F.M. PritchardB, R.M. NortonA, H. A. EaglesAB, P.A. SalisburyCB and M. NicolasC

ABSTRACT

The effects of region, year and weather on the quality of canola grown across the Australian state of Victoria were determined from data of advanced breeding experiments between 1985 and 1994. Residual maximum likelihood was used to estimate means from these unbalanced data. Environmental effects were large relative to genotypic effects for oil and seed protein content, while the reverse occurred for glucosinolate concentration. High oil contents (and low seed protein contents) were correlated with cooler temperatures and higher rainfall in spring, and were highest in canola from the cooler and wetter south-west and north-east regions of Victoria, and lowest in canola from the hotter regions, such as the Mallee, or in drought years. On average, oil content fell by 0.38% per 1.0^oC increase in spring maximum temperatures.

Fatty acid composition varied with year and region. Oleic acid content averaged 60.3±0.4% and was consistently higher in canola grown in central Victoria and the Wimmera, and in most years, in north-east Victoria. Oleic acid content declined with low temperatures and low rainfall. Linoleic (19.7± 0.3%) and linolenic (10.4±0.3%) acids showed low variability across regions, although higher maximum temperatures and lower rainfall in spring reduced the linolenic acid content of canola.

KEYWORDS Fatty acid, oleic, linoleic, linolenic, protein, oil, glucosinolate, temperature, rainfall.

INTRODUCTION

In Victoria, canola production is based mainly in the Wimmera and central Regions, but is expanding into the cooler and wetter south and south-west, and into the hotter and drier Mallee. These areas range from 300 to 500 mm average annual rainfall and average daily maximum temperatures in spring (when flowering occurs) range from 23oC in the Mallee to 19oC in the south-west. Under glasshouse conditions, high temperatures during flowering (consistently >16 oC) have reduced oil content and increased oleic acid from 23% to 40% in high erucic acid rapeseed (Canvin 1965). Water deficit during seed-filling has reduced oil content in the field (Mailer and Pratley, 1990) and also reduced oleic acid by up to 7% in the B. napus cultivar ‘Cérès’ in a pot experiment (Champolivier and Merrien, 1996). However, relatively little is known of the effects of changes in temperature and water deficit on the fatty acid composition of Brassica oilseeds under field conditions.

This paper describes the seed oil, protein and glucosinolate content and the fatty acid composition of oil from canola grown across the production regions of Victoria. It also identifies the effect of climate on those quality components.

MATERIALS AND METHODS

These data were derived from advanced breeding experiments by Agriculture Victoria, which included 10 - 40 lines of canola located at 4 - 10 sites across Victoria each year from 1985 to 1994. All sites were rainfed, except the Kerang site, which was irrigated during flowering. Sites were grouped into regions (Figure 1).

Figure 1. Victorian rainfall zones, regions and experiment site locations.

Seed samples were analysed for oil, seed protein and glucosinolate content and fatty acid composition. Fatty acids were measured by gas chromatography. Total glucosinolate concentration was measured by a modification of the glucose method. Oil, protein and moisture content were measured by NIRS from 1991. Prior to 1991, oil content was measured by the Sohxlet Extraction method and seed protein content estimated from the Kjeldahl method. Seed protein content was measured from 1989 to 1994. Oil content is expressed at 8.5% moisture.

Generally, the same breeding lines/cultivars were grown at each site for each year, but none were grown in all years. To allow for this, regional means and variance components were estimated using residual maximum likelihood (REML) (Payne et al., 1993) with cultivar and years as the random model, and regions as fixed. To estimate year means, region and years were fixed and the cultivars were again considered as random.

Weather data were collected from Monthly Weather Reviews (Bureau of Meteorology) for the recording station nearest the experimental site. Weather variables were regressed against quality components by linear or multiple regression, as appropriate. The dependent variables were the estimated site means of the quality component derived from REML.

RESULTS AND DISCUSSION
Seed oil, protein and glucosinolate levels

The regional values of oil, seed protein and glucosinolate content are shown in Table 1. Oil and seed protein content had a strong inverse relationship (r=-0.75, p<0.001).Table 1. Regional mean oil, protein and glucosinolate content and major fatty acids (and standard errors) of Victorian canola 1985 -1994.

Region	Oil	Seed protein	Glucosinolates	Oleic acid	Linoleic acid	Linolenic acid
	(%)	(%)	(μmol/g)	(%, C18:1)	(%, C18:2)	(%, C18:3)
Mallee	41.3±0.6 (19)*	20.5±0.1 (11)	15.5±1.6 (19)	60.6±0.4	19.6±0.3	10.0±1.0 (17)
Wimmera	41.3±0.6 (18)	19.8±0.1 (13)	14.7±1.6 (18)	60.7±0.4	19.3±0.3	10.2±1.2 (16)
North-east	42.5±0.6 (8)	18.5±0.2 (6)	14.4±1.6 (8)	60.6±0.4	19.6±0.3	10.0±1.3 (7)
South-west	43.3±0.6 (14)	18.3±0.2 (7)	14.5±1.6 (14)	60.2±0.4	19.5±0.3	10.7±1.2 (12)
Southern	40.0±0.6 (4)	21.3±0.2 (3)	13.3±1.7 (4)	58.5±0.4	20.9±0.3	11.6±1.3 (3)
Central	41.7±0.6 (17)	19.8±0.1 (11)	16.0±1.6 (17)	61.3±0.4	19.2±0.3	9.8±1.3 (15)
Irrigated	39.4±0.6 (6)	22.4±0.2 (3)	16.8±1.7 (6)	59.9±0.4	20.2±0.3	10.7±0.8 (4)
Mean and se	41.4±0.6 (86)	20.1±0.1 (54)	15.6±0.6 (86)	60.3±0.4	19.7±0.3	10.4±0.3 (74)

*The number of site years are in brackets.

Oil contents were generally higher in canola grown in the wetter and cooler areas of the state, such as south-western (SW) Victoria, while seed protein was lower in those areas. For example, the mean oil content was 43.3% from the SW and 41.3% from the Mallee (Table 1). Importers of Australian canola seed buy on a standard of 42% oil content, and have the option to reject seed below this level.

Year had a greater effect on oil content than region. Assuming a random model, the variance component for years was 3.51+1.73, compared with only 1.02+0.61 for regions. Generally, there were state-wide “high” and “low” oil years, with the range from 37.9% to 43.3%. Oil content varied from 39.4% to 43.3% among regions (Table 1).

Seed protein content increased with average maximum spring temperatures (r=0.40, p<0.01), but glucosinolate concentration did not correlate with any weather variable. Oil content was most highly correlated with cool temperatures in spring, when seed maturation occurs. The regression coefficients for various temperature configurations ranged from –0.29 (p<0.05) to 0.42 (p<0.001) and would suggest a 3.5% range in oil content in the Wimmera over the 10 years. Oil content was (weakly) positively correlated with spring rainfall (r=0.24, p<0.04). The range of spring rainfall over the 10 years from 43 mm to 305 mm (Pritchard, 1998) would suggest a difference of 2.6% oil content. Increased water availability has generally been associated with higher oil contents (Mailer and Cornish, 1987; Mailer and Pratley, 1990).Environmental variation of oil and seed protein content greatly exceeded cultivar variation (Table 2). However, the opposite was true for glucosinolate concentration - considered to be due to the successful breeding effort to reduce glucosinolate levels over the study period.

Table 2. Estimates of variance components (σ²) of environmental (e) and cultivar (g) factors for canola quality components.

Quality component	σ2g	σ2e
Oleic acid (C18:1)	2.98±0.57*	2.92±0.62*
Linoleic acid (C18:2)	0.76±0.11*	1.13±0.20*
Linolenic acid (C18:3)	0.23±0.04*	0.82±0.14*
Oil %	0.43±0.14*	7.66±1.34*
Seed protein %	0.18±0.03*	7.34±1.42*
Glucosinolates (μmol/g)	46.9±6.33***	18.1±3.41***

Symbols in second and third columns indicate significance of variance component: * p<0.05; ***p<0.001.

Fatty acid composition

Regional values of the composition of major fatty acids of canola shown in Table 1. Saturated fatty acid content averaged 6.4±0.1%. Erucic acid levels were very low in all years and regions, with means ranging from 0.3 to 0.5%.

A level of 60% oleic acid could be proposed as a reasonable standard, and this was achieved in most regions of Victoria (Table 1). Oleic acid content of canola was generally highest in regions that had warm and wetter springs, such as the Wimmera and central Victoria. It was lowest in the cool, moist South (although there were only 4 site-years for that region). Although oleic acid content of canola was generally highest with warm and wet springs, the irrigated site at Kerang in northern Victoria was the exception (Table 1), possibly due to the soil salinity, which has been demonstrated in safflower (Irving et al., 1988).

Genotypic and environmental variance estimates for most of the fatty acids were similar, although there were small differences (Table 2). The importance of environment relative to genotype increased with the degree of unsaturation of the 18-carbon fatty acids, from a ratio of 0.98 for oleic, to 1.49 for linoleic, to 3.48 for linolenic, suggesting that the desaturation process is influenced by environmental factors.

Despite low levels of oleic acid in some dry years (data not presented), there was no consistent effect due to rainfall across the data set (p>0.05). Warmer average maximum temperatures in spring favoured higher oleic acid levels, although this relationship is weak (r=0.23, p<0.05) and the variability of the oleic acid levels was quite low. Linolenic acid content increased with rainfall (r=0.47, p<0.001) and declined with increased highest maximum temperatures in spring (r=-0.41, p<0.001).

Post-flowering temperatures are considered to have a major effect of the fatty acid composition of a range of oilseeds. The degree of unsaturation of oils generally increases as post-flowering temperatures decrease in a range of oilseeds (Canvin, 1965). Unlike temperature effects, the effect of water deficit on the fatty acid composition of low erucic acid B. napus has been received little attention until recently and research results have been inconsistent (Bouchereau et al., 1996; Champolivier and Merrien, 1996).

CONCLUSION

This paper shows that oil, seed protein and glucosinolate content and composition of the major fatty acids vary significantly with season and across the state of Victoria. Year and region were far more important factors than cultivar on oil and seed protein content. Oil content varied considerably more with year than region. The glucosinolate results reflected the genetic improvement in canola over the years of study. Oleic acid levels were consistently highest in canola from central Victoria, the Wimmera and to a lesser extent, north-east Victoria. Linoleic and linolenic acid levels were generally low in these regions. Environmental effects on the levels of individual fatty acids increased with an increasing degree of unsaturation.

ACKNOWLEDGMENTS
The authors would like to recognise the efforts of the staff of Agriculture Victoria who assisted with the management of these experiments, as well as the many landholders on whose properties these experiments were conducted. The authors also sincerely thank Dr Ray Flood for his constructive comments on the manuscript.

REFERENCESBouchereau, A., Clossais-Besnard, N., Bendaoud, A., Leport, L., and Renard, M. (1996)

2. Water stress effects on rapeseed quality. European Journal of Agronomy. 5, 19 – 30.

3. Canvin, D.T. (1965) The effect of temperature on the oil content and fatty acid composition

4. of the oils from several oil seed crops. Canadian Journal of Botany, 43, 63 - 69.

5. Champolivier, L., and Merrien, A. (1996). Effects of water stress applied at different growth

6. stages to Brassica napus L. var. oleifera on yield, yield components and seed quality. European Journal of. Agronomy 5 (3 - 4), 153 - 160.

Irving, D.W., Shannon, M.C., Breda, V.A., and Mackey, B.E. (1988). Salinity effects on yield

7. and oil quality of high-linoleate and high-oleate cultivars of safflower (Carthamus tinctorius L.). Journal of Agricultural and Food Chemistry 36, 37 - 42.

8. Mailer, R.J., and Cornish, P.S. (1987) Effects of water stress on glucosinolate and oil

9. concentrations in the seeds of rape (Brassica napus L.) and turnip rape (Brassica rapa L. var. silvestris [Lam.] Briggs). Australian Journal of Experimental Agriculture, 27, 707 – 711.

10. Mailer, R.J., and Pratley, J.E. (1990). Field studies of moisture availability effects on

11. glucosinolate and oil concentration in the seed of rape (Brassica napus L.) and turnip rape (B. rapa L. var. silvestris (Lam.) Briggs). Canadian Journal of Plant Science, 70 (2), 399 - 407.

Payne, R.W., Lane, P.W., Digby, P.G., Harding, S.A., Leech, P.K., Morgan, G.W., Todd,

12. A.D., Thompson R., Tunnicliffe Wilson, G., Welham, S.J., and White R.P. (1993) REML estimation of variance components and analysis of unbalanced designs. In ‘Genstat 5 Release 3 Reference Manual’. (Eds. R.W. Payne, P.W. Lane, P.G. Digby, S.A. Harding, P.K. Leech, G.W. Morgan, A.D. Todd, R. Thompson G. Tunnicliffe Wilson, S.J. Welham, and R.P. White.) pp. 539 - 584. (Oxford Science Publications, Oxford, UK.)

13. Pritchard, F.M. (1998). ‘The effect of growing location and time of sowing on the production of premium quality oilseeds in south-eastern Australia’. M. Appl. Sc. thesis, the University of Melbourne, 195 pp.