Thesis (Ph.D. Agriculture (Soil Science)) -- University of Limpopo, 2017. / Climate change and variability pose a significant challenge to future global food security due to changes in mean climatic conditions which are rendering many areas marginally suitable for crop production. Erratic rains, drought, heat stress and declining soil fertility are some of the factors limiting crop productivity in many arid and semi-arid areas. Furthermore, water resources that could be used to mitigate drought and erratic rains, are also threatened by both scarcity and overuse. To ensure food security in the future, adaptation strategies at crop and management levels should be explored. Adaptation to future drier and warmer climates calls, among others, for improvement in drought resistance of crops through measures such as screening traits for drought tolerance and water use efficiency. This PhD study was carried out at University of Limpopo experimental farm, Limpopo Province, South Africa, a region with a semi-arid climate and mean annual rainfall of ca. 500 mm. The main aim of the study was to evaluate the use of plant δ13C and δ18O as screening traits for potential yield and water use efficiency of crop genotypes under drought conditions. The study also determined the agronomic performance of cowpea (Vigna unguiculata L. Walp.) and triticale (x. Triticosecale Wittmack) genotypes under varying soil moisture conditions including biological nitrogen fixation for cowpea. Both crops were grown under field conditions and the following four moisture levels (averages per level) were applied: well-watered (ca. 420 mm), moderately well-watered (ca. 350 mm), medium stress (ca. 290 mm), and severe stress (ca. 220 mm). The triticale experiment evaluated the agronomic performance and the spectral response of triticale to water stress under semi-arid conditions. The results showed a significant (P < 0.05) influence of moisture levels on the spectral reflectance, as well as on biomass and grain yield performance of triticale. However, these measured parameters did not significantly (P > 0.05) respond to genotypes probably due to the pre-screening of the genotypes or the lack of distinct genetic diversity in the studied parameters. Under well-watered conditions, triticale produced a grain yield of 3.9 t ha-1 in 2013 and 4.9 t ha-1 in 2014. These yields were however, found to be low when compared to other studies. Even though, no statistical differences were observed among the genotypes, Agbeacon showed a tendency of higher performance compared to the other genotypes. Of the four spectral indices tested, water based indices i.e. the water index (WI) and normalised difference water index (NDWI) were found to be more effective in detecting leaf water status compared to greenness based indices (normalised difference vegetation index and nitrogen reflectance index). This is because NDWI and WI respond to short term changes in water content of leaves. Overall, the performance of triticale showed a good adaptation to semi-arid conditions. The evaluation of δ13C and δ18O as screening traits for potential yield and water use efficiency under drought conditions produced interesting results. The findings showed that, Δ13C was positive and strongly related to grain yield and thus has potential to be used as a surrogate for grain yield in triticale under water stress. We also found a negative relationship between Δ13C and intrinsic water use efficiency (WUEintrinsic), which suggests that breeding for higher WUEintrinsic in triticale may not necessarily yield the desired improved grain yield. Measured grain Δ13C and flag leaf Δ13C suggested minimum contribution of pre-anthesis assimilates to grain filling under water stress, contrary to what is reported in literature. However, for concrete conclusions on the source of assimilates to the grains under drought conditions, further studies are still needed. Combining δ13C and δ18O provided more information on the physiological responses of triticale to varying moisture levels. The δ13C and δ18O were used to test the dual isotope model by Scheidegger et al 2000 and the results showed that, vapour pressure deficit (VPD) of air plays an important role in the operation of the model. The study indicated that the model worked only under high VPD when stomatal conductance limits transpiration rate but failed to work when VPD was relatively low and limiting transpiration rate. In order to address the declining soil fertility in the smallholder farming sector as well as the predicted loss in productivity of the commonly grown dry bean, this PhD study evaluated cowpea genotypes for biomass yield, grain yield and biological nitrogen fixation (BNF) under varying moisture levels. The results showed that soil moisture levels indeed affect biomass production, grain yield, nodule formation and the ability of cowpea to fix atmospheric nitrogen. BNF and nodule formation were the most sensitive to water stress compared to the other parameters. Severe water stress reduced BNF by 57% relative to well-watered conditions while nodule mass was reduced by 80% for the same soil moisture levels. Genotype TV4607 was superior in most of the parameters determined except for grain yield. As a result, TV4607 produced the highest biomass and returned the most nitrogen back to the soil compared to the other genotypes. However, IT00K-1263 emerged as the superior genotype due to its ability to produce the optimum balance of biomass, grain yield and BNF. Stomatal behaviour of cowpea under varying moisture levels was also investigated and the results showed sensitivity of stomatal conductance to soil moisture levels. As expected, stomatal conductance was high under well-watered conditions compared to water stressed conditions. Genotypic variation in stomatal conductance was only observed at early stages of cowpea growth i.e. at 47 and 54 days after planting (DAP). In conclusion, this PhD has shown that δ13C has potential to be used in breeding for drought resistance in triticale and probably other small grain crops. This study also revealed that there is minimum contribution of pre-anthesis assimilates to grain filling under water stress, contrary to what is reported in literature and hence more research is needed.The evaluated genotypes of triticale and cowpea showed tolerance to drought stress under semi-arid conditions. Agbeacon for triticale and IT00K-1263 for cowpea were identified as the most promising genotypes and hence their adoption in the smallholder farming system could be a step towards adapting to future warmer and drier climates