There are many definitions of agroforestry, most of which are based on early ideas presented by Nair (1989, 1993a,b), all relating to the many ways of integrating trees into farming systems as stand-alone agronomic technologies (Atangana et al., 2014); some of which are described in Chapter 4, (Cooper et al., 1996). These can be modified by the use of different species that suit the social, economic, and physical environment of the site. The magnitude of this variability has led to Nair (2014) saying that agroforestry cannot be defined exactly, but that it basically “refers to the purposeful growing of trees and crops in interacting combinations for a range of objectives including a variety of products and commodities and a vast array of environmental and other ecological services.”

The difficulty I have with this technological approach to defining agroforestry is that it overlooks the ecological functions that underpin agroforestry in ways that allow it to both enhance the environmental sustainability, as well as the productivity of farming systems. The functional definition of agroforestry, presented here (Chapter 1 [Leakey, 1996]), which is independent of any particular practice, agronomic system or species combination, was accepted and adopted by the ICRAF in its Medium-Term Plan (ICRAF, 1997). This definition sees agroforestry as applied agroecology in which the planting of trees in any combination or configuration initiates phases of an agroecological succession with increasing ecological integrity. Within mixed species plantings, especially if some are tall perennials, niches are formed above- and belowground for natural organisms to colonize. Through their interacting life cycles and food webs, these organisms create, support and maintain the ecological equilibrium between species, and importantly, enhance the function of the nutrient, carbon and water cycles. Furthermore, over time and with increasing scale, the landscape becomes a complex mosaic of patches in different phases of ecological succession, as in natural ecosystems. Both these processes are akin to the normal dynamics of natural ecosystems, processes which are at the heart of ecological and environmental sustainability.

The niche-forming principle within this functional definition also encompasses a mechanism for enhancing the social and economic sustainability of farming systems, as Leakey (1999b) suggested that farmers should enrich these agroecosystems by filling as many niches as possible with trees and other plants producing useful and marketable products (the “planned biodiversity”). Enrichment in this way increases farm total productivity while further creating niches for colonization by natural organisms—the “unplanned biodiversity.” Productivity can then be further intensified by the domestication of these useful, marketable and culturally important indigenous species, especially for nutritious food products to support rural livelihoods and local trade, as described in Section 3. This intensification then provides an added incentive for farmers to practice agroforestry. These concepts are examined further in Sections 4 and 5.

At the turn of the millennium relatively little was known about how to harness the ecological benefits by the deliberate enrichment of farming systems through agroforestry at either early or late phases of agroecological succession. So, there were many unanswered questions (Leakey, 1999b):

To address some of these questions, Leakey (1999b) called for three types of research:

To date, prototype systems (best-guess species combinations) have mainly been used in soil-plant interaction studies for productivity (Szott et al., 1991) and have not been widely used to examine any of the previous ecological questions, although Matos et al. (2003) have looked at the configuration of mixed species plots to investigate the susceptibility of mahogany to damage by Hypsipyla shoot tip moths. To define some ecological principles based on hypotheses to elucidate some of the relationships between planned and unplanned biodiversity, agroecological functions and total productivity, Leakey (2012b; 2014e) has presented experimental designs (Figs. 2.1 and 2.2). However, complex and long-term experiments of this sort require a multidisciplinary team of scientists and consequently are very costly and demanding. This has, so far, been a major obstacle to progress. Thus it has been the third of these approaches where most advance has been made, as seen in Chapter 2 (Leakey, 2014e). This review summarizes work which has thrown light on some key aspects of the agroecological function of early (or pioneer) stages of succession; and late (or mature) stages of succession. In the former, leguminous trees and shrubs importantly improve soil fertility and structure through the symbiotic nitrogen-fixing bacteria on their roots, while initiating the diversification of the agroecosystem both above- and belowground, while in the latter the predominant focus has been on cocoa and coffee agroforests in Latin America and Indonesia in which shade trees provide roosts for insectivorous birds and bats, habitat for ants and other small predators that prey on pests and pathogens.

In early-stage agroecosystems, relatively simple crop combinations between legumes and some grasses have been found to either deter pests (including some parasitic weeds) or to attract the predators of pests themselves. In late-stage phases of agroecological succession, the studies have taken two main forms: enumeration of different organisms (Table 2.1), and a start to unravel the complex interactions of different food webs (Table 2.2). This is beginning to provide evidence that explains the important role of trees in agroecosystem functions. This role is founded on the effects of the perennial above- and belowground structure of trees and tree canopies to create ecological niches for an extraordinary array of organisms, from the minutest microbe to the largest top predators and herbivores (including Homo sapiens). It is the level of activity of the food chains/webs and life cycles of these organisms that embodies the ecological balancing trick between organisms, and so determines the sustainability of the system. At all agroecological phases there is growing evidence of the benefits of species diversity on the success of the nutrient, water and carbon cycles that are crucial for healthy agroecosystems. It is the functioning of the carbon cycle resulting from the integration of woody perennials into agricultural and multifunctional landscapes that gives agroforestry the ability to mitigate climate change by both the reduction of greenhouse gas emissions and the sequestration of carbon in standing biomass (especially of perennial plants) and soil organic matter (van Noordwijk et al., 2011).

To update the literature presented here, there has been growing research activity to understand the dynamic interactions of organisms in agroforestry systems (Bennett et al., 2015; Garbach et al., 2014; Lavelle et al., 2014; Perfecto et al., 2014; Leakey, in press) and in landscapes (Scherr et al; 2014). Understanding the roles of this biodiversity is critical to understanding how to manage agroecosystems sustainably. This is needed in order to both capture the benefits of functioning agroecosystems that are important for the production of the food and numerous other useful products that provide the livelihoods of the farming households; and to promote the international public goods and services (biodiversity conservation, watershed protection, mitigation of climate change, etc.) important for humanity. In Chapters 37 and 39 (Leakey, 2014 and Leakey and Prabhu, 2017), we will see further evidence that the trade-offs that some believe to occur between production and biodiversity conservation (Garnett et al., 2013; Godfray and Garnett, 2014) can be avoided (and certainly are not inevitable) to create more sustainable farming systems.

Finally, to conclude, I think we can emphasize that it is the flexibility provided by the diversity of species and practices accorded by the ecological concept of agroforestry that gives it a unique advantage over what are often the rather prescribed land-management practices of conventional modern farming systems—which have been subject to economic boom and bust, environmental degradation and social deprivation. The delivery of a highly adaptable generic model to address and reverse these undesirable outputs of many farming systems will be presented in Chapters 31 and 34 (Leakey and Asaah, 2013; Leakey, 2013). This model provides much needed hope for the future (Leakey, 2014f). Importantly, it recognizes that improved agroecological functions alone are not sufficient to deliver the 2015 Sustainable Development Goals, and that farmers also need a good source of income to allow them to intensify and maximize production and to improve their livelihoods (Fig. 2.3, Leakey, 2012a; 2013). In the final chapter of this book (Leakey, 2017i) we will see that global sustainability has to embrace Homo sapiens within the functioning of intensified agroecosystems.