Does salinity affect body proportions and “size/mass” ratios of highly halotolerant Baeotendipes noctivagus larvae (Diptera, Chironomidae)?

Baeotendipes noctivagus (Kieffer, 1911). For those samples where 10 or more larvae of a sufficient size range were measured, the parameters of equation (2) were calculated (Table 2). A high correlation between these characteristics was found in all samples; the average correlation coefficient was 0.972 (SD = 0.022, CV = 0.023). No significant differences were found in the values of the power coefficient “m”; its average value was 0.812 (SD = 0.144, CV = 0.177) and significantly differed from 1 (p = 0.001). The head capsule width of 7 mm long larvae was calculated using equation (2). Table 2 shows that the maximum differences computed for different samples did not exceed 20%. No direct effect of salinity on the calculated value of the head capsule width was found. On average, the width of the head capsule in larvae of this length was 0.254 mm (SD = 0.023, CV = 0.090). Based on the available data, no correlation was found between the width of the head capsule and the body length (or their ratio) and temperature and pH (Table 1). When the array of measurements was analyzed as a whole, all points were grouped around two lines (Fig. 2). They can be approximated by two equations, the first one (R = 0.934, p = 0.0001):

Figure 2

and the second one (R = 0.921, p = 0.0005):

The calculated width of the head capsule in one-size larvae (7 mm) was 0.551 mm in one group and 0.253 mm in the other; the values differed 2.2 times. The authors analyzed the effect of salinity on the width of the head capsule in two groups (Fig. 3). Salinity affected the width of the head capsule in the group with wider head capsule; it may be approximated as (R = 0.997, p = 0.005):

Figure 3

where S– salinity, PSU.

A weak positive correlation was found for another group (R = 0.569, p = 0.05). Salinity, temperature and pH did not determine what group would be exist in the lake. It should be noted that in one particular water body, representatives of only one group with a smaller head capsule width were usually present.

Cricotopus gr. sylvestris (Fabricius, 1794). This species was represented by a sufficient number of larvae to perform the calculations for only three samples (Table 2). The correlation coefficient averaged 0.867. The calculated coefficients of three equations differed significantly (p = 0.01), despite the fact that salinity was nearly the same – 9 and 10 PSU. The calculated widths of the head capsule for larvae of 7 mm in length in these samples differed 2.8 times. All points were grouped around two lines when the whole array of measurements was analyzed (Fig. 4). Salinity does not explain this separation.

Figure 4

Paratanytarsus confuses Palmen, 1960. This species was represented by a sufficient number of larvae to perform calculations for only two samples collected at salinity 0 and 10 PSU (Table 2). There was no significant difference in the relationship.

Baeotendipes noctivagus. Equation (3) was used for 16 samples collected in 10 lakes at different times (Table 3). On average, the correlation coefficient was 0.950 (SD = 0.037, CV = 0.039). The exponent “b” varied from 1.43 to 3.06 and was an average 2.214 (SD = 0.502, CV = 0.227). There was no significant correlation between this parameter and salinity. Based on the obtained parameters of the equations for each sample (Table 3), the mass of 7 mm long larvae (Table 3, W₇) was calculated. The mass ranged from 0.384 mg to 2.139 mg; the maximum mass was 5.6 times higher than the minimum one. The range of differences in the weighed mass of one-size larvae in all samples was almost the same. A general trend of mass reduction with increasing salinity was observed in one-size larvae. Based on the determination coefficient (R²), no more than 40% of the total variability in the body mass of one-size larvae can be explained by changing salinity. The analysis of the salinity effect on the calculated mass of 7 mm long larvae shows that the data are grouped around two curves (Fig. 3). The first curve can be approximated by the equation (R = 0.946, p = 0.001):

where W₇ – the mass of 7 mm long larvae, mg; S – salinity, PSU, and the equation (R = 0.902, p = 0.001):

No correlation was found between the mass and its ratio to size traits and temperature and pH.

The authors also calculated the equation using all measurements (1167) for this species (R = 0.898, p = 0.0005):

The mass of 7 mm larvae, calculated using this equation, was 1.039 mg. This value is 2 times smaller than the largest mass in the analyzed samples and 2.7 times larger than the smallest mass in those samples (Table 3).

Cricotopus gr. sylvestris. Calculations were performed for three samples only, collected at salinity 9 and 10 PSU (Table 3). Significant differences were found in the equation parameters (p = 0.005) and in the calculated mass; larvae with a length of 7 mm differed 1.6 times.

Paratanytarsus confuses. Calculations were performed for two samples only, collected at salinity 0 and 10 PSU (Table 3). Calculated mass of the 7 mm long larvae was significantly higher in freshwater than in saline water.

Table 4 presents the parameters calculated by equation (3). Parameters for different samples differed greatly in the case of B. noctivagus. Correlation coefficients between the body mass and the width of the head capsule were on average lower than between the length and mass; the average value was 0.903 (SD = 0.113, CV = 0.125). The body mass of larvae with a head capsule width of 0.250 mm was calculated using the obtained equations (Table 3). The calculated mass ranged from 0.052 mg to 2.025 mg, averaging 0.897 mg (SD = 0.477, CV = 0.532). The calculated maximum mass was 8.7 times higher than the calculated minimum mass. The calculated mass was not directly correlated with salinity. This procedure was also applied to three samples of C. gr. sylvestris (Table 3). The parameters of these equations also differed significantly.

Based on the available published data, the power coefficients “m” in parabolic equation (2) describing the relationship between the width of the head capsule and the body length of chironomid larvae representing seven species were calculated: Chironomus decorus Johannsen, 1905 – 0.679 (Maier et al. 1990), C. ornatus (Meigen, 1818) – 0.856 (Swanson & Hammer 1983), C. xanthus Rempel, 1939 – 0.664 and 0.639 (Fonseca & Rocha 2004), C. zealandicus Hudson, 1892 – 0.856 (Graham & Burns 1983), Diamesa incallida (Walker, 1856) – 1.23 (Nolte & Hoffmann 1992), Orthocladius nitidoscutellatus Lundström, 1915 – 1.028 (Kravtsova 2007), Orthocladius sp. – 0.976 (Kravtsova 2007), and P. baicalensis (Tshernovskij, 1949) – 0.835 (Kravtsova 2007). For 10 species, including those listed in Table 2, the average value of the exponent “m” is 0.807 (SD = 0.171, CV = 0.214). The exponent “m” is in most cases significantly below 1. This suggests that the rate of increase in the body length of chironomid larvae during their growth is generally significantly higher than the growth rate of the head capsule width. One-size larvae of different species have a varying head capsule width, which means that, in general, the equations of relationships between the body length and the head capsule width are species-specific, as previously reported (Stoffels et al. 2003). Our data on B. noctivagus and C. gr. sylvestris suggest that intraspecific dimorphism in the head width may exist among chironomid larvae. The factor/factors determining this dimorphism are unknown. The second coefficient “k” in equation (2) varies; it cannot be considered as a definite constant. Data on B. noctivagus and C. gr. sylvestris (Table 2) show that environmental conditions affect the coefficient “m” and the width of the head capsule in one-size larvae. It is possible that salinity affects the ratio of the body length and the head capsule width but without one general trend. No effect of salinity on the body proportions was previously found in other arthropods (Anufriieva & Shadrin 2015). It is likely that the trophic factor has a stronger effect on the “body length/head capsule width” ratio. This conclusion is supported, in particular, by experiments on larvae cultured on different types of nutrient medium (Fonseca & Rocha 2004; Rech et al. 2014).

The exponent “b” in equation (3) describing the relationship between the body mass and the length of chironomid larvae significantly varied in our data (Table 3) – from 1.566 to 3.057. Values of this parameter provided in the available literature and calculated by the authors of this paper using published data also vary over a wide range: P. baicalensis – 2.282 (Kravtsova 2007), O. nitidoscutellatus – 2.716 (Kravtsova 2007), Orthocladius sp. – 2.239 (Kravtsova 2007), C. zealandicus – 2.571 (Graham & Burns 1983), C. decorus – 2.860 (Maier et al. 1990), Diamesa cinerella Meigen in Gistl, 1835 – 2.982 (Schütz & Füreder 2018), and from 2.340 to 3.347 in six species of New Zealand chironomids (Stoffels et al. 2003). Based on available data, i.e. our data and those derived from the literature (Graham & Burns 1983; Maier et al. 1990; Stoffels et al. 2003; Kravtsova 2007; Schütz & Füreder 2018), the mean value of “b” calculated for 15 chironomid species from different regions was 2.494 (SD = 0.833, CV = 0.334). In the past, two values of coefficients “b” were also calculated for larvae of chironomids as a whole – 2.04 (Toderash 1984) and 2.781 (Balushkina 1987). However, our data and the analysis of the published results allow us to conclude that it is inappropriate to always use one coefficient “b” for all chironomid larvae, because this may produce high under- or overestimation. This parameter is species-specific and depends on the conditions in which the larvae develop. It was suggested earlier, based on a more limited data set, that the coefficient “b” decreases with increasing salinity (Belyakov et al. 2018). However, our new data on B. noctivagus show that there is no overall trend in the salinity effect on this coefficient. It follows from our data (Fig. 3) as well as from equations (3) and (4) that the mass of one-size larvae of B. noctivagus and P. confuses decreases with increasing salinity. Equations (8) and (9) show that some other factors may affect this relationship. It is likely that the trophic factor is more important in determining the “body length/ mass” ratio and this assumption was supported by experiments (Fonseca & Rocha 2004; Rech et al. 2014) and field study (Schütz & Füreder 2018). This effect is more pronounced in larger larvae (Shadrin et al. 2017; Belyakov et al. 2018). The decrease is probably associated with an increase in energy expenditure to maintain metabolism and a decrease in the growth rate at high salinity (Kokkinn 1990; Cartier et al. 2011).

The exponent “d” in equation (4) of the relationship between the body mass and the width of the head capsule in chironomid larvae (Table 4) varied significantly in our data – from 1.130 to 3.378, i.e. in a wider range than in the equation of the relationship between the mass and the body length. In the literature, the following values of this coefficient are available: P. baicalensis – 3.960 (Kravtsova 2007), O. nitidoscutellatus – 2.720 (Kravtsova 2007), Orthocladius sp. – 2.440 (Kravtsova 2007), and from 1.641 to 4.020 in six species of New Zealand chironomids (Stoffels et al. 2003). The average value of “d” for 12 species of chironomids from different regions is 2.477 (SD = 0.844, CV = 0.341). In the past, the value of the coefficient “d” was calculated for the chironomid larvae as a whole – 3.280 (Balushkina 1987), which, as demonstrated above, is impractical for chironomid larvae under real conditions. This parameter is species-specific and depends on the conditions in which the larvae develop. At the same time, our data on B. noctivagus show that there is no overall trend in the effect of salinity on this coefficient. Probably, the trophic factor has a strong influence on the coefficient of the equation as well as some other factors (temperature, oxygen concentration, etc.). When comparing the calculation of the body mass with the body length and the width of the head capsule, it can be observed that – as in the case of B. noctivagus, according to our data, and in the case of other species, according to published data (Stoffels et al. 2003; Kravtsova 2007) – the variability in values is larger and the determination coefficient (R²) is smaller in the case of the mass/head capsule width ratio. Therefore, it is better to calculate the body mass using the body length rather than the width of the head capsule. Given the close relationship between the width of the head capsule and the body length, one can confidently use one of the parameters to calculate the other.

Two phenotypically distinct morphs differing in body size and body proportion were found among chironomid larvae of the studied species. A similar population polymorphism was observed in various animal taxa and it is assumed that it increases the adaptability of populations (Huston & DeAngelis 1987; Anufriieva & Shadrin 2014b). Such polymorphism may result from the interaction of different environmental characteristics with a population gene pool and the critical factors, which include: (1) growth rate variations among individuals resulting from intra-population metabolic communications; (2) polymorphism in a population gene pool; (3) epigenetic diversity resulting from alternative gene expression and epigenetic mechanisms (Huston & DeAngelis 1987; Goldberg et al. 2007; Anufriieva & Shadrin, 2014b). “The phenotype can be represented as a branching system of trajectories in a phase space spreading along the time axis” (Waddington 1940). B. noctivagus lives under a wide range of salinity, and probably it is possible only realizing different adaptive norms. It is easier to conclude that polymorphism results from genetic diversity or alternative discrete realizations of the same genotype. The available data do not provide an explanation as to whether this polymorphism is caused by genetic diversity or phenotypic plasticity. It is an intriguing issue and further research is needed.

When comparing the results obtained with fixed and non-fixed larvae of chironomids, it should be considered that with prolonged storage in formalin or alcohol, larvae can lose on average 20–30% of their original mass (Donald & Paterson 1977). Taking this into account, it should be assumed that in our study there were cases of larvae mass underestimation. It is likely that the “a” and “c” indicators in equations (3) and (4) should be increased by 20–30%. We assume that when the larvae of different sizes were fixed, their mass loss was at the same rate, and, consequently, the coefficients “b” and “d” in equations (3) and (4) for the living body mass remained unchanged.

General conclusions: when analyzing populations and taxocenes of chironomid larvae in different water bodies, it is inappropriate to use general equations for the mass–body relationship with linear traits. Salinity affects the larvae mass and “length–mass” ratio but cannot be considered as a single factor affecting the size and mass of chironomid larvae in water bodies. All factors act together. More detailed studies of the effects of various factors (salinity, temperature, oxygen concentration, quality and quantity of food, etc.) on the parameters of the equations are needed.