The relationship of on-farm biodiversity with food web structural properties and ecosystem services remains to be explored. To understand the functional significance of species richness and ecosystem complexity of rice farms, I examine here the architectural properties of rice food we from West Bengal, based on replicated plots of folk variety (organic) and modern (chemicalised) rice systems. All rice food webs, constructed from observational data collected over three years, show prominent scale dependence of dietary
links, link density, web height, diversity of natural enemies to pests, predator-pest ratio, and the numbers of omnivores and omnivory levels. Organic folk rice webs tend to have greater mean species richness, predator diversity, predator-pest ratio and chain length than modern rice farm webs, yet both systems show homogeneity of distribution of the web properties. Analyses of 16,400 computerized analog webs, following non-random rules of species association drawn on real-life, seasonally distinct rice food webs, validate the robustness of conclusions.

This study, based on empirically observed ecosystems as well as their simulated analogs, confirms scale dependence of (a) link density, (b) basal species fraction, (c)predator density (PR/PT), (d) web height, (e) omnivory, and (f) levels of omnivory. Furthermore, this study links the rice
food web architectural properties to selected agroecosystem functions, such as the diversity of predators and omnivores.

While predation on herbivores can effectively control pest populations on host plants, predation from higher
trophic levels on the herbivores and intermediate predators may have varying effects on suppression of herbivore
populations (Rosenheim and Corbett 2003). This study shows that the rice field houses an abundance of actively
foraging omnivores - coccinelid beetles, ants, predatory wasps, odonates, fireflies, ranid frogs that climb the rice plant to eat insects, Calotes versicolor, insectivorous birds
like myna, pond heron and bee eater. Unless eradicated by the use of pesticides (as in MV systems), the rich predator diversity is likely to override the pest species richness (as in FV systems) with a higher PR/PT ratio. Recent empirical studies indicate that actively foraging omnivores in complex ecosystems are likely to enhance herbivore suppression
(Rosenheim and Corbett 2003). A comparison of crop yield loss (due to pest damage) in FV vs. MV systems (Deb,
unpublished data) is beyond the scope of this paper. However, analyses of the food web architecture here do reveal (i)
a significant increase in predator diversity with both S and pest diversity (PT), (b) a direct relationship of the predator richness (PR) with the pest richness (PT) in both the real and model webs, and (c) a strong direct relationship between S
and the ratio of predator to pest diversity (PR/PT) (Table 4). These three relationships conjointly indicate that predator diversity tends to increase with both system size and pest
diversity, and the ratio of predator- to pest richness declines progressively with on-farm biodiversity.
Scale dependence of PT/ PR ratio in this study is consistent with the notion that greater on-farm biodiversity
can enhance (1) more natural enemies of insect pests, and (2) biotic compensation for the absence of synthetic pesticides (Letourneau and Bothwell 2008). Greater biodiversity may
also contribute to the stability and resilience of the farm ecosystem against large perturbations from climatic vagaries or pest and pathogenic outbreaks (Kinzig et al. 2002; Tilamn et al. 2002). Recent empirical and theoretical studies (Dunne
et al. 2002, 2004; Worm and Duffy 2003; Tilman et al. 2006) suggest that enrichment of biodiversity and complexity improves temporal stability and resilience, and that the loss of highly connected ontospecies will tend to induce higher levels of secondary extinctions than loss of random taxa. Furthermore, greater complexity of food webs may “intrinsically lower the likelihood of chaotic community
dynamics” (Fussmann and Heber 2002).
Biodiversity is depleted in modern rice production systems through monocultures and the use of pesticides and
herbicides, which simplify the on-farm ecosystem structure.
Systemic pesticides are known to impact severely on populations of a large number of arthropod and vertebrate
predators, leading to a collapse of vertical diversity of the rice food web (Heong and Schoenly 1998; Heong et al.
2007). This study indicates that more species-rich farm ecosystems, with longer food chains and more trophic
omnivores, tend to have a greater diversity of predators than do species-poor, simplified agro-
ecosystems. Less ontospecies
diversity in MV systems is associated with truncation of web complexity (e.g. shorter WH), which may have adverse impacts on ecosystem productivity (Memmott et al. 2005; Duffy et al. 2007), via loss of supporting ecosystem
services (e.g., generation of soil fertility, pest control, pollination)
(Tilman et al. 2002; Hooper et al. 2005; Tscharntke et al. 2005; Heong et al. 2007). In light of this understanding, it is imperative to put a strong policy emphasis on enhancing biodiversity and ecological complexity in agroecosystems.