Canola variety, nitrogen, phosphorus, and sulfur fertilization affect yield, quality, and fatty acid profile

Abstract Canola yield and quality are important for food, feed, and industrial end-uses. There may be trade-offs between the agronomic and quality aspects of canola production depending on varietal traits and management. The objective of this work was to assess the effects of nitrogen (N), phosphorus (P), and sulfur (S) fertilization on agronomic and quality properties of canola varieties with distinct oleic acid contents. Nitrogen fertilization rates were 0, 25, 50, or 100 kg·ha−1, P rates were 0 or 30 kg·ha−1, and S rates were 0 or 20 kg·ha−1. Canola was grown in 2003, 2004, and 2005 at Brandon, a private farm close to Brandon, and at Lacombe, Canada. Canola yields averaged 2.36 t·ha−1 for conventional, 2.53 t·ha−1 for low, and 2.2 t·ha−1 for the high oleic acid varieties at maximum fertilization of N, P, and S. The high oleic acid variety averaged 75% oleic acid content, whereas the low variety averaged 65%, and the conventional variety 62%. Total saturated fatty acids were greatest with the conventional oleic acid variety, and tended to increase with N, decrease with S, and were not influenced by P. The high oleic acid variety yielded slightly less than the other two varieties but tended to have lower glucosinolate and saturated fatty acid contents. This work could have implications for human nutrition or other end-uses. Current canola varieties and fertility management should be studied to produce canola with quality tailored for the end use.


Introduction
Canola (Brassica napus L.) is one of Canada's most important crops.Canola production in Canada averaged about 17 million tonnes per year from 2010 to 2020, second only to wheat over this period (Statistics Canada 2022).Canola oil is used as biofuel, industrial oil, an edible oil, and as animal feed (McVetty and Duncan 2016).
Canola oil has a lower saturated fatty acid content than other vegetable oils such soybean, sunflower, corn, olive, almond, peanut, and coconut (Orsavova et al. 2015;Ghazani and Marangoni 2016).Human consumption of saturated fatty acids can increase cholesterol levels and can impact overall health (Ghazani and Marangoni 2016).Canola oil also contains a relatively high level of oleic acid (Orsavova et al. 2015;Ghazani and Marangoni 2016).In diets high in fat, canola oil that is high in oleic acid can prevent the development of diastolic dysfunction in rats (Thandapilly et al. 2017).The U.S. Food and Drug Administration suggested that consumption of high levels of oleic acid can reduce the risk of coronary heart disease (FDA 2018).
Over the past 20 years in Canada, canola oil and oleic acid content have been increasing slightly, whereas cholorophyll in the seed, total glucosinolates, erucic acid content, and total saturated fatty acids have slightly decreased (see Table 1) (DeClercq and Daun 2000;Barthet 2020).High oleic fatty acid content canola varieties continue to be developed, and at least one such cultivar entered the Canadian market in 2022 (Halsall 2021).
The canola industry in Canada has prioritized lower saturated fatty acid contents as a key quality trait in the crop (McCartney et al. 2004).Both genetic and environmental factors influence canola quality and fatty oil profiles (McCartney et al. 2004).However, the extent to which management practices influence canola quality and fatty acid content remains unclear.
Selection of canola variety is often based on yield, weed management, maturity, plant height, lodging, disease tolerance, as well as premiums for canola varieties that produce oils with fatty acid profiles for specific functions such as human consumption (Hammond 2011).
Canola has a high demand for crop nutrients (Grant and Bailey 1993;Harker et al. 2012;Ejack et al. 2021).An adequate supply of nitrogen (N), phosphorus (P), and often sulfur (S) applications is needed to optimize crop yield (Karamonas et al. 2005).Therefore, proper nutrient management is important for efficient canola production, with fertilizer applications making up the difference between soil supply and crop demand.Nitrogen application has been shown to in-Table 1.Comparison of average canola quality data for varieties used in the current work (i.e., CNH501R, DKL3455, and IMC304RR) with means for canola exported from Canada before the study was conducted (1990-1999) and example data from years previous to the publishing of this work (2015)(2016)(2017)(2018)(2019)  Barthet (2020)   ‡ Total saturated fatty acids are the sum of the percentages of all saturated fatty acids present, including palmitic (C16:0), stearic (C18:0), arachidic (C20:0), behenic (C22:0), and lignoceric (C24:0).crease protein content of canola but decrease oil content and may also increase chlorophyll and glucosinolate levels (Grant and Bailey 1993).High nitrogen fertilization rates can lower canola oil content (Harker et al. 2012;O'Donovan et al. 2014).Phosphorus and sulfur may either increase or decrease canola oil content depending on the balance between crop demand and nutrient supply (Grant and Bailey 1993;Malhi and Gill 2006).
While information is available on the impact of nutrient management on oil content, there is little information in the literature on the impact on oil quality.Studies in Ontario found that increased N fertilization levels increased the free fatty acid content in spring canola and that cultivar and environmental stress influenced the size of response (May et al. 1994); however, they did not examine the fatty acid profile of the crop.Nitrogen and sulfur fertilizer applications have been found to affect fatty acid composition of canola oil and decrease oil content, but effects depend on environmental conditions (Gao et al. 2010).Indeed, available water and temperature can play a greater role than soil nutrient supply on canola yield and quality (Hammac et al. 2017).Temperature during canola development can influence the fatty acid composition in the seed (Deng and Scarth 1998).Even though environmental influences can have a large impact on canola yield and quality, N and S fertilizer applications can also influence fatty acid contents, in particular where high soil N levels do not dilute effects (Hammac et al. 2017).
Low saturated fatty acid canola varieties may be important in improving canola quality.Changes in genetics may also alter the fertility requirements of a crop; therefore, agronomic information is needed on the fertility responses of low saturated fatty acid varieties of canola.However, there is a gap in understanding the interrelations of variety selection and soil fertility management with canola quality properties including fatty acid profiles, glucosinolate, oil content, and protein.The aim of this work was to determine the effect of rate of application of N, P, and S for three canola varieties, bred for disparate fatty acid profiles, on agronomic and crop quality properties.Fertility management and variety selection may be used to produce canola with fatty acid and quality properties that would promote positive human nutrition.It is therefore important to manage canola crops to maximize properties beneficial to human nutrition while balancing crop yield.

Materials and methods
Canola was grown at three locations in Manitoba and Alberta, Canada, in 2003, 2004, and 2005. In 2003and 2004, data were collected from the Phillips farm of the Brandon Research and Development Centre (BRDC) of AAFC in Manitoba (50.020167 • , −99.880107 • ), a private farm approximately 5 km west from the Phillips farm, and AAFC Lacombe in Alberta (52.453826 • , −113.751732).In 2005, canola was grown at the Phillips farm, Lacombe, and at the BRDC in Brandon, Manitoba (49.865774 • , −99.989915 • ).The soils at the Phillips farm, the private farm, and Lacombe are loamy Orthic Black Chernozems, and the soil at the BRDC is a coarse loamy to fine loamy Gleyed Cumulic Regosol.Plots were arranged in a randomized complete block design with four blocks at all sites.Sites low in soil-test N (20 ppm NO 3 − -N), P (10 ppm Olsen-P), and S (20 ppm SO 4 −2 -S) at each location were selected each year for this research study.Canola varieties included a conventional open-pollinated Roundup Ready (glyphosate resistant) variety from Limagrain/Monsanto (DKL3455), a low oleic acid Roundup Ready variety (CNH501R in 2003 and2004 (discontinued) or V1032 in 2005), and a high oleic acid Roundup Ready open-pollinated variety (IMC304RR for all but 2005 at Lacombe which was IMC209RR).There were 13 fertilizer treatments at each site-year.Nitrogen rates were 25, 50, and 100 kg•ha −1 , P rates were 0 and 30 kg•ha −1 , and S rates were 0 and 20 kg•ha −1 .There was also a 0 kg N ha −1 , 0 kg P ha −1 , and 0 kg S ha −1 control.These N fertilizer treatments were close to the Manitoba fertilizer recommendations of 80-100 kg N ha −1 , 15-20 kg P ha −1 , and 22 kg S ha −1 where no soil test data are available (Manitoba Agriculture Food and Rural Initiative 2007).Fertilizer P was side-banded as mono-ammonium phosphate.The N and S were applied as a pre-plant band, with N applied as ammonium sulfate and N as urea.Canola was direct seeded with either a ConservaPak or SeedHawk on 23 cm row spacing, at a rate of 4-6 kg•ha −1 , at 1-2 cm depth, in early May.Plots were 2 m by 5 m.Weeds were controlled preseeding with glyphosate at 1 L per 0.405 ha and as needed during the growing season.Crop emergence, seed yield, and dockage were measured at all sites in all years.Plant biomass at flowering was measured at Phillips, BRDC, and the private farm, but is not included here.Canola emergence was determined by counting the number of plants in a 1 m length in two rows at representative locations, 2 to 3 weeks after emergence.Plot combines were used to harvest the canola and yield was adjusted to 10% moisture content.Dockage is the percentage of material in a sample that is not a kernel of seed, measured after sieving through calibrated screens.Fatty acid data were collected for canola at Phillips and the private farm in 2003 and 2004, as well as at Lacombe in 2004 and analyzed at Cargill in Saskatoon, Canada.Fatty acid composition in oil was determined by gas chromatography (ISO 5509 2000;ISO 15304 2002).Near-infrared spectroscopy was used to measure canola quality, using calibrations from chemical methods for chlorophyll (ISO 10519 1997), glucosinolate (ISO 9167 1992), oil (ISO 659 1998), and protein (AOCS Ba 4e-93, N concentration × 6.25).
Statistical analyses were performed in version 4.0.3 of R (R Core Team 2020).A multivariate ordination method, redundancy analysis (RDA), was performed to explore the canola quality data for the five site-year combinations and plotted to highlight interrelations of explanatory and response variables (Oksanen et al. 2020).Nitrogen, phosphorus, sulfur, canola variety, year, and site were included as explanatory variables.However, nitrogen, phosphorus, sulfur, and canola variety were included as constraints to be able to investigate the interrelations of canola quality and agronomic response data, and year and site were included as conditions as their inclusion was not the focus of this work.Fertilization influence on canola emergence, yield, and dockage for each variety were evaluated using linear mixed-effects models with maximum likelihood (Pinheiro et al. 2021).Fixed effects were N, P, and S fertilization and variety individually and as interactions.The 0 kg N ha −1 , 0 kg P ha −1 , and 0 kg S ha −1 was omitted from the linear mixed-effects models to balance the design; however, contrasts were used to confirm control versus fertilization treatments (Lenth 2021).Blocks were designated as random effects.Sites and years were also considered as random effects because our goal was to make inferences over space and time (Yang 2011), similar to Harker et al. (2012).Least-squares means and differences between combinations were calculated using the mixed-effects models described above for each variety and fertilizer treatment (Lenth 2021; Tsegelskyi 2022).Assumptions for ANOVA were tested using a combination of visual and quantitative techniques.Homogeneity of variance was tested with Levene's test (Kassambara 2021), normality with quantile plots (Kassambara 2020) and Shapiro-Wilk (Kassambara 2021), and outliers were identified (Kassambara 2021).All ANOVA assumptions were met.

Results
Growing conditions varied during the 3-year study (Table 2).In 2003, it was warm and dry, relative to the long-term average, at Brandon and Lacombe.Brandon was cooler and wetter than average in 2004 and wetter than average in 2005.Conditions at Lacombe were close to average in both 2004 and 2005.

Canola agronomic data
Over the nine-site years of agronomic data, canola yield and dockage were significantly affected by variety, N, P, and S (P < 0.01 for each) (Table 3).There were significant interactions of N and S for both canola yield (P = 0.01) and dockage (P = 0.04).Canola emergence was not significantly influenced by variety or fertilizer treatments.Canola yield was greatest at the maximum rate of fertilization of 100 kg N ha −1 , 30 kg P ha −1 , and 20 kg S ha −1 for the conventional (2.36 t•ha −1 ), low (2.53 t•ha −1 ), and high oleic acid varieties (2.20 t•ha −1 ).Canola yield for the conventional variety at 100 kg N ha −1 , 30 kg P ha −1, and 20 kg S ha −1 was significantly greater than where 100 kg N ha −1 was applied without P or S, as well as most other fertilization treatments (Table 4).There was a significant interaction between N and S rate for canola yield (P < 0.01, Table 3).Relative to 100 kg N ha −1 alone, P added 0.19 t•ha −1 , S added 0.13 t•ha −1 , although not significantly different, and P and S together significantly increased canola yield by 0.34 t•ha −1 .For the low oleic acid variety (Table 4), the 100 kg N ha −1 , 30 kg•ha −1, and 20 kg S ha −1 treatment was significantly greater than all other fertilizer treatments where 100 kg N ha −1 was applied with either P or S. Relative to 100 kg N ha −1 alone, S significantly increased yield by 0.30 t•ha −1 , and P and S together significantly increased yield by 0.34 t•ha −1 , for the low oleic acid variety.For the high oleic acid variety (Table 4), P and S together significantly increased yield by 0.46 t•ha −1 relative to 100 kg N ha −1 without P and S.
Dockage tended to decrease (not always significantly so) with increasing fertilizer rates relative to the 100 kg N ha −1 with no P or S treatment for all varieties (Table 4).
Redundancy analysis was run on the site-year combinations where canola quality data were available in addition to agronomic data (Fig. 1).The site-year combinations were Phillips Farm 2003 + 2004, the private farm 2003 + 2004, and Lacombe 2004.The RDA model was significant (P < 0.01) overall, and the constraining factors of nitrogen, phospho-  Growing Degree Days (GDD) were calculated following the method described by Agroclimate Maps of Agriculture and Agri-Food Canada (https://www.agr.gc.ca/eng/agriculture-and-climate/drought-watch/agroclimate-ma ps/about-the-agroclimate-maps/?id=1463576686512).The daily maximum temperature is added to the daily minimum temperature and divided by two and then the temperature threshold for canola (5 Negative values were set to zero.The GDD were summed from the beginning of May to the end of August following Harket et al. (2012).
Table 3. P values for linear mixed-effects models for nitrogen (N), phosphorus (P), and sulfur (S) fertilization, and variety (V) for nine site-years of canola emergence, yield, and dockage.rus, sulfur, and canola variety were all significant (P < 0.01).However, since the fertilizer treatments had much lower relative importance compared to the variety effects, the treatment centroids were too close to the origin to be visually apparent and were therefore excluded from the triplot.Year and site were 39% of total variance, nitrogen + phosphorus + sulfur + canola variety were 22% of total variance, and 38% of total variance was unexplained.The proportion of accumulated constrained eigenvalues for the first RDA axis was 66% and the second was 24%.The variance inflation factors were all less than 4, so the full model was retained.A total of 14 fatty acids were assessed.Saturated fatty acids displayed are myristic acid (C14:0), palmitic acid (C16:0), stearic acid (C18:0), arachidic acid (C20:0), behenic (C22:0), and lignoceric acid (C24:0).Mono-unsaturated fatty acids displayed are palmitoleic acid (C16:1), oleic acid (C18:1), gadoleic acid (C20:1), erucic acid (C22:1), and nervonic acid (C24:1).Polyunsaturated fatty acids displayed are linoleic acid (C18:2), αlinolenic acid (C18:3), and eicosadienoic (C20:2).The three varieties were distinctly grouped and are separated primarily by the three fatty acids with the greatest concentrations (i.e., oleic, linoleic, and α-linolenic acid) (Fig. 1).The conventional variety had greater α-linolenic acid, the high variety had greater oleic acid, and the low variety had slightly greater linoleic acid contents.Other quality properties displayed are total saturated fatty acids, chlorophyll, glucosinolate, moisture, oil, and protein.Agronomic properties displayed are canola yield, plant emergence, and dockage.The high oleic acid variety tended to have greater chlorophyll, oil, and protein and less glucosinolate but lower yield than the other two varieties.Total saturated fatty acids was positively correlated with palmitic and stearic acid; however, these two were negatively correlated with myristic and behenic acid.Canola variety significantly affected all fatty acids and quality properties (P < 0.01), except erucic acid (P = 0.42), accord-   indicates significant difference of no-fertilizer control to the average of all of the fertilizer treatments ing to the linear mixed-effects models (Table 5).Canola variety also significantly affected yield and dockage (P < 0.01) but not emergence (P = 0.18).Nitrogen application significantly affected myristic, palmitic, palmitoleic, stearic, oleic, arachidic, gadoleic, eicosadienoic, behenic, lignoceric, and nervonic fatty acid.Nitrogen application also significantly affected chlorophyll, oil, protein, and total saturated fatty acids in the linear mixed-effects models.Phosphorus application significantly affected myristic, palmitic, stearic, oleic, linoleic, arachidic, and gadoleic fatty acids.Phosphorus significantly affected glucosinolate, moisture, oil, and protein in the linear mixed-effects models.Sulfur application significantly affected myristic, palmitic, palmitoleic, stearic, oleic, linoleic, arachidic, gadoleic, eicosadienoic, and behenic fatty acids, as well as total saturated fatty acid contents, moisture, and oil content (Tables 5-8).
Canola oil content over all site-years and fertilizer treatments for the conventional variety was 47.2% and 47.5% for the low, and 48.0% for the high (Table 1).Nitrogen fertilization tended to reduce oil content in all varieties.The oil content of the 100 kg N ha −1 rate with no P or S was approximately 2% lower than the lowest N rate for each variety.For the conventional variety, the 100 kg N ha −1 rate had significantly lower oil content than the 25 kg N ha −1 rate where no P and S were applied, where just P was applied, and where both P and S were applied (Table 9).The same pattern present in the low oleic acid variety also occurred where sulfur was applied.For the high oleic acid canola variety, oil content significantly decreased with increasing N rate where no P or S and where only P was applied.
Protein levels in all varieties were greatest where 100 kg N ha −1 with no P but with S was applied.Conventional had 25.4%, low had 26.5%, and high had 27.2% protein on average (Table 9).Where 100 kg N ha −1 was applied, regardless of P or S rate, protein contents were significantly greater than the 25 kg N ha −1 rate for both P and S rates for the conventional and low oleic acid varieties.This pattern was also apparent in the high oleic acid variety where P and/or S was also applied, but not where neither was applied with 25 kg N ha −1 .
Glucosinolate averaged 13.3 μmol•g −1 in the conventional oleic acid canola variety, 12.3 μmol•g −1 in the low, and 10.1 μmol•g −1 in the high (Table 9).There were no significant differences in glucosinolate amongst fertilizer treatments for any of the canola varieties.
Chlorophyll content in the canola seed averaged 13.3 mg•kg −1 in the conventional variety, 12.3 mg•kg −1 in the low, and 24.1 mg•kg −1 in the high (Table 9).There were no significant differences amongst fertilizer treatments found in the conventional or low oleic acid varieties.Chlorophyll tended to increase with N application rate; however, the high oleic acid variety showed a significant increase of 1.42 mg•kg −1 from the 25 kg N ha −1 30 kg P ha −1 20 kg S ha −1 rate to the 100 kg N ha −1 30 kg P ha −1 20 kg S ha −1 rate.
N ha −1 relative to the same N rate with P (3.7) and with P and S (3.7).

Discussion
Canola yield, emergence, and dockage were not strongly associated with any one canola variety (Fig. 3), though the high oleic acid variety yielded slightly lower and had slightly higher dockage.Although it was not the objective of the current study, site and year contributed to the interrelations of the agronomic and quality properties (39% of total variance in RDA).It is apparent that there are trade-offs between agronomic performance and canola oil quality properties that relate to human nutrition.For example, canola yield tended to increase with fertilization but so did saturated fatty acid contents (Figs. 2 and 3).The high oleic acid variety tended to have greater oleic and gadoleic acids and chlorophyll and had less glucosinolates, but yielded slightly less than the conventional and low oleic acid varieties, whereas the conventional variety had greater total saturated fatty acids.It is apparent that the canola varieties differed in their saturated fatty acid profiles.The conventional variety had greater amounts of the two saturated fatty acids (palmitic and stearic) with greater concentrations overall, whereas the high and low canola varieties had greater amounts of the saturated fatty acids with lower overall concentrations (myristic, arachidic, and lignoceric).Saturated fatty acids are often grouped together into a single property in discussing implication for human nutri-tion from canola oil (Lunn and Theobald 2006); however, it is apparent from the current data that canola varieties have statistically different saturated fatty acid profiles that are influenced by fertilization.In the current study, where maximum N was supplied to the canola plants, S had an effect on yield and saturated fatty acid contents, whereas P effected only yield.This is similar to Appelqvist (1968) who found N effects but not P on fatty acid composition.
Average canola yields in the current study were 2.2 t•ha −1 for the high oleic acid, 2.4 t•ha −1 for conventional, and 2.5 t•ha −1 for low oleic acid for the maximum fertilizer rate (100 kg N ha −1 , 30 kg P ha −1 , 20 kg S ha −1 ).These yields were similar to Karamonas et al. (2005) that presented data from 1999, just previous to the current study, of between 2.4 and 3.0 t•ha −1 at 100 kg N ha −1 , and so the canola yields in the current study were representative for the time.In that study, canola yields were dependent on N fertilization and optimum canola yields occurred when there were no N, P, or S deficiencies (Karamonas et al. 2005).
Nitrogen tended to decrease oil content but increase protein, chlorophyll, and saturated fatty acid contents.Since oil and protein are the two main components of the seed, and are reported as percentages, increasing one will decrease the other.Increasing N fertilization rates may increase total oil yield per hectare, but still decrease the percentage of oil in the seed.Normally, as protein in canola seed increases, oil content decreases (Grant and Bailey 1993).
Nitrogen fertilization has been found to lower oil content and saturated fatty acid contents (Gao et al. 2010).Gao et

Table 5.
Probabilities for linear mixed-effects models for nitrogen (N), phosphorus (P), sulfur (S) fertilization, and variety (V) for five site-years of canola fatty acid, total saturated fatty acids (TSF), and quality.

Gluc
Canola yield was always greatest in the overall site-year data where maximum fertilization was used (i.e., 100 kg N ha −1 30 kg P ha −1 20 kg S ha −1 ).Where P and S were applied, N increased yield at 100 kg N ha −1 compared to 25 and 50 kg N ha −1 for conventional, and compared to 25 kg N ha −1 for the low and high oleic acid varieties.The 100 kg N ha −1 30 kg P ha −1 20 kg S ha −1 reduced oil content relative to 25 kg N ha −1 30 kg P ha −1 20 kg S ha −1 in both the conventional and low varieties; however, for the high oleic acid variety the maximum N fertilization of 100 kg N ha −1 reduced oil content relative to the same N content with P and S. Protein significantly increased in all varieties for 100 kg N ha −1 30 kg P ha −1 20 kg S ha −1 relative to 25 kg N ha −1 30 kg P ha −1 20 kg S ha −1 and increased protein compared to 25 kg N ha −1 30 kg P ha −1 0 kg S ha −1 .For the conventional variety at the maximum fertilization rate, total saturated fatty acid content was significantly greater than 25 kg N ha −1 0 kg P ha −1 20 kg S ha −1 , eicosadienoic acid content at 25 kg N ha −1 30 kg P ha −1 20 kg S ha −1 , arachidic acid at 25 kg N ha −1 0 kg P ha −1 20 kg S ha −1 and 25 kg N ha −1 30 kg P ha −1 0 kg S ha −1 .For the low oleic acid variety at maximum fertilization, myristic acid was significantly greater than at 25 kg N ha −1 30 kg P ha −1 20 kg S ha −1 .For the high oleic acid variety, the maximum fertilization rate had significantly lower total saturated fatty acids than 100 kg N ha −1 without P and S, had lower gadoleic acid than at 25 kg N ha −1 30 kg P ha −1 0 kg S ha −1 , and greater myristic, palmitoleic, and stearic acid contents.
There were some P and S fertilization effects on canola agronomic and quality properties; however, they were more subtle than for N.In most cases where sufficient N was supplied to the canola plant (i.e., 100 kg N ha −1 ), many effects of P and S were overwhelmed.For example, for the conventional canola variety, yield was significantly increased from 2.02 t•ha −1 for 100 kg N ha −1 to 2.36 t•ha −1 with the addition of P and S. In general, fertilizer management influenced canola agronomic and quality properties; however, both environmental factors and genetics play roles that can be larger than fertilizer management, similar to the results reported by Appelqvist (1968), andHammac et al. (2017).Where canola yield was not greatest or significantly different from the greatest value (i.e., 100 kg N ha −1 30 kg P ha −1 20 kg S ha −1 , 100 kg N ha −1 0 kg P ha −1 20 kg S ha −1 , and 100 kg N ha −1  30 kg P ha −1 0 kg S ha −1 , for all varieties), oil and protein were effected by increasing N without P or S additions.Fatty acids such as myristic, palmitic, palmitoleic, stearic, arachidic, lignoceric, and nervonic were influenced by N, P, and/or S where no canola yield increase was observed from fertilization.For the high oleic acid variety only sulfur applied with the lowest N rate significantly increased oleic acid content relative to the maximum N rate without P or S. Brassica napus oil content, protein, and saturated fatty acid content can be influenced by N, P, and S (Appelqvist 1968), as well as genotype and environment (McCartney et al. 2004).

Conclusion
This work showed that nitrogen, phosphorus, and sulfur fertilization can influence agronomic and quality properties, as well as fatty acid profiles of both conventional canola varieties and those bred for greater oleic acid contents.The total amount and range of the quality property (e.g., oleic acid or saturated fatty acid content) is mostly determined by the genetics of the variety; however, fertilization can also play a role, albeit a more subtle one.Canola yield and protein tended to increase with fertilization but so too did saturated fatty acid contents, those fatty acids associated with negative human nutrition.Individual saturated fatty acids were significantly affected by fertilization in many cases, but were not uniformly associated with a specific variety.This implies that fertilization and variety influence both the total content and profile of saturated fatty acids.Future work should include fertilization rates that create response curves using current canola varieties considering crop nutrient uptake, as well as an economic analysis of the use of high oleic varieties.This could demonstrate the potential for genetic by management interactions to produce canola oil with properties better suited to the end uses, including oil for better human nutrition.
Precipitation and growing degree days (GDD) during the growing season (May to August) relative to long-term average.The two farms near Brandon, MB are close enough together that the weather conditions are considered similar and thus only weather data from Brandon Municipal Airport are reported.* ), and dockage (%) of CNH501R (low oleic), DKL3455 (conventional (Conv.)oleic), and IMC304RR (high oleic) canola averaged over sites and years.

Fig. 2 .
Fig. 2. Mean canola yield by variety (low, conventional, and high oleic acid contents) and nitrogen (N), phosphorus (P), and sulfur (S) fertilization over all site years.Different letters beside data points indicate significance.

Fig. 3 .
Fig. 3. Mean canola saturated fatty acid content by variety (low, conventional, and high) and nitrogen (N), phosphorus (P), and sulfur (S) fertilization over all site years.Different letters beside data points indicate significance. . *

Table 2 .
Summary of growing conditions during the experiments.

Table 4 .
Effect of nitrogen (N), phosphorus (P), and sulfur (S) fertilizer (kg•ha −1 indicates significant difference of no-fertilizer control to the average of all of the fertilizer treatments. *