The efficiency of the analyses of host blood meal source differs widely between studies (i.e., 17.5%-92%, see [19, 27]). In this study, the importance of two sources of variation was quantified. Digestion status of blood meals, visually estimated according to the Sella score, strongly affects the success of host identification using DNA sequencing with efficiencies ranging between 84.5% and 25.0% depending of digestion status. Similar results were obtained by  and  based on mosquitoes kept at the laboratory, with a significant decrease in the identification success 30 to 36 hours after feeding . Although the time interval between insect feeding and collection was unknown in this study, for a recently fed individual (Sella stage 2) it may take about one day to reach the Sella stages 3 and 4, and 1 or 2 additional days to reach the Sella stages 5 and 6, respectively . Results from the present study support those from previous studies where authors reported a reduction of the proportion of reactions yielding sequences as the Sella score on field-caught mosquitoes increased [30, 31]. In this study, a significant drop in success of host identification was found for mosquitoes containing a blood meal in an advanced stage of digestion (Sella stages >5), a similar pattern found in mosquitoes from South Carolina . Obviously, including mosquitoes with blood meals in the highest stages of digestion (scored as 5 and 7 according to the Sella´s method), the overall success of host identification may be reduced, and this may partially explain discrepancies between studies in the rate of host identification success. In addition, in this study, using QIAGEN kit for DNA extraction, the success of host identification significantly increased by 10.2-35.3% depending of the blood meal digestion status. The increase in performance was especially important for the mosquitoes with more digested blood meal (scored from 5 to 7 according to the Sella´s method). Using the QIAGEN kit 47% of blood meal sources was identified while only 12% of those extracted using the HotSHOT procedure was identified. This is also a higher percentage of success than those reported by Tuten et al.  where authors, using the DNAzol BD Direct Extraction Kit (Molecular Research Center, Cincinnati, OH, USA), identified 27% (6/22) blood meals from mosquitoes with 5 to 6 Sella´s scores. Consequently, large improvements in blood source determination may be obtained by using more efficient DNA extraction methods. This increase in efficiency is not obtained free as the economic cost of extraction per sample is much higher when using commercial kits, but the extra cost may be worth investing when the number of blood-fed females to analyse is limiting, as used to be the case in most vector ecology studies.
At least 25 vertebrate host species of mosquitoes potentially involved in the transmission of pathogens by mosquitoes have been identified. Anopheles atroparvus showed a clear preference to feed on mammals of different sizes, from rats to horses, than on avian species in spite of the presence of a high diversity and abundance of birds in the studied area, supporting results from previous studies [33, 34]. Curiously, as recently reported, there is no information on the feeding preference of this species to bite indoors or outdoors . Results obtained in this study clearly indicate that this species feed on surrounding animals located outdoors but use human-made shelters for resting after feeding, adding valuable information to current knowledge on the biology of this species . On the other hand, Cx. perexiguus, the second more extensively sampled species in this study, fed on different bird species in addition to mammals and turtles. Its role as bird feeders, as is the case of other Culex species in this study, supports their importance in the transmission of wildlife diseases in Europe, i.e., West Nile and Usutu virus [9, 18–20].