BackgroundHost identification is an essential step in studies on the transmission dynamics of vector-borne disease. Nowadays, molecular tools allow the identification of vertebrate hosts to the species level. However, the proportion of successful identifications is variable and may be affected by the quality of the samples and the laboratory protocols. Here, the effect of two of these factors, namely the digestion status of mosquito blood meal and the DNA extraction procedure, on the success of host identification by amplification and sequencing of a fragment of the cytochrome oxidase 1 gene were tested.MethodsMosquitoes collected both outdoors and indoors during 2012 in southern Spain were identified to species level and their blood meal digestion status recorded using the Sella score, a visual estimation of the digestion status of mosquito blood meals. Each mosquito was assigned randomly to one of two DNA extraction procedures: the quick and cheap HotSHOT procedure or the QIAGEN DNeasy Blood and Tissue® kit and their hosts identified by a molecular method.ResultsThree hundred and forty-seven blood-fed mosquitoes belonging to Anopheles atroparvus (n=171), Culex perexiguus (n=84), Culex pipiens (n=43), Culex theileri (n=39), Culex modestus (n=5), Ochlerotatus caspius (n=4), Culiseta sp. (n=1) were included in this study. Overall, hosts were identified from 234 blood meals compromising at least 25 species including mammals, birds and a single reptile. The success of host identification was lower in mosquitoes with an advanced stage of blood meal digestion and for blood meals extracted using the HotSHOT procedure.ConclusionsThe success of host identification decreases with the advanced stage of mosquito blood meal digestion, from 84.5% for recent blood meals to 25.0% for more digested ones. Using the QIAGEN kit, the identification success improved by 17.6%, with larger increases at more advanced stages of blood meal digestion. Availability of blood-fed females used to be very limited for studies of vector ecology, and these results may help to increase the efficiency of blood meal analyses. In addition, results obtained in this study clearly support that the potential malaria vector An. atroparvus feeds on animals located outdoors but use human-made shelters for resting after feeding.