The shift from the Paleozoic to the Modern (post-Paleozoic) Fauna involved a major influx of benthic molluscs (gastropods and bivalves) into offshore marine environments, resulting in mixed brachiopod–benthic mollusc paleocommunities as early as the late Paleozoic Era; this change might be expected to have affected contemporaneous predators. Gastropods and productidine brachiopods from five upper Pennsylvanian shales in Texas (United States) show greater repair frequencies than do other taxa, suggesting that crushing predators readily consumed gastropods early in the faunal transition but that some brachiopod taxa still composed an important component of predator diets. In contrast, drilling predators continued to prefer brachiopods almost exclusively over benthic molluscs; Paleozoic drillers may have been incapable of taking mobile prey. Contrary to previous hypotheses, brachiopods were probably not “mistaken” prey, at least during the Paleozoic. The addition of gastropods to the diet of crushing predators suggests a major change in ecological processes in response to the new fauna.
One of the most important evolutionary and ecological developments in the history of metazoan marine life is the transition from the Paleozoic Evolutionary Fauna (PEF, typified by brachiopods) to the Modern Evolutionary Fauna (MEF, typified by benthic molluscs) (Sepkoski, 1981, 1996). Although facilitated by the Permian and Triassic mass extinctions (Sepkoski, 1996), this transition was not abrupt, but rather occurred over tens of millions of years: by the Pennsylvanian and Permian (318–251 Ma), benthic molluscs (gastropods and bivalves) had already invaded offshore marine systems previously dominated by brachiopods (Sepkoski and Miller, 1985; Miller, 1988; Jablonski and Bottjer, 1990; Olszewski and Patzkowsky, 2001; Forcino et al., 2010). Here we demonstrate the initial effects of this invasion of new prey on the diet of marine predators using a temporal series of exceptionally well preserved paleocommunities from Pennsylvanian shales of Texas, United States. In addition to exploring how predation may have changed in response to the development of a more modern prey fauna, the research also tests the hypothesis that brachiopods may have been “mistaken prey”, i.e., that they were only taken either by mistake or because preferred prey were unavailable (Kowalewski et al., 2005), by directly comparing predation on brachiopods and benthic molluscs from the same communities. If predators switched to molluscs, then the argument for brachiopods as mistaken prey would be corroborated. In contrast, if brachiopods continued to be an important part of predator diets, then the mistaken prey hypothesis must be questioned.
Sepkoski (1981) documented three evolutionary faunas: the Cambrian Fauna, the PEF, and the MEF. For the post-Cambrian Paleozoic Era, benthic marine communities were commonly dominated by crinoids and brachiopods, whereas many members of the MEF, such as benthic molluscs, were largely restricted to nearshore habitats (Sepkoski and Miller, 1985; Jablonski and Bottjer, 1990). Most marine predators (fish and arthropods) lived offshore, and so brachiopods were frequent prey (Ausich and Gurrola, 1979; Alexander, 1981, 1986; Elliott and Bounds, 1987; Kowalewski et al., 1998; Leighton, 2001, 2003). However, by the late Paleozoic many bivalve and gastropod taxa moved offshore to become part of mixed brachiopod-mollusc communities (Sepkoski and Miller, 1985; Jablonski and Bottjer, 1990; Olszewski and Patzkowsky, 2001). Although some workers (Gould and Calloway, 1980) have argued that the faunas are divided by the Permo-Triassic extinction, the mixing of brachiopods and molluscs suggests that the transition was in progress millions of years before the mass extinction.
This replacement of the PEF by the MEF is arguably the most important ecological transition for marine invertebrates in the past 400 m.y., as it involves a change in ecological dominance from sedentary, epifaunal suspension feeders to mobile, often high-metabolism organisms including infauna and predators (Wagner et al., 2006). Benthic molluscs generally have a greater ratio of organic tissue to shell volume than do brachiopods, and all other things being equal, predators switch to such energetically desirable prey (Stephens and Krebs, 1986). Alexander (1986) documented a decrease in fracture repairs on brachiopods through the Paleozoic, and interpreted this to be due to increasing predator strength through time: late Paleozoic predators were larger and stronger (Moy-Thomas and Miles, 1971; Schram, 1981; Brett and Walker, 2002) and their attacks less likely to fail, so there were fewer instances of repairable damage. However, it is also possible that the decline in repairs was due to a decrease in attacks, as durophagous (shell-crushing) predators switched to more profitable prey. In a case study of a Permian mixed community (Hoffmeister et al., 2004), drilling predators drilled bivalves more than brachiopods, whereas other studies showed that mid-Paleozoic communities, in which bivalves were rare, sometimes exhibited high drilling frequencies on brachiopods (Kowalewski et al., 1998; Leighton, 2001, 2003). Kowalewski et al. (2005; Simoes et al., 2007) suggested that drilling on brachiopods has generally been low, and that these attacks were cases of “mistaken predation.” This makes sense because of the difference in tissue yield between brachiopods and benthic molluscs. However, as brachiopods were sedentary and often had thinner shells than mobile molluscs, the energetic cost of penetrating brachiopods may have been substantially less than that of pursuing and attacking molluscs. The ratio of benefit minus cost to time, a proxy for optimal foraging by the predator (Stephens and Krebs, 1986; Tyler et al., 2010), may have been greater for brachiopods, making them desirable prey. Based on all of this evidence, the transition from the PEF to the MEF may have had a major impact on marine predators, as predators not only had a greater choice of prey but may have been exposed to new natural-selection pressures.
We collected 9773 brachiopod, gastropod, and bivalve specimens from five upper Pennsylvanian (upper Moscovian–Gzelian) shales of Texas: East Mountain Shale, Placid Shale, Colony Creek Shale, Finis Shale, and Wayland Shale (Fig. 1; Fig. DR1 in the GSA Data Repository1), spanning 8–12 m.y. These are among the oldest units containing abundant, diverse, and sufficiently well preserved brachiopods and benthic molluscs in the same paleocommunities. The fossils are exceptional; the aragonitic mollusc shells are preserved as well as those of calcitic brachiopods, revealing fine-scale features such as ornament, teeth, muscle scars, and predation traces. Articulation rates are exceptionally high (87% brachiopods, 97% bivalves). The taxonomic composition of these communities is similar to that of intermediate-depth, mixed brachiopod-bivalve communities from Kansas (Cluster B; Olszewski and Patzkowsky, 2001, their figure 5). Within each unit, samples were collected from Grossman et al.’s (1991) ammonoid biofacies, which they interpreted as an intermediate-depth, offshore setting, probably below storm wave base, based on: a high-diversity fauna; a lack of ichnofabrics, phylloid algae, or sedimentary structures indicative of wave energy; the presence of ammonoids and phosphate nodules; and stable isotopic signatures.
We further divided brachiopods, gastropods, and bivalves into morphological categories (Fig. 1). These categories are not always consistent with taxonomy; however, from a predator’s standpoint, the shape and size of the prey is more relevant than taxonomy. One category, the non-chonetidine concavo-convex brachiopods (NCCC), was subdivided into small (adults ≤2.5 cm in the longest dimension) and large (adults >2.5 cm). NCCC brachiopods consist of members of the suborder Productidina and one genus of the suborder Orthotetidina. With the exception of several NCCC taxa, none of the taxa exceed 2.5 cm in these units. As larger specimens may have greater repair frequencies, due to a longer life, being more desirable prey (greater yield), or greater resistance because of their size (Vermeij, 1987), we analyzed small and large NCCC taxa both separately and together, to determine if prey size affected predation.
Crushing and drilling predation frequencies were calculated and compared for each morphological division. Examples of predation traces (Fig. DR2) and details on the criteria for identifying predation traces and on the method used for calculating frequencies are described in the Data Repository.
Total repair frequency within each unit varied from 5.6% to 18.0% (total for all units = 12.7%) (Fig. 1; Table DR1 in the Data Repository). Overall, gastropods had a significantly (χ2 test, p << 0.001) greater frequency of repair than did brachiopods, which in turn had a significantly (χ2 test, p << 0.001) greater repair frequency than did bivalves. These results hold for each unit except the Finis Shale, in which gastropod and brachiopod repair frequencies were not significantly different. Repair frequencies on brachiopods were influenced by the relative abundance of the suborder Productidina (approximately equivalent to NCCC brachiopods), which among brachiopods generally had the greatest predation frequencies. NCCC brachiopods composed 23% of the Finis Shale assemblage, the only unit in which brachiopod repair frequency was comparable to that of gastropods. In all other units, NCCC composed <6% of the prey. The data are summarized at the genus level in Table DR2.
Drilling frequencies were low for all groups, but greater for brachiopods (1.2%) than for gastropods (0.1%) or bivalves (0.0%). NCCC brachiopods experienced greater drilling frequencies (3.8%) than any other group (χ2 test, p << 0.001). Significant differences in drilling frequency were not always observed within individual units (four units significant for brachiopods versus bivalves, two units significant for brachiopods versus gastropods) because of very low drilling frequencies in some units (Fig. 1; Table DR1). However, all of the drilled specimens in this study were sessile individuals—either brachiopods or the sessile, planispiral gastropods. Drilling success, as determined by complete borings, was generally high (67% for sessile gastropods, 59% for brachiopods) but drilling frequencies were too low to permit statistical tests for differences in success among taxa.
Among brachiopods, NCCC brachiopods were preyed upon more frequently than other morphologies by both crushers (42.0%, compared to 7.4% in chonetidines and 2.1% in rhynchonellates [ = biconvex]; χ2 test, p << 0.001) and drillers (3.8%, compared to 0.7% in chonetids and 0.4% in rhynchonellates; χ2 test, p << 0.001). Summarized across all units, NCCC brachiopods were preyed upon more frequently than any other group, including any morphology of gastropods; however, significance of this result varied from unit to unit, and in some units, specific gastropod groups, especially the sessile, planispiral gastropods (Straparollus), experienced predation frequencies greater than that of NCCC brachiopods (Tables DR1 and DR2). There was no consistent statistical difference in repair frequency between small and large NCCC taxa; in two of the units, small NCCC taxa even had greater repair frequencies.
Gastropod morphological divisions experienced more similar overall repair frequencies (5.9%–22.1%) than did the brachiopod divisions. The rank order of morphological divisions in terms of repair frequencies also varied among units. The sessile, planispiral gastropods had the greatest overall repair frequency among gastropods (22.1%) and also had the greatest repair frequencies in four of the five units. As noted above, these were the only gastropods to experience drilling predation. In contrast, high-spired gastropods had the lowest repair frequency overall and in three of the units.
In conjunction with earlier work (Schindel et al., 1982; Signor and Brett, 1984; Huntley and Kowalewski, 2007), the present results suggest that gastropods were acceptable and frequent prey for crushing predators by the late Pennsylvanian. The transition from the PEF to the MEF may have resulted in prey-switching by durophages as early as the Pennsylvanian. Alternatively, predators had less success against gastropods and so repair frequencies for gastropods were greater (Alexander, 1986; Leighton, 2002) but this does not change the fact that gastropods were frequently attacked.
As brachiopods were also heavily preyed upon in the Finis Shale, the relative abundance of NCCC brachiopods (essentially productidines) in these units may have influenced repair frequencies. The abundance of one prey may influence the frequency of attack on other prey (Stephens and Krebs, 1986; Leighton, 2002). Productidines had a proportionally larger body mass (in terms of actual tissue) than did most brachiopods; thus productidines, when available, may have continued to be an important component of predatory diets. However, there was no consistent difference in repairs between small and large productidines, so size by itself was not the only basis for prey choice. Regardless of productidine abundance, gastropod repair frequencies were >10% in each unit, and even exceeded 20% in the Finis Shale where productidines were most abundant, thus demonstrating that gastropods were frequent prey.
In contrast, drilling predators never attacked mobile gastropods; drillers attacked brachiopods and more rarely sessile gastropods. These results suggest that drillers were either incapable of, or unwilling to, attack mobile prey. Unlike modern naticid gastropods, which envelop their prey within a large foot prior to drilling, Paleozoic drillers may have lacked the ability to grapple with prey that could escape. Previous work (Hoffmeister et al., 2004; Kowalewski et al., 2005) has suggested that brachiopods were not primary prey of drillers, and may have been mistaken prey based on consistently low drilling frequencies compared to modern drillers. This assumes that Paleozoic drillers attacked other prey more frequently. While the drilling frequencies on brachiopods reported here are low, they are even lower for molluscs. Drilling predators are specialized for attacking shelled prey, even if they do not always drill, so it is unlikely that the preferred Paleozoic prey were soft-bodied and thus unpreserved. Drilling on Paleozoic echinoderms has been reported (Baumiller, 1990) but is always associated with kleptoparasitic platyceratids. The drill holes in the Texas shale specimens are cylindrical and located over the body cavity, indicating true predation, as opposed to beveled holes over the mantle cavity made by platyceratids (Baumiller et al., 1999). Cylindrical drill holes on brachiopods have been reported frequently (e.g., Ausich and Gurrola, 1979; Smith et al., 1985; Leighton, 2001), but such holes are rarely seen on other Paleozoic prey. Given our results on crushing and drilling, the hypothesis that Paleozoic brachiopods were mistaken prey seems unlikely. One alternative is that Paleozoic predatory drillers were metabolically less active and so required less food per unit time than more recent drillers, and so drilling frequencies were lower. Similarly, the repair frequencies reported here are lower than post-Paleozoic repair frequencies (Vermeij, 1987). This hypothesis is consistent with previous theoretical and practical studies (Vermeij, 1987; Bambach, 1993). Other possible, and perhaps simpler, hypotheses are that (1) there were fewer or smaller predators in the Texas systems, (2) the environment was a high-risk setting for predators, or (3) the climate was cooler and so metabolic needs of the predators were less.
One surprising result of the study is that bivalves were preyed upon rarely, if at all, in all of the units. An injured bivalve may be less likely to survive an attack and repair (Vermeij, 1983), thus decreasing the likelihood of preservation, but this argument applies to brachiopods too. The Texas shale bivalves were generally unornamented and would have provided more body tissue to volume than most other taxa, so bivalves would be expected to be frequent prey. However, the great majority of bivalves in these units were shallow-infaunal burrowers such as nuculoids and heteroconchs. Whether bivalves infaunalized in response to predation is still an open question (see arguments for [Stanley, 1977; Vermeij, 1987] and against [McRoberts, 2001]). While it is beyond the scope of the present study to test the initial cause for infaunalization, the result is consistent with the hypothesis that infaunalization reduced predation on these bivalves. Paleozoic predators, including crushers, may not have been able to detect or excavate infaunal prey.
Cherns and Wright (2009) argued that aragonite dissolution may have biased the Paleozoic record against molluscs; as such, differences between Paleozoic and modern faunas may be exaggerated. However, it is unclear how aragonite dissolution would lead to the documented increase in molluscs in offshore settings through time (Sepkoski and Miller, 1985; Jablonski and Bottjer, 1990), which is a hallmark of the transition between faunas. While dissolution may be a general problem, the present results are focused on the effects of the offshore transition, and are based on well-preserved molluscan assemblages for which dissolution was inconsequential.
Vermeij (1977, 1987) documented a Mesozoic Marine Revolution, during which new clades of increasingly powerful durophages (e.g., crabs, rays, teleost fishes) appeared. Prey clades, especially gastropods, responded morphologically and behaviorally to the new threat (Vermeij, 1977). As a consequence, the Mesozoic marine revolution may have finalized the transition from the PEF to the MEF. Although the present study examines the earliest stages of this transition, the invasion of benthic molluscs, a definitive component of the MEF, into Pennsylvanian offshore systems had an effect on crushing predators. These Paleozoic communities already exhibited crushing predation of a more modern aspect. The transition from the PEF to the MEF involved not only a change in the fauna but also a change in ecosystem processes.
Our findings indicate that (1) Pennsylvanian crushing predators attacked gastropod prey, raising the possibility of a potentially broad-scale prey switch in response to the availability of new prey; (2) although gastropods were frequent prey for crushing predators, the predators continued to attack NCCC (mostly productidine) brachiopods at a frequency similar to that of gastropods; and (3) drilling predation was rare, but in contrast to crushers, drillers continued to attack brachiopods almost exclusively. Given the above results, the hypothesis that brachiopods were mistaken prey seems unlikely; brachiopods were preyed upon too frequently not to be a normal part of the predators’ diets. The results also suggest that Pennsylvanian predators were probably less likely to detect or excavate infaunal prey, and drilling predators may have been incapable of taking mobile prey.
Funding for this research was provided by a Natural Sciences and Engineering Research Council of Canada Discovery award, and U.S. National Science Foundation grant EAR-0746072, awarded to Leighton. We thank Patricia Kelley, John Huntley, and an anonymous reviewer for their thoughtful reviews, and Chris Schneider and Ann Molineux for their assistance and expertise in the field.