Critical thermal maxima of three small-bodied fish species (Cypriniformes) of different origin and protection status

Abstract Environmental changes related to global warming are both a threat to cold-water fishes and simultaneously create favourable conditions for the spread of eurythermic or warm-water species. In both cases, it is necessary to develop conservation strategies supported by precise ecological data, including thermal requirements. In this study, we determined the upper thermal tolerance thresholds and the critical maxima for three small, non-commercial Eurasian freshwater fish species; i.e. lake minnow Eupallasella (=Rhynchocypris) percnurus, sunbleak Leucaspius delineatus and topmouth gudgeon Pseudorasbora parva using the Critical Thermal Method at different acclimation temperatures, i.e. 18, 23, 28, and 33°C (highest treatment for topmouth gudgeon only). We hypothesized that lake minnow would have the lowest thermal tolerance and the topmouth gudgeon the highest. The response to temperature rise and the upper thermal limits were species-specific and correlated with the acclimation temperature, but not with fish length. Sunbleak showed the lowest thermal tolerance, though at 28°C both E. percnurus and L. delineatus reached a similar critical thermal limit. Topmouth gudgeon showed distinctly higher upper thresholds of thermal tolerance and at the highest acclimation temperature the critical upper limit for the species was close to 42°C. The results obtained for L. delineatus were surprising as we predicted that this leuciscid fish would be more tolerant of high temperatures than E. percnurus. We discuss the results in relation to the threat of extinction and the risk of species spreading beyond their natural range in the context of a warming environment.


Introduction
Among abiotic factors, temperature has a major impact on the life functions of aquatic organisms, especially ectothermic species. In fishes, development, growth rate, the onset and duration of the spawning period, as well as survival rates are strictly temperature-dependent (Fry 1971;Elliott 1981;Wieser 1991). Individual fish species have different thermal requirements and acknowledging these differences allows implementing both economic and protective measures to improve their living environment, breeding, or cultivation. This is particularly important in the face of increasing climatic instability, which potentially affects the thermal conditions of aquatic environments (Harrod 2016;Dokulil et al. 2021). Climate changes observed in recent years include, inter alia, deterioration of available habitats in the case of freshwater fishes, mostly as a result of disturbances to flow dynamics or increase in water temperatures. Consequently, the assemblage structure of freshwater fishes may undergo transformations, for example by a gradual disappearance of native, cold-water species and the expansion of alien, eurytopic ones Jeppesen et al. 2010;Comte et al. 2013;Rolls et al. 2017). In Central European water gibel carp Carassius gibelio, Chinese sleeper Perccottus glenii and topmouth gudgeon Pseudorasbora parva, is clearly visible. Simultaneously, disappearance of native species such as lake minnow Eupallasella (=Rhynchocypris) percnurus and sunbleak Leucaspius delineatus is observed (Gozlan et al. 2005;Witkowski & Grabowska 2012;Wolnicki et al. 2022), which can partly be explained by changes in the environment related to the warming process.
Climate change has promoted a research interest in the thermal requirements of fishes. However, small-bodied fish species of limited commercial value, such as E. percnurus, L. delineatus or P. parva, do not attract as much research attention as species with greater economic significance, such as salmonids. Lake minnow is currently protected by national and European law as the priority species within the European Ecological Natura 2000 network, while natural populations of sunbleak are protected by the Bern Convention (Freyhof & Brooks 2011;Wolnicki et al. 2022). In contrast, the topmouth gudgeon is considered a serious biological threat to many species in European waters (Gozlan 2012;Baltazar-Soares et al. 2020). However, it could be concluded that all three species have an underestimated biological and economical potential, either for their conservation or as a biological threat when invasive (Zięba et al. 2010;Interesova 2012;Reshetnikov et al. 2017). Data on the thermal requirements of these species, particularly critical lethal or sublethal temperature thresholds, are sparse or imprecise. The exceptions are studies related to the optimal temperature for growth and egg development of lake minnow (Wolnicki et al. 2004;Kamiński et al. 2006). Nevertheless, the knowledge gap of these species' thermal requirements limits the ability to develop successful conservation strategies, either in terms of protection against extinction or invasion.
Of the various methodologies for determining critical temperatures for fish, the critical thermal method (CTM) has the advantage of the possibility of establishing the temperature limits (CTmax or CTmin) without reaching a lethal endpoint, in contrast to an incipient lethal temperature (ILT) or chronic lethal temperature (CLT) (Fry 1971;Beitinger et al. 2000). Moreover, CTM is a rapid and convenient method which does not require any complicated apparatus or a large sample size (Currie et al. 1998;Rajaguru 2002;Mora & Maya 2006). It is also recommended for small-bodied fish species (e.g. Beitinger et al. 2000;Carveth et al. 2006;Mora & Maya 2006). The CTM, classed as a dynamic method, consists of raising (or lowering) the ambient temperature for fish at a constant, strictly defined rate. The critical limit of thermal tolerance is defined as a pre-death thermal point at which locomotory functions become disorganized, preventing fish from escaping conditions that result in death (Fry 1971;Becker & Genoway 1979;Beitinger et al. 2000). In this case, a precise definition of the sublethal endpoint is required, which is most often expressed as loss of equilibrium in fish (LOE). After the fish have reached the sublethal phase (heat coma) and have been transferred back to an acclimation temperature, the individuals undergo recovery of vital functions (Currie et al. 1998, Beitinger et al. 2000Noyola et al. 2015). After undergoing this process, tested fish can be released back into their natural environment. Therefore, the method is recommended, especially for endangered and protected species (Beitinger et al. 2000;Carveth et al. 2006;Dalvi et al. 2009). The CTM allows the establishment of an upper tolerance threshold, similarly to lethal methods, but without killing test fish specimens (Fry 1971;Beitinger et al. 2000).
In this study, we aimed at identifying upper critical temperatures of three freshwater fishes, lake minnow (E. percnurus), sunbleak (L. delineatus), and topmouth gudgeon (P. parva), to assess their capacity to withstand thermal stress in their natural environment and predict their capability to populate new habitats. We hypothesized that E. percnurus would exhibit the lowest thermal tolerance, L. delineatus will be at an intermediate level, and P. parva will show the highest thermal tolerance. We discuss our results in the context of the potential outcomes for the studied species in a warming environment.

Fish collection and acclimation
The tested fish species were sampled between June and September 2019 in various types of water bodies located in northern Poland. Lake minnow specimens were collected in a natural isolated pond situated near the village of Egiertowo in the Kashubian Lakeland (54.228614 N, 18.196273 E). Sunbleaks were obtained from a small channel in the River Pasłęka system (Vistula Lagoon basin;53.907250 N, 20.195778 E), while topmouth gudgeon was caught in the Struga Rychnowska stream, a tributary of the River Drwęca (lower River Vistula system; 53.095285 N, 18.826229 E). After capture, fish were transported to the laboratory and placed in aquaria connected with a recirculation aquaculture system (RAS). This system consisted of three independent units, each of 900 l capacity, with the ability to regulate water temperature. Total length of each fish specimen was measured to the nearest 0.1 cm and only adult specimens (TL > 4.0 cm) were used for further procedure. All individuals of fish species were randomly divided into separate groups for thermal acclimation in 40 l aquaria. For all three species, individuals were acclimated gradually (ca 1°C day −1 ) to reach the target temperatures of 18.0°C, 23.0°C, and 28.0°C (± 0.2°C). In addition, topmouth gudgeon was acclimated at 33.0°C. The highest acclimation temperatures were chosen on the basis of our own experience and the published literature, with the aim that the normal physiological functions of the fish would not be compromised. Previous attempts to acclimate fish at temperatures higher than those selected failed due to weight loss of the fish. A photoperiod of 16 h: 8 h light/dark was imposed. These conditions were maintained for two weeks prior to CTM testing. Fishes were fed with frozen chironomid larvae and commercial aquarium dry food, and were fasted for 24 hours before testing.

CTM testing procedure
After acclimation, CTmax tests were conducted in 10 l plastic heating chambers equipped with a 300 W submersion heater (Eliko, Poland). During experiments, water was lightly aerated through two finely perforated air stones, which permitted equalization of the temperature in the chamber, but did not force the fish to expend additional energy in swimming against water movement. Prior to each trial several fish (no more than five) were randomly selected from a particular acclimation temperature and placed in a small basket submerged in the heating chamber. Test fish were subsequently allowed to settle for 10-15 minutes. Water temperature was recorded using a calibrated mercury thermometer with an accuracy of 0.1°C, and oxygenation was measured with an oximeter (OxyGuard Handy Polaris ver.1.26) placed inside the basket. During the experiments the oxygen saturation of the water was maintained at 80-105%, to prevent the oxygen concentration from falling below 5 mg l −1 .
For each acclimation temperature, experiments were carried out in three trials, except in the case of lake minnow acclimated at 23°C in which two trials were performed. Each fish was used only once for tests. The experiments were started at a given acclimation temperature, and the temperature rise was maintained at the recommended rate for small-bodied fish species of 0.3°C min −1 (Becker & Genoway 1979;Currie et al. 1998;Beitinger et al. 2000;Carveth et al. 2006). During tests, fish behaviour was carefully observed and any disturbance at a specific temperature was registered independently by two persons. The first sign of a loss of balance (failure to maintain dorsoventral orientation) of a fish individual was considered as an initial loss of equilibrium (ILOE). However, for most individuals, this phase was only temporary and they quickly recovered normal buoyancy during temperature increase. Thus, the upper thermal limit (CTmax) was considered as a final loss of equilibrium (Becker & Genoway 1979;Carveth et al. 2006;Eme & Bennett 2009), and the moment a fish lost the ability to maintain dorso-ventral orientation for 15 second was defined as the critical endpoint. This criterion was established to represent the point at which the individual lost the capacity to escape unfavourable conditions while also avoiding the risk of reaching a lethal endpoint (Beitinger et al. 2000). This allowed us to release experimental fish back to their natural habitat after surveys, which is crucial in the case of lake minnow since this is an endangered and protected species. When an individual reached the CTmax endpoint, the temperature was precisely recorded and the individual was removed from the basket, measured (total length) and returned to an aquarium at its original acclimation temperature. Each trial was continued until the final loss of equilibrium was recorded in all tested individuals. ILOE and CTmax for all species were calculated as an arithmetic mean of the collective thermal points at which the given endpoint was reached. After the termination of the tests, all specimens were monitored until recovery. Lake minnow individuals were then released into the same water body in which they had been caught.

Statistical analysis
CTmax values were regressed on acclimation temperature and analysis of covariance (ANCOVA) was used to establish the differences in coefficients of regressions among species. The significance of observed variation was analysed with the LSD Fisher post-hoc test. To test the effects of body size on CTmax, total length (TL, in cm) was included in the Gaussian Generalized Model (GLM) as a covariate with fish species (i) and acclimation temperature Tacc. (j). The model was specified as: The model was fitted using R version 4.1.2 (R Development Core Team 2021). CTmax -ILOE differences were analysed using nested analysis of variance (ANOVA), where the temperature of acclimation was nested in species (Zar 2010). This difference is a function of time and determines the rate at which the sublethal response of the fish is achieved with a constant temperature increase. To meet assumptions of analyses, diagnostic plots of residuals against fitted values and QQ plot of residuals were used to assess the normality of residuals. Homoscedasticity was tested using the Levene test. If the results of ANOVA showed a significant effect, the LSD Fisher post-hoc test was used to identify which factors differed significantly. Probability values (p) < 0.05 were considered significant. Data are reported as mean ± standard deviation (SD). Statistical analyses were conducted using STATISTICA 13 (Dell Inc 2016).

Results
During the tests lake minnow and sunbleak frequently appeared stressed by the temperature increase, in contrast to the topmouth gudgeon which failed to express a behavioural response. The lowest ILOE value was observed for sunbleak acclimated at 18°C and occurred at temperatures slightly above 30.0°C, which was about two degrees below ILOE for lake minnow (Table I). For topmouth gudgeon, ILOE values were highest, and the first symptoms of loss of balance were observed at a temperature exceeding 33.0°C. Acclimation temperatures of 18.0°C and 23.0°C in case of sunbleak resulted in the lowest upper thermal tolerance threshold (CTmax) among all tested species, whereas both the sunbleak and lake minnow acclimated at 28.0°C exhibited a similar upper critical temperature, which was in the range of 36.6-36.9°C (Table I).
Among the three species, topmouth gudgeon showed the highest thermal tolerance. Fish acclimated to 28.0°C reached CTmax slightly above 40.0°C, i.e. over 3.0°C higher temperature than the other two species. In the case of the highest tested acclimation temperature (33.0°C), topmouth gudgeon had a threshold of critical temperature close to 42.0°C (Table I).
CTmax of all species was positively correlated to acclimation temperature, and analysis of covariance revealed differences in slopes among species (F 2,111 = 39.06, p < 0.001). Multiple comparisons (LSD Fisher post-hoc test) showed that all species differed significantly from one another (Table II). The highest increase of CTmax in relation to acclimation temperature (b value in Table II as an equivalent of acclimation response ratio, ARR) was recorded for sunbleak, and the lowest for lake minnow. Topmouth gudgeon increased its CTmax by 0.46 ± 0.01°C (mean ± SD) with each Celcius degree increase in acclimation temperature, whereas sunbleak and lake minnow increased their CTmax by 0.52 ± 0.03 and 0.28 ± 0.02, respectively. For topmouth gudgeon and sunbleak, the relationships were almost parallel (Figure 1). The results are confirmed by the GLM, showing the effect of acclimation temperature on fish species, but no relation between fish size (TL) and CTmax was noted (Table III). The overall model fit was highly significant (R 2 = 0.952, df = 4.112, p < 0.001, residual standard error: 0.657) and the value of Akaike Information Criterion (AIC) was 240.5485.   Differences between CTmax and ILOE were significantly species-dependent (two-way nested ANOVA: F 2,107 = 7.245, p < 0.001), as well as on the acclimation temperature nested in species (F 7,107 = 39.404, p < 0.0001). In general, the differences between CTmax and ILOE decreased with temperature of acclimation, and therefore, for all species the largest differences were noted at 18.0°C (Figure 2). The LSD Fisher post-hoc test revealed that at the lowest acclimation temperature top-mouth gudgeon had the highest value and there was no difference between sunbleak and lake minnow. In the case of the two latter species, the CTmax-ILOE difference was insignificant at acclimation temperatures of 23 and 28°C, though the lowest value was recorded for sunbleak at 28°C (Figure 2). Surprisingly, lake minnow had a higher mean value of CTmax -ILOE differences at 28°C than at 23°C. No significant differences between mean values of the CTmax -ILOE difference were found for topmouth gudgeon at acclimation temperatures of 23°C, 28°C, and 33°C (Figure 2).

Discussion
The upper thermal tolerance thresholds in fishes differ at the levels of family, genus and species (Elliott 1981;Beitinger et al. 2000;Hasnain et al. 2013). Physiological plasticity and intra-specific variation in thermal tolerance is well established in the cyprinid group, often considered as a thermophilic or warm-water fish species group (Elliott 1981;Wieser 1991;Beitinger et al. 2000), and represented in our study by a leuciscid, L. delineatus, and a gobionid, P. parva. Many species in the genus Rhynchocypris (family: Leuciscidae), represented here by E. percnurus, are considered as cold-water species, which is related to their postglacial distribution and general ecological performance (Yu et al. 2014;Park et al. 2019). Our results revealed an unexpected response in CTmax values, which differs between lake minnow and sunbleak depending on acclimation conditions. However, the thermal tolerance of both species coincided at the highest acclimation temperature. It can be assumed that acclimation to warmer water is suboptimal in lake minnow as opposed to sunbleak due to a different physiological efficiency associated with long-term thermal stress. Physiological processes regulating acclimation to warmer temperature involve the metabolic rate, cardiovascular capacity, and oxygen consumption, and are typically species-specific (Fry 1971;Wieser 1991;Illing et al. 2020;McKenzie et al. 2021). Nevertheless, the sunbleak had a relatively low thermal tolerance of about 30°C in the case of the lowest acclimation temperatures. In contrast, the CTmax-ILOE difference values varied at the highest acclimation temperature and were higher for E. percnurus. For juvenile lake minnow, an optimum growth temperature (OGT) was determined as 22.0-25.0°C in laboratory conditions,  Figure 2. Effect of acclimation temperature on the CTmax -ILOE difference values (±95% CL) for the studied fish species. The same letters next to marks denote groups that did not statistically differ (the LSD Fisher post hoc test).

1356
G. Radtke et al. though at a constantly maintained 28.0°C the fish grew slower (Wolnicki et al. 2004;Kamiński et al. 2011). In the wild, lake minnow suspended feeding activity over 29.0°C in contrast to its most common cohabitant of gibel or crucian carp Carassius carassius (Radtke 2011). Field surveys revealed that E. percnurus occupies habitats in which the maximum temperature does not exceed 30.0°C ). There are no available data regarding the thermal optimum for sunbleak (e.g. OGT), but we can assume that our results correspond with the maximum temperatures occurring currently in small water bodies in central Europe Radtke 2011). The occurrence of sunbleak in Europe, like that of most cyprinid fishes, is a consequence of a postglacial expansion from the Ponto-Caspian refugee region. The region extends from the Volga basin in the East to the Rhine basin in the West, and to the River Rioni drainage basin in the South, as far as 42° latitude (Kottelat & Freyhof 2007). The sunbleak is relatively ubiquitous and more widespread than E. percnurus in Europe. Its natural habitats are stagnant isolated water bodies, or those connected to rivers during floods. In recent years, the species is known to be declining throughout its occurrence range (Gozlan et al. 2005;Andreou et al. 2011), but often spreads to waters connected with carp farms. Simultaneously, L. delineatus is considered as an invasive species in parts of Europe, such as Great Britain or the Asian part of Russia, where it was accidentally introduced with game fish (Zięba et al. 2010;Interesova 2012;Reshetnikov et al. 2017).
In contrast to the sunbleak, our predictions were supported in the case of P. parva. Topmouth gudgeon proved to be highly resistant to elevated temperatures ( Figure 1, Table II), and therefore can be classified as a warm-water or eurythermal species. Its native range is East Asia, as far to the south as the Zhujiang River in China (Kottelat & Freyhof 2007;Baltazar-Soares et al. 2020). The species has high reproductive, and exhibits feeding and growth plasticity (Onikura & Nakajima 2013;Záhorská et al. 2014;Davies & Britton 2015;Rolla et al. 2020). In Europe, the first specimens of P. parva were introduced to the Danube basin (Romania) in 1961. According to Gozlan (2012), they came from the Cháng Jiang River (River Yangzi system) at Wùhàn, a large town on the east coast of China. The species quickly spread to carp farms and, as a consequence, also to open waters. In Europe, topmouth gudgeon is treated as a highly invasive species, which can affect the structure of natural fish communities (Britton et al. 2007;Charrier et al. 2016). The species is considered as an important parasite vector, and consequently poses a serious threat to native fish faunas at a global scale (Gozlan 2012;Spikmans et al. 2020).
Lake minnow occupies water bodies across a large area of the northern hemisphere (Kusznierz et al. 2017). According to the latest reports, in the European part of its distribution the most numerous populations currently exist in the Kashubian Lake District, in northern Poland (Wolnicki et al. 2022). In other regions of the country, the species has already become extinct or its populations are vestigial. In adjacent countries, e.g. Belarus, Ukraine, and Lithuania, where lake minnow sites were identified in the past, the species most likely survived exclusively at a few sites in the Lithuania (Tsaryk et al. 2002;Kaupinis 2006;Wolnicki et al. 2020). In the context of habitat loss, another symptom of contemporary change that is gaining importance in freshwaters is the appearance of alien fish species (Bernery et al. 2022). In Poland, lake minnow populations are often accompanied by alien invasive gibel carp or its hybrids with native crucian carp, less often with sunbleak Wolnicki et al. 2022). Thus, the emergence of a new invasive species might exert a decisive impact on the survival of E. percnurus or L. delineatus in a changing climate. Recent studies have demonstrated the impact of invasive topmouth gudgeon on native lake minnow populations in their eastern range (Ishiyama et al. 2020). Due to a high biological plasticity P. parva is probably the most opportunistic fish among the studied species. Like sunbleak, P. parva exhibits parental care, an additional competitive advantage and valuable trait for an invasive species (Pinder & Gozlan 2003). In consequence topmouth gudgeon are able to invade a wider range of different habitats and are currently reported from subtropical waters, such as Vietnam (Karabanov & Kodukhova 2013) and alpine streams on the Tibetan Plateau (Jia et al. 2019). Simultaneously, as mentioned above, the topmouth gudgeon can be a vector of parasites in newly inhabited areas, and consequently poses a particular threat to lake minnow or sunbleak populations (Andreou et al. 2011). Nevertheless, it should be noted that currently both P. parva and L. delineatus owe their success in expanding their range primarily to the aquaculture activities, including pond cultivation and stocking (Pinder & Gozlan 2003).
Thermal requirements have, or may have, a substantial impact on the future fate of particular species and the formation of inland fish assemblages in response to increasing water temperature due to climate change (Jeppesen et al. 2010;Pletterbauer et al. 2015;Lynch et al. 2016;Myers et al. 2017). The effect of temperature rising is relatively well described for tropical and marine fishes (e.g. Clark et al. 2017;Illing et al. 2020), but its impact on different life stages and body sizes is ambiguous and varies by species (Ospina & Mora 2004;McKenzie et al. 2021).
The smallest water bodies are strongly influenced by external environmental factors, which determine their trophic state and biota (Tonn et al. 1990;Jackson et al. 2001;Scheffer & van Nes 2007;Matthews 2010). Among other outcomes, water temperature determines both the likelihood of survival of local native fish and the possibility and range expansion of alien species (Comte et al. 2013;Rolls et al. 2017;Kärcher et al. 2019). Thermoregulatory behaviour as a physiological response to thermal stress enables fish to adapt to changing environmental conditions (Crawshaw 1977;Goyer et al. 2014;Haesemeyer 2020). However, in small, isolated water bodies, fish have a limited ability to exhibit thermoregulatory tactics, e.g. through emigration. According to our results risk of physiological disorders and direct fish survival in the face of a temperature increase to values above 30°C will create a significant threat to native populations of sunbleak and lake minnow, but will exert a much lower effect on invasive P. parva. Currently, such "overheating" scenarios for L. delineatus and E. percnurus habitats are likely (Lynch et al. 2016;Woolway et al. 2020;Dokulil et al. 2021). Moreover, rising temperatures as a result of climate change accelerate the eutrophication of shallow lakes in the temperate region (Jeppesen et al. 2014), as well as overgrowth, drying, and terrestrialisation of peat bogs and oxbow lakes. Resulting habitat loss might be a more important threat to both species (Sowińska-Świerkosz & Kolejko 2019; Wolnicki et al. 2022).
In conclusion, our results demonstrated that P. parva have the greatest resistance to temperature increase, while L. delineatus the lowest. The results obtained for the latter species were unexpected as we predicted that this leuciscid fish would be more tolerant of higher temperatures than E. percnurus. The following questions arise: (i) how to protect best the last E. percnurus habitats in Europe against P. parva invasions to avoid impacts already observed in the eastern part of lake minnow occurrence range (Ishiyama et al. 2020), and (ii) which species, L. delineatus or P. parva, will be more competitive in extending its range to the north, especially if they occur in sympatry. The present study was restricted to the upper thermal tolerance limits only, but the literature shows that P. parva can also adapt well to low temperatures, as was observed in the Tibetan Plateau (Jia et al. 2019). Our field observations indicate that occurrence of sunbleak in small, isolated ponds is often an ephemeral phenomenon. The maximum water temperatures in such water bodies in summer may be close to the upper tolerance threshold of this species, making temperature a key limiting factor. A similar explanation is probable in the case of some lake minnow extinction events. For this reason, the fate of the studied species is dependent on the degree of temperature change in the future. Since juvenile forms have not been tested in this work, further research is needed on the different ontogenic stages of the species shown.