5.

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

5.1

National soybean cultivar trials

Mr JL Erasmus, Mr HSJ Vermeulen and Mr NW Mogapi

A total of 20 soybean cultivars were evaluated nationally in selected localities over the soybean production areas and an additional eight were evaluated as a phase-1 test at 11 additional 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 optimum amount of data could be collected. The comprehensive data set comprises phenological data such as 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 entered into a computerised data basis and used to calculate yield reliability and regression lines for the three main production areas differentiated as cool, moderate and warm. At the Cultivar Evaluation Committee meeting held during September 2008 one entry was approved for registration on the variety list namely, Lex D 2103 R. The 2007/08 report on the national soybean cultivar trials was also accepted for publication. Mean seed yield at all localities evaluated was 2 456 kg ha^-1 (1 717 kg ha^-1 for 2007 and 2 811 kg ha^-1 for 2006). Highest yields were obtained at Vaalharts (4 399 kg ha^-1), Cedara (3 394 kg ha^-1) and Koedoeskop (3 190 kg ha^-1). The cultivar LS 678 (2 611 kg ha^-1) yielded best on average overall and SNK 500 (2 171 kg ha^-1) the poorest. The cultivar with the highest percentage protein content was Stork with 42.9% and lowest was LS 6150 R with 40.0%.

5.2

The increase in soybean production on the Highveld

Mr WF van Wyk

The aim of this project is to define guidelines for the cultivation of soybeans on the Highveld.

The following three trials were executed under this project:

• Evaluation of row width, plant density and cultivar;
• Evaluation of fertilisation (potassium and phosphate levels); and
• Evaluation of different pesticides.

Row width, plant density and cultivar

The various plant densities showed significant differences in terms of yield between low and high densities in the 2006/2007 season. However in this season there were no significant differences. The yield on the 25 cm row width was significantly better than the 100 cm row width, but not significantly better than the 50 and 75 cm row widths. The plant and seed heights varied significantly between the different cultivars, row widths and plant populations. The protein and oil contents were lower than last season.

The difference in yield from a plant population of 400 000 pl/ha at a row width of 25 cm (3965 kg/ha) and a plant population of 400 000 pl/ha at a row width of 100 cm (3435 kg/ha) was 530 kg, which is not significant. However, if a monetary value were to be assigned to this (at R3500/ton) then an additional R1855/ha would be realised for the 25 cm row width in the high population. The straight-type cultivar which was planted this season (LS 6162) did not cover the leaf roof in the 75 and 100 cm rows in either the 300 000 or 400 000 plant stands. This is also a very fast cultivar and could possibly be used in the future to fight sclerotinia due to the fact that flowers form before any infection and also due to the fact that the micro-climate within the rows is not optimal for the formation of sclerotinia.

Fertilisation trial

In the fertilisation trial, significant differences in yield resulted from different potassium levels. It is also evident that yield correlates positively with the quantity of potassium, but that there is no pattern with regard to phosphate levels. This observation indicates once again the importance of the amount of calcium in the soil for absorption by the plant. The lack of reaction with the phosphate levels can probably be ascribed to the fact that the soil P (13 mg/kg P) is still sufficient even at the phosphate 0-level. All treatments were applied on the same research area. The Phosphate 0-level would therefore decline over time. Since potassium and phosphorous levels are currently being built up, it is now necessary to let this trial run in order to determine the optimum levels of both these minerals in the soil. This information will answer many questions on the fertilisation of soybeans.

Pesticide trial

The pesticide trial was planted with a better understanding than last season of the nature of the pest spectrum in the soil. Specific pesticides were used alone or in combination in order to test their efficacy on pests and their effect on the yield of soybeans. Unfortunately, most of the problem (or hard-to-manage) pests such as edible bulbs (a few), morning glory (none) and wandering Jew (none) did not occur and therefore certain pesti­cides could not be tested to their full potential. There were no significant differences in the yield between the different treatments.

It is a given that pests in soybeans should be controlled, whether there is good yield or not. Pests lead to many secondary problems such as harvesting problems, multiplication of the pest population and the difficulty of controlling these in the following seasons.

5.3

Effect of SO₂ and the interaction of growth, physiology and biochemistry of soybean (Glycine Max), studied in an OTC system

Mr JM Berner, S Lindeque, Dr PDR van Heerden and Prof GHJ Kruger

The effect of SO₂ fumigation on two soybean cultivars, LS 6164 and PAN 1666, were investigated. The aim of this study was 1) to establish exposure-response relationship for SO₂ in soybean and to determine the critical levels leading to yield reduction under South African growth conditions, and 2) to determine the extent of genotypic variation in tolerance against these air pollutants in the South African soybean gene pool.

The dry weight of the roots of LS 6164 fumigated with 25 ppb increased by 42%. After a period of 70 days of fumigation, the 75 and 150 ppb treatments resulted in a 25 and 26% decrease in root biomass. PAN 1666 showed decreases in root biomass at all three levels of SO₂ of between 50-60%. The number of nodules of both cultivars decreased as the levels of SO₂ increased. Decreases for LS 6164 were between 8 and 33%. For PAN 1666 the 150 ppb treatment resulted in an 80% decline in nodules while the 25 and 75 ppb treatments showed declines of 21 and 24%. The photosynthetic performance index (PIabs) measured after 4 weeks indicated that the 150 ppb treatment was 33% lower compared to the control and the 75 ppb treatment was only 6% lower. There were no differences between the control and the 25 ppb treatment. However, with the PAN 1666 cultivar the PIabs of the 75 and 150 ppb treatments dropped by 37 and 28% respectively. From these results it appears that the PAN 1666 is more susceptible to SO₂ pollution especially when levels reach 150 ppb.

5.4

Phenotypic markers for nodulation capacity of soybean cultivars

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

Fertilisation

Nitrogen fixation in nodules offers an important advantage in that soybean nodules fix the nitrogen required for plant growth resulting in the production of high-protein seeds. The life of a nodule is, however, remarkably short (11-13 weeks). Nitrogen fixation declines rapidly as the nodules age, and has normally almost ceased by the time pod-filling starts. Soil drying and high temperature stress can cut short the nodule lifespan affecting seed production, crop quality and yield. The objective of the project is therefore to identify easily measurable physiological and biochemical markers which might be applicable in a soybean breeding program to allow soybean cultivars to be selected for growth in the drier areas of South Africa.

Research is carried out at the University of Pretoria in collaboration with North West University and the 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.

Research activities have focused on optimisation by conducting experiments using special glass-fibre wigs located at different vertical positions in potting medium to control water supply to plants. This method allows for precise control and standardisation of plant growth/soil water status. Experiments have shown that drought stress reduces stomatal conductance and net photosynthetic assimilation rate of the tested cultivars Prima, A-54 and Jackson. However, reduced stomatal conductance and net photosynthetic assimilation rate were immediate and the decline was high in A-54 when compared to Jackson and Prima. The quantum yield of photosynthetic electron transport efficiency measurement was less affected during the course of stress period but was affected at the end of the drought treatment in all cultivars. Despite the fact that comparable effects of drought stress in nodule parameters in A-54 and Jackson were observed, the decrease in nodule number and fresh weight were higher in A-54 than in Jackson. This decrease in nodules, fresh weight and number was not related to nitrogenase activity in Jackson. The nitrogenase activity of A-54 under well-watered conditions was higher than in Jackson. However, at the onset of the drought period, activity declined which was not found in Jackson. The drought-induced ageing effect might be less severe in Jackson than in A-54. Overall, Prima was superior for most parameters with A-54 the most sensitive. In our experiments so far carried out, Jackson showed under drought higher stomatal conductance, net photosynthetic assimilation rate and root biomass when compared to A-54.

5.5

Establishment of an early warning system for soybean rust

Dr M Craven and Mr TM Ramusi

Due to the destructive nature of soybean rust and the apparent lack of adequate resistance to the disease an effective disease-management system is required until such time that resistant cultivars become available. Disease scouting to detect the presence of the pathogen as early as possible is an important factor in an effective management programme. It has been used in the USA with great success.

Since April 2005, ARC-GCI together with the Protein Research Foundation (PRF) and other co-workers such as PANNAR, the KwaZulu-Natal provincial department of agriculture and a number of soybean producers have attempted to create a soybean-rust-early-warning system that is based on trap crop trials, also referred to as indicator plots. During 2007/08, 10 localities were identified, representative of the major soybean production areas in the Free State, KwaZulu-Natal, Mpumalanga and Gauteng provinces. During 2007/08 rust was first detected at Greytown (28/01/08) followed by Vryheid (01/02/2008) and Normandien (05/03/08). Kestell, Kinross, Greylingstad and Potchefstroom remained free of rust during this season.

5.6

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

Prof NW McLaren and Me C Botha

A study to identify key driving variables in rust epidemics and to relate these to disease severity and concomitant yield and quality losses has been completed and is being prepared for publication.

Sequentially planted field trials over a five year period included 90 and 45 cm row spacings and three varieties that differed in maturity group i.e. PAN494, LS666 and SCS1. Plots were duplicated to include applications of fluzalizole/ carbendazim against which to measure yield and quality losses in unsprayed plots. At maturity, grain yield, 1000 kernel mass and oil and protein concentrations were determined. No relationship between first detection and planting date could be determined. Significant, negative relationships between Relative Life Time at first detection and Area Under the Disease Progress Curve (AUDPC) (R²=-0.53, -0.71, -0.52, in PAN494, LS666 and SCS1 respectively) were recorded. Similarly, the rate of rust development subsequent to onset differed depending on prevailing conditions and this criterion was positively related to AUDPC (R²=0.64, 0.59, 0.53, in the three genotypes respectively). Pooling of these criteria into a multiple regression model increased the relationship to R²=0.82, 0.81, and 0.67 in the short, medium and long season varieties, respectively indicating that these are major driving variables in rust epidemics. Since rust occurred in the fungicide sprayed plots, the AUDPC x yield loss analysis was performed by relating the difference in AUDPC in corre­sponding fungicide and untreated plots with the difference in yield loss percentage. Regression relationships of R²=0.72, 0.85 and 0.77 in PAN494, LS666 and SCS1 between AUDPC and yield loss were recorded. In addition, the ratios of AUDPC in the respective sprayed and unsprayed plots were used to determine the spray efficiency in each plot and this ranged from 16.4 to 93.9%. Application of the respective yield loss models to rust development in fungicide sprayed plots allowed the calculation of yield losses associated with fungicide treated plots and these ranged from 4.0 to 64.8% indicating that despite fungicide application, some degree of yield loss can be expected depending on spray efficiency. No significant relationship between disease development and oil or protein content was recorded.

This study has indicated that selection of genotypes or promoting conditions that delay the onset of disease and reduce the rate of disease devel­opment can have a significant effect on the severity of rust epidemics. Despite fungicide applications, losses of varying magnitude may still occur and careful attention needs to be given to spray efficiencies.

A collection of isolates of P. pachyrhizi has been initiated to determine races of the pathogen. These will be analyzed using AFLP analysis as well as a differential set of lines. Isolates will also be used to determine differential infection on a set of selected lines.

5.7

The use of silicon to control soybean rust (phakopsora pachyrhizi)

Dr Dael Visser and Dr Pat Caldwell

Aim to determine the effect of root application of silicon for the control of soybean rust.

Materials and Methods

For pot trials, soybean seed (LS 6161) will be planted in pots and placed in a greenhouse. Si (0, 50, 100, 250, 500, 1000, 1500 and 2000mg/l) will be applied by drenching pots for 6 weeks. Plants will be naturally infected in the glasshouse with urediospores during the summer months. There­after plants will be rated for percentage disease and the AUDPC calculated. For field trials soybean seed (LS 6161) were planted at Baynesfield (2007/2008) and Cedara (2008/2009) in 3 x 1.75 m plots. Plants were treated with Si at 0, 50, 100, 250, 500, 1000, 1500 and 2000 mg/l at Baynes­field and 500, 1000 and 2000 mg/l (2 and 4 times applications), a slow release Si fertilizer spread throughout the rows and broadcast over the plot. A Punch C and untreated control was also included. Soil samples were taken before Si was applied and after the first and last Si application and analyzed for Si concentration. Plants were rated for percentage disease and the AUDPC calculated. Yield was also determined.

Status

Field trials at Baynesfield in the 2007/2008 season have been conducted and analyzed. The 2007/2008 soil results from Baynesfield show no significant difference in Si after the first Si application. After the last Si application, the 2000 mg/l showed a significantly higher concentration of Si than all other treatments. The untreated control had the highest AUDPC, with 100, 250, 500 and 1000 mg/l being significantly lower. Plants treated with Punch C has had significant higher yields. No applications of Si resulted in increased yield but Punch C had the significantly highest yield. Possibly the first application of Si was not applied early enough and a third application should have been applied. This trial was repeated at Baynesfield in the 2008-2009 season but plants were not infected with soybean rust and so the trial was not harvested. A similar trial was run at Cedara in 2008-2009 and results are currently being analysed. Pot trials will be run in November 2009.

5.8

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 worldwide and in South Africa. Research has been done on the epidemiological characteristics of SBR, resulting in an understanding of the effects of temperature and moisture conditions on Phakopsora pachyrhizi urediospore 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 completed in order to gain insight into history, geographic extent and most importantly the epidemiological charac­teristics of the pathogen. An overview of different SBR models was done in order to better understand the advantages and short-comings of present models and to gain understanding of the current SBR models applicable to Southern Africa. It was found that little had been done to map the extent and frequency of risk from SBR in South Africa.

An algorithm (Eq. 1) was created based on first principles gathered from literature regarding the conditions needed for SBR. The algorithm was developed using 2001/2002 SBR outbreak data from fungicide trial data and weather data from an automatic weather station at Cedara for the same period.

I _T = m a x [ 0 , m i n ( 1 , i f ( T < 2 2 , ( T - 1 5 ) x 0 . 2 , 1 - ( T - 2 5 ) x 0 . 3 3 3 3 ) ) ] x D ( 1 )

This algorithm was tested against the 2002/2003 SBR outbreak trial data and weather data from the Cedara automatic weather station. The results were presented to a few SBR experts who agreed that it appeared accurate. Unfortunately no other data record could be found against which the model could be validated.

The algorithm was then programmed and run over a 50-year historical weather dataset, for the period 1950-1999 for relatively small climate zones throughout the country. The results of this were mapped using Arcview 3.3. In addition a time series graph was created for all the following soy­bean production areas, Bergville, Cedara, Vryheid, Delmas, Groblersdale, Marble Hall, Marken, Naboomspruit, Parys, Potchefstroom and Rustenburg.

An analysis of the result is currently being completed after which the forecasting framework will be developed. A hind casting chapter will be included, where the algorithm will be tested against weather data from an independent location where outbreaks of SBR have been recorded. A statistical analysis of the accuracy of the algorithm to predict conditions suitable for SBR outbreaks will be done.

5.9

Study of Inoculation and Disease Evaluation Techniques for Sclerotinia Stem Rot of Soybeans

Prof NW McLaren and Me C Botha

The objective of the study is to evaluate cultivars and alternate germplasm (imported lines) for environment x physiological stability and Scle­rotinia stem rot resistance.

Soybean lines and commercial cultivars were planted in Greytown (Pannar) for Sclerotinia stem rot evaluation. Five planting dates were used spaced approximately two weeks apart to ensure environmental diversity during the critical susceptible stage of host growth. At flowering, plants were inoculated with Sclerotinia sclerotiorum. Plants were scored at the end of the season for disease severity and host responses are being quan­tified using the regression methodology (McLaren & Craven, 2008). Results of screening trials to date indicate that most germplasm is susceptible to Sclerotinia stem rot even under less than optimal conditions for disease development although some cultivars, notably A 5601 RG, may be regarded as lower risk cultivars.

Field trials are also being used to model the relationship between weather variables and Sclerotinia stem rot severity. Temperature and leaf wet­ness duration studies indicate the optimum temperature for disease development at 20-25ºC although greenhouse trials have indicated that dis­ease may occur within the range of 18-28ºC providing that sufficient moisture is available. High humidity or free moisture is critical for disease development with optimal disease development at relative humidity (RH) >95% whereas disease incidence and severity decreased rapidly with decreased RH with little or no disease at RH<90%. Although the relationship between disease severity and weather variables could be modelled, more data points are required to optimize the relationship and ensure more reliable conclusions.

Samples of S. sclerotiorum were collected from areas where the disease was prominent and are being used to determine the genetic relationship between isolates from the different areas and their effect on pathogenicity. Laboratory evaluations of oxalic acid production by S. sclerotiorum isolates were carried out and a series of tests are being conducted to evaluate the importance of oxalic acid produced by S. sclerotiorum and their relevance to pathogenicity, while also focusing on the ability of plants to resist the activity of this acid i.e. oxalate oxidase is generally associated with a resistant response to the pathogen. Growth studies on the different isolates also suggest that there may be different ecotypes adapted to different temperature ranges. AFLP analysis will be used to quantify population diversity.

5.10

Super Soya Competition, KwaZulu-Natal

Mr C Havenga

The Super Soya Competition in KwaZulu-Natal is now in its nineteenth 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 as well as knowledge about its production has increased significantly over the past 19 years. However, the extent to which soybeans are 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 17 entries into the competition for S-KZN and 33 entries of N-KZN.

In S-KZN yields varied between 2.6 and 4.8 t/ha with an average of 4.0 t/ha which is an average of 1.0 ton/ha higher than the previous season.

Five fields (29%) were under irrigation with an average yield of 4.1 t/ha ranging between 3.6 and 4.8 t/ha. Yields under dryland ranged from 2.6 to 4.6 t/ha with an average of 3.7 t/ha.

The yields for this season's Super Soya Competition in N-KZN varied between 1.2 and 5.1 t/ha with an average yield of 2.8 t/ha. The average yield under irrigation was 3.7 t/ha which is considerably higher than the average dryland yield.

In general the protein content of the soybeans in S-KZN was good to very good, with an average protein content of 39.8% (dry matter basis) which varied between 37.6% and 41.2% (dry matter basis).

In N-KZN the protein content of the soybeans was satisfactory to very good. The average protein content was 39.9% (dry matter basis) which is higher than the previous seasons 38.8%. The protein content ranged from 36.5 to 46.7% (dry matter basis).

The oil content of the soybeans in S-KZN varied between 20.5 and 22.3% (dry matter basis), with an average of 21.3% which is a big improvement on the previous season. There was no significant difference between the average oil content of soybeans under irrigation (21.1%) and those under dryland (21.4%).

The average oil content of N-KZN soybeans was 20.8% (dry matter basis) and varied between 13.2% and 28.1%.

This competition also creates an opportunity to record and evaluate production practices and assists in identifying production problems which can, in turn, lead to research projects targeted at solving these problems.

5.11

Cultivar evaluation of oil and protein crops in the winter rainfall region

Mr PJA Lombard

National cultivar trials: Canola and Lupins

The Department of Agriculture Western Cape conducted several canola and lupin trials in the Swartland and Southern Cape during 2008. The trial sites at Riversdal and Tygerhoek were planted at the end of May. The rest of the Southern Cape trials were planted late in May and beginning of June due to a drought period in May. The germination in those trials was slow and not good.

Canola

The average yield of the trials in the Southern Cape were 1.75 and 1.78 ton/ha at Riversdal and Tygerhoek respectively. The average yield of the triazine tolerant (TT) and Clearfield (CL) cultivars were on average 19 and 12% lower than the conventional cultivars.

The yield from cultivar Garnet (2.8 ton/ha) was significantly better at Tygerhoek than from the other cultivars in the trial. Bravo TT (1.92 ton/ha) was the best TT and 45Y77 (2.04 ton/ha) the top producing Cl cultivar. The top producer at Riversdale was NPZ SR 10308 (2.34 ton/ha), followed by 44Y06 and Garnet. There was no significant difference between the top 5 cultivars in the trial.

The average yield in the Swartland differed between 2.39 ton/ha (Darling) and 2.67 kg/ha at Langgewens 2. The yield of the first planting at Lang­gewens 1 (14 May) was 8.8% lower than the second planting (27 May). This was the first time since 1999 that the yield of the second trial was higher than that of the first. The reason was the low average temperatures in September with good rainfall.

Garnet was the best producing cultivar in 3 of the 4 trials that were included in Canola Focus. The cultivar 44Y06 (3.15 ton/ha) was the top pro­ducer at Langgewens followed by 44C11 and Garnet. Bravo TT was the best TT and 45Y77 (2.04 ton/ha) the top producing Cl cultivar.

Lupins

The average trial yield in the Swartland was above average in 2008. The cultivar Mandelup (2.3 ton/ha) was, as in 2006 and 2007, the highest yield producer followed by Quilinock (2.2 ton/ha) and Tanjil (1.7 ton/ha). The highest yielding line was 2224 (2.35 ton/ha).

The new lines performed very well in the Southern Cape. The lines 42 (2.75 ton/ha) and 23 (2.79 ton/ha) gave a higher seed yield than Mandelup (2.7 ton/ha). The trials at Riversdale and Klipdale were planted on 24 April and 5 June and their respective average yields were 2.42 and 2.58 ton/ha.

5.12

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

Dr J Strauss, Dr MB Hardy and Mr W Langenhoven

Climatic conditions during 2008 were favourable for dry-land crop and pasture production at Langgewens. Above average rainfall was experienced over the growing season (April to October) and rainfall was well distributed. Temperatures remained mild and the hot dry periods that often occur during the growing season did not materialise. The average temperature in September was lower than that in August, which contributed to the high yields obtained in the canola and wheat crops.

Mean wheat yield over all systems was 3818 kg/ha. This was slightly lower than the previous season. As has been recorded over almost all sea­sons (2005 was an exception) there were clear differences in yield among the different crop sequences. Wheat yields following legume pasture and lupins were greater than wheat yields in all other systems.

Canola was harvested at the beginning of November. The yields obtained during harvesting ("straight") ranged from 1381 to 2901 kg/ha. The average crude protein (% CP) over all plots was 23% and the oil content averaged 34%.

The project is progressing according to the research proposal and protocols. Management and production data are regularly presented to the local farming community in popular publications and on occasions such as farmer's days. The information is also made available to technical advisors of the various Agri-businesses that operate in the area.

5.13

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

Dr MB Hardy, Dr JA Strauss and Mr W Langenhoven

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 economically sustainable land-use in the Southern Cape.

The 2008 year was the 7th 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.

All trial areas were planted and managed according to the planned protocols (including appropriate weed, disease and insect control measures).

Low and erratic rainfall occurred for much of the season throughout the southern Cape region. The Swellendam site experienced its worst rainfall season since the trial started in 2002.

Poor soil moisture before and after the start of the planting season limited the emergence, establishment and growth of crops at the Tygerhoek site, with the crops in the continuous cropping treatment experiencing less moisture stress than those in the other treatments. It appears that there was greater soil moisture availability in the continuous cropping treatments (their crop residues not being disturbed or removed by sheep) and the soils were not compacted (as occurs with sheep grazing).

Soil moisture availability remained favourable at the Riversdale site throughout the season resulting in excellent yields at the site.

Canola production

Canola (Spectrum) was planted at Riversdale following wheat. A seeding rate of 3.0 kg/ha was used and a total of 48 kg N/ha was applied to each plot (24 kg N/ha at planting and 24 kg N/ha top-dressing).

Jade was planted at Swellendam at 5 kg seed/ha and with 24 kg N/ha at planting. Canola (Thunder) was planted at Tygerhoek. A seeding rate of 3.8 kg/ha was used and a total of 54 kg N/ha was applied to each plot (24 kg N/ha at planting and 30 kg N/ha top-dressing).

Canola yields at Riversdale ranged from 1900 to 2100 kg/ha. These yields were approximately 50% of the yields obtained from wheat and Triticale (see below).

No canola seed was harvested at the Swellendam site because of the severity of the dry conditions at the site.

The average canola yield over all plots at Tygerhoek was 1016 kg/ha, which is acceptable in comparison to the expectations of the region but still only about 44% of the average yield for wheat at the same site. Estimated average plot yield was, however, about 100% greater than the harvested yield i.e. an estimated yield loss of 50% was observed. Note that we need to refine the sampling technique used to estimate potential yield, but over that past few seasons the yield loss estimates have been consistently high (in the region of 20 to 50% loss).

As has been reported for previous seasons the average oil content of seed from the Riversdale site, as determined at the SSK facility in Swellen­dam, was again higher (41% oil) than for duplicate samples analysed at the Elsenburg Laboratory (33% oil).

Lupin production

Lupins (Quillinock) were planted at the Riversdale and Tygerhoek sites at a seeding rate of 100 kg/ha. The Riversdale site produced the highest yield ranging from 2100 to 2350 kg/ha. At Tygerhoek yields were more variable ranging from 1100 to 1950 kg/ha. Soil physical limitations in one of the plots are thought to have contributed to the low yield recorded in that plot. Lupin yields were higher than expected considering the poor rainfall experienced at the Tygerhoek site.

5.14

The identification of soil parameters as indicators of sustainable dry-land crop production systems for the shale derived soils of the Western Cape: tillage practice, crop rotation, soil quality and crop production

Dr J Labuschagne

The aim of this project is to quantify the effects of tillage practice and crop sequence on soil physical and chemical properties, and soil biological activity with a view to gaining a better understanding of soil parameters that will promote sustainability in crop production systems on the shale derived soils of the Western Cape.

Tillage treatments and planting

At the Langgewens site the scarifying (tine) operations in the conventional and minimum tillage sub-plots were done on 12 February and 13 May respectively. At the Tygerhoek site the scarifying (tine) and plough operations were done on 29 January and 3 May. The loose seedbeds in the conventional treatments caused some difficulty in achieving uniform seeding depths.

The wheat monoculture crop sequence treatments at Tygerhoek had serious herbicide resistant annual ryegrass infestations. Since the trial's main objective is to determine the effects of tillage and crop sequence on the soil, the wheat zero till plots were killed off using a broad-spectrum herb­icide (glyphosate) after wheat ears had emerged but before the ryegrass seed heads could produce viable seed. It is assumed that any effect that a wheat crop would have on soil physical and chemical status would be expressed due to the wheat being present for the greatest part of the growing season.

Rainfall

Climatic conditions during 2008 were favourable for dry-land wheat production especially at the Langgewens site. September was however wetter and cooler than normal, at both sites. The Langgewens site experienced periods of waterlogged conditions early in the growing season causing dying off of lupin and canola plants leaving open areas ideal for weed infestation.

Crop production data

Seedling/plant densities

The mean seedling density for wheat at Langgewens was 111.8 seedlings/m² and later in the season mean ear counts of 278 ears/m² were re­corded. The lowest seedling density (92 plants/m²) was observed in the conventional till sub-plot of the LKCK rotation system. The plants com­pensated as 260 ears/m² were recorded on the same plots in October 2008. As a result of continuous high rainfall during the first part of the grow­ing season at Langgewens, lupin and canola sub-plots became waterlogged causing dying off of the crops mentioned. Weed infestation occurred in the bare spots and it was decided to use a non-selective herbicide to kill all growth (except one canola replicate) on these treatments.

Plant density for wheat was much lower at Tygerhoek at 61.2 plants/m², excluding the zero-till sub-plots where only 26.4 plants/m² were counted. Unfavourable seedbed and soil water conditions could be the reason for the low seedling counts. The zero-till plots were sprayed with a non-selective herbicide to minimise weed seed production on these plots.

At Tygerhoek a mean of 30.69 canola seedlings/m² were counted.

Grain yield

Wheat yields of 4.56 and 2.92 tons ha^-1 were recorded for Langgewens and Tygerhoek respectively. At Langgewens the highest yield of 4.96 t/ha recorded on the conventional plots was not significantly higher than the 4.90 t/ha harvested on the no-till sub-plots. At Langgewens the lowest yield of 3.99 t/ha was recorded on the zero-till plots. The minimum till canola plot yielded 2.06 t/ha (highest) and the lowest yield was recorded on the conventional till sub-plot (1.47 t/ha).

Wheat yields at Tygerhoek from the conventional, minimum and no-till plots were 3.08, 3.01 and 2.68 t/ha, respectively. Canola at Tygerhoek yielded 1.59 t/ha, with the conventional plots the highest at 1.64 t/ha. Minimum and no-till canola yielded 1.57 and 1.58 t/ha respectively.

Remarks

The project is progressing according to the research proposal and protocols. The severe ryegrass problem is being managed without compro­mis­ing the integrity of the treatments and their potential effects on the physical and chemical status of soils. Short-term effects of tillage system, crop sequence and site on crop yields should become more apparent over time.

5.15

Assessing canola seed losses during seed ripening and harvesting in the Western Cape Province

Dr J Strauss

The farming community of the Western Cape has raised questions about the difference between the potential yield of canola (as indicated by several seasons of canola cultivar trials) and the actual (or realised) yields achieved by producers and delivered to the silo. An average of 21 seeds on an area the size of a matchbox relates to 400 kg/ha, where the seed has a thousand kernel weight of 4g. Avoidable seed losses currently exac­erbate the perception that canola production is less profitable than wheat production.

The trial consisted of a farm survey and statistical trial that compared canola seed losses between the two types of harvesting methods, namely swathing and straight. In the first production season high losses in canola yield of between 31 and 43% were measured in the farm survey and between 24 and 28% in the statistical trial. In the farm survey, swathed canola showed the lowest losses, while the straight canola had a higher yield in the statistical trial. The biggest losses occurred during the harvesting process. The survey and statistical trial will be repeated in 2009.

5.16

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

Mr IFV Slabbert and Mr CL van Rooyen

A. Swartland canola competition 2008

Thirteen farmers entered the Canola Competition in the Swartland and information from all competitors could be used.

Although the first rains were somewhat late to ensure a good canola potential, favourable weather conditions persisted till the beginning of September. Thus the average yield of the participants was 1.34 t/ha, which was 0.24 ton less than the previous year (2007).

Mr Frikkie Russouw from Moorreesburg was the participant with the highest yield (1.7 t/ha) and Mr Michiel Smuts from Malmesbury recorded the highest gross margin (R3532/ha).

The runners up, Mr Daniel Rossouw (1,62 t/ha) and Mr Abrie Richter (R2601/ha) are both from Piketberg.

Canola was more profitable than wheat during 2008.

B. Southern Cape canola competition 2008

The competition started out with 22 respondents but the drought whittled the number down to 12 at the end. The climate was not conducive to early sowing and the bulk of the rain only fell from mid-June. The shorter growing season cultivars performed better than anticipated because of the late rains. Early sown canola took 75-90 d before emerging. The pre-emergence weed control failed due to lack of rain in the critical periods. The yields were better in the western parts because the canola could make better use of the late rains.

Some of the triazine-tolerant cultivars have been around for a long time and needs to be replaced. The availability of the best cultivars were lack­ing and producers sowed more farm seed than normal. In the Clearfield types the situation was even worse in the 2009 planting season as seed for certain areas were to late for early sowing. Canola looks to have reached a point of stagnation in hectares planted and only an increase in price will contribute to further expansion of canola in the Southern Cape.

The winner of 2008 was Mr Andrew Beukes of the farm Vrede with a yield of 2.4 t/ha and in 2nd place, the Blom brothers of Bredasdorp.

5.17

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

Prof GA Agenbag

Field experiments were again conducted at Langgewens-, Elsenburg-, Welgevallen- and Roodebloem Experimental Farms during 2008 to evaluate the response of canola to increasing (0-120 kg N ha^-1) N application rates. Unfortunately no results were obtained at Welgevallen because the experiment had to be abandoned due to very wet and flooded conditions. Soil and plant analyses were conducted at Langgewens and Elsenburg.

At both localities, sub-optimal soil pH (<5.0 pHKCl) and S (<6.0 mg kg^-1) content occurred, while B was shown also to be low (<0.2 mg kg^-1) at Langgewens. Plant analysis showed sub-optimal S contents (<0.5%) at both localities. In comparison to general norms for Na, Al and Fe of 300 - 5000 mg kg^-1, 200 mg kg^-1 and 150-200 mg kg^-1 respectively, high contents were found at both localities.

As was the case during previous years, yield responses varied considerably between localities. In contrast to Langgewens and Elsenburg, low rain­fall and very dry conditions were experienced at Roodebloem for most of the growing season. At Roodebloem, grain yields of 582 kg ha^-1 were obtained without any nitrogen fertiliser, compared to yields of 890 kg ha^-1 with 120 kg N ha^-1. There were no statistical differences in yield with 90 or 120 kg N ha^-1. Yields at Elsenburg varied between 1435 with no nitrogen application and 2235 kg ha^-1 with an application of 120 kg N ha^-1. In comparison, a yield of 2649 kg ha^-1 was obtained at Langgewens without any N fertiliser. At this locality the highest grain yield of 2961 kg ha^-1 was produced when 60 kg N ha^-1 was broadcast at planting and a further 60 kg N-1 applied as topdressing 60 d after planting. At both Langge­wens and Elsenburg, no statistical increas in yield was achieved when N application rates increased from 60 to 120 kg N ha^-1, on condition that N application at lower rates were band placed at planting. Due to high rainfall and, presumably, large leaching losses, similar yields were obtained when 30 kg N ha^-1 was band placed at planting and 30 kg N ha^-1 applied as topdressing 30 d after planting, compared to a total broadcast application of 120 kg N ha^-1.

The large differences in grain yield produced without any N application again emphasises the need to consider the N mineralization potential of the soil when N fertiliser requirements are determined.

5.18

Optimal soil tillage methods to be used in a wheat / canola / wheat / lupin crop rotation system in the Swartland production area

Prof GA Agenbag

In this experiment on Langgewens Experimental Farm, the effect of soil tillage methods on soil fertility and crop yields has been studied for the past 30 years and that of two crop rotation systems for more than 15 years.

During 2008, an average grain yield of 2474 kg ha^-1 was produced when wheat was planted in a four year crop rotation with lupins (2007) and canola. In comparison to this, wheat planted in a monoculture system yielded only 2062 kg ha^-1 on average. From these results it became clear that although wheat is planted only every second year in the crop rotation systems, grain yields are increased by 20%. In the crop rotation system where wheat was planted after lupins, the highest yield of 2567 kg ha^-1 was produced with only 20 kg N ha^-1 band placed at planting. In the mono­culture system, the highest wheat yield of 2157 kg ha^-1 required an application of 100 kg N ha^-1 which was divided between 20 kg N band placed at planting and two topdressings at 30 and 60 d after planting. In the monoculture system high yields were achieved only when seedbed preparation was done by intensive, time consuming and costly mouldboard ploughing, but in the crop rotation system, no-tillage planting resulted in similar yields compared to mouldboard ploughing. This tendency can to a very large extent be ascribed to the more effective chemical wheat control which is only possible when crop rotation is practiced.

These results showed that when sound crop rotation systems that include crops such as canola and lupins are used, increases in wheat yield can be obtained with reduced fertilizer and soil tillage inputs.

5.19

Management of herbicide resistance in the Western Cape

Dr PJ Pieterse

The testing for herbicide resistance and record keeping progressed satisfactorily. Only 43 of 47 seed samples harvested in 2008 were tested due to insufficient seed being received or to poor quality seed. Currently only two plant samples for the 2009 season are tested.

Counts of weed seedlings in the rotational cropping and soil tillage trials on Langgewens experimental farm were carried out in June and August 2008. Samples of wheat and weeds in these two trials were collected in October 2008 to determine biomass. Results showed that from a ryegrass control point of view, mouldboard ploughing was the superior tillage method and in terms of crop rotation, annual pastures such as medics and clover were superior to any other cropping system to reduce ryegrass numbers in fields on Langgewens. Seed from weeds in these plots was har­vested and the resistance profile will be investigated later in this season.

The glasshouse trial to investigate the competitive ability of canola and wheat cultivars was completed in 2008 and a trial was sown on Welge­vallen experimental farm where four wheat cultivars with varying degrees of competitiveness will be evaluated under different weed pressure levels. The first data have been collected but not yet analysed. The canola trial will be planted in 2010.

5.20

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

Dr SC Lamprecht

In this study several aspects with regard to the management of Rhizoctonia diseases of canola and lupin such as disease resistance, seed treat­ment and biological control using binucleate Rhizoctonia anastomosis groups (AGs) were investigated under glasshouse conditions. The resist­ance of eight canola and eight lupin cultivars/selections to damping-off and hypocotyl/root rot caused by the multinucleate Rhizoctonia AG-2-1, 2-2, 4 and 11 was evaluated. The canola cultivars included were the conventional cultivars, Comet, Outback, Spectrum and 44C11, the clearfield cultivars Rocket and 44C73 and the triazine tolerant cultivars Thunder and Tornado. Cultivars/selections of three lupin spp. were included viz. Cedara 6150 and E16 (L. albus), Mandelup, Tanjil, Quilinock and E42 (L. angustifolius) and E80.1.1.2 and E82.1.1 (L. luteus). Three isolates of each multinucleate AG were tested and trials were conducted in a glasshouse. The AGs most virulent on canola were AG-2-1 and 4. Canola cultivars did not differ significantly in their susceptibility to AG-2-1. However, Spectrum and 44C11 were more resistant to AG-4 than the other cultivars. On lupin, AG-2-2 and 4 were most virulent and the cultivar Cedara 6150 and selection E16 were most resistant to AG-2-2, and Cedara 6150, E16, Man­delup and Quilinock were also more resistant to AG-4 than the other cultivars/selections. The L. luteus selections, E80.1.1.2 and E82.1.1 were most susceptible to AG2-2, 4 and 11. Seed treatment with Cruiser OSD and SA-combination significantly increased survival of canola (cvs Muster, Rocket and Thunder) and lupin (cultivar/selection Cedara 6150, E82.1.1 and Mandelup) seedlings, decreased hypocotyl/root rot and improved the percentage healthy seedlings, with SA-combination being significantly more effective than Cruiser OSD. The binucleate Rhizoctonia AGs (A, Bo, K and I) significantly increased the survival of lupin seedlings inoculated with AG-2-2 and 4. A low inoculum ratio of the binucleate AGs [1:1 (multi­nucleate AG :binucleate AG)] was less effective than a high inoculum ratio [1:10 (multinucleate AG :binucleate AG)]. It appeared that AG-Bo and I were more effective in protecting lupin seedlings against multinucleate AGs than the other binucleate AGs. The binucleate AGs did improve sur­vival of canola inoculated with AG-2-1, 2-2 and 4, but it was not statistically significant. This is the first report of the potential of binucleate AGs to protect lupin seedlings against multinucleate AGs. Results obtained in this study can contribute significantly to management strategies that can be employed against Rhizoctonia diseases of canola and lupin in the Western Cape province of South Africa.

5.21

Insects and other pests of canola

Dr G Tribe

The most promising treatments resulting from the 2008 trial were slug pellets containing both metaldehyde and carbaryl broadcast on the surface at planting, followed by treatment of seeds with Cruizer. A combination of the two treatments promises to give satisfactory protection.

Isopods once again were shown from the trial data to be the most important organism responsible for the loss of canola seedlings, where treat­ments which caused the death of isopods also had the highest number of surviving seedlings, and vice versa. The slug population in the 2008 season was relatively low with the result that no definite conclusions could be drawn with respect to the different treatments on them. Slugs were classified into four size cohorts and 89.3% of the slugs occurred in the 'small' and 'medium' cohorts which are a result of eggs laid in autumn and winter of the previous season. This indicates that it is those slugs that successfully over-summered from the previous season which are to blame for the losses of seedlings.

Three species of exotic slugs were identified within the canola fields but one species, Deroceras panormitanum, comprised ±85% of all the slugs. The size of this slug when elongated would allow it to enter the tunnels of earthworms and this may be the reason why it predominated. Soil excavated under traps at the end of the season revealed no slugs or isopods, and indications are that both species do not occur as deep in soil as previously thought.

Trials comparing four slug pellet formulations showed that all were extremely effective in attracting and killing slugs, but that only those which contained a carbamate insecticide killed isopods as well. However, carbaryl also killed millipedes and other insects but did not kill crickets, neither of which have been shown to injure canola seedlings and therefore could be considered beneficial in recycling stubble.

The marking and recapture of isopods revealed that populations within a canola field may vary between 51 and 170/m². Of the few marked isopods that were recaptured, most had remained under the traps where they had been marked, and only two were found to have migrated four metres to adjacent traps. The vast majority of marked isopods simply disappeared.

Plants containing the cabbage stem weevil, Ceutorhynchus pallidactylus, had on average 17.8 larvae per plant where most were aggregated within a small area of the plant, usually near unconsumed pith. The inner pith of the entire plant was often totally consumed making translocation of nutrients in the plant impossible.

5.22

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

Dr M Ciacciarello

The aim of this project is to determine whether NIR technology could be used to determine the protein, lipid and digestible amino acid content of feed ingredients, mixed feeds and carcass samples rapidly, accurately and inexpensively. A wide range of feeds and carcass samples has been analysed over the years in Animal and Poultry Science at the University of the KwaZulu-Natal for protein, energy, total and digestible amino acids, and these were used as the basis of the evaluations. If this technology could be used successfully, this would reduce considerably the cost and time involved in such analyses.

Calibration curves for protein and total amino acid contents of fishmeal, lupins, maize, full-fat soya and soya oilcake have thus far been performed, with promising results. Further calibrations for the digestible amino acid contents of the above ingredients are currently being conducted, whilst carcass samples will also have been analysed by the end of this year.

5.23

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

Two broiler trials were conducted during the year, making use of a total of 7960 Cobb and Ross broiler males, which were reared to 35 d on various photoperiods between 2 and 21 h or under continuous illumination. There were no significant influences on either feed intake or growth for photo­periods less than 6 h, but significant depressions in intakes and growth for daylengths shorter than this. Feed conversion efficiency was maxi­mised by 12-h photoperiods, with decreases in efficiency above and below 12 h. Mortality was unaffected by photoperiods of less than 12 h, but increased proportionately with longer photoperiods. The incidence of Sudden Death Syndrome (SDS) decreased between 2 and 10 h, but was positively related to photoperiod above 12 h. The European Efficiency Factor was curvilinearly related to photoperiod; with the highest efficiency occurring at 12 h.

By 5 d, birds given less than 15 h illumination had learned to eat in the dark at a rate inversely proportional to photoperiod; further increases in nocturnal feeding rates occurred between 5 and 20 d in birds on these shorter photoperiods. The hourly rate of nocturnal feeding was consistently lower than diurnal feeding, irrespective of photoperiod.

There are thus many advantages in using daylengths shorter than those conventionally used in South Africa at present: bird welfare is improved, with a lower incidence of SDS and mortality and with bone strength being increased; electricity usage would be decreased by 1.6 mWh p.a., and because the efficiency of utilisation of feed is improved, annual feed protein usage in South Africa could be decreased by as much as R400m.

5.24

The evaluation of raw soybeans in diets of growing ostriches

Prof TS Brand

Due to the presence of trypsin inhibitor, raw soybeans are either excluded from or included at a very low level in diets fed to farm animals. The effect of the inclusion of raw soybeans in diets of ostriches is however unknown.

Two experiments are therefore currently being executed at Elsenburg to determine the effect of including raw soybeans in diets on the perfor­mance of ostriches. In the first experiment diets with varying amounts of raw soybeans were presented to birds on a free choice basis to deter­mine whether the birds would be prepared to eat the diet containing raw soybeans. In the second experiment birds will be fed diets with an in­creasing percentage of raw soybeans and a decreasing percentage of heat-treated soybeans. Production (feed intake, growth and feed conversion ratio) will be monitored. This experiment will enable soybean producers that cannot heat-treat the beans to use this valuable protein source unprocessed.

5.25

Income and cost budgets

Mr JSG Joubert

During the report year, the services of a contractor were utilised to assist with the compilation of economical and technical/biological information used in the calculation of income- and expenditure budgets of crops in which the PRF have an interest. In the Winter rainfall region, information for the calculation of budgets were obtained from Agri-organisations for the southern Cape regions, whilst group discussion forums were utilised in the Porterville and Malmesbury areas for obtaining information for the Swartland. Information from the Moorreesburg area was once again obtained from the Cooperative. Given the new policy adopted by the Moorreesburg Cooperative, the group discussion method will probably be used in future to obtain this information.

With regard to soya, maize and sunflowers in the summer rainfall regions, there was also a deviation from the standardised work method. All the budgets in the Mpumalanga area were re-done by assimilating new information from producers via the group discussion method.

The areas for which information was assimilated for Mpumalanga were also adapted. Regarding KwaZulu-Natal, the numbers from the Department of Agriculture were not used and information was assimilated through the group discussion technique. Information for the Free-State was obtained from the "Vrystaat Koöperasie Beperk (VKB)", as in the past.

Sunflower as a crop was included in the budgets for the first time in areas where it is deemed to be an important competitive crop.

Income- and cost estimations of soya, maize and sunflower (not for all areas listed below) were compiled in the following summer rainfall areas:

In the Winter rainfall area of the Western Cape province, income- and expenditure estimates of canola, wheat and lupins (last mentioned not compiled for Malmesbury and Porterville) were compiled for:

The income- and expenditure estimations contain information on a per hectare basis of gross incomes, directly allocatable variable costs and gross margins. Managers and prospective producers can utilise this information to compile or adapt their own budgets in order to be empowered to make informed decisions. The information reflects typical situations under normal climatic conditions and is not an average.

5.26

Models: APR- and Nieuwoudt/McGuigan Model

During the 2008 report year, the APR- and Nieuwoudt/McGuigan models were once again utilised to make projections of the protein requirements for animal consumption for 2010 and 2015. The APR-model was used to project the total protein requirement in the basis year 2006/2007 by animal species after which the Nieuwoudt/McGuigan model was used to make projections of protein requirements by animal species as a whole for 2010 and 2015.

For the purposes of this exercise, the PRF was interested only in the projections of oilcake requirements for the abovementioned periods and the other protein sources, such as fishmeal, gluten, poultry by-product meal etc., were disregarded.

The projected oilcake consumption for 2010 was estimated to be 1 918 763 ton and for 2020, 2 681 414 ton.

The 2007 research report mentioned that the PRF was interested in investigating what could be delivered by the BFAP model of the University of Pretoria. This model analyses international and local markets through base-line projections and the development of scenarios based on possible market- and policy variables. The impact of changes is then translated/relayed to local markets regarding farming profitability and sustainability. When scenarios are developed for organisations such as the PRF, the following variables, amongst others, are considered:

• Urbanisation and population growth;
• Increasing concentration of protein markets;
• Availability of research capacity, knowledge and experience;
• Efficiency in protein consumption by animals;
• Profitability and risk of soya production versus, for example, maize production;
• Tariff policies;
• Economic growth both internationally and in the RSA;
• Oil expressing capacity;
• Climatic changes;
• Policy directions the Government might take; and
• Energy consumption and types of energy.

The PRF has reached an agreement with the University of Pretoria to develop ready-made scenarios for the PRF that will generate information leading ultimately to more accurate predictions with regard to the protein market for animal consumption. Initially the idea was to integrate the BFAP-model with the existing APR- and Nieuwoudt/McGuigan models in order to make more accurate predictions. Unfortunately, this does not seem to be a practical possibility and therefore the results from the BFAP-model will be used in the other two models for the annual projections of protein requirements.

PRF Board members, as well as other invited experts from various disciplines, attended a workshop in 2008 with the BFAP personnel and developed concept scenarios. Unfortunately, the PRF continues to have issues with certain assumptions that were used and these are currently receiving attention, prior to the finalisation of the scenarios which will hopefully be completed for the next report year.

5.27

PRF website

The website remains one of the most powerful and important vehicles of communication of the PRF. It is for this reason that the PRF is continu­ously marketing the website to a wider audience and striving to make it increasingly user friendly.

During the report year, the website was evaluated by peers in order to determine whether large scale changes were required. The conclusion of the peer evaluators was that the website design was neat and functional and well programmed. A few cosmetic suggestions were made, such as font size, colours, etc, of which some were implemented.

During the report year, much attention was again paid to the database which is now well established and to which new research information is entered as it is received. We have also started loading historical information into the database. News snippets remain a popular item; it is also now managed by removing old news on an ongoing basis and also adding news items from work groups. Links with other websites remain a high prior­ity and during 2008 we created reciprocal links with an Australian website following a visit to that country by one of our Board members. This specific website has since become an important source of reference for visitors to the PRF website.

The popularity of the PRF website can be measured by the number of unique visitors over time which numbers are as follows:

Search engines remain the most important source of reference and currently around 70% of all visitors are referred via search engines. One of the most important search engines is www.aardvark.co.za.