African leafy vegetables have been shown and suggested to have potential to contribute to human diets and alleviate malnutrition; however, their levels of utilisation are currently low especially in South Africa. This is because there is limited access to these crops due to low availability in the market. Limited access is attributed, in part, to the lack of commercialisation as a result of limited agronomic information describing optimum management options for these leafy vegetables. Availability of such information would contribute to successful commercialisation of these crops. The primary objective of this study was to establish optimum agronomic management factors for Amaranthus cruentus, Corchorus olitorius, Vigna unguiculata and Brassica juncea for irrigated commercial production in South Africa. Seeds of Amaranthus cruentus, Corchorus olitorius were obtained from the Agricultural Research Council seed bank; Vigna unguiculata were obtained from Hydrotech and Brassica juncea seeds were obtained from Stark Ayres. The project consisted of three field studies whose overall objective was to evaluate growth and yield responses of the selected African leafy vegetables to agronomic factors under irrigated commercial production. These field studies comprised of two single factors; summer trials (planting density and nitrogen on three selected crops) and a combined winter trial (nitrogen, irrigation, plant density and planting date on a winter crop). Chapter three (3) investigated the effect of plant density on growth, physiology and yield responses of Amaranthus cruentus, Corchorus olitorius and Vigna unguiculata to three plant densities under drip irrigated commercial production. The plant density levels of 100 000, 66 666 and 50 000 plants/ha were used in the 2011/12 and 2012/13 summer seasons. Parameters measured included chlorophyll content index (CCI), chlorophyll fluorescence (CF), stomatal conductance (SC), leaf number, leaf area index (LAI) and biomass. Amaranthus cruentus and Corchorus olitorius showed better leaf quality at lower plant density of 50 000 plants ha-1 than at 66 666 plants ha-1 and 100 000 plants ha-1. These results are based on bigger leaves expressed as leaf area index (LAI), better colour expressed as chlorophyll (CCI) and higher biomass per plant observed in these crops at 50 000 plants ha-1 in comparison to 66 666 plants ha-1 and 100 000 plants ha-1. In Vigna unguiculata there were no responses observed in LAI and CCI. In Amaranthus cruentus, Corchorus olitorius and Vigna unguiculata fresh and dry mass yield of leaves were higher at 100 000 plants ha-1 compared to other treatments. In A. cruentus and C. olitorius, higher leaf quality parameters (CCI, plant height, leaf number, biomass per plant and LAI) indicated that these crops can perform better at lower densities of 50 000 than at 66 666 plants ha-1 and 100 000 plants ha-1 Therefore, using 50 000 plants ha- 1 is suitable for commercial production of A. cruentus and C. olitorius. In Vigna unguiculata, a plant density of 100 000 plants ha-1 produced the highest fresh and dry mass per unit area without compromising quality in terms of the leaf size (LAI) and colour (CCI). Therefore 100 000 plants ha-1 is a density recommended for commercial production in V. unguiculata.Chapter four (4) was conducted to investigate growth, physiology and yield responses of A. cruentus, C. olitorius and V unguiculata to nitrogen application under drip irrigated commercial production. Three nitrogen treatments levels were used viz. 0, 44 and 88 kg N ha- 1 in 2011/12 season and four nitrogen treatments levels viz. 0, 50, 100 and 125 kg N ha-1 were used in 2012/13 summer season. The nitrogen levels selected for each season were based on recommendations for Amaranthaceae species, Swiss chard (Beta vulgaris L.var cicla) derived from soil analysis of the trial (field) site. Parameters measured included chlorophyll content index (CCI), chlorophyll fluorescence (CF), stomatal conductance (SC), leaf number, leaf area index (LAI) and biomass. Results showed that application of nitrogen at 44 kg N ha- 1 in 2011/12 summer season and 100 kg N ha-1 in 2012/13 summer season improved LAI, CCI, biomass per plants and yield in A. cruentus. A similar trend was observed in C. olitorius except that 44 kg N ha-1 improved stem fresh yield. Further increase in nitrogen fertiliser above 44 kg N ha-1 during the 2011/12 season and above 100 kg N ha-1 in 2012/13 summer season reduced leaf quality and yield in both crops. In V. unguiculata, nitrogen application showed a slight increase in yield values from 0 to 44 kg N ha-1 followed by decrease at 88 kg N ha-1 in 2011/12 summer season; however, this increase in yield was not significant. During the 2012/13 summer season, yield in terms of fresh weight was significantly (P<.001) reduced by applying nitrogen at various levels. However, leaf dry matter content increased significantly (P<.001) with increase in nitrogen from 0 kg up to 100 kg N ha-1, then remained unchanged at 125 kg N ha-1. Therefore, the current study recommends that C. olitorius and A. cruentus could be commercialised at 44 kg N ha-1 and 100 kg N ha-1 which were lower nitrogen application rates than those recommended for Amaranthaceae species. In V. unguiculata, 50 kg N ha-1 improved leaf number; however, this did not translate to any fresh yield advantage, implying that the optimum rate for nitrogen application might be lower than 50 kg N ha-1. Therefore, nitrogen rates less than the ones used in the current study are recommended for V. unguiculata. Chapter five (5) was conducted in winter and it was necessitated by observations made primarily in the previous studies which focused on the effects of single factors such as plant density, planting date and nitrogen deficits. Therefore, there was a need to address interactions between irrigation, nitrogen, spacing and planting date. The objective of this study was to evaluate growth, physiology and yield responses of Brassica juncea to different agronomic and management factors in the 2012 and 2013 seasons. The treatments were as follows: two planting dates in main plot (1 June and 18 July, 2012); two irrigation frequency in sub main plot (once and three times a week); three nitrogen levels (0, 50, 100 kg N ha-1) and three plant densities (133 333, 80 000, 50 000 plants ha-1) as subplots. Parameters measured included chlorophyll content index (CCI), chlorophyll fluorescence (CF), stomatal conductance (SC), leaf number, leaf area index (LAI) and biomass. Results from this study showed a significant interaction effect on plant height, LAI, CCI and CF. Crops irrigated thrice or once a week with 50 kg N ha-1 combined with 50 000 plants ha-1 produced tall plants and bigger leaves (LAI) in the early planting date (1 June) compared to other combinations. Irrigating three times a week combined with nitrogen application at 100 or 50 kg N ha-1 improved CF for late planting date (18 July) in comparison to other combinations. Irrigating once a week combined with nitrogen application at 100 kg N ha-1 increased CCI. There was no significant interaction effect on yield. Application of nitrogen at 50 and 100 kg N ha-1 significantly (P>0.05) increased yield in early and late planting dates compared to the control (0 kg N ha-1), in 2012 and 2013 winter season. Irrigating three times a week led to a significant (P<0.05) increase in yield in the late planting date (18th July) and early planting date (1st June) in 2013 season. Higher plant density of 133 333 plants ha-1 resulted in significantly (P<0.05) higher yield in terms of fresh mass and leaf number in the late planting date 18 July in 2012 and 2013 seasons. However, leaf quality parameters such as leaf size and colour was compromised at 133 333 plants ha-1 relative to 50 000 plants ha-1. Therefore, farmers are recommended to plant early, apply 50 kg N ha-1, irrigate thrice a week and utilise a spacing of 50 000 plants ha-1. The current study indicates that growth and yield of traditional leaf vegetables can be optimised through improved agronomic practise. / Agriculture and Life Sciences / D. Litt. et. Phil. (Agriculture)

Les conséquences des flux de gènes et de l'introgression entre les cultures transgéniques et leurs apparentés sauvages sont encore au cœur des débats associés à la commercialisation des plantes génétiquement modifiées. J'ai développé mon étude sur les conséquences écologiques et évolutives du flux de gènes entre le colza (Brassica napus) et ses apparentés, la moutarde brune sauvage (B. juncea) et la ravenelle (Raphanus raphanistrum), en réalisant une série d'expériences en serre, au jardin et au champ à Beijing et à Dijon. En premier, j'ai présenté une revue synthétique de la littérature publiée sur les flux de gènes et ses effets sur la fitness chez les Brassicées. En second, j'ai cherché à mettre en évidence le rôle de la taille des semences hybrides entre du colza transgénique Bt et la moutarde. La petite taille des semences a réduit les capacités de croissance et de reproduction, mais l'effet sur la fitness était variable en fonction des fonds génétiques ou spécifiques. Les rétrocroisements sur le colza étaient plus faciles et productifs que pour les autres types de descendants. La plupart de ces plantes avait une morphologie de colza. Liée à la résistance à l'herbicide, cette caractéristique pourrait permettre aux descendants de survivre dans les champs et de disséminer les transgènes aux repousses et aux autres colzas, ce qui serait peut être plus gênant que de voir l'introgression réelle dans le génome du parent sauvage. Troisièmement, j'ai simulé le phénomène d'herbivorie chez la moutarde pour étudier la compétition entre des plantes résistantes et des plantes sensibles indépendamment des problèmes de fitness des hybrides interspécifiques. Les plantes résistantes ont un avantage compétitif évident sous la pression d'herbivorie, et cet avantage est exacerbé sous des conditions difficiles telles que de faibles ressources du milieu et l'intensité de l'herbivorie. L'utilisation d'insectes pour attaquer des populations mixtes composées de rétrocroisements sensibles et Bt-résistants aux insectes a confirmé ce résultat et a montré que le transgène n'avait pas de coût en l'absence d'insectes. La productivité totale des populations a augmenté avec la proportion de plantes résistantes. Quatrièmement, des populations de ravenelles ont été échantillonnées dans quatre régions éloignées entre elles, dont une ayant une longue histoire de coexistence avec le colza et donc ayant plus de chance d'avoir été soumise à l'hybridation interspécifique avec le colza. J'ai interprété la divergence des traits et leur polymorphisme dans le cadre d'une hypothèse d'introgression stabilisée en opposition au simple hasard, bien que les différences avec les autres populations n'étaient pas assez marquées pour faire sortir ces populations du domaine de variation décrit pour les ravenelles. Ces études soulignent plusieurs facteurs qui peuvent accroître le risque des flux de transgènes et l'introgression entre les cultures génétiquement modifiées et leurs apparentés sauvages, et cela doit être pris en compte dans les procédures d'évaluation des risques de l'usage de ces plantes. A savoir : la morphologie cultivée qui rend confuse l'identification des introgressants dans le cadre de la bio-surveillance, les petites semences hybrides avec une dormance et une dispersion supérieures, et l'intensité de l'herbivorie et de la compétition qui exacerbe l'avantage adaptatif des plantes transgéniques résistantes aux insectes. Cependant, l'hypothèse de la formation de " super mauvaises herbes " ne semble pas justifiée.

African leafy vegetables have been shown and suggested to have potential to contribute to human diets and alleviate malnutrition; however, their levels of utilisation are currently low especially in South Africa. This is because there is limited access to these crops due to low availability in the market. Limited access is attributed, in part, to the lack of commercialisation as a result of limited agronomic information describing optimum management options for these leafy vegetables. Availability of such information would contribute to successful commercialisation of these crops. The primary objective of this study was to establish optimum agronomic management factors for Amaranthus cruentus, Corchorus olitorius, Vigna unguiculata and Brassica juncea for irrigated commercial production in South Africa. Seeds of Amaranthus cruentus, Corchorus olitorius were obtained from the Agricultural Research Council seed bank; Vigna unguiculata were obtained from Hydrotech and Brassica juncea seeds were obtained from Stark Ayres. The project consisted of three field studies whose overall objective was to evaluate growth and yield responses of the selected African leafy vegetables to agronomic factors under irrigated commercial production. These field studies comprised of two single factors; summer trials (planting density and nitrogen on three selected crops) and a combined winter trial (nitrogen, irrigation, plant density and planting date on a winter crop). Chapter three (3) investigated the effect of plant density on growth, physiology and yield responses of Amaranthus cruentus, Corchorus olitorius and Vigna unguiculata to three plant densities under drip irrigated commercial production. The plant density levels of 100 000, 66 666 and 50 000 plants/ha were used in the 2011/12 and 2012/13 summer seasons. Parameters measured included chlorophyll content index (CCI), chlorophyll fluorescence (CF), stomatal conductance (SC), leaf number, leaf area index (LAI) and biomass. Amaranthus cruentus and Corchorus olitorius showed better leaf quality at lower plant density of 50 000 plants ha-1 than at 66 666 plants ha-1 and 100 000 plants ha-1. These results are based on bigger leaves expressed as leaf area index (LAI), better colour expressed as chlorophyll (CCI) and higher biomass per plant observed in these crops at 50 000 plants ha-1 in comparison to 66 666 plants ha-1 and 100 000 plants ha-1. In Vigna unguiculata there were no responses observed in LAI and CCI. In Amaranthus cruentus, Corchorus olitorius and Vigna unguiculata fresh and dry mass yield of leaves were higher at 100 000 plants ha-1 compared to other treatments. In A. cruentus and C. olitorius, higher leaf quality parameters (CCI, plant height, leaf number, biomass per plant and LAI) indicated that these crops can perform better at lower densities of 50 000 than at 66 666 plants ha-1 and 100 000 plants ha-1 Therefore, using 50 000 plants ha- 1 is suitable for commercial production of A. cruentus and C. olitorius. In Vigna unguiculata, a plant density of 100 000 plants ha-1 produced the highest fresh and dry mass per unit area without compromising quality in terms of the leaf size (LAI) and colour (CCI). Therefore 100 000 plants ha-1 is a density recommended for commercial production in V. unguiculata.Chapter four (4) was conducted to investigate growth, physiology and yield responses of A. cruentus, C. olitorius and V unguiculata to nitrogen application under drip irrigated commercial production. Three nitrogen treatments levels were used viz. 0, 44 and 88 kg N ha- 1 in 2011/12 season and four nitrogen treatments levels viz. 0, 50, 100 and 125 kg N ha-1 were used in 2012/13 summer season. The nitrogen levels selected for each season were based on recommendations for Amaranthaceae species, Swiss chard (Beta vulgaris L.var cicla) derived from soil analysis of the trial (field) site. Parameters measured included chlorophyll content index (CCI), chlorophyll fluorescence (CF), stomatal conductance (SC), leaf number, leaf area index (LAI) and biomass. Results showed that application of nitrogen at 44 kg N ha- 1 in 2011/12 summer season and 100 kg N ha-1 in 2012/13 summer season improved LAI, CCI, biomass per plants and yield in A. cruentus. A similar trend was observed in C. olitorius except that 44 kg N ha-1 improved stem fresh yield. Further increase in nitrogen fertiliser above 44 kg N ha-1 during the 2011/12 season and above 100 kg N ha-1 in 2012/13 summer season reduced leaf quality and yield in both crops. In V. unguiculata, nitrogen application showed a slight increase in yield values from 0 to 44 kg N ha-1 followed by decrease at 88 kg N ha-1 in 2011/12 summer season; however, this increase in yield was not significant. During the 2012/13 summer season, yield in terms of fresh weight was significantly (P<.001) reduced by applying nitrogen at various levels. However, leaf dry matter content increased significantly (P<.001) with increase in nitrogen from 0 kg up to 100 kg N ha-1, then remained unchanged at 125 kg N ha-1. Therefore, the current study recommends that C. olitorius and A. cruentus could be commercialised at 44 kg N ha-1 and 100 kg N ha-1 which were lower nitrogen application rates than those recommended for Amaranthaceae species. In V. unguiculata, 50 kg N ha-1 improved leaf number; however, this did not translate to any fresh yield advantage, implying that the optimum rate for nitrogen application might be lower than 50 kg N ha-1. Therefore, nitrogen rates less than the ones used in the current study are recommended for V. unguiculata. Chapter five (5) was conducted in winter and it was necessitated by observations made primarily in the previous studies which focused on the effects of single factors such as plant density, planting date and nitrogen deficits. Therefore, there was a need to address interactions between irrigation, nitrogen, spacing and planting date. The objective of this study was to evaluate growth, physiology and yield responses of Brassica juncea to different agronomic and management factors in the 2012 and 2013 seasons. The treatments were as follows: two planting dates in main plot (1 June and 18 July, 2012); two irrigation frequency in sub main plot (once and three times a week); three nitrogen levels (0, 50, 100 kg N ha-1) and three plant densities (133 333, 80 000, 50 000 plants ha-1) as subplots. Parameters measured included chlorophyll content index (CCI), chlorophyll fluorescence (CF), stomatal conductance (SC), leaf number, leaf area index (LAI) and biomass. Results from this study showed a significant interaction effect on plant height, LAI, CCI and CF. Crops irrigated thrice or once a week with 50 kg N ha-1 combined with 50 000 plants ha-1 produced tall plants and bigger leaves (LAI) in the early planting date (1 June) compared to other combinations. Irrigating three times a week combined with nitrogen application at 100 or 50 kg N ha-1 improved CF for late planting date (18 July) in comparison to other combinations. Irrigating once a week combined with nitrogen application at 100 kg N ha-1 increased CCI. There was no significant interaction effect on yield. Application of nitrogen at 50 and 100 kg N ha-1 significantly (P>0.05) increased yield in early and late planting dates compared to the control (0 kg N ha-1), in 2012 and 2013 winter season. Irrigating three times a week led to a significant (P<0.05) increase in yield in the late planting date (18th July) and early planting date (1st June) in 2013 season. Higher plant density of 133 333 plants ha-1 resulted in significantly (P<0.05) higher yield in terms of fresh mass and leaf number in the late planting date 18 July in 2012 and 2013 seasons. However, leaf quality parameters such as leaf size and colour was compromised at 133 333 plants ha-1 relative to 50 000 plants ha-1. Therefore, farmers are recommended to plant early, apply 50 kg N ha-1, irrigate thrice a week and utilise a spacing of 50 000 plants ha-1. The current study indicates that growth and yield of traditional leaf vegetables can be optimised through improved agronomic practise. / Agriculture and Life Sciences / D. Litt. et. Phil. (Agriculture)

Les conséquences des flux de gènes et de l’introgression entre les cultures transgéniques et leurs apparentés sauvages sont encore au cœur des débats associés à la commercialisation des plantes génétiquement modifiées. J’ai développé mon étude sur les conséquences écologiques et évolutives du flux de gènes entre le colza (Brassica napus) et ses apparentés, la moutarde brune sauvage (B. juncea) et la ravenelle (Raphanus raphanistrum), en réalisant une série d’expériences en serre, au jardin et au champ à Beijing et à Dijon. En premier, j’ai présenté une revue synthétique de la littérature publiée sur les flux de gènes et ses effets sur la fitness chez les Brassicées. En second, j’ai cherché à mettre en évidence le rôle de la taille des semences hybrides entre du colza transgénique Bt et la moutarde. La petite taille des semences a réduit les capacités de croissance et de reproduction, mais l’effet sur la fitness était variable en fonction des fonds génétiques ou spécifiques. Les rétrocroisements sur le colza étaient plus faciles et productifs que pour les autres types de descendants. La plupart de ces plantes avait une morphologie de colza. Liée à la résistance à l’herbicide, cette caractéristique pourrait permettre aux descendants de survivre dans les champs et de disséminer les transgènes aux repousses et aux autres colzas, ce qui serait peut être plus gênant que de voir l’introgression réelle dans le génome du parent sauvage. Troisièmement, j’ai simulé le phénomène d’herbivorie chez la moutarde pour étudier la compétition entre des plantes résistantes et des plantes sensibles indépendamment des problèmes de fitness des hybrides interspécifiques. Les plantes résistantes ont un avantage compétitif évident sous la pression d’herbivorie, et cet avantage est exacerbé sous des conditions difficiles telles que de faibles ressources du milieu et l’intensité de l’herbivorie. L’utilisation d’insectes pour attaquer des populations mixtes composées de rétrocroisements sensibles et Bt-résistants aux insectes a confirmé ce résultat et a montré que le transgène n’avait pas de coût en l’absence d’insectes. La productivité totale des populations a augmenté avec la proportion de plantes résistantes. Quatrièmement, des populations de ravenelles ont été échantillonnées dans quatre régions éloignées entre elles, dont une ayant une longue histoire de coexistence avec le colza et donc ayant plus de chance d’avoir été soumise à l’hybridation interspécifique avec le colza. J’ai interprété la divergence des traits et leur polymorphisme dans le cadre d’une hypothèse d’introgression stabilisée en opposition au simple hasard, bien que les différences avec les autres populations n’étaient pas assez marquées pour faire sortir ces populations du domaine de variation décrit pour les ravenelles. Ces études soulignent plusieurs facteurs qui peuvent accroître le risque des flux de transgènes et l’introgression entre les cultures génétiquement modifiées et leurs apparentés sauvages, et cela doit être pris en compte dans les procédures d’évaluation des risques de l’usage de ces plantes. A savoir : la morphologie cultivée qui rend confuse l’identification des introgressants dans le cadre de la bio-surveillance, les petites semences hybrides avec une dormance et une dispersion supérieures, et l’intensité de l’herbivorie et de la compétition qui exacerbe l’avantage adaptatif des plantes transgéniques résistantes aux insectes. Cependant, l’hypothèse de la formation de « super mauvaises herbes » ne semble pas justifiée. / In the framework of commercial release for transgenic crops with novel traits, consequences of gene flow and introgression are still one main concern. I explored the ecological and evolutionary consequences of gene flow between oilseed rape (Brassica napus) and its wild relatives, brown mustard (B. juncea) and wild radish (Raphanus raphanistrum), through several experiments carried out in greenhouse, common garden and field in Beijing and Dijon. First, I revised a comprehensive review of the literature about gene flow and its effect on plant fitness in the Brassiceae. Second, I investigated the effects on gene flow of seed size of hybrids between Bt-transgenic oilseed rape and mustard. Small seed size significantly reduced plant growth and reproduction, but its influence on plant fitness varied among genetic backgrounds. Backcrosses to oilseed rape were easier and more productive than other types of progeny of hybrids. Most of these plants exhibited oilseed rape morphology. Together with herbicide-resistance, this trait could help the progeny to survive in the field and disseminate the transgene to volunteers and feral populations, which could be more troublesome than completing introgression into the genome of the wild parent species. Third, I simulated herbivory on mustard to study the competition between insect-resistant and susceptible plants independently to the fitness of the interspecific hybrid. Resistant plants held a competitive advantage under herbivory pressure, and this advantage was magnified in harsh conditions, such as low resources and high simulated herbivory pressure. The use of insects to attack mixed populations composed of transgenic Bt-resistant and susceptible backcrosses confirmed the same conclusion and provided evidence of no cost due to the transgene in the absence of insect. The overall population production increased with the increasing proportion of insect-resistant plants in the presence of insects. Fourth, wild radish populations were sampled from four geographically distant regions, of which one region had a long history of oilseed rape cultivation, and, therefore, higher chance to have been submitted to interspecific hybridization with the crop. Traits divergence and polymorphism in the putative introgressed populations could be supported as alternate hypothesis to random variation, although the differences were not marked enough to place these populations out of the range of variation described in wild radish. These studies pointed out different factors that could enhance the risk of transgenic flow and introgression from transgenic crops to wild relatives, and they must be taken into account in the risk assessment of the use of GM crops: crop traits to identify the hybrid progeny and perform monitoring, small seed size to account for seed dispersal and dormancy, and intensity of herbivory and competition that magnify the fitness advantage of insect-resistant transgenic plants. However, the impact of introgression to create super-weeds was not supported.