Perricone and Collin, Page 1 of 30 1 2 Larvae of Caribbean Echinoids Have Small Warming Tolerances for Chronic Stress in 3 Panama 4 5 6 Valentina Perricone1,2 and Rachel Collin1,3 7 8 9 10 1 Smithsonian Tropical Research Institute, Apartado Postal 0843-03092, Balboa Ancon, Panama. 11 12 2 Department of Biology and Evolution of Marine Organisms, Stazione Zoologica Anton Dohrn, 13 Villa Comunale, I-80121 Napoli, Italy 14 15 16 3Corresponding author: e-mail: collinr@si.edu; (202) 633-4700 x28766. Address for 17 correspondence: STRI, Unit 9100, Box 0948, DPO AA 34002, USA. 18 19 Biological Bulletin 20 21 Keywords: Echinometra, Lytechinus, thermal tolerance, coral reefs, Bocas del Toro, seagrass, 22 23 Short Title: Warming tolerance of echinoid larvae 24 25 October, 2018 26 27 Abbreviations: Eucidaris tribuloides: Et; Echinometra viridis: Ev; Echinometra lucunter: El; 28 Lytechinus williamsi: Lw; Tripneustes ventricosus: Tv; Clypeaster rosaceus: Cr; Clypeaster 29 subdepressus: Cs 30 Perricone and Collin, Page 2 of 30 Abstract 31 In species with complex life cycles, early developmental stages are often less thermally 32 tolerant than adults, suggesting they are key to predicting organismal response to environmental 33 warming. Here we document the optimal and lethal temperatures of larval sea urchins and use 34 those to calculate the warming tolerance and the thermal safety margin of early larval stages of 7 35 tropical species. Larvae of Echinometra viridis, Echinometra lucunter, Lytechinus williamsi, 36 Eucidaris tribuloides, Tripneustes ventricosus, Clypeaster rosaceus and Clypeaster 37 subdepressus, were reared at 26°C, 28°C, 30°C, 32°C and 34°C for 6 days. The temperatures at 38 which statistically significant reductions in larval performance are evident are generally the same 39 temperatures at which statistically significant reductions in larval survival were detected, 40 showing that the optimal temperature is very close to the lethal temperature. The two 41 Echinometra species had significantly higher thermal tolerance than the other species, with some 42 surviving culture temperatures of 34°C and showing minimal impacts on growth and survival at 43 32°C. In the other species, larval growth and survival were depressed at and above 30°C or 32°C. 44 Overall these larvae have lower warming tolerances (1°C to 5°C) and smaller thermal safety 45 margins (-3°C to 3°C) than adults. Survival differences among treatments were evident by the 46 first sampling on day 2, and survival at the highest temperatures increased when embryos were 47 exposed to warming after spending the first 24 hours at ambient temperature. This suggests that 48 the first days of development are more sensitive to thermal stress than are later larval stages. 49 50 Perricone and Collin, Page 3 of 30 Introduction 51 Environmental warming is one of the greatest current threats to biodiversity. Many 52 studies of thermal limits and thermal performance curves have been conducted on adult 53 ectotherms (Sunday et al., 2011; 2012; Nguyen et al., 2011). However, the majority of metazoan 54 animals have complex life cycles and, in general, early developmental stages are considered to 55 be more susceptible to environmental stressors (Byrne, 2012; Keshavmurthy et al., 2014). For 56 example, studies of crustaceans and echinoids have found that larval stages have lower thermal 57 tolerance than adults (e.g., Miller et al., 2013; Schiffer et al., 2014; Collin and Chan 2016) and 58 late larval stages show less tolerance to thermal stress than do early larval stages (Storch et al., 59 2011; Fitzgibbon and Battaglene, 2012). However, a few species show the opposite pattern, with 60 embryos and early life stages showing a higher tolerance to thermal stress than adults (Diederich 61 and Pechenik, 2013; Tangwancharoen and Burton, 2014). In general, the thermal tolerances of 62 embryonic and larval stages of marine invertebrates remain poorly documented, especially in the 63 tropics. 64 To fully understand the vulnerability of different life stages to warming, comparisons of 65 thermal tolerances need to be placed in an environmental context that accounts for the changes in 66 habitat between larval and adult stages (e.g., Lu et al., 2016). Two metrics have commonly been 67 used to do this. Warming tolerances (WT) are the differences between the environmental 68 temperature and the lethal temperature and thermal safety margins (TSM) are the differences 69 between the optimal temperature and the environmental temperature. Sufficient data on 70 terrestrial animals and marine fishes exist to form the basis of meta-analyses of these metrics. 71 They show that tropical marine organisms are more vulnerable to warming than are temperate 72 organisms (e.g., Deutsch et al. 2008; Comte and Olden 2017). This has been primarily based on 73 estimates of warming tolerances, which are generally around 5°C for tropical marine fishes and > 74 10°C for temperate marine fishes (Comte and Olden 2017). 75 Unfortunately, similar comparative data are not available for other groups of marine 76 organisms. This is especially true for information on optimal temperatures and the TSM, which 77 are vital to understanding the negative impacts of environmental warming on populations prior to 78 evidence of a lethal effect. Generally experimental studies of thermal tolerance detect non-lethal 79 reductions in performance at temperatures lower than those at which survival is impaired. These 80 negative effects can become apparent at temperatures much lower than the lethal temperature. 81 Perricone and Collin, Page 4 of 30 For example, larvae of Pollicipes elegans barnacles show a reduction in swimming activity at 5-82 10 °C below the lethal temperature (Walther et al., 2013) and larval spiny lobsters growth rate 83 decreases above 21°C, while reduced survival is not evident until temperatures exceed 25°C 84 (Fitzgibbon et al., 2017). Although large difference between the optimal and lethal temperatures 85 have been documented for temperate organisms, this may not be the case for tropical organisms, 86 where the more extreme left skew of the thermal performance curve may reduce the difference 87 between the optimal and lethal temperatures. If this difference is very small in tropical animals, 88 the more easily measured lethal temperatures and WT could be used in place of more 89 complicated measures of performance. 90 The objective of the present study was to document the optimal and lethal temperatures 91 of larval sea urchins and to use those in combination with environmental temperatures to 92 calculate the warming tolerance and the thermal safety margin of early larval stages of 7 tropical 93 species. To do this we documented the impact of rearing temperature on larval growth and 94 survival over the first 6 days of larval life and used this to answer the following primary 95 questions: As temperature increases, does performance decrease before reduced survival can be 96 detected and how different is the optimal temperature from the lethal temperature (i.e, how 97 different are WT and TSM)? In addition, we conducted a follow-up short-term rearing 98 experiment to determine if upper limits of thermal tolerance are altered when thermal stress is 99 imposed after the first day of development. 100 We focused on tropical sea urchins for several reasons. Sea urchins are well-known eco-101 system engineers that play important roles as herbivores in both reef and seagrass ecosystems 102 and can be important for erosion and bioturbation (Birkeland, 1989; Valentine and Heck, 1991; 103 Heck and Valentine, 1995; Perkins et al., 2015; Ling et al., 2015; Davidson and Grupe, 2015; 104 Lessios, 2016; Kuempel and Altieri, 2017). There is a significant body of literature on the 105 thermal tolerance of the developmental stages of temperate sea urchins (e.g., Fujisawa, 1995; 106 Sewell and Young, 1999; Byrne et al., 2009; 2011; 2013; Pecorino et al., 2013; Delorme and 107 Sewell, 2013; 2014; Gianguzza et al., 2014; Karelitz et al., 2017) upon which to base 108 comparisons, and the thermal tolerance of the adults of our focal species has previously been 109 documented (Collin et al. 2018). Finally, sea urchins are abundant charismatic megafauna in 110 shallow water tropical environments and their transcriptomic and cellular responses to 111 environmental stressors are relatively well-understood, for marine invertebrates, making them 112 Perricone and Collin, Page 5 of 30 potential candidates to serve as easily surveyed sentinel species (e.g., Bonacci et al., 2007; 113 Pinsino and Matranga, 2015). 114 115 Materials and Methods 116 We studied larvae of 7 species of common sea urchins from the Caribbean (the cidaroid 117 Eucidaris tribuloides; the echinometrid euechinoids Echinometra viridis and E. lucunter, the 118 toxopneustid euechinoids Lytechinus williamsi and Tripneustes ventricosus and the clypeasteroid 119 euechinoids Clypeaster rosaceus and C. subdepressus). These all develop through an obligatory 120 planktotrophic larval stage, except for C. rosaceus, which produces larger eggs that develop into 121 facultative feeding larvae (Schroeder, 1981; Emlet, 1986; Sewell and Young, 1999; Wolcott and 122 Messing, 2005; Vellutini and Migotto, 2010; McAlister and Moran, 2013). There is either little 123 or no published information on larval thermal tolerance for any of these species except for E. 124 lucunter (Sewell and Young, 1999), although some information on the thermal tolerance of 125 earlier developmental stages is available for 3 others (Cameron et al., 1985). 126 Adult sea urchins were collected from seagrass meadows and fringing reefs in the 127 shallow waters around Isla Colon and Isla Solarte, in Bocas del Toro, Panama. The 128 environmental conditions in Bocas del Toro have been well-documented over the last 10 years 129 (D’Croz et al., 2005; Kaufmann and Thompson, 2005; Collin et al., 2009; Neal et al., 2014; 130 Seemann et al., 2014). Adults were collected by hand from the same locations as described in 131 Collin et al. (2018) while snorkelling to a depth of less than 5 meters. Animals were maintained 132 submerged in buckets while they were transported to the Smithsonian Tropical Research 133 Institute's Bocas del Toro Research Station. They were maintained, unfed, outside in 250L tanks 134 of aerated running seawater at environmental temperature (28.5 ± 0.5 °C). 135 Animals were spawned no more than 7 days after collection following standard methods 136 (Strathmann, 1987). Adults were induced to spawn by injecting 2-5 ml 0.55M KCl into the 137 perivisceral cavity of each sea urchin. For each cross, we pooled the gametes of 3 females and 3 138 males to produce a mixed population of larvae, with up to 12 alleles at any one genetic locus. 139 The eggs were washed gently and diluted with filtered sea water to bring the final volume to 140 500ml. Then, 0.3 ml of diluted sperm from each of 3 males were added to each beaker of eggs. 141 After 15 minutes, embryos with fertilization envelopes were counted into culture vessels and 142 placed in waterbaths to gradually come to the treatment temperatures. 143 Perricone and Collin, Page 6 of 30 144 Experiment 1: Chronic Thermal Stress During Development 145 To understand the tolerance of larvae to chronic elevated temperatures, we reared larvae 146 for 6 days at six different temperatures: 26, 28, 30, 32, 34, 36°C. These temperatures were 147 selected as studies of adult thermal tolerance show that some Caribbean sea urchin species can 148 survive short-term exposures to temperatures as high as 36°C (Collin et al. 2018) and that 149 fertilization may occur up to 37°C (Sewell and Young, 1999). Large water baths were made 150 using 40L aquaria with heaters, and bubblers, to mix the water. Temperature was measured twice 151 daily to ensure that the experimental temperatures were maintained throughout the trial. For each 152 trial, we added 50 cleaving embryos to 150ml of 0.45 m-filtered seawater with 20 g/ml 153 ampicillin in a 250 ml Erlenmeyer flask. 15 such flasks were placed in each water bath on day 0. 154 Water was changed every 2 days (i.e., on days 2 and 4) by gently concentrating the larvae into a 155 volume of ~10-20ml using reverse filtration. The larvae were poured into a clean flask with new 156 filtered sea water. The larvae were fed by adding 1250 cells/ml of Rhodomonas sp. and 1250 157 cells/ml of Dunaliella sp. 12 hours after fertilization, and every day thereafter. 158 Starting on Day 2 (i.e., 48 hours post-fertilization) and every day thereafter, 3 flasks were 159 removed from each treatment. The number of surviving larvae in each flask was counted and, 160 when present, 20 individuals were photographed under a compound microscope for subsequent 161 measurement. The total body length, the stomach length and the average length of the postoral 162 arms from the posterior margin of the larva to the arm tip were measured using ImageJ. Previous 163 studies of echinoplueti have shown these measurements to reflect growth of larvae and differing 164 ratios of these measures to reflect allometries induced by differing food rations (Bertram and 165 Strathmann, 1998; Sewell et al. 2004). In this way, we obtained estimates of growth and survival 166 over 5 days (Days 2-6). We conducted two trials of this experiment for each species, using 6 167 different parents for each trial. The trials were conducted on different dates at least a month 168 apart. Unfortunately we could only complete one trial for C. subdepressus. Measurements were 169 taken from the first survival trial for all the species, and we also measured larvae in the second 170 trial for E. viridis (Table 1). 171 172 Experiment 2: Survival After Early Development 173 Perricone and Collin, Page 7 of 30 Because the first experiment showed that most of the survival differences between 174 temperature treatments were evident by Day 2 (see results section), we designed a follow-up 175 experiment to determine if early development is more sensitive to elevated temperatures than are 176 the hatched larvae. Such a difference has been previously reported for E. lucunter (Sewell and 177 Young, 1999). We followed the same spawning and rearing protocols as described above, with 178 the following modification. We held the bulk cultures from each female at 28°C (ambient 179 environmental temperature) for 24 hours before allocating the embryos into the 250ml flasks and 180 rearing them at 28, 30, 32, 34, and 36°C. Larvae were counted after 48-hour exposure to the 181 experimental temperature. This exposure means that the larvae were kept at the experimental 182 temperatures for the same duration as the Day 2 sample in Experiment 1, but that we counted 183 them at the same age as those counted on Day 3 in Experiment 1. Higher relative survival of 184 larvae exposed to stressfully warm conditions in this experiment compared to Experiment 1 185 would indicate that the first 24 hours of development are particularly susceptible to thermal 186 stress. We assessed this by calculating the survival of the larvae raised at 30°C, 32°C and 34°C 187 after 48 hours relative to larvae from the same experiment raised at 28°C. If the timing of 188 thermal stress is important these ratios should differ more at higher temperatures than at cooler, 189 non-stressful temperatures. 190 191 Statistical Analyses 192 Analyses of variance were used to determine how the fully factorial combination of 193 temperature treatment and day affected survival and growth. Survival data are counts of the 194 number of larvae found in the flasks on each day and were analysed as such rather than being 195 transformed into proportions. These data were analysed in JMP version 12, using a REML 196 (Restricted Maximum Likelihood) ANOVA, the preferred method for model fitting, which uses 197 maximum likelihood estimation of a restricted likelihood function that is independent of fixed-198 effect parameter (Patterson and Thompson 1974; Searle et al. 1992). The REML ANOVA 199 model included a fully factorial combination of temperature and day, and trial was included as a 200 random effect to account for variation due to differences between the 2 trials. The distributions 201 of the residuals were checked for normality and data were transformed if necessary. Only the 202 survival data for E. viridis could not be adequately transformed in this way, but it should be 203 noted that ANOVA analyses are generally robust to the violation of the normality assumption 204 Perricone and Collin, Page 8 of 30 with similar group sizes (Glass et al. 1972 ; Schmider et al. 2010). Larval measurements were 205 analysed with the same fixed factors, and included jar as a random effect nested within day and 206 temperature, to avoid over-inflated degrees of freedom because we measured multiple larvae 207 from each jar. Post-hoc Tukey HSD comparisons were used to determine which treatments 208 differed significantly from each other. Repeated measures ANOVA was not used as each jar was 209 independent and only counted and/or measured once. 210 211 Results 212 Survival in Experiment 1 - Larvae of each species survived in cultures at 26-32°C through the 6 213 days of the experiment. In addition, some larvae of both Echinometra species also survived at 214 34°C (Table 1 and Figure 1). No larvae of the Echinometra species survived to Day 2 at 36°C 215 and no larvae of the other species survived to Day 2 at 34°C or at 36°C. 216 For all of the species ANOVA analysis showed significant effects of temperature on 217 larval survival (Table 2; Figure 2). In all but 2 species (T. ventricosus and C. subdepressus) there 218 was a significant effect of day on survival (Table 2; Figure 2). This slow reduction in the number 219 of larvae across all of the treatments is most likely the result of gradual larval mortality and 220 possibly larval loss during water changes. The rate of reduction was independent of temperature 221 treatment in all but E. lucunter where a significant interaction between day and temperature was 222 detected (Table 2). Differences between trails was due to different rates of initial survival 223 between the trials. 224 Post-hoc Tukey HSD tests for the 6 species for which we did not detect significant 225 interactions show that for all of the species except for C. subdepressus survival did not differ 226 between the 26°C and 28°C treatments (Figure 2). In 3 of these species survival at 30°C was 227 significantly lower than survival at 26°C (Table 1). In L. williamsi survival was significantly 228 reduced at 32°C but not at 30°C. In E. viridis survival was significantly reduced at 34°C 229 compared to 26-32°C. The significant interaction between day and temperature makes 230 interpretation of the results for E. lucunter slightly more complicated (Figure 2). However post-231 hoc tests conducted on each day show that survival at 34°C was significantly lower than survival 232 at 28°C on every day (Figure 2). Survival at 32°C was significantly lower than survival at 26°C 233 on Days 3 of the 5 days. In all of the species the significant reduction of survival at the higher 234 temperatures was already evident on Day 2 (Figure 1). In addition, in all of the species except for 235 Perricone and Collin, Page 9 of 30 E. lucunter, there was no indication of differential mortality rates across the temperature 236 treatments as would be evidenced by a significant interaction between treatment and day. 237 238 Growth in Experiment 1 – Some aspects of larval size was significantly affected by day, 239 temperature and the interaction between day and temperature for all of the species examined here 240 (Table 3). In general, the statistical effect of replicate beakers accounted for 5-30% of the 241 variance in the data. All species showed sustained growth in all 3 aspects of morphology during 242 the 6 days of the experiment, except for C. rosaceus in which growth more or less ceased after 243 Day 4, after which some reduction in size was evident (Figures 3 & 4). This species is a 244 facultative planktotroph which can metamorphose after 5-7 days (Emlet, 1986) and therefore a 245 reduction in arm length and larval body size was not unexpected. 246 The impact of rearing temperature was evident from visual inspection of the growth 247 trajectories (Figure 3 & 4). Post-hoc Tukey HSD tests comparing the size measurements from 248 each temperature treatment on day 6 showed how the treatments differ for each species (Table 249 1). In the 2 Echinometra species postoral arm length, stomach length and body length showed no 250 significant differences at 26°C, 28°C, 30°C and 32°C. In E. viridis larvae raised at 34°C were 251 significantly smaller in all 3 measures. So few larvae of E. lucunter survived at 34°C that it was 252 not possible to compare larval size on Day 6. But the few that could be measured on Days 2 and 253 5 were small and misshapen. Among the other species, L. williamsi had reduce body length and 254 stomach length at 32°C compared to 28-30°C. Eucidaris tribuloides, T. ventricosus, and C. 255 rosaceus had reduced arm length and body length at 32°C. Similar to the results for survival, the 256 irregular sea urchin C. subdepressus showed significantly longer bodies and stomachs at 26°C 257 than at 28 °C, while these lengths did not differ significantly at 30°C and 32°C (Table 1; Figure 3 258 & 4). 259 260 Experiment 2 – Embryos raised for 24 hours at ambient sea temperatures (28°C) prior to a 48-261 hour exposure to one of the experimental temperatures show higher survival after exposures of 262 the same duration at stressfully warm temperatures (Figure 5; Tables 4&5) than those that were 263 placed in the temperature treatments immediately after fertilization (i.e., Day 2 larvae). For all 5 264 species, there was a significant effect of temperature on survival, and for all of them survival was 265 significantly reduced at 32°C compared to 28°C and 30°C. In addition, survival of T. ventricosus 266 Perricone and Collin, Page 10 of 30 larvae had reduced survival at 30°C compared to 28°C. Larvae had greater survival at higher 267 temperatures relative to 28°C when exposed to heat stress later in development in this 268 experiment compared to Experiment 1. In 3 species some larvae survived in a warmer treatment 269 in Experiment 2 than they did in Experiment 1 (Figures 1 & 5). A few larvae of L. williamsi 270 survived at 34°C in Experiment 2, but no larvae survived at 34°C in Experiment 1. A similar 271 pattern was evident in both species of Echinometra. In Experiment 2 some larvae survived a 48-272 hour exposure to 36°C, while no larvae survived a 48-hour exposure to 36°C in Experiment 1. In 273 both Echinometra species, the relative survival at 34°C compared to 28°C was much higher in 274 Experiment 2 (~30%) than in Experiment 1 (~5%) (Table 5). Likewise, in T. ventricosus the 275 relative survival at 32°C compared to 28°C was much higher in Experiment 2 (~50%) than in 276 Experiment 1 (~5%) (Table 5). 277 278 Discussion 279 Larval WT and TSM 280 Calculations of warming tolerance and thermal safety margins following Deutsch et al. 281 (2008) show that warming tolerance and thermal safety margins for these tropical sea urchin 282 larvae are minimal (Table 6). We calculated warming tolerance as the difference between the 283 average habitat temperature using 29°C, the grand average of the temperatures at the 2 Bocas del 284 Toro reefs over 10 years reported in Collin & Chan (2016), and the temperature treatment at 285 which our larval cultures experienced >50% mortality relative to the temperature with the 286 highest survival. This is similar to a LT50 generated by acute warming experiments, which is 287 often used as an estimate of the critical temperature. We calculated the thermal safety margin, 288 the difference between the optimal temperature and the habitat temperature in two ways, using 289 estimates of optimal temperatures based on growth or on survival. Because temperature-specific 290 growth and temperature-specific survival were often statistically indistinguishable across several 291 experimental temperatures and because curve fitting approaches failed in many cases to estimate 292 an optimum (results not shown and Collin et al. 2018) we used the warmest temperature at which 293 (1) body length or (2) survival were statistically indistinguishable from the value in the treatment 294 with the highest values. When several temperatures resulted in the same values, we chose the 295 warmest temperature because performance curves are expected to be left skewed with a rapid 296 reduction in performance once the optimum has been exceeded. This approach gives estimates 297 Perricone and Collin, Page 11 of 30 of warming tolerances ranging from 5°C for the warm tolerant Echinometra viridis to 1°C for the 298 least tolerant C. subdepressus, and thermal safety margins ranging from 3°C for E. viridis to -299 3°C for C. subdepressus (Table 6). 300 These estimates need to be interpreted with these caveats in mind. Since early larval 301 development is relatively rapid and since it appears that the earliest few hours of development 302 are the most sensitive to thermal stress (see below and Sewell and Young, 1999; Kapsenberg and 303 Hofmann, 2014; Collin and Chan, 2016), average annual temperatures may not represent the 304 typical thermal experience of early development. This may be particularly true if reproductive 305 effort varies seasonally. Several of the species studied here reproduce throughout the year in the 306 San Blas Islands of Panama (Lessios, 1985), while E. viridis, L. williamsi and C. rosaceus reduce 307 reproductive effort during the coolest times of the year (Lessios, 1985). Likewise, the 2 308 Echinometra species have been shown to focus reproductive effort only during the two warmest 309 months of the year in Puerto Rico (Cameron et al., 1985). If anything, this suggests that 310 development may be taking place at temperatures above the annual average, in Panama, and 311 raises the possibility that our WT and TSMs may be underestimated. The most extreme result for 312 C. subdepressus, that it is already experiencing negative thermal safety margins in Bocas del 313 Toro, is unfortunately the least well-supported result in this study. We have found this species to 314 be particularly difficult to spawn successfully and repeated attempts resulted in only a single 315 successful rearing trial, from which the larvae had low overall survival at all temperatures 316 compared to the other species. This species may be difficult to spawn precisely because it is 317 already living under stressfully warm conditions in Bocas del Toro. However, the possibility 318 remains that the low thermal tolerance measured in this study could be due to low gamete quality 319 resulting from some other stressor that resulted in artificially low estimates of thermal tolerance 320 in this species. 321 Regardless of these caveats, it is clear that these tropical sea urchins have very small 322 warming tolerances and thermal safety margins for early development in Bocas del Toro. This is 323 likely to put them at risk from even moderate environmental warming. Their warming tolerances 324 are lower than any reported in a review of latitudinal patterns in thermal tolerances of insects 325 (Deutsch et al., 2008), but are similar to WTs reported for larval frogs from warm subtropical 326 environments (Duarte et al., 2012). Other comparable data from sea urchins is sparse and 327 variation in methods make them not strictly comparable, but the available data do suggest a 328 Perricone and Collin, Page 12 of 30 latitudinal trend similar to that observed in other groups (e.g., Deutsch et al. 2008; Comte and 329 Olden 2017). In the Antarctic Sterechinus neumayeri developmental stages have an WT of at 330 least 20°C for acute 2 h exposures (Kapsenberg and Hofmann, 2014), and larvae of the temperate 331 Strongylocentrotus purpuratus also show an acute WT of approximately 20°C along the west 332 coast of the USA (Hammond and Hofmann, 2010). In the subtropics, adult Arbacia stellata 333 studied in Baja California, Mexico had an estimated WT of 14.5-16.4°C but a much smaller 334 TSM of 3-4.3°C (Díaz et al., 2017). In the tropics, TSM for adult Diadema antillarum is 5.7°C 335 and for adult E. lucunter is 3.7°C in the Cayman Islands (Sherman, 2015), slightly higher than 336 those obtained from chronic thermal stressors experienced during development for our species in 337 Panama. 338 339 How Do Larvae and Adults Compare? 340 Despite limited interspecific variation in upper thermal tolerance, the rank order of larval 341 thermal tolerance is similar in some respects to the rank order of adult thermal tolerances. The 342 warmest acute (2 hour) temperature exposures survived by adult sea urchins breaks the species 343 into 3 groups. The two Echinometra species can survive at the highest temperatures (37°C), 344 followed by E. tribuloides and C. rosaceus which can survive up to 36°C, with L. williamsi, T. 345 ventricosus and C. subdepressus all surviving only to 35°C (Collin et al., 2018 Figures 3&4). 346 The rankings of the maximum temperatures survived are as follows: 347 Adult Survival: Ev, El > Et, Cr > Lw, Tv, Cs 348 Larval Survival: Ev, El > Et, Cr, Lw, Tv, Cs 349 Ranking adults with respect to righting performance gives a similar ranking to that based 350 on survival. Unfortunately, neither of the Clypeaster species can right, so this metric could not be 351 measure those species (see Collin et al., 2018 Figure 5). 352 Adult Righting Performance: Ev, El > Et > Tv > Lw 353 Larval Growth Performance: Ev, El > Tv, Et, Lw 354 The larval and adult rankings are similar in that the two Echinometra species clearly have higher 355 warming tolerance, and that the tolerances of the other species are all very similar to each other. 356 The fact that adult survival was determined to a precision of 1°C, while larval survival was 357 determined with a precision of 2°C probably explains why adult tolerance could be more 358 effectively ranked. 359 Perricone and Collin, Page 13 of 30 Comparisons of adult and larval WT and TSM show that larvae are more sensitive to 360 warming than adults (Table 6). In general, larval WTs are 3-5°C smaller than adult warming 361 tolerances and larval TSMs are 0-4°C smaller than adult thermal safety margins. However, it is 362 important to emphasize that these assays used different exposure times. Assays of adults were 363 conducted after a 2-hour exposure, while our differences in larval survival were assayed after a 364 minimum of a 2-day exposure. Although lethal thermal limit often decreases with extended 365 exposures, possibly biasing this comparison towards finding lower larval tolerances, two lines of 366 evidence suggest that this cannot account entirely for this result. First, a comparison of 2-hour 367 exposures of embryonic development, 4-day old larvae and adults for Lytechinus variegatus, 368 another sea urchin species from Bocas del Toro, showed that the LT50 for embryonic 369 development was 3°C lower than it was for adults and 2°C lower than it was for larvae (Collin 370 and Chan 2016), a result roughly similar to our comparisons of 2 hour adult and 2 day larval 371 exposures. Secondly, preliminary trials of adult thermal tolerance after 12-hour exposures gave 372 largely similar results to those of 2-hour exposures (Collin et al., 2018) for E. lucunter and E. 373 viridis, and very slightly lower (<1°C) estimates of LT50 for L. williamsi (unpublished data). 374 Despite these caveats, a major result is that distantly related genera and families of echinoids 375 with similar geographic ranges do not differ from each other by more than ~2-3°C in their acute 376 thermal tolerance as adults or in their growth under chronic stress as larvae, and that larvae or 377 early developmental stages do appear to be more sensitive to warming than are adults. 378 379 How different is the thermal optimum from the lethal thermal limit? 380 Studies of thermal tolerance often detect non-lethal reductions in performance at 381 temperatures just lower than those at which survival is impaired. This can include reductions in 382 ecologically relevant organismal function like sprint speed, or physiological function like cardiac 383 performance. For example, larvae of Pollicipes elegans barnacles show a reduction in swimming 384 activity at 5-10 °C below the lethal temperature (as measured by LT50), followed by a decrease in 385 oxygen consumption 0-7 °C below the lethal temperature (Walther et al., 2013). This is 386 explained by the general energy-limited tolerance to stress (Sokolova 2013) and the more widely 387 known but controversial special case, the oxygen and capacity limited thermal tolerance 388 (OCLTT) model which was developed primarily with data from arctic and temperate ectothermic 389 organisms (Pörtner and Knust, 2007; Pörtner, 2010; Jutfelt et al. 2018). These models posit that 390 Perricone and Collin, Page 14 of 30 the pattern of reduced performance followed by death as temperatures increase is due to a 391 reduction in aerobic scope due to rising temperatures and increased stress. The OCLTT posits 392 that this is due to increased metabolic rate which cannot be matched by the ability of the 393 organism to deliver oxygen to the tissues (Pörtner and Knust, 2007; Pörtner, 2010). This general 394 pattern fits a large body of empirical data showing reduced measures of ecological and 395 physiological performance at temperatures below the lethal temperature and is consistent with 396 the somewhat left-skewed thermal performance curve (Pörtner and Knust, 2007). However, 397 recent studies have found that cardiorespiratory capacity and aerobic scope can be maintained up 398 to the critical thermal limit in some tropical species (e.g., Ern et al., 2013; 2015; Fitzgibbon et 399 al., 2017). Even in these species where aerobic scope continues to increase until the critical limit 400 is reached, the optimal temperature for other aspects of performance often still occurs well below 401 that critical limit (Clark et al., 2013). For example, in Sagmariasus verreauxi metabolic rates and 402 aerobic scope increase through 25°C, and no mortality is evident at 25°C, yet the optimal growth 403 rate is around 21°C (Fitzgibbon et al., 2017). Our data, however do not fit this common pattern 404 of reduced performance at temperatures below the lethal temperature. The temperatures at which 405 statistically significant reductions in larval size are evident, are generally the same temperatures 406 at which statistically significant reductions in larval survival were detected (Table 1). In several 407 species survival was impaired at a lower temperature than at which we observed reduced growth. 408 The similarity between the temperature at which performance declines and the lethal temperature 409 suggest that WT could be used as an estimate of TSM. 410 411 Is the timing of exposure to elevated temperatures important? 412 Previous work with echinoid embryos suggest that the timing and duration of exposures 413 to elevated temperatures can impact estimates of upper thermal limits (Sewell and Young, 1999; 414 Kaspenberg and Hofmann, 2014; Collin and Chan, 2016). For two tropical species E. lucunter 415 and Lytechinus variegatus the acute thermal tolerance of fertilization and of larvae exceed the 416 tolerance of early developmental stages (cleavage-gastrulation). Likewise, the early embryos of 417 the Antarctic sea urchin Sterechinus neumayeri appear to be more susceptible to stress during 418 early embryological development (Kaspenberg and Hofmann, 2014). This suggests that the 419 pattern that most of the mortality in our rearing experiments occurred prior to the first survival 420 counts on Day 2 is due to a more pronounced impact of rearing temperature earlier in 421 Perricone and Collin, Page 15 of 30 development. Our follow-up experiment which maintained embryos at ambient (28°C) for the 422 first 24 hours followed by a 2-day exposure to elevated temperatures showed an increase in 423 tolerance of extreme temperatures. This provides further evidence that early developmental 424 stages may be one of the stages most susceptible to thermal stress and should be the focus of 425 future acute tolerance studies. It is important to note, however, that 2-hour acute exposures to 426 thermal stress show that early hatchlings (blastula and beyond) can survive short exposures of 427 temperatures 4-5°C warmer than their upper limit for normal development based on continuous 428 exposures during the first 2 days of development (Sewell and Young, 1999), suggesting that tests 429 using very short exposures may over-estimate the upper thermal tolerance limits by ecologically 430 meaningful amounts. 431 432 Conclusion 433 Results of this and previous studies concur that early developmental stages of echinoids 434 prior to the late 4-armed larval stage may be more susceptible to thermal stress than are later 435 stages. Larvae of 7 species of sea urchins representing 5 genera from 3 major clades which co-436 occur in the Caribbean show very similar thermal tolerances, with the 2 Echinometra species 437 being consistently more tolerant of warmer conditions than the other species, but little difference 438 between the 5 others. Comparisons with environmental conditions show that warming tolerance 439 and thermal safety margins of these larvae are similar to or smaller than those reported for 440 tropical terrestrial organisms. 441 442 Acknowledgements 443 We thank Sean Murphy, Adriana Rebolledo, and Valerie Goodwin and especially Emily 444 Smith for vital help with laboratory trials, and the staff of the Bocas del Toro Research Station 445 for logistic support. This work was supported by a Competitive Grant from the Smithsonian 446 Institution, and was performed with permission from the Autoridad de Recursos Acuáticos de 447 Panamá in 2015, and the Ministerio de Ambiente de Panamá in 2016. Part of this work was 448 completed by VP as part of a Masters in Marine Biology at the University of Bologna, Italy. 449 Two anonymous reviewers helped improve the manuscript. 450 451 Perricone and Collin, Page 16 of 30 Literature Cited 452 453 Bertram, D.F., and R. R. Strathmann. 1998. Effects of maternal and larval nutrition on growth 454 and form of planktotrophic larvae. Ecology. 79: 315–327. 455 456 Bonacci, S., Iacocca, A., Fossi, S., Lancini, L., Caruso, T., Corsi, I., and S. Focardi. 2007. 457 Biomonitoring aquatic environmental quality in a marine protected area: a biomarker 458 approach. AMBIO. 36: 308-315. 459 460 Byrne, M., Lamare, M., Winter, D., Dworjanyn, S. A., and S. Uthicke. 2013. The stunting 461 effect of a high CO2 ocean on calcification and development in sea urchin larvae, a synthesis 462 from the tropics to the poles. Phil. Trans. Roy. Soc. B. 368: 20120439. 463 464 Byrne, M. 2012. Global change ecotoxicology: identification of early life history bottlenecks in 465 marine invertebrates, variable species responses and variable experimental approaches. 466 Mar. Environ. Res. 76: 3-15. 467 468 Byrne, M., Ho, M., Selvakumaraswamy, P., Nguyen, H. D., Dworjanyn, S. A., and A. R. 469 Davis. 2009. Temperature, but not pH, compromises sea urchin fertilization and early 470 development under near-future climate change scenarios. Proc. Roy. Soc. B. 276: 1883-1888. 471 472 Byrne, M., Selvakumaraswamy, P., Ho, M. A., Woolsey, E., and H. D. Nguyen. 2011. Sea 473 urchin development in a global change hotspot, potential for southerly migration of 474 thermotolerant propagules. Deep Sea Research Part II: Topical Studies in Oceanography, 58: 475 712-719. 476 477 Cameron, R. A., Boidron-Metairon, I., and O. Monterrosa. 1985. Does the embryonic 478 response to temperature and salinity by four species of Caribbean sea urchins parallel the 479 reproductive synchrony? Proc. Fifth International Coral Reef Congr. 5: 273-278. 480 481 Clark, T. D., Sandblom, E., and F. Jutfelt. 2013. Aerobic scope measurements of fishes in an 482 era of climate change: respirometry, relevance and recommendations. J. Exp. Biol. 216: 2771-483 2782. 484 485 Collin, R., Rendina, F., V. Goodwin, and S. McCabe. 2018. Do tropical specialist sea urchins 486 have higher thermal tolerances and optimal temperatures than their more widely distributed 487 relatives? Mar. Ecol. Prog. Ser. 589: 157-170. 488 489 Collin, R., D’Croz, L., Gondola, P., and J. B. Del Rosario 2009. Climate and hydrological 490 factors affecting variation in chlorophyll concentration and water clarity in the Bahia Almirante, 491 Panama. Pp.323-334 in Proceedings of the Smithsonian Marine Science Symposium, M.A. Lang, 492 I.G.MacIntyre and K. Ruetzler eds. Smithsonian Institution Scholarly Press, Washington DC. 493 494 Collin, R. and K. Y. K. Chan. 2016. The sea urchin Lytechinus variegatus lives close to the 495 upper thermal limit for early development in a tropical lagoon. Ecol. Evol. 6: 5623-5634. 496 doi:10.1002/ece3.2317 497 Perricone and Collin, Page 17 of 30 Comte, L., and J. D. Olden. 2017. Evolutionary and environmental determinants of freshwater 498 fish thermal tolerance and plasticity. Global Change Biol. 23: 728-736. 499 500 Davidson, T. M., and B. M. Grupe. 2015. Habitat modification in tidepools by bioeroding sea 501 urchins and implications for fine‐scale community structure. Mar. Ecol. 36: 185-194. 502 503 Delorme, N. J., and M. A. Sewell. 2014. Temperature and salinity: two climate change stressors 504 affecting early development of the New Zealand sea urchin Evechinus chloroticus. Mar. 505 Biol. 161: 1999-2009. 506 507 Delorme, N. J., and M. A. Sewell. 2013. Temperature limits to early development of the New 508 Zealand sea urchin Evechinus chloroticus (Valenciennes, 1846). J. Therm. Biol. 38: 218-224. 509 510 Deutsch, C. A., Tewksbury, J. J., Huey, R. B., Sheldon, K. S., Ghalambor, C. K., Haak, D. 511 C., and P. R. Martin. 2008. Impacts of climate warming on terrestrial ectotherms across 512 latitude. Proc. Natl. Acad. Sci. 105: 6668-6672. 513 514 Díaz, F., Re, A. D., Galindo-Sanchez, C. E., Carpizo-Ituarte, E., Perez-Carrasco, L., 515 González, M., Sanchez G.A.L. and C. Rosas. 2017. Preferred temperature, critical thermal 516 maximum, and metabolic response of the black sea urchin Arbacia stellata (Blainville, 1825; 517 Gmelin, 1791). J. Shellfish Res. 36: 219-225. 518 519 Diederich, C. M., and J. A. Pechenik. 2013. Thermal tolerance of Crepidula fornicata 520 (Gastropoda) life history stages from intertidal and subtidal subpopulations. Mar. Ecol. Prog. 521 Ser. 486: 173-187 522 523 Duarte, H., Tejedo, M., Katzenberger, M., Marangoni, F., Baldo, D., Beltrán, J. F., ... and 524 A. Gonzalez‐Voyer. 2012. Can amphibians take the heat? Vulnerability to climate warming in 525 subtropical and temperate larval amphibian communities. Glob. Change Biol. 18: 412-421. 526 527 Emlet, R. B. 1986. Facultative planktotrophy in the tropical echinoid Clypeaster rosaceus 528 (Linnaeus) and a comparison with obligate planktotrophy in Clypeaster subdepressus (Gray) 529 (Clypeasteroida: Echinoidea). J. Exp. Mar. Biol. Ecol. 95: 183-202. 530 531 Ern, R., Phuong, N. T., Wang, T., and M. Bayley. 2013. Oxygen delivery does not limit 532 thermal tolerance in a tropical eurythermal crustacean. J. Exp. Biol. 217: 809-814. 533 534 Ern, R., Phuong, N. T., Madsen, P. T., Wang, T., and M. Bayley, M. 2015. Some like it hot: 535 thermal tolerance and oxygen supply capacity in two eurythermal crustaceans. Sci. 536 Reports, 5:10743. 537 538 Fitzgibbon, Q. P. and S. C. Battaglene. 2012. Effect of water temperature on the development 539 and energetics of early, mid and late-stage phyllosoma larvae of spiny lobster Sagmariasus 540 verreauxi. Aquaculture, 344: 153-160 541 542 Perricone and Collin, Page 18 of 30 Fitzgibbon, Q. P., Simon, C. J., Smith, G. G., Carter, C. G., and S. C. Battaglene. 2017. 543 Temperature dependent growth, feeding, nutritional condition and aerobic metabolism of 544 juvenile spiny lobster, Sagmariasus verreauxi. Comp. Biochem. Phys. A. 207: 13-20. 545 546 Fujisawa, H. 1995. Variation in embryonic temperature sensitivity among groups of the sea 547 urchin, Hemicentrotus pulcherrimus, which differ in their habitats. Zool. Sci. 12: 583-589. 548 549 Gianguzza, P., Visconti, G., Gianguzza, F., Vizzini, S., Sarà, G., and S. Dupont. 2014. 550 Temperature modulates the response of the thermophilous sea urchin Arbacia lixula early life 551 stages to CO2-driven acidification. Mar. Environ. Res. 93: 70-77. 552 553 Glass, G.V., Peckham, P.D. and J.R. Sanders. 1972. Consequences of failure to meet 554 assumptions underlying fixed effects analyses of variance and covariance. Rev. Educ. Res. 42: 555 237-288. 556 557 Hammond, L. M., and G. E. Hofmann. 2010. Thermal tolerance of Strongylocentrotus 558 purpuratus early life history stages: mortality, stress-induced gene expression and biogeographic 559 patterns. Mar. Biol. 157: 2677-2687. 560 561 Heck, K. L., and J. F. Valentine. 1995. Sea urchin herbivory: evidence for long-lasting effects 562 in subtropical seagrass meadows. J. Exp. Mar. Biol. Ecol. 189: 205-217. 563 564 Jutfelt, F., Norin, T., Ern, R., Overgaard, J., Wang, T., McKenzie, D. J., Lefevre, S., 565 Nilsson, G. E., Metcalfe, N.B., Hickey, A.J. and J. Brijs. 2018. Oxygen-and capacity-limited 566 thermal tolerance: blurring ecology and physiology. J. Exp. Biol. 221: jeb169615. 567 568 Kapsenberg, L. and G. E. Hofmann. 2014. Signals of resilience to ocean change: high thermal 569 tolerance of early stage Antarctic sea urchins (Sterechinus neumayeri) reared under present-day 570 and future pCO2 and temperature. Polar Biol. 37: 967-980. 571 572 Karelitz, S. E., Uthicke, S., Foo, S. A., Barker, M. F., Byrne, M., Pecorino, D., and M. 573 Lamare. 2017. Ocean acidification has little effect on developmental thermal windows of 574 echinoderms from Antarctica to the tropics. Glob. Change Biol. 23: 657-672. 575 576 Kaufmann, K. W. and R. C. Thompson. 2005. Water temperature variation and the 577 meteorological and hydrographic environment of Bocas del Toro, Panama. Caribbean J. Sci. 41: 578 392-413. 579 580 Keshavmurthy, S., Fontana, S., Mezaki, T., del Cano González, L., and C. A. Chen. 2014. 581 Doors are closing on early development in corals facing climate change. Sci. Rep. 4: 5633. 582 583 Kuempel, C. D., and A. H. Altieri. 2017. The emergent role of small-bodied herbivores in pre-584 empting phase shifts on degraded coral reefs. Sci. Rep. 7: 39670. 585 586 Lessios, H. A. 2016. The great Diadema antillarum die-off: 30 years later. Annu. Rev. Mar. 587 Sci. 8: 267-283. 588 Perricone and Collin, Page 19 of 30 589 Lessios, H. A. 1985. Annual reproductive periodicity in eight echinoid species on the Caribbean 590 coast of Panama Echinodermata, Proceedings of the Fifth International Echinoderm Conference. 591 24: 303-311. 592 593 Ling, S. D., Scheibling, R. E., Rassweiler, A., Johnson, C. R., Shears, N., Connell, S. D., ... 594 and S. Clemente. 2015. Global regime shift dynamics of catastrophic sea urchin 595 overgrazing. Phil. Trans. Roy. Soc. B. 370: 20130269. 596 597 McAlister, J. S., and A. L. Moran. 2013. Effects of variation in egg energy and exogenous food 598 on larval development in congeneric sea urchins. Mar. Ecol. Prog. Ser. 490: 155-167. 599 600 Miller, N.A., Paganini, A.W., and J. H. Stillman. 2013. Differential thermal tolerance and 601 energetic trajectories during ontogeny in porcelain crabs, genus Petrolisthes. J. Therm. 602 Biol. 38: 79-85. 603 604 Neal, B. P., Condit, C., Liu, G., Dos Santos, S., Kahru, M., Mitchell, B. G., and D. I. Kline. 605 2014. When depth is no refuge: cumulative thermal stress increases with depth in Bocas del 606 Toro, Panama. Coral Reefs. 33: 193-205. 607 608 Nguyen, K. D. T., Morley, S. A., Lai, C. H., Clark, M. S., Tan, K. S., Bates, A. E., and L. S. 609 Peck. 2011. Upper temperature limits of tropical marine ectotherms: global warming 610 implications. PLoS One. 6: e29340. 611 612 Pecorino, D., Lamare, M. D., Barker, M. F., and M. Byrne. 2013. How does embryonic and 613 larval thermal tolerance contribute to the distribution of the sea urchin Centrostephanus rodgersii 614 (Diadematidae) in New Zealand? J. Exp. Mar, Biol. Ecol. 445: 120-128. 615 616 Patterson, H. D., and R. Thompson. 1974. Maximum Likelihood Estimation of Components of 617 Variance. In Proceedings of the Eighth International Biometric Conference, 197–207. 618 Washington, DC: International Biometric Society. 619 620 Perkins, N. R., Hill, N. A., Foster, S. D., and N. S. Barrett. 2015. Altered niche of an 621 ecologically significant urchin species, Centrostephanus rodgersii, in its extended range revealed 622 using an Autonomous Underwater Vehicle. Estuar. Coast. Shelf S. 155: 56-65. 623 624 Pinsino, A., and V. Matranga. 2015. Sea urchin immune cells as sentinels of environmental 625 stress. Dev. Comp. Immunol. 49: 198-205. 626 627 Pörtner, H. O., and R. Knust. 2007. Climate change affects marine fishes through the oxygen 628 limitation of thermal tolerance. Science. 315: 95-97. 629 630 Pörtner, H. O. 2010. Oxygen-and capacity-limitation of thermal tolerance: a matrix for 631 integrating climate-related stressor effects in marine ecosystems. J. Exp. Biol. 213: 881-893. 632 633 Perricone and Collin, Page 20 of 30 Searle, S.R., Casella, G., and C.E. McCulloch. 1992. Variance Components. New York: John 634 Wiley & Sons. 635 636 Seemann, J., González, C. T., Carballo-Bolaños, R., Berry, K., Heiss, G. A., Struck, U., and 637 R. R. Leinfelder. 2014. Assessing the ecological effects of human impacts on coral reefs in 638 Bocas del Toro, Panama. Environ. Monit. Assess. 186: 1747-1763. 639 640 Schiffer, M., Harms, L., Lucassen, M., et al. 2014. Temperature tolerance of different larval 641 stages of the spider crab Hyas araneus exposed to elevated seawater pCO2. Front. Zool. 642 11: 87. 643 644 Schmider, E., Ziegler, M., Danay, E., Beyer, L., and M. Bühner. 2010. Is it really 645 robust?. Methodology. 6:147-151. 646 647 Schroeder, T. E. 1981. Development of a "primitive" sea urchin (Eucidaris tribuloides): 648 irregularities in the hyaline layer, micromeres, and primary mesenchyme. Biol. Bull. 161: 141-649 151. 650 651 Sewell, M. A., and C. M. Young. 1999. Temperature limits to fertilization and early 652 development in the tropical sea urchin Echinometra lucunter. J. Exp. Mar. Biol. Ecol. 236: 291-653 305. 654 655 Sewell, M. A., Cameron, M. J., and B.H McArdle. 2004. Developmental plasticity in larval 656 development in the echinometrid sea urchin Evechinus chloroticus with varying food ration. J. 657 Exp. Mar. Biol. Ecol. 309: 219-237. 658 659 Sherman, E. 2015. Can sea urchins beat the heat? Sea urchins, thermal tolerance and climate 660 change. PeerJ 3: e1006. 661 662 Sokolova, I.M. 2013. Energy-limited tolerance to stress as a conceptual framework to integrate 663 the effects of multiple stressors. Integr. Compar. Biol. 53: 597-608. 664 665 Storch, D., Fernández, M., Navarrete, S. A. et al. 2011. Thermal tolerance of larval stages of 666 the Chilean kelp crab Taliepus dentatus. Mar. Ecol. Prog. Ser. 429: 157-167 667 668 Strathmann, M. F. 1987. Reproduction and Development of Marine Invertebrates of the 669 Northern Pacific Coast: Data and Methods for the Study of Eggs, Embryos, and Larvae. 670 University of Washington Press. 671 672 Sunday, J. M., Bates, A. E., and N. K. Dulvy. 2011. Global analysis of thermal tolerance and 673 latitude in ectotherms. Proc. Roy. Soc. B. 278: 1823-1830. 674 675 Sunday, J. M., Bates, A. E., and N. K. Dulvy. 2012. Thermal tolerance and the global 676 redistribution of animals. Nat. Clim. Change. 2: 686-690. 677 678 Perricone and Collin, Page 21 of 30 Tangwancharoen, S., and R. S. Burton. 2014. Early life stages are not always the most 679 sensitive: heat stress responses in the copepod Tigriopus californicus. Mar. Ecol. Prog. Ser. 517: 680 75-83. 681 682 Valentine, J. F., and K. L. Heck. 1991. The role of sea urchin grazing in regulating subtropical 683 seagrass meadows: evidence from field manipulations in the northern Gulf of Mexico. J. Exp. 684 Mar. Biol. Ecol. 154: 215-230. 685 686 Vellutini, B. C., and A. E. Migotto. 2010. Embryonic, larval, and juvenile development of the 687 sea biscuit Clypeaster subdepressus (Echinodermata: Clypeasteroida). PloS One. 5: e9654. 688 689 Walther, K., Crickenberger, S., Marchant, S., Marko, P. and A. L. Moran. 2013. Thermal 690 tolerance of larvae of Pollicipes elegans, a marine species with an antitropical distribution. Mar. 691 Biol. 160: 2723-2732. 692 693 Wolcott, R., and C. G. Messing. 2005. A comparison of diets and water agitation methods for 694 larval culture of the edible sea urchin, Tripneustes ventricosus (Echinodermata: 695 Echinoidea). Bull. Mar. Sci. 77: 177-190. 696 697 Perricone and Collin, Page 22 of 30 Table 1: Summary of the results of Experiment 1 showing the temperatures at which larval 698 growth and survival are significantly reduced based on Tukey HSD Post-hoc tests from the 699 ANOVA analyses. 700 701 702 703 # Significant differences based on Tukey post-hoc tests, compared to 28°C. 704 * Significant differences based on Tukey post-hoc tests, compared to 28°C on Day 6 for all species except for C. 705 subdepressus where the largest size was observed at 26°C. 706 NA = not applicable; Stomach is not clearly visible. 707 708 709 Species 100% Lethal by Day 2 Reduced survival# Reduced PO arm length* Reduced stomach size* Reduced larval length* Echinometridae Echinometra viridis 36 °C 34°C 34°C 34°C 34°C Echinometra lucunter 36 °C 32, 34 °C 34°C 34°C 34°C Toxopneustidae Lytechinus williamsi 34, 36 °C 32 °C 26°C 26, 32°C 26, 32°C Tripneustes ventricosus 34, 36 °C 32 °C 32°C 26, 32°C 32°C Cidaroidea Eucidaris tribuloides 34, 36 °C 30, 32 °C 32°C none 32°C Clypeasteroida Clypeaster rosaceus 34, 36 °C 30, 32 °C 32°C NA 32°C Clypeaster subdepressus 34, 36 °C 32 °C none 28°C 28, 32°C Perricone and Collin, Page 23 of 30 Table 2: ANOVA tables from analysis of larval survival in Experiment 1. 6 out of 7 species had 710 multiple trials which were included in the model as a random effect. Survival was counted on 711 days 2-6. Bold highlights statistically significant results. 712 713 Factor df F Ratio p Echinometra lucunter R2 = 0.82; N = 117 Temperature 4 67.12 <0.0001 Day 4 10.43 <0.0001 Temperature X Day 16 2.36 0.0056 Trial [Random] -2LogLikelihood = 688.07 4.5% of total variance explained Echinometra viridis R2 = 0.80; N = 150 Temperature 4 51.65 <0.0001 Day 4 2.88 0.03 Temperature X Day 16 1.19 0.28 Trial [Random] -2LogLikelihood = 979.17 77.6% of total variance explained Lytechinus williamsi R2 = 0.73; N =120 Temperature 3 28.98 <0.0001 Day 4 36.43 <0.0001 Temperature X Day 12 1.53 0.13 Trial [Random] -2LogLikelihood = 109.08 21.1% of total variance explained Eucidaris tribuloides R2 = 0.63; N = 119 Temperature 3 16.64 <0.0001 Day 4 2.85 0.03 Temperature X Day 12 0.53 0.53 Trial [Random] -2LogLikelihood = 803.05 62.4% of total variance explained Tripneustes ventricosus R2 = 0.80; N = 117 Temperature 3 119.63 <0.0001 Day 4 2.12 0.08 Temperature X Day 12 0.60 0.83 Trial [Random] -2LogLikelihood = 750.26 16.0% of total variance explained Clypeaster subdepressus r2 = 0.54; N = 60 Temperature 3 12.36 <0.0001 Day 4 0.64 0.63 Temperature X Day 12 0.66 0.78 Clypeaster rosaceus Perricone and Collin, Page 24 of 30 r2 = 0.82; N = 120 Temperature 3 127.16 <0.0001 Day 4 5.61 <0.0004 Temperature X Day 12 1.42 0.17 Trial [Random] -2LogLikelihood = 752.46 20.1% of total variance explained 714 715 Perricone and Collin, Page 25 of 30 Table 3: Nested ANOVA results of 3 morphological measurements from Experiment 1. 716 Replicate beakers were treated as nested random effects and generally account for little of the 717 overall variance. Fixed factors with a significant effect are highlighted in bold. 718 719 Factor df F-Ratio arms P-value arms F-Ratio stomach P-value stomach F-Ratio body P-value body Echinometra lucunter N=1049 Temperature 3 4.85 0.0062 4.16 0.012 4.18 0.013 Day 4 130.83 <0.0001 70.65 <0.0001 133.13 <0.0001 Temperature X Day 12 1.56 0.15 2.30 0.027 2.09 0.047 Jar [Day, Temperature] Random -2 loglike. = 11929.22 Var. explained =15.51% -2 loglike. = 9624.92 Var. explained = 9.17% -2 loglike. = 10721.06 Var. explained = 11.36% Echinometra viridis 1 N=1296 Temperature 4 62.60 <0.0001 76.68 <0.0001 134.08 <0.0001 Day 4 26.73 <0.0001 17.40 <0.0001 18.18 <0.0001 Temperature X Day 16 2.09 0.026 1.18 0.32 0.47 0.47 Jar [Day, Temperature] Random Var. explained =35.24% -2 loglike. = 13662.57 Var. explained = 15.45% -2 loglike. = 10911.52 Var. explained = 15.39% -2 loglike. = 12643.03 Echinometra viridis 2 N=739 Temperature 4 265.83 <0.0001 140.35 <0.0001 206.90 <0.0001 Day 4 30.15 <0.0001 31.94 <0.0001 41.12 <0.0001 Temperature X Day 16 2.35 0.037 5.45 <0.0001 4.08 0.0008 Jar [Day, Temperature] Random Var. explained =4.47% -2 loglike. = 9207.25 Var. explained = 1.11% -2 loglike. = 5664.88 Var. explained = 4.13% -2 loglike. = 7212.49 Lytechinus williamsi N = 1116 Temperature 3 55.05 <0.0001 72.23 <0.0001 42.72 <0.0001 Day 4 216.98 <0.0001 202.46 <0.0001 281.60 <0.0001 Temperature X Day 12 6.33 <0.0001 4.02 0.0004 4.62 0.0001 Jar [Day, Temperature] Random -2 loglike. =12021.16 Var. explained = 5.90% -2 loglike. = 9202.96 Var. explained = 6.49% -2 loglike. = 10972.16 Var. explained = 8.88% Eucidaris tribuloides N = 1395 Temperature 3 27.03 <0.0001 24.09 <0.0001 19.59 <0.0001 Day 4 (3) 72.02 <0.0001 56.65 <0.0001 28.29 <0.0001 Temperature X Day 12 (9) 2.78 0.015 1.58 0.14 2.11 0.039 Jar [Day, Temperature] Random Var. explained = 18.45 -2 loglike. = 8254.93 Var. explained = 10.51% -2 loglike. = 7982.38 Var. explained = 17.21% -2 loglike. =9931.80 Tripneustes ventricosus* N= 758 Temperature 3 25.25 <0.0001 97.72 <0.0001 35.35 <0.0001 Perricone and Collin, Page 26 of 30 Day 3 208.66 <0.0001 64.30 <0.0001 45.80 <0.0001 Temperature X Day 9 11.44 0.0014 7.42 <0.0001 3.56 0.0031 Jar [Day, Temperature] Random -2 loglike. =6388.11 Var explained = 8.01% -2 loglike. =7517.52 Var explained =14.23% -2 loglike. = 7034.85 Var. explained =15.51% Clypeaster subdepressus N = 664 Temperature 3 16.19 <0.0001 14.05 <0.0001 17.15 <0.0001 Day 4 47.72 <0.0001 19.56 <0.0001 47.00 <0.0001 Temperature X Day 12 1.84 0.081 2.90 0.0060 2.06 0.049 Jar [Day, Temperature] Random --2 loglike. =7677.37 Var. explained =18.65% -2 loglike. = 3196.43 Var. explained =18.76% --2 loglike. = 5914.64 Var. explained = 18.03% Clypeaster rosaceus N= 970 Temperature 3 12.80 <0.0001 - - 18.09 <0.0001 Day 4 31.43 <0.0001 - - 41.01 <0.0001 Temperature X Day 12 4.77 <0.0001 - - 6.17 <0.0001 Jar [Day, Temperature] Random -2 logLike. = 9677.74 Var. explained =17.36% - - -2 logLike. = 10227.49 Var. explained =16.14% 720 * Only for Days 3-6, as insufficient larvae survived at 32°C on day 2 to calculate the interaction 721 effect. 722 723 Perricone and Collin, Page 27 of 30 Table 4: One-way ANOVA results from Experiment 2, showing the effect of a 48-hour 724 temperature exposure after the first 24 hours of development were spent at 28°C. 725 726 Species df F-Ratio p Echinometra viridis R2 = 0.97; N = 20 Temperature 4 105.13 <0.0001 Echinometra lucunter R2 = 0.97; N = 20 Temperature 4 127.07 <0.0001 Lytechinus williamsi R2 = 0. 94; N = 16 Temperature 3 65.85 <0.0001 Tripneustes ventricosus R2 = 0.89; N = 12 Temperature 2 37.39 <0.0001 Clypeaster rosaceus R2 = 0.96; N = 12 Temperature 2 110.02 <0.0001 727 728 729 Perricone and Collin, Page 28 of 30 730 Table 5: Comparisons of survival after a 2-day exposure to experimental temperatures in 731 Experiments 1 and 2 calculated as the ratio of survival at the experimental temperature relative to 732 survival at 28°C in the same experiment. Letters indicate Tukey HSD post-hoc tests for 733 significant differences in survival at the different experimental temperatures, where temperatures 734 linked with the same letter do not differ significantly from each other. Larvae in Experiment 2 735 were 3-days old but larvae from both experiments had experienced the experimental 736 texmperatures for 2 days. 737 738 Species 28°C 30°C 32°C 34°C Echinometra viridis Experiment 2 A 0.88 (A) 0.53 (B) 0.34 (C) Experiment 1 Day 2 A 0.90 (A) 0.84 (A) 0.08 (B) Echinometra lucunter Experiment 2 A 0.88 (A) 0.51 (B) 0.30 (C) Experiment 1 Day 2 A 0.89 (A) 0.46 (B) 0.03 (C) Lytechinus williamsi Experiment 2 A 0.83 (A) 0.48 (B) 0.07 (C) Experiment 1 Day 2 A 0.87 (AB) 0.71 (B) - Tripneustes ventricosus Experiment 2 A 0.77 (B) 0.46 (C) - Experiment 1 Day 2 A 0.89 (A) 0.06 (B) - Clypeaster rosaceus Experiment 2 A 0.93 (A) 0.28 (B) - Experiment 1 Day 2 A 0.80 (A) 0.16 (B) - 739 740 Perricone and Collin, Page 29 of 30 Table 6: Warming Tolerance (WT) and Thermal Safety Margin (TSM) for each species, using 741 habitat temperature of 29°C. Optimal temperature as the warmest temperature in Experiment 1 742 with the longest body length on day 6 or the warmest temperature with the highest survival. 743 744 Warming Tolerance (T50 – Thab) Thermal Safety Margin (Topt-Thab) Species Larval Adult* Larval Body Length Larval Survival Adult* Echinometra viridis 5°C 8.0°C 3°C 3°C 5°C Echinometra lucunter 3°C 7.5°C 3°C 1°C 7°C Lytechinus williamsi 3°C 6.0°C 1°C 1°C 1°C Eucidaris tribuloides 3°C 7.6°C 1°C 1°C 5°C Tripneustes ventricosus 3°C 5.9°C 1°C 1°C 3°C Clypeaster rosaceus 3°C 7.0°C 1°C -1°C - Clypeaster subdepressus 1°C 6.1°C 1°C -3°C - 745 *From Collin et al. 2018; Post-hoc Tukey test results indicated in Figure 4, indicating the 746 warmest temperature where righting was not significantly different from the optimal 747 temperature. 748 Perricone and Collin, Page 30 of 30 Figure 1: Bargraph of larval survival for the 7 species of sea urchins raised for 6 days over the 749 array of 5 experimental temperatures and pooled across the 2 trials. Error bars show standard 750 errors. 751 752 Figure 2: Bargraph of larval survival represented as the least square means of survival on each 753 day and at each temperature generated by the ANOVA. Letters above the bars show the results 754 of the Tuckey post-hoc test and significant differences are indicated when bars do not share a 755 letter. There was no significant interaction between day and temperature except for E. lucunter, 756 for which we show the Tuckey post-hoc test for the full factorial combination of temperatures 757 and days. 758 759 Figure 3: Larval growth trajectories of body length, stomach length and average postoral arms 760 for 4 species of sea urchins raised over the array of experimental temperatures. Error bars show 761 standard errors which are often too small to be visible. 762 763 Figure 4: Larval growth trajectories of body length, stomach length and average postoral arms 764 for 3 species of sea urchins raised over the array of experimental temperatures. Error bars show 765 standard errors which are often too small to be visible. The stomach is not clearly visible in C. 766 rosaceus and was not measured. 767 768 Figure 5: Bargraph showing the survival of 3-day-old larvae of 5 species of sea urchins raised 769 for 2 days at experimental temperatures after a day at 28°C in Experiment 2. Letters represent the 770 results of the Tukey’s HSD post-hoc test conducted separately for each species. Error bars show 771 standard errors. 772 773