Introduction

Biological and physical techniques that can serve as components of multitactic weed management strategies abound. Examples throughout this book illustrate how greater knowledge of ecological processes can maintain or improve crop yields while decreasing dependence on herbicides. We suggest that by reducing the need for herbicides, ecologically based weed management strategies can help farmers reduce their input costs, reduce threats to the environment and human health, and minimize selection for herbicideresistant weeds.

Despite the potential benefits of ecological weed management, most farmers in industrialized countries continue to rely heavily on herbicides, and the use of herbicides in developing countries is increasing. Many agricultural analysts question the ability of the world's farmers to produce enough food for a burgeoning human population without continued emphasis on herbicides and other agrichemical technologies. Some analysts argue that it will be possible to protect natural habitats and wildlife only by increasing production per unit of farmland through the intensive use of pesticides, synthetic fertilizers, genetically engineered seeds, and other purchased inputs.

In this chapter we examine reasons why ecological weed management has not been widely embraced and address whether ecological weed management is indeed consistent with the goals of increasing food security and protecting nature. We then suggest ways to promote research on ecological weed management. Finally, we address ways to foster ecological weed management on farms in both industrialized and developing countries.

Weeds from an ecological perspective

Weeds share certain ecological characteristics that distinguish them from other plants. Specifically, weeds are plants that are especially successful at colonizing disturbed, but potentially productive, sites and at maintaining their abundance under conditions of repeated disturbance. That is, weeds are the plants that thrive where soil and climate are favorable to plant growth, but disturbance frequently reduces competition among plants to low levels. Unlike previous conceptions of weediness (Baker, 1965; Harlan & de Wet, 1965; Buchholtz et al., 1967), this ecologically based definition lacks reference to humans and human disturbance. The species people refer to as weeds mostly existed prior to human disturbance, and the repertoire of behaviors that makes them invasive and persistent in human-dominated habitats largely evolved independently of human society. Nevertheless, as discussed in Chapter 10, human activities selectively modify weed characteristics such that weeds are becoming better adapted to human disturbance regimes.

The subcategory of weeds dealt with in this book consists of the weeds of agriculture – specifically, the plants that colonize and increase in the disturbances created by farming. These are sometimes termed agrestal weeds, as distinguished from the ruderal weeds of roadsides, waste piles, and other non-agricultural disturbances (Baker, 1965). Agricultural weeds share certain life-history characteristics that adapt them for life on farms (Table 2.1). The thesis of this chapter is that understanding life-history characteristics provides insights into how weed management practices work and how they can be improved.

Introduction

Cattle, sheep, goats, and other domesticated vertebrates graze more than 50% of the earth's total land area, 20% in managed pastures and 30% in rangelands (Snaydon, 1981). Animal production and cropland management are also frequently linked. Animals graze the herbaceous understory in tree crops and feed on residues and remnant vegetation in annual crop fields. Animal manures are applied to croplands, and pastures and forage crops are rotated with annual crops.

Domesticated herbivores can accentuate weed problems for humans. They disperse weed seeds (Chapter 2). They graze preferred species heavily, but leave unpalatable species to grow and reproduce. They compact soil around watering holes, at resting sites, and along trails, which fosters grazing- and trampling-tolerant unpalatable weedy vegetation. Introduced forage species naturalize to become weedy invaders (Low, 1997).

However, through managed grazing animals can also reduce weedy vegetation and promote desirable forage species. This chapter illustrates three principles for the use of livestock to reduce weeds in annual and perennial crops and on grazing lands:

1. A weed's susceptibility to control by grazing depends on its growth habit, its life cycle stage and the growing conditions at the time of grazing, and its palatability to different herbivore species. The identification of a weed's particular vulnerabilities to grazing contributes to understanding why it has become a problem or might become a problem. Whether a weed is vulnerable to control by grazing also depends on the other plant species in the same grazed area and their ability to tolerate and avoid grazing. In fact, a weed in one context may be a primary forage species in another.

2. […]

Introduction

One of the defining characteristics of an ecosystem is the diversity of plant species it contains. In agricultural systems, diversity of the dominant plant species – crops – can vary in both spatial and temporal dimensions. Crops can be sown in pure stands (sole crops), but can also be sown in multispecies mixtures (intercrops or polycultures), a practice that probably began with the development of tropical agriculture (Plucknett & Smith, 1986). Temporally, a crop can be sown continuously in the same field (continuous monoculture) or sown only intermittently, in sequence with other crops (rotation), a practice known from ancient Greece, Rome, and China (Karlen et al., 1994). Rotation sequences often contain only food, feed, and fiber crops, but may also include cover crops to improve and conserve soil during seasons when “main” crops are absent. In temperate areas, rotation cycles typically extend over several years, with only annual changes of crops, but in areas with long or continuous growing seasons, farmers may plant a sequence of several crops within a single year (multiple cropping), or overlap the late growth period of one crop with the interplanting and early development of another (relay cropping).

Spatial and temporal diversity in agricultural systems may also result from growing trees and shrubs with herbaceous species (agroforestry).

Of the many books that have been written about weed management, most have focused on the use of herbicides. This volume is different. Instead of providing information about chemical weed control technologies, the emphasis here is on weed management procedures that rely on manipulations of ecological conditions and relationships. By focusing on ecologically based methods of management, we have been able to provide in-depth treatment of subjects that most weed science books treat only briefly.

Although the reader will find much information on the ecology of weeds here, the primary purpose of the book is not to explain weed ecology. Rather, our intent is to elucidate the role of ecological principles in weed management. We believe that ecology can provide a theoretical basis for weed science, much as physics provides a theoretical basis for engineering and biology acts as the theoretical basis for medicine. Accordingly, throughout this book we show ways in which insights into ecological processes provide explanations for the successes and failures of weed management and avenues for developing better management strategies.

This volume could be used as a textbook for an advanced course in weed management, but it was not written primarily for that purpose. Rather, we have attempted to offer the reader a critical analysis and synthesis of the literature on ecological weed management and relevant aspects of weed ecology. Several goals motivated this review process. First, we wanted to identify clearly the principles that underlie ecological management practices.

Introduction

Physical removal of weeds by soil disturbance prior to planting, and by hoeing and hand-weeding during crop growth are undoubtedly the oldest forms of agricultural weed management. Farmers and agricultural equipment manufacturers continue to develop this ancient tradition of mechanical weed control through the refinement of hand tools and the invention of new tillage and weeding machinery. The purpose of this chapter is to explore the ways in which tillage before crop planting and mechanical weed control methods after planting interact with the ecology of weeds, and to use that understanding to suggest strategies for weed management.

Tillage and cultivation affect weeds in three distinct ways. First, they uproot, dismember, and bury growing weeds and dormant perennating organs. Second, they change the soil environment in ways that can promote germination and establishment of weeds or, less commonly, inhibit germination and establishment. Third, they move weed seeds vertically and horizontally, and this affects the probability that seedlings emerge, survive, and compete with the crop. The second of these effects was discussed in Chapter 2. The first and third are addressed in this chapter.

Each of the tools used for tillage and cultivation disturbs the soil in a unique way. In particular, tools vary with respect to their working depth and the degree to which they invert the soil column, break up soil aggregates, and shake weed roots free from the soil.

Introduction

Many cultural practices, including crop density, arrangement, planting date and choice of cultivar affect the crop's ability to compete with weeds. However, most recommendations for the planting of crops are based on the assumption that weeds are absent. This is a result of the scientific and economic context in which recommendations are developed. Variety trials, fertility rate trials, and many other agronomic experiments are usually run in weed-free conditions to avoid the confounding effect of weed competition. For the agronomist or horticultural scientist, keeping a particular experiment free of weeds is a practical possibility. Given the high spatial and temporal variability in density and composition of weed communities, a weed-free trial may also be the easiest way to generate results that are applicable over a wide area. In addition, weeds generally decrease yield regardless of other parameters. Consequently, weeds are usually excluded from experiments unless they are specifically the object of investigation. However, weed-free fields are rarely practical on the farm, and as explained in the following sections, the presence of weeds generally changes the optimal choices for cultural practices relative to those developed in weed-free conditions.

The central thesis of this chapter is that the density, arrangement, cultivar, and planting date of the crop that maximize the rate at which the crop occupies space early in the growing season usually minimize competitive pressure of weeds on the crop.

Introduction

Agriculture is the process of managing plant communities to obtain useful materials from the small set of species we call crops. Weeds comprise the “other” set of plant species found in agroecosystems. Although they are not intentionally sown, weed species are well adapted to environments dominated by humans and have been associated with crop production since the origins of agriculture (Harlan, 1992, pp. 83–99).

The ecological role of weeds can be seen in very different ways, depending on one's perspective. Most commonly, weeds are perceived as unwanted intruders into agroecosystems that compete for limited resources, reduce crop yields, and force the use of large amounts of human labor and technology to prevent even greater crop losses. In developing countries, farmers may spend 25 to 120 days hand-weeding a hectare of cropland (Akobundu, 1991), yet still lose a quarter of the potential yield to weed competition (Parker & Fryer, 1975). In the USA, where farmers annually spend $6 billion on herbicides, tillage, and cultivation for weed control (Chandler, 1991), crop losses due to weed infestation currently exceed $4 billion per year (Bridges & Anderson, 1992).

At the other end of the spectrum, weeds can be viewed as valuable agroecosystem components that provide services complementing those obtained from crops. In India (Alstrom, 1990, pp. 25–9) and Mexico (Bye, 1981; Mapes, Basurto & Bye, 1997), farmers consume Amaranthus, Brassica, and Chenopodium species as nutritious foods before crop species are ready to harvest.

Introduction

Weed scientists usually cite pervasive crop yield losses due to weeds and substantial direct and indirect costs of weed control to justify research and extension budgets (see Chapter 1). Reductions in costs and yield losses should also be used to evaluate the progress of scientists in solving weed problems. Ultimately weed costs to agriculture are determined by how farmers and ranchers manage weeds, not by papers published or field days organized. In temperate and tropical regions, field crop farmers who use mechanization, cattle ranchers, dairy farmers, vegetable and fruit growers, and smallholders on hillsides all devote time and resources to weed management. What is the role of research and extension in enabling this wide diversity of farmers to manage their weeds better?

This chapter examines the implications of farmer–extensionist–scientist interactions for the development of improved weed management. The first sections review historically how humans have learned to manage weeds. The chapter then analyzes scientist, extensionist, and farmer perspectives on weeds. The final sections describe how farmers, extensionists, and scientists can collaborate to develop field- and farm-level weed management strategies better adapted to weed patchiness and uncertainty. Case studies from the USA and Central America illustrate possible working relations among scientists, extensionists, and farmers.

Introduction

One of the distinguishing characteristics of terrestrial plants is that they spend a significant portion of their lives unable to travel farther than they can grow. As a consequence of the sessile, fixed root habit, the resource environment in which plants grow and reproduce is a very local phenomenon and interactions among neighboring plants are common (Harper, 1977, p. 4). The sessile habit makes it possible to suppress weeds through manipulations of soil conditions.

Given the similarity of most terrestrial plant species in their requirements for sunlight, water, and nutrients, it is not surprising that weeds compete with crops for resources and reduce crop yields. Conversely, crop plants exert a large competitive effect on associated weeds (see Chapter 6). A key insight from ecology, however, is that outcomes of competitive interactions between plants are highly dependent on environmental conditions, especially soil-related factors. As Harper (1977, p. 369) noted, “there is a very extensive literature in which it is demonstrated repeatedly that the balance between a pair of species in mixture is changed by the addition of a particular nutrient, alteration of the pH, change in the level of the water table, application of water stress or of shading.”

Ecological studies have also revealed that plant abundance and distribution are affected by the availability of appropriate sites for germination and establishment (Grubb, 1977).

Introduction

Most weed management practices are motivated by short-term goals: reduction of weed impact on the current crop and prevention of seed production that could pose problems in succeeding crops. A slightly longer perspective may enter considerations of crop rotation and its impact on weeds (see Chapter 7), but weed management planning horizons of farmers rarely exceed five years. In contrast, important phenomena relating to weed diversity, community composition, and weed evolution affect weed communities on time scales of five years to centuries. In principle, these processes could be managed, though at present they largely are not. This chapter explains why long-term management of these phenomena may be needed, and outlines some tentative strategies.

The nature of long-term changes in weed species and communities has not been well documented and proposals for managing these changes are therefore necessarily speculative. Consequently, most of this chapter focuses on the ecological and evolutionary processes governing the changing nature of weed species and communities, with most suggestions for management reserved for the final sections. Three general points will be made.

First, evolutionary and community responses of the earth's flora to the resources available in farm fields leads to a continuous increase in the global diversity of agricultural weeds. Simultaneously, long-distance colonization events and local spread of species to new locations create a tendency toward increase in regional and local weed diversity.

Introduction

Weeds and other plant species are susceptible to attack by a diversity of invertebrate herbivores and pathogens. Virtually every plant organ provides a niche for some type of insect, mite, nematode, fungus, bacterium, or virus (Harper, 1977, p. 484). Protection of crop plants from these organisms is a major issue in crop production. Conversely, the promotion of herbivory and disease to suppress weed recruitment, growth, and reproduction is a major objective of biological control programs.

Biological control of weeds requires that sufficiently high densities of herbivores and pathogens are present when weeds are at susceptible developmental stages. For this to happen, herbivores and pathogens used as biological control agents must be well adapted to abiotic components of the environment, such as temperature and precipitation regimes (Crawley, 1986; Cullen, 1995). To control weeds effectively, they must also largely escape the effects of predation, parasitism, disease, competition, and chemical interference (Newman, Thompson & Richman, 1998).

Three approaches are used in efforts to regulate weed populations with herbivores and pathogens (Andow, Ragsdale & Nyvall, 1997). Conservation methods involve modifying the environment to retain or increase populations of resident control agents and intensify the damage they inflict on weeds. Inoculation methods involve introducing relatively small numbers of biological control agents that will suppress a target weed species as their populations establish, increase, and disperse.

Seeds

Seed size

The size of seeds interests comparative ecologists because it is so variable among species. The dry mass of seeds ranges over at least six orders of magnitude across species of tropical rain-forest tree. The Melastomataceae and Rubiaceae include tropical tree species with seeds of dry mass as little as 20 µg (Metcalfe & Grubb 1995; Grubb & Metcalfe 1996). At the other extreme, the seeds of a number of trees, notably legumes, approach 100g dry mass. Within any tropical forest site, most studies have shown ranges of least five orders of magnitude for tree seed mass (Hammond & Brown 1995; Metcalfe & Grubb 1995; Grubb & Coomes 1997; Lord et al. 1997). Of course, mass is a volume-dependent property and so will rise with the cube of the linear dimensions involved, which will rapidly exaggerate size differences between species, but a million-fold range in offspring size is still enormous when compared with animal groups.

Seed size might be under allometric control of other characters. There is evidence of correlations with other size variables. The difficulties of small plants producing big seeds and of small fruits containing big seeds will probably always lead to some degree of positive correlation between plant size and seed size (Fig. 5.1). For tropical trees, a number of studies have shown increases in seed size with adult stature within a particular forest (Hilty 1980; Foster & Janson 1985; Metcalfe & Grubb 1995; Hammond & Brown 1995; Kelly 1995; Grubb & Coomes 1997).

Trees: form, mechanics and hydraulics Tree stature

Individual trees of a large range of size are to be found in the tropical rain forest (see the profile diagram in Fig. 2.1). Each tree species also has a characteristic size at maturity and species are often referred to various stature classes, such as understorey trees, canopy trees and emergents, but, as Figure 2.2 shows, there is no discrete clustering of species in size classes. Maximum diameter for species on 50 ha in Pasoh Forest, Peninsular Malaysia (Fig. 2.2), was approximately a truncated log-normal distribution with a modal class in the 10–20cm maximum dbh range. A breakdown of species into height classes (Table 2.1) shows about half of the tree species to have maximal heights of 20m or less. The tallest trees at Pasoh probably reach to about 60m in height, showing that the community of understorey specialists in the forest is as rich in tree species as that of the canopy, and is definitely richer on a unit depth basis. A similar pattern is seen in the forest at La Selva, Costa Rica (Hartshorn 1980), and on Barro Colorado Island, Panama (Hubbell & Foster 1992).

The vertical distribution of tree species diversity in the forest may reflect the relative illumination at different heights above the forest floor. At Pasoh, relative illumination increases logarithmically with height (Fig. 2.3). The understorey species in the bottom 20m of the forest rarely receive more than 5% of the radiation arriving at the top of the canopy.

Trees make a forest: they are both the constructors and the construction. To understand the forest we must know about the trees. This book is about the trees of the tropical rain forest. It was written with the aim of summarising contemporary understanding of the ecology of tropical rain-forest trees, with particular reference to comparative ecology. The analysis of patterns of variation among species is a valuable technique for identifying possibly adaptive trends and evolutionary constraints. It may also provide a means of classifying species in ecological terms. A workable ecological classification might mean that the rain-forest community could be conceptually simplified and made more amenable to analysis.

The organisation of the book follows the life cycle of a tree. The living, growing mature tree is introduced with reference to form and process. Reproduction, including pollination and seed dispersal, follows. Then come consideration of seed germination, seedling establishment and growth, and the completion of the life cycle. At each stage a range of different characteristics and phenomena relevant to tree species growing wild in the tropical rain forest are considered. I have tried to give some idea of what is typical, and what is rare, the range and central tendency exhibited among species, and whether discrete groupings, or a continuous variation, are observed within the forest, and also whether one character tends to be correlated with another.

Height at maturity

Forest ecologists have generally understated the rather obvious distinction between species in their height at reproductive maturity. Species are often divided into height classes in ecological analyses, but it seems more with the purpose of comparing like with like within the stature groups, than of making comparisons among the groups. However, some general trends do emerge from the literature, and these are summarised in Table 6.1 as a series of characteristics of small-statured species in comparison to those of larger size at maturity.

The factor that appears to determine mature height is the size at onset of reproduction. Small-statured tree species start reproduction at smaller size (Thomas 1996b; Davies & Ashton 1999). Thomas (1996b) found this both in absolute terms and in size relative to asymptotic height. Allocation to reproduction probably requires a compromise in height growth rate, and hence these trees are left behind in the height growth race by the taller-growing species. Thomas (1996b) found that the change in the slope of the H–D regression from linear to asymptotic for a species generally coincided with the onset of reproduction.

Reproduction at small size need not necessarily imply reproduction at an earlier age. A very shade-tolerant understorey tree may grow very slowly and so be as old as, if not older than, a canopy tree that grew up in a gap. However, it is likely that on average small-statured trees do reproduce earlier in life than large-statured ones.