INTRODUCTION
Conservation tillage and direct seeding mulch-based management strategies have shown success in the preservation or restoration of biodiversity of both soil fauna and microflora (Blanchart et al., Reference Blanchart, Villenave, Viallatoux, Barthès, Girardin, Azontonde and Fellera2006, Reference Blanchart, Bernoux, Sarda, Siqueira Neto, Cerri, Piccolo, Douzet, Scopel and Feller2007; Brévault et al., Reference Brévault, Bikay, Maldes and Naudin2007; Marasas et al., Reference Marasas, Sarandon and Cicchino2001; Maria de Aquino et al., Reference Maria de Aquino, Ferreira da Silva, Mercante, Fernandes Correia, Fatima Guimaraes (de) and Lavelle2008; Wilson-Rummenie et al., Reference Wilson-Rummenie, Radford, Robertson, Simpson and Bell1999). Mulch serves as a suitable refuge for soil and litter organisms by protecting the habitat against water and wind erosion, regulating environmental conditions and maintaining food availability (Kladivko, Reference Kladivko2001; Stinner and House, Reference Stinner and House1990; Thorbek and Bilde, Reference Thorbek and Bilde2004; Tian et al., Reference Tian, Brussaard and Kang1993). In addition, mulch may significantly reduce insect infestations. For example, mulch interferes with the visual orientation of winged aphids during host plant location (Doring and Chittka, Reference Doring and Chittka2007). However, the effects of a direct seeding mulch-based system on crop pests are difficult to predict (Stinner and House, Reference Stinner and House1990; Ratnadass et al., Reference Ratnadass, Michellon, Randriamanantsoa and Séguy2006). For example, Madagascar black beetles (Heteronychus spp.) are reported to cause significant damage to mulch-grown rice at the onset of the growing season (Ratnadass et al., Reference Ratnadass, Michellon, Randriamanantsoa and Séguy2006). Conversely, mulch reduces the infestation of soil-inhabiting herbivores such as the Colorado potato beetle (Leptinotarsa decemlineata) by disrupting the emergence and migration of adults (Teasdale et al., Reference Teasdale, Abdul-Baki, Mills and Thorpe2004).
Cotton (Gossypium hirsutum) occupies more than 30% of the agricultural landscape in northern Cameroon and is cultivated by roughly 300 000 farmers (Mbetid-Bessane et al., Reference Mbetid-Bessane, Havard and Djondang2006). Traditional soil management systems include systematic exportation or crop residue burning and reductions in fallow periods. These practices have led to the large-scale depletion of organic matter and soil degradation. As a result, a consistent decrease in cottonseed yields has been observed for over a decade in conventional cotton cropping systems. To increase crop yield sustainability, national research programmes and extension services have implemented soil restoration measures through the ‘Eau Sol Arbre’ Project. A no-till with mulch soil management strategy is being promoted. The recommended mulch cover applied to cotton comprises plant residues from the previous crop (usually cereal straw) intercropped with a cover crop (Naudin et al., Reference Naudin, Husson, Rollin, Guibert, Charpentier, Abou Abba, Njoya, Olina and Seguy2003). The introduction of this soil management strategy significantly restores soil fertility (Erenstein, Reference Erenstein2003; Ito et al., Reference Ito, Matsumoto and Quinones2006); however, few detailed studies have focused on mulch pest dynamics, particularly at the cotton seedling stage.
The objective of this study was to evaluate the levels of damage caused by major cotton seedling pests, including soil arthropods, soil fungi and aphids (Aphis gossypii). We hypothesized that the presence of mulch (i) would support a higher density of phytophagous arthropods, particularly millipedes (Julidae), as well as pathogenic fungi that cause severe damage to cotton seedlings, and (ii) would reduce early aphid infestations. The vigour of cotton seedling stands and arthropod damage were assessed according to soil cover (with or without mulch) and seed dressing (with or without insecticide or fungicide) treatments. The results provide valuable data to use in designing effective crop management practices for cotton farmers interested in direct seeding mulch-based systems.
MATERIAL AND METHODS
Experimental design
Two independent field experiments were conducted during the 2001 and 2002 growing seasons at the IRAD and SODECOTON Research Stations near Garoua (9°23′N, 13°45′E). In both experiments, three factors were studied: (i) presence or absence of mulch, (ii) seed dressing with Marshal 35 DS, a systemic insecticide (carbosulfan, carbamate with contact and stomach action to control a wide range of soil dwelling pests, including millipedes, 3 g active ingredient (a.i.) kg−1 lint seeds), and (iii) seed dressing with Caltir 80 WS, a broad-spectrum fungicide (thiram, dimethyldithiocarbamate with protective contact action to prevent damping-off from Pythium and Fusarium spp., 1 g a.i. kg−1 lint seeds). The study employed a total of eight treatments designed from the following combinations: non-tilled soil with (M) or without (S) cover-crop mulch, seed dressing with insecticide (I), fungicide (F), insecticide and fungicide (IF) or untreated control (C). Six and five replicates of each treatment were applied to six 4.8 m wide and 10 m long rows, in 2001 and 2002, respectively. The experiment was designed to apply a randomized complete block for subsequent data analysis.
Cover-crop mulches included Calopogonium mucunoides (Fabaceace) in the 2001 study, and Brachiaria ruziziensis (Poaceae) in the 2002 study, intercropped with maize established late in the previous season. Seven days prior to cotton planting, living plants were killed by the application of a post-emergence herbicide (glyphosate, 720 g a.i. ha−1) on the entire experimental surface; the residue was left on the soil surface. A superficial scraping was performed to clean the plots assigned to direct seeding in the absence of mulch. A previously coated 10 l vessel with a lid was used to dress 2 kg of seeds with a powder formulation. The cotton variety ‘Irma A1239’ was planted on 17 July 2001 and 22 June 2002. Twenty-five mounds per row were constructed with special hoes equipped with a flap to obtain a 2 cm homogeneous planting depth. Five seeds were deposited in each of the 25 mounds (40 cm between successive mounds).
Sampling and data analysis
Cotton seedling stands (i.e. number of seedlings) were observed at 10, 15 and 20 days after sowing (DAS) along the three central rows of each plot, to avoid interactions with neighbouring plots. Damaged cotton seedlings from pests evidenced by a wounded or dried hypocotyl were observed at 7, 10 and 15 DAS along the same rows. Major injuries to cotton seedlings are observed at this period. Dead millipedes, the most harmful arthropods for cotton seedling, were collected daily within the same rows. To assess the effect of treatments on the plant vigour, the height of 20 randomly selected plants along the three central rows was measured at 20 DAS. In addition, aphids were counted from five terminal leaves from four sets of five consecutive plants per plot. Data were analysed using analysis of variance (ANOVA), with SAS GENMOD for binomial and negative binomial distributions and SAS GLM for Gaussian distributions (SAS Institute, 1989).
RESULTS
Cotton seedling stands
In both experiments, the presence of mulch cover crop negatively impacted cotton seedling stand by 13.3 and 13.9 % at 20 DAS in 2001 (C. mucunoides) and 2002 (B. ruziziensis), respectively (Table 1). In both experiments, seedling stand increased significantly when the insecticidal seed dressing was applied (Table 1). In 2002, a significant interaction between mulch and insecticidal seed dressing was observed: insecticide seed application had a greater effect on seedling stand in mulched plots (+25.2%) than in non-mulched plots (+9.8%). In 2002, the fungicidal seed dressing significantly increased cotton seedling stand (Table 1). In addition, a significant interaction was observed between mulch and fungicidal seed dressing: fungicide seed application significantly increased seedling stand only in non-mulched plots (13.1 %).
Table 1. Cotton seedling stand (%) as a function of mulch and seed dressing under two cover crops.
DAS: days after sowing.
I: insecticide; F: fungicide; M: mulch.
SAS ANOVA GENMOD, *p < 0.05 and **p < 0.01.
Cotton seedling vigour
In both experiments the proportion of damaged seedlings was significantly higher in mulched plots than in non-mulched plots (Table 2). However, the number of dead millipedes collected was significantly higher in mulched plots than in plots without mulch only in 2001 (B. ruziziensis). In both experiments, insecticidal or fungicidal seed dressing had no effect on the proportion of damaged seedlings. However, in both experiments, significantly more dead millipedes were collected on plots where seeds had been coated with insecticide (Table 2). In 2002 (B. ruziziensis), a significant interaction between mulch and insecticidal seed dressing showed that insecticide had a greater effect in mulched plots. In both experiments, mulch had a positive effect on seedling height in (Table 2). Insecticide had a positive effect on seedling height only in 2001 (C. mucunoides) while fungicide had no effect in either experiment.
Table 2. Cotton seedling vigour according to mulch presence and seed dressing under two cover crops.
DAS: days after sowing.
I: insecticide; F: fungicide; M: mulch.
SAS ANOVA GENMOD, *p < 0.05 and **p < 0.01.
Aphid infestation
The temporal observation of aphids on cotton seedlings did not show any significant effect of mulch or seed dressing on the dynamics of field colonization, compared to control plots. However, in the 2001 experiment, qualitatively more aphids were observed on cotton leaves in plots covered with C. mucunoides mulch than in bare soil plots at 38 DAS (Table 3).
Table 3. Dynamics of aphid infestation according to mulch presence and seed dressing.
DAS: days after sowing.
I: insecticide; F: fungicide; M: mulch.
SAS ANOVA GENMOD, *p < 0.05 and **p < 0.01.
DISCUSSION
In both experiments, cotton seedling stands were negatively affected by the presence of mulch cover crop. In addition, the proportion of damaged seedlings was higher in mulched than in non-mulched plots, supporting the hypothesis that mulch favoured the presence of soil arthropod pests. In both experiments, insecticidal seed dressing increased the seedling stand and the number of dead millipedes collected, particularly in 2002 in B. ruziziensis-mulched plots, As reported by House and Del Rosario Alugaray (Reference House and Del Rosario Alugaray1989), mulch provides a suitable habitat for soil arthropods, including detritivores and, less often, herbivores. A disequilibrium between phytophagous and predator communities may follow mulched cropping systems and the adoption of certain mulched cropping systems by farmers may lead to pest outbreaks, particularly of soil-associated pest populations (Brown et al., Reference Brown, Pasini, Benito, Aquino and Correia2001). In Cameroon, the major risks associated with sowing in mulch-based systems are soil-borne pests, particularly millipedes, which in traditional crop management systems have been controlled by insecticidal protection with carbosulfan (SODECOTON, 2004). Ratnadass et al. (Reference Ratnadass, Michellon, Randriamanantsoa and Séguy2006) point out that following implementation of such crop management practices, it may take time to reach a new biological equilibrium among functional arthropod communities. In Australia, it has been reported that reduced or zero-tillage did not increase the incidence of insect pests on emerging seedlings (Wilson-Rummenie et al., Reference Wilson-Rummenie, Radford, Robertson, Simpson and Bell1999).
In both experiments, fungicide had little or no effect on seedling stand and vigour, indicating that the fungicide used did not suppress seedling diseases caused by soil fungi or that there was no seedling disease. It was, however, observed in 2002 that the fungicide seed dressing had a particular positive effect on seedling stand in non-mulched plots. Accordingly, it may be hypothesized that fungi may have been naturally inhibited under B. ruziziensis mulch. Consistent with these results, Doupnik and Boosalis (Reference Doupnik and Boosalis1980) reported that mulch protected sorghum seedlings from Fusarium moniliforme by reducing water stress and the effects of high temperatures that usually favour its activity. Another explanation assumes that mulch-based systems enhance natural microbial antagonist regulation (Altieri, Reference Altieri1999).
The hypothesis suggesting that natural mulch would have been a deterrent to aphids was not clearly supported by this research. Several authors have suggested that mulch background spectral reflectance impedes migrant aphids from landing on cotton plants (Costello, Reference Costello1995; Doring and Chittka, Reference Doring and Chittka2007; Summers et al., Reference Summers, Stapleton and Mitchell2004) or limits aphid population growth by increasing light intensity on the underside of plant leaves (Rummel et al., Reference Rummel, Arnold, Slosser, Neece and Pinchak1995). In our two-year experiments, the dynamics of aphid colonization were not influenced by the presence of mulch. In 2001, taller seedlings were documented in mulched plots than in non-mulched plots, probably due to greater water and nutrient availability in the mulched treatment. The more robust seedlings may have enhanced the attractiveness or appetence to aphids. The risk of aphid infestation could be reduced by early sowing and adequate systemic seed protection (Deguine et al., Reference Deguine, Goze and Leclant1994).
This study identified risks associated with direct mulch seeding. The results of this work should aid farmers in transitioning to new farming practices. The preliminary results of this research provide a springboard to further evaluate the effects of different parameters such as mulch composition, biomass, previous crop and age of the cropping system on seedling pest constraints. In addition, attention should be given to soil-inhabiting natural enemies that may assist in the regulation of aerial pests (Peachey et al., Reference Peachey, Moldenke, William, Berry, Ingham and Groth2002; Rypstra and Marshall, Reference Rypstra and Marshall2005). In cotton systems, Tillman et al. (Reference Tillman, Schomberg, Phatak, Mullinix, Lachnicht, Timper and Olson2004) reported that ants and Geocoris punctipes contribute to the management of Heliothis virescens and Helicoverpa zea in conservation tillage with crimson clover and rye cotton cover crops. Similarly, in vegetable crops it was demonstrated that fire ants were more abundant in cover-crop mulch plots and were important predators of weed seed and insect pests (Pullaro et al. Reference Pullaro, Marino, Jackson, Harrison and Keinath2004). Accurate sampling and measures of the impact of soil pests and beneficial organisms should be assessed to further define effective pest control strategies for sustainable crop management systems.
Acknowledgements
We are grateful to the Projet ESA (Eau-Sol-Arbre) and SODECOTON (Société de Développement du Coton au Cameroun) for providing financial support for this project.
