Research articlesRapid induction of the heat hardening response in an Arctic insect
Abstract
The ability to cope with increasing and more variable temperatures, due to predicted climate changes, through plastic and/or evolutionary responses will be crucial for the persistence of Arctic species. Here, we investigate plasticity of heat tolerance of the Greenlandic seed bug Nysius groenlandicus, which inhabits areas with widely fluctuating temperatures. We test the heat tolerance and hardening capacity (plasticity) of N. groenlandicus using both static (heat knock down time, HKDT) and dynamic (critical thermal maximum, CTmax) assays. We find that N. groenlandicus is able to tolerate short-term exposure to temperatures up to almost 50°C and that it can quickly increase heat resistance following heat hardening. Furthermore, we find that this hardening response is reversible within hours after hardening. These findings contrast with common observations from temperate and tropical insects and suggest high thermal plasticity in some Arctic insects which enables them to cope with extreme temperature variability in their habitats.
1. Introduction
Terrestrial arthropods in Arctic and subarctic regions are exposed to extreme and variable temperatures, compared to those of most other climatic zones [1–3], and climate change is predicted to be especially pronounced in these regions [4,5]. In such fluctuating environments, species can adapt to their thermal environments through evolutionary changes across generations and through adaptive plasticity within the lifetime of an organism [6,7]. Evolutionary adaptation to increasing temperatures is slow and can be evolutionarily constrained [8,9]. Therefore, plasticity may be the key to survival for many species facing variable thermal conditions and for Arctic and subarctic species in particular. Knowledge of plastic responses to heat stress in arthropods is typically based on studies of model organisms from temperate regions [10–13], whereas studies on Arctic and subarctic arthropods are few (but see [14,15]). Studying the ability of species to cope with extreme and variable temperatures in Arctic and subarctic regions through short-term plastic responses is key to understand present and future species distribution and abundance [10,16–19].
In the present study, we investigate heat tolerance and plasticity in the Greenlandic seed bug Nysius groenlandicus (Zetterstedt, 1838). We do this by looking at the rate of inducible heat tolerance through heat hardening and the reversibility of the hardening response. We use both static (heat knock down time, HKDT) and dynamic (critical thermal maximum, CTmax) methods for assessment of thermal tolerance, as these assays may capture different aspects of heat tolerance [20,21].
Nysius groenlandicus is widely distributed in Greenland and alpine areas of Scandinavia [2]. The species is univoltine, generally hatching in June, and then progressing through its five nymphal stages to reach the adult stage in July, before laying eggs throughout August. During summer months, temperatures in the Arctic and subarctic zones are extremely variable and studies have shown that temperatures in habitats of N. groenlandicus span the range of 0–50°C on a daily scale [3]. The species will thus be exposed to extreme upper temperatures, suggesting that thermal plasticity is important to its survival.
2. Material and methods
(a) Sample collection and maintenance
We collected adult N. groenlandicus in August 2017 in Narsarsuaq, South Greenland (61.180° N, 45.400° W). Daily microhabitat temperatures at the site of collection were highly variable (5–46°C) (electronic supplementary material, figure S1). Individuals used for the experiments were kept in the laboratory for 19 days before experiments were conducted. In the laboratory, the individuals were maintained in large Petri dishes (145 × 20 mm) in a climate room at 20 ± 0.5°C and a relative humidity of 52 ± 4% with a 12 : 12 light/dark cycle. Individuals had access to water (11.5 ml, 100 × 15.7 mm tubes, plugged with cotton), structure (open-cell polyether foam), and were fed raw sunflower seeds. Water, cotton balls and seeds were changed 2–3 times a week to avoid mould.
(b) Study design
To investigate aspects of the heat tolerance and hardening capacity of N. groenlandicus, we used both static (HKDT) and dynamic (CTmax) assays. For hardening capacity, we tested the importance of hardening duration and recovery time on subsequent heat tolerance.
(c) Assays and hardening treatments
Before testing, adults (individuals were not sexed in any of our experiments and a mix of males and females were used) were individually transferred to 4 ml screw-cap glass vials (45 × 14.7 mm). Screw caps contained a solidified droplet of a 2% agar solution to prevent desiccation during experiments [22]. All hardening treatments and tests for heat tolerance were carried out in glass vials that were submerged into temperature-controlled water baths (PolyScience MX Immersion Circulator: MX-CA12E).
CTmax was determined by exposing individuals to gradually increasing temperatures at a rate of 0.2°C min–1 from a starting temperature of 25°C as previously described [22]. Loss of movement was used as the endpoint value [20]. In the HKDT assay, individuals were exposed to a constant temperature of 48°C, where individuals had stopped movement within a short period of time, making the temperature suitable for an acute assay. Heat knock-down time was noted for each individual as the time taken until all visible movement had ceased.
The hardening treatment consisted of either 15, 30, 45 or 60 min exposure to 42°C. Pilot experiments showed that this temperature was non-lethal and led to induction of a heat hardening response in N. groenlandicus. Recovery time after hardening varied in different tests, but individuals were always kept at 20°C during recovery.
(d) Thermal exposure time
To investigate the effect of hardening time on heat tolerance, individuals were either exposed to a heat hardening treatment at 42°C for 15, 30, 45 or 60 min or control conditions (20°C). Immediately after hardening (no recovery period) HKDT of 15 individuals of each hardening treatment and control group was determined.
(e) Heat hardening and recovery time
To investigate the effect of recovery time on HKDT, individuals were exposed to a hardening treatment at 42°C for 1 h. This was followed by a recovery period of 0, 1 or 2 h, and 0, 2, 6 or 12 h in a second set of experiments, where the first experiment investigated the impact of short recovery periods on HKDT and the second longer recovery periods. From each recovery period, 15 hardened individuals and 15 control individuals (kept at 20°C) were tested for their HKDT. Similarly, to investigate the importance of recovery time on CTmax, individuals were exposed to a hardening treatment at 42°C for 1 h, followed by a recovery period of 0 (no recovery), 6, 12 or 24 h at 20°C. Subsequently, CTmax estimates of 10 individuals from each recovery period group and from the control group (kept at 20°C) were established. For individuals stored in glass vials longer than 6 h, the screwcaps were loosened slightly during the recovery period to avoid hypoxia.
(f) Hardening during temperature ramping
To investigate if individuals hardened during temperature ramping, i.e. when estimating CTmax, we exposed individuals to a hardening treatment at 42°C for 1 h (no recovery period) or control conditions (20°C). Subsequently, individuals were exposed to temperature ramping, with a start temperature of 25°C and a ramping rate of 0.2°C min−1. Fifteen hardened and 15 control individuals were subsequently sampled at three different temperatures during ramping: 26°C (5 min), 35°C (50 min), and 44°C (95 min), and HKDT at 48°C was determined immediately after sampling.
(g) Data analysis
Non-hardened controls were represented in each run for all experiments, which allowed us to correct for run effects. CTmax and HKDT values of controls varied between runs and therefore we corrected for run effects based on values of controls. This was done by using the formula:
Subsequently, data were tested for normality using the Shapiro–Wilk normality test and for homogeneity of variances using Levene's test. For analysis of differences between treatment groups, one-way or two-way analyses of variance (ANOVA) were conducted, while comparisons of individual samples were done using Student's t-tests. IBM SPSS Statistics was used for statistical analyses (v. 25).
3. Results
(a) Heat hardening and recovery time
Mean HKDT of N. groenlandicus was higher for individuals exposed to a heat hardening treatment and following different short post-hardening recovery durations (0 or 1 h) compared to control individuals (figure 1a). There was a significant positive effect of heat hardening on HKDT (F2,84 = 20.3, p < 0.01) and the 0- and 1-h recovery time groups had significantly higher HKDT compared to the 2-h recovery time group (PTukey's HSD< 0.05).
Figure 1. (a) Effect of post-hardening recovery duration (0, 1 and 2 h) on mean heat knock-down time (HKDT) at 48°C. Hardened individuals were exposed to 42°C for 1 h followed by recovery prior to determination of HKDT. Black bars represent corrected means of hardened samples (± s.e.) and grey bars represent corrected means of control samples (± s.e.). N = 13–15 per sample. (b) Effect of hardening duration (15, 30, 45 and 60 min) at 42°C on HKDT at 48°C. Hardened individuals were assessed for HKDT immediately after the hardening treatment. Black bars represent corrected means of hardened samples (± s.e.) and grey bars represent corrected means of control samples (± s.e.). N = 13–15 per sample. Asterisks indicate significant differences between hardened individuals and corresponding control; *p < 0.05, **p < 0.01, ***p < 0.001.
In a second set of experiments, individuals were allowed to recover for a longer period of time (0, 2, 6 and 12 h) (electronic supplementary material, figure S2). There was again a significant positive effect of heat hardening on HKDT after 0 h of recovery (t28 = 2.1, p < 0.05), but not after 2, 6 and 12 h of recovery (p > 0.05 in all cases).
Mean CTmax of N. groenlandicus for hardened individuals was 49.4°C and slightly lower for control individuals (electronic supplementary material, figure S3). In accordance with this, there was no significant effect of recovery time (0, 6, 12 and 24 h) on CTmax (F3,63 = 0.043, p > 0.05), nor any difference in CTmax between hardened and control individuals for any of the recovery periods (p > 0.05 in all cases).
(b) Thermal exposure time
Individuals exposed to hardening durations of 45 or 60 min had higher mean HKDT than individuals exposed to hardening durations of 15 or 30 min (figure 1b). There was a significant effect of heat hardening duration on HKDT (F7,111 = 9.73, p < 0.01). HKDT of individuals exposed to a 45-min hardening treatment was significantly higher compared to individuals exposed for 15 or 30 min, and individuals exposed for 60 min had significantly higher HKDT compared to individuals exposed for 30 min (PTukey'sHSD < 0.05). HKDT of individuals hardened for 15 and 30 min were not significantly different from the corresponding controls (p > 0.05), whereas individuals hardened for 45 and 60 min had significantly higher HKDT compared to the controls (t28 = 5.6, p < 0.01; t28 = 4.36, p < 0.01).
(c) Hardening during temperature ramping
Mean HKDT of pre-hardened samples were higher compared to HKDT of control individuals, but this difference diminished as exposure time and temperature increased towards the end-ramping temperature of 44°C (figure 2). A two-way ANOVA showed a significant effect of hardening treatment on HKDT (F1,82 = 4.25, p < 0.05), but no significant effect of end-ramping temperature (F2,82 = 0.406, p > 0.05) or temperature × treatment interaction (F2,82 = 0.78, p > 0.05).
Figure 2. Effect of heat ramping end-point temperature (26, 35 and 44°C) on mean heat knock-down time (HKDT) at 48°C. Prior to ramping, hardened individuals were exposed to 42°C for 1 h, whereas control individuals were kept at 20°C. Thereafter, hardened and control individuals were exposed to heat ramping, and individuals were sampled when temperatures reached 26, 35 and 44°C. Thereafter, HKDT was assessed at 48°C. Black bars represent means of hardened samples (± s.e.) and grey bars represent means of control samples (± s.e.). N = 13–15 per sample. No significant differences were observed between hardened individuals and corresponding control individuals.
4. Discussion
The seed bug N. groenlandicus is an Arctic and subarctic specialist herbivore exposed to highly variable and occasionally very high temperatures (electronic supplementary material, figure S1) [3]. Our results showed a surprisingly fast and reversible hardening response in N. groenlandicus. Few studies have addressed heat tolerance and dynamics of heat hardening in Arctic insects, and our results are thus important for understanding how well they will cope with future warmer and more variable temperatures.
Studies on plastic responses to heat stress in arthropods are typically based on studies on model organisms from temperate regions [10–13,27,28], whereas studies on Arctic and subarctic arthropods are few [14,15]. We found a strong effect of short-term heat hardening (45 min) on HKDT (figure 1b). This suggests a very quick induction of the heat stress responses in N. groenlandicus at temperatures close to what they experience in the field [3] (electronic supplementary material, figure S1). Furthermore, the results show that the effect of heat hardening vanishes within two hours of recovery (figure 1a), indicating quick reversibility of the heat stress response in N. groenlandicus. These patterns are distinct from observations in other species. For example, studies on Drosophila melanogaster have shown that the maximum gain of heat hardening is observed more than 8 h post-hardening and that heat hardening benefits are observable until 32 h post-hardening [23]. For the soil-dwelling arthropod, Orchesella cincta, adapted to highly stable environments, maximum gain of heat hardening is observed greater than 27 h post-hardening [11]. Thus, the optimal hardening temperature and duration is very dependent on the species, the temperatures they experience and their ability to behaviourally avoid thermal extremes [29], and suggest that extrapolation of results across species and ecosystems can be problematic. Typically, the methods used for assessing heat tolerance are either dynamic or static [10,30,31]. The use of these assays has been widely discussed, including their ecological relevance, and starvation and low humidity levels during long experiments have been suggested to accumulate deleterious effects during ramping assays [32–34]. Dynamic critical thermal limit assessment methods are generally considered more ecologically relevant, as slow ramping to potentially stressful temperatures is a more accurate representation of what species in the wild typically experience [35,36]. However, the fast hardening response found in the present study suggests that heat hardening for some species will take place during temperature ramping. We observed a clear trend that differences in HKDT between hardened and control individuals diminished during temperature ramping. This could also explain why we do not detect differences in CTmax between control individuals and hardened individuals. Thus, our results suggest that hardening will take place more so during dynamic thermal assays compared to static assays, such as HKDT.
In conclusion, our findings suggest a fast and dynamic heat hardening response in N. groenlandicus. We suggest this is an adaptation to the extreme and highly fluctuating Arctic and subarctic temperature conditions, and that N. groenlandicus may be able to cope with future warmer and more variable summer temperatures through adaptive plasticity.
Data accessibility
Data are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.tm06n0p [37].
Authors' contributions
M.H.S., T.N.K. and S.B. designed the experiments. M.H.S., T.N.K., J.M.S.L., N.K.N. and S.B. performed the experiments. M.H.S., T.N.K., T.T.H. and S.B. analysed the data and interpreted the results. M.H.S., T.N.K. and S.B. wrote the first version of the manuscript. All authors contributed to revision of the manuscript and approved the final version. All authors agree to be held accountable for the content of this manuscript.
Competing interests
We declare we have no competing interests.
Funding
We thank the Danish Council for Independent Research (grant no. DFF-8021-00014B), the Carlsberg Foundation (CF17-0415), North2North and the Aalborg Zoo Conservation Foundation (AZCF) for supporting this work.
Footnotes
References
- 1. IPCC. 2014Climate Change 2014 Synthesis Report. IN Contribution of working groups I, II and III to the fifth assessment report of the intergovernmental panel on climate change (eds Core Writing Team,
Pachauri RK, Meyer LA ). Geneva, Switzerland: IPCC. Google Scholar - 2.
Böcher J, Kristensen NP, Pape T, Vilhelmsen L . (eds.) 2015The Greenland entomofauna. Leiden, The Netherlands: Brill. (doi:10.1163/9789004261051) Crossref, Google Scholar - 3.
Böcher J . 1972Feeding biology of Nysius groenlandicus in Greenland, with a note on oviposition in relation to food source and dispersal of the species. Meddelelser om Grønl. 191, 3-41. Google Scholar - 4.
Lemke P, Jacobi H-W . 2013Arctic climate change: the ACSYS decade and beyond. Atmos. Oceanogr. Sci. Libr. 43, 49-6895. (doi:10.5860/choice.49-6895) Google Scholar - 5.
Bindoff NL et al. 2013Detection and attribution of climate change: from global to regional. In Climate change 2013: the physical science basis. Cambridge, UK: Cambridge University Press. Google Scholar - 6.
Walther GR, Post E, Convey P, Menzel A, Parmesan C, Beebee TJC, Fromentin JM, Hoegh-Guldberg O, Bairlein F . 2002Ecological responses to recent climate change. Nature 416, 389-395. (doi:10.1038/416389a) Crossref, PubMed, ISI, Google Scholar - 7.
Aitken SN, Yeaman S, Holliday JA, Wang T, Curtis-McLane S . 2008Adaptation, migration or extirpation: climate change outcomes for tree populations. Evol. Appl. 1, 95-111. (doi:10.1111/j.1752-4571.2007.00013.x) Crossref, PubMed, ISI, Google Scholar - 8.
Araújo MB, Ferri-Yáñez F, Bozinovic F, Marquet PA, Valladares F, Chown SL . 2013Heat freezes niche evolution. Ecol. Lett. 16, 1206-1219. (doi:10.1111/ele.12155) Crossref, PubMed, ISI, Google Scholar - 9.
Hoffmann AA, Chown SL, Clusella-Trullas S . 2013Upper thermal limits in terrestrial ectotherms: how constrained are they?Funct. Ecol. 27, 934-949. (doi:10.1111/j.1365-2435.2012.02036.x) Crossref, ISI, Google Scholar - 10.
Hoffmann AA, Sørensen JG, Loeschcke V . 2003Adaptation of Drosophila to temperature extremes: bringing together quantitative and molecular approaches. J. Therm. Biol. 28, 175-216. (doi:10.1016/S0306-4565(02)00057-8) Crossref, ISI, Google Scholar - 11.
Bahrndorff S, Mariën J, Loeschcke V, Ellers J . 2009Dynamics of heat-induced thermal stress resistance and hsp70 expression in the springtail, Orchesella cincta. Funct. Ecol. 23, 233-239. (doi:10.1111/j.1365-2435.2009.01541.x) Crossref, ISI, Google Scholar - 12.
van Dooremalen C, Berg MP, Ellers J . 2013Acclimation responses to temperature vary with vertical stratification: implications for vulnerability of soil-dwelling species to extreme temperature events. Glob. Chang. Biol. 19, 975-984. (doi:10.1111/gcb.12081) Crossref, PubMed, ISI, Google Scholar - 13.
Allen JL, Clusella-Trullas S, Chown SL . 2012The effects of acclimation and rates of temperature change on critical thermal limits in Tenebrio molitor (Tenebrionidae) and Cyrtobagous salviniae (Curculionidae). J. Insect Physiol. 58, 669-678. (doi:10.1016/j.jinsphys.2012.01.016) Crossref, PubMed, ISI, Google Scholar - 14.
Martinet B, Lecocq T, Smet J, Rasmont P . 2015A protocol to assess insect resistance to heat waves, applied to bumblebees (Bombus Latreille, 1802). PLoS ONE 10, e0118591. (doi:10.1371/journal.pone.0118591) Crossref, PubMed, ISI, Google Scholar - 15.
Coulson S, Block W . 1996Can Arctic soil microarthropods survive elevated summer temperatures?Funct. Ecol. 10, 314-321. (doi:10.2307/2390278) Crossref, ISI, Google Scholar - 16.
Sunday JM, Bates AE, Dulvy NK . 2012Thermal tolerance and the global redistribution of animals. Nat. Clim. Change 2, 686-690. (doi:10.1038/nclimate1539) Crossref, ISI, Google Scholar - 17.
Addo-bediako A, Chown SL, Gaston KJ . 2000Thermal tolerance, climatic variability and latitude. Proc. R. Soc. Lond. B 267, 739-745. (doi:10.1098/rspb.2000.1065) Link, ISI, Google Scholar - 18.
Buckley LB, Kingsolver JG . 2012The demographic impacts of shifts in climate means and extremes on alpine butterflies. Funct. Ecol. 26, 969-977. (doi:10.1111/j.1365-2435.2012.01969.x) Crossref, ISI, Google Scholar - 19.
Deutsch CA, Tewksbury JJ, Huey RB, Sheldon KS, Ghalambor CK, Martin PR, Haak DC . 2008Impacts of climate warming on terrestrial ectotherms across latitude. Proc. Natl Acad. USA Sci. 105, 6668-6672. (doi:10.1073/pnas.0709472105) Crossref, PubMed, ISI, Google Scholar - 20.
Terblanche JS, Hoffmann AA, Mitchell KA, Rako L, le Roux PC, Chown SL . 2011Ecologically relevant measures of tolerance to potentially lethal temperatures. J. Exp. Biol. 214, 3713-3725. (doi:10.1242/jeb.061283) Crossref, PubMed, ISI, Google Scholar - 21.
Castañeda LE, Rezende EL, Santos M . 2015Heat tolerance in Drosophila subobscura along a latitudinal gradient: contrasting patterns between plastic and genetic responses. J. Evol. Biol. 69, 2721-2734. (doi:10.1111/evo.12757) Google Scholar - 22.
Overgaard J, Kristensen TN, Sørensen JG . 2012Validity of thermal ramping assays used to assess thermal tolerance in arthropods. PLoS ONE 7, e32758. (doi:10.1371/journal.pone.0032758) Crossref, PubMed, ISI, Google Scholar - 23.
Dahlgaard J, Loeschcke V, Michalak P, Justesen J . 1998Induced thermotolerance and associated expression of the heat-shock protein Hsp70 in adult Drosophila melanogaster. Funct. Ecol. 12, 786-793. (doi:10.1046/j.1365-2435.1998.00246.x) Crossref, ISI, Google Scholar - 24.
Calabria G 2012Hsp70 protein levels and thermotolerance in Drosophila subobscura: a reassessment of the thermal co-adaptation hypothesis. J. Evol. Biol. 25, 691-700. (doi:10.1111/j.1420-9101.2012.02463.x) Crossref, PubMed, ISI, Google Scholar - 25.
Sørensen JG, Dahlgaard J, Loeschcke V . 2001Genetic variation in thermal tolerance among natural populations of Drosophila buzzatii: down regulation of Hsp70 expression and variation in heat stress resistance traits. Funct. Ecol. 15, 289-296. (doi:10.1046/j.1365-2435.2001.00525.x) Crossref, ISI, Google Scholar - 26.
Nygaard V, Rødland EA . 2016Methods that remove batch effects while retaining group differences may lead to exaggerated confidence. Biostatistics 17, 29-39. (doi:10.1093/biostatistics/kxv027) Crossref, PubMed, ISI, Google Scholar - 27.
Alemu T, Alemneh T, Pertoldi C, Ambelu A, Bahrndorff S . 2017Costs and benefits of heat and cold hardening in a soil arthropod. Biol. J. Linn. Soc. 122, 765-773. (doi:10.1093/biolinnean/blx092) Crossref, ISI, Google Scholar - 28.
Jensen A, Alemu T, Alemneh T, Pertoldi C, Bahrndorff S . 2019Thermal acclimation and adaptation across populations in a broadly distributed soil arthropod. Funct. Ecol. 33, 833-845. (doi:10.1111/1365-2435.13291) Crossref, ISI, Google Scholar - 29.
Huey RB, Hertz PE, Sinervo B . 2003Behavioral drive versus behavioral inertia in evolution: a null model approach. Am. Nat. 161, 357-366. (doi:10.1086/346135) Crossref, PubMed, ISI, Google Scholar - 30.
Lutterschmidt WI, Hutchison VH . 1997The critical thermal maximum: history and critique. Can. J. Zool. 75, 1561-1574. (doi:10.1139/z97-783) Crossref, ISI, Google Scholar - 31.
Jørgensen LB, Malte H, Overgaard J . 2019How to assess Drosophila heat tolerance: unifying static and dynamic tolerance assays to predict heat distribution limits. Funct. Ecol. 33, 629-642. (doi:10.1111/1365-2435.13279) Crossref, ISI, Google Scholar - 32.
Beitinger TL, Bennett WA, Mccauley RW . 2000Temperature tolerances of North American freshwater fishes exposed to dynamic changes in temperature. Environ. Biol. Fishes 58, 237-275. (doi:10.1023/A:1007676325825) Crossref, ISI, Google Scholar - 33.
Rezende EL, Tejedo M, Santos M . 2011Estimating the adaptive potential of critical thermal limits: methodological problems and evolutionary implications. Funct. Ecol. 25, 111-121. (doi:10.1111/j.1365-2435.2010.01778.x) Crossref, ISI, Google Scholar - 34.
Santos M, Castañeda LE, Rezende EL . 2011Making sense of heat tolerance estimates in ectotherms: lessons from Drosophila. Funct. Ecol. 25, 1169-1180. (doi:10.1111/j.1365-2435.2011.01908.x) Crossref, ISI, Google Scholar - 35.
Somero GN . 2005Linking biogeography to physiology: evolutionary and acclimatory adjustments of thermal limits. Front. Zool. 2, 1. (doi:10.1186/1742-9994-2-1) Crossref, PubMed, Google Scholar - 36.
Terblanche JS, Deere JA, Clusella-Trullas S, Janion C, Chown SL . 2007Critical thermal limits depend on methodological context. Proc. R. Soc. B 274, 2935-2942. (doi:10.1098/rspb.2007.0985) Link, ISI, Google Scholar - 37.
Sørensen MH, Kristensen TN, Lauritzen JMS, Noer NK, Høye TT, Bahrndorff S . 2019Data from: Rapid induction of the heat hardening response in an Arctic insect . Dryad Digital Repository(https://doi.org/10.5061/dryad.tm06n0p) Google Scholar
