Mysids are an important component in the food web of coastal zones, linking benthic and pelagic systems by nutrient exchange and biomass transfer (Roast et al., 1998). They can switch between two feeding modes (suspension and raptorial feeding), allowing them to feed on detritus and phytoplankton, but also to actively select zooplankton (Viitasalo and Rautio, 1998). In turn, they serve as a protein-rich (>70% of dry weight, Raymont et al., 1968) food source for higher trophic levels. We used Neomysis integer Leach, 1814 (Mysidacea: mysidae) in our experiments. N. integer occurs in the hyperbenthic zones of estuaries and is also abundant in shallow coastal waters of the Baltic Sea  (Margoñski and Maciejewska, 1999). In estuaries and brackish water systems, such as the Greifswalder Bodden, thermal variability occurs on both daily and seasonal scales (Seifert, 1938; Schiewer, 2008).

N. integer were collected at water depths <0.5 m in a shallow bay (coordinates: 54.097495, 13.455123) of the Greifswalder Bodden, in the southwest of the Baltic Sea on 04^th May 2021. At collection, the water temperature was 11.3°C and the salinity was 7.5 psu. The bay is very shallow (<5 m), making it likely that mysids can experience strong thermal variability at short time scales. To avoid acute stress responses, especially in the warmer temperature treatments of the experiment, 150 mysids were kept in a climate chamber at constant 15°C for the next 16 days. Mysids stayed in aerated 40 L-aquaria and were fed ad libitum with freshly-hatched Artemia salina nauplii and fish pellets (MultiFit). The culture medium consisted of a 50:50% mixture of artificial seawater (Tropic Marin Pro-Reef Salt) and filtered (0.2 µm) water from the sample site. The salinity was set to 10 psu, somewhat higher than the ambient value, to establish a reference for future experiments and for comparative purposes with other experimental studies (e.g. Roast et al., 1999; Roast et al., 2000). At least 50% of this culture medium was exchanged every 3-4 days. The light regime in the climate chamber and during experiments was set to a 12:12 h dark:light photoperiod.

Before the ingestion measurements, mysids were exposed simultaneously for 7 days to 4 constant temperature treatments and to 2 fluctuating temperature treatments ( Figure 1 ). They were kept in 6 L-aquaria placed in different climate cabinets under the respective treatment temperature (1 climate cabinet per treatment) and fed ad libitum with freshly-hatched A. salina nauplii. Mysids in constant treatments were kept on 5 (♂ n = 13; ♀ n = 4), 10 (♂ n = 15; ♀ n = 2), 15 (♂ n = 5; ♀ n = 1), or 20°C (♂ n = 3; ♀ n = 10), whereas mysids in the two fluctuating treatments experienced 9 ± 5°C (♂ n = 12; ♀ n = 5) and 14 ± 5°C (♂ n = 11; ♀ n = 6) with a frequency of once per day (see Figure 2 for shape of the temperature fluctuation profile). The amplitude of ± 5°C was chosen to elicit a response and is likely larger than what the animals would experience on a diurnal basis in the environment. Highly controllable and precise climate cabinets were used (± 0.5 Kelvin) to avoid potential differences caused by cabinet identity.

Vivid N. integer were starved for one day prior to the measurements of ingestion rates (Lindén and Kuosa, 2004). Ingestion rates of N. integer individuals from both the constant and fluctuating treatments were measured in the respective constant mean temperatures of 5, 10, 15, and 20°C, and in case of the fluctuating treatments when temperature was close to the respective mean value. One climate cabinet was used for each constant mean temperature. Ingestion rates of mysids from the fluctuating treatments were measured in the same climate cabinets as ingestion rates of mysids from the respective constant temperature treatment to ensure comparability. For the measurements, N. integer were placed individually in small beakers containing 50 ml of the culture medium described above at an initial food density of 30 A. salina nauplii. After 1.5 h, N. integer individuals were removed and stored in ethanol for later sex and length determination. The remaining A. salina nauplii were counted, and ingestion rates per mysid per hour were calculated ( Figure 1 ). The sex of the mysids was identified based on sexual characteristics using a stereo-microscope. All females had a developed marsupium. Body length (from the base of eyestalks to the end of the last abdominal segment, excluding uropods and telson; Fockedey et al., 2005) was determined from photographs using the software ImageJ (mean body length ± SD, females: 11.11 ± 1.65 mm, males: 9.46 ± 0.87 mm). Since all mysids used in the experiment were adults and were sampled in May, i.e. at the beginning of the growing season providing ample of food, we expected all individuals being in a reproductive state.

Data analyses were performed in R version 4.0.2 (R Core Team, 2020). Individual ingestion rates at constant temperatures were analyzed using Analysis of Covariance (ANCOVA) with a negative quadratic function of temperature as continuous independent variable and sex of animals as factor. Preliminary inspection of the data revealed that our data were distributed rather in the upper part of the TPC above a potential inflection point where the second derivative of the TPC is always negative, i.e. the concave downward part of the TPC. For this region, a negative quadratic function is a reasonable simplification and fits the typical unimodal responses in ectotherm TPCs (Angilletta, 2006). The full model was simplified based on AIC using the R function step(). A potential decrease of the ingestion rates measured under fluctuating temperatures was tested against the performance prediction of the ANCOVA for the respective constant temperature with a one-sided, one sample t-test. We also tested for any deviation from the performance predicted under fluctuating temperature by performing a two-sided, one sample t-test.

To calculate ingestion rates predicted at fluctuating temperature regimes, we used the deviations of the actually realized temperatures (measured in the climate cabinet) from the mean temperature of the 9 ± 5°C fluctuation treatment. This treatment was chosen exemplary since the realized temperature deviations from the mean in the 14 ± 5°C treatment were very similar and resulted in similar calculated ingestion rates at the end. In detail, we conducted the following steps ( Figure 2 ):

Under constant temperature conditions, ingestion rates of N. integer showed a unimodal relationship to temperature for both sexes ( Figure 3A ), characterizing a TPC_Constant that can be well described by a negative quadratic function (ANCOVA; F_1,48 = 5.02, p = 0.030). The female TPC_Constant reached a higher maximum than the male TPC_Constant ( Figure 3A , sex: F_1,48 = 5.44, p = 0.024), and the thermal optimum (14.1°C and 10.7°C for females and males, respectively) was shifted towards higher temperatures (interaction Temp × Sex, F_1,48 = 9.03, p = 0.004).

The measured ingestion rates of female mysids exposed to the 14 ± 5°C regime ( Figure 3B , open symbol), close to their thermal optimum, were lower than the ingestion value predicted by the TPC_Constant (one-sided t-test, p = 0.018), but did not deviate positively or negatively from the value of the TPC_Fluctuation (two-sided t-test, p = 0.11). For female mysids exposed to the 9 ± 5°C regime, ingestion rates neither decreased compared to the value of the TPC_Constant (one-sided t-test, p = 0.11) nor did they deviate from the corresponding TPC_Fluctuation value (two-sided t-test, p = 0.72). In contrast to females, the ingestion rates of male mysids exposed to fluctuating temperatures ( Figure 3C ) were very similar to the values predicted by the TPC_Constant for both corresponding temperatures (9 and 14°C, one-sided t-tests, p > 0.63) and did also not deviate from the two values of the TPC_Fluctuation (two-sided t-tests, p > 0.10).

Our study showed first results that temperature fluctuations can affect ingestion rates, a process influenced by metabolism and behavior of marine crustaceans, and that differences in responses even occur between females and males of the same population. In our case, we found evidence for differences in sex-specific ingestion rates of N. integer. At constant temperatures, the thermal optimum of ingestion rates based on fitted TPCs was higher for females than for males, and at warmer temperatures, females showed higher ingestion rates than males. Temperature fluctuations of ± 5°C around the observed thermal optimum led to reduced ingestion rates for females, but not for males, when compared to ingestion rates observed at constant temperature.

In our study, ingestion rates of female and male mysids show a unimodal pattern across the tested temperature range, reflecting the expected non-linear nature of TPCs (e.g. Angilletta, 2006; Dowd et al., 2015). Whilst the TPC_Constant for females and males showed similar responses to lower temperatures, sex-specific differences were apparent at higher temperatures ( Figure 3A ). Female mysids showed both a higher thermal optimum and higher total ingestion at the optimum temperature compared to males. Simmons & Knight (1975) described sex-specific differences in performance for the mysid Neomysis intermedia under various temperature and salinity that were weight-dependent. The respiratory response of N. intermedia was also dependent on the status of gravidity (Simmons and Knight, 1975). Similarly, Simmons and Knight (1975) indicated linkages between the respiration of mysids and the level of reproduction. The higher ingestion rates of females observed at higher temperatures in our study could be therefore related to the reproductive status of the used individuals. For females, such a relation between a higher ingestion and a higher energy demand for reproduction, in form of glycogen, has also been discussed for other marine invertebrates, such as the oyster Pinctada (Chávez-Villalba et al., 2013). Further, Winkler and Greve (2002) found that mysids’ reproduction rates and growth factors for smaller females seem to be higher at 15°C than at 10°C, resulting in shorter intermoult periods, which could lead to higher energy demands and thus to higher ingestion rates at 15°C compared to 10°C, as observed in our study. In the context of reproduction, a higher lipid demand, necessary for gamete production, may have led to the observed higher thermal optimum of females and their higher total ingestion of Artemia. Support for this assumption arises from a previous study showing that ovigerous females of N. integer contained almost double the amount of lipids compared to males due to the lipid-rich eggs (Linford, 1965). The lipid content and the energy budget were dependent on the season and therefore also influenced by variations in prey and temperature (Linford, 1965; Verslycke and Janssen, 2002). The lipid amount also correlates with the wet weight of an animal, with sex-specific relative lipid concentrations (Linford, 1965). This may also have led to the observed higher ingestion rates of females compared to males in our study, as female N. integer were slightly larger than males. However, the relationship between ingestion rates and length in the different temperature treatments was neither significant for females nor for males (p > 0.4). Alternatively, differences in ingestion rates between males and females may be influenced by differences in thermal tolerance. For instance, lower thermal tolerance in males compared to females may explain lower ingestion rates of males at higher temperatures. So far, however, nothing is known about sex-specific differences in thermal tolerance of mysids based on fitness-related traits such as growth rate or survival, making it difficult to draw conclusions about this option.

Similar to our observed ingestion rates for females, Roast et al. (2000) observed a significant increase of egestion rates for N. integer with increasing temperature up to 15°C. In another study, Roast et al. (1999) found an interactive effect of temperature and sex on respiration, with males showing a higher oxygen consumption than females. Even though we also found a significant interaction between temperature and sex, our results showed the opposite pattern, with higher ingestion rates of females observed at higher temperature. However, Roast et al. (1999) did not measure feeding in their study, and therefore the direct link between respiration and feeding rates in N. integer remains unclear. In studies using the blue mussel Mytilus, it was shown that respiration and feeding are two different functional traits, which at times are decoupled from each other, suggesting that they should be compared with caution (Vajedsamiei et al., 2021).

Ingestion rates of female N. integer exposed to temperature fluctuations around their optimum were lower than ingestion rates at constant temperature, but not lower than ingestion rates predicted under thermal fluctuations, suggesting females behaved as predicted by Jensen’s Inequality. In contrast, ingestion rates of male mysids exposed to fluctuating temperatures were similar to the values of the TPC_Constant ( Figure 3C ). This suggests that males are not affected by time-dependent effects of thermal fluctuations, whilst females were affected by such effects at their thermal optimum. Based on our current knowledge, we can only speculate that males instantaneously respond to the changing conditions and therefore show a TPC_constant-predicted ingestion rate, while females deviate as they either are capable of cognitively integrating their feeding behavior across the temperature fluctuations or suffer from some time-dependent effect, such as a stress-related response.

The longer-term thermal histories of females and males used in our study should be similar as N. integer lives in swarms (Lindén, 2007), and our animals were collected from the same swarm. Swarms might experience similar temperatures in the field, but could also show sex-dependent habitat preferences (Vesakoski et al., 2008). The extent of thermal variability the mysids actually experience in the study region, in particular in high spatial and temporal resolution, is a matter of further investigations.

There are many different types of thermal variability in natural environments. Thermal fluctuations are not as regular or periodic as in laboratory studies, but rather stochastic (e.g. Pansch et al., 2018) and vertical or horizontal movements of mobile ectotherms can change the amplitudes and frequencies of fluctuations they experience (Koussoroplis et al., 2019). Further complexity is also added by the interplay of temperature with other factors, such as acidification, hypoxia, food availability, that may all vary in time and space (Cornwall et al., 2013; Comeau et al., 2014; Koussoroplis and Wacker, 2016). Thus, there is a lot to learn about predicting ectotherm performance in naturally fluctuating multifactorial systems, especially in the context of changing variability patterns and frequencies of extreme events due to climate change. To our knowledge, this study is the first investigating a TPC for mysids’ ingestion rates under constant temperatures, and combining it with ingestion rates under fluctuating conditions. We observed differences in ingestion rates between females and males in warmer temperatures at constant conditions and that females responded to temperature fluctuations whilst males did not. Investigations on the underlying mechanisms, that may be related to the reproductive status or other physiological differences in females and males, are of considerable interest for future research. We further advise caution when using TPCs based on measurements in constant conditions to predict performance in fluctuating environments, since time-dependent effects such as stress-induced damage or compensation can influence the ectotherms’ responses at variable conditions. Since we observed differences in sex-specific ingestion rates within one species, we propose that future studies should put more emphasis on studying potential differences between female and male individuals of studied populations. This would allow a better understanding of variation observed in thermal performance curves and possible effects of fluctuating conditions, and is crucial for improved predictions of species’ performance under climate change.