Suplement XLV

Ichthyologica: The Aquarium Journal, Volume 39(2): 57-70
Author, San Francisco Aquarium Society.
Publisher, T.F.H. Publications, 1967.

April-June 1967


 George W. Barlow

Department of Zoology
Museum of Vertebrate Zoology
University of California
Berkeley, California

It is the habit of illustrators of fishes to present their subjects as seen from the side. This is reasonable because fishes, being typically  compressed, are most distinctive viewed in that way. Yet many fishes have markings on their heads that can only be appreciated in an adaptive context when seen from the front. Baerends and Baerends-van Roon (1950) created some awareness of this when they studied the hostile displays of cichlid fishes. They illustrated (p. 43, fig. 14) the frontal displays of certain species as they would be seen by their opponents, facing with the opercles extended; then the significance of the opercular flaps as large pseudo-eyes was fully appreciated. In a similar fashion, Abel (1964) figured the masks on the fronts of the faces of various male blennies.

Many other special signals would be discovered if more fishes were viewed from the front. As an example in passing, certain surf perches (Embiotocidae) have white lower lips. When seen from the side this feature is hardly noticed. But when a school turns and faces a diver theeffect is dramatic, his eye being deflected to the many white spots in the water.

The focus of this article, however, is on the striking interorbital stripe as exemplified by a South American leaf fish, Polycentrus schomburgkii Muller and Troschel. An argument will be advanced for its function. This argument will be extended to other fishes evincing similar color patterns. Finally, some exceptions will be described.

Acknowledgments: I would like to take this opportunity to thank the people the meeting of the American Society of Ichthyologists and and Herpetologists who responded so enthusiastically (see below). For a stimulating discussion of the paper presented there I am particularly grateful to John Magnuson and Henry Feddern. Edmund S. Hobson generously made available his knowledge and photographs garnered from underwater observations. Karel Liem offered supplemental observations on the feeding behavior of another leaf fish; Thomas Frazzetta, John Mertz, and Peter Skaller read the manuscript. The basic observations were made under the auspices of a post-doctoral fellowship (MF 8244) from the National Institute of Mental Health. This was undertaken at the Max-Planck-Institut fur Verhaltensphysiologie; I am eternally indebted to Konrad Z. Lorenz who made this possible, and who inspired me to learn about the behavior of animals. The writing itself was done under the support of a grant (GB 2210) from the National Science Foundation.

MATERIALS AND METHODS: The conditions under which the Polycentrus were kept are of little consequence to the behavior to be described here. Suffice it to say that the fish were kept in well-planted aquaria with conditions of water, temperature, and light appropriate to their natural environment. Details can be found in another article, one dealing with the social behavior of Polycentrus (Barlow 1967).

This paper was presented at the annual national meeting of the American Society of Ichthyologists and Herpetologists, during June, 1965, in Lawrence, Kansas. Taking advantage of the assembled body of knowledge present in the audience, I invited further examples of the kind of color pattern to be discussed in what follows (designated A.S.I.H.). For some time thereafter I continued to receive comments through the mail from ichthyologists who had heard about the paper; these will be marked pers. comm. Recently I observed several examples at the Steinhart Aquarium in San Francisco.

OBSERVATIONS: For approximately three months the diet of the Polycentrus had consisted largely of small amphipods, tubificid, worms, and insect larvae. There was nothing remarkable about the feeding behavior during that time. The fish simply swam to the object to be eaten, paused, fixed on it, and engulfed it. Then the fish were fed a school of young cichlids, Etroplus maculatus (Bloch). These small cichlids were difficult for the Polycentrus to catch by comparison with their previous food items. The cichlids hid among the plants, and fled and dodged artfully when approached by the Polycentrus. For the first time I observed what is evidently the full feeding response of Polycentrus. An idealized account follows, but it is one based on repeated observations of feeding behavior.

When hunting, Polycentrus drifts slowly toward its intended prey. The only moving parts of its body are the lobes of the dorsal and anal fins and the pectoral fins; occasionally the caudal fin may also contribute to the locomotion. All of these structures are transparent. As the Polycentrus glides toward the prey the median fins and the paired pelvic fins become fully extended. A slight partial flexing and extending of the spines of the dorsal fin can sometimes be seen. When about one to two body lengths away from its prey, the fish pauses. A slight lateral curvature forms in the region of the caudal peduncle, imparting what I have termed a sickle-shape in another article (Barlow 1961).  The Polycentrus then darts forward, engulfing the prey if successful. The lunge finishes in a sharp turn to one side, starting at about the point where the prey is grasped. Details of the strike cannot be observed because of its extreme speed. Both sexes may show various white and black spots on this dark background (Fig. 1).  The head in particular becomes nearly black, although the heavy black markings that run through it  are still apparent. In juveniles the body may remain pale while the head darkens.

Fig. 1 — Side view of a foraging Polycentrus schomburgkii showing the spotted pattern, the darker top of the head, and  the
inconspicuousness of the interorbital stripe  seen from this angle. Photo by the author.

The singular feature of the color change is that which takes place on the top of the head. A cream-colored stripe develops, running from the junction of the premaxillaries, back between the eyes, and up the nape, ending at the insertion of the dorsal fin. It is narrowest, less than one eye diameter, on the snout, widening to a little more than an eye diameter on the nape. In frontal view (Fig. 3), Polycentrus gives the appearance of a vertical black elipse, split by a white stripe running from its middle up to its top, and having two thin white stripes leading the pelvic fins.

DISSUSSION: More than any other single effort, Cott’s book (1940) on adaptive coloration has persuaded biologists to accept the principle of disruptive coloration in animal camouflage. His arguments of interest to us on this point can be summarized in a few lines: „It is this continuity of surface, bounded by a specific contour or outline, which enables us to recognize any object with whose shape we are familiar.” (p. 48).

So if Polycentrus is to minimize detection by its prey, it must break up its outline in some way.

Fig. 2.  — Front view of a foraging Poloycentrus schomburgkii showing the extension of the white stripe onto the edge of the dorsal fin.
Photo by the author.

Fig. 3 — Front view of o lurking Polycentrus schomburgkii moments before  engulfing a young Etroplus maculatus (slightly out of
focus, directly before the mouth of the Polycentrus). Photo by the author.

Again, as Cott put it on page 52, „Broadly speaking . . . we may say that white marks on dark animals living in dark surroundings (such as forests) . . . will be the most effective in breaking up the continuity of their surface, and in masking by contrast tell-tale half-tones of surface structure and modelling.”On pages 48 to 49 Cott discussed a frog (Plate 9) that has a prominent mid-dorsal stripe, in many respects the functional counterpart of the interorbital stripe in Polycentrus. The emphatic stripe was said to achieve its effect in three ways: (1) it stands out, detracting the eye from the profile of the frog; (2) this strong incidence of color tends to flatten the half-tones by which the frog is recognized; and (3) the stripe bisects the frog — two half frogs do not resemble a frog.

In Cott’s logic, the interorbtial stripe of Polycentrus serves to break up its characteristic frontal profile. Evidently it is an adaptation to reduce detection while slowly approaching its prey.

In addition to the frog, Cott also discussed some insects (p. 79) that employ median stripes for disruptive effects. (His comments here about coincident disruptive patterns may also be extended to Polycentrus but need not concern us.) In these instances there is no suggestion of camouflage for predation, but just for protection. For the sake of economy the discussion should be confined to fishes, yet one instance among mammals is especially worth noting. Charles Long drew my attention to the case of the badger. It is a predaceous burrower (Richardson 1929) with a split-head color pattern, and one that may lie in ambush for ground squirrels (Balph 1961).

Illustrations of fishes with disruptive stripes running up the interorbital space are difficult to find, but are not rare. Not only are fishes conventionally shown in side view, but they are usually drawn from dead specimens. If the pattern is seen only in the live fish, as in Polycentrus, even a figure depicting the frontal aspect will be of no avail. The grass pickerel (Esox americanus vermiculatus LeSueur) bears a permanent split-head pattern (Crossman 1966; pers. observ.), as must fishes that retain the interorbital stripe as museum material, e.g., Acanthoplesiops indicus (Day) (Smith 1950, fig. 409).

 Longley (in Longley and Hildebrand 1941) photographed in nature two species with disruptive interorbital stripes, a trumpet fish Aulostomus maculatus Valenciennes, and a Nassau grouper Epinephelus striatus (Poey). An unidentified marine eel also has such an interorbital stripe in life (Roughly 1951, pl. 40) as does the cleaner wrasse Labroides dimidiatus (Clupaty 1963, fig. 3).

The presentation of this paper at the meeting of the American Society of Ichthyologists and Herpetologists stimulated the following further examples: The split-head pattern is found in the pomocentrid fish Amphiprion akallopisus Bleeker, and in some long-snouted chaetodontid fishes (L. P. Woods). It has also been seen (D. P. DeSilva) in the tripletail, Lobotes surinamensis (Bloch). Further examples are found in eelpouts of the genus Lycodes occurring in the Bering Sea (N. Wilomovsky), in at least three gobiesocid fishes (J. C. Briggs), and in some gobies from the Philippine Islands (G. S. Myers). Preserved specimens of a sunfish, Pomoxis nigromaculatus (LeSueur), have been found with  a „. . . distinct brown line extending from the gular region up across the snout onto the dorsal spines. This bisecting line is not characteristic of most of the members of this species . . .” (W. H.  Walker, Jr., A.S.I.H. and pers. comm.). Finally, isospondylous fishes of the genus Dixonina lack the split-head color pattern but do have a conspicuous pale spot on the tip of the snout which may serve a similar end (F. H. Berry). A visit to the Steinhart Aquarium in San Francisco resulted in several variations on the same theme. In the large squirrel fish Holocentrus spinifer (Forskål), the forehead is lighter red than the rest of the fish and is split by a brown-red line running from the mouth to just behind the interorbital space. The split-head effect is extremely pronounced in the lobotid fish Datnioides microlepis Bleeker; the leading edges of the pelvic fins are white, the posterior portions black.

Fig. 4 — A Nassau grouper Epinephelus striatus. Photo courtesy T.F.H. Pub.

Fig. 5 — Datniodes microlepis. Photo by Peter Tsang.

Fig. 6 — Promicrops lanceolatus. Photo by H. Hansen.


Adults of the serranid fish Grammistes sexlineatus (Thunberg) are black with many white stripes and lack an interorbital stripe; it is present in the younger specimens, however. A reef-dwelling jack, Caranx sp., has a thin but distinct median stripe starting just above the eyes and running to the top of the head; I tried to view this species head on, but the fish always dropped down and looked slightly up at me, which placed the stripe in the middle of the obliquely-presented frontal profile. Two striking examples were seen among the anemone fishes of the genus Amphiprion: In both species a white middorsal stripe runs the length of the orange fish; in A. akallopisus Bleeker the stripe starts at the mouth, but in A. perideraion Bleeker it starts further back at the interorbital space. The stripe bisecting the face of the long-snouted  butterfly fish Chelmon rostratus (Linnaeus) is noteworthy for its border; the face is clear white and the stripe is gold, but set off from the white by a dark brown margin. A short but distinct split-head stripe was also observed in a pipefish Syngnathus griseolineata (Ayres), a klipfish Gibbonsia elegans (Cooper), a file fish Monacanthus sp., and in an agonid fish. Among several eel bennies only one species, the pholid Apodichthys flavidus Girard, had a split-head pattern, a pronounced thin white line running middorsally from the mouth to the tail; the same pattern, however, was noted in another pholid, Xererpes fucorum (Jordan & Gilbert), seen at the Bodega Marine Laboratory of the University of California.

In most of the cases just cited the fishes are thought to be predaceous, but useful observations are not readily available. In the following some examples are given where the split-head pattern has been seen, together with observations on the feeding habits of the animals.

The feeding behavior and color patterns of another leaf fish, and a closely related one, Monocirrhus polyacanthus Heckel, has been described by Cott (1940: 311-313, fig. 63). Cott dwelled on the appearance of the side of this fish in relation to its hunting behavior, although the prey must view the front end of the leaf fish (pers. observ.). The side view of Monocirrhus in Cott shows the pale interorbital region, although he failed to mention it. Upon hearing the presentation of the paper in hand, G. S. Myers ventured the suggestion that this leaf fish may be so thin that the split-head pattern could not function because the white stripe would occupy the entire frontal aspect of the head. With this in mind I asked Karel Liem to observe his leaf fish to establish whether or not this is so. He reported that the split-head pattern operates perfectly well in Monocirrhus. In fact, when it feeds the head may blacken around the interorbital stripe, increasing the split-head effect, while the color of the rest of the body may remain relatively unchanged.

I have observed in a related nandid fish, Nandus sp., that a split-head pattern appears during feeding. Polycentropsis abbreviata Boulenger, is an African nandid fish and a species that looks much like Polycentrus; it also manifests the split-head pattern when feeding (G. S. Myers, A.S.I.H.). A further example is the serranid fish Promicrops lanceolatus (Bloch). The young are black with yellow blotches and have a pronounced split-head marking. This species has been characterized as a lurking predator (Clupaty 1963, fig. 2).

Dermatolepis punctatus Gill is another serranid fish, having a split-head pattern (fig. 7). Hobson (pers. comm.) observed that this grouper is secretive by day but that comes out to feed during the crepuscular periods. „It is usually active close to the bottom during these periods, yet it not infrequently ranges up into midwater. Although I have not observed  this fish feeding, it is a sluggish swimmer with a manner that certainly suggests it to be a stalking predator.” As in the case of Promicrops, the split-head pattern is more evident in the smaller fish (less than 350 mm.).

Fig. 7 — Dermatolepis punctatus in its lair. Photographed by E. S. Hobson near Laz, Baja,  California.

Hobson has also written to me about another serranid fish with a striking split-head pattern, the soap fish Rypticus bicolor (Valenciennes). It too is a stalking predator that is active  primarily during the crepuscular or nocturnal periods, eating mainly benthic crabs and crustaceans, although some fishes are taken.

Still another marine fish showing the split-head pattern is the giant kelpfish, Heterostichus rostratus Girard (Hobson 1965, figure on p. 510). Hobson (pers. comm.) found that H. rostratus stalks small crustaceans and fishes, moving very slowly and not darting forward for the capture until within a few inches of its prey.

The sternarchid fish Apteronotus albifrons (Linnaeus) is a velvety black animal having a creamy white stripe running dorsally from the snout back about one-third of the body length (pers. observ.). It is nocturnal, perhaps crepuscular in habits, feeding on small fishes, shrimp, and insect larvae (Walker 1965).

These examples, coupled with the observations on Polycentrus and Nandus, led me to believe that fishes with a disruptive interorbital stripe might be characterized collectively as lurking or slow-moving predators. This generalization, however, had to be modified in light of a report by Magnuson and Prescott (1966; my thanks to them for sending the unpublished manuscript).

Part of their study dealt with the bonito, Sarda chiliensis (Cuvier), a schooling pelagic fish. When fed, the bonito manifests 9 to 10 vertical black bars on the body. More germane, a prominent yellow stripe appears extending from the tip of the snout to the base of the first dorsal fin. Nakamura and Manguson (in Magnuson and Prescott) have observed nearly the same color change in the skip-jack tuna, Euthynnus pelamis (Linnaeus). In both of these oceanic species the bars and the interorbital stripe fade as the animals finish feeding. The male bonito also expresses these body bars and the interorbital stripe when engaging with another male in lateral display, a hostile act, or when maneuvering to spawn with a female. It is important to note here that bonito do not display frontally, probably because they cannot back up well.

Nakamura, Magnuson, and Prescott all endorse the idea that the primary signal value of this color change relates to hostile and spawning behavior. But they suggest it may have acquired the secondary significance of indicating the presence of food.

The bars may communicate something in a school of bonito, whatever the message sent. That the interorbital stripe has signal value, however, is unlikely. Seen from the side, as it must be most of the time by other bonito, it is a poor signal; its effect is only to flatten the head slightly. It must be viewed from in front to be of use, and only the prey regularly sees this aspect of the bonito. The split-head color pattern of the bonito probably serves to disrupt its frontal profile, just as it does in Polycentrus.

If this is so, then why does the interorbital stripe appear when the bonito is involved in social, nonfeeding behavior? The reason might be that this is the simplest solution for the bonito. Any major arousal could activate the color change, only one kind of change being necessary. Since the interorbital stripe is not seen well, it is of little consequence in intraspecific interactions.

Now in Polycentrus a similar situation prevails. When spawning they develop the interorbital stripe, strikingly so in the female. This, too, could be accounted for by general arousal. But when two males fight, the interorbital stripe is lacking, the space being darkly colored as is orbital stripe is lacking, the space being darkly colored as is the rest of the „face.”

The explanation is that Polycentrus engages in frontal displays when aggressive. A distinctive intraspecific signal, in this case the head plus extended opercles, should be relatively solid in color to emphasize the profile, and/or have an uninterrupted profile. An interorbital stripe would work against this because of the splitting of the head and disruption of the profile.

The spawning female, who must swim toward the male, takes advantage of the interorbital stripe to avoid the hostile signal that the uninterrupted face conveys. Similarly, the various female toothcarps of the genus Cyprinodon show a split-head color pattern in the context of sexual or aggressive behavior (R. K. Liu, pers. comm.).

Another indication of the role of arousal lies in the closer observation of feeding behavior. It is only when the prey is difficult to capture, or the Polycentrus very hungry, that the interorbital stripe appears. When they have been fed small fish a few days in a row, the stripe is turned on as the feeder enters the aquarium room. But lest the point be made too strongly, recall that arousal involving aggressive behavior inhibits the appearance of the interorbital stripe.

Thus the contrast-rich interorbital stripe, the split-head color pattern, seems to characterize certain predaceous fishes, particulary those lying in ambush. In some it occurs only when foraging. In others it is a fixed part of the color pattern.

This discussion of the split-head color pattern would be incomplete if I did not point out some exceptions. There are fishes possessing such a color pattern that one would not ordinarily conceive of as being predators. One of the best examples is presented by browsing marine angel fishes. Henry Feddern (A.S.I.H.) has pointed out to me this remarkable exception, based on his studies of these fishes in the Caribbean Sea. In Pomocanthus paru (Bloch) yellow encircles the mouth, then runs up from the circle through the interorbital space back onto the nape, reaching the insertion of the dorsal fin; the background color is dark brown or blue. In P. arcuatus (Linnaeus) there is also a yellow circle around the mouth; similarly a stripe runs up from the circle between the eyes and back to the insertion of the dorsal fin; but the stripe also extends ventrally along the chin. Juvenile fishes of this genus feed on sponges and algae; they are blue with a yellow split-head pattern. When either the juveniles or adults are approached they tend to face the intruder. Thus while they may not be using the split-head pattern as a concealment device  to allow them to approach prey, they may be using it to permit them to look directly at a potential predator while making themselves more difficult to locate visually.

I have observed another interesting exception in the loach Botia horae Smith. This is a pale-colored fish with a black line running from the tip of the snout back along the entire dorsal surface, curving down along the insertion of the caudal fin, and terminating at the ventral surface of the caudal peduncle (Sterba 1962, fig. 452). Seen from the side, it has a black line at the base of the short dorsal fin and a black bar along the base of the caudal fin. The habits of the fish could be summarized as follows: It nibbles on plants and along the bottom, probably finding its food largely by virtue of the taste buds on the small barbels around the mouth. Prepared commercial fish food is taken visually, however, as it falls through the water. It is a fish that might be characterized as an active, nervous swimmer, but it is not a lurking predator. On the other hand, it frequently chases and apparently disturbs large fish in an aquarium. It is also vicious in the way it persistently attacks members of its own species. Conceivably, it could be nipping mucus from large fishes, although I doubt this. Perhabs more significant, the snout is long and relatively flat. The fish is mostly on or near the bottom. Thus the split-head effect could be part of a more general pattern which is to bisect the fish altogether when seen from above. (Anableps dowei Gill is a dark fish with a pale middorsal stripe running from the back of the head to the tail; living just under the water surface, it would often be seen from above by potential predators such as birds.) The black at the insertion
of the dorsal and caudal fins of B. horae probably serves in species recognition.

The pomocentrid fishes of the genus Amphiprion do not fit the scheme of a lurking predator. While carnivorous, they tend to hover, dart and turn, occurring together in small groups in the open. The clue is that the several species differ most immediately in the distribution of a few bold white marks on an orange to brownish background. The number of possible placements of these few white marks on the sides of the bodies of the species must reach some limit set by the abilities of the species to distinguish them.

Having a white stripe in the middorsal area increases the number of possible combinations. Since members of the species move slowly, hovering and facing one another, recognition from the front presents no problem. A parallel pattern of species-specific markings can be seen in the white discs on the foreheads of pomocentrid fishes of the genus Dascyllus; they are remarkably similar in behavior.

Two further comments need to be made. The first is to point out that in three serranid fishes (Grammistes sexlineatus, Promicrops lanceolatus, and Dermatolepis punctatus) the young animals have the split-head pattern while it is lacking in the adults. Knowledge of any associated change in feeding habits would illuminate the significance, or lack of it, of the split-head pattern in these fishes.

The final comment is to draw attention, though fleetingly, to the problems of phyletic and ecologic correlations. Why is it that some kinds of ambush predators are not known to show the split-head pattern? A  good example is the large genus Sebastodes. I suspect this is simply another example of different phyletic groups having different solutions to the same problems. Also, other considerations may prevail, such as species recognition. It is also my impression that the split-head pattern is more widely distributed among marine fishes than among freshwater fishes. This may be no more than a consequence of the great species diversity of the tropical shore fishes.

In conclusion, whenever the typical thin splitting coloration is seen, one can probably be safe in assuming first that it is for the purpose of concealment. When this is on the front end of the fish, chances are that the fish is a predator, and more likely of the lurking, ambush type than an active feeder such as the bonito. In some instances, a fish will have a split-head color pattern in order to observe a potential predator while remaining unseen. In a few species it may even be important as an intraspecific signal. And finally, some fishes, and particularly those that tend to feed on the bottom or near the surface in open areas, may have a dorsal line splitting them along their length to better conceal them.


Abel, E. F., 1964. Freiwasserstudien zur Fortpflanzungsethologie zweier Mittelmeerfische, Blennius canevae Vinc. und Blennius inaequalis C. V. Z. Tierpsychol., 21: 205-222.

Baerends, G. P., and J. M. Baerends-van Roon, 1950. Introduction to the ethology of cichlid fishes. Behavior, Suppl. 1: 1-242.

Balph, D. F., 1961. Underground concealment as a method of predation. J. Mamm., 42: 423-424.

Barlow, G. W., 1961. Ethology of the Asian teleost Badis badis. I. Locomotion, maintenance, aggregation and fright. Trans. III. State Acad. Sci., 54: 175-188.

— 1967. Social behavior of a South American leaf fish, Polycentrus schomburgkii, with an account of recurring pseudo-female behavior. Amer. Midl. Nat., in press.

Clupaty, P., 1963. Promicrops lanceolatus (Bloch). Aquar. u. Terrar., 10: 202-204.

Cott, H. B., 1940. Adaptive coloration in animals. Methuen, London, 508 pp.

Hobson, E. S., 1965. Forest beneath the sea. Animals (London), 7: 506-511.

Crossman, E. J., 1966. A taxonomic study of Esox americanus and its sub-species in Eastern North America. Copeia, 1966: 1-20.

Longley, W. H., and S. F. Hildebrand, 1941. Systematic catalogue of the fishes of Tortugas, Florida, with observations on color, habits, and local distribution. Pap. Tortugas Lab. Carnegie Inst., 34: 1-331.

Magnuson, J. J., and J. H. Prescott, 1966. Courtship, locomotion, feeding, and miscellaneous behavior of Pacific bonito, Sarda chiliensis. Anim. Behav., 14: 54-67.

Richardson, J., 1829. Fauna Boreali-Americana. Murray, London, 300 pp.

Roughly, T. C, 1951. Wonders of the Great Barrier Reef. Angus and Robertson, London, 13th ed., 283 pp.

Smith, J. L. B., 1950. The sea fishes of Southern Africa. Central News Agency, Capetown, 550 pp.

Sterba, G., 1962. Freshwater fishes of the world. Longacre, London, 878 pp.

Walker, B., 1965. Black ghost fish. Aquar. Jour., 36: 452-454.