Each year, corals of the Western Atlantic genus Orbicella (formerly Montastraea) deposit a skeletal band composed of high and low-density portions of calcium carbonate (Highsmith, 1979; Barns and Lough, 1993). The width of each band (linear extension) varies between years and is sensitive to natural and anthropogenic environmental changes (Buddemeier et al., 1974; Carilli et al., 2010; Rico-Esenaro et al., 2019). For example, changes in time series of the linear extension of Orbicella annularis at various sites in Puerto Rico correlated positively with water temperature (Dodge, 1981), and at the East Flower Gardens Bank, O. annularis extension rates varied inversely with annual discharge of the Atchafalaya river (Dodge and Lang, 1983). Mean linear extension value and variability depend on (i) the size of the coral colony and the length of the sampled time series, (ii) genetic differences between colonies (Osinga et al., 2011), and (iii) acclimation to the environment (Castillo et al., 2011; Groves et al., 2018). Furthermore, under moderately adverse conditions, species of Orbicella avoid a loss in linear extension by decreasing skeletal density (Dodge and Brass, 1984; Carricart-Ganivet and Merino, 2001). However, if conditions are severe enough, the mean linear extension is depressed (Dodge and Thomson, 1974; Dodge and Brass, 1984). The mean linear extension of Orbicella spp. can decrease with increasing temperature (Carricart-Ganivet, 2004), low light due to water turbidity and depth (Hubbard and Scaturo, 1985; Huston, 1985), sedimentation (Dodge et al., 1974; Torres, 2001; Carilli et al., 2009) and water quality (Dodge and Brass, 1984).

Several studies have examined the mean linear extension of fossil corals and made ecological and environmental inferences. For example, Gischler et al. (2009) reported a lower-than-present mean linear extension in massive Pleistocene Orbicella corals from lagoonal environments in Florida. They mentioned higher temperature as a probable cause during the Last Interglacial (LIG) which lasted ~10 ka (~130 to 120 ka). This interval was characterized by enhanced summer insolation that forced large polar deglaciation, leading to a reduced latitudinal temperature gradient and a higher sea level (Stirling et al., 1998; Denton et al., 2010). These conditions favored the extensive development of coral reefs as tropical waters expanded polewards and warm seas transgressed further across shallow shelves (Greenstein and Pandolfi, 2008; O’Lear et al., 2008; Blanchon, 2010; Kiessling et al., 2012). At that time, species richness in the Caribbean was slightly higher than present (Budd et al., 1994; Budd, 2000) and the Orbicella annularis species complex included three extant species, as well as a columnar form, and the organ-pipe Orbicella nancyi (Pandolfi, 2007). O. nancyi was widely distributed in shallow waters of San Andrés, Barbados, Lesser and the Greater Antilles, and Curaςao (Geister, 1977; Pandolfi and Jackson, 2001; Pandolfi et al., 2002). Pandolfi et al. (2002) proposed that the now-extinct O. nancyi had a high linear extension and that other species of Orbicella reduced their linear growth when this coral was present.

To investigate the extent to which environmental or ecological factors control on linear extension, we compared rates in massive and columnar species of Orbicella from a 125 kyr-old reef deposit at Xcaret, Quintana Roo, México (Jordán-Dahlgren, 1997; Blanchon et al., 2009; Blanchon, 2010), with two modern Orbicella species from shallow habitats of the wider Caribbean. We assess autocorrelation and linear trends of the coral’s time series and mean differences within and between fossil species in Xcaret and modern O. faveolata and O. annularis.

Neither modern nor fossil corals show consistent lag-1 autocorrelation on their linear extension time series. As shown in Table S1 , of the 44 colonies sampled, only 11% showed lag-1 autocorrelation. Similarly, no consistent linear trends were found. Only 25% of the coral colonies showed significant trends, with three modern corals showing a positive trend, five showing a negative trend, and two fossil corals showing positive linear trends ( Table S1 ). More interestingly, results show that fossil colonies of Orbicella have linear extension rates that are ~40% lower than modern corals ( Figure 5 , Table 1 ). The mean linear extension rate in fossil corals from Pleistocene Xcaret vary between 0.52 (0.48 to 0.56) for the fossil O. faveolta and 0.66 (0.58 to 0.74) cm/year for O. nancyi (mean and 95% credible intervals, Figure 4 ). In contrast, the mean linear extension rates in modern corals are similar for O. annularis and O. faveolata and vary between 0.87 (0.84 to 0.89) at the Gulf of Mexico and 0.99 (0.94 to 1.06) cm/year at the Mexican Caribbean. The mean extension rate is slightly higher in the Caribbean (Car) than the Gulf of Mexico (GoM) (mean effect size: GoM vs Car. -0.099 (-0.15 to -0.04) cm/year) and higher in both modern seas compared to Pleistocene Caribbean (PCar) (mean effect size: GoM vs PCar 0.377 (0.32 to 0.42) cm/year and Car vs PCar 0.47 (0.4 to 0.54) cm/year).

Differences in the mean linear extension exist between the fossil corals with O. faveolata having credibly lower linear extension than O. nancyi with a mean effect size of 0.14 (0.05 to 0.23) cm/year. Compared to their modern counterparts, the corals O. nancyi and fossil O. faveolata have lower extension rates ( Figure 5 ). In general, modern coral species showed no significant differences in their growth rates or had small effect sizes ( Table 2 ). However, the effect sizes of the differences in growth rates are large between modern and fossil corals and intermediate between the two fossil species at Xcaret ( Table 1 ). Compared to other Atlantic coral fossils of the same genus or species, Xcaret´s O. nancyi shows a lower mean linear extension and massive forms show a range from 0.5 to 0.9 cm/year ( Table 3 ).

Small differences in mean linear extension can, in a few years, have a significant effect on the colony size in terms of the amount of tissue grown; in addition, a columnar growth form like O. nancyi’s has an even more significant effect on the rapid growth of living tissue ( Figure 6 ).

The fossil coral at Xcaret with the highest linear extension was O. nancyi, followed by O. faveolata ( Figure 5 ). Although O. nancyi has a higher linear extension than fossil O. faveolata, both species have a mean linear extension that is ~ 0.5 cm lower than expected (~ 50% less) when compared to modern corals and the fossil O. nancyi from Barbados (Tomascik, 1990; Holcomb et al., 2004). Slow growing fossil corals have also been reported elsewhere, including massive forms in Florida (Gischler et al., 2009), columnar forms in Barbados (Holcomb et al., 2004) and organ pipe forms in Curaςao (del Valle, 2012). By contrast some Orbicella spp. fossils show high growth rates (> 0.8 cm/year), including massive forms from Curaςao (Johnson and Pérez, 2006), and organ pipe forms from Curaςao and Barbados (Holcomb et al., 2004; del Valle, 2012). Both fast and slow-growing fossil corals come from paleo-habitats that were less than 6 m deep.

Two possible explanations have been given for low extension rates on fossil corals. First, Gischler et al. (2009), working with massive Orbicella from the LIG in Florida, considered temperature as a possible environmental factor inhibiting linear extension. Second, Pandolfi et al. (2002) and Pandolfi (2007) proposed that the linear extension of Pleistocene Orbicella corals sympatric with O. nancyi was depressed through competition for space and light. According to this competition hypothesis, the extinction of O. nancyi at ~82 kyr caused a character release in the O. annularis species complex, partly because O. nancyi was a fast grower and, through competition, suppressed the growth of sympatric Orbicella spp. Upon extinction, O. nancyi released the extant species from the competition, and linear extension rates increased presumably to modern values.

Neither of these hypotheses, however, sufficiently explains differences in the linear extension data between Xcaret and modern corals. Our modern O. faveolata and O. annularis extension data occur in two regions with different temperature regimes: the Mexican Caribbean Sea, which has a higher (~1.3° C) mean sea surface temperature than the Gulf of Mexico (Carricart-Ganivet, 2004). This temperature difference is like the estimated higher sea temperature of the LIG (Otto-Bliesner et al., 2006), and yet differences in linear extension between modern corals in both regions are not as large compared to fossil and modern corals. Modern temperature differences between the regions are likely within the acclimation capacities of these corals (Dodge and Brass, 1984; Carricart-Ganivet and Merino, 2001; Carricart-Ganivet, 2004). In addition, Helmle et al. (2011) showed a recent increase in linear extension for O. faveolata that made coral growth tolerant to the ongoing increase in temperature due to climate change. In fact, Orbicella corals can exhibit a positive correlation between linear extension and temperature (Dodge, 1981), yet higher water temperatures can lead to reduced rates of skeletal extension (Gischler et al., 2009). This suggests that there may be a temperature threshold beyond which coral extension is negatively impacted. The exact mechanisms behind the threshold are not fully understood but might be related to the physiology of the coral. Under moderate stress Orbicella avoids a loss in linear extension by decreasing skeletal density (Carricart-Ganivet and Merino, 2001), but Mendes and Woodley (2002) showed a decrease in extension rate of O. annularis after a bleaching event possibly related to an impairment in reproduction, the effect was harsher the longer the extreme water temperature lasted. Maintaining high linear extension rates might be advantageous to the corals to compete for light and space but at high physiological cost (Rinkevich, 1996).

Nevertheless, competition could have the opposite effect. Corals compete using various mechanisms (Richardson et al., 1979) and their growth rates have been shown to decrease in the presence of macroalgae (Lirman, 2001), turf algae (Quan-Young and Espinoza-Avalos, 2006) and other corals (Tanner, 1997; Lapid and Chadwick, 2006). As a consequence, colony growth rates can influence community structure and development through competition (Maguire and Porter, 1977). Comparing the extant Orbicella species, interspecific ecological interactions show that O. annularis is the least aggressive, followed by O. faveolata and O. franksi (Weil and Knowlton, 1993). This aggressiveness between competing species can result in tissue mortality and sometimes the suppression of mean linear growth rates, with faster-growing species escaping competition by size (Sebens, 1982; Tanner, 1997; Cruz-Piñon et al., 2003). For our modern corals, which share the same habitat, we found no evidence that the stronger competitor, O. faveolata, had systematically higher linear extension rates ( Figure 5 , Table 1 ). This finding is consistent with Cruz-Piñon et al. (2003), who reported higher skeletal extension rates in O. faveolata than in O. annularis, whereas Tomascik (1990) reported a reverse trend in the same corals from Barbados, and Van Veghel and Bosscher (1993) found no significant differences between the two species. This lack of consistency in species linear extension differences is likely due to interspecific competition being rare on modern reefs (van Woesik, 2002), and the large variability in reef environments and growth plasticity in species of Orbicella (Foster, 1979). Given that Xcaret’s O. nancyi shows a lower extension rate than modern and fossil corals in other locations, it is possible that other characteristics, such as growth form ( Figure 5 ), made O. nancyi a stronger competitor (Pandolfi et al., 2002). Despite being in the same stratigraphic unit and level, it is uncertain if these coral species were cohorts that lived together. However, the clear facies on the deposit point to similar environmental conditions during the time the corals grew ( Figure 2 ). Also, members of the Orbicella complex seem to have coexisted at different locations and periods of the Pleistocene (see Pandolfi et al., 2002) and it is likely that this also occurred at Xcaret.

The lack of a consistent temperature or species effect on extension rates implies that the lower linear extension rates of fossil Orbicella at both Xcaret and Florida (Gischler et al., 2009) is more likely to have resulted from deleterious habitat conditions. Data from O. nancyi samples collected by del Valle (2012), showed lower extension rates on the leeward side of Curaçao. Lower extension rates in back-reef corals from Xcaret are consistent with those of lagoonal corals reported by Gischler et al. (2009). Corals at both of these sites grew in or adjacent to large lagoons that resulted from extensive coastal inundation during a sea-level highstand at least 6 m higher than present (Blanchon et al., 2009). As a result, the lagoons and adjacent back reef environments must have been subject to high sediment flux and were associated with the development of sediment-tolerant biota (Blanchon, 2010). If coral growth in and near these lagoons was indeed depressed by sedimentation, then the same species from different habitats should show higher extension rates, which seems to be the case. More growth data from Pleistocene corals in different paleo-habitats are required to test this hypothesis and better understand their response to changing environments.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

AJ-G, EJ-D collected data. AJ-G, EJ-D performed the statistical analysis. AJ-G, PB, EJ-D, JC-G analysed the results and wrote the manuscript. All authors contributed to reviewing and editing the manuscript. All authors contributed to the article and approved the submitted version.