Welcome to the International Institute of Tropical Agriculture Research Repository
What would you like to view today?
Comparative evaluation of the contributions of soil physicochemical properties to variations in the yields of four major staple food crops in eastern Nigeria
MetadataShow full item record
The contributions of soil variables to the variations in the yields of cassava (Manihot esculenta), yam (Dioscorea rotundata), maize (Zea mays) and pigeon pea (Cajanus cajan) were evaluated over 2 years in this study. The data were from three replicates of two randomized complete block design experiments sited in a newly cleared forest and on previously cultivated land both in Nsukka, eastern Nigeria. The 28 soil physicochemical properties and six crop yield parameters measured were partitioned between location and year before applying a stepwise regression procedure to analyze them. The study showed that soil variables accounted for >70% of the variation in cassava root yields and harvest index. Both soil physical and chemical properties contributed but the former (particularly macroporosity, microporosity, total porosity and bulk density) contributed most. Selected soil variables also accounted for >70% of the variation in yam tuber yield and shape index of tubers especially in 1998. In both crop years chemical properties appeared to dominate over the physical ones. Soil variables accounted for between 51 and 99% of the variation in maize grain and stover yields. The only exception was the figure of 44% obtained at the forest location in 1998. Soil pH, total exchangeable acidity and microporosity were particularly important contributors to the variations in both maize yield parameters. The contributions of soil variables to pigeon pea yield parameters were low (<50%) except in 1999 at the forest location where seven soil variables accounted for over 85% variation in seed yield. It was obvious generally from the study that soil variables were important determinants of yield variations in the four crops. It was also shown that physical properties should always be included in this kind of analysis. Also the number of soil variables which were of significance generally increased when the level of soil properties was low, as was the case with the cultivated site versus forest site, and 1999 versus 1998 analysis. Thus increasing the number of soil variables used and partitioning them into more homogeneous units helped to improve the results obtained using the procedure.
Multi standard citation
Permanent link to this itemhttps://hdl.handle.net/20.500.12478/3790
Digital Object Identifier (DOI)