, and vector competence for WNV, USUV, or SINV, Culex pipiens pipiens, Cx. pipiens molestus, and Cx. torrentium are potentially important vectors of arboviruses in Europe [26,27,28,29,30]. The transmission cycles promoted by these vectors are shaped by their host-feeding patterns, i.e., maintaining enzootic cycles within one host group (e.g., birds) or leading to a spill-over from one host group to another.

We did not observe a significant attraction for mouse, grey canary, or human lure for Cx. pipiens pipiens and Cx. torrentium. In similar experiments conducted in the USA, Cx. pipiens pipiens showed a significant attraction for birds against mammals [69, 70]. For the USA it is especially discussed that hybridization between Cx. pipiens pipiens and Cx. pipiens molestus is the driver of host attraction with intermediate host acceptance for the hybrid taxon [70]. However, we did not find any differences in the host attraction between Cx. pipiens pipiens and Cx. torrentium either, which do not hybridize.

Host feeding patterns can differ from host choice experiments under laboratory conditions, that is, they are expected to depend on the availability and abundance of the hosts [8]. Many studies have been conducted worldwide to identify the blood hosts of more than 20,000 Cx. pipiens specimens [33], but only a few have differentiated the bioforms of Cx. pipiens s.s., and none included Cx. torrentium. Nevertheless, in the literature, Cx. pipiens pipiens is regularly referred to as ornithophilic/-phagic, whereas Cx. pipiens molestus is described as mammalophilic/-phagic or anthropophilic/-phagic [16,17,18, 71]. Unfortunately this terminology is not based on a standardized classification and is generally used without a clear definition [19].

Studies from the literature differentiating Cx. pipiens s.s./Cx. torrentium were collated here and showed that Cx. pipiens pipiens fed predominantly on avian hosts. Much less data were available for Cx. pipiens molestus and Cx. pipiens pipiens × molestus, but showed a similar pattern with a high proportion of birds. No data were available for Cx. torrentium. Nevertheless, there were considerable differences between the countries, with some combinations of countries and taxa reaching more than 62% mammalian hosts, for example, Cx. pipiens pipiens collected in the Netherlands [67] and Cx. pipiens molestus collected in Argentina. Additionally, for the field-collected specimens analyzed in our laboratory, a broad host use was observed with up to 50% mammalian hosts. The reasons for these differences can be manifold. First, only very few studies differentiated the Cx. pipiens s.s./Cx. torrentium. Worldwide, more than 20,000 undifferentiated Cx. pipiens specimens were analyzed and revealed a broad host feeding pattern with one-third of the blood meals from each human, avian, and nonhuman mammalian host [33]. Our studies on the host feeding patterns in Germany, Iran, and Moldova increased the total number of available taxa-specific information on the host feeding patterns of Cx. pipiens s.s./Cx. torrentium by two-thirds. Another factor might be the species identification of the different Cx. pipiens s.s./Cx. torrentium taxa, that is, Cx. pipiens s.s. host attraction is considered to be the result of genetic introgressive hybridization between Cx. pipiens pipiens and Cx. pipiens molestus populations [25]. In addition, host availability is often assumed to drive the host feeding patterns observed in the field [8], but this information is mostly not collected in the field. Our data from Germany, Iran, and Moldova analyzed with the same laboratory workflow showed statistically significant differences for the proportions of the different host groups, e.g., lower proportion of nonhuman mammals for Germany or lower proportion of birds for Iran. However, the underlying drivers of these differences remain unclear and need further evaluation in further work. Our previous studies in Germany and Iran showed that land-use as most obvious driver might not explain these differences in host feeding patterns [22, 23].

The birds mainly detected in blood meals of Cx. pipiens pipiens belonged especially to the species Gallus gallus Columba palumbus, Hirundo rustica,, and Turdus merula. The latter was also present in the feeds of Cx. pipiens molestus Cx. pipiens pipiens × molestus and dominated the feeds of Cx. torrentium. Of these bird species, especially the blackbird Turdus merula in particular is known to be part of the transmission cycle of WNV and USUV in Europe, as it was found to die in large numbers during USUV outbreaks [72,73,74]. At the same time, we observed considerable proportions of human hosts for each Culex taxon, highlighting their potential role as enzootic and bridge vectors.

In the field-collected Culex specimens analyzed in our laboratory, mixed blood meals were detected in 41 Cx. pipiens pipiens and one Cx. torrentium specimen. Up to now, only a few mixed blood meals have been described in the literature, for example, for Cx. pipiens pipiens or Cx. pipiens molestus [17, 49]. The detection of mixed blood meals is interesting information, as it is evidence of the transmission potential transmission risk between two host species. However, the frequency of mixed blood meals must be interpreted with caution. Generally, gel PCRs with subsequent Sanger sequencing were used to identify the blood meal hosts. Different primers have been shown to have different specificity [35], potentially influencing the sensitivity for different host taxa. The presence of gene fragments of two or more hosts could lead to overlapping signals after sequencing, which are difficult to distinguish from low-quality signals, for example, requiring advanced techniques using next-generation sequencing [75]. Thus, actual amounts of specimens with ingested blood of more than one host could be higher than observed.

Conclusions

Cx. pipiens pipiens, Cx. pipiens molestus, and Cx. torrentium were found to feed with a significant proportion on each avian, human and nonhuman mammalian host. Thus, the classification of Cx. pipiens pipiens and Cx. pipiens molestus as strictly ornithophilic/-phagic and anthropo- or mammalophilic, respectively, should be reconsidered. The broad host range of these taxa combined with a high vector competence suggests a high relevance as both enzootic and bridge vectors in the transmission cycles of various mosquito-borne pathogens, for example, WNV, USUV, and SINV [26,27,28,29,30]. At the same time, we observed significant differences between data collected from different countries. Future studies especially should focus on the underlying intrinsic and extrinsic factors, e.g., the influence of population genetics, host availability, or general environmental conditions on the host feeding patterns.