Page 22

The Creation of Biodiversity

Edward O. Wilson
Museum of Comparative Zoology, Harvard University, 26 Oxford Street, Cambridge, MA 02138-2902

The term biodiversity, short for biological diversity, was introduced by the National Research Council staff at the first National Forum on BioDiversity, held in Washington in September 1986; and it gained rapid global currency after the publication of the forum proceedings in 1988. Biodiversity means, in simplest terms, the variety of life found in the Creation; it is the entirety of life on the planet.

Biologists rescue this conception from vacuity by analyzing biodiversity at different levels of organization, from biosphere downward to gene, and integrating the information to address the fundamental questions of its breadth and origin. More recently, with growing alarm, they have widened their focus to include the causes of the accelerating decline of biodiversity in the human-saturated environment. The first process, creation, is the concern of evolutionary theory; the second is the subject of the new discipline of conservation biology. I will now address the first process.

Researchers have found it most useful to stress diversity at just three levels of biological organization, namely, ecosystem, species, and gene. An ecosystem is a local community of species organisms plus their physical environment. Familiar examples are a New England pond, an old-growth forest in Oregon, and a deepsea thermal vent off the Pacific coast. Although broad types of ecosystems, such as old-growth conifer forests and thermal vents, can be roughly defined by properties they have in common, no two particular ecosystems belonging to a given type are ever exactly alike, either in their species composition or in their physical environment. Throughout the world, individual ecosystems are highly endangered or have disappeared. When a forest is cut, others of the same ecosystem type can

Page 23

persist nearby, but its unique properties have vanished forever. Moreover, many ecosystems contain endemic species, native to that place and environment and found nowhere else. A threatened individual ecosystem or aggregate of ecosystems with many endemic species is called a “hot spot.” The rain forest of Kauai is a hot spot; and because so many other kinds of ecosystems on Kauai and the surrounding islands contain threatened endemic species, all of Hawaii is justifiably called a hot spot.

Because ecosystems are difficult to classify and even in many cases to delimit geographically, they are seldom used in quantitative studies of biodiversity. The unit of choice is the species; species are relatively easy to describe and have been the focus of more than 2 centuries of research in classification and biogeography. The traditional definition of the species is the one given in the “biological species concept”: a population or series of populations of individuals capable of freely interbreeding with one another under natural conditions—in short, a closed gene pool. The occurrence of an occasional hybrid is not enough to combine two species into one under this definition; only free interbreeding can do that. Also, the ready production of hybrids in zoos and botanical gardens—for example, between lions and tigers—does not suffice. Gene flow must occur under natural conditions, which apparently never occurred between lions and tigers where they coexisted in the past.

The biological species concept works very well for most kinds of animals and for some plants, such as the orchids, but it has serious problems. In a large percentage of cases, there is no way to know whether two populations that occupy different geographic ranges would interbreed if somehow they met under natural conditions. A population of birds on Oahu, for example, cannot be judged with certainty against a somewhat different population on Kauai. The usual taxonomic solution to the dilemma under the biological species concept is to classify the two populations as subspecies, or geographic races, of the same species.

Yet another problem with the biological species concept is its irrelevance to the vast assemblage of life forms that do not reproduce sexually—or else do so rarely enough to reduce sexuality to marginal importance in the life cycle. Thus bacteria, which with the asexual Archaea are both the most primitive and the most numerous organisms on Earth, cannot be classified by the biological species concept. A bacterial species is instead defined as a lineage with 30% or more difference from other lineages in DNA base pairs or else subjectively different enough in traits of biochemistry and structure to justify such recognition. As a result, and with insufficient technology to impose even these loose criteria, no one knows to within a factor of less than 100 how many bacterial species exist on the planet.

Understandable dissatisfaction with the biological species concept has encouraged the devising of an alternative definition, that of the phylogenetic species. In this view, the most meaningful species is a distinctive population with a monophyletic lineage—in other words, derived from a single ancestral species. It is of little concern in this view if the populations have indeterminate breeding potential with other populations. As long as the population comprises individuals of the same coherent lineage that are distinguishable to a subjectively agreed-on degree from that of other populations, it can be ranked as a species.

Page 24

The advantage of the biological species concept is its recognition that closed gene pools are entities that have for the most part been irreversibly launched on an independent course of evolution by mutation and recombination. Given enough time, and without regression by hybridization, species that are at first barely distinguishable are destined to become very distinguishable. The advantage of the phylogenetic species concept is that it reflects rigorously the history of groups of related species without reference to their hypothesized future.

The two cross-cutting criteria, breeding and phylogenetic, can be joined to create a synthetic species concept as follows. A sexual species is a population that is both reproductively isolated and monophyletic. Suppose that a monophyletic sexual population is geographically isolated, so that its reproductive status vis-à-vis similar populations is indecipherable. It can be called a species if it is markedly distinct, or a subspecies if only slightly distinct.

What the new emphasis on molecular markers and phylogenetic analysis comes down to is, I believe, the prospect of increasing the number of formally recognized species through ever finer analyses of the phylogeny of populations, especially such analyses based on DNA sequencing. More subspecies, once they have been found to have substantial differences that are concordant across their ranges, will be raised to species rank. And more sibling species, which are hard to detect with conventional anatomical characters, will be recognized and named. The new emphasis does not, however, in my opinion represent a fundamental shift away from the species concept already used by most practicing taxonomists. As a rule, they have embraced the concepts of both reproductive isolation and monophyly while recognizing as guesswork the assignment of reproductive relationships among closely related but geographically isolated populations.

The current trend of systematic theory is toward a higher degree of objectivity and consensus than existed in the past. A synthetic, truly biological species concept, providing considerable information about each genetically distinguishable population, seems attainable. This aim is of central importance in ecology and conservation biology. How species are delimited and classified determines the number recognized, as well as the number of genera and other higher categories into which they can be defensibly grouped, and hence the magnitude of both local and global biodiversity. It affects the evaluation of the status of individual populations in conservation planning, that is, whether the populations are ranked as species, subspecies, or neither. And finally, the refined species concept conforms more closely to the emerging picture of how biodiversity is created.

Biodiversity is the product of two complementary processes of evolution. The first is vertical evolution of individual populations by changes in chromosome composition and gene frequency. During the process, biodiversity at the level of this hereditary unit grows or declines. But the number of species, the next level up, does not necessarily change as a result. The second evolutionary process, then, is the multiplication of species, often called speciation. In the course of vertical evolution, some species split into two or more daughter species; others do not.

Virtually all biologists closely familiar with the details of vertical evolution give natural selection the dominant role in evolution. In simplest terms, it begins when

Page 25

different forms of the same genes, or alleles, originate by mutations, which are random changes in the long sequences of DNA that compose the genes. In addition to such point-by-point scrambling of the DNA, new mixes of alleles are created by the recombining processes of sexual reproduction. Other forms of mutation occur when entire chromosomes, the carriers of the genes, are duplicated, deleted, broken, fused, or otherwise reconfigured. The mutations, genic and chromosomal, that enhance survival and reproduction of the carrier organisms spread through the population.

The ones that do not enhance fitness fall to very low frequencies or disappear altogether. Chance mutations are the raw material of evolution. Environmental challenge, deciding which mutants and their combinations will survive and reproduce, molds the population from this protean genetic clay.

Although natural selection has the commanding creative role, another force must be mentioned in any account of evolution. By chance alone, substitutions occur through long stretches of time in some of the genes. The continuity of random change is often smooth enough to measure the age of different evolving lines of organisms. But this genetic drift, as it is called, while altering the diversity of genes, adds little to evolution at the level of cells, organisms, and populations. The reason is that the mutants involved in drift must be neutral, or nearly so, in the crucible of natural selection; in other words, they can have little or no effect on the details of higher biological organization on which organisms depend for survival and reproduction.

Driven by natural selection, some species break into daughter species. By the criterion of reproductive isolation, species multiply when populations acquire genetic differences that interfere with mating or the healthy growth of hybrid offspring. These differences are called intrinsic isolating mechanisms. They affect various parts of the life cycle concerned with sexual reproduction, such as differences between populations in times or places of mating, in courtship and mating procedures, and in the developmental physiology of offspring. They can occur singly or in any combination, depending on the biological nature of the species and vagaries of natural selection affecting its evolution.

The classic model of species formation is geographic speciation. Its principal steps, which have been richly documented, are the following. A single population of interbreeding individuals is split into two or more populations by a geographic barrier. Because the barrier is not part of the genomes of the populations, it can consist of almost any feature of the physical environment. It can be the drying of a mesa when the climate enters an arid phase, causing the forest that once covered it to break into fragments sheltered by scattered canyons. It can be the straits that separate two islands of an archipelago. A bird species might only rarely cross this permanent water barrier, but when the event does occur, individuals from one island are able to invade the other island, where the colonists form a population almost entirely isolated from the source population.

As the two populations separated by geographic barriers of whatever nature diverge, they progress from being genetically identical or nearly so to slightly or moderately different, at which point they can be called subspecies—or, meaning the same thing in this context, geographic races. At this stage, the systematist

Page 26

who emphasizes interbreeding capacity says, “The differences are worthy of recognition but not strong enough or involved enough with reproductive traits to call the populations species. If the diagnostic traits were stronger and especially of this nature, I'd call them species.” Another systematist, concerned more with phylogenetic criteria, might respond, “All right, but I'll call them species if there are multiple and well-marked diagnostic characters throughout the population, and if the species with which they are compared share an immediate common ancestor and possess their own well-marked and consistent traits. I am not so interested in trying to predict their future.”

Even with such clarification, however, the distinctions between subspecies and species are filled with residual ambiguities difficult to explain to impatient students or members of Congress. Here are several:

• A subspecies, or geographic race, can contain genes and traits that are even more distinctive than those of otherwise similar reproductively isolated species, yet not be reproductively isolated or coherent enough to meet the criteria of a phylogenetic species.

• Two species can be separated by numerous genetic differences that nevertheless produce no outward traits easily discerned by investigators. Examples include odors used in communication and internally hidden physiological processes. These “sibling species,” even though important elements of biodiversity, are nevertheless consistently undercounted.

• Some species, especially on continents or large islands, are broken into numerous local populations that vary genetically from one another. The temptation exists to recognize many subspecies among the populations, but two outstanding difficulties are often encountered in such cases. First, the geographic limits of each population are often difficult, if not impossible, to define. Second, the traits typically vary discordantly. To take an imaginary but realistic example of discordance, size might decrease from north to south, color from east to west, food preference from northwest to southeast, and so on, indefinitely. The number of geographic character lines that can be drawn and hence the number of subspecies recognized in such discordantly varied species depend on which traits are chosen to follow them. Still, in spite of this difficulty, a large percentage of species comprise local populations that can be easily delimited and whose diagnostic traits are concordant enough to justify subspecific or, by stress on the phylogenetic criterion, specific status.

• To add to the many complications inherent in geographic differentiation, species can also multiply in the absence of geographic barriers. Almost half of living plant species and a smaller number of animal species have arisen by polyploidy, the multiplication of entire chromosome sets. The idea of polyploidy can be quickly grasped as follows. If the number of chromosomes in the egg or sperm of a nonpolyploid organism is N (haploid), then the number in the fertilized egg and ensuing organism is 2N (diploid). In a polyploid, the number in the fertilized egg and ensuing polyploid organism is 3N (triploid), or 4N (tetraploid), and so on. A polyploid with 4N chromosomes can in some cases breed with its 2N ancestor, but the hybrid offspring, which carries 3N chromosomes in each cell, is ordinarily

Page 27

unable to complete the steps of meiosis and hence to produce viable sex cells of its own. As a result, the 2N ancestor and its 4N derivative are distinct species. The splitting of one species into two species in this case occurred across only two generations—a near instant in evolutionary time. A variation of the process can occur when two species create a hybrid that is also a polyploid. With two of each kind of chromosome thus provisioned in each cell of the organism, the chromosomes can pair off with exact equivalents in the first meiotic division, permitting the production of normal sex cells. The polyploid hybrids can as a result breed successfully with one another, but not with their diploid parents; so they are established as a new, reproductively isolated species.

Another form of sympatric speciation, or species multiplication in the absence of geographic barriers, is through host races. The process is hard to detect and harder to prove, but it might be far more important in nature than previously appreciated. It unfolds when a species of say, an insect is specialized to feed on the leaves or fruit of one species of tree, a common situation in nature. It also mates exclusively on this same host plant. A few individuals, either because they are mutants in food preference or because they make an error in plant selection (and then become imprinted on the wrong tree species), move to an alternative host, where they proceed to feed and mate in isolation. As a result, two populations coexist in the same locality. At first, when the differences among them are slight, they are legitimately called host races, or ecological subspecies. But as they diverge genetically, and especially if the host preferences have a hereditary basis, they are classifiable as distinct species.

No one at this time can confidently evaluate the prevalence of sympatric speciation by host races or other highly local splitting of populations. But given that insect species alone number in the millions, many of them specialized as herbivores on plants or as inhabitants of microhabitats, the process might in time prove to be one of the most important in the origin of biodiversity.

As new species originate—sometimes across only two generations, sometimes during a period of hundreds or thousands of generations—other species die. Over large geographic areas and spans of time, the balance of birth and death maintains a roughly equilibrial number of species in major groups, such as birds, ants, conifers, and mosses. The number appears to be a complex correlate of factors summarized by the acronym ESA, not for the Endangered Species Act or the Ecological Society of America, but for Energy, Stability, and Area. In general, the greater the amount of energy available to the ecosystems, the larger the number of species; thus, high levels exist in the energy-rich coral reefs worldwide and the great tropical moist forests of South America, Africa, and Asia. The more environmental stability, as in the tropical forests and bottoms of the oceans, the greater the number. And, finally, the larger the area, the more species that can be sustained within it.

The role of area in particular can be described by the following broad rule: the number of species occurring in physically well-demarcated habitats—such as islands of an archipelago, patches of woodland in a fragmented forest, or clusters of lakes—varies from the sixth to the third root of the area of the habitats. The

Page 28

exact value varies with the kinds of habitat and organisms studied and the part of the world in which they occur. A common central value is the fourth root, which translates to an easily recalled rule of thumb: a 10-fold increase in area results in a doubling of the number of species.

Where the ESA factors combine, an astonishing number of species have typically accumulated. The greatest biodiversity overall in the world appears to occur in the upper Amazon Basin, which is notably high in all the ESA factors. For example, the largest number of butterfly species in the world observed at a single locality is 1,300, recorded within 3,925 hectares of mostly lowland rain forest at Pakitza, Parque Nacional del Manu, Peru, by Robbins and co-workers (1996). By comparison, only 380 species are known from all of western Europe (Higgins and Riley 1970). Similarly, the world record for ants is 365 species, collected within only 8 hectares of lowland rain forest at Cuzco Amazónico, also in Amazonian Peru, by Stefan Cover and John Tobin (personal communication). That diversity can be instructively compared with the 555 species found in all of North America (Bolton 1995).

The assembly of biodiversity at the level of ecosystems encompasses two complementary principles of organic evolution. The first is adaptive radiation, the expansion of multiplying species from individual stocks into niches available to them. The second is convergent evolution, the increasing similarity in anatomy, physiology, or behavior, singly or in combination (but not in the underlying genetic codes), of radiating groups found in different parts of the world.

The Hawaiian archipelago, the most isolated islands on Earth, are appropriately cited as a natural laboratory that displays the two complementary principles with exceptional clarity. Its roughly 8,000 known endemic land and freshwater species (Eldredge and Miller 1995) have been derived from only a few hundred ancestral species that managed to cross the immense barrier of the Pacific Ocean from continents and islands on both sides. Many of the colonists, arriving over a period of several million years, found an array of major niches open that were closed by competitors in other parts of the world. Among the insects that converged dramatically to adaptive types in other places are geometrid moths whose caterpillars abandoned herbivory to become ambush predators of other insects and a dragonfly whose nymphs have left freshwater streams to forage on land. One lineage of ducks, the moa-nalos, now extinct, evolved into large flightless forms with tortoise-like bills. The fullest and best-known radiation among animals is in the Hawaiian honeycreepers of the family Drepanidinae, whose 23 species (living and recently extinct) were derived from a single ancestral fringillid bird species. In anatomy and behavior, they have variously filled the niches of warblers, woodpeckers, finches, nectar-feeding sunbirds, and parrots. The most striking example among plants is in the tarweeds of the sunflower family Asteraceae, whose numerous species vary from low, herbaceous mats to shrubs and trees, to the spectacular silversword of Maui's Haleakala Crater.

It is by countless such radiations and exchanges of species among their own evolutionary headquarters that the ecosystems have assembled. On a grand scale, much of the history of life can be viewed as a succession of adaptive radiations during which major groups displaced previous assemblages or were able to spread

Page 29

into wholly new adaptive zones made possible by the increasing complexity of preexisting ecosystems. Life has always expanded to fill the space and use the energy offered to it. The glory of this creative process is the biosphere, billions of years old, over which humanity has lately taken command. The tragedy is that we are thoughtlessly tearing it down before we fully understand its origin, how it is sustained, and the essential role that it plays in human welfare.

References

Because this essay is a primer of a broad array of topics, it is appropriate to recommend three general texts, of the many available, for a more detailed introduction:

Raven PH, Evert RF, Eichhorn SE. 1999. Biology of plants, 6th ed., New York NY: Worth Publ. Futuyma DJ. 1997. Evolutionary biology, 3rd ed., Sunderland MA: Sinauer Assoc. Wilson EO. 1992. The diversity of life. Cambridge MA: Harvard Univ. Pr.

Several specialized citations not covered in these general works are given below.

Bolton B. 1995. A taxonomic and zoogeographical census of the extant ant taxa (Hymenoptera: Formiddae). J Nat Hist 29:1037–56.

Eldredge LG, Miller SE. 1995. How many species are there in Hawaii? Bishop Mus Occas Pap 41:1–18.

Higgins LG, Riley ND. 1970. A field guide to the butterflies of Britain and Europe. Boston MA: Houghton Mifflin.

Robbins RK, Lamas G, Mielke OHH, Harvey DJ, Casagrande M. 1996. Taxonomic composition and ecological structure of the species-rich butterfly community at Pakitza, Parque Nacional del Manu, Perú. In: Wilson DE, Sandoval A (eds). Manu: the biodiversity of Southeastern Peru. Washington DC: Smithsonian Inst Pr. p 217–52.