Egg masses of the Chironomus Brazilian population were collected in tap water accumulated in a 10 L plastic container, for several days (5^th, 6^th, 8^th, 10^th and 14^th), in January 2011 in the urban area of Manaus municipality Amazonas, Brazil (03°06'50.17"S, 59°58'30.99"W). The egg masses were placed individually in 80 mL plastic vials, which were labeled with collection information, covered with a screen and observed daily until the larvae hatched and abandoned the gelatinous mass. Larvae from each egg batch were transferred to a plastic tray (19.5×31×6.5 cm) containing burned sand as substrate and 1.5 L of water (pH = 5.9; electrical conductivity of 20.7 µS cm^-1). The trays were covered with wooden structures (40×21×32 cm) serving as frames for screens (2 mm mesh), following a model modified from Fonseca and Rocha (2004) for retention of adults and behavioral observations.

Larvae were fed fish food (TETRAMIM®) every 48 hours. The colony established using this collected material was kept in the insect-raising facility at the Coordenação de Biodiversidade, Instituto Nacional de Pesquisas da Amazônia (INPA), at environmental conditions similar to those of the climate in Manaus during the months from May to September 2011: temperature of 26 ± 0.3 °C; air humidity of 75 ± 7.7% and photoperiod of 12/12 hours (data obtained from http://www.inmet.gov.br). The specimens analyzed in the present study were obtained from this colony.

Emerged adults with pupal and larval exuviae from the colony were dissected and mounted on slides with Euparal® following the procedures outlined by Epler (1988). The morphological characteristics analyzed were male genitalia, anal spur and distribution of shagreen in pupae and, in structures present on the larval head capsule (antenna, pectin epipharyngis, premandible, mandible, mentum, labral setae). Comparisons between the Brazilian Chironomus population (n = 40; 10 adult male, 10 adult female, 10 pupae and 10 larvae) and other Chironomus species were made using available species descriptions and related literature (Kieffer 1910; Tokunaga 1936; Chaudhuri et al. 1992; Martin 2014). We also examined specimens of C. striatipennis (n = 4; 1 male adult, 1 pupa and 2 larvae) from India (lent by Dr. Jon Martin, University of Melbourne, Australia) and specimens of C. kiiensis (n = 30; 10 male adult, 5 adult female, 5 pupae and 10 larvae) from Japan (lent by Dr. Masaru Yamamoto, from Yamaguchi Prefecture). The measurements and images presented in this study were obtained using an Olympus compound microscope with a mounted digital photographic camera, model Olympus DP72, using Cell D (Olympus) software and a stereomicroscope (Leica M165C), with Leica software (auto montage, Application Suite V3).

Molecular analyses were done using the DNeasy Blood & Tissue (Qiagen) kit following the manufacturer’s recommendations. Amplifications of the extracted DNA from three specimens (2 larvae and 1 adult male) of the Brazilian Chironomus population were made using primers developed by Folmer et al. (1994) that are specific for the Cytochrome Oxidase I (COI) gene in the mitochondrial DNA. The amplified fragments were purified by the EXO-SAP (Exonuclease I-Shrimp Alkaline Phosphatase) method and sent to the Centro de Estudo do Genoma Humano (Universidade de São Paulo), where they were sequenced using an ABI 3730 DNA Analyzer.

Sequences for 14 Chironomus species and Lipiniella fujiprimus (Sasa, 1985) available at the GenBank and Barcode of Life Data Systems (BOLD) were used in the analysis (accession numbers in Table 1). The alignment and editing of the sequences were done in BioEdit software v.5.0.6 (Hall 1999). A tree was constructed using the neighbor-joining method with 1000 bootstrap replicates, and genetic distances were estimated using the Kimura-2-parameter (K₂P) model in the Mega5 program (Tamura et al. 2011).

Voucher specimens of the Brazilian population are deposited in the Coleção de Invertebrados do Instituto Nacional de Pesquisas da Amazônia. Haplotype sequences are deposited in GenBank under the accession numbers KJ424334–KJ424336.

The egg masses were characterized by their shape, length, width and number of eggs; the maximum length and width of eggs were measured. In order to determine the development time of the egg stage, five egg masses were isolated and observed every hour until first instar hatched.

Ten egg masses were isolated to determine the development time of each of the four larval instars; starting from the moment at which the first instar hatched, three larvae were fixed (in 80% ethanol) daily, until the last larva of the egg mass pupated. To identify the instar of each of the fixed larvae; they were mounted between slide and coverslip (using Hoyer as the mounting medium) to measure the ventral length of the head capsule, following the methodology of Strixino (1973). To classify larvae into one of the four larval instars, the measurements obtained were subjected to a frequency-distribution analysis; each peak in the graph indicates a larval instar (Strixino 1973). The development time of each instar was determined based on the size limits of each larval instar obtained in the frequency-distribution graph and the day when they were preserved. The development time of each larval instar was determined by combining information gathered from the frequency distribution graph with the measurements of the head capsule of the larvae collected daily.

To estimate pupal development time, 50 pupae were observed every 12 hours from the moment of pupation until adult emergence. Longevity of adults was estimated using 50 adults that emerged in the laboratory on the same day; these were isolated in pairs (male and female) in cages made of PET bottles and observed until there were no more survivors.

The morphology of the Brazilian Chironomus population (adult, pupal and larval) is identical to that presented in the original descriptions of C. striatipennis and of C. kiiensis (Kieffer 1910; Tokunaga 1936), and in other taxonomic papers on these two species (Chaudhuri et al. 1992; Martin 2014), it being impossible to distinguish them from each other morphologically. The examined specimens of C. striatipennis (Fig. 1A, D, G) from India and of C. kiiensis (Fig. 1B, E, H) from Japan were also morphologically indistinguishable from the Brazilian Chironomus population (Fig. 1C, F, I).

Figure 1. 

Adult male and pupae. Chironomus striatipennis, Indian population. A Wing D Hypopygium, dorsal view G Anal spur, dorsal view. Chironomus kiiensis, Japanese population B Wing E Hypopygium, dorsal view H Anal spur, dorsal view. Chironomus striatipennis, Brazilian population C Wing F Hypopygium, dorsal view I Anal spur, dorsal view. Scale bar: 500 µm (A, B, C, G, H, I); 200 µm (D, E, F).

In the neighbor-joining tree (Fig. 2), we observed that the Brazilian Chironomus population (C. sp1BRA, C. sp2BRA and C. sp3BRA) grouped, with 94% bootstrap support, with others specimens of C. striatipennis and of C. kiiensis from different regions of East Asia. The analysis resulted in two groups, each with 100% bootstrap support, which we named “Group A” and “Group B”. Group A is composed of the Brazilian Chironomus population and specimens identified as C. kiiensis from South Korea and Japan and C. striatipennis from Malaysia and India. Group B included specimens identified as C. kiiensis, also from South Korea and Japan.

Figure 2. 

NJ tree based on the COI sequences of the mtDNA of Chironomus (Diptera: Chironomidae) species. The sequence of Lipiniella fujiprimus was used as the outgroup. Bootstrap values > 50% are shown on branches. Accession numbers and countries are provided beside the species names. Species flagged with an asterisk (*) are neotropical species. Brazilian Chironomus population: C. sp1BRA; C. sp2BRA; C. sp3BRA

Chironomus species from other regions, including the three specimens from Brazil, were included in a distinct clade (Fig. 2). Considering that Groups A and B represent monophyletic groups, based on the intraspecific genetic distance between their members, we observe that in Group A the genetic distance varied from 0.0 to 3.3%, while in Group B the distance varied from 0.6 to 2.0% (Table 2). The average genetic distance between sequences from Groups A and B was 9.6% (Table 3), and within each group the average genetic distance was 1.3%.

Within Group A there are three specimens of C. striatipennis from Malaysia and India with a mean genetic divergence of 2.6%, a higher value considering that the mean divergence between the remaining group members was 0.6% (Table 3). The mean genetic distance between Groups A, B and other Chironomus species, including Neotropical species, was 15.6%, ranging from 10.5 to 19.5% (Table 3).

Egg masses of the Brazilian Chironomus population were found attached by a stem in the wall of a plastic tray with water. The egg masses measured 16.6 mm (SD = 3.1; n = 5) in length and 1.59 mm (SD = 0.04; n = 5) in width (median region). Each mass contained an average of 600 eggs (SD = 104; n = 5). The eggs were elliptical in shape and measured, on average, 0.24 mm (SD = 0.02; n = 50) in length and 0.10 mm (SD = 0.01; n = 50) in width. The mean incubation time of the eggs was two days (SD = 0.5; n = 50) at 26 °C, with a hatching rate of 89.9%. The eggs were distributed in a pseudo-spiral pattern, in alternate parallel rows along the primary axis of the gelatinous mass (Fig. 3A) and eggs were eliptic format (Fig. 3B).

Figure 3. 

Chironomus striatipennis, Brazilian population. A Egg mass B Egg.

The four instars of the Brazilian Chironomus population were well defined using the ventral length of the head capsule. Mean length of the head capsule of the 1^st-instar is 54.7 µm (SD = 4.2; n = 81); the 2^nd 93.7 µm (SD = 4.5; n = 53); the 3rd 157.1 µm (SD = 7.2; n = 67) and the 4^th 257.6 µm (SD = 16.7; n = 316) (Fig. 4).

Figure 4. 

Frequency of occurrence of the ventral length of the cephalic capsule of a Brazilian Chironomus population (Diptera: Chironomidae) showing the four larval instars.

Development times of the 1^st (SD = 0.6; n = 81), 2^nd (SD = 1.0; n = 53) and 3^rd instars (SD = 1.1; n = 67) were similar, each averaging three days; the 4^th instar was the longest, with mean development time of 10 days (SD = 2.5; n = 316). Mean time for complete larval stage development was 19 days (SD = 5.2; n = 517). Mean development time for the pupal stage, at 26 °C, was two days (SD = 0.24, n = 50). The life span of adults, on average, at 26 °C, was three days for both males (SD = 0.70, n = 50) and females (SD = 0.65, n = 50). Development time of the Brazilian Chironomus population from the time larvae hatch to the adult stage was 27 days at 26°C. The emergence percentage was 42.9% for females and 57.1% for males.

High morphological similarities observed between the Brazilian Chironomus population and both C. striatipennis and C. kiiensis (Kieffer 1910; Tokunaga 1936) corroborate the studies that reported the difficulty in distinguishing the latter two species at the morphological level, including the one that proposed C. kiiensis as a synonym of C. striatipennis (Martin 2014). However, cytotaxonomic studies have indicated that they are sibling species that belong to the same cytocomplex (Martin 2014). Our molecular analysis showed that both species names have been applied to specimens with high genetic divergence, and that the identification based on geographical distribution, which is currently in common usage to circumscribe both species (e.g., Nath and Lakhotia 1989; Yong et al. 1999; Jeong et al. 2004; Nandi et al. 2011), is not a good taxonomic practice.

Studies on the genetic intraspecific divergence in some Chironomidae genera have reported values between 0.5 and 2.3% (Sinclair and Gresens 2008; Silva et al. 2013; Proulx et al. 2013; Trivinho-Strixino et al. 2012). Similar values were observed in studies on other insects, including some Diptera families (e.g., Beckenbach and Borkent 2003; Hamada et al. 2010; Hernández-Triana et al. 2012). Although Silva et al. (2013) attributed low intraspecific divergence (mean 0.91%) to the specimens having been collected in the same place, our results and other studies, for example, with Lepidoptera and Simuliidae, do not corroborate this hypothesis (e.g., Hebert et al. 2004; Hernández-Triana et al. 2012).

The interspecific genetic divergence observed in our data is in accordance with results for other groups of Chironomidae and other aquatic Diptera families, with values around 15% (Beckenbach and Borkent 2003; Sinclair and Gresens 2008; Proulx et al. 2013; Silva et al. 2013), corroborating the hypothesis that Groups A and B represents at least two distinct species. Group A is composed of specimens identified as C. striatipennis and C. kiiensis from different regions of Asia. Within this group, there are three specimens identified as C. striatipennis from Malaysia and India (the country where this species was described) with high genetic divergence (mean 2.62%) when compared to other members in Group A (Table 3.). This fact might be an indication of the presence of two sibling species in Group A. But, since we have no detailed information on the morphology of all life stages of the populations included in Group A, we suggest that this group should be treated as a species complex. Based on ICNZ rules (Principle of Priority – Article 23.3), and as encouraged by Martin (2014), we named Group A as the C. striatipennis species complex.

The large genetic distance (mean 9%) between Group B specimens (composed of specimens identified as C. kiiensis from Japan and South Korea) and Group A is a clear indication that each clade represents a distinct species. This is also corroborated by the genetic distance observed between specimens in each species group (maximum = 3.3%), and by the interspecific genetic distance values reported by other studies. For example, a study on three Podonomus populations observed that the genetic distances between them were greater than 7%, indicating the presence of three species (Trivinho-Strixino et al. 2012). Since C. kiiensis was described from Japan, we hypothesized that the specimens in Group B may represent this species. On the other hand, because of the lack of information on the morphology of all life stages of these populations, and based on genetic information obtained from GenBank for specimens collected near the type locality of C. kiiensis, we choose do not propose any name for Group B specimens.

The placement of the Brazilian Chironomus population in Group A and its small genetic distance from South Korean specimens (0.8%) indicates that this might represents a recently introduced species in Brazil, probably, from a population from South Korea. This could have occurred due to the fact that Manaus is located in a port zone, receiving many cargo ships from different continents, including Asia, due to the presence of industries in the Manaus Tax-Free Zone (SUFRAMA). We therefore assume that specimens of the Brazilian Chironomus species arrived in Brazil by ship. Other exotic species have been entering the country by this way, perhaps using ballast water or some other source of standing water (e.g. Santos and Lamonica 2008; Brodin and Anderson 2009; Raunio et al. 2009; Jensen 2010).

Environmental conditions where the Brazilian Chironomus population and the Japanese C. kiiensis population were collected were different: the Brazilian population was observed inhabiting oligotrophic water and the Japonese population eutrophic water (Inoue et al. 2008; Al-Shami et al. 2010). The Brazilian Chironomus population has double the number of eggs (~ 600 eggs) compared to C. kiiensis from Japan, although they were reared at similar water temperatures (25–26 °C), feeding methodology and food type were not the same (Maeda and Yano 1988). Larval development time of the Brazilian Chironomus population was shorter (10 days) than that of the Japanese C. kiiensis population (14–20 days) (Nandi et al. 2011). Development times of the 1^st, 2^nd and the 3^rd instars of the Brazilian Chironomus population was shorter than that of the 4^th instar, as has also been observed in other Chironomus species (Fonseca and Rocha 2004; Zilli et al. 2008). The 4^th instar’s long larval development time is probably due to the greatest body growth occurring in this period (Tokeshi 1995), as well the production and maturation of the oocytes (Strixino and Trivinho-Strixino 1985).

The size of last-instar larvae, represented by the ventral length of the head capsule, of the Brazilian Chironomus population was smaller than that of the Japanese C. kiiensis population (Maeda and Yano 1998). Male survival time of the Brazilian Chironomus population under laboratory conditions was shorter (three days) than in the Japanese C. kiiensis population at similar air temperature (25 °C) (Maeda and Yano 1988; Nandi et al. 2011). Environmental conditions can have an effect on species biology and information in this area is important for characterizing the habitat of each species in a species complex and can help to delimit each species in the complex.

In this study we addressed the taxonomic problem involving the species named C. striatipennis and C. kiiensis from Asia, since these names have been used for specimens morphologically similar and closely related. Regional variation among Asian populations of the two above mentioned species was observed, but only at molecular level, based on the genetic distance estimated using a partial COI sequence. We hypothesize that these two names encompass at least two, and perhaps three, species, based on sequences deposited in GenBank and BOLD. We indicated that the usual way to identify these two species, which are practically indistinguishable at the morphological level, based on geographical distribution is not a feasible approach, since specimens identified with the same name (C. kiiensis) from Japan and South Korea have highly genetic divergent (genetic distance of 9%). A detailed study, including the morphology of all life stages and a multilocus molecular analysis of populations from the entire distribution needs to be done to solve this taxonomic problem. The presence of an Asiatic Chironomus in Brazil might be the consequence of human globalization, where fast and easy global transportation systems are available to any organisms with the minimum characteristics of alien species. This example also demonstrates that the geographic limits of species cannot be considered in isolation, but rather need to be examined from a broad perspective to avoid mistakes.