Limitations of cross- and multigenerational plasticity for marine invertebrates faced with global climate change

2.1 Crustacea

Crustaceans are considered to be tolerant of acidification due to their ability to physiologically buffer the negative effects of hypercapnia and because their chitinous exoskeleton makes them less vulnerable to decreasing CaCO₃ saturation than other calcifying taxa (Wang, Jeong, Lee, & Lee, 2018; Whiteley, 2011). Several multigeneration studies have availed of the short generation time of copepods and amphipods where F1 offspring are readily isolated from brooding females to establish additional generations. These crustaceans have been a focus because they are immensely important in marine ecosystems, often representing the highest biomass and play critical roles in trophic energy transfer.

Data from cross- and multigeneration studies are available for 12 copepod species in response to acidification and/or warming (Table 1), allowing some comparisons to be made for the trait most reported on, hatching success (Figure 2). Studies that use extreme pCO₂ levels to emulate carbon capture and storage scenarios show the high tolerance of several copepods (pH < 7.1, ~8,000+ µatm pCO₂; see Table 1). For the meta-analysis, we only used data pH ≥ 7.3, to be closer to climate change scenarios (Figure 2).

Most studies involve calanoid and acartid copepods (Table 1). Cross generation studies of acidification effects on Acartia bifilosa and Acartia tonsa report decreased hatching success and offspring development at fairly moderate levels of acidification (Table 1). The nauplii were particularly sensitive with higher mortality (Cripps, Lindeque, & Flynn, 2014). However, very low pH (pH_NBS 7.3) did not have negative effects on hatching, development rate and offspring size for Acartia tsuensis (Kurihara & Ishimatsu, 2008). The only study that also included exposure of the male parent prior to mating found strong negative cross generation effects on hatching, indicating the importance to consider paternal effects (Cripps, Lindeque, & Flynn, 2015).

In a study of A. tonsa over 80 generations (3.5 years), beneficial phenotypic change in offspring in response to acidification was not observed (Langer et al., 2019). The negative legacy of acidification was seen in the reduced developmental rate and early hatching, but not on hatching success. After 80 generations, copepods reared in 800 µatm pCO₂ and fed algae cultured in the same conditions were smaller in length but had a bigger body mass, pointing to food as a stressor offset. Food offsetting of stress effects is seen in acidification and thermal stress studies of marine invertebrates (e.g. Bellworthy, Spangenberg, & Fine, 2019; Thomsen et al., 2010; Uthicke et al., 2016).

For Pseudocalanus acuspes, multigenerational exposure to acidification (at two levels pH_T 7.54 and 7.75) resulted in decreased fecundity in the F2 generation and reduced gene expression, including downregulation of RNA transcription genes (De Wit, Dupont &Thor, 2016; Thor & Dupont, 2015). For the F2 generation maintained at pH_T 7.54, there was some alleviation (~50%) of the decreased fecundity seen in progeny from ambient conditioned parents. This change was associated with a change in metabolism (Thor & Dupont, 2015).

For Calanus finmarchicus, the effect of acidification on development of the F1 generation was ameliorated at pH_T 7.3, but not at pH_T 7.15 (Pedersen et al., 2014). The latter pH level was deleterious across generations, elevating metabolism and reducing growth. There was no effect on hatching to the F2 at either pH level. With regard to hatching success, the threshold for deleterious effects of acidification in C. finmarchicus was estimated to be between 2,080 and 3,080 µatm pCO₂ (Pedersen et al., 2014). High tolerance to acidification (pH_NBS 6.95) in this species (Mayor, Matthews, Cook, Zuur, & Hay, 2007) is suggested to reflect its adaptation to low pH (pH 5.0) during diapause (see De Wit et al., 2016) and vertical migration through pH strata (see below). Hatching of C. glacialis embryos is also resilient to pH_NBS 7.6 (Table 1).

There are several studies of acidification across multiple generations of benthic copepods. Tisbe battagliai reared to F3 in acidification conditions exhibited negative effects on larval size and production and changes to the exoskeleton (Fitzer et al., 2012). Reduced growth and skeletal changes were associated with a reallocation of resources to maintain reproduction. For Tigriopus japonicus, pH 7.1 had no effect on hatching in a cross generation study (Kita, Kikkawa, Asai, & Ishimatsu, 2013) or on embryo survival, development and larval production in a multigeneration study (Li, Wang, & Wang, 2017).

For the amphipod Gammarus locusta, exposure to pH_T 7.7 enhanced female fecundity, but this was followed by negative effects in the F1 generation with reduced fecundity, reduced survival and altered behaviour (shorter mate guarding time; Borges et al., 2018). Reduced hatching success was observed in a cross generation study of G. locusta at pH 7.6 (Cardoso et al., 2018). A recent study where several stress markers were analysed (e.g. lipid peroxidation, DNA damage) showed strong negative effects in the F1 generation of G. locusta (Lopes et al., 2019). Warming (+4°C) also reduced the number of embryos, but those that did hatch developed faster. In a study of Sunamphitoe parmerong where brooding females were maintained in ambient and +4°C through progeny development and F1 rearing, the warm conditioned line had higher thermal tolerance (Campbell, 2018). A multigeneration study of this species showed that warming (+2–4°C) had no effect on F1 development, but increased F2 brood number with the magnitude of change depending on diet (Ledet, Byrne, & Poore, 2018). Finally, for the commercially important Tanner crab, Chionoecetes bairdi, cross generation exposure to pH_T 7.5 had negative effects on the larvae (Long, Swiney, & Foy, 2016).

One multigenerational study that investigated temperature as a single stressor revealed that the high intertidal benthic copepod, Tigriopus californicus, has little capacity to adapt to warming even after 10 generations (Kelly, Sanford, & Grosberg, 2012). Although this appears surprising considering the thermally dynamic habitat of this species, other studies suggest that intertidal invertebrates living in extreme habitats have low acclimatory capacities (Somero, 2010; Stillman, 2003). T. californicus lives in high intertidal pools, a habitat associated with low genetic diversity and this is suggested to contribute to its poor adaptive capacity (Kelly et al., 2012). In a follow-up study where warm and cool adapted populations of T. californicus were cross-bred, heat tolerance was observed in selected lines (to F5), but with a cost of reduced fecundity (Kelly, DeBiasse, Villela, Roberts, & Cecola, 2016).

The importance to consider habitat and vertical migration in interpreting the sensitivities of copepods to acidification is shown in a study of Arctic under ice conditions (to 200 m depth; Lewis, Brown, Edwards, Cooper, & Findlay, 2013). Calanus species are exposed to a wide pH range during their daily migration and were tolerant of low pH (pH_T 7.6). In contrast, surface dwelling (non-migratory) copepods were more sensitive. For Arctic copepods, sensitivity to acidification parallels the natural levels experienced. Similarly, along a latitudinal gradient of acidification due to upwelling along the coast of Chile, A. tonsa from lower and more variable pH areas were more tolerant of acidification than conspecifics living at higher pH (Vargas et al., 2017). These studies provide the link between environmental variability and tolerance of copepods to acidification (see also Thor et al., 2017), but the key selective forces associated with this variability (e.g. pH mean, range, minimum, maximum) are not known (see below, Section 5).

For copepods, the weight of evidence from eight cross and six multigeneration studies across 12 species indicates that acidification (≥pH 7.6) has little impact on hatching success (Table 1). Meta-analyses of hatching success revealed that female conditioning did not affect embryo hatching rates. The effect sizes overlapped zero (Figure 2a). In the meta-analyses of hatching data over multiple generations, acidification effects were more variable (Figure 2b). Negative effects were only seen for P. acuspes. Overall, the analyses show no significant effect of acidification indicating that copepods have considerable phenotypic plasticity with regard to pH level for the hatching trait (Figure 2b). For copepods, it will be important to not only quantify the response of the progeny from adults (males and females) conditioned in control treatments but also those from adults conditioned in acidification treatments to execute a fully crossed factorial experimental design to allow a better understanding of the impacts of stressors (see Supporting Information for experimental designs). This would allow a full estimation of whether conditioning reduces the impacts of low pH on copepods.

2.2 Annelida

Cross- and multigeneration studies of the Annelida with respect to climate change have largely involved one polychaete species, Ophryotrocha labronica using populations from the Mediterranean (Table 1). This cosmopolitan invasive species occurs in fouling communities (Simonini, Massamba-N'Siala, Grandi, & Prevedelli, 2009). With its short generation time (17 days), weekly production of egg masses and parental care, this species is a good model for multigeneration studies (Chakravarti et al., 2016).

A laboratory strain of O. labronica maintained for ~30 generations in fluctuating pH (pH_NBS 7.4–8.4) and then exposed to unvarying pH_NBS 7.7 for a further six generations showed initial (F1, F2) negative effects of acidification (Table 1). By F3–F7, however, fecundity levels were restored and offspring performance improved (Rodríguez-Romero, Jarrold, Massamba-N'Siala, Spicer, & Calosi, 2016). Thus, this strain restored its fitness in an acidification environment through phenotypic plasticity within two generations and maintained this fitness for the additional four generations investigated. There was also no difference in metabolic rate in the F7 generation (acidification vs. control), indicating physiological acclimation and perhaps some genetic selection over the 30 generations in fluctuating pH.

In a study of a different population of O. labronica, cross generation exposure to warming or acidification had negative effects on fecundity, egg volume, hatching success and juvenile development (Chakravarti et al., 2016). Some of these negative effects were ameliorated over subsequent generations in offspring reared in the same conditions (Gibbin, Chakravarti, et al., 2017; Gibbin, N'Siala, et al., 2017). Combined warming and acidification had a dampening effect on negative responses, indicating that they acted as opposing vectors of stress. Warming accelerated development through generations while acidification had a retarding effect. However, while the O. labronica developed faster at higher temperature, they reproduced at a smaller size, resulting in lower fecundity. This indicated a trade-off between reproducing earlier but with a lower output. In a recent study that investigated warming (+3°C) as a single stressor (Jarrold et al., 2019), negative effects observed in the cross-generation part of the study were reversed in the F2 generation. However, as trade-offs were evident in Gibbin, Chakravarti, et al. (2017) and Gibbin, N'Siala, et al. (2017), beyond F2 deleterious effects would be expected.

Over generations, the responses (positive or negative) of O. labronica to acidification in seven traits (adult, developmental, physiological) depended on number of generations. The initial benefits of parental conditioning to increased temperature were lost as warming became broadly deleterious. A consistent change or improvement over generations was not evident (Gibbin, Chakravarti, et al., 2017; Gibbin, N'Siala, et al., 2017). After five generations, reactive oxygen species production (ROS) was double that of controls indicating high stress levels, as the polychaetes coped with warming (+3°C). Most negative effects were attributed to increased temperature (Table 1). While the benefits of parental conditioning were evident in the early generations, this was lost over generations leading to a strong decline in fitness (Gibbin, Chakravarti, et al., 2017; Gibbin, N'Siala, et al., 2017).

We need more studies of polychaetes, especially groups with planktonic larvae such as serpulids that form calcareous tubes that are vulnerable to acidification (Lane, Mukherjee, Chan, & Thiyagarajan, 2013). In a cross-generation study of Spirobranchus triqueter, the larvae generated by adults conditioned at pH_T 7.6 and reared in this treatment were smaller and had lower survival than those from ambient parents (Díaz-Castaneda et al., 2019). At pH_T 7.6 and 7.8, the tubes of the F1 worms were smaller than those of controls. A study of Hydroides elegans, a locally abundant serpulid, produced an unexpected result showing that juvenile growth and calcification were faster if the acidified conditioned parent was a male, with the opposite result for maternal effects and no effects if both parents were conditioned acidification (Lane, Campanati, Dupont, & Thiyagarajan, 2015). For H. elegans, there were negative COE of acidification from larva to juvenile stages (Lane et al., 2013). Juveniles resulting from larvae reared at pH_NBS 7.7 and 7.9 were unable to make their tube. Serpulids are major biofouling species across the globe and the results for H. elegans and S. triqueter indicate that their dominance in fouling communities may change due to ocean acidification (Díaz-Castaneda et al., 2019). Reduction in serpulid density is evident at CO₂ vent sites (Kroeker, Micheli, Gambi, & Martz, 2011).

In a cross-generation study of the effects of temperature as a single stressor on the calcareous tube worm Galeolaria caespitosa where some adults were conditioned at cool (~−4°C), and others conditioned at warm (+2°C) temperature, offspring survival was higher when parent and offspring environmental temperatures matched (Chirgwin et al., 2018).

Multigeneration studies as undertaken for O. labronica are rare yet are needed to provide insights into the response of species to changing marine environments. These studies quantified many traits including cellular and physiological health markers (Table 1). The health markers (e.g. ROS) were particularly insightful with regard to how fitness tracked over generations, showing the importance to consider these traits, as well as morphological ones. Studies with O. labronica also incorporated reciprocal transfers between experimental and control environments and showed that some of the trait changes were reversible (indicating acclimation), while others were fixed (indicating adaptation; Gibbin, Chakravarti, et al., 2017; Gibbin, N'Siala, et al., 2017). Due to the mixed outcomes for progeny, these authors suggest that six generations are insufficient to be able to discern between plasticity and adaptive capacity.

2.3 Mollusca

All cross-generation climate change studies of molluscs have involved bivalves, including clams, scallops, mussels and oysters (Table 1). These bivalves have received emphasis due to their ecological and commercial importance. Most studies have involved adult conditioning.

Conditioning of wild type and selected families of the Sydney rock oyster (Saccostrea glomerata) in pH_NBS 7.9 during gametogenesis produced beneficial effects for offspring (Parker et al., 2012). The progeny of both oyster types were less sensitive to low pH than the larvae of ambient treatment adults, with the larvae of the selected families being more resilient (Figure 3). The larvae of acidification treatment oysters had a shorter development time and were larger. The F1 generation juveniles of the wild-type oysters were transferred to common garden field ambient conditions until the adult stage and were then similarly conditioned in pH_NBS 7.9. In the resulting F2 larvae, the positive effects of the original parental exposure were maintained, showing beneficial MGP (Parker, O'Connor, Raftos, et al., 2015). Moreover, offspring generated from the acidification treatment adults also performed better at ambient pH.

A multistressor study showed that while S. glomerata larvae produced by acidification conditioned adults, were, as expected, resilient to acidification as a single stressor, they were more sensitive when thermal stress, reduced salinity and low food conditions were combined (Parker, O'Connor, Byrne, et al., 2017). Adult conditioning was maladaptive when parental and offspring environments did not match. The authors suggest that this maladaptive response might not have occurred if the parents were also conditioned to the same multistressor environment.

In a recent study of Ostrea lurida, where juveniles generated from adults that were maintained in acidification treatments (pH 7.3) through gametogenesis (60 days) were outplanted at several sites, survival was higher in the acidification treatment progeny (Spencer et al., 2019). The magnitude of this effect depended on the parental source and out planting site.

There are several cross and multigeneration studies of mytilids. For Mytilus edulis, a study comparing the progeny of adults from the naturally low pH environment of the Western Baltic Sea and nearby high pH areas showed that the larvae of Baltic mussels were more tolerant to acidification, an indication of CGP to acidification in nature (Thomsen et al., 2017). In a 2 year, multigenerational experiment, families (individual male–female crosses) from the Baltic Sea mussels were reared in three pH levels (Table 1). One-third of these families were tolerant to low pH (pH_T 7.46) and were then reared to maturity in control and acidification conditions to generate the F1, and so on, to the F2. Beneficial effects were evident in the larger initial size of the larvae, their calcification performance and increased survival of the F1 and F2 generation (Figure 3c,d). However, over time, the beneficial effects of parental exposure to acidification with respect to progeny performance were not realized (Thomsen et al., 2017).

Interestingly, the larvae generated by M. edulis conditioned to 1,000 µatm pCO₂ over 6 months, and reared in these conditions, metamorphosed into juveniles with an altered shell minerology (Fitzer, Phoenix, Cusack, & Kamenos, 2014). The juveniles produced calcitic rather than aragonitic shells, the latter being more soluble in acidification conditions. If such calcification plasticity is a genetic trait of M. edulis, this may be an important adaptation to resist habitat acidification. For M. chilensis, parental conditioning to acidification resulted in larger larvae when reared at ambient pH and in low pH-warm treatments (Table 1).

For the invasive fouling mussel Musculista senhousia, cross generation parental exposure to acidification (pH 7.7) was broadly beneficial for all larval traits investigated (development rate, survival and metamorphosis; Table 1). These positive effects were linked to the larger eggs spawned by acidification conditioned females (Zhao et al., 2019). The positive effects of parental conditioning for progeny continued to the juvenile stage which had higher survival. The enhanced performance of the juveniles appeared to be due to a change in metabolism. The broad ranging positive effects are suggested to be associated with the high stress tolerance and natural phenotypic plasticity of M. senhousia (Zhao et al., 2019).

For the scallop Argopecten irradians and the clam Mercenaria mercenaria conditioning of adults at low pH (pH_T 7.4) did not benefit offspring (Griffith & Gobler, 2017). These adults produced larvae that were equally or more sensitive to low pH than larvae originating from ambient pH-treatment adults (Figure 3c,d). The larvae were smaller, had lower survival and were more vulnerable to other factors (low food, harmful algae).

For the Manila clam, Ruditapes philippinarum parental conditioning (pH 7.7) resulted in faster growth in the resulting juveniles and adults grown in the same conditions (Zhao, Schöne, Mertz-Kraus, & Yang, 2017; Zhao, Yang, et al., 2018). The juveniles switched to a less costly ion regulation mechanism and availed of a less costly source of carbon, facilitating calcification and growth. It was suggested that phenotypic plasticity in R. philippinarum was associated with its adaptation to fluctuating conditions in its intertidal habitat.

Carry-over effects of acidification have also been investigated in the embryo to larva and larva to juvenile transition in bivalves. For O. lurida, there were strong negative COE (Hettinger et al., 2012). Juveniles generated by larvae cultured in acidification and out planted into the natural environment did poorly (40% lower growth rate) compared with juveniles generated in ambient conditions. In contrast, for A. irradians and M. mercenaria, the COE from larva to juvenile were beneficial (Gobler & Talmage, 2013). For M. mercenaria, although 80% of the larvae died in acidification culture, the survivors had greater success as juveniles when transferred to natural conditions. These juveniles grew faster and had higher survival compared to those sourced from ambient cultures. A similar result was initially seen for A. irradians, but after 10 months, the juveniles were smaller than those generated by ambient treatment adults (Gobler & Talmage, 2013). For the Ostrea angasi, the low pH that surrounds brooded offspring during valve closure (to pH_NBS 7.46) is suggested to convey larval resilience to acidification (Cole et al., 2016), but this was not the case for O. chilensis (Chaparro, Montory, Segura, & Pechenik, 2009).

The results of eight cross- and two multigeneration acidification studies across seven bivalve species are mixed with respect to outcomes for offspring (Table 1). The weight of evidence indicates that depending on the pH level used, parental conditioning is beneficial for few species. Meta-analyses of the pH effect (See Supporting Information for experimental design) revealed that parents conditioned in acidification treatments produced larvae that were still significantly negatively affected by acidification with respect to survival and growth (Figure 3a,b). Overall, conditioning of parents did not remove the negative impact of low pH on offspring reared in the same conditions.

The effect of conditioning was assessed by comparing the response to low pH of offspring produced by conditioned and non-conditioned parents (Figure 3c,d; see Supporting Information for experimental design). These data from studies that used a fully orthogonal design are insightful with regard to the influence of conditioning on offspring performance. Parental exposure to low pH reduced the magnitude of the impact of acidification on offspring survival for M. edulis, but not for the other species (Figure 3c,d). Larval growth was rescued for M. edulis from the Western Baltic Sea at low pH (pH_T 7.46) and to some extent for S. glomerata (especially from the selected family line) at moderate acidification (pH_NBS 7.9; Figure 3d). Overall, conditioning of parents significantly reduced the negative impacts of acidification on larval growth, likely reflecting the subset of resilient survivors.

2.4 Echinodermata

For echinoderms, cross generation studies have involved eight echinoid species. While most of these studies have involved conditioning of mature adults, one study of Tripneustes gratilla reared animals from the juvenile stage in warming (+2°C) and acidification (pH_T 7.8) treatments in a fully crossed experimental design for the entirety of gonad/germ line development and gametogenesis to the mature adult (F0 parents) (Karelitz et al., 2019). These adults were used to generate the F1 generation. Progeny of the acidification and the acidification + warming parents were resilient to low pH, while those from the warming and the acidification + warming parents were resilient to increased temperature. The larvae of experimental treatment parents were better able to grow their skeleton in acidification/warming conditions than those of control treatment parents. However, negative effects were evident as the larvae from experimental treatment parents were overall smaller than control larvae (Karelitz et al., 2019). The eggs of the warm treatment urchins were smaller and this may have influenced this result.

For Strongylocentrotus droebachiensis, there was a difference in the outcome for progeny depending on the time that parents were exposed to acidification (pH_NBS 7.7). The progeny generated by parents after 4 months exhibited increased mortality, decreased settlement and low juvenile survival (Dupont, Dorey, Stumpp, Melzner, & Thorndyke, 2013). After 16 months, many of these negative effects were ameliorated. Parental fecundity was lower in the acidification treatment at 4 months, but not at 16 months, indicating that gametogenic timing influenced these results. After 16 months, the urchins would have gone through their annual gametogenic cycle and the gonads would have fully matured under experimental conditions, providing an opportunity for the gametes to be ‘imprinted’ to the progeny environment thereby promoting beneficial ‘anticipatory’ plasticity (positive CGP) for offspring (see Box 1).

Exposure of the Antarctic sea urchin Sterechinus neumayeri to a range of acidification and warming conditions also had different outcomes for offspring depending on parental exposure time (Suckling et al., 2015). The S. neumayeri were conditioned in control, warm (+1.5°C) and two acidification-warming treatments (pH_NBS 7.7/+1.8°C; 7.5/+2.1°C). With this design, it is not possible to tease out the impacts of warming and acidification as individual stressors. After 6 months, the effect of adult conditioning was negative, with reduced egg size and embryo hatching (Table 1). Decreased larval survival was seen in the warm-control pH treatment. After 17 months of adult conditioning, the outcomes were beneficial with the largest eggs produced by urchins maintained at pH_NBS 7.5/+2.1°C, although with high larval abnormality. There was no difference in larval size indicating that adult conditioning in warm-acidification conditions ameliorated the negative effects of acidification on growth and calcification. This result is likely due to the effect of +2.0°C countering the stunting effect of acidification on larval calcification in S. neumayeri (Byrne, Ho, et al., 2013). As expected, increased temperature resulted in faster development. With the longer conditioning time, the gonads of S. neumayeri would have fully matured in treatments and imprinted on future offspring environment, promoting beneficial CGP (Box 1).

The study by Suckling et al. (2014) where Psammechinus miliaris was conditioned in acidification (pH_NBS 7.7) for different durations (28–70 days) also showed that outcomes differed between incubation times (Table 1). There were positive effects for progeny after 42 and 70 days conditioning but not after 28 days. The marked reduction in larval survival across all adult conditioning times, however, indicated deleterious effects. Following the longer conditioning period, the surviving larvae were larger, indicating they were better able to calcify in low pH conditions, but it is not known how they would have performed at ambient pH as this treatment was not included (Suckling et al., 2014). The incubation times used would not have allowed the gonads of P. miliaris to mature fully in conditions with potential to generate negative effects on gamete quality (Box 1).

Following a 7-week conditioning of Echinometra mathaei in pH_NBS 7.7, the larvae produced did not differ from those of control parents grown in the same conditions (Uthicke, Soars, Foo, & Byrne, 2013). Adult conditioning did not ameliorate the negative effects of acidification on larval skeletogenesis and abnormality. As gametogenesis in E. mathaei takes months, the conditioning period used would have been too short to provide beneficial outcomes for progeny and may have reduced gamete quality (Box 1), as seen in poor spawning of males.

For Strongylocentrotus purpuratus from the California Current Upwelling System (CCUS), there is evidence of adaptation to local temperature and pH conditions (Kelly et al., 2013; Pespeni, Chan, et al., 2013). In a reciprocal breeding experiment using parents from high and low CO₂ sites, the impact of acidification on larvae was lower if the maternal parent originated from the high CO₂ site, indicating the presence of phenotypic plasticity in oogenesis (Kelly et al., 2013). In a recent study where adult S. purpuratus were conditioned to a range of acidification and temperature treatments to emulate the upwelling and non-upwelling conditions (Table 1), there were minimal to no effects of any treatment for progeny morphology (Wong, Kozal, Leach, Hoshijima, & Hofmann, 2019). This outcome reflects the resilience of the S. purpuratus from populations that routinely experience upwelling and appear to be genetically adapted low pH conditions (Pespeni, Chan, et al., 2013; Pespeni, Sanford, et al., 2013). The larvae of this species exhibit only a slight decrease in arm growth at pH 7.7 (Kelly et al., 2013; Yu, Matson, Martz, & Hofmann, 2011). In a global meta-analysis, S. purpuratus from the CCUS showed the highest tolerance to acidification, whereas conspecific larvae from other areas were more sensitive (Byrne, Lamare, et al., 2013). Similarly, for Loxechinus albus along the coast of Chile, larval growth in populations exposed to variable pH due to upwelling was less sensitive to acidification (Gaitán-Espitia et al., 2017). More variable pH is correlated with exposure to lower extreme pH, but we do not know which aspect of this variability makes S. purpuratus and L. albus larvae less sensitive to acidification (see below, Section 5).

Although the morphological responses of the progeny of S. purpuratus were muted, gene expression in gastrulae differed and when grouped according to parental treatment, the highest expression levels were found in embryos derived from upwelling treatment adults (Wong, Johnson, Kelly, & Hofmann, 2018). There were also differences in DNA methylation (Strader, Wong, Kozal, Leach, & Hofmann, 2019; see Section 3). These findings highlight the importance to consider mechanistic (cell and molecular) underpinnings in cross generation climate change studies.

For the juvenile offspring of Heliocidaris erythrogramma that had been conditioned to acidification and warming treatments in a fully crossed experimental design over the entire gametogenic cycle, only the highest temperature used (+6°C) influenced juvenile performance, causing increased respiration (Harianto, 2019). A study on the effects of warming as a single stressor on Strongylocentrotus intermedius across one generation found decreased hatching and larval size (Zhao, Zhang, et al., 2018).

Carry-over effects of acidification have been investigated in the larva to juvenile transition in sea urchins. For S. droebachiensis, larval exposure to acidification reduced juvenile survival (by 95%; Dupont et al., 2013). For Arbacia lixula, the larvae were robust to acidification surviving to metamorphosis, but the resulting juveniles were smaller than those generated in control conditions (Wangensteen, Dupont, Casties, Turon, & Palacín, 2013). Similarly, warming and acidification experienced during the larval stage of H. erythrogramma had negative COE for the juvenile stage with an increased abnormality and decreased calcification, although the negative effects of acidification were ameliorated by modest warming (+2°C, but not +4°C; Byrne et al., 2010, 2011). The impacts of these stressors differed with larvae being more sensitive to pH and temperature and juveniles more resilient to both stressors, although this also reflects the subset of larvae that survived and were able to metamorphose under these conditions (Foo, Dworjanyn, Poore, Harianto, & Byrne, 2016). In the single study of the juvenile to mature adult transition, T. gratilla was reared from the very small juvenile to adulthood in acidification and warming treatments. Those reared at pH_NBS 7.6 and +3°C gave rise to smaller adults with poor gonads, while those in pH_NBS 7.8 and +3°C were similar to controls and had well-developed gonads (Dworjanyn & Byrne, 2018).

For sea urchins, the weight of evidence (Table 1) indicates that the effects for progeny depend on acidification/warming level and the duration of adult conditioning. A global analysis indicated that (pCO₂ 1,000 µatm) approximates the tipping point for significant effects of acidification on reduced larval calcification (Byrne, Lamare, et al., 2013). The outcomes for sea urchins (Table 1) show the importance of gamete imprinting over the complete gametogenic cycle so that the gonads fully develop and mature under experimental conditions (Box 1).

2.5 Cnidaria

Given their economic, cultural and scientific importance, the response of reef building corals to climate change has been a prominent area of study, with a focus on mechanisms of rapid adaptation (Torda et al., 2017). Three studies have examined CGP in response to adult conditioning in pocilloporids (Table 1) that brood larvae of both sexual and asexual origin (Combosch & Vollmer, 2013), making the identification of CGP or COE difficult. Two studies of Pocillopora acuta/damicornis documented positive responses to climate change stressors in offspring when their parents were conditioned. Adult conditioning during brooding to combined low pH and increased temperature resulted in metabolic enhancement of larvae (Putnam & Gates, 2015). In Putnam, Ritson-Williams, Cruz, Davidson, and Gates (2019), adult conditioning to low pH had positive impacts on larval survival, settlement and growth, but this enhanced performance was no longer detectable by 6 months post settlement. Conditioning of adult brooding Stylophora pistillata to warming (+3–5°C) and reduced pH (pH_T 7.6) for 26 and 33 days resulted in reduction in symbiont cell density, chla/c₂ and photosynthetic performance in the F0 (Bellworthy, Menoud, Krueger, Meiboom, & Fine, 2019). This reduced photosynthetic performance was also present in the F1 generation, but no significant CGP/COE were observed in host physiology, recruitment and mortality due to parental treatment, which the authors interpreted as stress resistance.

The dearth of studies for reef building corals reflects their long generation time and challenging nature of these experiments. It is essential, however, to understand the eco-evolutionary responses in such an important and threatened group, especially as modelling efforts that include the potential for acclimatization and adaptation differ vastly from those that do not (Logan, Dunne, Eakin, & Donner, 2014).