A classic key innovation constrains oral jaw functional diversification in fishes

Abstract

Modifications to the pharyngeal jaws-a prey processing system located posterior to the mouth cavity-are widely considered a key innovation that enhanced diversification within several prominent fish clades. Seen in cichlids, damselfishes, wrasses, and a few other lineages, these musculoskeletal alterations are believed to increase the evolutionary independence and, thus, the diversification of the oral and pharyngeal jaw systems. To test this classic hypothesis, we conducted comparative phylogenetic analyses to assess the effect of the pharyngeal novelty on the diversification of feeding morphology and kinematics across a taxonomically diverse sample of spiny-rayed fishes. We quantified movements of the oral jaws and other craniofacial structures from 689 suction-feeding strikes using high-speed videos collected from 228 species with and without the pharyngeal jaw novelty. Contradicting long-held predictions, we find significantly greater disparity across all traits and faster rates of oral jaw functional evolution in fishes without the specialized prey processing system. The modified pharyngeal jaw is undoubtedly a functional innovation as it enhances the strength of the prey processing system, facilitating exceptional transition rates to feeding on hard and tough prey. However, it also restricts the diversification of the feeding system, revealing that the impact of pharyngognathy is more nuanced than previously thought. In light of these and other recent findings, a reinterpretation of the macroevolutionary consequences of the pharyngeal jaw novelty is needed.

Keywords: Acanthomorpha; evolutionary integration; feeding kinematics; geometric morphometrics; morphology; pharyngognathy.

Figure 1.

(A) Illustrations depicting the oral (lateral view; shaded in black) and pharyngeal (lateral view; shaded in gold) jaw structures involved in prey capture and processing functions, respectively. Though the prey processing system is similar across spiny-rayed fishes, several lineages exhibit a modified pharyngeal jaw (MPJ) where major musculoskeletal changes enable greater strength and mobility in prey manipulation and transport. Modifications include (1) fusion of the paired lower pharyngeal jaw bones into a single skeletal structure, (2) a mobile joint between each of the paired upper pharyngeal jaw bones and the neurocranium, and (3) a muscular sling suspending the lower pharyngeal jaw structure from the neurocranium. Bars below pharyngeal jaw structures represent 10mm scales. (B) A representative stochastic character map shows the relatedness of the 228 study species across 43 acanthomorph families and the evolutionary character history of the MPJ. Note that Serranidae includes Epinephelidae, Liopropomatidae, and Anthiadidae; Labridae includes Scaridae. Pie charts show the frequency of the MPJ and non-MPJ states at each node across 1,000 stochastic character maps and changes in branch color reveal that the novelty independently evolved four times across our sampled species between ~48 and 93 Ma. Images display closed mouth (left) and maximum gape (right) stages of feeding motions for a subset of MPJ and non-MPJ species.

Figure 2.

(A) Average shape trajectories for suction-based feeding motions in fishes with modified or unmodified pharyngeal jaws. Trajectories are comprised of 10 craniofacial shapes sampled at equally spaced time points from closed mouth to maximum jaw expansion. Deformation grids show theoretical shapes at the extremes of PCs 1 and 2, where filled circles represent 10 fixed landmarks (red) and 8 sliding semi-landmarks (pink; along the ventral margin of the head) used to track feeding movements. PC1 primarily captures variation in interspecific head shape, while PC2 describes motion-based shape change. (B) Average motion trajectory for Plectropomus laevis (Grouper; non-MPJ; highlighted in black in panel A). Images depict craniofacial shapes at different time points throughout the feeding motion. Total craniofacial kinesis is the sum of Procrustes distances (d_i) between successive shapes, describing the amount of shape change achieved during prey capture. Kinesis skew is the natural logarithm of kinesis across the last five motion shapes (∑(d₆:d₉)) divided by total craniofacial kinesis (∑(d₁:d₉)). Note that kinesis and kinesis skew are computed with full-dimensional shape data but are shown here in two-dimensions for visualization.

Figure 3.

Species lacking modified pahryngeal jaws show greater disparity than species with a specialized prey processing system for all traits, including (A) interspecific craniofacial shape (3.4-fold more disparity), (B) the movements of key skeletal structures that contribute to feeding kinematics (2.7-fold more disparity), (C) total craniofacial kinesis (4.7-fold more disparity), and (D) kinesis skew (2.9-fold more disparity). Select species exemplifying morphological and functional diversity in MPJ and non-MPJ fishes include (a) Caquetaia kraussii (Cichlid), (b) Crenicara punctulata (Cichlid), (c) Stegastes partitus (Damselfish), (d) Aulostomus chinensis (Trumpetfish), (e) Liopropoma eukrines (Painted Basslet), (f) Lepomis gulosus (North American Sunfish), (g) Cirrhitichthys oxycephalus (Hawkfish), (h) Centrogenys vaigiensis (False Scorpionfish), (i) Chaetodon kleinii (Butterflyfish), (j) Antennarius hispidus (Frogfish), (k) Epibulus insidiator (Wrasse), and (l) Terelabrus flavocephalus (Wrasse; all species depicted in Figure 1B).

Figure 4.

(A) Evolutionary correlation matrices reveal that MPJ (top triangle) and non-MPJ (bottom triangle) taxa have strong pairwise relationships between total craniofacial kinesis and all six motion components. Dashes indicate that pairwise relationships are not greater than those found between traits simulated under Brownian Motion. (B) A matrix depicting the difference in absolute values of pairwise correlations between MPJ and non-MPJ fishes reveals that species with a modified pharyngeal jaw show stronger coevolution of kinesis with premaxillary protrusion and lower jaw rotation. Circle size, color, and asterisks represent the difference in correlation strength, the group with the stronger correlation, and a significant difference between the two pairwise correlations (p ≤ 0.05), respectively.

Figure 5.

(A) Branch-specific estimates of the Brownian rate parameter for total craniofacial kinesis show 2.3-fold faster, state-dependent evolution in fish species with generic pharyngeal jaws compared to fishes with modified pharyngeal jaws. (B) Posterior estimates of the Brownian rate parameter in each group show minimal overlap, suggesting that the state-dependent rate of craniofacial kinesis evolution (total shape change along a motion trajectory between the closed mouth and maximum gape positions as depicted in non-MPJ species Plectropomus laevis) is significantly impacted by the state of the discrete trait—pharyngeal jaw configuration. We recover a 98.3% posterior probability of state-dependence, where 491,500 of 500,000 generations in our model return significantly different evolutionary rates of total craniofacial kinesis for the MPJ and non-MPJ groups.