- A-Z Publications
- Journal of Applied Science in Southern Africa
- OA African Journal Archive
- Volume 7, Issue 2, 2001
Journal of Applied Science in Southern Africa - Volume 7, Issue 2, 2001
Volume 7, Issue 2, 2001
Source: Journal of Applied Science in Southern Africa 7, pp 73 –82 (2001)More Less
Desmodium uncinatum, Stylosanthes guianensis and Macroptilium atropurpureum, grown at the University of Zimbabwe Farm, were harvested serially in the early (December), mid (February) and late (April) growing season in 1992/93 and oven-dried. Field-cured hays were also made from the legumes in April 1993 and 1994. The legumes differed in phenological development which, in turn, influenced crude protein (CP) levels. The CP content of S. guianensis (132 to 167 g/kg dry matter) was the least and first rose and then fell, that of D. uncinatum (195 to 234 g/kg DM) decreased, while that of M. atropurpureum (189 to 229 g/kg DM) increased with the season. Neutral detergent fibre (NDF) was higher in the early (364 to 440 g/kg DM) than in the mid to late season (470 to 559 g/kg DM).
Source: Journal of Applied Science in Southern Africa 7, pp 83 –96 (2001)More Less
DDT and its metabolites were analysed in a Mount Darwin terrestrial environment. River sediments, water, soil, house dust, cow dung, grass, leaves, treebark and fish were taken from accessible sites along the Mount Darwin-Rushinga highway during the period March to November 1999. Results obtained indicate widespread contamination from trace residues of DDT. River sediments and house dust had high levels of DDT varying from 0.14 ng/g to 506.2 ng/g and 2.10 ng/g to 276.5 ng/g respectively. Other matrices had generally low residue levels.
Author J. MafodyaSource: Journal of Applied Science in Southern Africa 7, pp 97 –112 (2001)More Less
This paper presents an alternative method of constructing Multivariate control charts. The method is based on a statistic referred to hereafter as the Studentised Hotellings statistic. From simulated data and industrial data, this method seems to be better than the usual methods in that it clearly shows the out-of-control observations. That is, it seems to have more power than the procedures currently in use. In addition it extends the work of Nola D. Tracey et al. (1992) on individual observations, to the more practical situation where subgroups are used.
Author J.T. GoneseSource: Journal of Applied Science in Southern Africa 7, pp 113 –120 (2001)More Less
A survey of small-holder tobacco farmers production practices, featuring the use of fertilisers and lime in their tobacco lands indicated that 21 out of a systematically drawn sample of 43 respondents from an area covering over 95 percent of the total small-holder tobacco area of the country, use the recommended rate of basal fertilizer. None of the 43 had their soils analysed for pH or nutrients before growing tobacco or had ever applied any lime before growing tobacco. Forty two soil samples collected from tobacco lands belonging to another group of 42 farmers, similarly selected from the same population of farmers, showed that more than 80 percent of the farmers may be growing tobacco at pH levels below 5.0, and that only about 6 percent of them may be growing it at levels near its optimum. The latter group was predominantly from the northern regions of the country. Most of the soils sampled fell in the pH range 4.0 to 4.5. This trend may suggest that tobacco productivity on small holder-farms may increasingly be threatened with acidity over time and that the fertilizer levels being applied on small-holder farms may not be cost-effective.
Author V.R. MundembeSource: Journal of Applied Science in Southern Africa 7, pp 121 –131 (2001)More Less
This paper presents an example of a typical calibration of a partially immersed Liquid-in-Glass (LiG) thermometer, by comparison to a standard precision digital thermometer, to obtain a correction to the LiG thermometer at a nominal temperature of 50C. The standard way of evaluating and expressing the measurement uncertainty in typical corrections and uncertainty sources for the calibration has been used.
Author K. MutsonziwaSource: Journal of Applied Science in Southern Africa 7, pp 133 –143 (2001)More Less
A simple and yet sound statistical procedure for parameter estimation based on probability transforms is developed. Probability transforms are well-known to be very important and consequently widely used in statistics to investigate properties of probability distributions or descriptionbe statistical properties of random variables. However, from the statistics literature, in particular Estimation Theory, probability transforms are not commonly used for parameter estimation and inference in general. The procedure is shown to give unbiased, consistent and asymptotic normal estimates.