The Comparative Effectiveness of Adult Mosquito Sampling Methods Dealing With Odor-baited Resting Box Traps in a Malaria-endemic Area in Southern Iran

Background Iran is under threat of a potential outbreak of mosquito-borne diseases, such as malaria, Dengue fever, Chikungunya, and Zika. The study aimed to determine the eciency and sustainability of some adult mosquito sampling methods for designing effective entomological surveillance systems in a malaria endemic area. Methods Different rates of tap water, sugar, and yeast (Saccharomyces cerevisiae) were mixed to provide CO 2 . Anopheles stephensi and Culex pipiens Bandar Abbas strains were reared in the insectary and used for tests. On a lab scale, CO 2 orientation experiments were performed using a Y-tube olfactometer on the insectary mosquito strains. In the eld trial, human landing catches (HLC), articial pit shelter (APS), CO 2 -baited trap (CO 2 -BT), human and cow odor baited resting boxes (HOBT, COBT), cow urine baited trap (CUBT), and colored un-baited box (UB) were studied in Bandar Abbas, Hormozgan Province, south of Iran. Mean densities of An. stephensi and Cx. pipiens insectary strains, which oriented to CO 2 as owrate of 170, and 300 mL/minute was signicantly higher compared to the control group (P < 0.05). The black un-baited inbox resting traps signicantly more collected wild mosquito species compared to others colored UB (P (cid:0) 0.05). s.l.(14.21%), 3.63%). Signicant differences were 0.009), CBBT (P=0.007), HLC (P=0.005), HBBT (P 0.005), UBT CO 2 most collected

The determination of host preference is a critical issue in the control and surveillance of blood-sucking mosquitoes; hence, we researched to extend our knowledge about the collection of mosquitoes and their response to attractants. Present research aimed at the evaluation of odor-baited resting inbox traps by testing their in uence on host-seeking behaviors of mosquitoes under laboratory, and eld conditions.
Rearing the species Anopheles stephensi and Cx. pipiens Bandar Abbas strains have been reared in the insectary of Bandar-Abbas Health Training and at average temperature (25 ± 2°C), relative humidity (70 ± 10 %), and photoperiod (10L:14D). CO 2 production and following lab scale trial in the insectarium We combined different rates of tap water, sugar, and yeast (Saccharomyces cerevisiae) to supply CO 2 in a plastic bottle (4 L) containing 2.5 L of water, based on literature [25]. In this approach, 6 concentrations of carbon dioxide had been provided fresh every day. A standard manual bubble owmeter, Sigma-Aldrich model is used to measure the amount of CO 2 . (i.e. 170, 300, 190, 250, 340, and 360 mL/min) ( Table 1).
The CO 2 orientation experiments were performed using a Y-tube olfactometer, as described by Guha et al. (2014) [26] with few modi cations. CO 2 was released inside the treatment arm of the Y-tube olfactometer. At the same time, nylon perforated mesh was placed in the control arm of the Y-tube olfactometer to provide the mosquitoes with free access to the external environment. The temperature and relative humidity of the experiments was maintained at 25 ± 2°C and 70 ± 10%, respectively. Five replicates of each 5-7-day old adult non-blood fed female, An. stephensi and Cx. pipiens mosquitoes were exposed for each concentration and the run time for each experiment was 5 min. For each replicate, 20-25 female mosquitoes were released into mosquito introduction ports at the stem of the Y-tube olfactometer and the acclimatization time was 5 min. The air ow was maintained at 1 L/min during the experiment by a pressure regulator, which was described by Seenivasagan et al., (2012) [27]. The number of mosquitoes was counted and then recorded in each Y-tube olfactometer arm.  (Fig. 1). Mosquitoes were collected once a month using this method from each locality with a suction tube in the morning from 5:30 am to 7:30 am, released to well-labeled paper cups. The captured specimens were transported to the laboratory and identi ed to species based on a valid diagnostic key of Iran at the species level [29].

Human landing catches (HLC)
This process was carried out on three young volunteers at each collection site once a month, from 5 p.m. to 6 a.m. [28]. Mosquitoes seeking for hosts were collected by the aspirator and sorted by one hour intervals, well released into paper cups, labeled, transported to the laboratory, then identi ed based on a valid diagnostic key of Iran at the species level [29].
Odor-baited resting boxes Trap (OBT) Mosquitoes attracted to odor hosts were collected at each locality site, once a month using black resting boxes (30 × 30 ×30 cm). Black cotton cloth material (30 × 30 cm) impregnated by sweat of the cow body (head, feet, and back) was used inside the resting boxes (RBs) as the mosquitos' attractant (Fig. 2). The RBs had been set in the eld at dusk and collected in the early morning. Mosquitoes were collected by the aspirator, transported to the laboratory, counted, and identi ed [29]. The same method was performed for boxes baited with human body odor (HOBT), cow odor baited resting boxes (COBT), and cow urine baited trap (CUBT). Sweat impregnated cloth has been discarded after each eld trial, and the new odor-baited fabric has been used for similar experiments. Similar inbox size was used as treated with carbon dioxide generated ow rates as 170 mL/min (Table1).

Un-baited box (UB)
The similar black, blue, green, red and white cotton cloth materials used in RBs (30 × 30 ×30 cm) without any odor as the un-baited box.

Data analysis
The obtained data were entered into SPSS 19.0 (SPSS, Inc., Chicago, IL, USA) and were analyzed by One-Way ANOVA following the Tukey test (post-hoc analysis) to determine the mean differences. P-values of ≤ 0.05 were selected as a point of signi cant difference.

Results
Attraction of mosquitoes to different CO 2 concentrations Mean densities of mosquito oriented to CO 2 treatment under the laboratory condition was signi cantly higher compared with the control and this was observed among the two combinations of sugar and yeast cultures (that mentioned in Table 1) in test numbers 1 and 2 (P < 0.05) ( Table 2). However, when the e cacy of the different mixtures of sugar and yeast in attracting mosquitoes was compared, more orientation was found in the control group than treatment in tests No. 3 to 6. Therefore, the most effective mixture should be of the lower concentration of 17.5 g of yeast plus 250 g of sugar plus 2.5-liter water, to release of CO 2 at a ow rate of 170 mL/minute (Table1). Table 3 shows the number and densities of mosquitoes collected in un-baited inbox traps with different colors in the eld. A total of 217 mosquitoes, belonging to 3 genera and 7 species, were collected during 6 months. Anopheles stephensi was the most frequent species (36.8%) followed by Cx. pipiens (16.1), An. dthali (13.4), An. uviatilis s.l. (13.4%), Cx. theileri (9.2%), An. sergentii (4.6%), and Ae. caspius (6.5%). Mean densities of female mosquitoes collected were signi cantly higher compared to males except for An. sergentii (P < 0.05) ( Table 3). Table 3 Mosquito number and densities within Un-baited traps.   Table  3). The black un baited RBs were collected both sexes, and the most physiological stages of the females were blood-fed females followed by unfed, semi gravid plus gravid respectively.
The number of gravid and semi gravid specimens of An. uviatilis s.l., An. suprpictus s.l. and An. dthali was higher than the number of unfed and blood-fed, which proportion of gravid plus semi gravid to unfed plus fed was more than 1. In contrast, the number of gravid and semi gravid of An. stephensi, and An. sergentii, was lower than that of unfed and blood-fed ones, which that proportion was less than 1.

Human landing catches
A total of 541 adult mosquitoes were collected comprising An. stephensi, An. dthali, An. uviatilis s.l., An. pulcherrimus, An. sergentii, An. culicifacies s.l., Cx. pipiens, Cx. sitiens, Cx. theileri and Ae. caspius. Culex pipiens (29.94%) and Aedes caspius (27.35%) were the most prevalent species in this method. Anopheles species were more collected during the second part of the night (9 pm -01 am) except An. sergentii, while the peak-biting activity of culicinae species, and Ae. caspius was in the rst third of the night from 1700 to 2100 hours (P=.0001) ( Table 4).  Table 5). The observation of the abdominal condition of the females is presented in Table 6. Overall, the most of mosquitoes collected in the CUBT were un fed, followed by blood-fed, semi gravid plus gravid respectively. Anopheles stephensi was most collected in October, December, and April respectively while the other species have been abundant in autumn.

Sampling data of control baited trap
The control traps have been concurrently set up in the eld for each test. The most collected species were Cx. pipiens (53.53%) followed by An. stephensi (21.21%) ( Table 5). The abdominal condition demonstrated that more fed females of An. stephensi and Cx. pipiens collected than gravid and semi gravid (P=0.0001).
Cow odor-baited box trap (CBBT) Seven species had been collected by cow odor-bated box traps ( Table 5). Out of 391 individual mosquitoes, 36.82% had been identi ed as unfed females. Culex pipiens were most collected in April-March and December.      (17), and An. uviatilis s.l. 19 (14%) ( Table 7). The number and percentage of unfed females of mosquitoes collected by this sampling method were higher than that of blood-fed followed by semi gravid and gravid (P=0.024).  Aedes caspius was most frequent in HLC followed by CUBT, CBBT, and HBBT. There was no collection of this species by CO 2 -BT, UBT, and APS. There were signi cant differences among the number of Ae. caspius collected by HLC, CUBT (P = 0.008), CBBT (P=0.007), and HBBT (P = 0.001).

Discussion
The useful mosquito trap was to be effective, non-toxic, and covers a wide area. The present investigation revealed that the attraction pattern of four anophelinae, two culicinae, and one aedeniae mosquitoes was varied and also highly depended on the type of the traps.
Anopheles stephensi is the main malaria vector in Iran, and three biological forms i.e. type, mysorensis, and intermediate of the species have been reported in the Hormozgan Province [30]. This study demonstrated that the cow urine-baited resting traps have been more attractive to An. stephensi than either odorbaited boxes or other methods. This observation was consistent with that of Kweka et al [23,31], who reported a higher density of An.arabiensis in CUBT followed by CBBT, and UBT respectively. The e ciency of CUBT is dependent on the place of the trap in indoors and outdoors. In our study, CUBT has a longer desirable effect on An. stephensi, which tend to rest inside the houses and shelter. On the other hand, the proportion of gravid and semi gravid of An. stephensi species collected by CUBT was lower than that of unfed and blood-fed ones. It seems that this species blood meal on both humans and cows rests in shelters to complete their gonotrophic cycle. Despite the widespread distribution of LLINs and ITNs among residents in southern Iran, most of them do not use a mosquito net when they sleep. This nding could be highlighted for programs of vector control such as IRS (indoor residual spraying), which impact on endophilic vector species as well as health promotion through the distribution of ITNs and LLINS. This nding agrees with a study performed on An stephensi in the south and southeastern Iran [32] and also control of malaria program in Africa [33].
Anopheles uviatilis s.l. is the malaria vector in Iran and only species T has been reported from Iran [34,35]. So far, at least three sibling species S, T, and U have been reported in the world [36]. In our study, An. uviatilis s.l. more collected by HLC followed by APS, CUBT, CBBT, CO2-Bt, HBBT, and CBT. The ratio of semi gravid plus gravid to unfed plus fed of female An. uviatilis s.l. was more than one indicated exophilic behavior. Agree with our nding, Basseri et al (2010) was relieved exophilic behavior with a high anthropophagic index for An. uviatilis s.l. species [32].
An. dthali has been reported as a secondary vector for malaria in Iran [37]. In our study, An. dthali was more abundant in arti cial pit shelters, followed by CUBT, CO2-Bt, HLC, CBBT, HBBT, and CBT. The abdominal condition of the female An. dthali during our study shows exophilic behavior. Agree with our nding, Basseri et al (2010) was reported the exophilic behavior of An. dthali [32].
An. culicifascies s.l. is the main malaria vector in Iran [38]. In the present study, this species was most abundant in UBT (n=4, 57.2%) followed by HLC (n=3, 42.8%). This species was captured by UBT, and found unfed, freshly fed, semi gravid, and gravid. Although we collected a few An. culicifascies s.l. during of this study, based on the ratio of abdominal condition was observed both endophilic and exophilic behaviors. The same result has been reported in a previous study [32]. In our study, this species was collected in the second third of the night by HLC. Agree with our study, more biting activity of this species was observed before midnight [39]. In contrast, this species was reported as the zoophilic behavior [6, 38]. We collected a few An. culicifascies s.l. during of this study, so at the present, we were unable to decide about comparable behavior.  [40]. This species had the highest abundance and distribution among Anopheles in Kashan County and the best sampling method for its adults was the BG trap [41,42].
An. pulcherrimus is a suspected malaria vector in Iran [39]. Our founding relieved the more abundant of this species in UBT (n=2, 22.2%) followed by HLC (n=7, 77.8%). In our study, this species was collected in the second third of the night by HLC. Agree with our study, more biting activity of this species was found to occur before midnight [39]. Also, this species was captured by UBT, and found unfed, freshly fed, semi gravid, and gravid. Although we collected a few An. pulcherrimus during this study, based on the ratio of abdominal condition was observed exophilic behaviour. Agree with our study Zaim et al (1993) was reported the exophilic behaviour of this species [38]. We collected a few An. pulcherrimus during of this study, so at the present, we were unable to decide about comparable behavior.
In our research, the use of various sampling methods to collect some culicinae sub-family and the tribe Aedini species from indoor and outdoor places helped us to know more about the behavior diversity and densities of these species.
Culex pipiens complex includes several members, more than 75 binomial names due to the complexity of synonyms [43]. Among the species complex, Culex pipiens, Cx. quinquefasciatus and Cx. pipiens form molestus are the most common members of this complex and are widespread in the world [44]. The molecular and morphological studies using ITS2, COI, Ace2 cDNA genes and DNA barcoding indicated some variations of the Culex pipiens complex in Iran [45][46][47]. Because of having morphological variations among the Cx. pipiens complex, sometimes it is not easy to morphologically separate from Cx. quinquefasciatus and Cx. torrentium and, need to molecular assay [47]. Recently, was reported the occurrences of Cx. pipiens pipiens biotype pipiens and a hybrid between Cx. pipiens pipiens biotype pipiens and Cx. pipiens pipiens biotype molestus in Iran [48,49]. Culex pipiens pipiens has a wide distribution in Iran and is also reported the steno-eurygamy, anautogeny, and ornithophilic behaviors [45,49,50]. West Nile and Sindbis viruses are predominantly transmitted between birds [51,52]. This species was reported a potential enzootic and bridge vectors of SINV and WNV already circulating in Iran [53,54].
Shahossaini et al. (2018) reported the occurrence of the West Nile virus and its subtype Kunjin in Iran [49].
In our study, more Cx. pipiens collected by CBBTfollowed by CUBT and, HBBT. It seems that, this species selected calf odor over human odor, despite similar weights and thus similar CO2 expiration rates. In agreement, Shahossaini et al. (2018) reported a high proportion of birds for Cx. pipiens pipiens f. pipiens of detected blood meal sources [49]. Similarly, Culicidae species are more oriented to mosquito Magnet traps treated with synthetic mixture than MM-X traps baited only with human odors in the eld [55]. In contrast, more attraction of Cx. quinquefasciatus to human skin odors were reported [56]. This nding agrees with previous studies, which indicated the Cx. pipiens has a wide range of hosts, whereas Culex quinquefasciatus more attract to selected human odor over calf odor [57][58][59][60][61][62].
Culex theileri is widespread in Iran and, was reported naturally infected with Diro laria immitis and Setaria labiatopapillosa [63,64]. In our study, Cx. theileri was most collected in APS followed by CUBT, HLC and, least in CBBT. This species was not collected during the present investigation utilizing HBBT, CO2-BT and UBT. This nding agrees with previous studies in Guilan Province, northern Iran, which indicated that this dominant species was more collected from cow shelter than human shelter using CDC light trap and Hand catch methods [65]. Shahossaini et al (2018) reported a relatively high proportion of non-human mammals for Cx. theileri of detected blood meal sources [49]. Prior studies carried out in Isfahan province, central Iran, was found that, Cx. theileri as the dominant species in both larval and adult collections with exophilic and exophagic behaviour [41,42,66,67]. In Kashan County, Cx. theileri and Cx. pipiens were the only species collected using the animal-baited bed net trap [41]. In contrast, we found low abundance among culicinae species. Hormozgan has a warmer climate than Isfahan province, and perhaps may explain these unexpected results, moreover there are widespread rice elds in Isfahan, so that we cannot observe them in Hormozgan. On the other hand, a lesser abundance of this species in our study, is perhaps due to the sampling method and time of collection. . We observed no evidence of the occurrences of Ae. aegypti and Ae. albopictus, the important vectors of arboviral diseases in Hormozgan Province. Aedes caspius was found only Aedini species, which were collected in our study. In contrast, Hassandoost et al (2020), reported two species, Ae. caspius and Ae. vexans by larval collection in Hormozgan Province [70]. It should be considered, adult sampling methods in our study, may explain the variation in the results.
The human landing catch was introduced as the standard method for measuring human exposure to mosquito bites. However, this method potentially exposes individuals to a range of nuisance mosquitoes as well as mosquito-borne diseases and created bias through the sampling. On the other hand, baited resting boxes were not a suitable alternative method to collect of this species, which are more important in addressing entomological surveillance due to arbovirus transmission.
The arti cial pit shelter was found more e cient than the baited resting boxes in the density and diversity of An. uviatilis s.l., An. dthali, Cx. pipiens, Cx. theileri and Ae. caspius species.
In agreement, Pombi et al. (2014) reported the low abundance of Afrotropical malaria vectors by Sticky resting box (SRB) than the pit-shelters [71]. Also, our nding agrees with a study performed on An. arabiensis, and An. funestus s.l. in south-eastern Tanzania [72], which the mean abundance of these species collected with SRB traps inside and outdoors was signi cantly lower than with Back Pack Aspiration and Resting Bucket traps. Overall, very low numbers of mosquitoes were capture from the RBs baited with human odour [73], an observation similar to what was found in the current study.
The proportion of gravid and semi gravid to unfed and blood-fed showed the exophilic behavior of these species. The higher density of semi-gravid and gravid mosquitoes using the APS method is useful in evaluating the oviposition behavior of malaria vectors and in determining the best control method in targeting ovipositing mosquitoes. This nding could be highlighted for programs of vector control such as larviciding and thermal fogging, which impact exophilic mosquito species.
Our nding indicated that, the case of black un-baited resting boxes was the priority for all species. Agree with our nding, Burkett et al. (1998) showing Aedes aegypti preferred black spots or traps [74]. Also, more attraction of Cx. quinquefasciatus, Cx. sitiens, and An. epiroticus to black trapping boxes were reported [75]. Khemrattrakool et al. (2019) demonstrated the most of the mosquitoes found in the back UBs were blood-fed females [76], evidence similar to what was observed in the current study. Our nding revealed that, baited RBs more collected unfed mosquitoes, which are important to identifying the parity rates and addressing the situation of malaria transmission., which are important to identifying of the parity rates and addressing the situation of malaria transmission.
The urine-baited traps provide a useful alternative tool for sampling malaria vector mosquitoes. This method can also be useful to survey adult mosquito mites and fungus infection that merged from any breeding site and then collected by baited RBs.
In the laboratory and eld experiments, CO 2 at a ow rate of 170 ml/min was the best dose and used to attracting mosquitoes than other doses. Culex pipiens was found the most frequent species followed by An. stephensi, An. dthali, and, An. uviatilis. Current data approved that Cx. pipiens with non-human preference behavior are more targeted to traps than Anopheles species. In Kashan County, the best method for adult sampling of Cx. pipiens was BG-Sentinel trap with CO2, followed by human baited bed net trap [41].
Agree with our nding, a study determined that the zoophilic species of An. coustani and, An. pretoriensis are more oriented to CO 2 rather than anthropophilic species [60] (Dekker and Takken 1998). In the laboratory experiments, we observed a variation among the different concentrations of CO2 for the attractiveness of An. stephensi and, Cx. pipiens. Agree with our nding, a study determined in Tanzania showed that no correlation among the different concentrations of CO 2 for the attractiveness of An. gambiae and Cx. quinquefasciatus [25]. Carbon dioxide is an activator of mosquito host-seeking, more ever, Zöllner et al. (2004) reported increase in ambient CO2 levels may reduce the effective range of attraction to their blood hosts [77]. Consequently, Grant et al. (1995) relieved that the variations in CO2 concentration could affect the ability of the vector to host-seeking behavior [78]. Our nding showed that carbon dioxide raised from yeast supply was not attracted by some mosquitoes. This nding agrees with a study performed by Smallegange et al (2005), which suggested an alternative method using carbon dioxide in the eld [20]. Agree with our nding, a study determined that, An. gambiae s.s. has been tended to CO2, also dry ice-baited trap was well-operated rather than yeast provided CO2 trap [60,79].
In contrast, was reported the strong response to carbon dioxide raised from molasses supply in the open eld [80]. It should be considered, that climatic factors, such as wind speed uctuations, affected the amount of released carbon dioxide. It means CO 2 was not released at durable concentrations to attract species. In our eld area, winds in the low levels become much more uniform at night to early morning, which perhaps may explain a lesser abundance of some species collected by this method.

Conclusion
Black CUBT and APS methods can be suggested as a perfect sampling strategy for malaria vectors surveillance. APS, and HLC methods were found useful to entomological surveillance systems for Arboviral and larial vector-borne diseases. Further modi ed sampling methods should be devoted to identifying effective sampling methods. In conclusion, the cow urine-baited RBs and Arti cial pit shelthers can be considered as a speci c-species sampling for malaria vector surveillance systems, Also, the human-baited double net trap, could be evaluated compared with human landing catches, for estimating human-biting rates of arboviral and larial vector-borne diseases. On the other hand, to the monitoring of larval stage and container-breeding species, using ovitrap is suggested.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.