2006), but broad interest in BIA is demonstrated by sales of more than 500,000 copies of a popular manual in its seventh edition (Jeavons 2006), by adoption of BIA by small-scale farmers in 130 countries around the world, and by a variety of papers presented at recent symposia (Miles et al. 2002, Doran et al. 2006, Omondi et al. 2006, Pia 2006, Mbugwa et al. 2006, Beeby et al. 2006, Bouyouris et al. 2006).

BIA systems use mixed plantings to increase land use efficiency (Jeavons 2006). Mixed plantings have been well-studied (e.g. Brown et al. 1985, Hart 1986, Theunissen 1997), and found to make more efficient use of land (Jolliffe 1997) and water (Kanton and Dennett 2004) than pure stands. According to the ecological concept of niche differentiation, mixtures are more productive than pure stands when mixing replaces stronger within-species competition with weaker between-species competition (Willey 1979, Vandermeer 1989, Jolliffe and Wanjau 1999). Since the choice of crops used in a mixture presumably has some effect on whether, or not, this occurs, certain crop combinations might be expected to consistently offer yield advantages over pure stands. Using the language of human relationships metaphorically, BIA manuals typically recommend specific crop combinations as beneficial “companions,” and discourage others as “antagonists” (Riotte 1975, Jeavons 2006). For example, basil (Osimum basilicum L.) mixed with tomato (Lycopersicon esculentum Mill.) is said to improve tomato growth, but brassicas, such as Brussels sprout (Brassica oleracea L. var. capitata), are considered antagonistic to tomato; brassicas are said to benefit from mixing with aromatic herbs. In the language of Riotte (1975), tomatoes love basil but hate Brussels sprouts. The reputed effects of mixing certain plant species are frequently vague (e.g. “tomatoes and all members of the brassica family repel each other,” Riotte 1975), and the rationale for these recommendations is often unclear: They are based on gardener experience and on laboratory assays involving the analysis of crystal formations produced by evaporating aqueous mixtures of plant extracts and copper chloride (Pfeiffer 1931, Riotte 1975). Although these assays show similar results when interpreted by blind panels or image analysis software (Andersen et al. 1999), a plausible mechanistic explanation of the relationship between copper chloride crystal formation and plant compatibility has not been proposed or tested. This may contribute to the fact that companion planting recommendations continue to be widely-circulated in the popular press, but largely untested in the horticultural literature.

A few studies published in the entomological literature report tests of specific mixtures recommended as companions in BIA manuals for their supposed ability to deter insect pests (Finch et al. 2003, Held et al. 2003). While a large body of literature supports the idea that crop diversity reduces arthropod pest pressure (Andow 1991, Smith and McSorely 2000) specific combinations recommended for BIA systems have not yet shown greater benefits than other mixtures.

Certain BIA practices may not be appropriate for some environments. According to Jeavons (2006), on well-structured soils the practice of ‘double digging’ -- deep cultivation with a spading fork -- “is not needed to maintain significant yields and may even deplete the quality of the [soil] structure.” Indeed, the only published peer-reviewed study of double digging was conducted in well-drained, fertile soil with sufficient rainfall, and found that the practice offered no yield benefit (Holt and Smith 1998). A more recent preliminary study reached a similar conclusion, leading to call for factorial studies to evaluate the effect of individual BIA practices in different environments based on the concern that “promoting such a highly labor constraining practice as an integral part of the BIA package could preclude farmers from adopting the other, potentially more useful BIA components (i.e. compost use and close