PHYSIOLOGICAL ECOLOGY
Seasonal Changes in the Critical Thermal Maxima of Four Species of Aquatic Insects (Ephemeroptera, Trichoptera) DAVID C. HOUGHTON1,2
AND
LOGAN SHOUP1,3
Environ. Entomol. 43(4): 1059Ð1066 (2014); DOI: http://dx.doi.org/10.1603/EN13344
ABSTRACT Seasonal changes in the critical thermal maxima (CTmax) of four species of aquatic insects were determined from February 2012 to February 2013 from a Þrst-order stream in northern Lower Michigan. Three of these species: Stenonema femoratum (Ephemeroptera: Heptageniidae), Hydropsyche slossonae (Trichoptera: Hydropsychidae), and Dolophilodes distinctus (Trichoptera: Philopotamidae) exhibited seasonal changes in CTmax, increasing through the spring and summer and then decreasing into the subsequent fall and winter. CTmax of these species correlated strongly with both the seasonal ambient stream temperature and with a series of different laboratory acclimation temperatures, suggesting that organisms adapt to laboratory acclimation in a similar manner as they adapt to seasonal changes. In contrast, the CTmax of Parapsyche apicalis (Trichoptera: Arctopsychidae) remained constant regardless of ambient or acclimation temperature. All species exhibited greater thermal sensitivity relative to ambient temperature during the summer than the winter. Our study indicates that thermal tolerance patterns can be different among species in the same environment. It also provides the Þrst winter and year-round thermal tolerance data for aquatic insects. KEY WORDS critical thermal maxima (CTmax), thermal, heat, tolerance, aquatic
Temperature is widely recognized as one of the most important variables inßuencing the distribution, life history, and ecology of aquatic organisms, nearly all of which are exothermic (Caissie 2006, Haidekker and Hering 2008, Dallas and Rivers-Moore 2012). Many anthropogenic activities increase the temperature of freshwater ecosystems, for example: thermal discharges from industrial sites or agriculture (Lessard and Hayes 2003), removal of riparian vegetation with subsequent increase in sunlight penetration (Rutherford et al. 1997, Houghton and Wasson 2013), removal of groundwater and other subsurface changes (LeBlanc et al. 1997), and the developing effects of anthropogenic climate change (Daufresne et al. 2004, Hoffman and Sgro 2011). These concerns necessitate studies that determine high temperature tolerances of freshwater organisms, especially in light of the current documented decline of such organisms (Ricciardi and Rasmussen 1999, DeWalt et al. 2005, Lynse et al. 2008, Houghton and Holzenthal 2010). Unfortunately, knowledge about high temperature tolerances of aquatic insects is still lacking. The most comprehensive study (Dallas and Rivers-Moore 2012) documented upper temperature tolerances for 23 South African insect families within the aquatic Coleoptera, Hemiptera, Diptera, Megaloptera, Plecop1 Department of Biology, Hillsdale College, 33 East College St., Hillsdale, MI 49242. 2 Corresponding author, e-mail: [email protected]. 3 Current address: Department of Zoology, Southern Illinois University, 900 S. Normal Ave., Woody Hall-MC 4716, Carbondale, IL 62901.
tera, Ephemeroptera, and Trichoptera. Many of their specimens, however, were not identiÞed to the species level. For North American aquatic insects, laboratory studies on upper temperature tolerance have been conducted on ⬇8 stoneßy species (Heiman and Knight 1972, Ernst et al. 1984, Poulton et al. 1989, Shoup and Houghton 2013), 13 caddisßy species (Nebecker and Lemke 1968, GauÞn and Hern 1971, deKozlowski and Bunting 1981, Moulton et al. 1993, Houghton et al. 2014), 3 mayßy species (deKozlowski and Bunting 1981), and 2 dragonßy species (Garten and Gentry 1976). For most of these studies, high temperature tolerance has been deÞned nonlethally, through the determination of the critical thermal maximum or CTmax (Cowles and Bogert 1944). CTmax is the temperature at which a predetermined nonlethal behavioral endpoint is reached in the laboratory after an experimental increase in temperature. For aquatic invertebrates, this endpoint usually involves a loss of equilibrium or grip on the substrate (Dallas and Rivers-Moore 2012). Variables that affect the experimentally determined CTmax of an organism include length of time held in captivity before experimentation (Mora and Maya 2006), body size (Ribeiro et al. 2012), concentration of oxygen or chemical pollution within the experimental environment, (Poulton et al. 1989, Galbraith et al. 2012), rate of experimental temperature increase (Mora and Maya 2006, Ribiero et al. 2012, Houghton et al. 2014), and, especially, acclimation temperature (Ernst et al. 1984, Moulton et al. 1993, Jumbam et al. 2008, Kumlu and Turkmen 2010, Galbraith et al. 2012).
0046-225X/14/1059Ð1066$04.00/0 䉷 2014 Entomological Society of America
1060
ENVIRONMENTAL ENTOMOLOGY
Vol. 43, no. 4
40 Air temperature Stream temperature 30
Temperature (°C)
20
10
0
-10
*
-20
2012
2013
Fig. 1. Air and water temperature for our study site during the study period. Arrows indicate approximate dates of seasonal CTmax trials. Arrow with asterisk indicates beginning date of CTmax laboratory acclimation experiment.
Not surprisingly, the thermal history of the organism is also important. Organisms that live in warmer environments tend to have a higher CTmax than those that live in cooler environments (Garten and Gentry 1976, Moulton et al. 1993, Hopkin et al. 2006, Nyamukondiwa and Terblanche 2010, Dallas and RiversMoore 2012). Due to the importance of acclimation temperature and thermal history, we suspect that seasonality should also inßuence CTmax. Seasonal changes in CTmax have been previously noted in terrestrial insects (Hu and Appel 2004) and in aquatic crustaceans (Layne et al. 1987). In the latter study, the authors determined a 14.7⬚C difference in CTmax between winter and summer in the rusty crayÞsh (Orconectes rusticus Girard) from a population exposed to ambient temperatures ranging from 0 to 30⬚C. No similar study has been done for aquatic insects. This study had two objectives. The Þrst was to determine differences in the CTmax of several species of aquatic insects based on seasonal changes in ambient water temperature. The second was to determine differences in the CTmax of these same insects based on changes in laboratory acclimation temperature to see if this response was similar to that of seasonal ambient temperature. Materials and Methods Study Site. Specimens were collected from a 100-m stretch of Fairbanks Creek (44.04⬚ N, 85.66⬚ W), a Þrst-order stream located in the northwestern portion of the Lower Peninsula of Michigan. A detailed description of this site can be found in Houghton and
Wasson (2013). As is the case with many low-order streams of the northcentral United States (Meyer et al. 2007), this particular site of Fairbanks Creek is relatively thermally unstable due to a past history of agricultural use and a resulting lack of riparian canopy. Thus, results obtained from it should be indicative of the thermal tolerance of low-order stream insects throughout the region. Air and stream temperatures were collected every Friday afternoon from January 2012 to May 2013 using a digital thermometer (Fig. 1). Collecting and Laboratory Maintenance. Four species were selected for CTmax studies due to their abundance at the site and their different types of life cycles and habitat preferences. Stenonema femoratum (Say) (Ephemeroptera: Heptageniidae) has an extended summer emergence and is common in nearly all types of streams (Edmunds et al. 1974). Hydropsyche slossonae (Banks) (Trichoptera: Hydropsychidae) and Dolophilodes distinctus (Walker) (Trichoptera: Philopotamidae) are most abundant in small and cold streams with a summer adult emergence, although the latter species may also have an early spring generation (Houghton 2012). Parapsyche apicalis (Banks) (Trichoptera: Arctopsychidae) is limited to small, cold streams and emerges in the late spring or early summer, although some cold-water Canadian populations have extended adult emergence into early fall (Singh et al. 1984, Williams and Hogg 1988). Specimens were collected by hand from February 2012 to February 2013. We were unable to test P. apicalis during much of the summer and early fall due to the absence of larvae in the stream. We were able to Þnd larvae of the other species throughout our study period, although we had to use early (⬇2/3 of
August 2014
HOUGHTON AND SHOUP: SEASONAL CHANGES IN CTmax IN AQUATIC INSECTS
maximum size) instars of S. femoratum in September, and third to fourth instars of H. slossonae and D. distinctus in July and September. In all other cases, Þfth (terminal) instars were used for the caddisßies. For S. femoratum, nymphs were at or near maximum size, although the high number of instars of this species precluded a deÞnite determination. In the laboratory, all specimens were housed without food in ßow-through containers within Frigid Units Living Stream environments (www.frigidunits.com), set to ambient photoperiod. Living Stream water was composed of ⬇20% stream water and ⬇80% unchlorinated well water. Caddisßies were removed from their Þxed retreats before housing them in the Living Stream habitats. The identity of specimens was conÞrmed in both the Living Stream and in the water bath before trials. In the cases of S. femoratum, P. apicalis, and D. distinctus, identiÞcation was simple, as there were no known congeners at the site and the generic characters found in Merrit et al. (2008) were easy to identify macroscopically, even in earlier instars. Because Fairbanks Creek has been sampled extensively for several years, it is unlikely that other similar species exist at the site. In the case of H. slossonae, the large central yellow frontoclypeal spot diagnosed in Schuster and Etnier (1978) was distinct in thirdÐÞfth instar larvae and visible macroscopically with proper lighting. This character appeared sufÞcient to differentiate H. slossonae from Hydropsyche sparna (Ross) and Hydropsyche betteni Ross, which were also known from this site. We also examined specimens under the microscope after trials to conÞrm their identity. Data from one February 2012 trial of H. slossonae are not included in this study due to taxonomic confusion with H. sparna. Subsequent trials contained only H. slossonae. Voucher specimens of all species have been deposited in the Hillsdale College Insect Collection. Experimental Trials. CTmax trials were conducted using a Julabo MB-13 circulating heater (www.julabo.com) set to 40% external and 60% internal circulation. The device was linked to a computer using Julabo EasyTemp software, allowing for precise programming and logging of temperature protocols. In each trial, specimens were placed into a bath containing Living Stream water, given both natural stream rocks and 1 by 1 mm latex window screen to use as substrate, and allowed to orient themselves relative to the current for 5 min before the temperature was raised. Water temperature began at ambient and was raised by 0.33⬚C per minute (Dallas and Rivers-Moore 2012, Houghton et al. 2014) until CTmax was reached for all trial specimens. CTmax was deÞned as the inability to cling to substrate and, thus, being dislodged by the current. Specimens temporarily dislodged by the current or by other specimens were left in the water bath if they were able to reattach and assume a normal posture on the substrate. Once CTmax was reached for a specimen, it was removed from the water bath and placed into an 850-ml bowl which was ßoated in the Living Stream to cool specimens back to acclimation temperature over
1061
a 30- to 60-min period. Once acclimation temperature was reached, specimens were returned to the Living Stream and their survival checked at 24 h. Seasonal Acclimation. To assess seasonal differences in thermal tolerance for the four species, CTmax trials were conducted on larvae throughout a 1-yr period year. SpeciÞcally, we conducted trials in February, April, June, July, September, October, and November 2012, and in January 2013. Due to different abundances of larvae in the Þeld, the number of specimens in a single trial ranged 10 Ð25 for S. femoratum, H. slossonae, and D. distinctus, and 5Ð10 for P. apicalis. In all cases, acclimation temperature was set to ambient temperature. All four species were collected on the same day, acclimated in the Living Stream for 12Ð18 h before CTmax trials, and all tested on the next day. If larvae were abundant enough in the stream, this procedure was repeated on subsequent days to yield several trials. Laboratory Acclimation. To assess differences in thermal tolerance based on acclimation temperature, trials were also conducted in March 2013. Water temperature in Fairbanks Creek during this period was 3Ð 4⬚C. As above, all four species were collected on the same day, gradually brought to acclimation temperature over a period of several hours, and then acclimated to their particular temperature for an additional 12Ð18 h. CTmax values were determined on the next day. Acclimation temperatures were 0.5, 5.0, 10.0, 15.0, and 20.0⬚C in random order. Two cycles of this procedure were done within a 4-d period. Sample size for these experiments was 10 specimens per trial for all species. Statistical Methods. To assess changes in CTmax based on seasonal and laboratory acclimation temperature, the mean CTmax of each individual trial of specimens was treated as an independent data point and correlated with acclimation temperature. As mentioned above, acclimation temperature was either ambient temperature in the stream or one of the Þve laboratory acclimation temperatures during the March 2013 experiment. To assess seasonal differences in CTmax between the individual species, all specimens for each species were grouped together, regardless of the number of trials, to determine a global mean for each species for each month tested. The stream temperature was subtracted from each monthly value, and the results were compared using a one-way analysis of variance with a post hoc Tukey test. Data were not transformed before analysis, as normality assumptions were met. By subtracting stream temperature from CTmax, differences in CTmax relative to seasonal ambient temperature could also be compared. Results Stream temperature ranged from ⬇0⬚C in the winter to 24⬚C in the summer (Fig. 1) and was strongly correlated with air temperature (r2 ⫽ 0.90; P ⬍ 0.001; Fig. 2). Mortality during the acclimation period was ⬍10% except for two trials of P. apicalis noted below. Posttrial mortality was ⬍5% for all species and accli-
1062
ENVIRONMENTAL ENTOMOLOGY
ter, spring, and fall (Fig. 4). During summer, this value was similar for S. femoratum, H. slossonae, and D. distinctus. It was smallest for P. apicalis in all months that the species was present. For all species, the difference between CTmax and ambient temperature was larger during the cooler months of the year and smaller during the warmer months (Fig. 4).
Stream temperature (°C)
30
y = 0.57x + 2.47 R² = 0.90 P
Seasonal Changes in the Critical Thermal Maxima of Four Species of Aquatic Insects (Ephemeroptera, Trichoptera) DAVID C. HOUGHTON1,2
AND
LOGAN SHOUP1,3
Environ. Entomol. 43(4): 1059Ð1066 (2014); DOI: http://dx.doi.org/10.1603/EN13344
ABSTRACT Seasonal changes in the critical thermal maxima (CTmax) of four species of aquatic insects were determined from February 2012 to February 2013 from a Þrst-order stream in northern Lower Michigan. Three of these species: Stenonema femoratum (Ephemeroptera: Heptageniidae), Hydropsyche slossonae (Trichoptera: Hydropsychidae), and Dolophilodes distinctus (Trichoptera: Philopotamidae) exhibited seasonal changes in CTmax, increasing through the spring and summer and then decreasing into the subsequent fall and winter. CTmax of these species correlated strongly with both the seasonal ambient stream temperature and with a series of different laboratory acclimation temperatures, suggesting that organisms adapt to laboratory acclimation in a similar manner as they adapt to seasonal changes. In contrast, the CTmax of Parapsyche apicalis (Trichoptera: Arctopsychidae) remained constant regardless of ambient or acclimation temperature. All species exhibited greater thermal sensitivity relative to ambient temperature during the summer than the winter. Our study indicates that thermal tolerance patterns can be different among species in the same environment. It also provides the Þrst winter and year-round thermal tolerance data for aquatic insects. KEY WORDS critical thermal maxima (CTmax), thermal, heat, tolerance, aquatic
Temperature is widely recognized as one of the most important variables inßuencing the distribution, life history, and ecology of aquatic organisms, nearly all of which are exothermic (Caissie 2006, Haidekker and Hering 2008, Dallas and Rivers-Moore 2012). Many anthropogenic activities increase the temperature of freshwater ecosystems, for example: thermal discharges from industrial sites or agriculture (Lessard and Hayes 2003), removal of riparian vegetation with subsequent increase in sunlight penetration (Rutherford et al. 1997, Houghton and Wasson 2013), removal of groundwater and other subsurface changes (LeBlanc et al. 1997), and the developing effects of anthropogenic climate change (Daufresne et al. 2004, Hoffman and Sgro 2011). These concerns necessitate studies that determine high temperature tolerances of freshwater organisms, especially in light of the current documented decline of such organisms (Ricciardi and Rasmussen 1999, DeWalt et al. 2005, Lynse et al. 2008, Houghton and Holzenthal 2010). Unfortunately, knowledge about high temperature tolerances of aquatic insects is still lacking. The most comprehensive study (Dallas and Rivers-Moore 2012) documented upper temperature tolerances for 23 South African insect families within the aquatic Coleoptera, Hemiptera, Diptera, Megaloptera, Plecop1 Department of Biology, Hillsdale College, 33 East College St., Hillsdale, MI 49242. 2 Corresponding author, e-mail: [email protected]. 3 Current address: Department of Zoology, Southern Illinois University, 900 S. Normal Ave., Woody Hall-MC 4716, Carbondale, IL 62901.
tera, Ephemeroptera, and Trichoptera. Many of their specimens, however, were not identiÞed to the species level. For North American aquatic insects, laboratory studies on upper temperature tolerance have been conducted on ⬇8 stoneßy species (Heiman and Knight 1972, Ernst et al. 1984, Poulton et al. 1989, Shoup and Houghton 2013), 13 caddisßy species (Nebecker and Lemke 1968, GauÞn and Hern 1971, deKozlowski and Bunting 1981, Moulton et al. 1993, Houghton et al. 2014), 3 mayßy species (deKozlowski and Bunting 1981), and 2 dragonßy species (Garten and Gentry 1976). For most of these studies, high temperature tolerance has been deÞned nonlethally, through the determination of the critical thermal maximum or CTmax (Cowles and Bogert 1944). CTmax is the temperature at which a predetermined nonlethal behavioral endpoint is reached in the laboratory after an experimental increase in temperature. For aquatic invertebrates, this endpoint usually involves a loss of equilibrium or grip on the substrate (Dallas and Rivers-Moore 2012). Variables that affect the experimentally determined CTmax of an organism include length of time held in captivity before experimentation (Mora and Maya 2006), body size (Ribeiro et al. 2012), concentration of oxygen or chemical pollution within the experimental environment, (Poulton et al. 1989, Galbraith et al. 2012), rate of experimental temperature increase (Mora and Maya 2006, Ribiero et al. 2012, Houghton et al. 2014), and, especially, acclimation temperature (Ernst et al. 1984, Moulton et al. 1993, Jumbam et al. 2008, Kumlu and Turkmen 2010, Galbraith et al. 2012).
0046-225X/14/1059Ð1066$04.00/0 䉷 2014 Entomological Society of America
1060
ENVIRONMENTAL ENTOMOLOGY
Vol. 43, no. 4
40 Air temperature Stream temperature 30
Temperature (°C)
20
10
0
-10
*
-20
2012
2013
Fig. 1. Air and water temperature for our study site during the study period. Arrows indicate approximate dates of seasonal CTmax trials. Arrow with asterisk indicates beginning date of CTmax laboratory acclimation experiment.
Not surprisingly, the thermal history of the organism is also important. Organisms that live in warmer environments tend to have a higher CTmax than those that live in cooler environments (Garten and Gentry 1976, Moulton et al. 1993, Hopkin et al. 2006, Nyamukondiwa and Terblanche 2010, Dallas and RiversMoore 2012). Due to the importance of acclimation temperature and thermal history, we suspect that seasonality should also inßuence CTmax. Seasonal changes in CTmax have been previously noted in terrestrial insects (Hu and Appel 2004) and in aquatic crustaceans (Layne et al. 1987). In the latter study, the authors determined a 14.7⬚C difference in CTmax between winter and summer in the rusty crayÞsh (Orconectes rusticus Girard) from a population exposed to ambient temperatures ranging from 0 to 30⬚C. No similar study has been done for aquatic insects. This study had two objectives. The Þrst was to determine differences in the CTmax of several species of aquatic insects based on seasonal changes in ambient water temperature. The second was to determine differences in the CTmax of these same insects based on changes in laboratory acclimation temperature to see if this response was similar to that of seasonal ambient temperature. Materials and Methods Study Site. Specimens were collected from a 100-m stretch of Fairbanks Creek (44.04⬚ N, 85.66⬚ W), a Þrst-order stream located in the northwestern portion of the Lower Peninsula of Michigan. A detailed description of this site can be found in Houghton and
Wasson (2013). As is the case with many low-order streams of the northcentral United States (Meyer et al. 2007), this particular site of Fairbanks Creek is relatively thermally unstable due to a past history of agricultural use and a resulting lack of riparian canopy. Thus, results obtained from it should be indicative of the thermal tolerance of low-order stream insects throughout the region. Air and stream temperatures were collected every Friday afternoon from January 2012 to May 2013 using a digital thermometer (Fig. 1). Collecting and Laboratory Maintenance. Four species were selected for CTmax studies due to their abundance at the site and their different types of life cycles and habitat preferences. Stenonema femoratum (Say) (Ephemeroptera: Heptageniidae) has an extended summer emergence and is common in nearly all types of streams (Edmunds et al. 1974). Hydropsyche slossonae (Banks) (Trichoptera: Hydropsychidae) and Dolophilodes distinctus (Walker) (Trichoptera: Philopotamidae) are most abundant in small and cold streams with a summer adult emergence, although the latter species may also have an early spring generation (Houghton 2012). Parapsyche apicalis (Banks) (Trichoptera: Arctopsychidae) is limited to small, cold streams and emerges in the late spring or early summer, although some cold-water Canadian populations have extended adult emergence into early fall (Singh et al. 1984, Williams and Hogg 1988). Specimens were collected by hand from February 2012 to February 2013. We were unable to test P. apicalis during much of the summer and early fall due to the absence of larvae in the stream. We were able to Þnd larvae of the other species throughout our study period, although we had to use early (⬇2/3 of
August 2014
HOUGHTON AND SHOUP: SEASONAL CHANGES IN CTmax IN AQUATIC INSECTS
maximum size) instars of S. femoratum in September, and third to fourth instars of H. slossonae and D. distinctus in July and September. In all other cases, Þfth (terminal) instars were used for the caddisßies. For S. femoratum, nymphs were at or near maximum size, although the high number of instars of this species precluded a deÞnite determination. In the laboratory, all specimens were housed without food in ßow-through containers within Frigid Units Living Stream environments (www.frigidunits.com), set to ambient photoperiod. Living Stream water was composed of ⬇20% stream water and ⬇80% unchlorinated well water. Caddisßies were removed from their Þxed retreats before housing them in the Living Stream habitats. The identity of specimens was conÞrmed in both the Living Stream and in the water bath before trials. In the cases of S. femoratum, P. apicalis, and D. distinctus, identiÞcation was simple, as there were no known congeners at the site and the generic characters found in Merrit et al. (2008) were easy to identify macroscopically, even in earlier instars. Because Fairbanks Creek has been sampled extensively for several years, it is unlikely that other similar species exist at the site. In the case of H. slossonae, the large central yellow frontoclypeal spot diagnosed in Schuster and Etnier (1978) was distinct in thirdÐÞfth instar larvae and visible macroscopically with proper lighting. This character appeared sufÞcient to differentiate H. slossonae from Hydropsyche sparna (Ross) and Hydropsyche betteni Ross, which were also known from this site. We also examined specimens under the microscope after trials to conÞrm their identity. Data from one February 2012 trial of H. slossonae are not included in this study due to taxonomic confusion with H. sparna. Subsequent trials contained only H. slossonae. Voucher specimens of all species have been deposited in the Hillsdale College Insect Collection. Experimental Trials. CTmax trials were conducted using a Julabo MB-13 circulating heater (www.julabo.com) set to 40% external and 60% internal circulation. The device was linked to a computer using Julabo EasyTemp software, allowing for precise programming and logging of temperature protocols. In each trial, specimens were placed into a bath containing Living Stream water, given both natural stream rocks and 1 by 1 mm latex window screen to use as substrate, and allowed to orient themselves relative to the current for 5 min before the temperature was raised. Water temperature began at ambient and was raised by 0.33⬚C per minute (Dallas and Rivers-Moore 2012, Houghton et al. 2014) until CTmax was reached for all trial specimens. CTmax was deÞned as the inability to cling to substrate and, thus, being dislodged by the current. Specimens temporarily dislodged by the current or by other specimens were left in the water bath if they were able to reattach and assume a normal posture on the substrate. Once CTmax was reached for a specimen, it was removed from the water bath and placed into an 850-ml bowl which was ßoated in the Living Stream to cool specimens back to acclimation temperature over
1061
a 30- to 60-min period. Once acclimation temperature was reached, specimens were returned to the Living Stream and their survival checked at 24 h. Seasonal Acclimation. To assess seasonal differences in thermal tolerance for the four species, CTmax trials were conducted on larvae throughout a 1-yr period year. SpeciÞcally, we conducted trials in February, April, June, July, September, October, and November 2012, and in January 2013. Due to different abundances of larvae in the Þeld, the number of specimens in a single trial ranged 10 Ð25 for S. femoratum, H. slossonae, and D. distinctus, and 5Ð10 for P. apicalis. In all cases, acclimation temperature was set to ambient temperature. All four species were collected on the same day, acclimated in the Living Stream for 12Ð18 h before CTmax trials, and all tested on the next day. If larvae were abundant enough in the stream, this procedure was repeated on subsequent days to yield several trials. Laboratory Acclimation. To assess differences in thermal tolerance based on acclimation temperature, trials were also conducted in March 2013. Water temperature in Fairbanks Creek during this period was 3Ð 4⬚C. As above, all four species were collected on the same day, gradually brought to acclimation temperature over a period of several hours, and then acclimated to their particular temperature for an additional 12Ð18 h. CTmax values were determined on the next day. Acclimation temperatures were 0.5, 5.0, 10.0, 15.0, and 20.0⬚C in random order. Two cycles of this procedure were done within a 4-d period. Sample size for these experiments was 10 specimens per trial for all species. Statistical Methods. To assess changes in CTmax based on seasonal and laboratory acclimation temperature, the mean CTmax of each individual trial of specimens was treated as an independent data point and correlated with acclimation temperature. As mentioned above, acclimation temperature was either ambient temperature in the stream or one of the Þve laboratory acclimation temperatures during the March 2013 experiment. To assess seasonal differences in CTmax between the individual species, all specimens for each species were grouped together, regardless of the number of trials, to determine a global mean for each species for each month tested. The stream temperature was subtracted from each monthly value, and the results were compared using a one-way analysis of variance with a post hoc Tukey test. Data were not transformed before analysis, as normality assumptions were met. By subtracting stream temperature from CTmax, differences in CTmax relative to seasonal ambient temperature could also be compared. Results Stream temperature ranged from ⬇0⬚C in the winter to 24⬚C in the summer (Fig. 1) and was strongly correlated with air temperature (r2 ⫽ 0.90; P ⬍ 0.001; Fig. 2). Mortality during the acclimation period was ⬍10% except for two trials of P. apicalis noted below. Posttrial mortality was ⬍5% for all species and accli-
1062
ENVIRONMENTAL ENTOMOLOGY
ter, spring, and fall (Fig. 4). During summer, this value was similar for S. femoratum, H. slossonae, and D. distinctus. It was smallest for P. apicalis in all months that the species was present. For all species, the difference between CTmax and ambient temperature was larger during the cooler months of the year and smaller during the warmer months (Fig. 4).
Stream temperature (°C)
30
y = 0.57x + 2.47 R² = 0.90 P
