Vertical Distribution, Biodiversity, and Some Selective Aspects of the Physicochemical Characteristics of the Larval Habitats of Mosquitoes (Diptera: Culicidae) in Chaharmahal and Bakhtiari Province, Iran

Background and aims: Mosquitoes (Diptera: Culicidae) are still a focus of research because of their role in the transmission of diseases and annoying biting behavior. Source reduction is an effective measure to control mosquito populations, which is based on good knowledge of larval habitats. This study was conducted to obtain that basic knowledge in Chaharmahal and Bakhtiari province. Methods: This study was carried out in 2011 and 2012. Geographical coordinates, altitude, pH, temperature, and the dissolved oxygen level of larval habitats were recorded by relevant devices, followed by documenting physical attributes by direct observation. In addition, the indices of biodiversity were calculated to analyze the vertical biodiversity of species. Finally, the affinity index was calculated to elucidate species co-occurrence. Results: Eighteen species were recovered from 92 larval habitats. Low(≤ 1400 m), mid(1401–2000 m), and high(≥ 2001 m) altitudes lodged 7, 17, and 14 species, respectively. Further, the indices of the species richness and biodiversity for these altitudinal categories were 0.93, 1.94, and 1.58, as well as 1.54, 2.13, and 1.96, respectively. Larval habitats were mostly natural, temporary, with standing but clear water, muddy substrate, sunlit, and with vegetation. Other physicochemical characteristics and affinity of species were described and discussed as well. Conclusion: To the best of our knowledge, this is the first report of vertical distribution and biodiversity of mosquito larvae in Iran. The relative uniformity of physicochemical characteristics of larval habitats was attributed to prevailing water resources in the studied area and sampling design. The oviposition site selection of gravid mosquitoes is still an unresolved problem which needs further investigations.


Introduction
Mosquitoes (Diptera: Culicidae) are of major importance not only because of their role in the transmission of the causal agents of many diseases but also because of their blood-sucking behavior which imposes considerable amounts of stress and discomfort to human beings and domesticated livestock. 1,2 Therefore, lots of resources are spent yearly to combat these annoying insects, even in the absence of any diseases they transmit.
Source reduction through environmental management is an effective supplementary tool in controlling mosquitoes operating at their potential larval habitats. The advantage of this method is that it not only affects endophagic and endophilic mosquitoes but also exophagic and exophilic populations that are harder to be eliminated by other control measures. However, the success of this intervention depends on the type, number, and accessibility of available larval habitats in the targeted area. 3 Mosquitoes do not lay their eggs in a random manner; rather they discriminatively select where to oviposit. 4,5 While environmental factors such as relative humidity, ambient temperature and wind speed are involved in the flight orientation of mosquitoes to locate a potential oviposition site; long-to short-range visual, olfactory and tactile cues are consecutively engaged to choose a suitable place for laying eggs. Once contact with the substrate is made, the right chemical profile originating from previously laid eggs, growing larvae, pupae, and even the substrate itself stimulate the female gravid mosquito to oviposit. Otherwise, the female mosquito flies away to find another place to lay her eggs. 6,7 Interestingly, it is showed that arboviral infection could affect the mosquito associate olfactory learning and results in the loss of oviposition site preference. 8 On the other hand, the growth and development of mosquito larvae in a given aquatic habitat is under the influence of biotic and abiotic surrounding environment. 9 Water movement, suspended or dissolved organic and inorganic materials, temperature, hydrogen ion concentration, dissolved oxygen, and the presence of food particles could be listed as abiotic factors. Vegetation type and the presence of predators or competitors are also considered as biotic parameters. [4][5][6][7] Although some studies report that there is a correlation between larval abundance and some sets of these variables, 10-13 others do not prove it. 14,15 In spite of considerable advances in the behavioral and chemical ecology of oviposition in mosquitoes, contradictory evidence on the role of environmental variables in the life history of immature stages demands further research both under controlled and field conditions.
Regardless of numerous reports on the larval habitat characteristics of mosquitoes in other countries, there are few studies characterizing the larval breeding places of mosquitoes in Iran. While early researchers have mostly concentrated on the general description of larval habitats, 16,17 late researchers have rather paid attention to the physicochemical characteristics of these places in detail. Several examples of the latter type of studies are those conducted by Yaghoobi-Ershadi et al 18 in Ardebil, Azari-Hamidian [19][20][21] in Guilan, Dehghan et al 22 in Hamedan, Hanafi-Bojd et al 23,24 and Soleimani-Ahmadi et al 13,25 in Hormozgan, Banafshi et al 26 12,31,32 in Mazandaran, Abai et al 14 in Qom, Sofizadeh et al [33][34][35] in Golestan, and Paksa et al 36 in East Azerbaijan. In a number of these studies, the biodiversity of mosquitoes 23,24,36,37 and associate species 12,[19][20][21][22][23][24]26,30,31,38,39 are considered as well. The biodiversity of the mosquitoes of Iran is relatively a young and an intriguing field for mosquito ecologists, which has demonstrated a growing trend in recent years. 23,24,37 Although evidence on the horizontal biodiversity of mosquitoes is increasingly accumulated, the vertical biodiversity of mosquitoes as a closely related subject has not received enough attention of entomologists. This need may be reinforced by the fact that it is newly showed that the prevalence of some mosquito-borne diseases is dramatically decresed at a certain level of elevation. 40 In a recent faunistic larval survey, all previously reported culicine mosquitoes (11 species) and six anopheline species were collected from Chaharmahal and Bakhtiari province. 41 The distribution maps of the recovered species were reported, and Culex territans was introduced as a new species to the fauna of this province. Briefly, while Cx. theileri (25.1%) and Anopheles superpictus (15%) were the most abundant and widespread species, Cx. territans, Cx. tritaeniorhynchus, and Ochlerotatus caspius s.l. (<0.01% each) were the least frequent and the most limited species in the province. There are some controversial debates on the classification of the tribe Aedini. Based on Reinert et al. [42][43][44][45] there are 82 genera in the tribe and based on Wilkerson et al. 46 just ten genera with numerous subgenera. Debates continue in this regard. 1,47 This study listed the species of the tribe based on the first classification. As far as we know, there is little or no information on the physicochemical characteristics of the larval habitats of mosquitoes in Chaharmahal and Bakhtiari province. There is also no information on the biodiversity and inter-specific association of this taxon in this region. Considering the above-mentioned explanations, the present study addressed some selective aspects of the larval habitat characteristics of mosquitoes in this province, followed by studying the biodiversity and inter-specific association of the identified species.

Study Area
This descriptive-analytical cross-sectional study was carried out in Chaharmahal and Bakhtiari province. The province with an area of 16 532 km 2 is located between 31º 9′ to 32º 48′ N and 49º 28′ to 51º 25′ E at the western part of Iran ( Figure 1). It borders with Isfahan province in the north and east, Kohgylooyeh and Boyerahmad province in the south and southeast in addition to Khuzistan province in the west and southwest and Lorestan in the northwest. Moreover, its population is 947 763 and the province officially includes 10 counties. The province with 2153 meters altitude above the sea level on average is mostly a highland area in which few plains are stretched between hilly and mountainous parts. 48 Additionally, the climate is the Mediterranean based on Köppen's classification so that winters are cold and humid and summers are relatively temperate. 48,49 The average annual rainfall is 1152 mm, and the minimum and maximum temperatures are 17.98ºC and 33.68ºC, respectively. Rainfall generally starts from October each year and peaks around December. Then, it gradually decreases until the next April. Water bodies are so rich that about 10% of the total resources of the country are ensured by Chaharmahal and Bakhtiari province. Regardless of special climate, landfilling and agricultural activities are the major determinants of the distribution of temporary surface waters in the province. It should be mentioned that July and December are the warmest and coldest months of the year, respectively. The plant coverage is relatively rich, and human settlements more or less follow the water course. 48 There are many livestock in the province and animal husbandry and agriculture are two common occupations of people.

Larval Collection
In this study, 92 larval habitats in 10 counties of Chaharmahal and Bakhtiari Province were sampled during the June-September periods of 2010 and 2011 (Table 1). The sampling sites (villages and cities) were selected based on their geographical location and altitude from the sea level. In each locality, various potential larval breeding places were checked for the presence of mosquito larvae. In addition, the geographical attributes of the collecting sites were recorded by a hand-held GPS device (Etrex, Garmin, Taiwan). Then, field surveys were carried out from the early morning up to late afternoon by the 30-minute dipping method in each sampling site. Next, the collected larvae were preserved in lactophenol and Berlese's medium for temporary and permanent fixation purposes, respectively. All specimens were morphologically identified at the species level by a valid local key 50. No attempts were made on the differentiation of sibling species.

Larval Habitats and Physicochemical Characteristics
Mosquito larval breeding places were generally classified into river edge, river bed, stream bed, spring bed, pond, rice field, drainage, agricultural irrigation pool, agricultural irrigation channel, and the man-made container. For each larval habitat, environmental variables including nature (natural/artificial), stability (temporary and permanent), water flow (slow running and standing), turbidity (clear and turbid), substrate type (muddy, sandy, and gravel), sunlight status (sunny, partially shaded, and shaded), and vegetation (with and without) were recorded separately. Vegetations included both aquatic and immersed terrestrial plants. It was assumed that all larval breeding places contain freshwater. Water temperature was measured by a hand thermometer and wherever possible, pH and dissolved oxygen levels were also recorded by a multimeter (Lutron, YK-2001 DO, Taiwan) equipped with special probes. The device was calibrated before each measurement to minimize the possibility of systematic errors.

Indices of Biodiversity and Species Co-occurrence
The species richness index was calculated based on Margalef index as D mg = (S-1)/ln (n) where S is the total number of species and n equals the total number of individuals in the sample. 51 Then, the Shannon-Wiener index of biodiversity, H' = -Σp i ×lnp i , in which p i denotes the proportional abundance of the ith species, was calculated for each stratum to compare species richness and evenness between the strata. 52 Evenness (J' or E or Pielou's index) was calculated by J' = H'/ H'max. H'max is the maximum possible Shannon's diversity that is computed by H'max = Log10k where k indicates the number of species collected in the sample. 53 The affinity index of species co-occurrence   where J and N A are the number of joint occurrences and the total number of the occurrences of species A, respectively. Further, N B is the total number of the occurrences of species B. The species were assigned to the letters so that N A <N B . This index ranges from -1.0 to +1.0. Eventually, values equal to or more than 0.5 were considered as significantly associated species in larval habitats. 54

Data Analysis
Data were entered into SPSS software, version 20.0, and the indices of central tendency or proportions were calculated for the physicochemical parameters of each type of larval breeding place, separately. Furthermore, independent samples t-test was used to compare the means of pH, dissolved oxygen, and temperature between anopheline and culicine larval habitats at P = 0.05 level of significance. The differences in the physicochemical properties of the larval habitats of mosquito species were determined using ANOVA. The association between altitudinal categories and the subfamily or species of mosquitoes was examined by chi-square and logistic regression tests, respectively. Additionally, the indices of species richness and biodiversity were calculated by a free package of Biodiversity Calculator on the web, followed by computing the affinity index of species by Microsoft Office Excel software, version 10.0. In this report, altitudinal data were arbitrarily classified into three relatively equal range strata to figure out the vertical distribution of mosquito species.

Vertical Distribution and Biodiversity
In total, 8335 mosquito larvae representing four genera and 18 species were collected in this study. Mosquito larvae were collected from a wide range of altitudes from 911 to 2589 meters above the sea level. Table 2 presents collected mosquito larvae, the location and the altitude of their larval breeding places, and the total occasion of larval habitats according to the altitudinal categories. For more convenience, henceforward, low-, mid-, and high altitudes are used instead of ≤1400 m, 1401-2000 m, and ≥2001 m to short writings. In general, 9.8%, 43.5%, and 46.7% of the total larval breeding places were placed in low-, mid-, and high-altitudes, respectively. These altitudinal ranges were represented by 7, 17, and 14 mosquito species, respectively ( Table 2). There was a significant association between the altitudinal strata of larval habitats and mosquitoes at the subfamily level (X 2 = 9.19, df = 2, P = 0.010) except for Cx. territans, Cx. tritaeniorhynchus, and Oc. caspius s.l. Table  3 provides the selective alpha biodiversity indices of the collected mosquito larvae based on the altitudes of larval habitats in the province.

Physicochemical Characteristics
In general, spring bed (32.3%), rice field (20.2%), and river edge (19.9%) were the most prevalent types of larval habitats for both anopheline and culicine mosquitoes while drainage (0.9%) and man-made container (1.6%) were the least ones. The agricultural irrigation channel was a bit more frequent to contain culicine compared to anopheline larvae (13% versus 4.3%). In addition, the single culicine larva of Cx. territans, Cx. tritaeniorhynchus, and Oc. caspius s.l. was collected from the rice field, man-made container, and stream bed, respectively (Table 4 and Figure 2). All mosquito larvae were collected from freshwater habitats ( Table 5). Regardless of few inter-specific differences, most anopheline and culicine larval breeding places were natural (67.6%) and temporary (70.7%) in nature with standing (60.5%) but clear water (94.4%), a muddy substrate (69.4%), sunlit (84.0%), and with some sorts of vegetations (96.6%). Few exceptions to this generalization are Cx. pipiens which was equally collected from both natural (50%) and artificial (50%) larval habitats. The larval breeding places of An. claviger (100%), An. turkhudi (66.7%), and Cx. laticinctus         The chi-square test did not show any significant association between larval habitat types and mosquitoes at the subfamily level. Table 6 presents the mean of water temperature, pH, and dissolved oxygen in anopheline and culicine larval breeding places. Based on the findings, a significant difference was observed between the mean temperature of anopheline and culicine larval habitats (P = 0.006). Excluding Cx. territans, Cx. tritaeniorhynchus, and Oc. caspius s.l. because of their single occurrence, acidity (P = 0.010), dissolved oxygen (P = 0.031), and temperature (P = 0.001) significantly differed between the larval habitats of different mosquito species.

Discussion
This study described the selective aspects of the larval habitats of mosquitoes in Chaharmahal and Bakhtiari province, including vertical biodiversity and distribution, species occurrence, species affinity, and the physicochemical characteristics of the larval habitats of 18 recovered mosquito species.

Vertical Distribution and Biodiversity
Referring to the species occurrence provides further  To make these inferences, the rationale was taken that the higher occurrence leads to the wider distribution and the higher probability of an abundant population. Anopheles dthali larvae were collected six times from low-altitudes. They were also collected 10 times from mid-altitudes while never from high-altitudes (P < 0.001). These findings show that mid-altitude is more favorable for its low population.

Anopheles claviger
Anopheles turkhudi larvae were collected only one time from low-altitudes. They were further collected two times from the mid-altitudes although they were never collected from high-altitudes (P = 0.025). It implies that the population of this species is so low that it does not allow the species to spread horizontally and vertically.
The horizontal distribution maps of An. dthali and An. turkhudi larvae revealed that the populations of these species are limited to the western parts of Chaharmahal and Bakhtiari province where high elevations interfere with their dispersion into adjacent areas. 41 In their study, Amani et al 27 collected all An. dthali and An. turkhudi larvae from the mountainous parts of Aligudarz county in Loristan province, but in this study, these larvae were collected from the low-to mid-altitudes of Chaharmahal and Bakhtiari province. Unfortunately, they did not define mountainous and plain areas by altitude in their report.
In addition, Cx. mimeticus larvae were collected 4, 20, and 8 times from low-, mid-, and high-altitudes, respectively (P = 0.124). The picture was taken 8, 20, and 8 times for Cx. perexiguus, respectively (P < 0.001), indicating that the vertical distribution of Cx. mimeticus and Cx. perexiguus larvae was relatively similar. Although they both more horizontal distribution in mid-altitudes, the populations of Cx. perexiguus had more adaptive capacity to spread in low-altitudes.
Furthermore, An. maculipennis s.l. larvae were collected 2, 18, and 13 times from low-, mid-, and high-altitudes, respectively (P = 0.356), implying that this species was well distributed in mid-to high-altitudes, and low-altitudes were also colonized by their medium size population.
In a study on the mosquito vectors of dirofilariasis in the northwest of Iran, the larvae of An. maculipennis s.l. and An. sacharovi were collected from higher (830-1650 m) and lower altitudes (40-80 m) of Ardebil province, respectively. 55 This finding is in contrast with those of studies in the low-altitude Caspian Sea littoral where An. maculipennis s.l. is much more abundant than An. sacharovi that is absent or rarely found in this area. 21,32,56 Probably, other factors such as rice fields, temperature, and humidity, instead of altitude, influence the composition of the species in the region. More accurate interpretation of the altitudinal data of An. maculipennis s.l. larvae in the present study needs further molecular identifications of the collected larvae.
It should be mentioned that Cx. theileri larvae were the most abundant and widely distributed species among all collected mosquito larvae in the province. They were collected 6, 20, and 24 times from low-, mid-, and highaltitudes, respectively (P = 0.189), which implies that their high population confers an ability to spread everywhere in the province.
The larvae of Cx. theileri are collected infrequently from Guilan, 20 Qom, 14 Hormozgan, 24 and Golestan [33][34][35] provinces that are low land areas in Iran. On the other hand, it is a frequently encountered larval species in high-altitude provinces such as Ardebil, 55 Isfahan, 30,57 and Zanjan. 58 This evidence indicates that the larval habitats of Cx. theileri are located at higher altitudes. Conversely, some other reports may not comply with this outcome. For example, the larvae of this mosquito species were collected in low numbers from the high-altitude provinces of East 36,59 and West Azerbaijan. 28 The findings of the present study do not fit any of these studies. Perhaps, a reasonable explanation needs further research by controlling possible confounders to rectify these discrepancies.
Anopheles superpictus s.l. larvae were the second most abundant and widely distributed species next to Cx. theileri. They were collected 9, 26, and 17 times from low-, mid-, and high-altitudes (P = 0.001), respectively, representing that this species is well adapted to mid-altitudes, but to some extent, breeds in low-or high-altitudes as well. However, the population size of An. superpictus s.l. larvae was 37% of Cx. theileri. 41 Given that the number of the occurrence of these two species is more or less the same, it means that the breeding places of Cx. theileri are more productive compared to An. superpictus s.l.
Nonetheless, An. claviger, Cx. laticinctus, and Cs. subochrea were never collected from low-altitudes. Although An. claviger was collected in one and three occasions (P = 0.107), Cx. laticinctus was collected two and one time(s) from midand high-altitudes (P = 1.000), respectively. Moreover, Cs. subochrea was similarly collected two and three times from mid-and high-altitudes (P = 0.520), respectively. This pattern shows that these mosquito species breed in higher lands but their populations are quite low.
Anopheles marteri, Cx. arbieeni, and Cx. pipiens also showed relatively similar vertical and horizontal distributions although none of these species were collected from low altitudes. Anopheles marteri larvae were collected in 5 and 2 occasions (P = 0.500) although Cx. pipiens and Cx. arbieeni larvae were collected 5 and 3 times (P = 0.819), and 7 and 3 times (P = 0.943) from midand high-altitudes, respectively. This indicates that these species breed in highlands but their populations are low.
A high number of Cx. pipiens larvae was collected from the low lands of the Caspian Sea littoral, including Guilan, 20 Mazandaran, 39 and Golestan 33,35 provinces. In comparison, the larvae of this species were collected in low numbers from the high land areas of Iran such as Ardebil, 55 Isfahan, 30 West Azerbaijan, 28 East Azerbaijan, 36,59 Kurdistan, 26 Kurdistan, and Kermanshah 38 provinces, implying that under suitable environmental conditions, Cx. pipiens tends to establish its population in low-altitudes. However, Dehghan et al 22 were able to collect a relatively high number of the larvae of this species from Hamadan province that is usually considered as a high land area in Iran. This report may be an indication of the influence of other factors in the establishment of the species (e.g., the adaptation for breeding and surviving around human settlements). 60 Based on the findings of the present study, the larvae of Cx. pipiens were found in man-made habitats more than other species (50%), which is in accordance with the findings of many previous investigations. 20,32,39,57 It seems that the most favorable larval habitats of the species in the cities are the house ponds, wells, septic tanks, and the sewage wells. 17,60-62 That is why Cx. pipiens is called the house mosquito, is found almost throughout Iran, and occurs most probably in cities with different altitudes more than any other mosquito species. Because of morphological similarities between female adults and the larvae of Cx. pipiens and Cx. torrentium in northern Iran where they may occur together, along with Cx. pipiens and Cx. quinquefasciatus in southern Iran for the same reason, 50 their compositions need to be investigated extensively with precise identification in the future.
Culex hortensis and Cs. longiareolata larvae were never collected from low-altitudes, but they were collected from mid-and high-altitudes 8 and 25 (P < 0.001), as well as 7 and 22 times (P < 0.001), respectively. The obtained data demonstrate that both species prefer higher, especially high-altitudes, to breed because their populations are well distributed in these areas.
Only one single specimen of Cx. territans, Cx. tritaeniorhynchus and Oc. caspius s.l., was collected in the present study. In addition, two species Cx. territans and Oc. caspius s.l. were collected in the central parts of the province. This may imply that some populations of these species were escaped from our sampling efforts. Therefore, it would be more logical to exclude these species from our explanations on the distributional patterns of the presented mosquito species.
Unfortunately, most literature regarding the mosquitoes of Iran either does not pertain to altitudinal information or contains so mixed data that cannot be easily retrieved and used in analytical comparisons. For example, no data are reported on the altitude of collected species in the studies conducted on the larval habitats of mosquitoes in Golestan, 33,35 Guilan, [19][20][21] Hormozgan, 13 Qom, 14 Hamedan, 22 Isfahan, 30 Kurdistan, 26 Kermanshah, 38 and Mazandaran 31 provinces. On the other hand, in studies carried out in Hormozgan 13,23,25 and East Azerbaijan 36 provinces, the altitudinal data were reported as a function of localities and sampling sites rather than mosquito species. Perhaps, in a number of these cases, such lack of data may actually come from the local ground features as there is little variation in elevations in low land areas. A few examples of this kind are found in studies conducted in Hormozgan, 24 Bushehr, 63 Mazandaran, 31 Golestan, 33,35 and Guilan [19][20][21] provinces. Similarly, data are presented by the topography of collection sites in few studies that paid attention to the vertical distribution of mosquito larval habitats. 27,64 In these investigations, there is no reference level to understand what they mean by mountainous or foothill areas.
In this study, larval habitats located in mid-altitudes showed the highest number of mosquito species, the highest value of the Margalef richness index, and the highest value of the Shannon-Weiner index of biodiversity while the least value of the Simpson index (Table 3). It is noteworthy that when Simpson index (D) decreases, the species diversity represents an increase, explaining why the Simpson index is sometimes shown as 1-D or 1/D. Interestingly, when the Pielou's evenness index was calculated for three altitudinal strata, the lowest altitude category (≤ 1400 m) displayed the highest value (Table 3). This shows the influence of the sample size of this stratum in which the numbers of collected larval specimens (637), species (7), and larval habitats (9) are the lowest among the strata. This is because Chaharmahal and Bakhtiari is a high altitude province and the most larval habitats are located in higher than 1400 m above the sea level. There is little information about the biodiversity of mosquitoes in Iran for comparison. For instance, Nikookar et al 37 mentioned some biodiversity indices of mosquitoes in three low-level sites (185-290 m) in Mazandaran province of the Caspian Sea littoral, northern Iran. Furthermore, Hanafi-Bojd et al compared the Shannon-Weiner index of biodiversity and the Pielous' richness index in five sites (450-1020 m) of Bashagard County of Hormozgan province, southern Iran. 23,24 However, none of the above-mentioned studies compared or discussed the biodiversity based on the altitudes of the sites.
The "latitudinal gradient of species richness" is a known and well-documented phenomenon in biogeography, which declares that biodiversity increases by decreasing the latitude. 65 In contrast, the "elevation diversity gradient" that is usually considered as a mirror of the previous phenomenon on a smaller scale is a more complex event and the scale and extent of the altitudinal gradient may be observed in different patterns depending on the targeted taxonomic group. 66 Our data are consistent with the "midelevation peak" pattern in which the species richness is higher in mid-altitudes. 67 To our knowledge, this is the first study to deal with the vertical biodiversity of mosquitoes in Iran.

Physicochemical Characteristics
Two sets of variables are commonly used in the description of the physico-chemical characteristics of mosquito larval habitats. The first set consists of qualitative nominal or ordinal variables that are normally employed to describe the physical attributes of mosquito larval habitats. These parameters are also referred to as environmental 9,13,25,68 or biological characteristics by some authors. 24,26,28,32,36 In the present study, spring bed pools and, to some extent, river edge and rice fields were the most prevalent types of larval habitats for all six anopheline and 12 culicine mosquitoes. They were mostly natural and temporary in nature, with standing and clear water, muddy substrate, sunlit, and with vegetation. It is proposed that this relative uniformity of data is partly due to prevailing water resources for breeding mosquitoes in the studied area. Other possibilities are the insufficiency of criteria and the problem of the sampling design.
Nikookar et al 31 reported 6 anopheline and 10 culicine mosquito species in a larval survey in Mazandaran province located in northern Iran. The larval habitats were mostly temporary, stagnant, with plants, shadow-sun, muddy floors, and turbid or clear freshwater. In another study, Hanafi-Bojd et al 23 presented the larval habitats of 8 anopheline species from Hormozgan province in the south of the country. Natural larval breeding places, without vegetation, in full sunlight, with a sandy substrate and fresh and clear water, were the dominant physical characteristics of mosquito larval habitats. A couple of years later, they also characterized the oviposition sites of 12 culicine species from the same area. 24 Natural, temporary, or permanent larval breeding sites without vegetation, in full sunlight, with a sandy substrate and fresh and clear water, were the predominant physical features of mosquito larval habitats.
These findings of many other studies may be an indication of slight intra-provincial variation in the characteristics of mosquito larval habitats.
It is noteworthy that most exceptional cases in such reports are those species with low abundance and low occurrence. For instance, in the latter example, the larval habitats of 8 culicine species had a sandy substrate. However, 100% of the larvae of Cx. arbieeni, Cx. theileri, Oc. caballus, and Oc. caspius were exceptionally collected from larval breeding places with a muddy substrate.
Looking at the companion data shows that all these four species were collected from only one larval habitat. This means that in the interpretation of the descriptive data of mosquito larval habitats, the population size and the number of occurrences should be taken into account in order to avoid unrealistic conclusions.
Probably, an example of mosquitoes with actually deviant physical characteristics of their larval habitats could be observed in the study of Azari-Hamidian, 21 who reported 5 anopheline species larvae from Guilan province although most other mosquito larvae (69.6%) were collected from sunlit oviposition sites. Anopheles claviger (66%) and An. plumbeus (100%) larvae were exceptionally collected from shaded habitats. These species constituted 6.3% and 13.1% of the total collected larvae with 14 and 12 occurrences, respectively. This validates that the obtained data about the sunlight status of the larval breeding places of these species in the studied context were not accidental.
We do not believe in the incompetence of currently used variables, but the sampling design could be arguable. Most studies on the larval habitats of mosquitoes, including the present investigation, do not intensively survey potential places for breeding mosquitoes. They neither report negative places nor characterize their attributes. Therefore, we have to admit that our knowledge on the oviposition sites of mosquitoes is highly preliminary, in a sense.
The second set of variables in the description of mosquito larval habitats is comprised of quantitative variables like pH, dissolved oxygen, along with the anions and cations of water resources. These are usually used in the description of the chemical characteristics of larval breeding places. Temperature is exceptionally classified in this group.
In this study, the larval habitats of anopheline and culicine larvae showed low levels of alkaline pH. Several studies in Iran reported the pH of mosquito oviposition sites. However, not all of them presented species-specific data. 13,23,25,36 In general, the mean pH level of mosquito larval habitats is reported to range from 7.0 to 8.0. [12][13][14]24 The minimum and maximum of the reported pH level are 6.9 12 and 8.9. 29 Other field and laboratory evidence shows that different mosquito species tolerate pH levels from lower than 4 up to 10.5. 69 Therefore, although there are some reports regarding the correlation between pH and the larval density of mosquitoes, 70,71 it is believed that it does not exert a direct effect on the distribution of mosquitoes. 69 The present study reported the mean dissolved oxygen level for the larval habitats of 18 mosquito species in Chaharmahal and Bakhtiari province. To the best of our knowledge, no other report is available for comparison in this regard in the context of Iran. Mosquito larvae get oxygen directly from the air by breathing through their respiratory siphons or a pair of spiracles. Thus, the dissolved oxygen of water has not been before a focus of much research by mosquito ecologists 5 . Nevertheless, it is an indicator of water quality and productivity in aquatic habitats. A few recent works have reported that there is a strong correlation between habitat types, 72 larval abundance, 11,73 and mosquito species 74 with the dissolved oxygen level.
In this study, it was found that the mean of the temperature of anopheline larval breeding places is around 1.4 degrees of centigrade higher than that of culicine ones (P = 0.006). This difference might be attributed to the altitudinal level of oviposition sites as the larval habitats of anophelines were about 140 m lower than those of the culicine ones. Temperature is an imperative factor for the growth and development of anopheline and culicine larvae. 69 It is also shown that female mosquitoes do not lay eggs on waters with higher or lower levels of temperature than a certain one. 4 There is a great deal of information about the temperature of mosquito larval habitats from Iran. However, in the analysis of these data, it should be noted that the temperature of water resources is subject to a considerable change by the time of the day, season, size, depth, as well as the movement of water and the type of the substrate. Perhaps, most of these confounders could be managed with more realistic image obtained by the advent of remote sensing and the geographical information system. 68,75 Several studies in Iran [12][13][14]25,29 and other countries 10,76 have explored the relationship between larval density and the physicochemical characteristics of mosquito larval breeding sites. Some of them reported a significant relationship between a variable and the larval density of a mosquito species although other studies did not find any relationship in this regard. Two points merit to be stressed in this respect. In these studies, it is not clear whether the effect of the number of mosquito species occurrence, as discussed above, is also incorporated or not. In addition, these relationships do not necessarily represent the preferred oviposition sites of a mosquito species. This is because the larval density is a post-oviposition phenomenon and is just an implication of the productivity of a habitat type. 5,7 The presence of even a single larva in a water body suggests that this place has been selected by a female mosquito for oviposition.
Mosquitoes adopt different strategies to lay eggs. Although some species do not touch the surface of the water, others make contact with it at least for some moments. It is believed that olfactory and chemical cues are the key elements of the oviposition behavior of these two mosquito groups, respectively. 7 The metabolites of decaying bacteria, plus other complex organic compounds released by predators and competitors, produce distinct odors and flavors that attract or repel a specific female mosquito to find a suitable place, and also stimulate or deter female mosquito to finally lay eggs there. 4,6,9 Despite remarkable progress in this regard, there are still many unanswered questions regarding this issue why two seemingly identical and adjacent water resources are differently selected by gravid mosquitoes. This knowledge would be quite useful in devising more efficient ovitraps.

Species Occurrence, Co-occurrence, and Affinity
In the present study, some mosquito species displayed shared larval habitats. Furthermore, the analysis of the species occurrence data, along with larval abundance, indicated that there is a positive correlation between these variables. 41 For example, Cx. theileri and An. superpictus s.l. were the most abundant and had the most frequent larval habitats. The same was true for the next abundant species Cx. perexiguus, Cx. hortensis, An. maculipennis s.l., Cx. mimeticus, and Cs. longiareolata, which is in agreement with the results of other studies in Guilan, [19][20][21] Kurdistan, 26 and Isfahan 30 provinces. However, in areas where the number of collected larvae is extremely low, this correlation loosens its strength. In these situations where larval breeding places are highly scattered (e.g., due to harsh conditions), it is not infrequent to collect a higher number of the larvae of a given species from just a single oviposition site. 14,24 In this study, An. superpictus s.l., An. maculipennis s.l., and Cx. perexiguus larvae were frequently collected with Cx. theileri. Similarly, Cx. perexiguus larvae commonly occurred with An. superpictus s.l. Moreover, Cx. theileri highly occurred with An. maculipennis s.l., An. superpictus s.l., Cs. longiareolata, Cx. perexiguus, and An. claviger in Kalaleh of Golestan 34 ; Cx. perexiguus, Cx. hortensis, Cs. longiareolata, and An. maculipennis in Isfahan 30 ; An. superpicus and An. maculipennis in Kurdistan 26 ; and An. maculipennis and Cx. tritaeniorhynchus in Guilan provinces. 20 Such co-occurrences could be the result of overlapping high populations of adults, common larval needs, enough nutritional resources, no interspecific competition, and the scarcity of available resources to deposit eggs.
It is reported that Cs. longiareolata larvae may be collected alone. 24,26 The predatory behavior of this mosquito species could be the reason. In this study, this species was collected with 3 anopheline and 7 culicine mosquito larvae. Others also reported its concurrent presence with other species. 19,22,28,30,34,38 This evidence signifies that the predatory behavior of Cs. longiareolata could not be all the reason. The general physical characteristics and the pH value of the larval habitats of this species overlap with others, 24,26 meaning that other factors must also be acted in those contexts.
The highest values of the affinity index in our study were recorded for the pairs of Cx. theileri/An. maculipennis, Cx. theileri/Cx. perexiguus, An. superpictus s.l./Cx. perexiguus, An. superpictus s.l./Cx. theileri, and Cx. perexiguus/ An. maculipennis s.l. larvae in a descending order. Considering that around two-thirds of the occurrences of Cs. longiareolata happened with Cx. hortensis, it is not surprising that this pair exhibited a high value of the affinity index. There is little information on the affinity of mosquito larvae in Iran. 23,24,30 Ladonni et al 30 did not find any affinity between 15 mosquito species larvae in Isfahan province. On the other hand, Nikookar et al 12 calculated the inter-specific correlation coefficient for collected mosquito larvae in Mazandaran province. In another study, Hanafi-Bojd et al 23 reported a significant affinity between 8 pairs of anopheline mosquito species larvae in Hormozgan province. Of those pairs, An. superpictus s.l. and An. dthali showed a 0.521 affinity. However, the current study did not find such an affinity between these species. The reason could be related to the fact that only 31% of the total occurrences of An. superpictus s.l. were accompanied with An. dthali.
Studies on the ecology of mosquito larvae like any other investigations are facing with a number of limitations. Banafshi et al 26 listed several limitations in this regard in their study conducted in Kurdistan province. It should be added that aquatic habitats do not stay constant but are subject to a change over time. Normally, most ecological studies take a lot of time, even longer than a season, to be conducted well. In the meantime, the quality and quantity of water bodies are considerably altered so that by a change in meteorological conditions, low nutritive and unsuitable water bodies convert to highly productive resources for breeding mosquitoes. Simultaneously, human activities like agriculture, substantially expand the potential site for mosquito breeding. Besides, the population of adult mosquitoes, which is, in turn, under the influence of the availability of preferred hosts, interferes with the distribution of larval habitats. A low number of collected specimens may come from really low populations of a mosquito species. This problem may not be resolved by intensive sampling and thus makes the generalization of the results difficult. Undoubtedly, all these constraints affect our assessments on the preferred oviposition sites of mosquitoes.

Conclusion
To the best of our knowledge, this is the first study to describe the selective aspects of the ecology of the larvae of mosquitoes in Chaharmahal and Bakhtiari province and to present species occurrence, species affinity, and physicochemical characteristics of the larval habitats of 18 recovered mosquito species. The data of the altitudinal distribution of these species are new to Iran and reveal higher vertical biodiversity in mid-altitudes. Additionally, the investigated physicochemical characteristics of larval habitats were indiscriminative. It is proposed that, regardless of the possibility of problems with sampling and the incompetence of criteria in characterizing larval habitats, this might be a reflection of typical water resources available for the breeding of mosquitoes in the studied area. On the other hand, the overlapping of the larval habitats of mosquitoes may be an indication of the generalist type behavior of some female mosquitoes in the selection of suitable places to lay eggs. The bottom line is oviposition site selection by female mosquitoes, like their host preference, is still an unresolved subject. More investigations, both under controlled and field conditions, are necessary to understand the basis of this vital behavior in mosquitoes.