The Species Concept

Biologists normally use what is called a "biological species concept". It is framed in different ways. At the organism level, it is defined as a set of actually or potentially interbreeding organisms: for example, all Homo sapiens can interbreed successfully with one another, but not with any other species.

At the genetic level, one talks of a gene pool: the set of genes that are contained within an interbreeding population. Thus all members of a species contribute to/are part of a common gene pool, which thus contains all the genetic diversity within the species.

However, in the real world, it is time-consuming and expensive to make the observations of organisms in their real habitat that would allow us to say with confidence that such-and-such a set of organisms really is a species. And in the fossil world, it is impossible.

So instead, most biologists and all paleontologists make a good-faith guess about the boundaries of the set of organisms they propose to name a species. Typically, the species is defined on the morphology it has, not on the genetics and behavior that is specified in the biological species concept. And when we use morphology, geologists and biologists are much on the same level, except that the biologists have soft parts as well as hard parts to work with. However, there are problems even for biologists working with living organisms:

Problems in applying the species concept to living organisms

Many organisms reproduce asexually, so we do not know whether they fit the concept or not. For example, bacteria usually clone themselves, and even when they do go through something like sexual reproduction, they do not exchange genes the way that most prokaryotes do.

This objection may apply even to "advanced" creatures. For example, some lizards and salamanders do not (ever) have sexual reproduction. Their populations are all females, and they lay eggs which have never been fertilized, and hatch as females which carry on the cycle. Among plants, asexual reproduction is even easier, as many plants reproduce by budding. Strawberries and Bermuda grass are easy examples, but dandelions also reproduce without sex, and redwoods can.

The Oregon junco is a small bird, Junco oreganus, that ranges from southern Alaska to Baja California. The gray-headed junco, Junco caniceps lives in the Great Basin; the slate-colored junco, Junco hyemalis, lives east of the Rockies, and the white-winged junco, Junco aikeni, lives in the Black Hills of South Dakota. In their particular regions, they look and behave like perfectly good species. However, there are narrow strips of territory where any two species overlap slightly. In those corridors they interbreed freely. So ornithologists have been forced to place all these birds into one species, the "dark-eyed junco", J. hyemalis, which happens to have at least four "subspecies". Local birders in central California typically still call our California bird the Oregon junco. This example is one of hundreds showing that the biological species is difficult to apply, even by gifted biologists looking at conspicuous and well-known birds.

We successfully defined the "dark-eyed junco" (above) as a single species. That wasn't a simple decision, because there may be more limited and local interbreeding between species, and here a "competent biologist" might prefer to keep the two species as species, and make up a different set of terms for the interbreeding sections of the population. Typically, individuals that contain some mixture of the species' genes are called hybrid individuals that live in a hybrid zone where the species overlap and interbreed. Usually hybrids are less fertile than "pure" members of the two species, so usually hybrids are not numerous and hybrid zones are narrow.

Hybrid zones are a problem. Although the parent populations are exchanging genes, they remain largely genetically distinct. There must be some sort of dynamic equilibrium between the adaptation of the species to their different habitats, despire some genetic exchange across the hybrid zone. Hybrid zones can probably tell us a lot about the relative genetic isolation of populations, and perhaps about mechanisms of speciation.

Obviously genes are being passed from one species to another, and obviously that gene flow does not result in homogenizing the two species into one (though one could imagine that sometimes it might).

Populations separated (or connected) by hybrid zones may be very distinct genetically, or very much alike, depending on the group involved. "Mimetic" butterfly species look very much alike: they may be separated only by a very few genes that give different wing patterns. On the other hand, the firebellied toads of Eastern Europe, Bombina bombina and B. variegata, interbreed freely over a hybrid zone 1000 km long even though the two species differ in coloration, life history, mating calls, habitat, enzymes, and mtDNA: they have probably been separating for 3-4 million years.

Obviously, it is not a good idea to breed with a member of another species: or is it? Nature being what it is, there is always going to be some situation where a "rule" breaks down. Take the example of two species of Swedish flycatchers: they normally live side by side as separate species, but under certain circumstances, it is advantageous to breed across the species "boundary", so of course, they do .

Ring Species
An older example, that featured gull species around the Arctic, turns out to be far more complex than the original observations suggested. See this Web essay with references.

One of the better examples mentioned there features birds from Asia: warblers from the mountains of Asia [story in San Francisco Chronicle, January 2001].

Species in Time

For biologists, species and populations are frozen: they are either species or they are not, and the morphological, genetic, and geographic separation between species can almost always be seen as a sharp boundary, easy to establish. Biologists typically don't bother to worry about species through time: why should they? They deal with living things.

But in paleontology we deal with time, and since evolution has happened, taxonomic boundaries must somewhere be crossed as one species evolves into another. The first bird hatched out of a dinosaur's egg, though it obviously was not much different from its parents, and it found plenty of contemporaries as potential mates, at about the same level of morphology that it had.

What would we do with transitional forms if we found them? We would have to make some sort of arbitrary distinction that was never there in the original populations. Archaeopteryx was classified as a bird because it has feathers, but every other character is "dinosaurian". (And we have just found a new Australopithecus that appears to have been making and using tools, a behavior that we had always thought was a character of Homo.)

Since species always grade into one another in time as they evolve one into another, we will increasingly be faced with problems like this. We are faced with the problem at higher taxonomic levels too: when does one family evolve into another? (It is still the point at which one species changes into another.)

As populations evolve through time, they may sometimes change enough that any biologist or paleontologist looking at them would decide they are two separate species. For example, no-one has ever doubted that chimps and humans are separate species. Now think about a time 6 million years ago, when the ancestor of humans and chimps was shambling round in East Africa. If we could arrange the populations that descended from that ancestor, we would find that two separate lines have descended from that single species. Each line contains multiple species. If we had a complete record, we could not tell when one species evolved into the next, because all along the way, parents had offspring, and the offspring found compatible mates, and so on. There would not have been some event when suddenly all the adults in a particular population had mutant offspring who were the first generation of a new species!

In reality, therefore, the paleontologist has to face the uncomfortable truth that there are no convenient gaps between species in time like those that the biologist sees between living species. In reality, the paleontologist faces the task of dividing fossils into groups that may be arbitrary. How do we decide where Homo erectus has evolved into Homo sapiens? (Currently, anthropologists are avoiding the issue by coining new names for "species" in between: Homo heidelbergensis, Homo antecessor, Homo neanderthalensis, for example, but that only postpones the day when they too will have to face the reality of what they are doing, attempting the illogical.)

Most of the time the paleontologist faces enough gaps in the record to give gaps in the evolutionary story, so most "species boundaries" are defined at the gaps. As the fossil rcord becomes more complete, species boundaries in time become more arbitrary, as they should. It's like drawing grade boundaries for a class: the bigger the class, the less chance there is that there will be a gap between B and C students.

Now that we understand the difficulty, let's face the problem of



. We cannot define fossil species using the genetic definition of a species. But then that's also true of most of the species that biologists define among living organisms. Instead, most biologists, and all paleontologists, use a morphological definition: a species is defined as a group of individuals that have some reliable characters distinguishing them from all other species. In biology, the characters can include behavioral or biochemical features, or morphological characters such as soft-part anatomy or coloration that are practically never found in the fossil record. Nevertheless, the procedures are practically identical, only differing in that the suites of potential characters used to define species are more limited in fossil material.

A paleontologist cannot look at the living creatures, so must interpret species on the basis of paleontological remains only. So there are potential problems. This does not mean that the problems are so difficult that a paleontologist cannot separate out species, but it does mean careful thought.

If you only collect a few specimens of a widely variable clam species, you might think you had several species. However, as you collect more and more specimens, you realize that in fact you are filling in the "gaps" between the original specimens. So you have only one species.

Taking material from another lecture, this actually happened with species of the dinosaur Triceratops. After 80 years of collecting, 30 species of Triceratops had been described, each one slightly different from the others. Eventually someone realized that all these "species" had been collected from only two counties in Wyoming. In fact, it was ONE species, with large and small males, large and small females, juveniles, and so on....

If you collect many specimens of scallops from a particular seashore, you will probably find that there is a range of numbers of ribs on the shell. In the example I showed, I believe the number ranged from 15 to 19, with an average of 17. Futhermore, most specimens had 17 ribs, and those with 15 and 19 were few in number. It is obvious that this is a sample from one species.

However, it is possible that you could find two good species that overlap slightly, say, in size. You would need a good number of specimens to plot a graph and show that there were TWO maxima, so two species.

People often plot graphs of measurements taken from shells or parts of shells, and then they try to decide whether there are two lines that can be plotted through the sata, or only one. Sometimes this can be a subjective judgment. You can often find a statistical program that will differentiate between two groups, but that doesn't necessarily mean that they are separate species. Imagine plotting the heights of basketball players and gymnasts. They would not overlap at all, yet that has no biological meaning.

Sometimes it helps that many organisms have determinate growth: they grow to a certain adult size, then stop. (Humans, for example). If you had a good collection of fossils, you might be able to decide that you were dealing with two species of mammals, for example, one of which reached a greater overall size than the other.

As creatures grow, they often grow disproportionately (the process is called allometry). Human babies are born with large heads and big goo-goo eyes, and their bodies grow much more than their heads do, until they reach adult size. And this is true for many other organisms. One should always be aware that ratios between structures may change through growth: this is not a good basis for separating out species.

Collecting only from one place may be very deceiving. For example, the California fur seal has an unusual yearly cycle in which males and females separate for many months, feeding in totally different habitats, but spend the birth and breeding season together. One could easily imagine being fooled by collecting only the localities where the sexes were separate: you would confidently describe two species of fossil seal, a big species and a small species, not realizing that you were describing male and female.

Sexual dimorphism in ammonites (shelled cephalopods). Though ammonites are extinct, we reconstruct some species as having small males with a rather ornate shell opening, and much larger females. (The reason would be that it takes a relatively large shell to store a lot of relatively large eggs.) These shells were called by completely different names because they look very different as adults (the baby shells are the same). This is a testable hypothesis: you would predict that the "male species" would have the exact same geological age range as the "female species". If so, you can go ahead, call them a species, and use only one name for them.

As we saw for Triceratops, the same confusion may occur among vertebrates, and if the difference in size or structure is subtle, it may be difficult to detect. For example, some people have suggested that the specimen "Lucy" of the early hominid species Australopithecus afarensis is actually a male.

Growth patterns may include interruptions for molting, as in crustaceans and trilobites. At each molt, the creature has the option of changing its soft body shape before it grows its new armored carapace. So an adult trilobite may have a noticeably different body plan and shape than it did when it was a juvenile. The paleontologist often needs a good collection of all most of the life stages, to avoid calling a juvenile and an adult of the same species by a different name.

Complex changes may go on inside a creature as it grows. Inside brachiopods, a loop of skeleton holds the feeding organ, and that loop changes dramatically through growth. The shape of the loop is considered important enough to be used in naming, and it is likely that many brachiopods have been called separate species when they are in fact growth stages of one another.

As a side story, large collections of molting trilobites have revealed details of their life history. Because each molt stage represents an individual that had already placed several previous stages into the fossil record, one can see how long-lived an average ostracod was.

And after a paleontologist has considered all the factors that may be telling her an incomplete story, she must look at her collection of specimens and decide whether to name a species or not. If it is a new species, previously unknown to science, she must decribe it clearly, give it a new name, specify how it differs from previously described species, draw or photograph it, and name a single specimen that will serve as a reference for the new species. Then it goes to a museum where others can examine it and challenge her interpretation if they have better evidence.


. Among living organisms, transitional populations between species are exceptional, with the Arctic gulls being a good example: and even then, it's really a matter of definition whether one regards that example as one species or two intergrading ones. Fossil material, on the other hand, carries the dimension of time, and often shows sequences of species that are obviously related to one another, that follow one another in time, and are generally interpreted as evolutionary lineages. Thus few people doubt that Homo habilis-H. erectus-H. sapiens form a sequence of evolutionary descent as well as a sequence of related species in time. In such cases, we might expect to see transitional populations quite regularly. How would we define species? If all living organisms evolved from some common ancestor in the early Precambrian, then there is full genetic continuity throughout all organic life on Earth-where would we draw the boundaries?

Last revised March 1, 2006.

Return to Evolution topics

Return to Geology Department home page