The seasonal patterns in plant quality in various ecological zones in Natal.

Abstract

The objectives of this study required that the following investigations be undertaken: 1) to determine if plant quality can be altered by modifying growing conditions; 2) to quantify the seasonal trends in plant quality from different sites; 3) to relate differences between sites to environmental variables; 4) to develop an objective classification of SWEETNESS; and 5) to plan future research. The majority of the commercial and subsistence livestock in southern Africa rely almost entirely on veld (rangeland) for their supply of nutrients. These rangelands are traditionally and conventionally managed according to their classification as 'SWEETVELD' and 'SOURVELD'. An intermediate group 'MIXED-VELD' is also recognised. The subjective classification is based on the quality (nutritive value) of the rangeland when it is mature (winter). Both extremes of 'sweet' and 'sour' rangeland contain many of the same species and this thesis considers the relationship between the soil ( chemical and physical) and the physical environment and plant quality of a single indicator species Themeda triandra (red grass). A glass house experiment was used to determine the effect of manipulating the soil environment on the quality of T. triandra. There were no significant differences ( P> 0. 05) between any of the six treatments (combinations of eutrophic, ameliorated and dystrophic soil together with 'sweet' and 'sour' T. triandra plan ts) . When compared as a group significantly (PsO. 01) higher in disappearance ( CDMD) than the SOUR group. the SWEET group were cellulase dry matter However the difference was only 1 . 5% CDMD uni ts and is believed to be biologically

unimportant. It was concluded that the quality of T. triandra can not be altered by manipulating its growing conditions. In a field investigation the seasonal pattern and the relationship between environmental variables and plant CDMD, N, P, K, Mg, Ca, Sand Zn are described for Natal, South Africa. Most models have significant (Ps0.01) R2 values but very few show any strong relationship between soil chemical status and forage j~-l quality. ALTITUDE appears most in the models as a related

variable. The models would have little predictive value outside Natal and do not contribute or describe adequately the factors determining seasonal patterns in plant quality at different locations. A multivariate approach is used to provide an objective index of 'SWEETNESS' (based on seasonal variations in plant quality at 31 sites over 23 months), and this is related to environmental variables. This analysis also showed that the soil environment was only weakly related to plant quality. The results are confusing given the wide variations in both plant quality and soil chemistry in the data presented.