Typescript Includes bibliographical references

This research focuses on developing a crop decision planning model to help farmers make decisions for an upcoming crop year. The decisions consist of which crops to plant, the amount of land to allocate to each crop, when to grow, when to harvest, and when to sell. The objective is to maximize the overall profit subject to available resources under yield and price uncertainties. To help achieve this objective, we develop yield and price forecasting models to estimate the probable outcomes of these uncertain factors. The output from both forecasting models are incorporated into the crop decision planning model which enables the farmers to investigate and analyze the possible scenarios and eventually determine the appropriate decisions for each situation. This dissertation has three major components, yield forecasting, price forecasting, and crop decision planning. For yield forecasting, we propose a crop-weather regression model under a semiparametric framework. We use temperature and rainfall information during the cropping season and a GDP macroeconomic indicator as predictors in the model. We apply a functional principal components analysis technique to reduce the dimensionality of the model and to extract meaningful information from the predictors. We compare the prediction results from our model with a series of other yield forecasting models. For price forecasting, we develop a futures-based model which predicts a cash price from futures price and commodity basis. We focus on forecasting the commodity basis rather than the cash price because of the availability of futures price information and the low uncertainty of the commodity basis. We adopt a model-based approach to estimate the density function of the commodity basis distribution, which is further used to estimate the confidence interval of the commodity basis and the cash price. Finally, for crop decision planning, we propose a stochastic linear programming model, which provides the optimal policy. We also develop three heuristic models that generate a feasible solution at a low computational cost. We investigate the robustness of the proposed models to the uncertainties and prior probabilities. A numerical study of the developed approaches is performed for a case of a representative farmer who grows corn and soybean in Illinois.

This thesis presents results from two separate studies. First, the impact of bearer length on yield components within the canopy was investigated in season 2005/06, on a commercially-managed, mechanically-pruned vineyard of Vitis vinifera L. Cabernet Sauvignon in Coonawarra, South Australia. Pruning resulted in the retention of bearers with 1-7 nodes, with the weighted average bearer length being two nodes for the canopy. As bearers of one to five nodes in length were the most common, these were studied. Yield components (on a per shoot basis) were analysed according to the node position on the bearer at which the shoot arose. Both budburst and inflorescence number per node were highest at the distal node positions on each length bearer, even if the nodes were at the same positions from the base of the bearer and would normally be expected to have similar fertility. Budburst appeared to act by modifying inflorescence number per node based on the relative location of each node from the apex of the bearer. Shoots that arose from the most distal node positions had the highest flower number per inflorescence and berry number per bunch. Flower number per inflorescence was significantly higher on two-inflorescence shoots than single-inflorescence shoots. The relationship between bunch size and node position, unlike that between inflorescence number and node position, was dependent on bearer length. The relative size of the inflorescence appeared to be affected more so by the node pOSition at which the shoot occurred on the bearer, as opposed to the actual node position on the shoot at which the inflorescence occurred. There was a positive, non-linear relationship between average fruit yield per bearer and bearer length. Although yield was highest from the bearer with the highest node number (five nodes), there was no significant difference in yield per bearer for the bearers of three to five nodes in length. If average bearer length was increased from two to three nodes, the potential yield gain per bearer is estimated at 38 per cent. The second study presents results of correlations between bunch number and components of bunch weight (flower number and berry number) to investigate co-development of bunch number and bunch size. These data were collected from 4 vineyards in the Limestone Coast Zone of South Australia from Vilis vinifera L. Chardonnay, Shiraz and Cabemet Sauvignon during seasons 2002/03 to 2006/07. The significant correlations found between fertility and both bunch weight and flower number per inflorescence suggest that the same factors that affect bunch number in a particular season will also affect bunch size. When inflorescence primordia were initiated and differentiated under cool conditions, actual bunches per node and flowers per inflorescence were low. Differences in climate between the vineyard sites were found to be minimal and therefore did not strongly affect the magnitude of the yield components at the vineyard sites. Cultural practices at each vineyard site were sufficiently variable to affect fertility levels. Genotype is thought to determine the range of flowers per inflorescence that a variety can potentially carry, whereas actual flower number per inflorescence is thought to be determined by inflorescence primordium initiation and differentiation temperatures, as well as temperatures during budburst. Despite significant correlations between flower number per inflorescence and berry number per bunch, flower number per inflorescence preflowering for Cabemet Sauvignon, Shiraz and Chardonnay is inversely related to actual percentage fruit set. This is possibly a survival mechanism for the grapevine as it allows the vine to maximise yield each season without detriment to its longevity. Bunches per vine accounted for the majority of the seasonal variation in yield per vine. Fluctuations in bunch number per vine (and therefore yield) are likely to be reduced by varying the number of nodes retained per vine according to the relative fruitfulness per node present pre-pruning. This practice is therefore likely to result in the seasonal variation of berries per bunch becoming a stronger driver of yield. The commercial impacts of these studies are two-fold. Data presented will assist growers to understand the reasons for which their pruning regimes are affecting yield production and how these pruning regimes may be modified to achieve a target yield-particularly when growers are faced with seasons of low predicted fertility. In addition, data presented will allow growers to improve their crop forecasting accuracy, with a greater understanding of the link between bunch number and bunch size. In the current situation of oversupply in the wine industry, wineries are adopting a tough stance towards growers over-delivering on their grape contracts. Therefore, any assistance that can be provided to growers on improving accuracy of yield estimates will be beneficial both to the grower and winery. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1352744 / Thesis (M.Ag.Sc.) -- University of Adelaide, School of Agriculture, Food and Wine, 2009

This thesis presents results from two separate studies. First, the impact of bearer length on yield components within the canopy was investigated in season 2005/06, on a commercially-managed, mechanically-pruned vineyard of Vitis vinifera L. Cabernet Sauvignon in Coonawarra, South Australia. Pruning resulted in the retention of bearers with 1-7 nodes, with the weighted average bearer length being two nodes for the canopy. As bearers of one to five nodes in length were the most common, these were studied. Yield components (on a per shoot basis) were analysed according to the node position on the bearer at which the shoot arose. Both budburst and inflorescence number per node were highest at the distal node positions on each length bearer, even if the nodes were at the same positions from the base of the bearer and would normally be expected to have similar fertility. Budburst appeared to act by modifying inflorescence number per node based on the relative location of each node from the apex of the bearer. Shoots that arose from the most distal node positions had the highest flower number per inflorescence and berry number per bunch. Flower number per inflorescence was significantly higher on two-inflorescence shoots than single-inflorescence shoots. The relationship between bunch size and node position, unlike that between inflorescence number and node position, was dependent on bearer length. The relative size of the inflorescence appeared to be affected more so by the node pOSition at which the shoot occurred on the bearer, as opposed to the actual node position on the shoot at which the inflorescence occurred. There was a positive, non-linear relationship between average fruit yield per bearer and bearer length. Although yield was highest from the bearer with the highest node number (five nodes), there was no significant difference in yield per bearer for the bearers of three to five nodes in length. If average bearer length was increased from two to three nodes, the potential yield gain per bearer is estimated at 38 per cent. The second study presents results of correlations between bunch number and components of bunch weight (flower number and berry number) to investigate co-development of bunch number and bunch size. These data were collected from 4 vineyards in the Limestone Coast Zone of South Australia from Vilis vinifera L. Chardonnay, Shiraz and Cabemet Sauvignon during seasons 2002/03 to 2006/07. The significant correlations found between fertility and both bunch weight and flower number per inflorescence suggest that the same factors that affect bunch number in a particular season will also affect bunch size. When inflorescence primordia were initiated and differentiated under cool conditions, actual bunches per node and flowers per inflorescence were low. Differences in climate between the vineyard sites were found to be minimal and therefore did not strongly affect the magnitude of the yield components at the vineyard sites. Cultural practices at each vineyard site were sufficiently variable to affect fertility levels. Genotype is thought to determine the range of flowers per inflorescence that a variety can potentially carry, whereas actual flower number per inflorescence is thought to be determined by inflorescence primordium initiation and differentiation temperatures, as well as temperatures during budburst. Despite significant correlations between flower number per inflorescence and berry number per bunch, flower number per inflorescence preflowering for Cabemet Sauvignon, Shiraz and Chardonnay is inversely related to actual percentage fruit set. This is possibly a survival mechanism for the grapevine as it allows the vine to maximise yield each season without detriment to its longevity. Bunches per vine accounted for the majority of the seasonal variation in yield per vine. Fluctuations in bunch number per vine (and therefore yield) are likely to be reduced by varying the number of nodes retained per vine according to the relative fruitfulness per node present pre-pruning. This practice is therefore likely to result in the seasonal variation of berries per bunch becoming a stronger driver of yield. The commercial impacts of these studies are two-fold. Data presented will assist growers to understand the reasons for which their pruning regimes are affecting yield production and how these pruning regimes may be modified to achieve a target yield-particularly when growers are faced with seasons of low predicted fertility. In addition, data presented will allow growers to improve their crop forecasting accuracy, with a greater understanding of the link between bunch number and bunch size. In the current situation of oversupply in the wine industry, wineries are adopting a tough stance towards growers over-delivering on their grape contracts. Therefore, any assistance that can be provided to growers on improving accuracy of yield estimates will be beneficial both to the grower and winery. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1352744 / Thesis (M.Ag.Sc.) -- University of Adelaide, School of Agriculture, Food and Wine, 2009

This thesis presents results from two separate studies. First, the impact of bearer length on yield components within the canopy was investigated in season 2005/06, on a commercially-managed, mechanically-pruned vineyard of Vitis vinifera L. Cabernet Sauvignon in Coonawarra, South Australia. Pruning resulted in the retention of bearers with 1-7 nodes, with the weighted average bearer length being two nodes for the canopy. As bearers of one to five nodes in length were the most common, these were studied. Yield components (on a per shoot basis) were analysed according to the node position on the bearer at which the shoot arose. Both budburst and inflorescence number per node were highest at the distal node positions on each length bearer, even if the nodes were at the same positions from the base of the bearer and would normally be expected to have similar fertility. Budburst appeared to act by modifying inflorescence number per node based on the relative location of each node from the apex of the bearer. Shoots that arose from the most distal node positions had the highest flower number per inflorescence and berry number per bunch. Flower number per inflorescence was significantly higher on two-inflorescence shoots than single-inflorescence shoots. The relationship between bunch size and node position, unlike that between inflorescence number and node position, was dependent on bearer length. The relative size of the inflorescence appeared to be affected more so by the node pOSition at which the shoot occurred on the bearer, as opposed to the actual node position on the shoot at which the inflorescence occurred. There was a positive, non-linear relationship between average fruit yield per bearer and bearer length. Although yield was highest from the bearer with the highest node number (five nodes), there was no significant difference in yield per bearer for the bearers of three to five nodes in length. If average bearer length was increased from two to three nodes, the potential yield gain per bearer is estimated at 38 per cent. The second study presents results of correlations between bunch number and components of bunch weight (flower number and berry number) to investigate co-development of bunch number and bunch size. These data were collected from 4 vineyards in the Limestone Coast Zone of South Australia from Vilis vinifera L. Chardonnay, Shiraz and Cabemet Sauvignon during seasons 2002/03 to 2006/07. The significant correlations found between fertility and both bunch weight and flower number per inflorescence suggest that the same factors that affect bunch number in a particular season will also affect bunch size. When inflorescence primordia were initiated and differentiated under cool conditions, actual bunches per node and flowers per inflorescence were low. Differences in climate between the vineyard sites were found to be minimal and therefore did not strongly affect the magnitude of the yield components at the vineyard sites. Cultural practices at each vineyard site were sufficiently variable to affect fertility levels. Genotype is thought to determine the range of flowers per inflorescence that a variety can potentially carry, whereas actual flower number per inflorescence is thought to be determined by inflorescence primordium initiation and differentiation temperatures, as well as temperatures during budburst. Despite significant correlations between flower number per inflorescence and berry number per bunch, flower number per inflorescence preflowering for Cabemet Sauvignon, Shiraz and Chardonnay is inversely related to actual percentage fruit set. This is possibly a survival mechanism for the grapevine as it allows the vine to maximise yield each season without detriment to its longevity. Bunches per vine accounted for the majority of the seasonal variation in yield per vine. Fluctuations in bunch number per vine (and therefore yield) are likely to be reduced by varying the number of nodes retained per vine according to the relative fruitfulness per node present pre-pruning. This practice is therefore likely to result in the seasonal variation of berries per bunch becoming a stronger driver of yield. The commercial impacts of these studies are two-fold. Data presented will assist growers to understand the reasons for which their pruning regimes are affecting yield production and how these pruning regimes may be modified to achieve a target yield-particularly when growers are faced with seasons of low predicted fertility. In addition, data presented will allow growers to improve their crop forecasting accuracy, with a greater understanding of the link between bunch number and bunch size. In the current situation of oversupply in the wine industry, wineries are adopting a tough stance towards growers over-delivering on their grape contracts. Therefore, any assistance that can be provided to growers on improving accuracy of yield estimates will be beneficial both to the grower and winery. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1352744 / Thesis (M.Ag.Sc.) -- University of Adelaide, School of Agriculture, Food and Wine, 2009

Thesis (MScFor)--University of Stellenbosch, 2000. / ENGLISH ABSTRACT: The windbreak species near Wolseley (WoSl and WoS3) was Casuarina cunninghamiana. The windbreak species near Saron (SaSl) and Villiersdorp (ViSl) was Eucalyptus cladocalyx and Pinus radiata respectively. A shelter effect was indicated at ViSI (from northerly winds in winter/early spring 1999), and WoS3 (from southerly winds in late spring 1999). Wind speeds were consistently reduced in the leeward sheltered zone ofViSl and WoS3. To improve the correlation of the shelter effect, mild contaminating winds (comprising approximately 28% of the total data recorded at each site) were excluded for the prediction equations of the shelter effect at each site. Compared to the more exposed wind speeds at 1.0 H to the windward side, wind speeds at ViS] were reduced by 32% at 3.0 H. Compared to 11.0 H, wind speeds at ViSl were reduced by 49 and 46% at 3.0 and 1.0 H respectively. With r2 values above 93%, accurate linear prediction equations were produced. The early barley damage assessments indicated that damage was absent or negligible at 11.1 H to the lee, and significantly highest (X < 0.05) at 13.7 H to the lee and beyond. The crop shelter effect significantly increased (p < 0.05) barley head number, transformed proportion of total barley head mass from above-ground mass, potato tuber mass, potato above-ground mass total potato tuber number, small tuber number, and disproportionate tissue damage to the windward section of each plant. For WoS3, the shelter effect became pronounced with severe south-easterly winds. The average hourly wind speeds at 3.0 H dropped from 0.9 m / s (in the previous and less windy sampling period) to 0.8 m / s, despite the sharp increase in exposed wind speeds. Compared to 13.0 H, wind speeds were reduced by 73 and 32% at 3.0 and 7.0 H respectively. With r2 values above 94%, accurate linear prediction equations were produced. The crop shelter effect significantly increased (p < 0.05) sub-sample mass of lOO grains. At WoS3, strong and sustained wind speeds caused leeward soil temperature increases of up to 4°C at 3.0 H, compared to 11.0 H. Brief strong winds (characteristic of winds at ViSl) had little effect on the soil temperature differences. At ViSl, a deviation of soil moisture content between 3.0 and 11.0 H, following periods of recharge, indicated a potential soil-moisture conservation effect in the sheltered zone. This did not occur at WoS3, due partly to a very low soil moisture content that had little scope for variation; From the crop variations and the microclimate variations at both ViS 1 and WoS3, the maximum shelter effect extended to approximately 4.0 H, followed by an intermediate zone of diminishing shelter that extended to approximately 9.0 H. For ViSl, WoS3 and SaSl, a shading effect significantly reduced yields (p<0.05) at 1.0 H from the respective northern windbreak, compared to yields at 2.0 H. At ViS 1 and WoS3, soil probes did not indicate a depletion of soil moisture resulting from the respective windbreaks. Soil moisture competition was indicated on the northern side of the WoSl windbreak with drier conditions; where soil moisture levels at 1.0 and 3.0 H diverged from a negligible level to a 22% lower level (p > 0.05) at 1.0 H, compared to 3.0H. / AFRIKAANSE OPSOMMING: Die boomsoorte wat as windbreke gedien het was Casuarina cunninghamiana naby Wolseley (WoS 1 en WoS3), en Eucalyptus cladocalyx en Pinus radiata by Saron (SaS 1) en Villiersdorp (ViS 1) respektiewelik. Beskutting is aangetoon by ViS 1 (teen noordelike winde in winter/vroee lente 1999) en WoS3 (teen suidelike winde in laat lente 1999). Windspoed is konsekwent aan die lykant sones van ViS 1 en WoS3 verminder. Ten dien einde die korrelasie van die beskuttings effek te verbeter, is matige kontaminerende winde (ongeveer 28% van die totale waargenome data by elke plek) buite berekening gelaat vir die voorspellingsvergelykings van die beskuttingseffek. In vergelyking met die meer blootgestelde windspoede by 1.0 H aan die windkant, is windspoede by ViSl met 32% verminder by 3.0 H. In vergelyking met 11.0 H is windspoede by ViSl met 49 en 46% by 3.0 en 1.0 H respektiewelik verminder. Met r2 waardes hoer as 93%, is akkurate lineere voorspellingsvergelykings verkry. Die vroee gars skade opnames het aangetoon dat geen of minimale skade by 11.1 H aan die lykant aangerig is, en beduidend die hoogste (x< 0.05) by 13.7 Hen verder aan die lykant was. Die gewas beskuttins effek het beduidende toenames (p< 0.05) in garsare, getransformeerde verhouding van totale gars-are massa van bo-grondse massa, aartappelknol massa, aartappel bo-grondse massa, totale aantal aartappeiknolle, en oneweredige weefsel beskadiging aan die windkant van elke plant tot gevolg gehad. Vir WoS3 het die beskuttingseffek beduidend geword met baie sterk suidoostelike winde. Die gemiddelde uurlike windsnelhede by 3.0 H het van 0.9/ (in die vorige en minder winderige toetstydperk) tot 0.8 m/" verminder, ten spyte van die skerp toename in blootgestelde windsnelhede. In vergelyking met 13.0 H is windsnelhede met 73 en 32% by 3.0 en 7.0 H respektiewelik verrninder. Met r2 waardes hoer as 94% 15 akkurate lineere voorspellingsvergelykings verkry. Die gewas beskuttingseffek het sub-monster massa van 100 graankorrels beduidend (p< 0.05) verbeter. By WoS3 het sterk en volgehoue windsnelhede grondtemperature aan die lykant met tot 4°C by 3.0 H verhoog in vergelyking met 11.0 H. Kortstondige sterk winde (tipiese winde by ViS1) het weinig uitwerking op grondtemperatuursverskille gehad. By ViS1 het 'n afwyking van grondwaterinhoud tussen 3.0 en 11.0 H na tydperke van aanvulling, 'n potensiele grondwaterbewaringseffek in die beskutte gebied getoon. Soortgelyke tendense het nie by WoS3 voorgekom nie, gedeeltelik as gevolg van 'n baie lae grondwater inhoud wat nie veel kon varieer nie. Van die variasies wat in gewasse en mikroklimaat voorgekom het, by beide ViSl en WoS3, kan afgelei word dat maksimum beskutting tot by ongeveer 4.0 H verleen is, gevolg deur 'n intermediere sone van verminderende beskutting tot ongeveer 9.0 H. Vir ViS1, WoS3 en SaSl het 'n skadu-effek gewasopbrengs beduidend (p< 0.05) by 1.0 H verminder in vergelyking met opbrengste by 2.0 H. By ViS 1 en WoS3 het grondwater strooiingspeilers nie 'n uitputting van grondwater as gevolg van die windbreke getoon nie. Kompetisie vir grondwater is getoon aan die noordelike kant van die WoS1 windbreek met droer toestande. Grondwatervlakke by 3.0 H het minimaal maar by 1.0 H met 22% beduidend (p> 0.05) teenoor 3.0 H gedaal.

No description available.

MSCAGR / Department of Plant Production

1. Seventy-five samples of soils were taken in Washington County, Virginia. These samples were taken while the soil survey was in progress. In this group thirty-one different soil types were included. The samples were taken from all parts of the county. Notes were made of the soil type, slope of the land, drainage conditions, erosion, and land utilization. When samples were taken from cultivated fields, the yield of the crop was estimated. When samples were taken from pasture, notes were made of the type and condition of the vegetation. 2. The following properties were determined and studied in relation to soil type and plant growth: pH, available phosphate, available potash, percent organic matter, percent colloids, base exchange capacity, exchangeable hydrogen and present base saturation. The results of these determinations were used in conjunction with the physical properties, which had been determined by the soil survey, in studying the various relationships. 3. The laboratory methods of determination were devised by soil investigators. The reliability of these methods was discussed by means of reference to literature. 4. Soils of the same textural class varied widely in percentage colloids. 5. Soils that are relatively high in available potash varied widely in base exchange capacity, pH, percent base saturation, and available phosphate. 6. Soils at like pH values showed no relationship of percent organic matter to percent base saturation. 7. Soils above pH 6.0 contained 125 pounds per acre or more available potash; they were above 65 percent base saturation and were widely variable in available phosphate and organic matter. Soils at similar pH values showed wide variations in available phosphate. 8. Within the same textural class, and under similar conditions of land utilization, soils varied widely in percent organic matter. 9. There was some variation in pH values of residual soils of limestone origin. However, the average values of five samples of each were very close together. The average for Dunmore was 5.4, Hagerstown 5.5, and Clarkesville 5.6. 10. Available phosphate and potash, as well as pH, were closely related to yield of corn and tobacco. 11. Available phosphate and closely related to quality of meadow. In soils where available phosphate was above 50 pounds per acre quality of meadow was good. Potash was present in sufficient quantity not to be a limiting factor. 12. A close relationship was found between some of the soil properties studied. The coefficient of correlation between base exchange capacity and percent organic matter was .7191 ± .0376; between pH and exchangeable hydrogen -.6500 ± .0450; between pH and percent base saturation .9008 ± .0147. This high degree of correlation of pH and percent base saturation, it appears, would hold only for a large number of soils. Within this group, soils at like pH values vary considerably in percent base saturation. There was little relationship between percent colloids and base exchange capacity. 13. In all cases, in pasture soils, where available phosphate was below nine pounds per acre, pastures were of poor quality. There was a gradual decline in the minimum range of available phosphate and potash in the soils from the good, medium and poor quality pastures. There is also a gradual decline in the average pH values as quality of pasture declines. 14. Evidence is shown that, within a soil type, the properties of the soil, under varying conditions of soil management vary widely. The soil type name gives no indication of the fertility of the soil at the present time. However, associated with the type name are certain physical properties which determine, largely, the possibilities or limitations of that soil type, or the degree of productivity that the type may be built up to under ideal conditions of soil management. / Master of Science

Thesis (M. Tech. (Agriculture)) -- Central University of Technology, Free state, 2014 / The main objective of the study was to determine the effect of different potassium concentrations and water quality (salt) compared with the current scientifically accepted potassium threshold level and standardised water quality on the yield, oil composition and leaf morphology of rose geranium (Pelargonium graveolens L.) when grown in different potting-bag sizes and root media under temperature controlled condition. To achieve this objective, two trials were conducted. The first experiment evaluated potassium concentrations at 1.3, 3.3, 5.3 and 7.3 mmol L-1 and potting-bag size of 5 and 10 L. Treatments were arranged in a randomised complete block design assigned in a split plot layout. The main plots consisted of potassium concentration and the pottingbag sizes were allocated to sub-plots. Plant height, potassium content, linalool, geraniol, geranyl formate and the citronellol to geraniol ratio (C:G) were affected by potassium. Plant height, number of branches, the branch to height ratio (B:H), foliar fresh mass (FFM) and oil yield were significantly increased when 5 L potting bags were used. Plant foliar mass was significantly increased by the interaction between 5.3 mmol K L-1 and 5 L potting bags. In the second experiment salt levels applied at 1.6, 2.4, 3.2 and 4.0 mS cm-1 and root media (sand and sawdust) were evaluated. A split plot experimental layout was also used in this trial, with the salt levels allocated to the main plots. The sub-plots were allocated to the root medium. High salt level of 4.0 mS cm-1 reduced the number of leaves, plant height, number of branches, B:H ratio, leaf area, chlorophyll content and foliar fresh mass significantly. The number of leaves, leaf area and FFM were significantly increased where sawdust was used. Time of the day significantly affected stomatal conductance, and the opening of most stomata occurred at 10:00. Geranyl formate and the C:G ratio were significantly affected by salt at 4.0 mS cm-1. Salt induced the development of capitates trichomes. The abaxial leaf surface had a higher number of trichomes than the adaxial leaf surface. A strong polynomial (r2=0.97) relationship was found between capitates trichomes and salt levels. High densities of capitates trichomes were found at high salt level of 4.0 mS cm-1. Although the development of asciiform trichomes was induced, it was at an insignificant level. Trichome densities are therefore not affected by salt. It was therefore concluded that the application of 5.3 mmol K L-1 concentration and the use of 5 L potting bags improves the yield and oil quality of rose geranium. It was evident from this study that rose geranium might Salt induced have some degree of tolerance to salt. It was therefore concluded that rose geranium is a moderately salt-sensitive crop.