S.J. Marcroft1, T.D. Potter4,   P.A. Salisbury1,2, W.A. Burton1

and D.J. Ballinger3

1 Agriculture Victoria, Victorian Institute for Dryland Agriculture, Private Bag 260, Horsham Victoria 3401 Australia

2Institute of Land and Food Resources, The University of Melbourne Parkville Victoria 3052, Australia

3DOVURO Pty. Ltd., RMB 3775, Horsham Victoria 3401 Australia

4South Australian Research and Development Institute, PO Box 618 Naracoorte SA 5271 Australia

Email: steve.marcroft@nre.vic.gov.au






Recommendations from the canola industry have been to use only certified seed when planting canola.  This advice has been well adopted and has resulted in 85-90% of growers using certified seed.  However, some farmers are still retaining and using their own seed in order to reduce the costs incurred when sowing the crop.  Major concerns with the use of farmer-retained seed include reduced seed viability, reduced vigour and genetic drift.  In Australia, canola crops ripen in the spring, which can be hot and dry, affecting the quality of the seed produced.  Results from 1998 trials showed that the use of farmer retained seed can result in poor seed viability and establishment failure in subsequent crops.  The average yield reduction from farmer retained seed was 12.4%, ranging between 0 and 100%.  Outcrossing in canola can result in slight genetic change from year to year and considerable change over a number of years.  Experiments have shown that over time farmer-retained seed can have reduced oil quality and agronomic performance.  Farmer-retained seed should be germination and vigour tested under less than ideal conditions to give an indication of its potential field performance.  It is strongly recommended that farmers use certified seed (which has been tested) in order to reduce the risk of establishment failure and to be sure that they are growing true to type canola.



KEYWORDSseed viability, certified seed, emergence, seedling vigour.





The Australian canola industry has experienced a rapid increase in production during the past five years, with much of the expansion occurring in short season environments.  In these areas, lower yield and therefore, lower profits are expected and there is more pressure on farmers to reduce costs. In some cases this is achieved by retaining seed rather than purchasing new seed each year. The quality of this farmer-retained seed is often unknown and may compromise the chances of the grower getting the best possible return from the following canola crop.  There are two concerns with farmer retained seed; potential genetic drift and reduced seed viability (reduced emergence, vigour and subsequent yield).


Canola is a crop which both self pollinates and outcrosses.  As a result of outcrossing of individual plants with different characteristics, the characteristics of a variety can drift slightly from one generation to the next. In most instances, the characteristics for which the variety was originally selected tend to regress, while the undesirable characteristics tend to become more prominent.  In a previous study, Kudnig (1998) found subsequent generations from basic seed had longer season length, greater height and less blackleg resistance.


The small seed size of canola and its subsequent relatively poor vigour compared with most other crops often results in canola plant populations varying widely even when sown at the same rate (Canola Check monitoring data).  Because of their size canola seedlings are particularly vulnerable to fungal infection, insect attack, fertiliser toxicity and other physical factors.  The potential for these factors to cause establishment problems is considerable. Salisbury et al. (1995) found that germination under optimal conditions did not reflect germination and emergence in field conditions.  As conditions for emergence become less than ideal, seed of lower quality will fall in germination percentage faster than seed of good quality.  The ability of the seed to produce a vigorous, healthy seedling, therefore is of crucial importance when attempting to establish a crop in most sowing conditions experienced in Australia.  If seed of inferior quality produces lower plant numbers and less vigorous crops, the potential yield loss is considerable.  The aim of this experiment was to compare crop performance of certified seed and farmer-retained seed.





The experiment was conducted during 1998 at Kybybolite in South Australia (sandy loam over clay, 1998 April-October rainfall was 400 mm)  and Horsham in Victoria (grey self mulching clay; 1998 April–October rainfall was 293 mm).  Eighteen seed samples of the same canola cultivar Karoo produced in different paddocks, were chosen from both certified and farmer retained seed. These samples were classified as good (96-100%), mid (91-95%) or poor (>85%) germination, on the basis of laboratory germination results (accelerated aging technique).  Certified seed is rejected if it has less than 85% germination, the certified seed tested therefore ranged from 85–100% germination. The farmer retained seed ranged from 37-100% germination.  The farmer-retained seed was chosen from seed which farmers had voluntarily sent to the Primary Industries and Resources South Australia seed testing service laboratory for assessment prior to sowing.  Both the certified seed and the farmer retained-seed were therefore seed which was intended for sowing.   

During the season all seed lots from both sites were assessed for germination, seedling vigour, biomass production, yield and oil quality.  The trial at Horsham was sown at the recommended depth of 2 cm and then at 5 cm in order to stress the seed and accentuate any differences resulting from seed quality, Kybybolite was sown only at 3 cm.  Grain yield was determined following machine harvest and oil content was determined by NIR.





Effect on emergence: The emergence was considerably higher at Kybybolite compared with Horsham (Figure 1), which can be attributed to differences in sowing time, soil type and sowing equipment.  Figure 1 illustrates that plant emergence was very consistent for all of the certified seed across both sites; there were no statistically significant differences between any of the certified seed lines. However, while farmer retained seed lines classified as good or mid were equal to the certified seed, seed classified as poor showed a significant lower plant population at both sites, producing approximately half the plant numbers of the certified seed.


Effect on emergence from depth:   In most cases, emergence from depth (5 cm) was lower than emergence from shallow (2 cm) sown seed.  There were significant differences between individual seed lines.  However, these differences did not relate to the laboratory germination classification (good, mid or poor) for either certified or farmer retained seed.

Effect on plant vigour:   Vigour scores correlated strongly across both sites (Figure 2.), all the treatments had relatively good vigour except for the farmer-retained poor lines which were significantly lower in vigour at both sites.  Lower vigour has many ramifications for crops attempting to establish under harsh conditions.


Effect on biomass:   Biomass data was only recorded at Horsham, with the results mirroring the emergence data; the only treatment to be significantly different from the certified seed was the farmer-retained poor seed (Figure 3).  The individual plants in plots established from the farmer-retained seed classified as having poor germination, were, however, larger than in other plots, due to the lack of inter-plant competition because fewer plants established in this treatment. 


Effect on yieldAt Kybybolite where conditions were more favourable, there were fewer significant differences between seed lines; poor emergence, vigour and biomass did not have a large effect on grain yield.  There was no significant difference (p<0.05) between the yield of certified seed lines and only one farmer-retained poor line resulted in a significant (p<0.05) reduction in yield.


The Horsham site, which had 107 mm less growing season rainfall, showed greater differences between treatments.  The farmer-retained seed had the largest effect with yield ranging from 98 to 1171 kg/ha, this yield difference equates to a cost to the farmer of $322/ha assuming a price of $300/tonne.  The lowest yielding farmer retained seed line had 3.7% site mean emergence, 9.8% site mean vigour and 0.9% site mean biomass. These results are consistent with Canadian research where certified seed gave a 7% yield advantage over farmer-retained seed (Harrison 1998).


Canola being a indeterminate crop is capable of compensating for poor establishment by extra branching during flowering.  The yield results from Table 1 illustrate that given reasonable growing conditions even seed of very poor quality can recover to give an acceptable yield.  However, if conditions for plant growth are not favourable, poor quality seed may result in reduced yield.


Effect on quality:   Oil content, protein, glucosinolates and erucic acid was measured on all seed lines before sowing and then after harvest.  There were no consistent differences between any of the seed lines.  Differences in emergence, vigour, biomass and yield had no effect on the final oil quality of the harvested seed.  This is consistent with the finding that environmental conditions during the pod fill stage are responsible for oil quality (Pritchard 1998).


The data illustrates that seed source can have a major effect on subsequent canola crops.  Poor quality seed can result in poor establishment and may cause yield loss depending on seasonal conditions.  Good quality seed results in very consistent establishment.  All of the certified and some of the farmer-retained seed lines tested were of good quality, resulting in acceptable plant establishment.  However, the results did show that retaining seed increases the risks associated with the crop and may in some cases lead to crop failure. 

Figure 1.   Emergence of Certified and Farmer Retained Seed for Horsham and Kybybolite (Plants per 0.25m quadrat)

                                LSD (p=0.05) = Horsham – 7.0, Kybybolite – 8.6,

CV% = Horsham – 43.1%, Kybybolite –19.0%


Figure 2.   Vigour of Certified and Farmer Retained Seed for Horsham and Kybybolite (Visual score 1 – 5)

LSD (p=0.05) = Horsham – 0.46, Kybybolite – 0.46,

CV% = Horsham – 22.3%, Kybybolite –23.9%

Figure 3.   Biomass Production of Certified and Farmer Retained Seed from Horsham

 (Grams per 0.25m quadrat)

LSD (p=0.05)  = Horsham – 9.7

CV% = Horsham – 23.5%


Table 1.   Grain Yield of Certified and Farmer Retained Seed Lines at Horsham and Kybybolite – kg/ha





Yield kg/ha


Yield kg/ha






Good 1





Good 2





Good 3





Mid 1





Mid 2





Mid 3





Poor 1





Poor 2





Poor 3





* Missing data

(results for Horsham are a combination of both sowing depths)

LSD (p=0.05)  = Horsham - 279kg, Kybybolite - 354kg,

CV% = Horsham -29.2%, Kybybolite -15.3%



Seed produced under varying growing conditions can have major effects on subsequent crops.  Farmers who retain seed should have the seed germination tested in a laboratory prior to sowing to avert the considerable risk of establishment problems and potential yield loss. These trials illustrate that laboratory screening of seed is able to identify lines of poor quality allowing farmers to determine the quality of their seed prior to sowing.  When considering retaining seed, farmers should be aware of the costs associated with germination testing, grading, storage, and potential weed problems.  Farmers should never retain seed from a crop which was planted from retained seed due to potential changes in varietal characteristics  (genetic drift).




We thank Jack Kay, Ian Ludwig, David Robson and Leisa White for there assistance with trial management and data collection and Audrey Leong for quality analysis. Initial seed testing was performed by Kevin Boyce and Heather Laurie of Seed Services, PIRSA. Certified seed samples were provided by DOVURO Pty Ltd. Patrick Lim provided statistical support.




Harrison, B. (1998) Bin-run canola seed costs you money. Canola Guide, April 1998, Winnipeg, Manitoba pp. 20-21.


Kudnig, J. (1998) Farmer retained canola seed shows decreased yields and lower potential financial return. DOVURO agronomic update, pp1-2.


Pritchard, F.M. (1998) The effect of growing location and time of sowing on the production of premium quality oilseeds in south-eastern Australia. Masters Thesis Melbourne University, Chapter 1.4 Environmental effects on oil and protein content pp 14-16.


Salisbury, P.A., Ballinger, D.J., Robson, D.J. and Murray, L. (1995) Effect of seed sources on sowing quality and yield of canola. Proceedings 9th GCIRC International Rapeseed Congress, July 1995, Cambridge, pp. 247-249.