Ocean acidification alters morphology of all otolith types in Clark’s anemonefish (Amphiprion clarkii)

Introduction

Since the advent of the industrial revolution, humankind has inadvertently relocated a significant volume of carbon to the troposphere, where it now resides as a greenhouse gas, warming the earth via radiative forcing (IPCC, 2013). Global warming, however, is not the sole consequence of surplus atmospheric CO₂: the surface ocean has absorbed approximately 30% of anthropogenic CO₂ emissions (Mikaloff Fletcher et al., 2006; Le Quéré et al., 2010), contributing to ocean acidification (Caldeira & Wickett, 2003). While this absorption is an important sink, abating the greenhouse effect (IPCC, 2013), it has consequences for marine ecosystems. Following diffusion, aqueous CO₂ impacts seawater chemistry by reducing pH and carbonate (CO${}_3^{2-}$) concentration (Doney et al., 2009). Both will impact the fitness of marine biota, with cascading effects up to the ecosystem level (Fabry et al., 2008; Queirós et al., 2015; Hoegh-Guldberg et al., 2017). From population abundances to community shifts, ocean acidification has the potential to alter the ecological landscape of the ocean (Gaylord et al., 2015).

The declining availability of free CO${}_3^{2-}$ is particularly worrisome due to its implications for marine calcifiers, which use calcium carbonate (CaCO₃) to form body structures including shells, teeth, and spines. Surface waters are normally supersaturated with CO${}_3^{2-}$, but as [CO${}_3^{2-}$] decreases, calcifiers may struggle to precipitate CaCO₃ (Gattuso & Buddemeier, 2000). Furthermore, if seawater is undersaturated with respect to calcium carbonate minerals (e.g., aragonite, Ω_Ar), existing structures may readily dissolve (Orr et al., 2005). A vast body of literature expounds ocean acidification’s anticipated effects on calcifier fitness in the future ocean, demonstrating variable degrees of severity (Hendriks, Duarte & Álvarez, 2010; Kroeker et al., 2013). Differential responses may depend on the specific biochemical pathways involved in calcification (Ries, Cohen & McCorkle, 2009), biological mechanisms for buffering pH changes in body fluids (Munday et al., 2011a), energetics limiting physiological acclimation (Seibel, Maas & Dierssen, 2012), or various ecological forces acting on an organism (Kroeker, Micheli & Gambi, 2012).

Teleostei is an extremely diverse infraclass of Actinopterygii representing the modern bony fishes, comprised of more than 30,000 species and dominating most aquatic habitats (Froese & Pauly, 2018). Teleosts are internal calcifiers, precipitating CaCO₃ in the intestinal lumen that aids water absorption and osmoregulation (Grosell, 2011), and precipitating otoliths in the inner ear that are critical for mechanoreception (Moyle & Cech, 2004). Heuer & Grosell (2014) reviewed numerous effects of acidification on marine teleosts, including respiratory acidosis leading to sustained elevation of blood plasma HCO${}_3^{-}$ (Esbaugh, Heuer & Grosell, 2012), cognitive disruption and behavioral changes linked to inhibited GABA_A neurotransmitter receptor function (Nilsson et al., 2012), mixed impacts on standard and maximum metabolic rates with implications for aerobic scope (Munday, Crawley & Nilsson, 2009), and increased otolith area (Munday et al., 2011b) and mass (Bignami, Sponaugle & Cowen, 2013; Bignami et al., 2013). As such, otoliths may be points of vulnerability for teleosts in the near-future ocean (Ishimatsu, Hayashi & Kikkawa, 2008; Munday et al., 2008; Heuer & Grosell, 2014).

Otoliths, or ear stones, are critical features located within the inner ear of teleost fishes, formed by precipitation of CaCO₃ around a protein-rich matrix and bathed in endolymph (Panella, 1971). CaCO₃ supersaturation is maintained in the endolymph by proton pumps in the epithelial cells adjacent to the site of crystallization, which maintain the pH gradient required for CO${}_3^{2-}$ - HCO${}_3^{-}$ balance (Ishimatsu, Hayashi & Kikkawa, 2008). Otoliths exist in three pairs (sagittae, lapilli, asterisci), with one from each pair contained within each otolithic end organ. When disturbed by fish movement or sound waves, otoliths trigger sensory maculae lining the interior wall of their chambers, converting the force into electrical impulses interpreted by the brain. Likewise, otoliths function as sensory organs for hearing and gravisense (Popper & Fay, 1993).

Researchers recognize the potential for ocean acidification to impact otolith growth in teleosts, especially during the sensitive larval phase, and many have demonstrated effects experimentally (Table 1). Contrary to the hypothesis that ocean acidification will inhibit otolith growth due to dwindling CO${}_3^{2-}$ availability (Ishimatsu, Hayashi & Kikkawa, 2008), elevated seawater pCO₂ stimulates growth of sagittae and/or lapilli in many taxa. This growth is attributed to elevated blood plasma [HCO${}_3^{-}$], retained to buffer acidosis and transported into the endolymph where it becomes substrate for CO${}_3^{2-}$ aggregation (Checkley et al., 2009; Munday et al., 2011b; Heuer & Grosell, 2014). Only one study (Mu et al., 2015) observed decreased otolith size in response to elevated pCO₂. Other studies (Franke & Clemmesen, 2011; Munday et al., 2011a; Simpson et al., 2011; Frommel et al., 2013; Perry et al., 2015; Cattano et al., 2017; Martino et al., 2017; Jarrold & Munday, 2018) observed no effects of pCO₂ on otolith morphology.

Evidence that acidification alters otolith size and shape has inspired hypotheses that this could interfere with otic mechanics, and thus impair sensory perception in teleosts (e.g., Munday et al., 2011b; Bignami et al., 2013; Bignami, Sponaugle & Cowen, 2014). Indeed, there is some evidence that asymmetry of otolith size, shape, and mass may impair auditory/vestibular function in some species with consequences for habitat detection and overall fitness (Lychakov & Rebane, 2005; Gagliano et al., 2008; Anken, Knie & Hilbig, 2017). Others have added that increased otolith size from ocean acidification could enhance auditory sensitivity to the benefit or detriment of the fish depending on life history (Bignami et al., 2013; Bignami, Sponaugle & Cowen, 2014; Réveillac et al., 2015).

While most available studies quantified simple morphometrics to analyze pCO₂ effects on otolith morphology, the most informative among them augmented morphometrics with other analyses, including complex shape analyses (e.g., Fourier analysis) (Munday et al., 2011a; Munday et al., 2011b; Simpson et al., 2011; Martino et al., 2017; Mirasole et al., 2017); mass, volume and density analyses (Bignami, Sponaugle & Cowen, 2013; Bignami et al., 2013); and compositional analyses (e.g., LA-ICPMS) (Munday et al., 2011b; Hurst et al., 2012; Martino et al., 2017; Mirasole et al., 2017; Coll-Lladó et al., 2018). Similarly, scanning electron microscopy can be used to screen for treatment effects on aspects of otolith morphology and composition that, although typically overlooked in simple morphometric analysis, may impact ear function. These may include: (i) lateral development, defined as the degree of convexity of an otolith’s lateral face; (ii) percent visible crystals, defined as an estimate of surface crystal density or grain, approximating surface roughness; (iii) crystal habit, here defined as any deviation in crystal shape from the predominant orthorhombic aragonite in sagittae and lapilli, or hexagonal vaterite in asterisci; and (iv) overall mineralogy, here defined as relative proportion of orthorhombic aragonite versus hexagonal vaterite visible on an otolith’s surface. The former two metrics estimate an otolith’s surface topography and texture, and the latter two estimate crystal features indicative of composition, density, and stability under environmental stress (Boulos et al., 2015). These metrics are intended as first-pass screening tools for efficiently identifying general trends in the data; should they yield compelling evidence of treatment differences, they could be followed with more rigorous methods to best quantify the variable (e.g., measuring otolith height directly or determining CaCO₃ polymorph composition with Raman spectroscopy (Coll-Lladó et al., 2018)).

In addition to standard morphometrics, the mineralogical metrics described above were used to investigate ocean acidification impacts on otolith morphology in larval Clark’s anemonefish, Amphiprion clarkii (Bennett, 1830). A. clarkii is a teleost reef fish belonging to Pomacentridae and inhabiting shallow reefs throughout the Indo-Pacific (Froese & Pauly, 2018). The species was chosen both as a novel taxon and to enable intragenus comparison with previous work (Munday et al., 2011b). Any impacts on its otolith morphology could have implications for teleost sensory perception and fitness in the future ocean.

Materials & Methods

Results

Otolith morphometrics and scoring

Otolith morphometrics and mineralogical metrics varied with pCO₂ treatment and according to otolith type and side (Fig. 1). Scoring for the crystal habit and mineralogy metrics never deviated from the norm for any otolith type in any treatment (i.e., sagittae and lapilli were consistently scored as predominantly aragonitic, exhibiting orthorhombic crystal habit; asterisci were assumed to be vateritic despite exhibiting little to no identifiable crystal habit).

Principal component analysis

For each otolith type and side, principal component analysis produced two components with eigenvalues greater than 1.0 (Table 3). From here on, the components are referred to as rotated components—RC1 or RC2—to reflect that their loadings were scaled with varimax rotation. Left Sagittae (LS): As pCO₂ increased, left sagittae were rendered overall larger and wider in circumference, with more pronounced lateral faces and rougher surface textures owed to greater visible crystal density (Tables 3 and 4; Fig. 2A). Right Sagittae (RS): As pCO₂ increased, right sagittae responded according to the same metrics as left sagittae, albeit with slightly stronger responses of area/SL and perimeter/SL (Tables 3 and 4; Fig. 2B). Whereas left sagittae responses were best represented as a linear model, right sagittae responses were best represented as a curvilinear (quadratic) model, with responses leveling out between the 1,600 µatm pCO₂/pH 7.60 and 3,000 µatm pCO₂/pH 7.30 treatments. Left lapilli (LL): As pCO₂ increased, left lapilli were rendered larger and wider in circumference (Tables 3 and 4; Fig. 2C). Right Lapilli (RL): In the most extreme pCO₂ treatment only, right lapilli were rendered rougher with more pronounced lateral faces despite remaining approximately the same size (Tables 3 and 4; Fig. 2D). Whereas left lapilli responses were best represented as a curvilinear (quadratic) model, with responses leveling out between the 1,600 µatm pCO₂/pH 7.60 and 3,000 µatm pCO₂/pH 7.30 treatments, right lapilli responses were best represented as a linear model. Left Asterisci (LA): In the most extreme pCO₂ treatment only, left asterisci were rendered increasingly elliptical (rather than circular), rougher, and wider in circumference (Tables 3 and 4; Fig. 2E). This was the only instance of 2-dimensional otolith shape change observed in response to treatment, as well as the only otolith metric that decreased rather than increased with increasing pCO₂. Right Asterisci (RA): Right asterisci were not observed to respond to increasing pCO₂ (Fig. 2F). This was the only otolith type/side that exhibited no response to treatment.

Table 3:

Component variances and loadings.

Loadings corresponding to various Amphiprion clarkii otolith morphological parameters, the variance of which composes the rotated components in Fig. 2 and other components excluded from the analysis. Also included are the variances associated with each component and the total variance associated with components.

Otolith	Component	Variance (%)	Area/SL	Perimeter/SL	Circularity	Lateral development	Percent visible crystals
Left Sagittae	RC1*	47	0.62*	0.37	0.05	0.97*	0.94*
	RC2	37	0.61	0.85	−0.86	0.08	0.17
	Total	84
Right Sagittae	RC1*	59	0.77*	0.72*	0.00	0.95*	0.96*
	RC2	29	0.39	0.61	−0.94	0.06	−0.07
	Total	88
Left Lapilli	RC1	46	−0.12	−0.30	0.67	0.90	0.96
	RC2*	36	0.89*	0.90*	−0.32	−0.29	−0.03
	Total	82
Right Lapilli	RC1	44	0.88	0.99	−0.65	−0.08	−0.01
	RC2*	34	0.02	0.06	0.31	0.88*	0.92*
	Total	78
Left Asterisci	RC1	44	0.81	0.81	−0.15	−0.75	−0.54
	RC2*	34	−0.11	0.53*	−0.92*	0.04	0.73*
	Total	78
Right Asterisci	RC1	50	0.89	0.98	−0.73	−0.06	0.48
	RC2	24	−0.16	0.08	−0.31	0.91	0.50
	Total	74

DOI: 10.7717/peerj.6152/table-3

Notes:

*signify which components are related to pCO₂ (p < 0.05) and which variables are strongly associated with each of those components (r ≥ 0.50).

Table 4:

Regression statistics.

Regression models are listed by row, including all models in the manuscript for which pCO₂ predicted the response variable (p ≤ 0.05), and truncated models with data from the 3,000 atm pCO₂/pH 7.30 treatment excluded (designated by subscript T). For the otolith morphological models, model name abbreviations are identical to those elsewhere in the manuscript (e.g., LS for Left Sagittae). FSC stands for Fish Settlement Competency (at 10 dph). FSL stands for Fish Standard Length. Component names (Comp.) are listed for otolith morphological models only. Statistics including degrees of freedom (DF), F-statistics (F), p-values (p), line equations, line slopes multiplied by 100 (b₁*100), 95% confidence intervals for slopes multiplied by 100 (CI*100), and R²s are listed. Only DF, F, and p are listed for truncated models, as the full models are considered to be more informative except for determining whether the 3,000 µatm pCO₂/pH 7.30 treatment disproportionately weighted them. Line slopes and confidence intervals should be read as “[response variable] increased by [b₁*100] [units] for every 100 µatm increase in pCO₂ (95% CI: [CI*100])”. Line slopes and confidence intervals were excluded for polynomial (quadratic) models, as they are not as easily interpreted.

Model	Comp.	DF	F	p	Equation	b1*100	CI*100	R2
LS	RC1	1,10	11.98	0.0061	y = (7.28E−4)x − 0.98	0.07	0.03–0.12	0.50
LST	RC1	1,7	10.67	0.0137	–	–	–	–
RS	RC1	2,9	20.56	0.0004	y = (2.51E−3)x − (5.08E−7)x2 − 1.98)	–	–	0.78
RST	RC1	2,6	14.64	0.0049	–	–	–	–
LL	RC2	2,9	10.47	0.0045	y = (3.48E−3)x − (8.86E − 7)x2 − 2.25)	–	–	0.63
LLT	RC2	2,6	10.91	0.0100	–	–	–	–
RL	RC2	1,10	8.21	0.0168	y = (6.62E−4)x − 0.89	0.07	0.01–0.12	0.40
RLT	RC2	1,7	0.11	0.7489	–	–	–	–
LA	RC2	1,9	5.61	0.0420	y = (6.64E−4)x − 0.80	0.07	0.00–0.13	0.32
LAT	RC2	1,7	1.04	0.3423	–	–	–	–
FSC	–	1,10	4.83a	0.0279	logit (π) = (−3.81E−4)x + 2.31	0.04	0.00–0.07	0.38
FSCT	–	1,7	7.26a	0.0071	–	–	–	–
FSL	–	1,10	17.77	0.0018	y = (−1.31E−4)x + 6.85	0.01	0.00–0.02	0.60
FSLT	–	1,7	2.43	0.1629	–	–	–	–

DOI: 10.7717/peerj.6152/table-4

Notes:

aFor FSC and FSC_T, which are binomial logistic regression models, the values listed as F are actually χ²-statistics.

Mortality, settlement competency, and somatic growth:

Fish mortality, competency to settle at 10 dph, and standard length varied somewhat with pCO₂ treatment (Fig. 3). Despite high fish mortality throughout the experimental trial, mortality was not associated with pCO₂ (Table 4, Fig. 4A). As pCO₂ increased, fewer fish were competent to settle at 10 dph (Table 4, Fig. 4B). In the most extreme pCO₂ treatment only, fish exhibited diminished somatic growth relative to those at lower concentrations (Fig. 4C).

Discussion

Otoliths exhibited diverse responses to treatment according to type and side. In response to increasing seawater pCO₂, all three otolith types exhibited increasing perimeter and percent visible crystals, sagittae and lapilli exhibited increasing area and lateral development, and asterisci exhibited differences in shape. While the sagittae changed according to the same metrics regardless of side, the lapilli and asterisci changed according to different metrics depending on side. These differences reveal important things about the nature of the metrics under investigation; for example, while both sagittae responded to treatment by growing larger with more pronounced lateral faces, these effects were segregated according to side in the lapilli; this suggests that otolith area and lateral development are uncoupled rather than being two immutably conjoined metrics of growth. As such, it is often informative to investigate each otolith independently rather than investigating one type or pooling by type without regard to side. Among the 24 studies reviewed here that analyzed ocean acidification impacts on otolith morphology, five investigated lapilli (Table 1), none investigated asterisci, and eight segregated otoliths by side during morphometric analysis (at least six of which pooled them after observing no evidence of asymmetry) (Franke & Clemmesen, 2011; Munday et al., 2011a; Munday et al., 2011b; Maneja et al., 2013; Bignami, Sponaugle & Cowen, 2014; Mu et al., 2015; Perry et al., 2015; Réveillac et al., 2015; Martins, 2017; Jarrold & Munday, 2018). However, the results suggest that responses of one or two otoliths cannot necessarily be extrapolated to the rest of the otolith system.

Some regression models were disproportionately weighted by the 3,000 µatm pCO₂/pH 7.30 treatment, as evidenced by the p-value change between the full model and truncated model with data from that treatment excluded. Although the p-values for all otolith morphological models increased with truncation (Table 4, attributable in part to a loss of power from fewer degrees of freedom), the p-values for the right lapilli and left asterisci models rose >0.05, indicating there is little to no evidence of pCO₂ impacts on those otoliths except in conditions of aragonite undersaturation. For the lapilli, this exercise indicates an asymmetry of resilience: left lapilli are impacted by lower pCO₂ concentrations, whereas right lapilli are not. Nevertheless, this still results in morphological asymmetry of lapilli at lower concentrations. For the asterisci, it indicates mutual resilience at lower concentrations, and morphological asymmetry at the highest concentration. Scenarios in which these relationships are ecologically relevant could include isolated coastal systems with poor mixing under additional pressure from eutrophication (Wallace et al., 2014; Fabricius, 2005) or volcanism (Vogel et al., 2015). However, lower concentrations are here considered more typical of the future ocean.

Researchers previously examined otolith development in teleost larvae reared under acidified conditions, and despite differences in methodology and model species, it is possible to draw comparisons. Notably, Munday et al. (2011b)’s study species (Amphiprion percula) enables intragenus comparison with A. clarkii. Here, the results are consistent with those of Munday et al. (2011b) and several others (Table 1) in that sagittae grew at elevated seawater pCO₂. However, Munday et al. (2011b) observed growth in left sagittae only, whereas A. clarkii exhibited growth in both sagittae. The results are further consistent with six of those studies (Table 1) in that lapilli also grew at elevated pCO₂ (albeit in the left ear only). Regarding otolith shape: the results are consistent with five studies (Table 1) in that otolith shape changed at elevated pCO₂, albeit in left asterisci only, and only in conditions of aragonite undersaturation. A caveat: whereas most other studies reared fish from two or more genotypes (e.g., Munday et al., 2011a; Munday et al., 2011b; Bignami, Sponaugle & Cowen, 2013; Bignami et al., 2013; Bignami, Sponaugle & Cowen, 2014), the present study used one clutch of eggs produced by one broodstock pair. While this eliminated lineage as a potentially confounding variable in the analysis, care should be taken when comparing the results of this study with others or extrapolating to larger populations.

Some of the observed effects of seawater pCO₂ on otolith growth in A. clarkii, including increasing area, perimeter, and lateral development, may be consequences of acid-base regulation triggered by respiratory acidosis. Fishes normalize internal pH disturbances by metabolic adjustment: blood plasma HCO${}_3^{-}$ is absorbed/retained and H⁺ excreted by modulating rates of transport across the gill epithelium. However, extracellular pCO₂ and HCO${}_3^{-}$ remain elevated following pH adjustment, and excess HCO${}_3^{-}$ is imported to the endolymph where it becomes substrate for CO${}_3^{2-}$ aggregation, enhancing net otolith calcification (Checkley et al., 2009; Munday et al., 2011b; Heuer & Grosell, 2014).

Since otoliths are critical components of the ears and vestibular organs (Fekete, 2003; Moyle & Cech, 2004), ocean acidification-driven changes to otolith development may challenge sensory perception in A. clarkii and other teleosts (e.g., Munday et al., 2011b; Bignami et al., 2013; Bignami, Sponaugle & Cowen, 2014). Indeed, there is some empirical evidence outside the context of ocean acidification that marine fish exhibiting abnormal otolith morphology or asymmetry suffer diminished sensory ability. In terms of hearing, fish with larger, vateritic, and/or otherwise asymmetrical sagittae exhibited reduced sensitivity or deafness (Oxman et al., 2007; Gagliano et al., 2008; Browning et al., 2012); some presumably because vaterite is less dense than aragonite, thereby reducing otolith displacement amplitude and effectiveness in the inner ear (Bignami et al., 2013; Reimer et al., 2016). Although there remains no evidence of ocean acidification-induced vaterite replacement in otoliths, including in A. clarkii, there is some evidence for calcite replacement in sagittae and lapilli at elevated pCO₂ (Coll-Lladó et al., 2018); calcite is similarly less dense than aragonite (Nakamura Filho et al., 2014). In terms of kinesthesia, some fish with abnormal and/or asymmetric sagittae/lapilli exhibited kinetoses (Söllner, 2003; Anken, Knie & Hilbig, 2017); however, other studies observed pCO₂ impacts on otolith morphology without observing impacts on behavior (Bignami, Sponaugle & Cowen, 2013; Bignami, Sponaugle & Cowen, 2014; Shen et al., 2016). It is possible the pCO₂ impacts on otolith morphology and asymmetry observed here could impair A. clarkii hearing and kinesthesia, but no sensory or behavioral assays were conducted, so hypotheses remain speculative.

In addition to corroborating reports of otolith growth along the x and y axes (i.e., increasing area and perimeter) in young teleosts in response to increasing seawater pCO₂, there was evidence for pCO₂-induced otolith growth along the z-axis (i.e., upward growth from the lateral face) in A. clarkii. Lateral development appears most conspicuous in sagittae, and linked to treatment in sagittae and right lapilli, although it was observable in asterisci as well. While lateral development occurs on the lateral face, which does not directly interact with maculae, it is possible this CaCO₃ aggregation will increase otolith mass at a magnitude greater than that which is evident from increased 2-dimensional area and perimeter. Thus, sagittae exhibiting advanced lateral development may have a wider displacement amplitude independent of area and perimeter, enhancing auditory sensitivity (Bignami et al., 2013); however, displacement amplitude was not measured here. This hypothesis is independent of otolith composition, for which there was no evidence of having changed, but which undermined auditory sensitivity in some studies (Oxman et al., 2007; Browning et al., 2012; Reimer et al., 2016). Also, since lateral development appears to occur on only one face of the otolith (though the medial face was not investigated here, all otoliths were imaged convex-side up, which was invariably the lateral face), its center of mass likely changed as well, with unknown consequences for otic mechanics.

Some of the otoliths appear visibly smooth on the surface, while others appear rougher due to the exposure of aragonite table edges and similar crystal activity. Estimating percent visible crystals is akin to estimating otolith surface roughness. The observation that percent visible crystals increased with increasing pCO₂ in sagittae, right lapilli, and left asterisci is consistent with the characterization of rough-type otoliths as abnormal in other species (Béarez et al., 2005; Ma et al., 2008; Browning et al., 2012). Increasing roughness could be a symptom of haphazard CaCO₃ aggregation, evidence of altered protein matrix deposition, and/or a snapshot of an evolving CaCO₃ crystal habit/polymorph baseline. While increasing roughness seems unlikely to affect otolith displacement amplitude, it could conceivably impact otolith-maculae mechanics with unknown consequences for function; in some fish, otoliths are observed to be rough on the ventral end only, driving maculae deformation by hooking them to the otolith surface (Ohnishi et al., 2002). More research concurrently investigating fish behavior, ocean acidification-induced otolith roughness, and maculae displacement is needed to explore this hypothesis.

As might be expected when rearing many hundreds of fish in the most delicate early stages of development, Amphiprion clarkii larvae experienced substantial mortality throughout the experimental trial. Although there is evidence in the literature of acute CO₂ toxicity in larval teleosts, this is typically observed at pCO₂ levels far exceeding those evaluated here (i.e., >48,000 uatm) (Kikkawa, Ishimatsu & Kita, 2003; Ishimatsu et al., 2004; Kikkawa, Kita & Ishimatsu, 2004); this is consistent with the absence of evidence that treatment instigated fish mortality. The sublethal impact of pCO₂ on settlement competency at 10 dph, however, was dramatic. There is some evidence of inhibited larval growth following delays in metamorphosis (Victor, 1986; McCormick, 1999), and some degree of inhibited growth was observed here, but the present analysis should be interpreted with caution due to the following caveats: (i) 10 dph was selected as the onset of metamorphosis and presumed settlement competency, but this threshold may be arbitrary to metamorphosis and settlement completion; (ii) metamorphosis and response to settlement cues may decouple at elevated seawater pCO₂ (Rossi et al., 2015), complicating metamorphosis as a reliable indicator for settlement competency; (iii) indeed, laboratory-reared fish were unexposed to settlement cues including reef sounds, lunar phase, and anemone presence; (iv) otolith microstructure evidence of settlement competency was not investigated; (v) inhibited larval growth due to delayed metamorphosis isn’t known to translate to inhibited growth post-settlement (McCormick, 1999). Finally, the impact on somatic growth was marginal: the reduction in fish standard lengths with increasing pCO₂ amounts to a small fraction of fish standard lengths, and this effect was only relevant in conditions of aragonite undersaturation.

Conclusions

This work corroborates evidence of otolith growth and altered shape with increasing seawater pCO₂ reported for other taxa in a novel taxon, Amphiprion clarkii. In addition, it reports evidence of increasing otolith lateral development and surface roughness with increasing pCO₂. Impacts were observed in all otolith types, including the previously uninvestigated asterisci. Each otolith type and side were investigated independently, indicating asymmetrical responses of lapilli and asterisci to pCO₂. The experimental design and analysis facilitated construction of pCO₂ dose–response curves, which were created for all otolith types and sides in A. clarkii excepting right asterisci. These curves outline changes to multiple morphometric and mineralogical variables and may be leveraged to predict responses to pCO₂ conditions not investigated here. These responses could impact auditory and/or vestibular sensitivity in teleosts, adding to previous observations and hypotheses involving sagittae and lapilli. In summary, the work adds to the existing knowledge base regarding otolith response to ocean acidification, which may aid in predicting and preserving teleost fitness in the near-future ocean.

Supplemental Information