Cichlid Room Companion

The State of Confusion in Discus Taxonomy

By & Marc Weiss, 1996. image

Classification: Taxonomy and phylogeny, South America.

Guest comment by Bob Goldstein.

In a paper that appeared in Tropical Fish Hobbyist in 1991, Dr. Warren Burgess posed two questions on the taxonomy and systematics of Discus: 1. Could the Symphysodon discus Heckel found outside of the Rio Negro be evidence for a much greater historical range of this species? 2. Could S. discus willischwartzi be a natural hybrid between the two discus species, viz. S. discus and S. aequifasciata Pellegrin?

We have in our possession wild fish that have the body coloration of a blue or brown Discus (S. aequifasciata) but also the wide accented 5th bar as is seen in the so-called "Heckel" Discus. What do these signify in the taxonomy of Discus? Are they natural hybrids of the two species or do they represent a new species or subspecies of the genus? In this article we want to examine these questions and then begin to describe the techniques involved in answering them.

Symphysodon aequifasciata
Symphysodon aequifasciata
Symphysodon aequifasciata
Symphysodon aequifasciata

These Symphysodon were all caught in the Rio Madeira and represent intermediate forms between S. discus and S. aequifasciata. Photos by Lo Wing Yat "Sunny".

Let us first look at the present status of Discus. At the present time two species of Symphysodon are recognised with the taxonomy based almost exclusively on colour and locality. S. discus has 9 vertical bars on the body with the 1st, 5th, and 9th bar being predominant, and the 5th the widest. The other bars tend to be very faint and at times not present at all. This species is found primarily in the tributaries of the Rio Negro. S. aequifasciata is found primarily in the western portion of the Amazon River drainage. Body counts (scale counts, number of fin rays, and number of vertebrae) tend to overlap. There are a few differences that will be discussed a little later. With only two species in the genus the taxonomy should be very clear-cut. Nothing could be further from the truth. The primary reason for this is that the full ranges of the two species and their described subspecies are still unknown. And in some cases the localities that have been published may be false. If you have found a beautiful wild Discus, would you want everyone else to know exactly where that fish came from? Also, locality data may be hard to actually pin down. One tributary looks like another in the Amazon basin. To pinpoint exactly where one morph is found may be close to impossible. This is one piece of the puzzle that we need to answer any questions on the taxonomy of Discus, and the information is hard to come by. Collections in every tributary of the Amazon basin will be needed to get a proper answer. This will take time, lots of time!

What is a species anyway?

This is a question that has perplexed science since the days of Linnaeus. You ask a number of people what a species is and you may get as many answers back. Science has recognised 5 basic species concepts. The Typological Species Concept says that species are entities that differ from all other entities. They have also been unchanged since the beginning of time. But this concept leads to much confusion in describing complex species: if a species has two different colour varieties, two different species names can be given to the two different forms.

Morphological Species are entities that differ sharply from others in structural characteristics (eg number of scale rows, number of vertebrae). How sharply do they have to differ before they will be considered two separate species? How many characters do you use? In reality, the more differences you find the more robust your argument is. So what about two characters? Is that enough?

For many years the standard species concept was the Biological Species Concept. This states that a species is a group of interbreeding natural populations that are reproductively isolated from other such groups. But what about those that reproduce by asexual means. There are many Poeciliids (livebearers) that reproduce by means of asexual reproduction (parthenogenesis). How can we relate these to this concept? Every individual could be considered a species. Also, how fertile or infertile do hybrids have to be? Many closely related species can produce some fertile offspring or offspring whose fertility is slightly reduced. How much fertility is needed to be considered a "good species"? So science had to come up with a concept that would take these species into account. That being the Evolutionary Species Concept. This states that a species is a lineage evolving separately from other species and with its own unitary evolutionary role and tendencies.

The concept that tends to be the most widely used nowadays is the Practical Species Concept. This concept says that species are those samples that a biologist can distinguish, and tell others how to distinguish. This tends to avoid the basic issues and says nothing about the mode of reproduction of the organism or how closely related it is to others in the group.

For sexually reproducing organisms one of the basic tests to determine if two species are indeed distinct is whether or not "naturally" occurring hybrids exist. If hybrids do exist, are these as fertile as the parental species? It has been known since the time Discus first appeared in the aquarium trade that, when crossed, the two species may produce fertile offspring. So again, is S. discus willischwartzi a natural hybrid? In order to test this using supposed hybrids, you must determine whether or not they have statistically the same reproductive viability as the parental species. If S. discus x S. discus have on average 150 viable offspring and S. aequifasciata x S. aequifasciata have on average 200, what does it mean if S. discus x S. aequifasciata have on average only 140 offspring? Is this enough of a reduction in viability to maintain the taxonomic status quo? Again, if the two species are indeed geographically separated, is there any reason for them to become totally reproductively isolated when there is no likelihood of them coming into contact with each other? Traditionally, these two species have appeared to be geographically separated from each other in their distribution: S. discus has been described primarily from the tributaries of the Rio Negro whereas S. aequifasciata is found in the western portion of the Amazon River drainage. But recent collections appear to put these original distributions in doubt.

Preliminary observations of Discus in their natural habitat have not yielded any evidence that more than one species of Symphysodon exists. These fishes appear to live in "family" groups and seem not to leave their immediate areas; the individuals in each group tend to look similar over a limited area. A short distance across open water in a suitable habitat, another group will be found that appears different. Different enough that they may be called something different. A similar observation can be made in other cichlids such as Mesonauta festivus and Heros severus. These differences arise because traits are much more easily fixed in a small breeding population. In biological terms this is referred to as the founder effect. If two populations differ slightly to begin with, then within a few generations, they may be dramatically different.

To give you an example using some simple numbers. You have one population of fishes with 20 individuals. This population then somehow gets separated into two. Sub-population 1 is composed of 6 typically blue fishes and 4 brown. Sub-population 2 has 9 browns and 1 blue. If each individual has an equal chance of breeding, within a few generations the two sub-populations will diverge even more. Thus if we collect the two sub-populations in, say 5 years, they may have diverged enough to call them different varieties; but in fact they are not.

But what about the fishes that appear to be hybrids? By taking time and making careful observations one can demonstrate that intermediate forms are not uncommon, living both between groups as well as among them; quite contrary to what most literature indicates.

Confusion abounds for hobbyists as well as scientists due to the fact that fishermen and exporters artificially categorize and sort Discus when they collect them. This is mainly to protect collecting sites.

Many well-intentioned scientists and aquatic journalists have been misled by these false collecting reports. Indeed, the S. aequifasciata haraldi Schultz (1960) based his description on is not typical of the type locality cited. The same is true for the so-called "Alenquer" Discus which is supposedly from a small town called Alenquer but very likely collected in the Rio Purus a vast distance away.

Heckel Discus without the central bar are also not hard to find. Similarly "Blue", "Red", and "Brown" Discus with a prominent central bar are quite common in some areas. "Hybrids" abound in certain drainages and are scarce in others. When collected, some are shipped as Heckels, others as Blue or Royal Blue, and some as plain old Brown Discus. Four different types; one forest stream!

Attempts have been made to categorize wild caught as well as domestic forms of Discus, but a few days in Amazonia will demonstrate that all the Heckel morphs purported to be from all over the Amazon can be found under one log! Blue heads, yellow fins, with or without a central bar, all in one batch of fishes. It is obvious that the idea of delineating subspecies on the basis of colour is more fantasy than good science. Some geographic variation does indeed exist and needs to be properly investigated, but it seems easier to prove it doesn't.

Interestingly, knowledge of biotic and abiotic factors that affect Discus can be used to "create" morphs even using one fish! Unilateral damage can result in a Discus with different coloration on either side, photographs of which appear to be of two different "subspecies."

By breeding these fishes in captivity, recombination of genes takes place and allows them to be saved by the breeder. In nature perhaps piranhas or even the parents themselves cull the spawn so that some of these "odd" specimens are lost. Not all are and occasionally they popped up in our nets. In captivity it is easier to observe "Green Discus" beget "Blue Discus" and Heckels produce Browns ad infinitum.

So, are there two species of Discus or just one? How do you go about establishing whether or not two fishes in your possession are the same or different species? If the two fishes are dead and preserved we have to use some method other than reproduction to determine whether or not we have more than one species. So how do we do it? There are number of techniques used in modern taxonomy. Some have been developed gradually since the beginning of fish taxonomy and some are very modern. The first type of comparison is by their meristic values, which are anything on the fish that can be counted: number of scale rows, vertebrae, lateral line pores, etc. Meristic values (body counts) are traditionally the preferred characters used in taxonomy of fishes because they tend to be the most reliable; they tend to change very little over time. With Discus a distinct picture does not emerge when utilizing meristic characters. Heckel's original description of S. discus was based upon only one specimen. When Schulz reevaluated the genus in 1960 he again relied on only one specimen of S. discus, but used more than 50 specimens of S. aequifasciata. Of the ten characteristics that Schulz used (see table) only one can be utilized to distinguish these two species: the number of scale rows from the operculum to the base of the caudal fin (table). If more specimens of S. discus had been evaluated would there be more variation in the counts and would this variation overlap with the variation observed in S. aequifasciata? This is a very important concept because all animals show variation. If you look at any species identification, the authors will always state such things as dorsal fin rays 9-14, meaning that the number of dorsal fin rays varies from 9 to 14 in members of the species. Thus demonstrating that not all individuals of the species have the same number of rays. But no variation can exist if only one individual is examined.

The second taxonomic method consists of measuring morphometric characters. These are measurements that can be made on a fish: total length, snout length, etc. Because fishes grow throughout their lives these measurements are usually expressed as ratios to standard length. There are many problems associated with this technique. A major problem is that in many species overall body size and shape (and hence ratios) changes throughout their lives. Comparison of different age groups may yield conflicting and confusing results. So it is best when making comparisons between fishes to compare specimens of approximately the same size ,just in case there is variation between different life stages. Another problem is that many species are sexually dimorphic: there are differences in overall body plan between males and females, so intersexual comparisons may also yield confusion. A third problem, especially with Discus, is that specimens from different types of habitat, such as flowing water or still backwaters, may have completely different body shapes. Hanel (1981) compared the different species and subspecies of Discus using morphometric characters. He found overlap in all of the characters examined. However, the pictures provided of the fish used in the research show the typical pointed-noses (longer than higher) and large-eyes that are indicative of weakened diseased fish. Apparently individuals used in this research were wild fish which had been kept in captivity for many months. Thus they may have lost their natural body shape due to inadequate diet.

Symphysodon from Rio Madeira
Symphysodon from Rio Madeira

In contrast to the previous reports, different colour morphs of Symphysodon are found sympatrically, questioning the vality of the subspecies and also that of the two nominal species S. discus and s. aequifasciata. Photos by Lo Wing Yat "Sunny".

The third taxonomic technique used involves anatomical characteristics: shape, completeness and position of the lateral line, position and size of internal organs, and secondary sex characteristics such as breeding tubercles. These characters can be definitive characteristics for a species but closely related species tend to have similar anatomical features. These characters are also very hard to quantify because there is so much variation between individuals of a single species.

Another group of characters which is sometimes used to distinguish species is colour patterns. These can be some of the most variable characters of all for any species. As we all know, there are many different colour varieties of wild Discus and the two different species have their own basic pattern. We have in our possession wild specimens of S. discus that have the front half of their body striated, but the back half (posterior to the fifth vertical bar) lacks striations and looks like that of a Brown Discus. What would you call these specimens, S. discus or S. aequifasciata? What do subtle variations between individuals signify? As stated previously, we also have wild S. aequifasciata with a very striated body and the accentuated 5th bar as in S. discus. What should we call these? Are they wild hybrids of the two species or are they a completely new species?

In one of the authors' (AIM) research laboratories, we are investigating the taxonomy of the two species of Discus. We are using 2 modern techniques to do this: karyotyping and electrophoresis. Both of these techniques involve the use of tissues and cells from fishes and tend to be labour intensive; they take time. In another 6 months to one year we should be able to finalise our results. In addition, there is need for more field work, as well as laboratory research. Perhaps some more breeding experiments. Environmental influences may need better definition. These studies are in progress. To our way of thinking, the currently accepted taxonomy of Symphysodon seems to be in error. In the end we may have to face being proved wrong as well. Time will tell. We will keep you informed.

So what does all this mean for the Discus breeder and hobbyist? Well, the fish that you have in your tank or in your breeding program may have the genes to produce an infinite varieties of Discus. It only takes time (much time) to bring those genes to expression.

S. discus * S. aequifasciata
Fin rays
dorsal (spinous)
dorsal (soft)
anal (spinous)
anal (soft)
Pectoral (unbranched)
Pectoral (branched)
Pectoral (unbranched)




Lateral line pores
total pores


No. scale rows 44 50-61
No. vertebrae




  • Anthony Inder Mazeroll, PhD, Deptartment of Life, Earth and Environmental Sciences, West Texas A&M University, Box 808, Canyon, TX 79016-0001, USA.
  • Marc Weiss; World Wide Fish Farm - 5935 Ravenswood Rd., E20 Ft. Lauderdale FL 33312 USA

Guest Comment by Bob Golstein

The Mazzeroll and Weiss article on discus taxonomy is a terrific summary of where we are and how we got here, with the clear implication that in being here, we're lost! The article suggests strongly that based on the current evidence, there can be no justification for retaining S. aequifasciata. But it is early days yet, as the authors note.

Two new techniques offer promise to clarifying discus and other cichlid taxonomy. The first and older system is specifically important for locating collecting sites over the broad geographic range of discus. It is the Global Positioning System and today operates over the entire world. Using a handheld unit, one can precisely determine location down to about 5 meters. When combined with a map generated for the region, it will be possible to clearly map every river and stream in the world and know with precision exactly where you are. Confusion of which water body you are in will be a thing of the past.

The second technique is being used now with darters, minnows, and other fishes. It is A statistical technique called PAUP, and allows much more precise analyses of any suite of characters than was possible using older statistical procedures. Today it can be used to compare far more kinds of data than ever before. Traditional analyses have compared bones, meristics, and variations in enzymes based on what they did and their electrophoretic mobility. Today, ichthyologists are looking a sequences of base pairs on a specific piece of DNA located in the mitochondrion. As one scientist noted, in the past we have been able to compare something like 40 characters, but with the new DNA techniques and PAUP analyses, we can handle 40,000! Similar techniques were used in the OJ Simpson trial, but the jury ignored or was incapable of grasping the evidence. In a few years, every school child will be familiar with the method and what it means. For the rest of us, it means fewer persons will be convicted wrongly, fewer criminals will go free, and most important of all, we'll know how many lines of evolution of discus are out there.

This article was originally published in "Cichlids Yearbook Volume 5, Cichid press; 1995; pp. 77-83. It is here reproduced with the permission of Marc Weiss, World Wide Fish farm.

References (3):


Mazeroll, Anthony Inder. (Jul 30, 1996). "The State of Confusion in Discus Taxonomy". Cichlid Room Companion. Retrieved on Dec 04, 2022, from: