The survey of ant–plant mutualisms in Chapters 3–7 presents a complicated picture. In Chapter 3 it was shown that careful studies of plant species that bear extrafloral nectaries have yielded contradictory results. Some show clear evidence of ant protection, others do not. In Chapter 5 the phenomenon of ant feeding of plants was confirmed in a couple of species, but the benefits conferred by ants on a great variety of others that harbor nests remained obscure. In Chapter 6 the discussion of ant dispersal showed that different authors can study the seed and seedling demography of the same elaiosome-bearing species and reach different conclusions as to the importance of ants at these life-history stages. The situation for ant pollination (Chapter 7) is also unclear. The data are scarce and claims of ant pollination, often based on anecdotal evidence, are generally inconclusive. Although it is possible that the differing conclusions reflect individual biases among authors, it seems more likely that they accurately reflect natural variation in function and effect. In this chapter the causes of this variation will be examined using ant protection and ant dispersal as examples.

The evolution of mutualisms is affected by demographic and life-history characteristics of the plant and ant populations. For the ants we know that particular foods are required at particular stages in colony development or reproduction. We also know that profound demographic changes such as the proliferation of worker castes or reproductives are likely to lead to increases and decreases in colony interest in plant rewards.

Plants offer two basic types of rewards for the services of ants: housing and food. Housing was discussed in Chapters 2 and 3 and will not be taken up at length here. Food, or nutrition, will be treated in a broad sense. Thus, in addition to carbohydrates, proteins, and lipids ingested for normal metabolism, development, and growth, substances necessary for social organization such as those required for the biosynthesis of mating pheromones, alarm, and defense will be included. Also it should be borne in mind that substances that attract ants to a reward are not necessarily those from which they derive the primary benefit. That is, the distinction between attractants and nutrients may be crucial (Marshall, Beattie, & Bollenbacher 1979).

Ant nutrition

Most ant species involved in seed dispersal or plant protection are omnivores, and the plant rewards harvested by them while performing these services are only part of their overall nutrition. Individual colonies may opportunistically exploit extrafloral nectar, or elaiosomes on seeds at an intensive level for short periods; but almost invariably some workers are foraging at other resources at the same time. A major consequence of this is that the intensity with which plant rewards are harvested varies enormously.

This scenario is not at all surprising when it is remembered that the nutritional requirements of ants vary with the state of the colony, and conversely, the food available to a colony influences the number of individuals or castes.

Walk through the vegetation almost anywhere on earth and you will see ants foraging on plants. The basic interaction has evolved into four ant–plant mutualisms in which ants: (1) protect plants from herbivores and other enemies, (2) feed plants essential nutrients, (3) disperse seeds and fruits, and (4) pollinate. Rewards produced by the plants, chiefly nest sites or food, are utilized by ants and the behavior patterns involved result in one or more of these services.

Much more is known about the benefits to the plants than the benefits to the ants. For example, the anatomy and morphology or chemistry of many of the rewards borne by the plants have been described. Critical field experiments have been performed to test the impact of ant services on plant growth, survivorship, and fecundity, and it is clear that ant services can profoundly affect plant fitness. On the other hand, although ants eagerly occupy plant-borne nest sites and harvest plant-borne food rewards, almost nothing precise is known about how this affects ant fitness. A major gap in our knowledge of ant–plant mutualisms is how the food rewards affect the physiology, growth, and demography of ant colonies. As a consequence of this situation, this book is written primarily from a “plant's point of view.” It also should be noted that fundamentally nonmutualistic ant–plant interactions, such as predation by seed-gathering ants and herbivory by leaf-cutter ants, are referred to only in passing.

When van der Pijl (1955) reviewed the relationships between ants and plants he perpetuated a number of terms, most of which were first proposed by Warburg (1892). These included myrmecophily for ant pollination, myrmecochory for ant dispersal of seeds, and myrmecotrophy for the feeding of ants by plants, principally by means of extrafloral nectaries. Because myrmecotrophy involves much more than the feeding of ants and has largely been replaced by the concept of ant protection of plants, I intend to put the term to another use. If the subject–object relationship of these terms remains consistent, then myrmecophily denotes the relationship in which ants benefit plants by acting as pollen vectors, myrmecochory describes the benefit to the plant conferred by ants dispersing seeds, and myrmecotrophy implies a relationship in which ants benefit plants by feeding them. As far as I can tell most biologists were unaware of this possibility at the time of van der Pijl's review. Since then, however, the existence of ant-fed plants has been documented.

The nutritional benefit to the plants has been clarified for two Rubiaceous genera from Southeast Asia and northern Queensland that house ants in large tubers derived from the embryonic hypocotyl (see Table 7). Janzen (1974b) observed that the ant Iridomyrmex myrmecodiae abandons the remains of prey in some of the cavities that ramify the tubers of Hydnophytum formicarium and Myrmecodia tuberosa (Figure 8). These cavities are lined with absorptive tissues.

Many species of ants gather seeds. Harvester ants store them in underground granaries and consume them during the winter or dry season. These ants are granivores and the net interaction is usually predation. Other ants gather seeds and fruits distinguished by the presence of external tissues, collectively called elaiosomes, which attract ants and stimulate them to carry the entire seed or fruit back to the nest (Figure 9). There the elaiosomes are removed and typically fed to the larvae. The seeds are then discarded, both intact and viable, either in an abandoned gallery of the nest, or close to an entrance in a refuse pile together with other organic waste. Because the elaiosomes contain ant attractants, and as the seeds are not harmed, the interaction, known as myrmecochory, has long been assumed to be a mutualism. Until recently, however, the advantages of the interaction, especially for the plants, have remained undocumented. Recent studies have shed light on the problem, and there appear to be five current hypotheses on the selective advantage to plants of dispersal of seeds and fruits by ants.

The predator-avoidance hypothesis

Seed predators are so diverse and abundant that plant species must be under great selective pressure to avoid them. The essence of this hypothesis is that seeds released from the parent plant are quickly taken by ants to their nests, where they find refuge from predators. The ants are rewarded with nutritive elaiosome tissue and then rapidly lose interest in the seed, which either is not a part of their diet, or is unavailable because it is protected by a tough seed coat.

Homopterans

Ants are notorious for the habit of maintaining colonies of homopterans on plants (Figure 7). The principal families involved are the Membracidae or treehoppers, the Cicadellidae (Jassidae) or leafhoppers, the Psyllidae or “lerps,” the Fulgoridae or planthoppers, the Aphididae or plant lice, the Coccidae or soft scales, and the Pseudococcidae or mealybugs. Together they represent thousands of different interactions with ants, the majority still undescribed and unstudied. Interactions between homopterans and ants have been reviewed by Way (1963) and many are fundamentally mutualistic. The Homoptera secrete honeydew on which the ants feed. In return, the ants provide a number of vital services to the homopterans.

Homopterans take sap directly from the phloem through the slender mouthparts. Phloem contents are under several atmospheres of hydrostatic pressure and little effort is required for ingestion. However, the animals are capable of regulating their intake (Kennedy & Fosbrooke 1972). Changes in the chemical constitution of the sap occur during its passage through the homopteran gut so that when it becomes available to ants as honeydew it contains a variety of sugars, organic acids, alcohols, plant hormones, salts, vitamins, amino acids, and amides (Brian 1977). The main nitrogenous components of honeydew are amino acids, but these fluctuate widely according to the condition of the host plant (Mittler 1958; Llewellyn, Rashid, & Leckstein, 1974). However, a low nitrogen content in the sap stimulates the homopterans to increase the rate of feeding.

In this chapter I first examine fossil ants and plants and primitive living ants and angiosperms to try to reconstruct the origin and early evolution of ant–plant interactions. The fossil record is fragmentary and even extant primitive species may be only remotely related to those that were involved in the first interactions. This discussion is therefore speculative.

Ants and plants in the Cretaceous

It is now generally agreed that the flowering plants (angiosperms) had spread across most of the land masses of the world and had diversified dramatically during the early part of the Cretaceous period. By mid-Cretaceous, about 100 million years before present, this plant group was dominant among terrestrial vegetation (Raven 1977; Doyle 1978). The many suggestions for the causes of this comparatively rapid ascent include changing physical, climatic, and geographical conditions (Axelrod 1970), the rise of major insect pollinator groups (Takhtajan 1969; Crepet 1979), the appearance of avian and mammalian seed-dispersal agents (Regal 1977), the proliferation of herbivores (Ehrlich & Raven 1964; Burger 1981), and the evolution of novel plant secondary compounds (Swain 1977, 1978). In this chapter, I argue that another factor, the ants, contributed significantly to the success and adaptive radiation of the flowering plants.

Given that the fossil record places the angiosperm rise to dominance in the early to mid-Cretaceous, it is appropriate to ask when the ants began to flourish. The order Hymenoptera, to which the ants belong, first appeared in the early Triassic, perhaps 100 million years before the angiosperm accession.

Discussion of theoretical ecology, like discussion of ecology generally, is plagued by the problem of determining what it is in order to ascertain how it got that way or even what it should be. Consideration of theoretical ecology is difficult in that it requires conceptions of what is theory, what is ecology, and how they combine. Some recent writers on theoretical ecology resolved the problems I suggest with disarming simplicity. E. G. Leigh, Jr. (1968), for example, opened an article with a “historical sketch of ecologic theory,” moving directly to the “pioneers of mathematical ecology, Lotka and Volterra,” the founders in the 1920s of the “Golden Age of theoretical ecology” as it was described by Scudo and Ziegler (1978). Robert May (1974a), the editor of the first book, entitled simply Theoretical Ecology (May 1976) offered a similar history, but overlooked Lotka in writing, “Theoretical ecology got off to a good start in the 1920s with Vito Volterra's seminal and still central contributions.” This view of the alliance of mathematical theory and ecology had been described earlier in the volume Theoretical and Mathematical Biology (Waterman and Morowitz 1965): “There are few areas of biology where theoretical mathematical studies have had as much impact as they have had in ecology” (Morowitz 1965). The author of a volume entitled An Introduction to Mathematical Ecology wrote, “Ecology is essentially a mathematical subject” (Pielou 1969:v). Mathematical theory was described as the big advance in ecology over Elton's conceptual contributions to ecology (Christiansen and Fenchel 1977).

The rapid, even “revolutionary,” emergence of self-conscious ecology from the amorphous body of classical natural history and the overshadowing presence of experimental laboratory-based physiology, which was the dominant aspect of late 19th century biology, is a remarkable, and poorly studied, phase in the history of biology (Frey 1963a; Coleman 1977; Egerton 1976; Lowe 1976; Mclntosh 1976, 1983a; Cox 1979; Cittadino 1980, 1981; Tobey 1981). Oscar Drude, an eminent German plant geographer and a major influence on the development of plant ecology in America, aptly described the sudden recognition of ecology at the Congress of Arts and Sciences meeting at the Universal Exposition in St. Louis in 1904:

In spite of the fact that ecology was coined in 1866 and had been, unnoticed, in the literature since then, it was, as Drude stated, essentially unknown in 1890.

Whatever may be said of the origins of ecology in the Greek science of Hippocrates, Aristotle, and Theophrastus, or in 18th-century natural history as exemplified by Linnaeus and Buffon, or even in Darwinian evolutionary biology, its rise as a named and “self-conscious” discipline with its own practitioners was essentially in the last decade of the 19th century (Allee et al. 1949). Ecologists began to define ecology by doing it and recognizing that they were doing it.

An attempt to write a general account of the origins, development and current problems of ecology, even within the constraints noted below, might well be thought foolhardy. Ecology built upon traditions of natural history beginning in classical antiquity but developed as a science in the context of late 19th-century biology, natural history surveys, and conservation. It became widely known to the general public, often in distorted forms, only in the 1960s. It has been called polymorphic because it appeared and continues in numerous and different forms appropriate to the enormous variability and complexity of the things studied by ecologists. Until recently, ecology has not excited the interest of historians of science, and detailed historical studies of ecology or biographical works about ecologists are few. This volume was not, however written to fill the need for careful historical analyses of ecology and its relation to biology and to environmental concerns, although it leans heavily on those now available. It is an attempt to provide an account of the background of ecology and suggest its relevance to current problems of ecology as a science. It has an underlying assumption that some of the difficulties and conflicts now manifest in ecology can be better resolved if ecologists, particularly younger ecologists, become familiar with what went before them and their mentors and outside their immediate interests.

Like the word ecology, ecosystem was applied to a concept with a long history. It also had a number of competing synonyms when it was coined, suggesting that the time was ripe for its appearance, and there was a lag period between its coinage and its widespread incorporation into ecological science. The British plant ecologist Tansley (1935) introduced ecosystem in the context of a discussion of the superorganism concept of the plant community and succession which was developed around 1905 by F. E. Clements and was still being strongly advocated in the 1930s by the South African botanist John Phillips. Tansley commented that Phillips's articles “remind one irresistably of the exposition of a creed – of a closed system of religious or philosophical dogma,” a flavor not entirely missing from some later expositions of “systems ecology,” facetiously called “theological ecology” (Van Dyne 1980).

Godwin (1977) noted that Tansley, like many biologists of his era, was “remarkably unspecialized” and a man “of wide culture and familiarity with many sciences,” just what the ecosystem concept called for. Tansley was also well read in philosophy and psychology, having studied with Freud, which, perhaps, influenced his ecological thought. He defined ecosystem as

Ecology in its early years was sometimes decried as not a science at all but merely a point of view. After nearly a century of trying to erect a conceptual, methodological, and theoretical framework for the most complex phenomena encountered in nature, ecology was familiar only to a relatively small number of academic biologists and applied biologists, range managers, foresters, and fishery and game managers. These shared an overlapping, but not coincident, network of concepts, methodologies, professional associations, publications, funding sources, and concerns about the relations of organisms as populations and communities to their environment. In the wake of widespread recognition, in the 1960s, of the “environmental crisis,” ecology was abruptly thrust into the public arena and widely hailed as an appropriate guide to the relation of humans, as well as other forms of life, to their environment. Strikingly, ecology became a watchword, even in high political circles, just when Paul B. Sears, one of its most articulate practitioners and expositors, described ecology as “a subversive subject” (Sears 1964). Sears's point was that the view of nature derived from ecological studies called into question some of the cultural and economic premises widely accepted by Western societies. Chief among these premises was that human civilizations, particularly of advanced technological cultures, were above or outside the limitations, or “laws,” of nature (Dunlap 1980b).

The 18th century produced at least the beginning of a change in natural history from a view of nature as a divinely ordered, essentially static system following a providential mandate to a dynamic “historically changing” entity endowed with “self-activating” and “self-realizing” powers (Lyon and Sloan 1981). This conception of natural history persisted in the early 19th century in “Humboldtian science” (Cannon 1978). Its canon of “accurate measured study of widespread but interconnected phenomena . . . to find a definite law and a dynamical cause” was continued by the rising science of ecology in the late 19th century. A major theme of functional, experimental biology as it developed in the 19th century was its emphasis on progressive change as the most significant characteristic of natural phenomena (Coleman 1977). Whether ecology is seen as emerging from 18th- or 19th-century natural history, from 19th-century mechanistic physiology, or some amalgam of these, the conception sometimes encountered of early self-conscious ecology as descending from and embodying a static, typological, descriptive progenitor needs to be reconsidered. Certainly in the view of most of its early proponents and practitioners, the key word for ecology was dynamic. This was explicit in the writings of the leading figures of ecology as it became a self-conscious science. If they or lesser ecologists failed immediately to emphasize all of the ideas that later came to typify the “new” dynamic or functional ecology, it should not be assumed that first-generation ecologists were content with description as the aim of ecology.