Postdispersal Weed Seed Predation and Invertebrate Activity Density in Three Tillage Regimes

Abstract

Field experiments were conducted near Boone, IA, to quantify postdispersal seed predation of common lambsquarters and common waterhemp in corn (2003) and soybean (2004) managed with conventional, reduced, and zero-tillage systems. Seed predation in each tillage regime was quantified using selective exclusion treatments during July through September 2003 and June through October 2004. In addition, the activity density of ground-dwelling invertebrates was estimated with pitfall traps. Choice and no-choice feeding trials were conducted in the laboratory using the most abundant weed seed predators found in the field to determine seed preferences of the potential predator organisms. The greatest seed loss occurred during July and August. In 2003, seed predation was lower in zero tillage than in conventional and reduced tillages, but no differences in seed predation between tillage regimes were observed in 2004. Maximum seed predation for common lambsquarters was 53% in 2003 and 64% in 2004. Common waterhemp seed predation reached 80% in 2003 and 85% in 2004. The majority of seed predation was by invertebrate organisms. The most common invertebrate species captured with pitfall traps were field crickets (Gryllus pennsylvanicusDe Geer [Gryllidae, Orthoptera]) and ground beetles (Harpalus pensylvanicusBurmeister [Coleoptera, Carabidae]). In 2003, field crickets were relatively more abundant in conventional and reduced tillage than in zero-tillage plots. In 2004, field crickets were more abundant in the reduced tillage than in the other two tillage regimes. No differences were detected for ground beetles among tillage regimes (P = 0.57). Choice and no-choice feeding experiments confirmed the preferences of field crickets and ground beetles for common lambsquarters and common waterhemp seeds over the larger seeds of giant foxtail and velvetleaf. Under field conditions, the activity density of field crickets was a significant predictor of common lambsquarters (r2= 0.47) and common waterhemp (r2= 0.53) seed predation. Positive relationships were also detected between the activity density of ground beetles and common lambsquarters (r2= 0.30) and common waterhemp (r2= 0.30) seed predation. This research demonstrated that weed seed predation is an important component affecting weed seedbanks and that crop management practices that favor the activity of predators such as field crickets or ground beetles could influence weed populations. Also, the results suggested that tillage is more important in determining the number of weed seeds available on the soil surface to predators than directly affecting predator activity density.