One of these is not like the others

The fact that suction feeding is nearly ubiquitous in the aquatic realm is undisputed. Suction feeding is so spectacularly abundant among fishes that it is striking to find groups that have evolved alternative prey capture mechanisms. Among these are the holocephalans, a group of mostly extinct aquatic vertebrates that dominated early marine ecosystems. Holocephalans, based upon most morphological accounts to date, are suited for an entirely different mode of food capture: prey crushing. A key feature shared by all holocephalans is a nonmobile upper jaw, and the teeth typically form a plate-like surface. The few living examples demonstrate little ability to generate even nominal water flow into the option mouth for capturing food (1). Thus, it is striking that Dearden et al. (2) have reconstructed from the fossil record evidence of suction feeding in Iniopera, suggesting a broader repertoire in the early history of the group than seen today.

Suction prey capture is a remarkably versatile mechanism for capturing unattached or elusive food items in the aquatic realm (3). Water is dense and viscous, and as such an organism moving through the water in an attempt to obtain a food item will tend to push the food item away from itself. This is due to the forces exerted on the water by the moving organism, which creates a “bow wave,” much like what can be seen in front of a moving boat on the surface of the water. A vast array of organisms, spanning multiple magnitudes of scale, can generate suction to capture or process food (4, 5). Suction is generated by drawing water, and hopefully the food item, into the mouth via rapid expansion of the region of the head posterior to the mouth. This generates a current of water into the open mouth in much the same way as flow is generated into the column of a pipette when the pipette bulb is released (after initial compression).

A key evolutionary innovation that facilitated rapid inertial suction feeding, or what is often termed “high-performance” suction feeding, for capturing fast and elusive prey was jaw protrusion (6). Jaw protrusion in fishes refers to movement of the premaxilla bone (the most anterior bone forming the upper jaw) away from the head and toward the food item, which can be performed by most teleost fish and which has evolved multiple times in fishes even outside the teleost group (7). Suction feeding is also used by many elasmobranchs (sharks, skates, and rays), which use labial cartilages on the sides of the mouth, plus a protrusile Meckel’s cartilage (the upper jaw in elasmobranchs), to perform motions analogous to bony fish jaw protrusion (8). This ability to move the jaw toward the prey item, instead of the whole body, is thought to aid in stealth as well as enhancing suction production (9) and has become more elaborate and more extreme over evolutionary time. Jaw protrusion is credited, at least in part, for the massive radiation of fishes into the most speciose group of vertebrates we see today (7).

The holocephalans are a sister group to elasmobranchs (10), which is always somewhat curious due to a suite of rather incongruent traits readily observed in extant members. Holocephalans are bony, not cartilaginous. Holocephalans have a single gill opening covered by an operculum, not five (or more) gill slits. They lack jaws filled with serially replaced teeth, and the upper jaw is not mobile. Yet, Gillis et al. (11) find evidence for multiple, vestigial gill arches in the developing embryos of one extant species. Also, Johanson et al. (12) document an evolutionary trend from multiple, separate shark-like teeth in extinct species to the fused teeth in extant holocephalans. Could it be that the holocephalans were the ecologically dominant suction feeders prior to the rise of bony fishes with their protrusile jaws, and did they coopt suction feeding before or after the early elasmobranchs?

A race is on to pinpoint the earliest high-performance suction feeder, given that until now there was very little evidence for suction feeding in any holocephalan. Enter the closest contender, a stem-group elasmobranch. Coates et al. (13) describe features consistent with elasmobranch high-performance suction feeding in an ancient fossil shark, Tristychius arcuatus, from the Late Mississippian interval (∼331 mya): a mobile upper jaw protruded by the depression of the lower jaw, lateral cartilages that occlude the sides of the open mouth to create a small mouth opening, and expansive branchial elements activated by a retractable pectoral gridle. Tristychius is a hybodont, a sister group to modern elasmobranchs, which date back to around ∼360 mya. Dearden et al. (2) describe in Iniopera a small, anteriorly oriented mouth, an expandable pharynx, and muscular linkages connecting the pectoral girdle with the ventral region of the pharynx, facilitating expansion of the region. Iniopera is from the Late Pennsylvanian interval (beginning ∼300 mya). However, Iniopterygians are known in the fossil record from at least ∼330 mya. Thus, suction feeding likely evolved twice within the Carboniferous Period; the two lineages diverged in the Late Cambrian (∼500 mya). The elasmobranch version of suction feeding does seem to have preceded the holocephalan by as much as ∼30 mya, if the ages of the fossils are correct. Currently, we are severely limited in the availability of three-dimensional fossils preserved with the detail of Iniopera and Tristychius, but we are certainly beginning to see just how old the origins of high-performance suction might be.

There is no doubt that suction generation, although maybe not high-performance suction generation, is old. A universal feature of suction feeders appears to be a circular mouth opening and an oral cavity that expands in an anterior-to-posterior “wave” of expansion, facilitating the flow of water to and through to the pharynx or analogous regions (4). These characters, and concomitant suction food capture, exist in jawless fishes and cephalochordates, albeit not powerful suction. This means that suction was likely the dominant feeding mode at the base of the vertebrate tree, with some interesting convergent pump-generated flows observed in some of the urochordates (4). Incredibly, suction has also been observed in the tiny, carnivorous, aquatic bladderwort plant (5). The evolution of structures facilitating some suction generation over and over again only serves to further emphasize the importance of this versatile feeding mode.

High-performance suction, however, still appears to be a uniquely vertebrate feature (4), and Iniopera seems to tick all the boxes (Fig. 1): a round, down-turned mouth and an expandable head that would facilitate a large wave of expansion. It is also possible that the jaw was protruded by a soft skeletal element that did not fossilize, given the apparent expandable nature of the head. However, even without obvious skeletal elements that would aid in jaw protrusion, such as labial cartilages, the features described for Iniopera are still consistent with high-performance suction feeding in many bony fishes. Furthermore, the immobile upper jaw of Iniopera lends itself to comparisons with tetrapod suction feeding, like in aquatic turtles, amphibians, and marine mammals (4). In species that lack gills or other openings posterior to the jaws, the water mobilized by the wave of expansion will rebound (i.e., reverse direction), such as in tetrapods. It is unclear the extent to which rebounding might occur in Iniopera, with its single opercular opening, but we know from tetrapods that this bidirectional flow does not limit high-performance suction.

Fig. 1.

Reconstruction of the cranial elements of Iniopera in a relaxed, nonfeeding state (A) and in an expanded state consistent with maximum expansion (B). Blue indicates the space, and therefore the volume of water that would be present, within the oral cavity in both states. Note the differences in the volume from A to B, suggesting water has been drawn into the oral cavity via suction. Image credit: R. Dearden, University of Birmingham, Birmingham, UK.

All of this begs the question of why holocephalans became the dominant durophagous predators of their day. Does competition for resources alone explain the holocephalan radiation into durophagy? Could suction have facilitated a transition to durophagy, allowing predators to suck up smaller, hard-shelled prey before crushing them? The aforementioned dearth of well-preserved elasmobranch and holocephalans fossils from this era means it will take some more digging, quite literally, to uncover the mysteries of ancient predator–prey interactions. In the meantime, the novelty of a suction-feeding holocephalan provides enough excitement to keep us interested in looking further.