Identity of the reverse drill-hole producer

Among drilling gastropods, Klompmaker (Reference Klompmaker2009) only found naticids in the studied Langenboom samples, with nearly all drill-hole traces preserved in co-occurring shell-bearing invertebrate prey also having been attributed to naticids (Klompmaker, Reference Klompmaker2009, Reference Klompmaker2011, Reference Klompmaker2012; Klompmaker et al., Reference Klompmaker, Karasawa, Portell, Fraaije and Ando2013). Naticids usually plough through the sediment (Kabat, Reference Kabat1990; Kelley and Hansen, Reference Kelley, Hansen, Kelley, Kowalewski and Hansen2003), although they can be found on the surface occasionally (Kelley and Hansen, Reference Kelley, Hansen, Kelley, Kowalewski and Hansen2003; Pahari et al., Reference Pahari, Mondal, Bardhan, Sarkar, Saha and Buragohain2016). Astartids are also known to be (shallow) burrowers (e.g., Seilacher, Reference Seilacher1990; Damborenea and Manceñido, Reference Damborenea and Manceñido2005). Thus, the habitat overlap of Astarte incerta and naticids, the abundance of naticids in the assemblage, and the parabolic cross-sectional shape of the drill hole combined suggest that the reverse drill hole is highly likely to be of naticid origin, and represents the first report of this behavior in the family.

Specimens of Astarte goldfussi most likely were shallow, infaunal burrowers vulnerable to both muricids and naticids, both abundantly present in the Miocene sediments at Miste (Janssen, Reference Janssen1984). Muricids are epifaunal drillers, although they also can shallowly dig for prey (Kelley and Hansen, Reference Kelley, Hansen, Kelley, Kowalewski and Hansen2003). The incompleteness of the drill hole, the fact that muricid and naticid drill holes can have overlapping morphologies (compare Kitchell et al., Reference Kitchell, Boggs, Kitchell and Rice1981, fig. 5 and Radwin and Wells, Reference Radwin and Wells1968, figs. 12–18), and the abundance of muricids and naticids in the same assemblage precludes attribution of this reverse drill hole to a particular gastropod group.

First fossil reverse gastropod drill holes

Some reports briefly mentioned reverse gastropod drill holes in extant bivalves. Dietl and Alexander (Reference Dietl and Alexander1995) reported on an empty valve of the bivalve Chione elevata (Say, Reference Say1822) with a reverse drill hole. Hancock (Reference Hancock1959) briefly mentioned that rare reverse drill holes, probably produced by Urosalpinx cinerea (Say, Reference Say1822), were found among well-stocked beds of oysters, probably Ostrea edulis Linnaeus, Reference Linnaeus1758. Finally, Carriker (Reference Carriker1955) described that the muricid oyster drill U. cinerea sometimes missed living oysters—probably Crassostrea virginica (Gmelin, Reference Gmelin and Gmelin1791)—and drilled instead into empty shells beneath, perhaps producing reverse drill holes. The specimens herein showcase the first instances from the fossil record of this gastropod behavior.

Possible reasons for reverse drill holes

Why did some gastropods drill from the inner side of the shell? Carnivorous gastropods locate their food through chemical cues in the water column (chemoreception) and mechanoreception, making use of vibrations to detect prey (e.g., Carriker, Reference Carriker1955; Morton, Reference Morton1960; Kohn, Reference Kohn1961; Kitching and Pearson, Reference Kitching and Pearson1981; Croll, Reference Croll1983; Chase, Reference Chase2002). Because the specimens herein were dead when drilled, mechanoreception to locate these shells can be excluded. One explanation for the occurrence of reverse drill holes could be that ill or diseased predators lost their capability to distinguish live from dead prey. This hypothesis is, however, unlikely because observations of dying naticids under experimental conditions indicate that they usually do not drill, probably because the drilling process is metabolically costly (personal observation, GPD, 2015). Chemical cues from nearby living potential prey could serve as an explanation. Based on the reports by Carriker (Reference Carriker1955) and Hancock (Reference Hancock1959), Carriker and Yochelson (Reference Carriker and Yochelson1968) hypothesized that the high amount of ectocrines (chemical cues released into the environment) and the close proximity of valves of many different oyster specimens could explain the reverse drill holes made by muricids in oyster valves. Similarly, Carriker and Van Zandt (Reference Carriker, Van Zandt, Winn and Olla1972) mentioned that an oyster drill (Muricidae) can penetrate the valve of a dead oyster when triggered by a chemical cue escaping from the gaped valves of an actively feeding prey individual. Although the reverse-drilled astartid valves described herein did not live in clusters of cemented specimens like some oysters do, the density of astartid specimens within the bottom can be high today (up to a few hundreds of individuals/m²; e.g., Sejr et al., Reference Sejr, Jensen and Rysgaard2000; Skazina et al., Reference Skazina, Sofronova and Khaitov2013) and was likely high too for the studied assemblages because of the high relative abundance of astartids in the bivalve assemblages of Langenboom and Miste (Janssen, Reference Janssen1984; Klompmaker, Reference Klompmaker2009). Therefore, we hypothesize that well-mixed chemical cues emanating from dense aggregations of living prey could have confused foraging predators (i.e., interfered with their search behavior) and stimulated them to initiate the drilling process on empty valves.

Although the chemosensory searching capacity for prey by drilling predators might have been hindered by well-mixed prey chemical cues, shifting to secondary sensory cues (e.g., tactile information) to make prey choice decisions does not seem to have offset such effects. Predators did not detect the smooth surface (relative to the ornamented outer side) and concave shape of the inner side of the prey's valve, which would have been a tactile cue signaling that the prey was dead and should be abandoned (assuming predators were adapted to differentiate between shell shapes). Still, the relative value of a secondary stimulus in eliciting a drilling response in the predator might also depend upon context. For instance, predators might use tactile information differently based on internal context (e.g., motivation). Given the slow drilling rate of 0.01–0.02 mm/hr for modern naticid and muricid drillers (Carriker, Reference Carriker1955; Carriker and Van Zandt, Reference Carriker, Van Zandt, Winn and Olla1972; Kitchell et al., Reference Kitchell, Boggs, Kitchell and Rice1981; Chattopadhyay and Baumiller, Reference Chattopadhyay and Baumiller2009), instances of mistaken predation were costly energetically. If a drilling predator was strongly motivated to undergo the energetically costly and time-consuming process of acquiring food due to hunger, it might have made decisions more rapidly and relied less on secondary tactile cues when a primary cue (i.e., chemical information) was obscured or unavailable. This scenario presents an alternative (albeit not mutually exclusive) hypothesis to explain reverse drilling behavior. However, we stress that reverse holes were rare in the two assemblages studied (< 1% of drill holes), implying that the driller's sensory and decision-making processes were generally reliable at distinguishing dead from live prey. The rarity of such reverse drill holes refutes claims that naticids cannot distinguish between live and dead specimens (Hoffman et al. Reference Hoffman, Pisera and Ryszkiewicz1974; Stanton and Nelson Reference Stanton and Nelson1980; Pek and Mikuláš, Reference Pek and Mikuláš1996, p. 112). Instead, more support is found for Kitchell et al. (Reference Kitchell, Boggs, Rice, Kitchell, Hoffman and Martinell1986, p. 297), who argued that “naticid predators can readily distinguish live prey from empty shells,” but with some exceptions.

Comparison to other types of failed drilling predation

Various authors have reported on multiply-drilled specimens and incomplete drill holes in mollusks produced by predatory gastropods in study systems from different regions and geological ages (e.g., Kelley et al., Reference Kelley, Hansen, Graham and Huntoon2001; Kelley and Hansen, Reference Kelley and Hansen2006; Sawyer and Zuschin, Reference Sawyer and Zuschin2010, Reference Sawyer and Zuschin2011; Mondal et al., Reference Mondal, Goswami and Bardhan2017; Harper et al., Reference Harper, Crame and Sogot2018; Goswami et al., Reference Goswami, Das, Bardhan and Paul2021). Although there are exceptions (see Kelley and Hansen, Reference Kelley, Hansen, Kelley, Kowalewski and Hansen2003; Chattopadhyay and Baumiller, Reference Chattopadhyay and Baumiller2007; Hutchings and Herbert, Reference Hutchings and Herbert2013), such holes are often considered as evidence of failed predation (e.g., Vermeij, Reference Vermeij1987; Kelley and Hansen, Reference Kelley, Hansen, Kelley, Kowalewski and Hansen2003). Multiply-drilled specimens and incomplete drill holes that represent evidence of failed predation do not always equate to mistaken predation by the predator as for reverse drill holes. Incomplete drill holes are mistakes if a predator selected a prey too large for it to handle or too thick to penetrate, but not if a driller was interrupted by another predator or by an environmental disturbance (e.g., storm). The creation of a second drill hole in a prey shell would be the predator's mistake if the prey shell was empty, having already been drilled by another predator or if the same driller abandoned its initial drilling site and started another drill hole, but not if a driller was interrupted by a competitor that initiated drilling at another site. Unfortunately, it is difficult to quantify what proportions of incomplete drill holes and drill holes in multiply-drilled specimens represent mistaken predation attempts. However, the commonness of these types of predation relative to reverse drill holes within the same assemblages indicates that reverse drill holes likely are rarer than mistaken predation via incomplete drill holes and holes in multiply-drilled specimens for both assemblages.

Another type of unequivocal mistaken predation was reported from the Langenboom locality previously. Gaemers and Langeveld (Reference Gaemers and Langeveld2015) attributed very low percentages of incomplete drill holes (generally < 0.4% of all otoliths within large samples) found in fish otoliths to naticids that mistook them for tellinid bivalves initially, then discovered their mistake and aborted all drill holes well before full penetration. This value is comparable to the frequency of reverse drill holes that we report herein.

The occurrence of reverse drill holes in bivalve prey begs the question of whether complete drill holes that were initiated through the outer side of valves by predators could also have been produced in already dead bivalves. Such instances of mistaken predation, although probably uncommon like reverse drill holes, would be considered successful drill holes when studying death and fossil assemblages. Unlike reverse drill holes, however, identification of such behavior solely from the dead, empty remains of drilled prey is not possible. If such behavior is confirmed with observations of living predator-prey interactions, some drilling intensities might have been slightly overestimated when interpreted as successful predation.

Future research on reverse drill holes could focus on when such holes first appeared in the fossil record; their frequency through time, also relative to other types of failed predation within assemblages; and/or whether such holes are restricted to prey taxa living in dense aggregations.