Integration of cover crop technology and imidazolinone resistant maize on weeds and maize yield in two ASAL areas of Kenya

Abstract

Weeds compete for nutrients and soil moisture resulting to low maize (Zea mays L) yields in drylands. Previous efforts to minimize weed related yield losses are hampered by limited knowledge on weeds and management technologies needed to boost maize productivity. This thesis reports four studies conducted to address four objectives. Study one was a survey conducted in 2009, to determine weed species diversity and density in maize growing fields in two counties: Machakos (Kalama where cover crop technology showed potential to suppress weeds and increase maize yields); and, Makueni (Kee a neighbouring region where cover crop technology was not introduced). A quadrat was placed randomly and weed species in the sample area counted and recorded. This was repeated in four other locations in a “W” designated pattern in each maize field. The procedure was replicated in 12 villages. Study two was done to examine sociological factors influencing cover crops adoption in Kalama and Kee using a multistage sampling technique. A semi-structured questionnaire was administered to 80 randomly selected participants to obtain sociological information including gender, age category, education levels, and cover crops adoption. Two binary logistic regression models were used to determine the factors affecting cover crops adoption by respondents. In order to enhance weed control under cover crop study three was done to evaluate effect of Imazapyr coated (IR-maize) on rape weed (Brassica napus L.) emergence and growth in green house for 35 days. Rape was chosen because it is highly sensitive to Imazapyr herbicide. Study four was done to evaluate the effect of dolichos (Lablab purpureus L.) and open pollinated imazapyr coated (IR-maize) on weed species composition, density and maize yield in three years at Kenya Agricultural and Livestock Research Organisation, Kiboko. Initial weeds were controlled using glyphosate at 1.6 kg ai ha-1. Twenty four plots were marked, each measuring 4 x 5 m. Six treatments 1) IR-maize coated, 2) IR-maize coated + brown dolichos, 3) IR-maize coated + black dolichos 4) IR-maize uncoated, 5) IR-maize uncoated + brown dolichos, 6) IR-maize uncoated + black dolichos were laid out in a randomized complete block design replicated four xxiii times. IR-maize was planted at a spacing of 90 x 45 cm with 2 seeds per hole. Weeds were sampled from a 1.0 m-2 quadrat 21 and 42 days after planting (DAP) and recorded for types of species and density .The data on weed species and densityand maize yield was subjected to analysis of variance using Genstat version 12.0. Study one established that 25 weed species infested maize fields; and, species were mainly annuals and few perennials. Oxygonum sinuatum (Meisner.) Dammer and Bidens pilosa (L.) were the most abundant among broadleaf species while Dactyloctenium aegytium (L.) Willd and Eleusine indica (L.) Gaertn. were abundant among grasses. An inventory of weed species was developed. Weed species were location specific; and, management technology (ies) were needed. Study two analysis indicated that 80% of the respondents had adopted cover crop technologies at Kalama compared to 57.5% at Kee. Men were less likely to adopt (P < 0.05). At Kalama age category affected cover crops adoption (P < 0.05); however, age had no significant effect at Kee. Education indicated mixed effects on cover crop adoption suggesting other factors not covered are influencing. Knowledge and skills, demonstration of gains and related cost had effect on cover crop adoption (P < 0.05). Majority of farmers, adopters or non-adopters used seeds from market. Non-adopters in Kee (37.5%) used relief seed suggesting other factors were required to give the threshold to influence adoption. Study three results showed that herbicide coated IR-maize reduced mean weed dry matter by 12 times compared to uncoated IR-maize (P < 0.05). Herbicide coated IR-maize with dolichos or beans or cowpeas reduced weed dry matter by 6 times compared to uncoated IR-maize with cover (P < 0.05). Weed dry matter in herbicide coated IR-maize watered once a week was lower than watered twice (P < 0.05). Watering twice a week decreased emerged weed count in herbicide coated IR-maize with dolichos, beans or cowpeas (P < 0.05) respectively. Study four established that eighteen (18) weed species infested Kiboko field. Interaction of dolichos and herbicide coated IR-maize had no effect on the number of weed species. Dolichos reduced the density of Portulaca quadrifida L. and Paraknoxia parviflora L., and increased Eleusine indica L (P < 0.05). Dolichos xxiv decreased weed species composition by 14% (21 DAP) and 33% (42 DAP) compared to no cover. Maize yields were higher in plots with dolichos than without (P < 0.05). From the results, knowledge on weeds could be used to determine appropriate adaptive weed management, pin point problem weeds that need concerted research efforts, and form a base for measuring weed changes in future.