Populations as defined here do not need to have distinct boundaries. Individuals can and do move between populations. Individuals almost never have uniform distribution within populations: there are marginal stragglers, making the boundaries fuzzy as populations intermingle. The population is a unit within which genetic change can take place, unlike an individual: the population has a gene pool.

A species is an array of Mendelian populations, and between each species there is supposed to be genetic reproductive isolation. A species is supposed to have absolute boundaries: if they are not geographical, they are behavioral or genetic.


Although the definition of a species is clear, there are practical difficulties. It would be prohibitive in time and money to establish whether each species on earth was a distinct entity, reproductively isolated from every other species. Instead, biologists describe species on the basis of differences in morphology, hoping that the morphological differences reflect genetic differences that are so important that they prevent interbreeding. In almost all cases, this approach is valid. But there are enough problem cases to show that the biological species concept is difficult to apply even among living organisms.

Sibling species

Sometimes there is no discernible morphological difference between two or more species, yet they do not interbreed. The differences may be behavioral, or genetic, and they may be geologically recent. There are cases among fruitflies, among snails, and among frogs, and such species pairs or groups are called sibling species.


Members of two species that are morphologically distinct may sometimes interbreed to produce a hybrid, an offspring that differs from both parents. The hybrid itself may be sterile, but may sometimes be fertile. Often hybrids are found in a narrow zone along the geographic boundary of a species range, where potential mates of one's own species may be rare. Hybrids may also occur where mistakes are made in mate recognition. Hybrids are often less fit than their parents, so that they reproduce poorly or not at all. Therefore there is usually strong selection for discrimination in mating: an individual should be able to recognize a conspecific. Usually, then, gene flow between species is small compared to the gene flow within species, so that hybrids are relatively rare. Nevertheless, they occur often enough to pose the problem of dealing with them relative to the species concept. Clearly they do not fall within either of the parent species, yet they do not form their own species either. We have to recognize that even the concept of the species as a discrete unit breaks down: one could argue that a hybrid zone is just a very steep cline within the range of a broadly defined species.

There are practical difficulties too. Sometimes hybrids between two species occur with such frequency and regularity that they are significant, permanent groups of individuals: such hybrids have often been described as species. Some hybrid populations may eventually become self-sustaining to the point that they do become true species: but the supply of new individuals from hybrid matings would have to stop before that stage was reached. Some lizard populations in the American Southwest may be close to this stage.

Sometimes hybrid populations are difficult to distinguish from species. Butterflies and birds that were described as separate species have been observed first to hybridize and then to merge completely into a continuously interbreeding species. (Obviously, they were not really separate species when first described: the point is that it can be difficult to decide whether a set of populations forms one species on the basis of short-term observations.)

Genetic vs. Ecological Species

Some butterfly species can show regional differences, based on factors such as differences in diet. Yet they may interbreed freely across regional boundaries, so that the geographical (or ecological) subspecies can be easily recognized as such in spite of their differences in habits and/or appearance. Other butterfly species may look very similar, and have similar ecological habits, yet are genetically rather different, so that they interbreed only rarely, or produce hybrids that are sterile. These would be classed as true sibling species in spite of their outward similarity. The complexities of defining species are very great.

Ring Species

The classic example of ring species is provided by Arctic gulls (Larus species). These birds form a set of populations which live around the Arctic Ocean. In Western Europe the lesser black-backed gull (L. fuscus) is a familiar species whose range extends east into the Soviet Arctic, through populations that are interbreeding but which can be arranged into four subspecies, each slightly different. The easternmost subspecies interbreeds with populations of the so-called Siberian skua. In turn, the Siberian skua interbreeds with populations farther east, through five more subspecies. The easternmost subspecies is so far east that it ranges into Western Europe as the herring gull, where it exists alongside the lesser black-backed gull. But the herring gull does not interbreed with the lesser black-backed gull in Western Europe, and no amateur ornithologist there would ever place these two in the same species. In fact the herring gull is called Larus argentatus. But there is gene flow round the 20,000 km of the Arctic ring that forms the range of these gulls. The interbreeding is so free, and the color patterns of the plumage and the behavior is so smoothly transitional between these successive sets of populations that it is clear they all belong to the same species, even though the end members are distinct.

At least 22 cases of ring species have been documented. Although their biology is not difficult to understand, these cases nevertheless pose great difficulty for operational taxonomy. We have to recognise that the natural world simply does not fit easily into the neat categories that humans would like to impose upon it.


In applying the species concept to the fossil record, there are two grades of difficulty:

We hardly ever see discussion of the first category of difficulty, because there is nothing to discuss: the difficulty is insurmountable. However, most biologists do not use the genetic species concept in practice, and it is perfectly reasonable for paleontologists to take the same approach and to define species on morphological grounds. The second difficulty is one of degree: there is never as much information available from a fossil as there would be from a living, or recently dead individual. Again, biologists sometimes have to deal with difficult material, such as specimens dredged from the deep sea floor. Paleontologists have to work on challenging material, but that does not prevent them from making reasonable inferences as they define species in fossil material. There are several ways to ensure that classification is useful.

A. Examine the Holocene and Pleistocene fossil record, which often contains good fossils of species that still survive. How does the fossil record preserve or obscure the real differences we can see between living individuals, populations, and species? One can show, for example, that the hundreds of dire wolves preserved in the La Brea tar pits contain an unusually large percentage of males. Once this bias is identified, one can be aware that it may also have affected smaller collections in which the bias would not so easily be recognised. In an extinct Pleistocene animal, the European cave bear, Kurten has been able to recognise differential selection between growing individuals that related to malformed or misplaced teeth; he has been able to work out mortality rates and life expectancies; he has worked out the mating patterns and social structure, and he has been able to show differences between populations in time and space within this species.

B. Understand geographic variation: two fossil groups that are slightly different but separated in space may in fact turn out to be members of the same species when new discoveries are made between the localities. For example, the robust australopithecines that are found in East Africa and South Africa (Australopithecus boisei and A. robustus repectively) may turn out to be members of the same species when the fossil record of Zambia and Zimbabwe is better known. Or, more accurately, the fossils may turn out to be no more different than one would expect of populations of living hominids, so that they can be more easily interpreted as one species than two.

C. Understand ontogenetic variation. Organisms usually change in size and shape as they grow, and an incomplete collection of individuals may show large gaps between two stages, suggesting a separation into two species. Larger collections can and do fill such gaps, and it may then be recognized that the specimens are better interpreted as one species. Systematic changes in shape with size are called allometry. Familiar examples include the tremendous development of horns with later growth in many ungulates, living and fossil, but almost all fossil groups show some sort of allometry.

Many invertebrates undergo some kind of metamorphosis, in which the body is drastically reorganized in a very short time. Unless the biology of the group is well known, the two developmental stages may be classified as different species. Arthropods in particular are prone to misinterpretation, because many of them molt their exoskeleton and take advantage of their temporary soft bodies to alter shape and size. Insects have developed metamorphosis to an extreme that is shared only by frogs and toads among amphibians. Fortunately, these animals all have a relatively poor fossil record, and we understand them well. More subtle cases of metamorphosis are likely to be more difficult to recognize.

D. Sexual dimorphism, or other cases of polymorphism, may develop during ontogeny. Humans show a mild degree of sexual dimorphism, and it can be much more extreme in many mammals, perhaps in dinosaurs, and especially in some invertebrate groups. Cephalopods, in particular, have been classified into different families on the basis of the difference between the sexes of adults of the same species.

E. Phenotypic variation resulting from environmental parameters may be extreme enough to cause a paleontologist to classify a species under several names. Among living animals, some groups such as oysters and corals are notoriously prone to wide variation because of variation in the substrate they grow on and/or the availability of food, light, clear water, and the presence or absence of turbulence and sediment. It's clear that this source of variation was significant among some fossil groups, and the paleontologist must be aware of it.

F. Taphonomic agents may remove or distort the morphology of the fossils to the point that vital information is lost. For example, young stages may be washed out of assemblages of clams before they are buried, or thin (young) shells may be preferentially dissolved.

G. Mistrust statistics. Paleontologists sometimes resort to statistical methods in order to gain some "objective" guide about variation. If there are "significant" morphological distinctions between two collections, the paleontologist may assign them to different species. But even among living organisms there may be statistically significant differences between populations: humans are a good example. Statistical difference is a real and objective fact, but it need not indicate anything about the biological difference between samples. Return to 107 main menu