Oviposition and Larval Development of Culicoides Insignis Lutz (Diptera: Ceratopogonidae) under Laboratory Conditions

Culicoides insignis Lutz (Diptera: Ceratopogonidae) is a conrmed vector of bluetongue virus (BTV) throughout the American tropics and a possible vector of epizootic hemorrhagic disease virus (EHDV) in Florida. Despite its importance, fundamental information on the biology and ecology of this species is lacking. In this study, we examined the oviposition and larval development of C. insignis under laboratory conditions and attempted colonization of this species.

Much of our knowledge on the larval ecology of Culicoides vectors in North America comes from studies on C. sonorensis, a midge species best studied in the arti cial wastewater ponds of California and also the only con rmed vector of BTV/EHDV successfully colonized to date worldwide [8][9][10][11][12][13]. As such, basic research on the biology/ecology and successful laboratory colonization of C. insignis and other Culicoides species associated with virus transmission in the southeastern US will be highly advantageous in understanding the transmission dynamics of BTV/EHDV in this region and will also be useful in the development of effective vector/disease management strategies in the long term. In this study, we examined the oviposition preferences and larval development traits of C. insignis under laboratory conditions and attempted colonization of this species. Culicoides insignis is a common vector of orbiviruses distributed across most of South America, Central America, and the Caribbean, with its northern boundary extending into Florida and the neighboring states [14,15]. This species, 1) is frequently associated with livestock facilities [16][17][18], 2) is a signi cant biting pest of livestock [19,20], and 3) is a con rmed vector of BTV in Florida and also likely plays a role in EHDV transmission in the state [5,21]. Overall, our study demonstrates effective methods for the blood-feeding, oviposition, and egg collection of C. insignis, and provides the rst insight into the oviposition preferences and larval development traits of this species. Unfortunately, male-biased sex-ratios in the progeny and di culties in inducing captive mating in the F1 generation prevented successful colonization of this species.

Methods
Live midge collection Live midges were collected using procedures described previously for C. stellifer [22]. Brie y, CDC miniature light traps tted with UV-LED arrays and insect collection containers were set up overnight at the Archbold Biological Station's Buck Island Ranch, Lake Placid, FL, USA (27°9'16.4"N, 81°11'51.7"W).
The eld-collected insects were brought to the laboratory the next morning where C. insignis females were anaesthetized using triethylamine (TEA), morphologically identi ed, caged in paper cups, and provided with 10% sucrose until use.
Oviposition studies Culicoides insignis females were starved for 6 -12 hours after which were introduced into 50 ml conical tube feeding chambers and allowed to blood-feed on the breast of a live chicken (University of Florida IACUC protocol #7682). The partially and fully engorged females were sorted under TEA anesthesia, placed individually in paper cups provided with two different substrates for oviposition (two-choice assays) and allowed to oviposit for 14 days. The oviposition dishes were checked daily, and the number of eggs deposited on them were counted. After the end of the two-week period or after midge death, females were dissected and the number of eggs retained, if any, were also counted. The females were provided with fresh cotton pads dampened with 10% sucrose solution daily. Each oviposition dish consisted of about 1g of natural substrate (see below) placed on the bottom of a Petri dish (35 × 10 mm) and was covered with a layer of cotton and a lter paper on the top. Moisture levels across all dishes was maintained constant by adding 1.5ml deionized water and replenishing with the same volume of water in both dishes presented to midges whenever necessary. The natural sources tested for oviposition preferences were selected based on personal observations and published reports [18,23]. The larval habitat of C. insignis was identi ed on Archbold Biological Station's Buck Island Ranch, Lake Placid, FL, USA through emergence sampling from this site. Surface mud samples (top few centimeters) along the waterline from this site were collected into Ziploc bags using a trowel. Non-habitat mud was collected in the same way from a shaded puddle midge habitat located on a commercial cervid facility in Quincy, FL, USA [23]. Emergence sampling from this habitat indicated that this site was a productive habitat of C. stellifer and no C. insignis emerged from this site. Previous studies reported C. insignis to be associated with cattle pastures and livestock facilities [16][17][18]. Therefore, mud mixed with fresh cattle manure (25.0% w/w) was also tested for its oviposition attractiveness. Overall, four different two-choice assays were conducted in various combinations: 1) Deionized water (DI) vs. Deionized water, 2) DI vs. mud, 3) mud + cattle manure vs. mud, 4) habitat mud vs. non-habitat mud. Each experiment had 5 -8 replicates using a single midge in each replicate. At least two trials were conducted per each experiment ( Table 1).
The eggs deposited during the oviposition studies were used for the larval rearing experiments. Midge larval rearing was conducted using methods described previously for C. stellifer [24]. Brie y, the eggs deposited during the oviposition studies (24 -36 h old) were placed in Petri dishes (60 × 15 mm) containing 0.3% (w/v) agar slants and allowed to hatch. The larvae were given a diet of Panagrellus redivivus Linnaeus nematodes (Carolina Biological Supply Company, Burlington, NC, USA) that were replenished every Monday, Wednesday, and Friday (~2 mg/day). The life history traits recorded were egg stage duration, egg hatch rates, larval survival rates to pupal stage, larval stage duration, pupal stage duration, adult eclosion rates, and sex-ratio of the emerged adults. Each Petri dish had 10 -20 eggs with four to six dishes per trial. Three independent trials were conducted overall ( Table 2). All laboratory experiments were conducted at 26 ± 1°C, 60 -80% RH, and 14:10 (L:D) h photoperiod cycle.

Statistical analysis
For the oviposition experiments, only gravid females were included in the statistical analyses. Variation in the number of eggs produced by gravid females and percentage of eggs retained by ovipositing females across the experiments was analyzed using generalized linear mixed-effects models (GLMM; package lme4) with the variation arising from trials and females incorporated as a random-effect under binomial or Poisson distributions. The oviposition preferences exhibited by C. insignis in the two-choice assays were analyzed using generalized estimating equations (GEE; package geepack) with a logit link function under binomial distribution taking into consideration the percentage of eggs retained by ovipositing females. As each midge was given two choices for oviposition, these dishes were regarded as a cluster and an exchangeable correlation structure was incorporated in the models [25]. Variation in the percentage of egg batch deposited (control + treatment) by females during different experiments was analyzed using generalized linear models (GLM) with a binomial distribution. For the larval rearing experiments, differences in the egg hatch rates, larval survival rates to pupal stage, larval stage duration, pupal stage duration, eclosion rate, and F1 adult sex-ratio between trials were analyzed using GLMM with the variation arising from females as a random-effect under binomial or Poisson distributions. All data were analyzed using R statistical software v.3.6.1 using the packages MASS, car, lme4, and geepack [26][27][28][29][30].

Discussion
Overall, our study demonstrates useful methods to collect live C. insignis midges from the eld, bloodfeed them under laboratory conditions, collect eggs from gravid females, and rear the larvae till adulthood, and provides valuable insight into the life history traits of C. insignis, an important vector of BTV/EHDV in Florida. Although considered a mammal biter, C. insignis females showed satisfactory blood-feeding rates on live chicken in the laboratory. These blood-feeding rates can possibly be increased further by using other laboratory animals (mammals) and/or altering starvation periods or environmental conditions during blood-feeding. In addition, fecundity of C. insignis may also be potentially increased by using a mammalian blood source (versus avian blood source used in this study) as host blood meal source can alter fecundity in hematophagous species [31,32]. However, further studies will be needed to test these hypotheses.
During the oviposition experiments, gravid females deposited a distinctly higher number of eggs on substrates with habitat mud over DI controls, suggesting that mud from the larval habitat provides strong oviposition cues to C. insignis. However, the number of eggs deposited on substrates during mud + cattle manure vs. mud trials and habitat mud vs. non-habitat mud trials was not signi cantly different, suggesting that organically enriched muds other than the habitat mud are also attractive for the oviposition of this species. However, these results should be interpreted cautiously as they could be an artifact of the small sample size of the study (only 12 females oviposited in these two experiments [ Table 1]). In addition, whether or to what extent mud from the larval habitat mud of C. insignis examined was already enriched with organic matter is currently unknown (cattle had open access to this site on the Archbold ranch). Moreover, if olfactory cues are involved in the oviposition site selection of C. insignis, the set-up of the experimental design could have caused errors in the recognition of preferred substrates as the two dishes were placed close to each other in the relatively small sized paper cups. Further studies using Y-shaped olfactometer bioassays could examine whether these results are biologically signi cant and determine whether these oviposition cues are olfactory/tactile in nature. Further studies are also needed to examine whether other natural sources from the habitat such as vegetation play a role in the oviposition of C. insignis (vegetated water bodies often harbor C. insignis larvae) [14,18]. Previously, mud and/or vegetation (Sphagnum spp. moss) from the larval habitat were found to strongly in uence the oviposition of C. stellifer and Culicoides impunctatus Goetghebuer under laboratory conditions [22,33]. Currently, very little is known regarding the oviposition preferences and/or habitat requirements of C. insignis and other important Culicoides species in North America [14,18,22,23]. Future studies characterizing the larval habitat of C. insignis, examining the physicochemical properties of the breeding site, and identifying the key biotic/abiotic factors in uencing oviposition site selection of this species in nature are warranted. This information, in the long term, can be potentially exploited to design novel sampling/control strategies targeting gravid females and to manipulate local habitats to discourage the oviposition of C. insignis.
The large variation in the number of eggs produced and the percentage of females that developed eggs in the study was not unexpected. This variation may have arisen due to variation in the blood meal size ingested by females (partially-engorged females were also included in the study) and due to variation in the mated status of the females used in the study (midges were eld-collected) respectively, patterns that have been observed in mosquitoes [34,35]. However, variation in the percentage of females that oviposited and variation in the percentage of egg batch deposited by ovipositing females likely represents a differential preference for the available oviposition substrates. For example, very few females deposited a very small percentage of their egg batch during the DI vs. DI trials compared to the other experiment trials, suggesting avoidance of DI substrates and a preference of habitat mud and other organically enriched muds. Future studies that require oviposition and/or collection of eggs from C. insignis in the laboratory may bene t by providing organically enriched substrates to gravid females.
It was interesting that among the 34 females that oviposited across the study, 44% (15/34) demonstrated skip oviposition as they deposited eggs on both the dishes available. Such behavior was documented previously in C. stellifer as well but appears to be more common in C. insignis than in C. stellifer (9% females) [22]. Skip oviposition has been well studied in numerous container-breeding mosquito species [22,[36][37][38]. It is believed that skip oviposition is advantageous in resource limited habitats (such as arti cial containers, plant pitchers, tree holes, or others) as it enhances larval survival by reducing larval densities. However, the role of skip oviposition on the survival of mud breeding species such as C. insignis and C. stellifer is currently unknown. Further studies will be needed to understand the role of skip oviposition on the ecology/survival and other life history traits of mud breeding Culicoides species. Further studies will also be needed to examine whether or to what extent skip oviposition occurs in dungbreeding and tree-hole dwelling Culicoides species.
The egg hatch rates of C. insignis varied signi cantly across the study, which was not unexpected. It is likely that fertilization status of the eggs varied possibly due to variation in the age of the eld-collected females and/or the mated status of the males these females mated with in the eld. Previously, the priormated status of males was found to affect egg fertilization rates in tephritid ies and butter ies [39,40]. The signi cant variation in larval survival rates and larval stage durations across the study were also not unexpected. It is possible that this variation arose due to, 1) variation in the age/nutritional status of females the eggs were obtained from as parental nutrition can affect larval development traits in insects [41,42], 2) variation in the number of eggs placed in the larval dishes (ranged from 10-20) as larval densities can affect insect larval development [43][44][45], and/or 3) age/condition of the nematodes used as midge larval diet as early instar midge larvae may have had di culties capturing/ingesting adult nematodes.
Overall, the agar/nematode method was convenient and effective for the larval rearing of C. insignis. All larval instars could be seen moving through the agar and in/out of the standing water freely. The late instars were frequently observed engul ng nematodes whole while the early instars probably fed on the nematode pieces/carcasses and/or microbial community of the medium. Interestingly, C. insignis larvae (late instars) were also observed to feed on dead conspeci c larvae, suggesting that Culicoides larvae are omnivorous opportunistic feeders. Pupation occurred mainly on the surface of the agar, but the pupae were also found oating in the standing water, albeit with less frequency. Although the larval development of C. insignis was successful, sex-ratio of the F1 adults was male-biased, which may not be desirable for potential colony maintenance. The reasons behind this outcome are currently unknown.
However, it is likely that the nematode diet used could not satisfy nutritional requirements of the female larvae potentially causing mortality. Previously, female mosquitoes were suggested to require more larval nutrition than males to pupate [46]. Moreover, previous larval rearing studies on Culicoides species using the agar/nematode method reported non-distorted sex-ratios in the progeny only for C. stellifer and C.
circumscriptus Kieffer while the sex-ratios of other species were found to be either male-biased or femalebiased [24,47,48]. It is likely that the larval nutritional requirements of Culicoides midges vary between species. Further studies will be needed, 1) to examine the nutritional requirements of male and female biting midge larvae, and 2) to improve production conditions of C. insignis by potentially incorporating nutritional supplements to the nematode diet or by using other larval diets.
Very little is known regarding the mating behavior of Culicoides species currently. Many species are believed to be eurygamous (need swarming to mate) while some species are stenogamous (will mate in restricted spaces) [49][50][51]. Our attempts at inducing swarming/mating in the F1 generation of C. insignis by using host cues (octenol), environmental cues (habitat [mud + cattle manure] and dawn/dusk conditions), varying light colors (blue, green, and red), and cage sizes (capillary tubes with terminalia in contact to large 47.5 × 47.5 × 47.5 cm BugDorm cages) were all unsuccessful (non-mating was inferred as F1 females did not deposit viable eggs post blood meal). The reproductive behavior of C. insignis has not been reported to date. Further studies will be needed to investigate the mating habits/cues of C. insignis in nature, which may offer clues towards providing conditions that encourage captive mating in this species.

Conclusions
Overall, our study provides valuable insight into the oviposition and larval development traits of C. insignis, an important vector of BTV/EHDV in Florida. Our study indicates that mud from the larval habitat and other organically enriched substrates provide strong oviposition cues to C. insignis gravid females and demonstrates that the agar/nematode method is effective for the larval rearing of C. insignis. Further studies are needed, 1) to determine the nature of the oviposition cues (olfactory/tactile), 2) to identify the key biotic/abiotic factors in uencing midge oviposition in the eld, 3) to resolve the issue of male-biased sex-ratios in the progeny, and 4) to examine the reproductive behavior of C. insignis in nature. Collectively, this information may possibly be used to create laboratory conditions that encourage captive mating in this species and, in the long term, may be potentially exploited to design novel sampling/control strategies targeting gravid females and/or to discourage the oviposition of C. insignis in important habitats around livestock facilities. Life history traits of Culicoides insignis adult females under laboratory conditions, A) blood-feeding rates on live chicken (± 95% CI), B) number of eggs deposited (mean ± SE), C) percentage of egg batch deposited (mean ± SE), and D) percentage of females that deposited eggs on one or both dishes (± 95% CI). Numbers above bars in A indicate number of midges blood-fed/total number of midges. Asterisk in B indicates signi cantly higher (marginally [P = 0.0780]) number of eggs deposited on mud substrates compared to DI controls, while asterisk in C indicates a signi cantly lower percentage of eggs deposited during DI vs. DI trials (on both dishes combined) than during other experiment trials. Numbers above bars in D indicate number of females depositing eggs on one or both dishes/total number of females that oviposited.

Figure 2
Life history traits of the immature stages of Culicoides insignis under laboratory conditions, A) egg hatch rates (mean ± SE), B) larval survival rates, C) larval stage duration, and D) pupal stage duration. Letters above bars indicate signi cant differences between trials (P < 0.05).