• <div style="background-image:url(/live/image/gid/32/width/1600/height/300/crop/1/41839_V14Cover_Lynch_Artwork.2.rev.1520229233.png)"/>


Influence Of Oral Morphology On Dietary Specialization In Two Herbivorous New Zealand Fishes

Ryan Drake
Department of Biology

Lake Forest College
Lake Forest, IL 60045
Download PDF


Herbivorous fishes are highly diverse organisms and have evolved characteristic features that enable them to feed. Most herbivorous fishes have minimized jaws with rigid skulls and suspensora, allowing them to feed continuously with small bites. In ecosystems with similar food options (e.g. algae), specific choices are typically selected for ingestion, sometimes with little competition from other species. Herbivorous fishes express morphological adaptations that differ between species that is correlated with this dietary behaviour. Morphological variation in the skeletal, muscular, and alimentary components promotes food preference and niche development. These mechanisms can explain the evolution of modern herbivorous fishes.

In this study, the alimentary contents of a New Zealand butterfish (Odax pullus) and a parore (Girella tricuspidata) were examined to determine dietary specialization. In addition, the skeletal and muscular structure of the jaws were also examined to achieve the same goal.


This study used the methods detailed in the BIOSCI 329 2013 Field Course Guide. The fishes examined consisted of a female O. pullus (FL = 380 mm) and a male G. tricuspidata (FL = 292 mm).


Alimentary tissue and contents

The alimentary track of O. pullus had a total length of 753 mm and was composed of a single, roughly z-shaped intestine and lacked a stomach. Including the ingested material, the weight of the intestine was 128.96 g. The contents of the intestine were dark green in color, resembled a paste, and had a slick and slightly chunky texture. The mass of intestinal tissue and contents are detailed in figure 1 below. The total fish weight was 912.2 g and the gutted weight was 726.3 g.

The alimentary track of G. tricuspidata had a total length of 910 mm and was intricately folded upon itself. A bulbous gastric stomach was present at the anterior end of the track. The contents of the stomach was a thick reddish- brown paste strongly resembling faeces of terrestrial mammals in shape. Separation revealed small bits of red algae, which appeared uniform upon first inspection. The thickness of the intestinal contents diminished posteriorly until the matter was a thick brown liquid near the anus. The mass of intestinal tissue and contents are detailed in figure 2 below. The total fish weight was 540.2 g and the gutted weight was 485.5 g. 

Table 1. Mass of intestinal tissue and contents of O. pullusTable 1. Mass of intestinal tissue and contents of O. pullusTable 2. Mass of intestinal tissue and contents of G. tricuspidataTable 2. Mass of intestinal tissue and contents of G. tricuspidataFigure 1. Alimentary contents weight (O. pullus)Figure 1. Alimentary contents weight (O. pullus)

Figure 1 shows the mean weight of the alimentary contents across each section from the entire class. There is a gradual decrease in the amount of ingested material as it progresses through the intestine. This decrease roughly begins in section III. The sampled specimen shows a decrease in intestinal matter posteriorly and is consistent with the class data. The decrease in hindgut material can be observed between sections III and IV.

In the class data, G. tricuspidata exhibited relatively high gut content in section I than in the remaining sections (fig. 2). This is no surprise since section I was the stomach and would contain most of the undigested material. There is a rapid decrease in gut content following section I. This is consistent with the sampled specimen, where a nearly 50% drop in gut content was observed in section II and decreased almost linearly through the remaining sections. 

Figure 2. Alimentary contents weight (G. tricuspidata)Figure 2. Alimentary contents weight (G. tricuspidata)

Upon examination of the gut content in O. pullus, brown algae was determined to be the dominate type consumed. Of the brown algae, only Carpophyllum species were observed. 50% of the ingested algae were Carpophyllum reproductive structures, and 18% was Carpophyllum thallae (total 68% Carpophyllum sp.). 20% of the ingested material was foliose red algae, and 4% was filamentous red algae. Red algae species present included Abroteia suborbiculare, Dasyclonium incium, Dipterosiphonia dentritica, and Ceramium apiculatum. Finally, 4% of the ingested material was the foliose brown alga Ecklonia radiata.

The gut content in G. tricuspidata was purely red algae. Foliose red algae made up 70% of the ingested material, and filamentous red algae made up the remaining 30%. The species present included Ceramium apiculatum, Abrateia suborbiculare, Ceramium diseordicatum, and Nancythalia humilis. 

Figure 3. Class gut content across speciesFigure 3. Class gut content across species

Figure 3 shows the amount of each type of ingested type of material. Rhodophytes appeared to have been ingested only in small quantities, and animal and chlorophytic material is negligible. G. tricuspidata consumed mostly rhodophytes. Only 10% of the consumed material where chlorophytes, and negligible animal material was consumed. No phaeophytes were ingested. In previous years, this trend was consistent in O. pullus, where the majority of consumed material consisted largely of phaeophytes and, to a lesser degree, rhodophytes (fig. 4). There is, however, a larger degree of variation in G. tricuspidata (fig. 5). It is apparent that the majority of the ingested material consists of rhodophytes and chlorophytes, but previous laboratories have documented a sizable amount of animal material in the gut. This does not reflect the findings of this study, but the fact that G. tricuspidata is a generalist feeder may account for this annual variation. Nonetheless, phaeophytes remain the least consumed material in each year. Despite the large amount of variation in G. tricuspidata, the gut contents of past studies are consistent with those found in the sampled species, in that O. pullus mainly feeds on phaeophytes and G. tricuspidata primarily feeds on rhodophytes and chlorophytes. 

Figure 4. Annual gut contents in segment I (O. pullus)Figure 4. Annual gut contents in segment I (O. pullus)

Figure 5. Annual stomach contents (G. tricuspidata)Figure 5. Annual stomach contents (G. tricuspidata)Figure 6. O. pullus jaw myology with the integument removed.Figure 6. O. pullus jaw myology with the integument removed.

Morphology of oral jaws and pharyngeal apparatus
The skeletal structure of the oral jaws in O. pullus is comprised of tightly-bound regions of articulation. The largest structures present are the premaxilla and the dentary that form a shape analogous to a beak. Both are relatively similar in size and are the densest bone structures observed in the oral region. The teeth on the premaxilla and dentary are small ridge-like structures and are serrated. (Clements & Bellwood, 1988). A single groove near the posterior ends of both the premaxilla and dentary serve as points of articulation. The premaxilla articulates with the maxilla, and the dentary articulates with the articulate. The maxilla is bound tightly to the premaxilla by the anterior maxillary-premaxillary ligament, resulting in limited movement of the maxilla relative to the premaxilla (Clements & Bellwood, 1988). The articulate is roughly triangular in shape, with the anterior descending process inserting into the dentary groove (Clements & Bellwood, 1988). The oral jaws and their associated bones are shown in figure 6. 

Figure 7. O. pullus jaw myology with the integument and A2 removedFigure 7. O. pullus jaw myology with the integument and A2 removed

There are four main muscles associated with dentary occlusion: the A1, A1β, A2, and A3. In O. pullus, the A1 is situated just below the eye and between the A3 and LAP muscles. The A1 is connected to the posterior region of the maxilla by the A1 tendon (A1t). The retraction of the A1t pivots the maxilla in a counter-clockwise direction, causing the curved anterior region to occlude with the premaxillary groove, thus allowing lifting and extension of the premaxilla. The A1β is located anterior to the A1, is slightly obscured by the A2, and converges with the A1t. The A2 is the largest oral muscle, extending from the articulate to the preoperculum. The A2t connects the A2 to the posterior region of the dentary. The A3 is located underneath (medially) the A2 and occupies much of the same space, albeit being smaller. The A3t connects it to the most dorsal arm of the articulate. The musculature of the oral region (with A2 excised) is shown in figure 7 below.

The oral skeletal morphology of G. tricuspidata is slightly different from O. pullus. The largest structures are the premaxillary and dentary, which are rounded and contain small tricuspid teeth that are raspy in texture. The premaxillary articulates with the maxillary in a pivoting fashion, where the maxillary can pivot the anterior arm into an L-shaped groove in the premaxillary. The articulation between the dentary and the articulate is very rigid in that the roughly triangular articulate rests tightly against the L-shaped dentary groove. The skeletal structure is shown in figure 8 below. 

Figure 8. G. tricuspidata jaw myology with integument removed.Figure 8. G. tricuspidata jaw myology with integument removed.

Figure 8: G. tricuspidata jaw myology with integument removed. G. tricuspidata utilize 3 main muscles in dentary occlusion: the A1, A2, and A3. Unlike O. pullus, the A1β is absent. The majority of the superficial musculature is composed of a complex of the A1 and A2. The A1 rests on the dorsal margin of A2, and both fill most of the region below the eye. The A1t runs along the dorsal margin of A1, connecting to the posteroventral arm of the maxillary. Contraction causes the dorsal arm to pivot into the premaxillary groove, where articulation occurs. The A2t extends from the anterior end of the A2 to connect with the dorsal arm of the articulate. The A3 rests underneath (medially) the A2. The A3t connects with the lower arm of the dentary and articulates it directly. The musculature of the oral region (with A2 excised) is shown in figure 9 below. 

Figure 9. G. tricuspidata jaw myology with integument and A2 block removedFigure 9. G. tricuspidata jaw myology with integument and A2 block removed

The pharyngeal jaws of O. pullus are partially obscured by four ceratobranchials and by muscular tissue. The lower pharyngeal bone (LPB) is visible immediately posterior to CB 4 and is a formed by the unification of the fifth ceratobranchials (Clements & Bellwood, 1988). The LPB is connected to the neurocranium by the levator posterior (LP), the postpharyngeal ligament (PPL), the pharyngocleithralis externus (PCE), and the pharyngohyoideus (PH). The levator posterior ascends to the dorsal surface of the neurocranium and is the largest of the muscular bundles. The postpharyngeal ligament extends laterally from the posterior edge of the lower pharyngeal bone and attaches to the cleithrum (Clements & Bellwood, 1988). The pharyngocleithralis externus attaches the posteroventral region of the lower pharyngeal bone to the cleithrum, passing the pharyngohyoideus which extends from the urohyal to attach to the lower pharyngeal adjacent to the pharyngocleithralis externus (Clements & Bellwood, 1988). These muscles are shown in figure 10. The excision of the ceratobranchials reveals the pharyngeal complex (figure 11). The levator internus (LI) connects the anterodorsal region of the upper pharyngeal bone material for each species from the entire class. There were significantly more phaeophytes in the gut of O. pullus than any to the prootic region of the neurocranium and is the only muscle connecting with the upper pharyngeal (Clements & Bellwood, 1988). The pharyngeal bones are roughly triangular in shape and have rough tooth-like surfaces along the pharyngeal tract.

Figure 11. O. pullus pharyngeal morphology with ceratobranchials,epibranchials, and associated musculature removed.Figure 11. O. pullus pharyngeal morphology with ceratobranchials,epibranchials, and associated musculature removed.

Like the butterfish, G. tricuspidata has four ceratobranchials that obscure the pharyngeal bones. The dorsal regions of each ceratobranchials are connected by individual levator externus bundles, numbering four in total, which connect to the dorsal surface of the neurocranium. The pharyngocleithralis internus (PCI) extends laterally from the posterior surface of the lower pharyngeal bone to the cleithrum, largely occupying the same space as the postpharyngeal ligament did in O. pullus. Like O. pullus, the lower pharyngeal bone is actually a specialized structure formed by the unification of the fifth ceratobranchials (Liem & Greenwood, 1981). The upper pharyngeal jaw does not have any significant muscular control and is largely free-floating (Liem & Greenwood, 1981). The pharyngocleithralis externus and pharyngohyoideus both connect the urohyal region with the ventral surface of the lower pharyngeal bone as they did in O. pullus, but they are, however, larger in diameter and more elongate, and the pharyngocleithralis externus extends more ventrally in G. tricuspidata (Liem & Greenwood, 1981). A levator posterior is absent in this species. The pharyngeal morphology is shown below in figures 12 and 13. 

Figure 12. G. tricuspidata pharyngeal morphologyFigure 12. G. tricuspidata pharyngeal morphologyFigure 13: G. tricuspidata pharyngeal morphology with ceratobranchials, epibranchials, and associated musculature removed.Figure 13: G. tricuspidata pharyngeal morphology with ceratobranchials, epibranchials, and associated musculature removed.


Alimentary tissue and contents

The high amount of phaeophyte material ingested by O. pullus indicates it as a preferential food option over both chlorophytes and rhodophytes. The presence of Carpophyllum species and Ecklonia radiata further suggests, in accordance with Clements & Bellwood (1988), that they are the primary sources of nutrition in O. pullus. Rhodophytes are occasionally consumed in adults, but are usually consumed only by juveniles of the species (Clements & Choat, 1993). Reduced ingestion of E. radiata in this sample is indicative of a higher preference for Carpophyllum species.This is supported in that the majority of Carpophyllum ingested were reproductive structures, which are high in glucose and are thus more nutritious to the fish than the thallae of E. radiata and Carpophyllum sp. It is possible that the ingestion of Carpophyllum thallae was for efficient supplication since the fish would have already been browsing the Carpophyllum. This is in contrast to the gut contents of G. tricuspidata, which was entirely rhodophytic in composition. The nutritional value in rhodophytes is in their storage polysaccharides, which are true starch and floridean glycogen (Clements & Choat, 1997). These compounds are readily digestible by most fishes, but laminarin and mannitol, the storage polysaccharides of phaeophytes, are not (Clements & Choat, 1997). The β-linked polysaccharide chains of phaeophytes can be digested in odacid fishes (i.e. O. pullus) because of the presence of symbiotic bacteria in the gut (Clements & Choat, 1997; Seeto et al., 1996; White et al., 2010). Mannitol can make up more than 50% of the frond weight in phaeophytes, and thus can be a significant food material for fishes that can ferment this molecule in their guts (White et al., 2010). In O. pullus, the ingested material progressively had a liquid state starting in segment III of the intestine. This corresponds with the increase in bacterial fermentation in the hindgut (White et al., 2010). This also corresponds with the abrupt decrease in gut content between segment III and IV. Due to the rhodophytic composition of the gut contents in G. tricuspidata, it is unlikely that significant amounts of mannitol were present. The progressively liquid state of the ingested material throughout the intestine is likely due to the breakdown of algal cell walls via the high pH of the gut (White et al., 2010). Much of this was likely to have been initially digested in the stomach, which had the highest amount of material, and then passed on through the intestine for further digestion and subsequent absorption.

Morphology of oral jaws and pharyngeal apparatus

The skeletal structure of the oral jaws are reflective of the individual’s diet and behaviours while feeding. In O. pullus, the premaxilla and dentary are beak-like in shape and contain serrated teeth functionally adapted for cutting (Clements & Bellwood, 1988). When feeding on thallae, O. pullus uses opercular suction to draw the algae into the oral cavity, where the bite would cut away a disk for ingestion (Clements & Bellwood, 1988). When feeding upon Carpophyllum reproductive structures, a sideways flick of the head following the bite would effectively remove the structure from the algae (Clements & Bellwood, 1988). In G. tricuspidata, the tricuspid teeth are adapted for the combing and scraping of substrate (Ferry-Graham & Konow, 2010; Yagishita & Nakabo, 2003). Girellids are capable of opening their jaws widely to maximize contact with the desired food material (Ferry-Graham & Konow, 2010). This allows for more material (i.e. algae) to be scraped off of the substrate for ingestion. This method of feeding likely contributes to the variation in ingested material in that the scraping of any given substrate may result in the consummation of any undesirable material, such as hydrozoans, in the immediate vicinity. This feeding mechanism in Girellids is also enhanced by the power of their bite (Ferry-Graham & Konow, 2010). The presence of an intermandibular joint (IMJ) connecting the anterior arm of the articulate with the dentary groove not only allows for increased gape expansion, but also to increase dentary closure strength when under flexion (Ferry-Graham & Konow, 2010). When the jaws open, the dentary rotates around the anterior arm of the articulate to which it connects (Ferry-Graham & Konow, 2010). The closure of the jaws is then derived from the retraction of the A2, which attaches to the articulate, and the A1, which attaches to the maxilla (Ferry-Graham & Konow, 2010). It is during this process that the dorsal rotation of the dentary around the IMJ generates a high velocity to increase the strength of the bite (Ferry-Graham & Konow, 2010). This allows for greater control of contacting the substrate and increased collection of food material (Ferry-Graham & Konow, 2010). The bite forces exhibited in G. tricuspidata exceeds those of O. pullus, which lacks an intermandibular joint. Furthermore, the biting force in O. pullus is reduced due to relatively thin bone structure in the dental plates and in the maxilla, in addition to a small adductor mandibulae complex (Clements & Bellwood, 1988).

The morphology of the pharyngeal jaws indicates a developed mastication process in O. pullus. The lower pharyngeal bone is tightly bound to the cleithrum by the post-pharyngeal ligament, which anchors the structure against the opposing musculature (Clements & Bellwood, 1988; Liem & Greenwood, 1981). The principle muscle involved in the labroid mastication process is the levator posterior, as it maintains muscular support for the lower pharyngeal bone (Liem & Greenwood, 1981). Further control is gained by the opposite forces directed by the pharyngocleithralis externus (Liem & Greenwood, 1981). This differs from G. tricuspidata in that the levator posterior muscle is absent. Without this muscle, there is little dorsal support in the mastication process, thus the forces exerted by the pharyngeal jaws are significantly reduced compared to O. pullus. Indeed, the gut contents of O. pullus were finely masticated and much of the algae’s original structure was deformed. In G. tricuspidata, the gut contents appeared to have been ingested without major alteration and they largely retained their physical structure, which indicates that the fish scrapes food material into the body without oral processing.

In summary, the differences between herbivorous fish diet and behaviour are functions of morphology. Both O. pullus and G. tricuspidata exhibit morphological adaptations that maximize food and energy intake that reflect their evolutionary development. As in the case of these two species, these evolutionary adaptations can lead to different dietary specializations within a shared habitat. 


Clements, K. D., & Bellwood, D. R. (1988). A comparison of the feeding mechanisms of two herbivorous labroid fishes, the tem perate Odax pullus and the tropical Scarus
rubroviolaceus. Marine and Freshwater Research, 39(1), 87-107.

Clements, K. D., & Choat, J. H. (1993). Influence of season, ontogeny andtideonthedietof thetemperatemarineherbivorous fish Odax pullus (Odacidae).Marine Biology, 117(2), 213-220.

Clements, K. D., & Choat, J. H. (1997). Comparison of herbivory in the closely-related marine fish genera Girella and Kyphosus. Marine Biology, 127(4), 579-586.

Ferry-Graham, L. A., & Konow, N. (2010). The intramandibular joint in Girella: a mechanism for increased force production?. J ournal of morphology, 271(3), 271-279.

Liem, K. F., & Greenwood, P. H. (1981). A functional approach to the phylogeny of the pharyngognath teleosts. American zoolo gist, 21(1), 83-101.

Seeto, G. S., Veivers, P. C., Clements, K. D., & Slaytor, M. (1996). Car bohydrate utilisation by microbial symbionts in the marine herbivorous fishes Odax cyanomelas and Crinodus lophodon. Journal of Comparative Physiology B,165(7), 571-579.

White, W. L., Coveny, A. H., Robertson, J., & Clements, K. D. (2010). Utilisation of mannitol by temperate marine herbivorous fishes. Journal of Experimental Marine Biology and Ecology, 391(1), 50-56.

Yagishita, N., & Nakabo, T. (2003). Evolutionary trend in feeding habits of Girella (Perciformes: Girellidae). Ichthyological Re search, 50(4), 358-366. 


Eukaryon is published by students at Lake Forest College, who are solely responsible for its content. The views expressed in Eukaryon do not necessarily reflect those of the College.

Articles published within Eukaryon should not be cited in bibliographies. Material contained herein should be treated as personal communication and should be cited as such only with the consent of the author.