A 120‐year record of resilience to environmental change in brachiopods

Abstract The inability of organisms to cope in changing environments poses a major threat to their survival. Rising carbon dioxide concentrations, recently exceeding 400 μatm, are rapidly warming and acidifying our oceans. Current understanding of organism responses to this environmental phenomenon is based mainly on relatively short‐ to medium‐term laboratory and field experiments, which cannot evaluate the potential for long‐term acclimation and adaptation, the processes identified as most important to confer resistance. Here, we present data from a novel approach that assesses responses over a centennial timescale showing remarkable resilience to change in a species predicted to be vulnerable. Utilising museum collections allows the assessment of how organisms have coped with past environmental change. It also provides a historical reference for future climate change responses. We evaluated a unique specimen collection of a single species of brachiopod (Calloria inconspicua) collected every decade from 1900 to 2014 from one sampling site. The majority of brachiopod shell characteristics remained unchanged over the past century. One response, however, appears to reinforce their shell by constructing narrower punctae (shell perforations) and laying down more shell. This study indicates one of the most calcium‐carbonate‐dependent species globally to be highly resilient to environmental change over the last 120 years and provides a new insight for how similar species might react and possibly adapt to future change.

for long-term acclimation and adaptation, the processes identified as most important to confer resistance. Here, we present data from a novel approach that assesses responses over a centennial timescale showing remarkable resilience to change in a species predicted to be vulnerable. Utilising museum collections allows the assessment of how organisms have coped with past environmental change. It also provides a historical reference for future climate change responses. We evaluated a unique specimen collection of a single species of brachiopod (Calloria inconspicua) collected every decade from 1900 to 2014 from one sampling site. The majority of brachiopod shell characteristics remained unchanged over the past century. One response, however, appears to reinforce their shell by constructing narrower punctae (shell perforations) and laying down more shell. This study indicates one of the most calcium-carbonate-dependent species globally to be highly resilient to environmental change over the last 120 years and provides a new insight for how similar species might react and possibly adapt to future change.

K E Y W O R D S
climate change, global warming, museum specimens, ocean acidification, shell characteristics 1 | INTRODUCTION Increased atmospheric carbon dioxide since the Industrial Revolution and subsequent rises in seawater temperature and decreases in pH have been well documented (Caldeira & Wickett, 2003, 2005IPCC, 2013;Orr et al., 2005). Biological implications of these changes are less well described and have largely been identified from organism responses in laboratory experiments lasting a few days to a few months (Riebesell & Gattuso, 2015). In recognition of the fundamental role played by seasonal phenotypic plasticity and genetic change across generations, long-term experiments which allow for acclimation (Cross, Peck, & Harper, 2015;Cross, Peck, Lamare, & Harper, 2016;Hazan, Wangensteen, & Fine, 2014;Suckling et al., 2014) and/or adaptation potential in organisms with short generation times (Andersson et al., 2015;Collins, Rost, & Rynearson, 2014) are now being made. Although information from long-term laboratory experiments is vital to reveal sensitivities of marine organisms, even they can still only predict responses from exposures of relatively short durations, of months or even a few years, to environmentally unrealistic conditions experiments, including in situ mesocosms (Nagelkerken & Munday, 2015) and CO 2 vent sites (Fabricius et al., 2011;Hall-Spencer et al., 2008;Uthicke et al., 2016), are another common approach which allows for the investigation of impacts on more long-term scales and also often include responses at the community level and the physical, chemical and biological variability in their natural environments that cannot be recreated in laboratory experiments. This method, however, has a lack of control of treatment conditions where organisms, for instance near vent sites, are locally exposed to significant short-term variation in pH levels as well as vents releasing other harmful substances (Gattuso et al., 2014). The newest methods in ocean acidification research are the free-ocean CO 2 enrichment (FOCE) systems, which are designed to assess the impact of lowered pH on biological communities in situ over weeks to months (Gattuso et al., 2014).
These systems include natural daily and seasonal pH changes as well as interspecific relationships and food webs (Barry et al., 2014;Kirkwood et al., 2015;Kline et al., 2012); however, the logistics of these systems are extremely challenging with replication a particular limitation due to cost and feasibility (Gattuso et al., 2014). A different and rarely used approach in the ocean acidification community is to evaluate changes over many decades in museum collections to determine how organisms have been affected by past environmental change.
This provides a historical record of the effects of changing environments on marine organisms (Hoeksema et al., 2011;Lister, 2011), which complements widely used laboratory and field experiments by allowing the assessment of possible long-term adaptation and presents a more holistic understanding of species responses.
Brachiopods are one of the best model groups of organisms to determine responses to climate change as they inhabit all oceans from intertidal to hadal depths (James et al., 1992;Peck, 2001) and are one of the most calcium-carbonate-dependent marine groups because their calcareous skeleton and other support structures usually make up >90%, and sometimes >95%, of their dry mass (Peck, 1993(Peck, , 2008, values which are amongst the highest reported for any marine invertebrate to date. They have also been locally important for shallow and deep-water communities for over 550 million years by providing a habitat for a diverse range of epifauna including encrusting sponges and algae (Barnes & Peck, 1996). A large-scale loss of brachiopods would therefore not only affect local communities, but could also have wider consequences which potentially could lead to changes or imbalances in benthic ecosystems (Peck, 2008).
Brachiopods are common in all the world's oceans. They are, however, only abundant in a few areas and in New Zealand many species are highly abundant at relatively accessible shallow depths (<30 metres), and at several different locations (Rudwick, 1962). This includes Calloria inconspicua in Paterson Inlet, Stewart Island (Figure 1;Doherty, 1979) where marine biological exploration began in the early 1900s because of its high biodiversity (Willan, 1981). As a result, there are excellent museum collections of this species deposited at regular intervals over the last century from this site, making C. inconspicua an ideal species to investigate variation in shell characteristics since the Industrial Revolution. Environmental change in New Zealand waters over the last two decades is also in line with global trends of a 0.1 pH unit decrease and 2°C warmer (Bates et al., 2014;Law et al., 2017). The aims of this study, therefore, were to determine whether past environmental change had affected shell morphology, structure, elemental composition and integrity in museum specimens of C. inconspicua collected from a single site Non-destructive morphometric measurements were made on specimens >16 mm in length as individuals become sexually mature at 14-16 mm length in C. inconspicua (Doherty, 1979 further destructive shell analysis depending on the available museum collection sample size. Details on the specific location and depth of each sample are given in Table S1. Long-term datasets of surface seawater temperatures were provided by Dr Doug Mackie and surface seawater pCO 2 by Dr Kim Currie.

| Shell characteristics
Eight key shell characteristics were analysed to determine any change over the past century. Shell morphology was assessed by measuring shell length, breadth and height of 389 individuals to the nearest 0.1 mm using Vernier calipers. A calcification index was calculated on 70 specimens (3-15 specimens per decade) that were donated for further shell analysis to quantify any variation in the efficiency of calcification. Calcification index was quantified as the amount of internal living space produced per unit of shell material deposited (Graus, 1974), therefore, calculated from the following equation: Calcification index = dry weight of the shell (g)/internal volume of the shell (cm 3 ) Dry weight was measured to 0.001 g on a Sartorius LA3200D weighing balance and the volume measurements were made according to Peck (1992).
To investigate any changes in shell structure, shell density, punctal (shell perforations) density and punctal width were measured.
Shell density was calculated for the same 70 individuals used in the calcification index analysis using the following equation: Shell density = dry weight of the shell (g)/shell volume (cm 3 ) Punctal densities (mm À2 ) were calculated from Scanning Electron Microscope (SEM; FEI QEMSCAN 650F) micrographs (1 mm 2 ) of the outer surfaces of 40 pedicle valves (2-5 valves per decade) from 10 different areas on each specimen ( Figure S1). Punctae are spatially distributed in terebratulids at regular intervals of~45 lm in a dominantly hexagonal, close packing pattern throughout each valve with minimal spatial variability (AE5 punctae per mm 2 ) (Williams, 1997). Punctal widths were measured from acetate peels (prepared according to Richardson, Crisp, and Runham (1979)) of cross sections of 40 brachial valves (2-5 valves per decade). Ten punctae were measured per specimen across the length of the individual to the nearest 0.1 mm on a Swift monocular petrological microscope with fitted micrometer. The percentage of shell that is punctae vs. shell matrix was then calculated from firstly calculating the mean area of a punctum for each year, where area is calculated as pr 2 and r is half the mean punctal width.
The mean area of a punctum (mm 2 ) and the mean punctal density of 152 mm À2 were then used to calculate the percentage area of shell occupied by punctae for each year by the following equation: Percentage area of shell occupied by punctae = (mean area of punctum x mean punctal density) 9 100 The percentage change in shell occupied by punctae compared to shell matrix was calculated by solving the linear regression equation relating shell area occupied by punctae for 1900 and 2014.
Solubility of skeletal structures is partly controlled by the elemental composition of the calcite crystal lattice (Harper, 2000;LaVigne et al., 2013). Any ion substitutions through changes in environmental variables, such as temperature (Chave, 1954), seawater composition (Ries, 2010) and seawater saturation state (Ries, 2011), could increase mineral solubility (LaVigne et al., 2013;Morse, Arvidson, & L€ uttge, 2007). Elemental composition of the shell was, therefore, analysed using a Cameca SX100 electron microprobe operated at 15 keV acceleration voltage, a 20 nA beam current and a 5 lm spot size. and Ba) were below detection limits and, therefore, excluded from further analysis. Matrix correction was performed following Pouchou and Pichoir (1984) (the PAP procedure). Standard analysis reproducibility was <1% for each element analysed. PAP corrected data were stoichiometrically calculated as carbonate (Reed, 1993).
To investigate any changes in shell integrity, a shell condition index and shell thickness were measured. Shell condition index was determined through measuring percentage areas of four types of shell condition (Table S2)

| Statistical analyses
Calcification index, shell density, punctal density, punctal width and shell thickness data were all normally distributed (Anderson-Darling test; p > .05). Parametric linear regression analyses were, therefore, performed on these characteristics to determine whether they changed over the last 120 years. The regression models for the calcification index and shell density datasets included year as the only factor as all measurements were made on different individuals. As punctal density, punctal width and shell thickness measurements were con-

| Shell characteristics
Non-destructive morphometric measurements of length, breadth and height in individuals >16 mm in length revealed that breadth increased almost proportionally to length with size (slope =0.90), but height increased more than length (slope=1.44) with size ( Figure 3; Table S3). Growth is, therefore, not isometric and shells get taller as they grow. The relationships of length to breadth and length to height did not change over time ( Figure 4b). This equated to a 1% decrease in shell occupied by punctae, which explains part of the 3.43% increase in shell density.
The majority of shell surfaces (>55% in each decade) of specimens throughout the 120-year study were intact with the protective periostracum layer undamaged and with the outer pitted layer present ( Figure S2). Only minimal shell dissolution (0%-13%) occurred in any specimen, and this did not vary throughout the time series (Figure 6.; Table 1).

| DISCUSSION
Long-term monitoring of environmental conditions in the southeast of New Zealand over the last 20-60 years (Figure 2) are in line with global trends of our oceans becoming 2°C warmer and 0.1 pH units more acidic since the Industrial Revolution (Caldeira & Wickett, 2003, 2005IPCC, 2013;Orr et al., 2005). The resilience of C. inconspicua to environmental change over the last century is clear from the data on various shell characteristics in this study. Six key aspects of the shells of this high calcium carbonate content species did not change since 1900 through to 2014 despite significant environmental shifts of a 0.6°C SST increase over the last 60 years and 35.7 latm increase in surface seawater pCO 2 over the last 20 years.
Morphometrics in this study, as expected, revealed that shells get taller as they grow, providing more space for larger gonads and larval brooding, particularly after individuals become sexually mature (14-16 mm length; Rickwood, 1977;Doherty, 1979). Variance in shell shape is a result of space constraints as larvae settle on or next to older individuals forming dense conspecific clusters. Morphometric relationships did not differ over the last 120 years, which is in contrast to other shell-bearing organisms in laboratory experiments where temperature and pCO 2 have been reported to impact phenotypic plasticity and alter shell morphology (Fitzer et al., 2015;Peyer, Hermanson, & Lee, 2010). Temperature had the greatest effect on shell morphology in the mussel Dreissena polymorpha in comparison to food quantity and water motion (Peyer et al., 2010).
Higher temperatures (~18 to 20°C) caused more rounded shells to be produced, whereas lower temperatures (~6 to 8°C) caused more laterally flattened shells. Increased pCO 2 conditions (750 latm and 1,000 latm) resulted in rounder and flatter Mytilus edulis shells which also had a thinner aragonite layer compared to ambient conditions (380 latm) (Fitzer et al., 2015). This new shell shape was explained as a compensatory mechanism to enhance protection from predators and changing environments due to the inability of this species to produce thicker shells under increased ocean acidity. The lack of a change in shell morphology of C. inconspicua over the last 120 years demonstrates the tolerance of this species to altered abiotic conditions. Shell density increased by 3.43% from 1900 to 2014, which cannot be explained by a change in shell morphology, elemental composition, shell thickness or the number of punctae as none of these shell characteristics varied over this period. This demonstrates the robust control of several aspects of shell production in C. inconspicua to changing environmental conditions. Punctal width, however, significantly decreased by 8.26% demonstrating that this species appears to have laid down more shell by constructing narrower punctae. This response may increase protection from their changing habitat or predation pressure by reinforcing the shell structure. Low shell repair frequencies, however, have been observed in C. inconspicua from Paterson Inlet [Harper, pers. obs.]. This response is, therefore, unlikely a result of any changes in their predator populations over the last 120 years. Producing narrower punctae is more likely a response to acidification by increasing calcification as seen in some species (Wood, Spicer, & Widdicombe, 2008). Producing narrower punctae, however, could have physiological implications for rhynchonelliform brachiopods as there is less space for extensions of soft tissue, called caeca, from the mantle into the punctae (Peck & Holmes, 1989a). As the function of punctae is still under debate (P erez-Huerta et al., 2009), the extent, if any, of the impact on the organism remains unknown.
The majority of the shell surfaces throughout the 120-year time period remained intact with the protective periostracum layer undamaged, the pitted layer present and only low levels of dissolution. The increase in seawater pCO 2 since 1900, therefore, did not cause extensive dissolution or impact the outer protective periostracum. This is in contrast to studies investigating the effects of end-century acidified conditions of a further increase to 1,000-1,300 ppm (IPCC, 2013) where dissolution is common amongst marine calcifiers including corals (Comeau, Carpenter, Lantz, & Edmunds, 2014), echinoderms (Dubois, 2014) and molluscs (Nienhuis, Palmer, & Harley, 2010). These rhynchonelliform brachiopods have therefore been unaffected in their abilities to construct and maintain their extensive skeletons by the change in ocean acidity and temperature over the last 120 years. Future conditions, however, could pose a threat to shell integrity and major questions around how far conditions need to change before significant impacts are evident remain to be answered.
Historical studies provide a different perspective to outstanding questions of how environmental change will impact marine life. (1,000-1,300 ppm) predicted for 2100 during an 8-week laboratory experiment (Anthony, Kline, Diaz-Pulido, Dove, & Hoegh-Guldberg, 2008).
The current study is the first to present data of a wide range of shell characteristics of a potentially vulnerable species from a single site every decade since 1900 to 2014, which covers the second half of the post-industrial revolution, when alterations in pCO 2 have been strongest. Six out of eight key shell characteristics measured in this unique collection were not affected by environmental change.
The only change observed was a decrease in punctal width which partially explained an observed increase in shell density, that may reinforce the structure of the shell. This indicates this highly calcium-carbonate-dependent species has been highly resilient to past changes in environmental conditions over the last 120 years and provides a novel insight into how similar species might react and adapt to future change.

ACKNOWLEDG EMENTS
The