Establishing isotopic turnover rates and trophic discrimination factors in tiger beetle (Coleoptera: Cicindelidae) larvae through a diet switch experiment

Introduction

Food webs are exceptionally complicated systems of energy flow between living organisms. By using ecological tracers such as stable isotopes, researchers can track this flow of energy and thus understand what trophic interactions are occurring in a given ecosystem (Parnell et al., 2013). This knowledge aids in conservation, especially in ecosystems with invasive species, as researchers are able to identify energy pathways for at-risk species and invasive species alike (Muñoz, Currin & Whitfield, 2011; Jackson et al., 2012; Holthuijzen et al., 2023). Heavier isotopes tend to become more enriched between trophic levels, a process termed trophic discrimination (Bowes & Thorp, 2015; Veselý et al., 2024). Stable isotopes are reported as a ratio of ratios using the following formula:

$${\delta }^{15}Nor{\delta }^{13}C=[(R_{sample}/R_{standard})-1]\times 1000$$where R = ¹⁵N/¹⁴N or ¹³C/¹²C (Perkins et al., 2014). The nitrogen-15 ratio (δ¹⁵N) increases with each upward step in the food chain and is used to determine an organism’s trophic position (Post, 2002). The carbon-13 ratio (δ¹³C), however, usually remains relatively unchanged throughout the food chain compared to δ¹⁵N and is commonly used to decipher the underlying energy pathways, such as the primary producer, of the particular food web that an organism is part of (Post, 2002; Phillips et al., 2014; Zalewski et al., 2014).

Most models assume δ¹⁵N and δ¹³C trophic discrimination factors (TDFs), which is the difference between the δ¹⁵N or δ¹³C of a consumer and its diet, to be 3.4‰ and 0.4‰, respectively (Post, 2002; Bowes & Thorp, 2015; Quinby, Creighton & Flaherty, 2020). Though isotope mixing models incorporate uncertainties associated with TDFs (Phillips et al., 2014), assuming a mean δ¹⁵N TDF of 3.4‰ and a mean δ¹³C TDF of 0.4‰ may result in incorrect conclusions. This is because TDFs often vary between taxa, and these assumed values were mostly derived from vertebrates and non-insect invertebrates (Post, 2002; Veselý et al., 2024). Additionally, δ¹³C TDFs can be influenced by the physiological status of the organism because lipids synthesized de novo are depleted in ¹³C (DeNiro & Epstein, 1977). This uncertainty is supported by several experiments with insects that have TDFs both higher and lower than the assumed δ¹⁵N TDF of 3.4‰ and δ¹³C TDF of 0.4‰ (Traugott et al., 2007; Jardine et al., 2008; Perkins et al., 2014; Matos et al., 2018). One caveat to this trend is phloem sap feeders, such as aphids, which tend to have negative TDFs (Madeira et al., 2013; Perkins et al., 2014; Zhang et al., 2015). By conducting stable isotope studies on organisms without first establishing species- or genus-specific TDFs, we risk inaccurate assessment of the basal food source of an organism’s food web and the organism’s trophic level, especially at higher trophic levels (Perkins et al., 2014; Phillips et al., 2014; Quinby, Creighton & Flaherty, 2020). For example, if an organism has a δ¹⁵N TDF of 1.7‰, then its trophic position will be underestimated by half when using the commonly applied value of 3.4‰. To achieve more reliable trophic position estimates of a consumer, controlled feeding experiments are needed to understand potential deviations from typical TDFs (Phillips et al., 2014; Quinby, Creighton & Flaherty, 2020).

In addition to TDFs, isotopic half-lives can provide insight into an organism’s metabolism and allow estimation of the duration over which the organism’s diet is represented (Vander Zanden et al., 2015). High tissue turnover rates indicate that the tissue or organism is highly metabolically active and/or that the organism is growing (Sacramento, Manetta & Benedito, 2016; Li, Roth & Detwiler, 2018). Half-lives range widely depending on the organism’s size, temperature, and whether an organism is endo- or ectothermic (Boecklen et al., 2011; Vander Zanden et al., 2015). Tissue type is also a determinant of half-life, but whole-body half-lives are typically calculated in small organisms such as insects and small fish (Vander Zanden et al., 2015). Vander Zanden et al. (2015) developed a dataset of articles published between 1982–2014 in which they found four articles reporting whole organism half-lives of arthropod species. These four articles reported a range of 1.5 days (Hemiptera; Jardine et al., 2008) to 9.3 days (Collembola; Larsen et al., 2011) for δ¹⁵N, and 1.5 days (Diptera; Overmyer, MacNeil & Fisk, 2008) to 10 days (Collembola; Larsen et al., 2011) for δ¹³C. Unlike TDFs, there is no consistent difference between δ¹⁵N and δ¹³C half-lives reported in the literature (Vander Zanden et al., 2015).

Tiger beetles (Coleoptera: Cicindelidae) are a diverse, cosmopolitan beetle family that can be found in many habitat types (Pearson, 1988; Duran & Gough, 2020). As beetles, Cicindelids have a holometabolous life cycle with three larval instars (Pearson, 1988). Between instars and between third instar and pupation they molt, leaving behind a chitinous exoskeleton termed exuvia (Tibbets, Wheeless & Del Rio, 2007). Tiger beetles are carnivorous throughout their entire life cycle and are characterized by their fast terrestrial speeds and long, sickle-shaped mandibles (Pearson, 1988). They may become more prominent bioindicators in the coming years because they often occupy small, localized niches and are sensitive to changes in the environment (Carroll & Pearson, 1998; Rodríguez, Pearson & Barrera, 1998; Svehla et al., 2023). Because they occupy localized niches, several Canadian species and subspecies are at-risk, such as Cicindela formosa gibsoni WJ Brown, 1940, Omus audouini LJ Reiche, 1838, and Cicindela marginipennis PFMA Dejean, 1831 (French et al., 2021; Klymko, 2021; McGregor, 2021).

The purpose of this article is to provide insight into tiger beetle stable isotope TDFs and half-lives. This was done by subjecting wild-caught tiger beetle larvae to a diet switch experiment in a laboratory setting where they were sampled periodically. Stable isotope analysis was done on these samples to observe the isotopic turnover over time and to calculate tiger beetle TDFs. With this knowledge, researchers may apply the values determined here in future food web studies of wild tiger beetles and related insects. To our knowledge, this is the first study to conduct a diet switch on any tiger beetles for this purpose. Based on the available literature, we hypothesized that (1) the δ¹³C TDF will be a small but positive non-zero number, (2) the δ¹⁵N TDF will be below the commonly applied literature value of 3.4‰ but greater than zero, and (3) the δ¹⁵N and δ¹³C half-lives will be rapid in the fast-growing larval life stage and will be comparable to other invertebrates.

Materials and Methods

Results

The tiger beetle larvae were reared successfully in the laboratory. The C/N ratio of the sampled larvae increased over time, with a 43% increase between day 0 and day 36 (Fig. 1A). The increase was linear, with an R² value of 0.66 (P < 0.001). Of the individuals that were not sampled as larvae (n = 29), 93% became pupae, and of those pupae that were not sampled (n = 23), 96% emerged as adults. The first adults emerged 70 days after we initially collected them from the wild, and the last beetle emerged on day 92. All three species of lab-raised adults were smaller on average compared to wild-caught adults, although these size differences were not significant (Table 2). The mean elytron length and width were 3.7% (P = 0.080) and 2.2% (P = 0.421) shorter, respectively, for lab-raised C. repanda. For lab-raised C. hirticollis, the mean length was 5.8% (P = 0.252) shorter, and the mean width was 0.2% shorter (P = 0.973). Finally, the single lab-raised C. duodecimguttata elytron length was 8.1% shorter in length and 5.4% shorter in width.

The nitrogen and lipid-corrected carbon isotopes, but not uncorrected carbon isotopes, varied between tiger beetle larvae and exuviae (Table 1). We found an effect of life stage for δ¹⁵N (F = 4.4, P = 0.012), although this difference was only significant between larvae and exuviae (P = 0.010, Fig. 2A) where the larvae were depleted in ¹⁵N compared to exuviae. Similarly, there was also an effect of life stage for δ¹³C_corr (F = 4.1, P = 0.017) with the difference also being between larvae and exuviae (P = 0.015, Fig. 2B) but the larvae were enriched in ¹³C compared to the exuviae. We found no effect of life stage for δ¹³C (F = 2.4, P = 0.089).

According to the models, the asymptotic values were: for δ¹⁵N, 7.0‰ (±0.3‰ SD), which was achieved on day 14.8, for δ¹³C –24.2‰ (±0.3‰ SD), achieved on day 12.3, and for δ¹³C_corr, –23.1‰ (±0.3‰ SD), achieved on day 14.6 (Fig. 1, Table 1). The Cox and Snell pseudo-R² values were 0.85, 0.66, and 0.66 for the δ¹⁵N, δ¹³C, and δ¹³C_corr models, respectively. When the mean T. ni ratios (Table 1) were subtracted from the asymptotic values, the δ¹⁵N TDF was 1.7 ± 0.2‰, the δ¹³C TDF was –0.6 ± 0.2‰, and the TDF for δ¹³C_corr was –0.2 ± 0.2‰. The TDF was significantly different from zero for δ¹⁵N (F = 48.2, P < 0.001) and for δ¹³C (F = 18.2, P < 0.001), but not δ¹³C_corr (F = 0.9, P = 0.352). The half-life for δ¹⁵N was 11.4 days, for δ¹³C, it was 3.9 days, and finally, the half-life for δ¹³C_corr was 4.7 days.

Discussion

We found support for our predictions that the δ¹⁵N TDF in tiger beetle larvae is lower than the literature-assumed value of 3.4‰ and that the turnover of ¹⁵N and ¹³C in tiger beetles occurs quickly (i.e., days). Similarly, our δ¹³C and δ¹³C_corr half-lives of 3.9 and 4.7 days, respectively, fall in the center of the literature range of 1.5–10 days for whole-bodied arthropods, whereas our δ¹⁵N half-life lies just above ranges reported in the literature (Vander Zanden et al., 2015). Our prediction of a small positive number for δ¹³C was not supported as we calculated TDFs of –0.6 ± 0.2‰ for δ¹³C and –0.2 ± 0.2‰ for δ¹³C_corr, with the latter not significantly differing from zero.

The δ¹⁵N TDF calculated in our study yielded a smaller value than what has been used for isotope mixing models in the literature (Post, 2002; Phillips et al., 2014). Our value of 1.7 ± 0.2‰ is in line with reported TDFs for Elateridae (δ¹⁵N TDF 1.1 ± 0.3‰ to 1.6 ± 0.2‰) but not for other beetles such as Coccinellidae (2.9‰), illustrating the importance of establishing taxon-specific TDFs (Ostrom, Colunga-Garcia & Gage, 1997; Traugott et al., 2007; Quinby, Creighton & Flaherty, 2020). Although the TDF for δ¹³C (–0.6 ± 0.2‰) was slightly negative, it was consistent with values reported by Jardine et al. (2008) and fell within the typical range of ±3‰ reported by Post (2002) After lipid correction, there was no difference between the beetles and their diets, suggesting that the negative TDF for uncorrected δ¹³C was driven by lipid content, as indicated by C/N ratios in beetles as high as 6.5 by the time they reached equilibrium with the new diet. The increase in fat stores may reflect the energy needed to overwinter, pupate, and avoid internal freezing (Jardine et al., 2008; Enriquez et al., 2022).

Future studies on wild tiger beetle food webs may use our reported δ¹⁵N value of 1.7 ± 0.2‰ to determine a more accurate diet composition and trophic position than if one were to use the traditional assumption of 3.4‰ (Bowes & Thorp, 2015; Morente & Ruano, 2022). The assumed TDF of 0.4‰ for δ¹³C should be replaced with a value of zero if lipid correction is applied, meaning that tiger beetles should closely match those of their prey’s underlying sources of primary production.

Tissue turnover rates are variable based on metabolism and growth rate, and thus, would change at different temperatures, and in adults as they cease to grow post-eclosion (Fry & Arnold, 1982; Sacramento, Manetta & Benedito, 2016; Li, Roth & Detwiler, 2018). Wild tiger beetles likely consume food less frequently than the larvae we reared in the laboratory which would translate to a slower growth rate in wild specimens and therefore slower turnover rates (Pearson & Knisley, 1985). Regarding temperature, lower or higher temperatures than our experimental 18–24 °C range would lead to slower or faster turnover rates, respectively, due to how temperature influences metabolism (Acar et al., 2001). Therefore, if our half-lives of 11.4 days for δ¹⁵N, 3.9 days for δ¹³C, and 4.7 days for δ¹³C_corr are to be used in future research, temperature and life stage must be considered. Nonetheless, our reported half-lives for δ¹³C and δ¹³C_corr are well within the arthropod half-life range of 1.5–10 days found in Vander Zanden et al. (2015), and our δ¹⁵N half-life is only slightly above their longest reported half-life of 9.3 days, but far below the half-lives of some vertebrate tissue. This is consistent with our hypothesis that tiger beetle larvae have rapid stable isotope half-lives because larvae are small and grow rapidly. These fast turnover rates may allow researchers to track recent dietary shifts in wild tiger beetle populations (Vander Zanden et al., 2015).

Past studies have found that adult holometabolous insects, apart from beetles (Tibbets, Wheeless & Del Rio, 2007), are enriched in ¹⁵N and depleted or unchanged in ¹³C compared to larvae and pupae (Barriga et al., 2013; Matos et al., 2018). Our findings that δ¹⁵N, δ¹³C, and δ¹³C_corr did not vary between tiger beetle adults, larvae, and pupae are consistent with Tibbets, Wheeless & Del Rio (2007) and support the application of our TDFs (δ¹⁵N: 1.7‰ and δ¹³C: 0‰) to both larval and adult tiger beetles. Hypothetically, this may be because the beetle species tested in Tibbets, Wheeless & Del Rio (2007) and our study do not excrete meconium, a metabolic waste product, prior to pupation, which tends to be depleted in ¹⁵N, resulting in adults being enriched. Alternatively, beetle meconium may simply not be depleted in ¹⁵N, which was not tested here or in Tibbets, Wheeless & Del Rio (2007). Or, more simply, it may have been due to our small sample sizes. Our results for the exuviae, which were significantly enriched in ¹⁵N compared to larvae, are consistent with Tibbets, Wheeless & Del Rio (2007) but not with Traugott et al. (2007), who found no significant differences between exuviae and larval segments. Traugott et al. (2007) also found no differences between larval segment δ¹³C and exuviae δ¹³C, which is consistent with our findings. Conversely, after correcting for lipids, ¹³C was significantly depleted in our exuviae compared to larvae. Future research should test stable isotope ratios before, during, and after pupation on other beetle species, including sampling exuviae and meconium, to fill in these knowledge gaps.

The methods used to collect and rear tiger beetles were effective for species that live by the edge of a large river. The larvae were able to pupate without having to overwinter as third instars as typically occurs under natural conditions (Hamilton, 1925). Although on average the lab-raised adults were smaller than wild-caught adults the difference was not statistically significant. This was likely due to our small sample size of lab-raised tiger beetles. The high frequency of successful emergence demonstrates the efficacy of the method used in our experiment.

Tiger beetles have the potential to become a useful bioindicator of ecosystem health because they are globally distributed, yet at the species level, they occupy habitats that are sensitive to change (Rodríguez, Pearson & Barrera, 1998; Cassola & Pearson, 2000). Currently, however, only a handful of studies have used them in this capacity (Carroll & Pearson, 1998; Svehla et al., 2023). The values reported here may be used in future studies examining tiger beetle feeding ecology and food web interactions as indicators of habitat and community health. To make accurate extrapolations about tiger beetle food web ecology, potential prey and primary producers should be analyzed alongside the tiger beetles (Phillips et al., 2014). This is so that energy pathways can be matched to the δ¹³C of tiger beetles (Zhang et al., 2015). This same principle applies to identifying tiger beetle prey and trophic position using δ¹⁵N (Phillips et al., 2014).

Conclusions

The results from this experiment supported two of our three predictions. The δ¹⁵N TDF of 1.7‰ is indeed lower than the 3.4‰ assumed in most literature models. Conversely, the prediction that we made about the δ¹³C TDF being a small, positive and non-zero value was not met. However, our values still fell within the wide range of literature values for δ¹³C TDFs, and correcting for lipids brought values closer to zero. Finally, the ¹⁵N and ¹³C half-lives in our study are comparable to other invertebrate studies. The values calculated in our study may be used to better understand the feeding ecology of other tiger beetle species and their food web dynamics.