5.

Projects completed successfully or in which significant progress was made (Annexures I and II)

5.1

Development of soybean cultivars with better adaptation and seed quality

Mr AJ de Lange

The soybean breeding programme at ARC-GCI was active between 1984 and 2007. The main objectives were breeding for increased yield and protein content. Attention was also given to adaptation to a wider range of biotic and abiotic stresses. Crosses were made, segregating popula­tions were developed and improved lines were selected. Stable lines were evaluated in yield trials. A number of different localities were used, i.e. Potchefstroom, Brits, Bethlehem, Bergville, Vaalharts and Cedara. Four cultivars were released from the programme, i.e. Egret, Stork, Heron and Ibis 2000. Lines with resistance to root-knot nematodes were identified and are in an advanced stage of evaluation. Pure breeding lines and segre­gating populations with rust and cold tolerance have been identified. All lines and populations of the 2006/2007 season were harvested and will be kept in cold storage. During 2007 the soybean breeder resigned from the ARC. The project has thus been suspended and will continue when a new plant breeder is appointed.

5.2

National soybean cultivar trials

Mr JL Erasmus

In the National Cultivar trials, a total of 30 cultivars were evaluated at selected localities across soybean production areas of South Africa. An additional 30 cultivars were evaluated in phase-1 trials at nine locations. Trials were planted at research stations where the necessary infra­structure exists in order to carry out a comprehensive agrometeorological study. Trial sites were also selected in areas where a need for infor­mation exists and where the maximum amount of data could be collected.

A comprehensive data set was collected on phenological data such as days to emergence, days to flowering, physiological and crop maturity, maximum plant height and minimum pod clearance, lodging, shattering and green stem, relative disease or pest susceptibility, seed character­istics and yield, physical and chemical quality, and soil and weather data. All data from the trials were analysed to calculate yield reliability and regression lines for cool, moderate and warm production areas. At the cultivar evaluation committee meeting during September 2007, five cul­tivars were approved for registration on the variety list namely, PAN 1664 R, PAN 1564, PAN 1669, PAN 1666 R and AS 4801. The 2006/2007 cultivar evaluation report was approved for publication.

Average seed yield across localities was 1 717 kg ha^-1 compared to 2 538 kg ha^-1 for 2006 and 2 582 kg ha^-1 for 2005 seasons. Highest yields were obtained at Groblersdal (3 589 kg ha^-1), Atlanta (3 007 kg ha^-1) and Glen (2 951 kg ha^-1). The cultivar LS 666 (1 888 kg ha^-1) yielded the highest and Sonop (1 538 kg ha^-1) the lowest. The cultivar with the highest protein percentage was Stork with 46.33% and lowest was PAN 538 R with 41.65%.

5.3

The increase in soybean production on the Highveld

Mr WF van Wyk

Three trials were collectively done at the UP-experimental farm in Hatfield Pretoria, namely:

• Row width, plant density and cultivar
• Fertilisation
• Weed control

The row width, plant density and cultivar trial showed some trends although there were no significant differences in some cases, possibly because of the very dry season. Some data (for example the percentage 4-seed pods) proved significant between cultivars but this could be attributed to the genetic trait of the specific cultivar. Interestingly, the bushy-type cultivar had a higher yield than the upright-type cultivar at the 100, 75 and 50 cm row widths but a lower yield at the 25 cm row width. The higher plant density (400 000 plants/ha) out performed the lower plant density (300 000 plants/ha) at all row widths. Researchers in Brazil and the USA found that in years with lower than average rainfall there were no significant differences between the various row widths.

The fertilisation trial was in its first year and due to high general fertility of the soil, no significant differences were expected. Differences, if any, will become evident in the following years. However, there was evidence that with high P and K applications the protein content increased.

The weed control trial was planted without any knowledge of the weed spectrum. Some herbicides were used on their own or in combination with others in order to determine their effect on the different weeds as well as on soybean yield. Some of the more "hardy" weeds like nutgrass, morn­ing glory and wandering jew were not prominent and therefore some herbicides could not be evaluated to their true potential. Where herbicides were used to target either grass or broadleaf weeds, the control was good on the weeds present. Due to the fact that it was a dry season, the weeds also did not flourish and therefore good yields were still possible. In treatments where glyphosate was applied 2 and 6 weeks after emer­gence, the application treatment at 2 weeks outyielded that at 6 weeks. Unfortunately there was not a control treatment where weed control was done by hand after emergence, so the difference in yield at the different application dates could be either from early weed competition (6 week treatment) or phytotoxicity of the glyphosate on more mature plants. A hand control treatment (where all weeds will be hand pulled from emer­gence till harvest) will be included in the 2007/2008 trial. The herbicide Frontier (S-dimethenamid) caused drastic phytotoxic symptoms by slow­ing down emergence by 7 days and reducing plant density by between 50-60%. A complete progress report was presented in October 2007.

5.4

Effect of SO₂ and the interaction with drought, on growth, physiology and biochemistry of soybean and cabbage, studied in an OTC system

Dr PDR van Heerden, Prof GHJ Krüger, Ms E Heyneke, Mr PR Smit and Mr R Strauss

Research during 2007 entailed the following: (i) erection and optimisation of a battery of 12 open top chambers (OTC) growth chambers, and (ii) exposure of Soybean (G. max) to different SO₂ concentrations over a full growth period while periodically measuring growth, chlorophyll fluores­cence, photosynthetic gas exchange, chlorophyll content as well as leaf and water potential. The working hypothesis was that moderate water stress would reduce the harmful effect of SO₂, due to a SO₂-induced decrease in stomatal conductance.

The excellent controllability of the OTC system was documented. A standard procedure for the control and management of the system was compiled which will be useful for continued research on air pollution impacts using the OTC facility.

An ultrastructural investigation of leaf material of soybean, exposed to SO₂ for 35 days, 7 hours per day, indicated that damage occurred mainly in the chloroplasts. An important finding was that physiological and biochemical constraints occurred before the appearance of visible symptoms. Though a moderate stimulation of some parameters occurred at the 50ppb treatment, SO₂ exposure decreased the biomass production of cab­bage and soybean significantly. In drought stressed soybean, exposed to 300ppb SO₂, the decrease in biomass production was 57%. Pod size and seed number were also decreased significantly. SO₂ exposure caused lower water use efficiency (WUE) in the test plants. Analysis of the gas exchange data showed that the SO₂-induced inhibition of photosynthesis was mainly due to biochemical constraints rather than to stomatal limitation. The contention that moderate water stress would reduce the inhibitory effect of SO₂ was proven wrong.

In addition to the SO₂ induced decrease in pod size and seed number, an indirect relationship occurred between SO₂ concentration of the treat­ment and ureid content of the nodules. Already at the 50ppb treatment, a decrease of 25% occurred, pointing at the strong inhibitory effect of SO₂ on symbiotic nitrogen fixation.

Through this investigation important information was also obtained on the effect of SO₂ on primary photochemistry (PSII function), photosyn­thetic gas exchange, the light response and carboxylation efficiency (Rubisco) of soybeans. A synoptic scheme showing the key physiological and biochemical aspects of photosynthesis and the sites where SO₂ exhibit its inhibitory effect in soybean, was compiled. The implication of our find­ings for crop production in highly polluted areas is being discussed.

5.5

Chemical control options for plant-parasitic nematodes associated with soybean in South Africa

Dr H Fourie

Significant yield losses in local soybean are imminent since no nematicides are registered for the crop. In addition, soybean production is currently extending into production areas with sandy soils where root-knot nematode infestations are known to cause maize yield losses. Consequently various nematicides were evaluated in support of endeavours to register products for the control of plant-parasitic nematodes on soybean. Three field trials were conducted during the 2005/2006 growing season in the Nelspruit, Vaalharts and Hammanskraal areas, on sites where natural infestations of root-knot nematodes occur.

Significant differences existed among the 14 nematicide treatments with regard to the number of root-knot nematode eggs and second-stage ju­veniles (J2) per 50 g roots at all three localities. Three nematicide treatments maintained significantly lower root-knot nematode numbers than un­treated controls at all three localities during the 2005/2006 growing season, thus showing potential to reduce root-knot nematode numbers effec­tively. Other treatments also maintained significantly lower root-knot nematode numbers but only at one or two localities during the 2005/2006 growing season. Those treatments could, however, also be considered for registration on soybean. It is important though to obtain data from residue analyses before final decisions are made with regard to registration. In terms of yield no statistical differences were obtained among the 14 nematicide treatments at any of the localities. Environmental conditions such as excessive rains and soil characteristics may have contributed to this situation and should not be ignored when these nematicides are considered for registration. Data obtained as a result of these trials indi­cate that nematicide treatment on soybean may offer substantial relief to producers where high root-knot nematode populations occur.

5.6

Phenotypic markers for nodulation capacity of soybean cultivars

Me Urte Schlüter, Dr Riekert van Heerden and Dr Karl Kunert

Nitrogen fixation in nodules offers an important advantage in soybean when compared with most grain crops due to the fact that soybean nodules fix the nitrogen required for plant growth and for the production of high-protein seeds.

The life of a nodule is, however, remarkably short (11-13 weeks) and nitrogen fixation declines rapidly as the nodule ages, and has normally almost ceased by the time pod-filling starts. Soil drying and high temperature stress limit nitrogen fixation by cutting short the nodule lifespan impacting seed production, crop quality and yield. The objective of the project is therefore to expand our current knowledge on efficient and easily measur­able physiological markers which might be applicable in a soybean breeding program to allow selecting soybean cultivars for growth in the drier areas of South Africa. A particular focus is on markers for extended nodule lifespan under water deficit conditions. Research is carried out at the University of Pretoria in close collaboration with North West University and also University of Limpopo comparing high-yielding soybean cultivars with potential for South Africa with a reference (benchmark) cultivar (Jackson) known to have water deficit tolerance. Current research activities focus on optimization of water deficit experiments particularly on the control of desiccation rates. Rapid desiccation rates, especially if inade­quate pot sizes are used in experiments under controlled conditions, frequently lead to unrealistic water deficit symptoms in plants making extra­polation to field conditions very difficult.

To circumvent this serious problem, a pilot experiment is currently carried out where plants are grown in large pots (15 l) containing a standardized potting medium with five special glass-fibre wigs located at different vertical positions in the potting medium. The wigs constantly pull water from a reservoir up into the potting medium through capillary force, keeping the potting medium at field capacity. We expect that this method allows for precise control and standardization of plant growth/soil water status. Water deficit experiments will then be carried out at the onset of flowering, which is an especially water deficit sensitive stage of plant development. Water deficit will be induced in half of the plants by reducing the number of glass-fibre wigs supplying water to the pots, while the remaining plants are kept at optimal soil moisture status. Various shoot and root nodule parameters will further be measured to quantify the water deficit-response of the various soybean genotypes throughout the water deficit treat­ment. Parameters measured include shoot growth, plant water status, leaf chlorophyll content, carbon assimilation and root nodule parameters. Initial experiments also include the basic characterisation of the root nodules (size, numbers, colour, shape). For validation of the visual markers this data is supplemented by biochemical analysis of key features of nodule metabolism, such as sucrose synthase activity, which is a very sensi­tive indicator of nodule metabolic potential, nitrogenase activity, leghemoglobin content, ureide and carbohydrate content.

5.7

Establishment of an early warning system for soybean rust

Dr M Craven and Mr TM Ramusi

Trial sites representative of major soybean production areas in the Free State, KwaZulu-Natal (KZN), Mpumalanga, Gauteng and North West provinces were selected to form part of the early detection programme for soybean rust for the 2006/2007 season. Kestell (Eastern Free State), Winterton, Greytown, Normandien, Vryheid (KZN), Piet Retief, Morgenzon, Kinross (Mpumalanga), Heidelberg (Gauteng) and Potchefstroom (North West) were consequently identified as suitable trial sites.

As in the past an attempt was again made to plant the trap crop trials approximately three to four weeks before the regular soybean planting date for the various areas. Two cultivars with different length of maturity were selected, viz. PAN 421 RR (short grower) and PAN 522 RR (long grower) to plant at all the localities. In total 2 000 m2 were planted at each locality. Planting and general maintenance of the trials were the responsibility of the respective producers. We visited the trials weekly since the third week of January 2007. Strict sanitary protocols were followed in order to limit the possibility of contaminating trials as we proceed.

Precautionary measures included spraying rain suits and gumboots worn during the screening process with a 70% Jik solution after each screen­ing. Once rust was detected, the relevant producers were advised to spray the trial with appropriate fungicides or to destroy it and remove all plant material afterwards in order to ensure that the trap crop trial does not serve as an inoculum source for infection to commercial crops. The Heidel­berg trap crop trial was excluded from the screening route as the producer concerned terminated the trial as a result of excessive hail and animal damage. Screening of the various localities commenced on 18 January 2007 and rust was first detected at the Greytown as well as Piet Retief sites on 26/01/07. Since then rust was detected at Vryheid (01/02/07), Morgenzon (02/02/07) and Normandien (22/02/07). Kestell, Winterton, Kinross and Potchefstroom remained free of rust. The lack of rust reports in the majority of the trials included in the study could be attributed to a very dry and hot climate during 2006/2007.

5.8

Determining the epidemiological value of resistance to rust caused by Phakopsora Pachyrhizi in soybean lines

Prof NW McLaren and Me C Botha

The aim of the current study is to quantify resistance to rust in soybean germplasm and concomitant losses associated with the disease. The study is a continuation of the identification of host characters associated with reduced rust severity and yield loss as determined in cultivar trials (McLAREN, N.W. (2008) Reaction of South African soybean cultivars to rust caused by Phakopsora pachyrhizi. South African Journal of Plant and Soil 25: 49-54). A principle component that affected the epidemics and the final disease level was the physiological stage at first detection of dis­ease which differed significantly between cultivars. Area Under the Disease Progress Curve was significantly (P>0.05), negatively related to growth stage at first detection. Yield loss was related to premature defoliation which was 16 to 21 days earlier in the longer maturity group. Cultivars also differed in the rate of disease development.

The study was extended to quantify yield loss in 32 diverse epidemics including a short, medium and a long season variety. Treatments within sequential planting included applications of fluzalizole / carbendazin, subsequent to flowering compared with untreated plots and 90 and 45 cm row spacing's. Rust onset was closely related to host development stage in the three varieties used and onset was significantly related to AUDPC (R²=0.59, R²=0.61 and R²=0.64 in the short medium and long season varieties respectively). Similarly, rate of disease development (sensu vdPlank, 1963) was significantly correlated with AUDPC (R²=0.69, R²=0.68 and R²=0.72 in the short medium and long season varieties respec­tively). Integration of these two variables into multiple regression models improved the relationships significantly (R²>0.80) and indicated that selection for these host characteristics could contribute significantly to reducing rust severity. Yield analyses were confounded by differing spray efficiencies and considerable disease levels were recorded in fungicide-sprayed plots. Spray efficiencies ranged from 41-93% and models were modified to determine actual yield losses. Results indicated that significant yield losses can occur despite the use of fungicides and spray efficiency remains a significant variable in the control of rust.

Line evaluation trials consisted of 478 entries and were conducted in field experiments at Cedara. Plots were monitored two-weekly for rust development. Rust infection was dependent on natural infection. Data is still being analysed. Epidemics are being modelled to determine relative lifetime at disease onset, rate of disease development and AUDPC. Twenty-two lines have been selected for inter-crossing to determine whether populations with improved rust resistance can be selected. Trials are currently being conducted in the greenhouse to determine the reaction of lines to rust based on latent period, lesion type, size and infection frequency. Associated with the study are rust collection, identification of races, determining isolate x host genotype interaction, mechanisms of resistance (phenols, anti-fungal proteins) and environmental effects.

5.9

The effect of silicon on the control of soybean rust (phakopsora pachyrhizi) on soybeans

Ms Dael Visser and Dr Pat Caldwell

Various concentrations of Si, in the form of potassium silicate, were applied at concentrations of 100, 250, 500, 1000 and 2000 mgl-1 at 30 day intervals from the V3 growth stage.

An untreated control and a treatment with Punch C (800 l ha^-1) were also included. Plots (3 x 3 m with 4 rows per plot) were arranged in a ran­domized complete block design with 5 replicates per treatment. Plants in the middle two rows were rated for percentage leaf area infected, area under disease progress curve (AUDPC) was calculated and yield determined.

Plants treated with Punch C had the lowest AUDPC (17.2), while the untreated control had the highest AUDPC (92.2). Plants treated with 2000 ppm Si showed a significantly lower AUDPC (64.8) than all other Si and the untreated control. AUDPCs from the remaining Si treatments were not sig­ni­ficantly different from one another, or the untreated control. Yield of plants treated with Punch C and 2000 ppm Si were not significantly different (834 and 697 g, respectively) from each other but were significantly different from the remaining Si treatments. These Si treatments were not significantly different to one another, or the control.

Future aims

When we re-run this trial in the 2008/2009 season, we would also like to include some other field trials investigating the effect of:

1. Increased number of Si applications.
2. Determining the residual effect of Si (i.e. run a trial on the original plot from the 2007/2009 season).
3. Determine whether Si can be applied together with fungicides, herbicides etc that are applied during the growing season to minimize application costs.

5.10

Generating management-orientated maps of long-term soybean rust susceptible areas in South Africa

Me Lauren van Niekerk and Dr Pat Caldwell

Soybean rust (SBR) poses a serious threat to soybean production world-wide and in South Africa. Research has been done into the epidemio­logical characteristics of SBR, resulting in an understanding of the effects of temperature and moisture conditions on Phakopsora pachyrhizi uredospore germination. However, limited knowledge exists as to whether certain locations in South Africa are susceptible to SBR infection. This means that farmers do not know whether SBR is a risk in their area and hence may not manage their risk appropriately.

The objectives of this research are to:

1. Create management-orientated spatial maps showing areas at risk to SBR infection and,
2. Develop a framework for near real-time rust susceptibility forecasting.

A literature review has been undertaken to gain knowledge of the environmental conditions required form SBR germination and to investigate relevant epidemiological algorithms. So far, one epidemiological algorithm has been identified that will be modelled. Modelling of SBR is done to try and predict the likelihood and severity of infection, primarily in relation to rainfall and temperature patterns. An overview of these models pro­vided knowledge into the methods, benefits and short-comings of modelling SBR. This overview also provided insight into some of the assump­tions made when modelling SBR, and whether these assumptions can be trusted or not. Some of these assumptions will be applied to this research.

The databases needed to test and model the SBR algorithms have been collected. These databases include:

1. automatic weather station database for Cedara, with data from 1998 to the present day, and
2. SBR trial data which was conducted during the 2001/2002 and the 2003/2004 soybean growing seasons.

In addition, another database of high resolution climate database containing temperature and rainfall data for a period of 50 years for South Africa (including Lesotho and Swaziland) will be used. This database was developed at the School of Bioresources Engineering and Environmental Hydrology.

An understanding of the techniques and benefits of Principle Component Analysis (PCA) was gained after attending a PCA course during April. These skills may prove useful in confirming the relationships between rust occurrences and weather variables. A course on Agrometeorological principles was also undertaken to gain a better understanding of the inter relationships between weather parameters.

Using identified environmental conditions as parameters, a new epidemiological algorithm will be developed, tested and modelled. All algorithms will be modelled using the weather database from Cedara and compared to a soybean rust database for the same area at the same time period. The most accurate algorithm will be selected and used to model SBR to the whole of South Africa, using the 50 year climate database. The results will be represented spatially using high-resolution maps. Depending on results, an additional map to delineate areas with frequent mist conditions may also need to be developed and incorporated into the final risk map.

Following this, a framework will be developed for near real-time rust susceptibility forecasting. This framework could, in the future, be integrated into an early warning system, informing farmers of the likelihood of SBR infection, based on current and forecasted weather conditions. This framework will incorporate the above-mentioned model, near-real time weather data collected automatically from weather stations around the country and weather forecasts. Based on this framework, a warning system could be developed to inform farmers, either through a system-generated email or sms, that current conditions are conducive for SBR infection. Farmers can then make an informed decision whether or not to spray their crops for SBR prevention.

So far no major setbacks have occurred. Currently, the second algorithm is being developed and testing of both algorithms will commence shortly thereafter.

5.11

Evaluation of fungicides for the control of sclerotinia stalk rot of soybean

Dr M Craven and Mr WG Khali

Eighteen soybean varieties were evaluated for resistance to Sclerotinia stem rot during the 2003/2004 to 2006/2007 seasons. Sequentially plant­ed field trials were conducted, consisting of three plantings at Bethlehem during the 2003/2004 and 2004/2005 seasons and four plantings at Bethlehem and Greytown during the 2005/2006 and 2006/2007 seasons. Planting was spread from mid-November to late-December to ensure a range of weather conditions subsequent to flowering and inoculation. At flowering, plots were inoculated with Sclerotinia sclerotiorum. This was repeated after two weeks during 2003/2004 and 2004/2005. During 2005/2006 and 2006/2007 the second inoculation was conducted using ground PDA-grown mycelium in a water suspension.

All cultivars were to a greater or lesser degree susceptible to Sclerotinia stem rot. Ranking of cultivars according to disease incidence did not correlate over planting dates, seasons or locality and differential responses by cultivars to the pathogen under different environmental conditions were evident. The practice of using mean Sclerotinia stem-rot incidence as a parameter for the determination of resistance proved questionable. Non-linear regression analysis was used to determine the relationship between Sclerotinia stem rot potential associated with inoculation dates and observed disease incidence within cultivars. Disease potential was quantified as the mean disease incidence associated with a planting date at a specific locality over all entries in that trial. Cultivar response could be classified into three categories, viz. those linearly related to disease potential, those highly susceptible at even low disease potentials and those with variable degrees of resistance despite increasing disease potentials.

5.12

Study of inoculation and disease evaluation techniques for sclerotinia stem rot of soybeans

Prof NW McLaren and Me C Botha

• The objectives of this study were to compare various inoculation techniques for S. sclerotiorum on soybeans in the greenhouse; determine the optimum conditions for infection in the greenhouse and in the field; evaluate commercial cultivars for resistance to Sclerotinia stem rot in the field and greenhouse; to determine an economically feasible and effective control method (chemical and biological) for Sclerotinia stem rot and to determine the genetic relationship and isolate variation of various S. sclerotiorum isolates collected from local production fields.
• Greenhouse evaluations of six inoculation techniques on four soybean cultivars were conducted. Leaf damage and wilting incidence were eval­uated using a 0–5 rating scale. Infection levels induced by various inoculation techniques differed and a spray mycelium method proved to be the most consistent and effective. An inoculation technique x cultivar interaction was recorded, especially when wounding of plants was included.
• Temperature and leaf wetness duration studies were conducted in the greenhouse. Optimum temperatures for disease development were determined at 20.90-22.75ºC. Results indicated that high humidity or free moisture is critical for disease development and disease incidence and severity increased with increasing RH. Field evaluations indicated that the optimum temperature and humidity were 22.75ºC and 95.37% respectively. Multiple regression analyses indicated that these were the only variables significantly related to disease potential.
• Field trials indicated that cultivars react differently to changing environmental conditions and disease potentials. High genotype x environ­ment interactions were obtained. Most cultivars were susceptible to Sclerotinia stem rot under less than optimal conditions for disease development. A regression methodology was used to quantify cultivar behaviour under changing disease potentials. Differences in disease intensity indicate that the desired levels of resistance have yet to be attained and emphasises the need for more efforts directed at breeding of resistant cultivars. (McLaren, N.W. & Craven, M. (2008) Evaluation of soybean cultivars for resistance to Sclerotinia stalk rot in South Africa. Crop Protection 27:131-135).
• Various chemical and biological control agents are commercially available but these fail to control the disease effectively. Laboratory results revealed that chemicals applied at the regulatory rate failed to control mycelium growth effectively as opposed to the double rates which were effective throughout the evaluations. Biological control has the potential to act as a mycoparasite and suppress S. sclerotiorum. However further research is needed to optimize effectiveness. Benomyl and procymidone proved to be the most effective treatments in the laboratory and greenhouse. Application time relative to inoculation had a small but significant effect, indicating the need for correct application time.

5.13

Super Soya Competition

Mr C Havenga

The Super Soya Competition in KwaZulu-Natal is now in its eighteenth year and interest in the competition is still high.

This competition is an extension programme to promote the production of soybeans in KwaZulu-Natal. Interest in soybean production has increas­ed significantly over the past 18 years. However, the extent of soybeans planted annually is, to a large degree, influenced by the maize/soya price ratio.

Winterton and Bergville still fall under the South Region (S-KZN). This season there were 15 entries into the competition from S-KZN and 31 entries from NKZN.

This seasons average yield for the two competitions were significantly lower than for the previous season. These lower yields can be attributed mainly to the low rainfall during the growing season, especially during the critical period of February and March.

Yields in S-KZN varied between 1.07 and 4.87 ton/ha with an average of 3.01 tons/ha which is an average of 1.22 tons/ha lower than the previous season.

Eight fields (53%) were under irrigation with an average yield of 3.29 tons/ha ranging between 1.83 and 4.87 tons/ha. Yields under dry land ranged from 1.07 to 3.42 tons/ha with an average of 2.68 tons/ha.

Yields for this season's Super Soya Competition in N-KZN varied between 1.39 and 4.16 tons/ha with an average yield of 2.62 tons/ha. The average yield under irrigation was 3.73 tons/ha which is considerably higher than the average dry land yield.

In general the protein content of the soybeans in S-KZN was good to very good, with an average protein content of 39.73% (dry matter basis) which varied between 36.11% and 43.34% (dry matter basis).

In N-KZN the protein content of the soybeans was satisfactory to very good. The average protein content was 38.83% (dry matter basis) which is less than the previous seasons' 42.1%. The protein content ranged from 35.64% to 43.35% (dry matter basis).

The oil content of the soybeans in S-KZN varied between 17.92 and 23.15% (dry matter basis), with an average of 21.32% which is a big improve­ment on the previous season. There was no significant difference between the average oil content of soybeans under irrigation (21.26%) and those under dry land (21.39%).

The average oil content of N-KZN soybeans was 20.53% (dry matter basis) and varied between 18.31% and 22.66%.

This competition also creates an opportunity to record and evaluate the results of production practices from which production problems can be identified and research projects designed to try and solve these problems.

5.14

Evaluation of canola cultivars under irrigation

Dr AA Nel

For economically viable canola production, farmers should optimise production inputs. This implies selection of cultivars that are well adapted to a specific environment. Limited information on the performance of cultivars in South Africa is currently available.

New cultivars are introduced annually, which emphasizes the need for continuous cultivar evaluation. The aim of this project is to evaluate com­mercially available canola cultivars in terms of yield and seed composition at five localities under irrigation.

During 2006, 17 cultivars were evaluated in seven localities-planting date field trials. Five trials were completed successfully. Maximum and mini­mum temperatures regimes were favourable for canola. Early plantings grew lush resulting in lodging during the reproductive stage. Grain yields of cultivars varied from 0.9 to 5.4 t ha^-1 at the different localities The mean yield across localities and planting dates was 3.2 t ha^-1. The moisture free oil content of seed varied from 34% to 42%, with an average of 38%. The protein content varied from 23% to 28%, with an average of 26%.

5.15

Nitrogen requirement of canola under irrigation

Dr AA Nel and Mr NA Nkosi

The animal feed industry in South Africa relied on more than 842 000 tons of imported oil cake to satisfy their protein demand for 2005/2006. An increase in the local production of canola would, therefore, alleviate this dependence and could serve as a rotational crop in cropping systems under irrigation where diseases are problematic. Fertilisation is the most important cost-of-production item of dry land canola production in South Africa. Currently there are no scientifically-determined fertilisation guidelines for irrigated canola in South Africa and recommendations from else­where do not agree on application rates. The aim of this project was to determine guidelines for nitrogen (N) fertilisation on irrigated canola.

Field trials were planted during the winters of 2004, 2005 and 2006 on a sandy soil at Vaalharts and a loamy soil at Potchefstroom. Six fertilisation rates varying from zero to 300 kg N ha^-1 were applied on two cultivars. Biomass samples were collected on three occasions during the vegetative to the beginning of flowering stages during the 2004 and 2005 and four samples during 2006. The biomass samples were analysed for N content. These data were used to validate an N-dilution curve developed for oil seed rape in Europe. At the appropriate stage the crops were harvested, yield determined and seed from both localities were analysed for oil and protein content. The N content data fitted the N-dilution curve well and could therefore be applied on canola in South Africa for the identification of an early N-deficiency. Calculated optimum yields and associated N rates showed a high degree of variability between seasons and localities. A general N-fertilisation rate of 204 kg ha^-1 for irrigated canola is recommended when no yield target is applicable. The deltayield procedure, which takes the soil N supply and crop yield potential into account, shows potential as a reliable method for determining N-fertilisation rates. In this procedure the N-fertilisation rate (kg ha^-1) = D0.687 where D = delta yield expressed in kg ha^-1. This equation is suitable for both dry land and irrigated canola. Seed oil content of the cultivars differed in two seasons and was also influenced by an N-fertilisation x locality interaction in 2004 and a locality x cultivar interaction in 2005. The seed protein content increased with increased N-fertilisation rates. It was also influenced in one or more seasons by cultivars, localities, seasons and several interactions. The inconsistency of treatment effects across seasons indicates that different factors played a role in the determination of seed oil and protein content.

5.16

An evaluation of crop rotation with canola under irrigation

Dr AA Nel

Canola is an oilseed with low fibre and high protein contents, characteristics preferred by the animal feed industry. An increase in local canola production will reduce domestic dependence on protein imports. Being part of the Brassica family, canola is not a host to cereal diseases and has the potential as rotational crop for improved sustainability of cereal cropping systems under irrigation.

1. To determine the effect of a canola crop on subsequent crops or the effect of those crops on canola in one- and two-year rotation systems,
2. To determine the effect of canola on the nematode population,
3. To monitor the incidence and severity of diseases on all relevant crops, and
4. To compare the profitability of the different crop rotation systems after completion of the field trial.

Results from four summers- and three winter seasons are as follows: Crop rotation system had a significant effect on the yield of maize (mean yield 10 102 kg ha^-1). Maize yield in the canola-maize, one-year system was 18% lower than that of maize following canola in the canola - maize - wheat - maize system and 12% lower than that of maize in the canola-groundnut-wheat-maize system. Yield of maize in the one-year wheat-maize system was 9% lower than that of maize in the two-year systems where canola and groundnut were part of the cropping sequence. Groundnut yields (mean 3 991 kg ha^-1) were affected by crop rotation system. As in the case of maize, groundnut grown in the most diverse crop system yielded best. The yield of groundnut following canola was 14% and 9% lower than the yield of groundnut preceded by wheat and barley respec­tively. Crop systems affected the yield of canola which varied from 872 to 3 142 kg ha^-1. Ranking of canola yields according to system changed from season to season. Barley yield was not influenced by crop rotation (mean yield 6 487 kg ha^-1). Wheat (mean yield 5 962 kg ha^-1) was affected by crop system. Ranking of wheat yields according to the rotation systems or preceding crop, varied from season to season. Across seasons the nematode population in the soil was influenced by crop rotation but the variation in numbers of nematodes in the roots of crops cannot be logically explained. Differences in nematode numbers were relatively small. Maize streak disease was observed in one season and a few wheat plants showed root rot symptoms in a second season. No diseases of any significance were observed in any of the other crops.

5.17a

Canola cultivar evaluation of oil and protein crops in the winter rainfall region

Mr PJA Lombard

The purpose of the project is to identify the best adapted canola cultivar with the highest protein content for a specific area and secondly to pro­mote canola as an alternative rotation crop for wheat. According to the Crop Estimates Committee, 32000 ha of canola was planted in the Western and Southern Cape during the 2007 season.

Climatic conditions were favourable for crop production in the Swartland with adequate rainfall during the production season. Heavy rainfall at the start of June caused water logged conditions, resulting in the scrapping of the trials at Grasrug (Malmesbury), Voëlvlei (Piketberg) and Poskantoor (Porterville). In the Southern Cape the moisture was favourable although some short dry periods occurred. The trial at Mosselbay was written off due to root rot at the seedling stage.

Trial yields in the Southern Cape varied between 1.6 ton/ha (Tygerhoek, first planting) and 2.36 ton/ha (Riversdal). The difference between the first and second planting at Tygerhoek was 12%. It was smaller than expected but the below average temperatures late in the season may be the expla­nation. The yield of the Clearfield and Triazine tolerant cultivars were respectively 10% and 15% lower than the conventional cultivars. The 3 top producing cultivars were Muster (2.27 ton/ha), Hyola 61 and Opal and they averaged over 2.1ton/ha. The difference between the highest producing cultivars and the cultivar in the 8th position was only 143 kg/ha. Thunder (1.97 ton/ha) was the top producer in the Triazine group and 45Y77 (1.95 ton/ha) was the best Clearfield cultivar.

Trial yields in the Swartland varied between 2.04 ton/ha (Langgewens, second planting) and 2.91 ton/ha (Darling). The difference between the first (8 May) and second planting (22 May) at Langgewens was 18%. There was a visible correlation between the vegetative growth and yield. The below average temperatures were favourable for plant development and yield.

The Clearfield cultivar 45Y77 was on average the second best of all cultivars. The yield of the Clearfield and Triazine tolerant cultivars were respectively 1% and 30% lower than the conventional cultivars. The three top producing cultivars were 44Y06 (3.15 ton/ha), 44Y77 (3.07 ton/ha) and 44C11 (2.87 ton/ha).

The big difference in yield between early and late planting at Langgewens is normal and is an annual occurrence. This highlights the fact that late planting is a risk and will most probably result in a lower yield. Conventional cultivars were the top producers but the options in controlling weeds are limited. The option of a high producing Clearfield cultivar is now available for the first time to producers.

5.17b

Lupin cultivar evaluation of oil and protein crops in the winter rainfall region

Mr PJA Lombard

The purpose of the project is to identify the best adapted cultivar for a specific area and secondly to promote lupin as an alternative rotation crop for wheat. According to the Crop Estimates Committee 14000 ha of sweet lupins was planted in the Western Cape during the 2007 season.

Climatic conditions were favourable for crop production in the Swartland with adequate rain during the production season. Heavy rainfall at the start of June caused water logged conditions, resulting in the scrapping of the second planting at Langgewens due to root rot. In the Southern Cape the moisture was favourable although some short dry periods occurred. The trial at Mosselbay was written off due to damage by bushbuck.

Ten lupin lines were evaluated for the first time for possible future registration as cultivars for the Western Cape. Trial yields in the Swartland varied between 1.6 ton/ha (Eendekuil) and 2.84 ton/ha (Philadelphia). Brown leaf spot infected the trial at Hopefield and the L. albus cultivars showed the worst infection, so much so that there was no yield. The three top lines were 42 (2.65 ton/ha), 35 (2.49 ton/ha) and 36 with 2.45 ton/ha. Mandelup (2.39 ton/ha) was the highest producer among the cultivars followed by Quilinock (2.14ton/ha) and Tanjil (1.94 ton/ha). At Eendekuil porcupines damaged the L. albus cultivars CED 6150 and HE 16.

The second planting (25 May, 1.78 ton/ha) at Tygerhoek in the southern Cape yielded over 200 kg/ha higher than the first planting (30 April, 1.57 ton/ha). The narrow leaf cultivars were the top producers in 2007. Mandelup (2.36 ton/ha) average the highest followed by Quilinock (1.96 ton/ha) and Tanjil (1.78 ton/ha) among the cultivars. Seven of the lines produced higher than the cultivar Quilinock. The lines, 24 (2.23 ton/ha) and 25 (2.15 ton/ha) were placed second and third in the trials behind Mandelup (2.36 ton/ha).

Several lines showed potential. The best lines in the Swartland (42, 35 and 36) and Southern Cape (24, 25 and 16) differed from each other. To draw a proper conclusion on which lines are best suited as cultivars, the promising lines must be evaluated for the next two seasons before making the final decision.

5.18

An investigation into the production dynamics of eight crop rotation systems including wheat, canola, lupins and pasture species in the Swartland, Western Cape

Dr MB Hardy

The aim of this study is to determine the short- and long-term effects of eight of the most feasible crop and crop/pasture rotation systems on: crop yields; weed control; disease suppression; soil production potential; sheep production; economically sustainable land-use in the Swartland.

No significant rainfall was recorded during February and March. The first significant rains fell on the 26th of April (25.6mm). A total of 333mm fell between 1 April and mid October made up of relatively consistent rainfall events. Moisture availability was therefore high for most of the growing season and the hot dry periods that often occur during the growing season did not materialize. Maximum daily temperatures varied from about 11 to 21ºC for most of the growing season. Excellent wheat yields were recorded for the season and the average lupin yields were unexpectedly higher than the average canola yields.

Canola and lupin production

The TT canola cultivar, ATR Stubby, was planted to increase the opportunities for controlling herbicide resistant ryegrass on the research site. Canola and lupins (Tanjil) were planted on 1st and 8th of May respectively. Seedling densities for canola and lupins (54 & 51 seedlings m²) were higher than expected, indicating a better than estimated seedling survival. Management protocols developed by the Technical committee were followed but adjusted during season as a function of variable climatic conditions. Favourable climatic conditions resulted in high yield potential.

Canola and lupins were harvested at the end of October. As in previous seasons canola ripening and "readiness" for harvesting was extremely variable and uneven. Each canola camp was therefore also sampled before harvesting to obtain estimates of yield potential.

Canola yields obtained during harvesting ("straight") ranged from 628 kg/ha to 1582 kg/ha with an average yield of canola over all plots of 1103 kg/ha. There were significant differences between the actual yields and estimated yields. The average difference in seed loss over all plots was 38 % with the average potential yield for 2007 being 1780 kg/ha. Given that the average yield of wheat over all wheat plots was 4396 kg/ha, canola yields were low viz. 25.1% of the wheat yield. Of interest is the fact that the estimated yields show that canola following medic pastures has a higher production than canola following wheat or lupins. ATR Stubby also had massive lodging problems due to the occurrence of Blackleg, which had a negative effect on harvesting.

Economics

All crop production data for the 2007 season will be analysed as soon as the financial program that is being developed by the Department of Agricultural Economics at the University of Stellenbosch becomes available.

5.19

Economic sustainability of short- and long-rotation crop / pasture production systems in the Southern Cape

Dr MB Hardy

The aim of this study is to determine the short- and long-term effects of a number of the most feasible short- and long-rotation, crop and crop / pasture rotation systems identified for the Southern Cape region on: crop yields, weed control, disease suppression, soil production potential, sheep production and the economically sustainable land use in the Southern Cape.

2007 was the 6th year of production. The trial comprises two main components namely: short-rotation systems that are being tested at Tygerhoek experimental Farm at Riviersonderend and long-rotation systems that are being tested on farms in the Riversdal and Swellendam districts respect­ively. Note that 2007 was the 1st year of the 2nd 5-year cropping phase that is being tested at the Riversdale and Swellendam sites.

All trial areas were successfully established and managed according to the planned protocols. As stated in the trial protocols, wheat was the only crop planted at the Swellendam site while wheat and canola were planted at the Riversdale site in 2007.

Rainfall during the latter part of the summer and early autumn (before planting) was favourable at all sites and resulted in good moisture levels in the soil. Daily maximum temperatures were also not excessive. These favourable climatic conditions persisted for most of the growing season, except at the Swellendam site where very little rain fell from August to October resulting in a dramatic decline in yield relative to the mid-season potential for the site.

Canola production

Canola (Comet) was included as the first crop following lucerne at Riversdale in two of the crop sequences being tested. This allowed for the po­tential to grow 2 crops of canola in the 5-year cropping phase before lucerne is re-established in the camp i.e. the 1st and the 5th year. A seeding rate of 3.0 kg/ha was used and a total of 39 kg N/ha was applied to each plot (15N/ha at planting and 24N/ha top-dressing).

The cultivar ATR Stubby was planted at the Tygerhoek site at a seeding rate of 3.8 kg/ha together with 20 kg N/ha, 15 kg P/ha and 10 kg S/ha on 7 May 2007. Simazol was applied at planting onto wet soils. Only one N top dressing was applied during the season. Weed and pest control meas­ures were applied at both sites as required. Refer to details on file for weed, pest and disease control measures.

Canola yields at Riversdale were excellent ranging from about 2000 kg/ha to 2500 kg/ha clearly showing the potential benefits of planting canola in the first year after lucerne in the Riversdale "Vlakte" region.

The average canola yield over all plots at Tygerhoek was 1466 kg/ha, which is acceptable in comparison to the expectations of the region. Inter­estingly, the average yield of canola following a pasture crop (1724 kg/ha) was approximately 500 kg/ha more than the average yield of canola following a cash crop (1208 kg/ha).

Average oil content of seed (as determined at the SSK facility in Swellendam) was 43.5% and 37.5% for the Riversdale and Tygerhoek sites respectively. This figure is considerably higher than the figures presented in previous reports where oil content was determined in the Animal Production at Elsenburg Laboratory.

Lupin production

Crop rotation treatments at the Riversdale and Swellendam sites did not include lupins. The cultivar Tanjil was planted at the Tygerhoek site. Lupin yields were much higher than those recorded in 2006 at the same site. Plant growth was vigorous during the season. There appeared to be suffi­cient branching and flowers per branch. The average production over all four plots was 1901 kg/ha. Crude protein content of lupin sample did not vary between plots (average = 28.5% CP) indicating that crop sequence in the previous season had no effect on this quality parameter.

Economics

A new database for capturing and analysing economic data has been commissioned to facilitate more efficient economic analyses of the data sets. The data base is being developed by the Department of Agricultural Economics of the University of Stellenbosch and will be available in mid 2008.

5.20a

The promotion of canola as a rotational crop within a conservation farming system in the dryland planting region of the Swartland, through the use of producer competition

Mr IFV Slabbert

Four farmers in the Moorreesburg, four farmers in the Piketberg / Porterville and five farmers in the Malmesbury / Durbanville areas entered the competition.

The decline in the number of regular entrants to the competition in the Morreesburg and Piketberg / Poterville areas can be ascribed to the low price of Canola in relation to wheat. Planting of Canola is more risky in these areas and participants who do not want to plant Canola anymore contend that the Canola yield is not adequate.

Rainfall during April was low, except in the Rooikaroo north of Piketberg where early thunder showers were experienced. This lead to early planting by some producers.

There was sufficient rain in May and all the other farmers could then sow their Canola and wheat. Canola was never under moisture stress; on the contrary there was the possibility of waterlogged conditions.

5.20b

The promotion of canola as a rotational crop within a conservation farming system in the dryland planting region of the Southern Cape

Mr C van Rooyen

The competition started off with 22 participants, but due to isopod damage some camps were eliminated. Canola producers affected by isopod damage have become demoralised and were demanding answers to the isopod problem. The cabbage stem beetle was found in some canola camps and on all of the Crambe trial plots. The Department of Agriculture Western Cape is monitoring the situation. A positive outcome of the rainy weather at harvest time was that canola could be harvested in the early mornings before other grains. The top third of the respondents in 2007 achieved an average yield of 1.9 tons/ha – a feat last seen in 2003. The price of canola versus barley should lead to an increase in the hectares planted to canola in 2008. The new cultivars that are now available could bring an end to the use of farm seed in the near future. Good margins achieved with canola production in 2007 have cemented its place in every rotation system in the Southern Cape.

5.21

N-fertilization of canola based on N-mineralisation and leaching

Prof GA Agenbag

Field experiments were conducted at Langgewens, Elsenburg, Welgevallen and Roodebloem Experimental Farms to evaluate the response of canola to increasing (0-120 kg N ha^-1) N application rates. Results were however very variable due to differences in prevailing climatic conditions, yield potentials, N-mineralization and nitrogen leaching. Although yields did increase significantly with increasing N application rates at several localities, optimum N application rates differed largely between localities and years. Agronomic efficiency at maximum yield level also proved to be generally low with yield increases of less than 4 kg for every kg of N applied. Plant analysis done at 90 days after planting at Langgewens and Roodebloem showed extremely high sodium contents, but more research is needed to determine whether these have any negative effect on grain yield and quality. Analysis done on soil samples collected at planting at Langgewens and Roodebloem revealed very low Boron and Sulphur con­tents at both localities. Sufficient supply of both elements is very important to ensure high yielding canola crops and this aspect will for this reason require some research.

5.22

Salinity tolerance of canola

Prof GA Agenbag

In 2007 research focused on the response of different canola cultivars to increasing salinity levels in pot trials as well as their response in field trials. The experiment done with 12 different cultivars showed that the emergence of most of the cultivars tested was reduced by an EC of 8 dS m^-1, but growth of the surviving seedlings was not affected. Although the final percentage emergence of cultivars Spektrum, Tornado and Hyola 61 was not significantly reduced with an increase in EC from 0 to 8 dS m^-1, Spektrum showed some higher tolerance during the early growth (seedling) phase. As the response to salinity may differ between the different crop growth stages, these results however did not necessarily indi­cate that Spektrum will also be more tolerant with regard to grain yield. Field experiments done by the Department of Agriculture Western Cape showed that Spektrum performed particularly well at localities with a low soil resistance (high EC). However, more detailed field studies are needed before final conclusions with regard to the salinity tolerance of canola in general and that of Spektrum in particular can be made. Such studies will have to involve soils with a wide variety of soil physical and chemical properties.

5.23

Characterization and management of Rhizoctonia on canola and lupin cropping systems in the Western Cape province

Dr SC Lamprecht

Rhizoctonia anastomosis groups (AGs) associated with canola and lupin grown in crop rotational systems in the southern and western production areas in the Western Cape province of South Africa, were recovered during 2007 and identified using sequence analyses of the rDNA internal transcribed spacer regions. Four crop rotation trials, two each at the Tygerhoek (southern Cape, Riviersonderend) and Langgewens (western Cape, Moorreesburg) experimental farms were included in this study. Four tillage practices (zero-till, no-till, minimum till and conventional tillage) were also evaluated in one trial each at the two localities. Canola planted after barley, medic/clover mixture and wheat, and lupin planted after barley and wheat were sampled at the seedling, mid-season and seedpod growth stages. A total of 298 isolates were obtained of which 72.8% were binucleate and 27.2% multinucleate Rhizoctonia isolates. The most abundant binucleate AG was AG-I (30.9%), followed by AG-K (23.8%), AG-A (10.7%) and AG-Bo (7.4%). The most abundant multinucleate AG was AG-2-1 (15.4%) followed by AG-2-2 (7.7%) and AG-11 (4.0%). All the AGs were obtained from both Langgewens and Tygerhoek.

Crop rotation affected the incidences of individual AGs as well as the binucleate and multinucleate group. AG-2-1 was only isolated from canola and AG-11 only from lupin. Significantly higher incidences of the binucleate group were recorded on lupin than on canola in trials 2 and 4 and significantly higher incidences of the multinucleate Rhizoctonia group on canola than lupin in the same trials. Tillage practices did not affect the incidences of the Rhizoctonia AGs at Tygerhoek, but at Langgewens, the binucleate Rhizoctonia group was more frequently obtained from plants from the conventional tillage treatment and AG-2-1 was more frequently obtained from the zero till than no-till and minimum till treatments. Inci­dences of Rhizoctonia AGs were generally higher in the beginning and end of the season than mid-season. This study confirmed results obtained in 2006 that important Rhizoctonia AGs such as AG-2-1, 2-2 and 11 occur in both the southern and the western production areas of the Western Cape province and that crop rotation affects the incidences of Rhizoctonia AGs.

5.24

Insects and other pests of canola

Dr G Tribe

A trial to determine whether the losses of canola seedlings caused by slugs and isopods could be circumvented by planting later, and the correct timing of the application of slug pellets was carried out on four sites in the Caledon district. The split plot design consisted of two planting dates with four treatments, viz. control, one and two applications of slug pellets, and an insecticide application. Each plot contained five melthoid traps and the numbers of the different organisms found under each was counted at weekly intervals together with the number of seedlings in four areas (1x 0.5m) within each plot.

There was an almost total loss of seedlings in three of the plots where they failed to emerge and these losses were ascribed to isopods feeding on the germinating seeds below the soil. The fourth plot lost between 15 -20% of the seedlings and this was due to slugs because no isopods were present due to the extremely wet soil. Although the 10 permanent melthoid traps in the pasture at Roodebloem recorded a 1.5 fold increase in the number of isopods and a 17 fold increase in the number of slugs compared to the previous year, the major losses were due to isopods. The slugs were recorded in four classes, and the majority were in the 'small' and 'medium' classes with relatively few in the 'tiny' class, indicating that more juveniles from the previous year had survived the dry summer months due to rainfall in November / December 2006. Because of the erratic rainfall experienced in 2007, early and late planted canola suffered equal losses.

Three slug pellet formulations were tested and all were shown to be highly effective in attracting and killing slugs. The carbaryl + metaldehyde formulation also killed 100% of the isopods and all other insects, including beneficial predators. The slugs are of Mediterranean origin and their control can be achieved through the judicious use of slug pellets but the timing and number of applications have yet to be accurately determined. The isopods on the other hand are indigenous and an integral part of the decomposition of stubble in conservation farming. The emphasis here will have to be on the protection of the seed and seedling over the 14 day period when they are vulnerable to attack, rather than an onslaught on the isopods themselves. Future trials will concentrate on seed treatments.

Continuing investigations of the insect pests of canola revealed that all samples of plants infested with the aphid Brevicoryne brassicae were 100% parasitized by the braconid parasitoid Diaeretiella rapae. Two different aphid species were present on the ubiquitous European weed Sonchus asper and were not parasitized. This indicates that S. asper harbours neither the pest aphid nor the parasitoid and therefore is neither a reservoir for the canola aphid nor the parasitoid.

Four parasitoid species of the Diamondback moth were recovered from samples, but only one species Diadegma mollipla accounted for almost all the parasitism which amounted to 30% of the moth larvae by all parasitoids combined.

There was a 100% infestation of the stalks of Crambe by the cabbage stem weevil Ceutorhynchus pallidactylus and a 68% infestation of canola stalks. It appears that the weevil has a single generation per year and that they pupate and over-summer in the soil.

A new beetle pest of canola was discovered which was active when most canola plants had been wind-rowed and for this reason does not appear to be a serious pest. But although tiny, they occur in large numbers and have the potential of destroying the foliage on which they feed. It has been identified by the PPRI-National Collection of Insects through Dr Biondi in Italy as Phyllotreta mashonana.

5.25

A comparison of the response to dietary protein by the Cobb and Ross broiler strains used in South Africa

Prof RM Gous

The two most common broiler strains used in South Africa, Cobb and Ross, have been shown to respond very differently to dietary protein supply, which raises the important issue of how to design feeds that will optimise performance in the two strains. Clearly, the feeds that will optimise performance in the one strain are unlikely also to be best for the other strain. In a follow-up to the trial reported last year, responses to a range of dietary protein levels were again measured. However, in this trial four strains were used and they were reared to 15 weeks of age rather than only to six weeks. Once again, in addition to measuring the conventional production variables such as weight gain and feed conversion, birds were sampled weekly and the weights of their physical parts were recorded before the birds were minced for chemical analysis. The weights of the physical parts of males and females of the Cobb, Hybro and two Ross strains on three levels of protein were regressed against the weight of body protein at the time of sampling, to determine allometric relationships between these parts and body protein.

The four strains and two sexes once again exhibited almost the same relationships between body protein weight and breast meat, thigh, drum and wing weight, suggesting that the conformation of broilers has not been altered by genetic selection, and that reported differences in breast meat yield between strains are likely to be the result of birds having been harvested and compared either at different degrees of maturity or at different body lipid levels. As with the previous trial, low dietary protein contents resulted in heavier component weights at a given protein weight than did higher dietary protein contents. Because the amount of lipid deposited in the body is inversely related to the dietary protein content, low protein feeds would be expected to result in heavier body components at the same protein weight. A further trial has therefore been started to determine the rate of lipid deposition in each of the body components on low protein feeds.

The conclusion from this research is that differences in efficiency of utilisation of dietary protein by the most common strains of broiler used in South Africa cannot be attributed to differences in the rates of growth of the physical parts.

Geneticists from the various breeding companies have used different selection criteria to improve performance in these strains, and the challenge is to describe these differences in such a way that simulation models can be used to determine the biological and economic optimum protein (amino acid) contents of feeds for the different strains.

5.26

Calibration of near infrared spectrometer for amino acid analysis in animal feed

Dr M Ciacciarello and Prof RM Gous

The need for accurate amino acid analysis in raw materials and feed for the animal industry is well known. The success of a feed in meeting the amino acid requirements of any farm animal depends largely on its composition, and ultimately on the correct evaluation of the ingredients' nutritive value. In the past few years, the use of near infrared spectroscopy (NIRS) to analyse raw materials and feeds has been on the increase. However, calibration of the equipment is a complex process and it determines the accuracy of the results. This technology has certain advantages when compared to the conventional chemical analysis, mostly related to the time required to obtain the results. Although the cost of purchasing the equipment could be considered a disadvantage, these instruments are becoming more affordable over time. Furthermore, considering the advantages and savings brought about by using this technology, the initial investment can be reasonably justified. Samples of fishmeal, maize, sorghum, full fat soya, soya oilcake and lupins (118, 106, 217, 152, 112 and 57 samples respectively) were collected. These samples were analysed for moisture, crude protein, crude fat, ash, gross energy, AME, total and digestible amino acids. The results of these analyses were then used to build the calibration curves that would be used to speed up the amino acid analysis for raw materials and animal feed, and to determine amino acid digestibility.

Twenty samples of each raw material were selected at equal intervals of protein concentration. For fishmeal the range of lysine concentration was 4.4–5.8%. The protein content ranged between 56.9% and 71.0%. The lysine content of maize was found to be between 0.21 and 0.42%. Its protein content was between 5.25 and 13.25%. The crude protein concentration in sorghum was between 5.76 and 16.67%. The lysine results for this raw material are still being analysed. Full fat soya showed a concentration of lysine from 2.24 to 2.68%. The protein content was between 36.04 and 41.58%. Soya oilcake samples contained from 2.4% to 3.4% lysine whilst the protein content was found to be between 39.8% and 49.4%. Lupins had a lysine concentration between 1.53 and 1.84%, with protein contents of between 28.1 and 37.6%.

Once the samples were analysed using conventional wet chemistry, the samples were tested in a FOSS NIR system. All samples were run through the system twice to give two readings on as-is bases. All samples were then dried and two further readings were obtained on dry matter bases. Calibrations were plotted against the known analysis on an as-is and a dry matter basis. Calibration curves showed higher accuracy where the raw materials were scanned on a dry matter basis compared with the as-is basis. Very good calibrations were achieved for fishmeal (protein, lysine, AME, fat, and ash) as well as sorghum. For the other raw materials the calibration was not as robust, but it is expected to improve with the analysis of more samples. Digestible amino acids and ME of the different protein sources are being calibrated at present.

5.27

Amino acid and protein utilization of broiler breeder hens fed maize-soya diets

Dr M Ciacciarello and Prof RM Gous

The amount of energy to be allocated daily to broiler breeders is difficult to prescribe because their daily energy requirements change with egg production and with fluctuations in environmental temperature. Being control-fed, these birds do not have the option of increasing feed intake as temperature decreases, as a means of meeting their increased energy requirements. When the environmental temperature falls, the amount of food or energy provided must therefore be adjusted, so full-fed laying hen models are not appropriate for broiler breeders.

An experiment involving 288 Cobb broiler breeder females housed in six environmentally controlled chambers at 44 weeks of age for a twelve week period, was designed to study the effect of temperature on the performance of broiler breeder hens fed fixed daily quantities of nutrients (other than energy), and to investigate how these hens partitioned their dietary ME as the environmental temperature decreased.

Egg production declined linearly throughout the range of temperatures used when hens were fed 11.9, 10.5 and 9.7MJ ME/kg feed, indicating that the birds were using increasing amounts of dietary energy for maintenance as the temperature declined, leaving less available for production, while birds fed 12.9MJ/kg (2000kJ/d) maintained their performance at all temperatures. Low temperatures have no effect on egg production in laying hens fed ad libitum, but where energy intake is restricted, as was the case with broiler breeders on the lower energy allocations, energy would need to be partitioned differently, the most likely scenario being that rate of laying would be reduced to accommodate the higher maintenance requirement.

It appears from this study that broiler breeders, on a fixed daily allocation of dietary energy, will reduce egg output when faced with an energy deficiency, which has the result of increasing lipid reserves when the feed is adequate in protein content due to the deamination of the protein not used for egg production. The additional energy required for cold thermo genesis amounts to 1.0kJ AME/kg W. ºC, which should be added to the maintenance requirement of broiler breeders for environmental temperatures below about 18ºC.

5.28

The effect of lighting on the efficiency of utilisation of dietary protein by Cobb and Ross broiler strains used in South Africa

Prof RM Gous

Rearing broilers on 6-h photoperiods and transferring them to 23 h at 21 d has been shown to reduce mortality and the incidence of leg disorders without adversely affecting final body weight or feed conversion efficiency. However, in many countries, welfare codes for meat chickens currently stipulate a minimum photoperiod of 8 h, and they are likely to specify a minimum uninterrupted dark period of 8 h in the future. In a follow-up to the research reported last year, the responses of two genotypes of male broilers were measured to constant 8- and 16-h photoperiods, and to an abrupt transfer from 8 to 16 h at 10, 15 or 20 d. Body weight, feed intake, and feed conversion efficiency were not significantly different at any stage during the 35 d study. Mortality and the incidence of Sudden Death Syndrome were similar for all lighting groups at 35 d.

When these data were pooled with previously reported data for female broilers, growth and feed conversion efficiency post 21 d and through to depletion for constant 8-h and birds transferred from 8 to 16 h at 20 d were significantly superior to constant 16-h birds. Constant 8-h birds ate about half their feed during the dark period, whilst 16-h birds consumed no more than 10%. Birds which had been started on 8 h and transferred to 16 h at 10, 15 or 20 d, reduced their rate of nocturnal feeding when changed to the longer photoperiod. However, they still consumed more feed in the 8-h dark period than birds that had always been given 16 h illumination. Cobb and Ross genotypes responded similarly to all lighting treat­ments. This work is especially important in demonstrating that the efficiency of utilisation of feed protein is improved with the use of short day lengths, and that the use of electricity would be considerably reduced if broiler producers made use of the results of this research.

5.29

Evaluation of Distillers Dried Grains and Solubles (DDGS) as a protein source for poultry and pigs

Prof RM Gous

No progress was made in this project during the year, mainly because it proved impossible to import DDGS into South Africa last year. Meanwhile, the South African Government decided against the use of maize for ethanol production in this Country, so the impetus to conduct this research was reduced. Nevertheless, the Protein Research Foundation agreed that there was merit in continuing with this research, given that there would be large amounts of DDGS available on the market in the future, and that this product may well be worth importing into the Country if the price were right.

5.30

Income and cost budgets

Mr JSG Joubert

During the year, information was gathered in the Winter rainfall regions with regard to canola, lupins and wheat and in the Summer rainfall regions with regard to soybeans and maize. The information in the Winter rainfall regions was obtained from Agri-businesses as was the information from Summer Rainfall regions with regard to the Free-State and North-West, while the information on KwaZulu-Natal was obtained from the Department of Agriculture in this province. Mpumalanga information was collected through the group discussion technique. The information, which is of an economic and technical / biological nature was processed and income- and cost budgets were compiled for the abovementioned crops. The main components of such a budget are made up of gross income, production costs and gross margin on a per hectare basis. The income- and cost budgets are highly in demand and can be applied for several uses by producers, agri-businesses and financial institutions. It also serves as an important source of technology transfer.

Income and cost budgets of soybean and maize were compiled in the Summer Rainfall regions with regard to the following areas:

In the Winter rainfall region of the Western Cape province, income and cost budgets were compiled for canola, wheat and lupins. (Last mentioned was not compiled for Malmesbury and Porterville).

Income and cost budgets attempt to reflect a typical situation, given specific cultivation practices. Recommended practices and data from study groups were used as sources of information.

5.31

Models: APR- and Nieuwoudt / McGuigan Model

During the report year, the APR and Nieuwoudt/McGuigan models were used supplementary to each other to develop projections of protein requirements for 2010 and 2020. Both models were developed by the University of KwaZulu-Natal. The APR model was developed for a Doctoral Thesis funded by the PRF and the Nieuwoudt/McGuigan model for a Masters Thesis. The latter was initiated and funded by the PRF.

The Department of Agricultural Economics of the University of the Free-State is currently running and maintaining these models under contract for the PRF.

The APR model is used to divide the total feed consumption, which includes protein between the animal species, while the Nieuwoudt/McGuigan model is used to make projections of protein requirements for 2010 and 2020.

During the report year discussions were also held with representatives of the Department of Agricultural Economics of the University of Pretoria regarding the BFAP (Bureau for Food and Agricultural Policy) model. The BFAP model, is a model with a wider reach that analyses global and local markets and, based on base line projections and scenarios, forecasts future trends and calculates the possible impact thereof on local markets and farms as well as profitability and survival capacity. For instance it forecasts what would happen with the gross value of specific sectors, areas under crops, human- and animal consumption of certain commodities as well as future prices of commodities.

The purpose of the discussion with BFAP was to determine the possibility of integrating the three models in order to make more accurate projec­tions on the future requirements of protein. It was evident that it did not make sense to integrate the models, but rather to use the BFAP model to supplement the existing PRF models. During the discussion, the PRF also provided certain inputs which were included into the BFAP model on a trial basis. It was also agreed that a similar work session would be held the following year, but that specialists from the most important industries would be included in order to provide better data for the model.

It is envisaged that the information from the BFAP model will be used as input into the APR and Nieuwoudt / McGuigan models.

5.32

PRF website

The PRF Website is still one of the most important vehicles used to communicate with stakeholders. The Website is also used to provide feedback to the PRF regarding issues such as bursary- and research project applications. A popular addition to the website is information on the Homepage that indicates what has recently been added to the website. There has also been an addition that provides important news snippets. The price graph of imported soya oil cake is also updated on a daily basis.

A list of important dates of upcoming events reminds visitors to plan in advance to attend these events. Flashes of important deadlines also appear on the Homepage.

In last year's report it was mentioned that a framework for a research database was added to the website into which details of research findings could be uploaded. The framework was divided by commodity (eg. soybeans), subject (eg. fertilisation) and sub-categories of subjects. The con­sumption of protein by animals is another section of the framework. During the report year, a project to upload documents containing information of interest available in the PRF offices was also initiated. Good progress has been made on this initiative. During the report year biofuels was also added as a new subject.

During the report year we added links to sixteen important websites, this being an effective way of increasing traffic to the website and also to increase automated web search visibility.

The most important source of reference of visitors to the PRF website remains search engines, with 70.7% of visitors to our website originating from here. Direct hits are 20.6% and visitors from other websites amount to 8.7%.

The most important sources of reference to the PRF website are AFMA, Aardvark, Google, SABA and SAGIS.

International visitors to the website make up as much as 31% of total hits each month.

Traffic to our website continues to increase, the numbers being:

In the first five months of 2008, the number of visits already total 5 205.