Crambe: Seed Yield and Quality in Response to Nitrogen and Sulfur—A Case Study in Northeastern Poland

The aim of this study was to determine the effect of nitrogen (0, 30, 60, 90, 120 kg ha−1) and sulfur (0, 15, and 30 kg ha−1) fertilization on the morphometric parameters of plants, seed yield components, seed and straw yield, N fertilizer use efficiency (NFUE), and quality of crambe seeds. The experiment had a randomized complete block design, and it was carried out in Bałcyny (northeastern Poland) in 2017–2019. In northeastern Poland, the average seed yields ranged from 0.96 to 1.64–1.82 Mg ha−1 (hulled seeds). Seed yield increased significantly in response to 120 kg N ha−1 and 15 kg S ha−1. The NFUE of crambe decreased by 28% with a rise in N rate. Hulled crambe seeds accumulated 324–394 g kg−1 DM of crude fat, 208–238 g kg−1 DM of total protein, and 118–137 g kg−1 DM of crude fiber. Nitrogen fertilization decreased the crude fat content (by 6%), and it increased the total protein content (by 11%) and the crude fiber content (by 14%) of crambe seeds. Sulfur fertilization increased crude fat content (by 4–5%) without inducing significant differences in the total protein content and the crude fat content of seeds.


Introduction
The oil from plants of the family Brassicaceae is a natural source of erucic acid (EA, C22:1), one of very long chain fatty acids (VLCFAs). In plants, VLCFAs and their derivatives are the precursors of suberin, pollen coat, cuticular waxes, and sphingolipids [1]. Erucic acid has numerous industrial applications; it is used in the production of polymers, polyesters, emulsifiers, detergents, ink, paper, cosmetics, pharmaceuticals, textiles, lubricants, food, and fuel [2]. The worldwide consumption of EA increased from 18 to 35 Tg between 1990 and 2010 [3]. The seeds of rapeseed cultivars (Brassica napus var. oleifera L.) with a high content of erucic acid and low concentrations of glucosinolates (high erucic acid rape, HEAR), as well as crambe seeds (Crambe abyssinica Hochst. ex R.E. Fries), are the most popular "green" sources of EA [2,4]. However, the cultivation of HEAR with other rapeseed cultivars (such as low erucic acid rape, LEAR, for food processing) is difficult due to high levels of gene flow mediated by pollen (cross pollination) and seeds (seed shedding during maturation and harvest, and a long period of secondary dormancy) [5,6]. In LEAR production, the risk of EA contamination from HEAR varieties is difficult to control and requires numerous adjustments, including appropriate crop rotations with a low share of Brassica crops, the establishment of buffer zones, and effective control of weeds, in particular of the family Brassicaceae [6]. At present, the oil from C. abyssinica seeds (synonym: Abyssinian oil) is favored over HEAR oil due to a higher content conditions. There is a general scarcity of published data on the effects of crop rotation on crambe cultivation. In a study conducted in the Northern Great Plains, crambe was a less suitable preceding crop for durum wheat (Triticum durum Desf.) than camelina (Camelina sativa (L.) Crantz) and Indian mustard (Brassica juncea (L.) Czern.) [55]. Crambe was not effective in reducing the population of soybean cyst nematodes (SCN) in soils, which suggests that this crop species is not suitable for sustainable management of pathogens in large-area soybean farms in SCN-infested regions [56]. However, due to its short growing cycle, the species could be a highly suitable preceding crop for winter cereals in Europe [46]. In Australia and New Zealand, Brassica crops are more widely used in rotation and intercropping with rice and wheat [57]. The oil and nonfat seed residues of C. abyssinica constitute renewable feedstocks for biofuel and bio-based products in the oleochemical industry [25,38,58].
Brassica crops accumulate large amounts of protein, crude fat, and GLS, and they have a high demand for fertilizers, in particular N. Oils are important plant metabolites with the highest energy density among all carbon reserves [59]. Nitrogen affects the distribution of nutrients to roots and assimilative organs, thus influencing photosynthetic capacity and crop stand productivity [60]. Nitrogen is a part of structural compounds, carriers of energy, and genetic information, as well as compounds that regulate plant metabolism. In Brassica crops, the synthesis of nutrients and biologically active compounds is a highly energy-intensive process, which could explain low N fertilizer use efficiency (NFUE). In rapeseed, Indian mustard, camelina, field mustard (Brassica rapa L.), white mustard (Sinapis alba L.), and Abyssinian mustard (Brassica carinata A. Braun), NFUE has been determined in the range of 11.1 to 26.4 kg seed kg −1 N [61][62][63]. In B. napus, the highest yielding Brassica crop, NFUE is 27-81% lower than in common wheat (Triticum aestivum L.) (26 vs. 33-47 kg seed kg −1 N) [61,64,65]. In B. napus, the production of 1 Mg of seeds with the corresponding yield of stems, leaves, and roots requires 50-60 kg N [66]. The remaining Brassica crops, including C. sativa [62,67,68], B. juncea and B. rapa canolas, B. juncea and S. alba mustards [61,69], and S. alba canola [69], have an equally high demand for N. In oilseed crops of the family Brassicaceae, crude fat content is the most important quality trait which is negatively correlated with total protein content [70,71]. Nitrogen decreases the content of crude fat and increases the total protein content of seeds in Brassica crops [68,72,73]. A negative correlation between N fertilization and the oil and protein content of seeds is not observed only when N does not exert yield-forming effects [63].
Brassica oilseed crops need 15-20 kg S to produce 1 Mg −1 seeds and the corresponding straw yield [74]. Sulfur is present in selected amino acids and glutathione, which are responsible for the structure of the protein chain [75]. Compounds containing sulfhydryl groups (-SH) participate in redox reactions [76] and determine the activity of many enzymes. Sulfur is also found in sulfolipids, the structural elements of cell membranes [77]. Brassicaceae crops have a high demand for S that participates in the biosynthesis of GLS [78]. All GLS contain S and glucose, but they differ in the structure of aglycone, which determines the compound's properties [79]. The effect of S fertilization on the metabolism of Brassica plants, including the content of essential nutrients, varies depending on SO 4 2 levels in soil, atmospheric deposition of S, the biological activity of soil which affects the transformation rates of S, and agronomic factors [80]. In recent years, S has emerged as an important fertilizer due to the progressive depletion of S from soils, increasing production of Brassica crops in agricultural ecosystems, and a significant decrease in annual SO2 deposition in soils [81]. In Poland, the annual SO2 emissions to ambient air have decreased nearly five-fold (2.53 vs. 0.56 million Mg) in the past 27 years (1990-2017) [82]. In crops of the family Brassicaceae, S fertilizers not only increase yields, but also influence the quality of agricultural products, including oil, fat-free seed residues, and harvest residues [81,[83][84][85]. Sulfur fertilization is much more likely to affect total protein concentration than crude fat concentration in the seeds of Brassica oilseed crops [86] because S actively participates in the synthesis of major S-containing amino acids (cysteine and methionine) [87]. The aim of this study was to determine the effect of N and S fertilization on plant parameters (plant height, shoot diameter at the base, number of productive branches), yield (seed yield, yield components, straw yield, harvest index), and the processing suitability of crambe seeds as a potential feedstock for bio-based products in the oleochemical industry (content of crude fat, total protein, crude fiber, fatty acids, acid detergent fiber (ADF) and NDF) in northeastern Poland.
The experiment had a randomized complete block design (RCB) with three replications. Plot size was 15 m 2 (10 by 1.5 m). In each year of the study, the preceding crop was spring wheat (Triticum aestivum L.). The applied tillage treatments were skimming, winter plowing, and soil loosening before sowing. Crambe cv. Borowski was sown at the beginning of April (9-11 April) with a plot seeder at 200 pure live seeds m −2 , with a spacing of 19 cm, to a depth of 1.5-2.0 cm. Directly before sowing, phosphorus (enriched superphosphate, 40% P2O5) was applied at 40 kg ha −1 P2O5, and potassium (potassium sulfate, 50% K2O, and/or potash salt, 60% K2O) was applied at 100 kg ha −1 K2O. Weeds were controlled with metazachlor, which was applied at 1000 g ha −1 immediately after sowing. Pesticide (6 g ha −1 lambda-cyhalothrin) was applied at the beginning of inflorescence emergence (50 BBCH). Crambe was harvested at physiological maturity (89 BBCH) using a small-plot harvester .
The experiment was established on Haplic Luvisol originating from boulder clay [89]. The chemical properties of the soil are presented in the Table 1.

Plant Materials
The biomass yield of crambe (hulled seeds and straw) from each plot was determined by weight after threshing and expressed in DM per hectare, taking into account the moisture content of samples after oven-drying at 105 °C. The main morphometric parameters (plant height, shoot diameter at the base, number of productive branches) and yield components (plants m −2 and hulled seeds plant −1 ) were determined immediately before harvest (89 BBCH). Plants were counted (plant m −2 and morphometric parameters) along a 1-m-long section of each of the two middle rows [90]. The number of hulled seeds plant −1 was calculated as the ratio between hulled seed yield and hulled seed weight [24]. Thousand hulled seed weight (TSW) was determined after harvest and expressed on a DM basis, taking into account the moisture content of samples after oven-drying at 105 °C. The harvest index (HI) was calculated using Equation (1): Nitrogen fertilizer use efficiency (NFUE) was calculated with Equation (2) [61]: Hulled seed samples were scanned in the NIR Systems 6500 monochromator (FOSS NIR Systems Inc., USA) equipped with a reflectance module. Intact seeds (approximately 5 g) were placed in a standard ring cup and scanned. The content of total protein, crude fat, crude fiber, ADF, and NDF in crambe seeds, and the fatty acid composition of Abyssinian oil, were determined according to the procedures described by Jankowski et al. [68].

Statistical Analysis
Data were analyzed in the general linear mixed model in Statistica software [91] with N and S fertilizers as the fixed effects, and the growing season (Y) and replications nested within years as the random effects. Post hoc multiple comparisons were performed with the use of Tukey's test (HSD) in subsequent stages of statistical analyses. Data were regarded as statistically significant at α = 0.05. The results of the F-test for fixed effects in ANOVA are presented in Table 2.

Weather Condition
The growing seasons (April-August) during the field experiment differed in temperature and precipitation levels ( Figure 1). In the first growing season, the average daily temperature approximated the long-term average . In the second growing season, the average daily temperature exceeded the long-term average. In the third growing season, the average daily temperature was similar to the long-term average, excluding June when this parameter exceeded the long-term average. Total rainfall in March-August was 264 mm in 2017, 331 mm in 2018, and 371 mm in 2019. At the experimental site, the average long-term precipitation during the growing season over the last 37 years (in 1981-2015) reached 342 mm. The first growing season was characterized by low total precipitation due to rainfall deficiency in May, July, and August (59, 65, and 31% of the longterm average, respectively). In the second growing season, rainfall levels approximated the long-term average . In the third year of the study, the wet months of May and June (with precipitation levels 66% and 28% higher than the long-term average, respectively) contributed to above-average precipitation levels during the entire growing season (March-August) ( Figure 1).

Plant Height
Crambe plants produced shoots with a length of 65-103 cm, base diameter of 5-6 mm, and 7-8 productive branches (Table 3). The vegetative growth rate of C. abyssinica was highest in the third year of the study (Table 3) when precipitation levels exceeded the long-term average in May and June ( Figure 1). Nitrogen fertilization stimulated the vegetative growth of crambe. Crambe plants fertilized with N produced significantly taller (by 10%) and thicker shoots (by 20%) with more productive branches (by 26%) relative to the control treatment (without N fertilization). Nitrogen stimulated the vegetative growth of C. abyssinica up to the rate of 30 taller shoots, 90 thicker shoots, and 120 kg ha −1 higher number of productive branches. Nitrogen had no beneficial influence on shoot thickness except in the dry year of 2017, in particular during seed setting and filling (July and August) ( Figure 2). Sulfur fertilization did not induce significant differences in the vegetative growth rate of C. abyssinica (Table  3).

Yield Components
The density of C. abyssinica stands before harvest ranged from 129 (2017) to 162-166 (2018, 2019) plants m −2 . In NE Poland, crambe produced 94 do 131 seeds plant −1 with TSW (hulled seeds) of 8.0 to 8.7 g (Table 4). Yield components were most favorable in the third year of the study that was characterized by above-average precipitation in the growing season. Yield components were least expressed in the dry year of 2017 (first year of the experiment). Water-stressed plants were characterized by lower density (by 21%) and set significantly fewer seeds (by 28%) with lower TSW (by 8%) ( Table 4). Nitrogen fertilization increased plant density before harvest (by 6%) and the number of seeds plant −1 (by 25%) ( Table 4). Nitrogen exerted a positive effect on plant density and the number of seeds plant −1 , up to the rate of 30 and 90 kg ha −1 , respectively. An interaction was found between N fertilization and weather conditions for the number of seeds plant −1 and TSW (Table 2). In the growing season characterized by high yields (high precipitation in the third year of the experiment), an increase in N rate to 120 kg ha −1 led to a 1.7-fold increase in the number of seeds plant −1 accompanied by a 12% decrease in TSW (Figure 3). In the year with average precipitation (2018), the number of seeds plant −1 increased by 16% in response to the N rate of up to 60 kg N ha −1 . In the driest year of 2017, N fertilization had no significant effect on the number of seeds plant −1 (Figure 3). Thousand-seed weight increased by 6% in response to the S fertilization rate of up to 30 kg ha −1 ( Table 4), regardless of agroecological conditions or N rate ( Table 2). The remaining yield components (plants m −2 and seeds plant −1 ) were not significantly affected by S fertilization (Table 2).

Biomass Yield and Harvest Index
In NE Poland, the average seed yield of crambe ranged from 0.96 to 1.64-1.82 Mg ha −1 (hulled seeds). Nitrogen rates up to 120 kg ha −1 contributed to a significant increase in seed yield (by 0.45 Mg ha −1 ), mainly due to the beneficial influence of N on the number of plants m −2 and seeds plant −1 (Table  4). Nitrogen rates up to 120 kg ha −1 improved yields (Figure 4) in years with average and aboveaverage precipitation during the growing season of C. abyssinica (2018, 2019). In the year with belowaverage precipitation (2017), N rates of 30-120 kg ha −1 were not productive (Figure 4). Sulfur rates up to 15 kg ha −1 induced a significant increase in seed yield (by 0.12 Mg ha −1 , i.e., 9%) relative to the unfertilized control (Table 4), regardless of agroecological conditions or N rate ( Table 2). The NFUE of crambe ranged from 2.1 (2017) to 5.6-5.9 kg of hulled seeds kg −1 N (2018, 2019). An increase in N rate from 30 to 120 kg N ha −1 decreased NFUE by 28% on average ( Table 4), regardless of agroecological conditions ( Table 2). It should be noted that S fertilization increased the efficiency of N rates of ≥30 kg ha −1 by 29-39% (60 kg N ha −1 ), 23-31% (90 kg N ha −1 ), and 22% (120 kg N ha −1 ) ( Figure 5). In treatments with low N fertilization levels (30 kg ha −1 ), the application of S caused a 1.7to 2.1-fold decrease in NFUE ( Figure 5).  Table 4). This parameter increased significantly up to the N rate of 120 kg N ha −1 (33%) and the S rate of 30 S ha −1 (14%) ( Table 4), regardless of agroecological conditions ( Table 2).
The content of C. abyssinica seeds in total harvested biomass (hulled seeds + straw) ranged from 35 to 48% (Table 4). Nitrogen rates of ≥60 kg ha −1 had a negative impact on the ratio of seeds to straw (the harvest index decreased from 0.49-0.41 to 0.31-37) in years with average and above-average precipitation (2018 and 2019, respectively) ( Figure 1). In the year with below-average precipitation (2017), N fertilization had no significant effect on the value of HI (Figure 4). The content of seeds in the total harvested biomass of C. abyssinica was not significantly affected by S fertilization levels ( Table 2).

Quality of Seeds and Oil
Hulled crambe seeds contained 324-394 g kg −1 DM of crude fat and 208-238 g kg −1 DM of total protein. Crude fat and total protein content were highest in the year with average precipitation (2018). In years with below-average (2017) or above-average (2019) precipitation, crude fat and total protein content was 15-18% and 10-12% lower, respectively. Nitrogen fertilization decreased the crude fat content of C. abyssinica seeds by 6% and increased total protein content by 11% (Table 5). Nitrogen rates up 60 kg ha −1 decreased the synthesis of crude fat in seeds ( Table 5), regardless of agroecological conditions ( Table 2). On average, the total protein content of seeds increased up to the N rate of 90 kg N ha −1 throughout the entire study (Table 5). However, total protein content was strongly influenced by the interaction between N rate and weather conditions. In the dry year of 2017, the total protein content of crambe seeds increased up to the N rate of 120 kg ha −1 . In years with average (2018) and above-average precipitation (2019), total protein content increased up to the N rates of 90 and 30 kg ha −1 , respectively ( Figure 6). Sulfur rates up to 30 kg ha −1 contributed to a 4-5% increase in the crude fat content of crambe seeds (Table 5), excluding in the year with above-average precipitation (2019) when the greatest increase in crude fat content was observed already in response to the S rate of 15 kg ha −1 (Figure 7). Sulfur fertilization had no significant effect on the total protein content of seeds, regardless of N rate or weather conditions ( Table 2). The crude fiber content of crambe seeds ranged from 118 to 137 g kg −1 DM, and the proportions of ADF and NDF were determined at 33% and 43-45%, respectively ( Table 5). The crude fiber content of seeds was significantly lower (by 9-14%) in the year with above-average precipitation (2019) in the treatment without N fertilization (118 g kg −1 DM). Nitrogen rates up to 90 kg ha −1 increased the content of crude fiber and NDF by 14% and 4%, respectively. Sulfur fertilization had no significant effect on the crude fiber content of seeds or the proportions of NDF and ADF (Table 2).  Erucic acid was the predominant component of Abyssinian oil, and it accounted for 57-65% of total fatty acids ( Table 6). Abyssinian oil also contained considerable amounts of oleic acid (15-19%) and linoleic acid (9-11%) ( Table 6). Abyssinian oil was most abundant in MUFAs which accounted for 79-82% of all fatty acids. Nitrogen fertilization increased the content of linoleic acid (from 9.2% to 9.8%) and decreased the content of linolenic acid (from 7.2% to 6.6%) ( Table 6). The proportions of SFAs, PUFAs, and MUFAs in Abyssinian oil were not significantly differentiated by N rate (Table 2). Sulfur fertilization significantly decreased the content of linoleic acid in Abyssinian oil (Table 6), but had no effect on the concentrations of the remaining FAs or the proportions of SFAs, MUFAs, or PUFAs (Table 2). Means with the same letters do not differ significantly at p ≤ 0.05 in Tukey`s test. The absence of superscript letters indicates that the main effects or interactions were not statistically significant. C16palmitic acid; C18-stearic acid; C18:1-oleic acid; C18:2-linoleic acid; C18:3-linolenic acid; C20:1-eicosanoic acid; C22:1-erucic acid; SFA-saturated fatty acids; MUFA-monounsaturated fatty acids; PUFA-polyunsaturated fatty acids.
The harvest index of C. abyssinica was determined at 0.29-0.37 by Fontana et al. [16] and Stolarski et al. [38], and it ranged from 0.39 to 0.48 in the work of Stolarski et al. [25] and in the present study (Table 4).
In the production of C. abyssinica, all agronomic requirements for the species, as well as adequate soil management and plant nutrition requirements, have to be met to achieve high and stable seed yields [54]. Detailed fertilization recommendations for the species discussed are not available in the literature, and the effect of fertilization on the development of crambe plants and seed yields is ambiguous in the existing research [25,26,93].
Crambe has similar soil fertility requirements to mustards and rapeseed [54]. According to Endres and Schatz [52], around 50 kg of N is required to produce 1 Mg of seeds and the corresponding amount of straw. In Brazil, crambe seed yields peaked in response to N rates of 40 [94] to 60-90 kg ha −1 [27,93]. In a study conducted by Cihacek and Gonzales [95] in North Dakota (northern USA), seed yield increased up to the N rate of 100 kg ha −1 . In the current study, seed yields continued to increase up to the N rate of 120 kg ha −1 , mainly due to an increase in the number of plants m −2 and seeds plant −1 . In another study conducted in NE Poland, crambe seed yields increased up to the N rate of 80 kg ha −1 due to an increase in the number of seeds plant −1 and TSW [96]. Stolarski et al. [25] reported high and stable crambe yields (2.1 Mg ha −1 ) in NE Poland in treatments without N fertilization where winter wheat was the preceding crop. In the present study, N fertilization did not increase the seed yield of C. abyssinica in the year with low precipitation.
In this study, the NFUE of crambe ranged from 2.1 to 5.9 kg seeds kg −1 N, subject to precipitation levels. As expected, C. abyssinica yields decreased with a rise in N rate (from 5.3 to 3.8 kg hulled seeds kg −1 N). Similar trends were reported in studies of C. sativa [63,68] and other Brassica crops (B. juncea, B. napus and B. rapa canolas, and B. juncea and S. alba mustards) [61].
Oilseed crops of the family Brassicaceae have high S requirements [81]. In soils with moderate levels of S (10-35 mg SO kg −1 soil [97], seed yields peaked in response to 20-30 kg S ha −1 in B. juncea (traditional and canola type cultivars), C. sativa, and C. abyssinica [68,81,84,85,98]; 40 kg S ha −1 in B. napus (spring cultivars) and S. alba; and 40 to 80 kg S ha −1 in B. napus (winter cultivars) [81,83,99]. The current study was conducted on Haplic Luvisol originating from boulder clay with moderate SO levels, and crambe yields increased up to the S rate of 15 kg ha −1 (5%), mainly due to the beneficial influence of S on TSW and, to a lesser degree, on the number of seeds plant −1 . Szczebiot [96] also reported a 5% increase in crambe seed yields in response to the S rate of 25 kg ha −1 in NE Poland. Sulfur fertilization plays a key role in the production of Brassica crops because it directly affects seed yield and yield components, and indirectly increases the efficiency of N fertilization [68,100,101]. In the present study, S fertilization increased NFUE when N was applied at rates higher than 30 kg N ha −1 , by 29-39% (60 kg N ha −1 ), 23-31% (90 kg N ha −1 ), and 22% (120 kg N ha −1 ). When N was applied at 30 kg ha −1 , S fertilization caused a 1.7-to 2.1-fold decrease in NFUE. Similar trends were reported in other studies, where NFUE increased in response to S fertilization, particularly at higher N rates, in the production of C. sativa [68,[102][103][104], B. juncea and B. rapa [86], and S. alba [105].
Hulled seeds of C. abyssinica cultivars grown in European Russia (Middle Volga Region) and Central Europe (Poland) accumulate 189-218 g kg −1 DM total protein [36,37]. In the Mediterranean Region, the total protein content of hulled C. abyssinica seeds was higher by 50-70 g kg −1 DM [33]. In the current study, the total protein content of C. abyssinica seeds ranged from 208 to 238 g kg −1 DM and was typical for colder regions with a humid continental climate.
Regardless of genetic factors that affect nutrient synthesis in the seeds of Brassica crops, nutrient levels can be modified by climatic and soil conditions, and agronomic management [81]. In the group of agronomic factors, N [68,72,102,103,105,109,110] and S fertilization [68,81,111,112] play a key role in the biosynthesis of major nutrients and bioactive components in the seeds of oilseed crops of the family Brassicaceae.
Nitrogen influences the synthesis of reserve compounds that determine the protein and crude fat content of oilseed crops [54,113]. In most Brassica crops, N fertilization decreases the crude fat content and increases the total protein content of seeds. This trend was reported in B. napus [72] and other Brassica crops, including C. sativa [68,102,103,110], B. juncea [109], and S. alba [105]. Similar observations were made in this study, where N fertilization affected the qualitative and quantitative composition of fatty acids in C. abyssinica seeds. High rates of N fertilizer increased the content of linoleic acid and decreased the content of linolenic acid in Abyssinian oil. In a study by Chaves et al. [93], the protein and crude fat content of crambe seeds was not modified by increasing N rates.
Sulfur exerts varied effects on the crude fat and total protein content of Brassica seeds. In B. napus [114,115], C. sativa [68,102,103,116], and other oilseed crops of the family Brassicaceae (B. napus, B. juncea, S. alba), the relationship between S fertilization and the crude fat and total protein content of seeds appears to be less obvious [81]. In the current study, S fertilization exerted a minor impact on the total protein and crude fiber content of C. abyssinica seeds. The crude fat content of crambe seeds increased with a rise in S rate, and the strength of this association was determined by weather conditions. In the year with above-average precipitation, oil concentration in crambe seeds was highest after the application 15 kg S ha −1 . In the year with below-average precipitation, the crude fat content of seeds increased in response to the S rate of 30 kg ha −1 . Sulfur fertilization caused a significant decrease in the content of linoleic acid in Abyssinian oil, but did not affect the proportions of the remaining fatty acids or the total content of SFAs, MUFAs, and PUFAs.

Conclusions
Nitrogen fertilization increased plant height, shoot diameter at the base, the number of productive branches, the number of plants m −2 , and seeds plant −1 . Sulfur fertilization improved TSW. The growth-promoting effects of N and S fertilization on C. abyssinica were more pronounced in the year with above-average precipitation, from inflorescence emergence to fruit development. In NE Poland, the average seed yields of C. abyssinica ranged from 0.96 Mg ha −1 in the dry year, to 1.64 Mg ha −1 in the year with average precipitation, and 1.82 Mg ha −1 in the year with above-average precipitation. The response of crambe plants to N fertilization was determined by precipitation. The N rate of 120 kg ha −1 delivered satisfactory results in years with average and above-average precipitation, whereas N fertilization was not effective in the dry year of the experiment. Nitrogen rates of ≥60 kg ha −1 decreased the HI in years with average and above-average precipitation. Nitrogen fertilization decreased the crude fat content and increased the total protein and crude fiber content of C. abyssinica seeds. Crambe seed yields continued to increase up to the S rate of 15 kg ha −1 , regardless of agroecological conditions in the experimental years. Sulfur fertilization was particularly effective when N was applied at >30 kg ha −1 (22-39% increase in NFUE). Sulfur increased the crude fat content, but did not affect the concentrations of the remaining nutrients in crambe seeds. Abyssinian oil was most abundant in erucic acid (57-65%), and it also contained considerable amounts of oleic acid (15-19%) and linoleic acid (9-11%). Nitrogen fertilization increased the content of linoleic acid and decreased the content of linolenic acid, whereas S fertilization significantly decreased the content of linoleic acid in Abyssinian oil.